Nature

RESEARCH HIGHLIGHTS
Nature Reviews Molecular Cell Biology | AOP, published online 10 May 2012; doi:10.1038/nrm3357
P R OT E I N F O L D I N G
Sequestration at the IPOD stops division

Proteins that contain intrinsically disordered regions (IDRs) often form amyloids -sheet-rich insoluble fibrous aggregates and are toxic to cells when expressed at high levels. Amyloid formation is associated with many human diseases, but the causes underlying toxicity are unknown. Treusch and Linquist now show that the IDR-containing yeast prion Rnq1 is toxic to yeast cells because it sequesters a component of the spindl e body and causes cell cycle arrest. Yeast prions contain IDRs that can stably exist either in a soluble, unstructured state or in a selfperpetuating amyloid state. The authors studied Rnq1, which can affect the conformation of other IDR-containing proteins when in its amyloid state (denominated [RNQ+]). Moreover, Rnq1 is known to be extremely toxic when its gene is overexpressed in cells in which endogenous Rnq1 has adopted the amyloid [RNQ+] state. In a genome-wide screen to identify suppressors of Rnq1 toxicity, the authors isolated nine genes, three of which are loosely connected to the cell cycle. Furthermore, microarraybased gene expression analysis revealed that Rnq1 overexpression in [RNQ+] yeast strains induced the downregulation of several genes involved in cytokinesi s, which suggest s that Rnq1 may affect cell cycle progression. Indeed, Rnq1 over expression in an [RNQ+] background led to the accumulation of large budded cells that had undergone DNA replication but were arrested at mitosis. Interestingly, depletion of a crucial component of the spindle checkpoint (namely Mad2) in these cells led to further rounds of DNA synthesis without cytokinesis. Thus, Rnq1 over expression causes cell cycle arrest by activatin g the spindlecheckpoint. Using electron microscopy, the authors found that arrested cells had a monopolar spindle and not a bi polar spindle, as occurs normally. This defect was due to a failure to duplicate the spindle pole body (SPB; the yeast microtubuleorganizing centre), which normally duplicates and then divides to generat e a bipolar spindle. So how does Rnq1 affect SPB duplication? In arrested [RNQ+] cells, most SPB components are localized at the unduplicated SPB, except for Spc42, which is localized at both the SPB and the insoluble protein deposit (IPOD; a site where misfolded proteins accumulate in yeast) in the mother cell. The IPOD also contains Rnq1, which suggests that Rnq1 overexpression blocks cellcycle progression by sequestering the SPB component Spc42 at the IPOD. Importantly, the authors show that Spc42 was sequestered by Rnq1 when in its unstructured state rather than in its amyloidstate. Taken together, this work suggest s that the toxic effect of IDRcontaining proteins is due to specific protein interactions that disrupt key cellular functions; thus, toxicity could possibly be alleviated by inhibiting these interactions.
Kim Baumann
ORIGINAL RESEARCH PAPER Treusch, S. & Lindquist, S. An intrinsically disordered yeast prion arrests the cell cycle by sequestering a spindle pole body component. J. Cell Biol. 23 Apr 2012 (doi:10.10183/jcb.201108146)
Spc42 was sequestered by Rnq1 when in its unstructured state
NATURE REVIEWS | MOLECULAR CELL BIOLOGY 2012 Macmillan Publishers Limited. All rights reserved
VOLUME 13 | JUNE 2012
MACMILLAN SOUTH AFRICA
RESEARCH HIGHLIGHTS
ORGANELLE BIOGENESIS
When two become one

Peroxisome biogenesis is unconventional compared with other organelles, occurring by a two-step process. First, peroxisomal membrane proteins (PMPs) and lipids from endoplasmic reticulum (ER)-derived preperoxisomal vesicles drive the formation of an immature peroxisome and the peroxisomal translocon. Second, thetranslocon mediates the import of matrix proteins from the cytosol to produce a fully functioning peroxisome. But how the translocon itself assembles has been the source of some debate. Van der Zand and colleagues find that multiple trafficking routes are used to ensure that the full complement of translocon components only comes together in the peroxisomal membrane, perhaps ensuring additional control over where and when new peroxisomes arise. The translocon consists of two halves: a docking complex formed by the PMPs Pex13 and Pex14; and a RING finger complex composed of Pex2, Pex10 and Pex12. Van der Zand and colleagues set out to determine how these PMPs traffic from the ER to establish the initial preperoxisome and mediate the formation of the translocon. First, they labelled individual endogenous PMPs in live yeast cells and confirmed that the proteins were functional and formed the expected subcomplexes. Next, they examined different Pex mutants that disrupt peroxisome biogenesis at distinct steps to track when and where PMPs interact with each other. Cells carrying mutations that prevent either PMP exit from the ER or that allow only immature ER-derived preperoxisomal vesicles to form contained translocon subcomplexes but not the full translocon. Therefore, although subcomplexes could form in the ER and in preperoxisomal vesicles, something was preventing the formation of a full translocon on these membranes. This led the authors to ask whether PMP subcomplexes leave the ER in distinct membrane carriers, ensuring their separation until a new peroxisome is formed. Indeed, biochemical characterization of the membrane fractions from Pex mutants that accumulate immature ER-derived preperoxisomal vesicles showed that RING finger PMPs and docking PMPs concentrate in different fractions. Moreover, co-immunoprecipitation and colocalization analyses verified that docking PMPs and RING finger PMPs are kept apart in different membrane carriers. To show that this separation of the two half-translocons is not specific to Pex mutants, the authors tracked fluorescently labelled newly synthesized PMPs by pulse-chase analysis. They found that, whereas PMPs from a common subcomplex interacted at each trafficking step, PMPs from the two halves of the translocon did not coincide until a later time point. They thus conclude that docking and RING finger PMPs traffic from the ER in
Working out how PMP sorting at the ER and subsequent vesicle fusion occur will therefore be key...
distinct preperoxisomal membrane vesicles and come together only in the peroxisomal membrane. How do these vesicle precursors come together to form mature peroxisomes? It has previously been proposed that this might occur by heterotypic fusion of preperoxisomal vesicles, in a process that depends on the peroxisome proteins Pex1 and Pex6. By combining fluorescence pulse-chase analysis and a cell fusion assay, van der Zand and colleagues confirmed that only preperoxisomal vesicles underwent heterotypic fusion and that mature peroxisomes could not fuse. They also confirmed that fusion required Pex1 and Pex6, which localized to distinct vesicle fractions. Finally, the authors asked whether the subcomplexes that form in cells expressing Pex1 or Pex6 mutants have the potential to form full translocons and are thus functional intermediates. They showed that mating the two mutants allowed fusion of the preperoxisomal vesicles to resume and, ultimately, led to the import of the matrix protein PTS1 (peroxisomal targeting signal receptor 1) labelled with cyan fluorescent protein and the formation of a mature peroxisome. As this was not the result of preperoxi somal vesicle fusion with existing peroxisomes, the authors conclude that heterotypic fusion allows the formation of new translocons and, consequently, of active peroxisomes. Thus, by sorting the translocon subcomplexes into distinct vesicles and allowing translocon formation only after their heterotypic fusion, the ER provides an additional layer of control over peroxisome biogenesis. Working out how PMP sorting at the ER and subsequent vesicle fusion occur will therefore be key for understanding peroxisome biogenesis.
Alison Schuldt
ORIGINAL RESEARCH PAPER van der Zand, A. et al. Biochemically distinct vesicles from the endoplasmic reticulum fuse to form peroxisomes. Cell 149, 397409 (2012)
DigitalStock
RESEARCH HIGHLIGHTS
CELL SIGNALLING
Preventing WNT signalling

WNT ligands interact with Frizzled receptors and the WNT co-receptors low-density lipoprotein receptor-related protein 5 (LPR5) and LPR6 to induce -catenin-dependent canonical WNT signalling. In addition, WNT binding to Frizzled proteins alone initiates non-canonical WNT signalling. Here, Hao et al. describe a role for zinc and RING finger 3 (ZNRF3) in the inhibition of both canonical and non-canonical WNTsignalling pathways. ZNRF3 was identified in a screen for -catenin gene targets that are potentially involved in the negative regulation of WNT signalling. ZNRF3 encodes an E3 ubiquitin ligase with a transmembrane domain and a cytoplasmic RING finger domain. Overexpression of ZNRF3 suppressed -catenin-dependent transcription, whereas silencing of ZNRF3 or overexpression of a dominant-negative form of ZNRF3 that lacked the RING finger domain increased LRP6 and Frizzled levels at the membrane and enhanced -catenin activation. Notably, ZNRF3 was found to form a complex with Frizzled and LRP6 and to ubiquitylate Frizzled, thereby inducing its degradation. Thus, ZNRF3-mediated inhibition of WNT signalling involves the degradation of WNT receptors. But how are the membrane levels of ZNRF3 regulated to ensure tight control of WNT inhibition? Hao et al. show that the activator of both WNT pathways R-spondin, which has been shown to interact with the orphan transmembrane receptor LGR4, stabilizes WNT receptors by binding to the extracellular domain of ZNRF3. This binding enables the association of ZNRF3 with LGR4, which in turn downregulates ZNRF3 membrane levels by promoting ZNRF3 autoubiquitylation and degradation. Thus, the E3 ubiquitin ligase activity of ZNRF3 mediates ZNRF3 autoubiquitylation and turnover following R-spondin signalling.
this ZNRF3 function is evolutionarily conserved
Finally, in vivo studies in zebrafish and Xenopus laevis embryos confirmed the central role of ZNRF3 in the regulation of canonical and non-canonical WNT signalling pathways during embryonic development. Moreover, analysis of Znrf3-knockout mice indicated that this ZNRF3 function is evolutionarilyconserved. Given the central role of WNT signalling in embryonic development and cancer, the authors predict that ZNRF3 and its functional homologue RNF43, which is often mutated in patients with pancreatic cancer, will be valuable therapeutic targets. Maria Papatriantafyllou
ORIGINAL RESEARCH PAPER Hao, H. -X. et al. ZNRF3 promotes WNT receptor turnover in an R-spondin-sensitive manner. Nature 29 Apr 2012 (doi:10.1038/nature11019)
MACMILLAN SOUTH AFRICA
RESEARCH HIGHLIGHTS
SMALL RNAS
miRNAs strict schedule

MicroRNAs (miRNAs) have been implicated in practically all cell biological processes by virtue of their role in regulating gene expression through translational repression and mRNA decay. Two studies now examine the kinetics of this process and find that miRNAs first block translation of their mRNA target and then mediate its decay. To determine the kinetics of miRNA-mediated gene silencing, Bazzini et al. used zebrafish embryos, which express miR-430 4 hours post-fertilizatio n to clear maternal mRNAs. They found that the number of ribosomes bound to miR-430 targets significantly decreased 4 hours post-fertilization in wild-type embryos but not in mutants lacking Dicer (which is required for miRNA biogenesis). This finding was indicative of miRNA-mediated translational repression. By contrast, target mRNA levels did not change until 6 hours post-fertilization, which suggests that translational repression occurs before mRNA decay. Djuranovic et al. approached the same question using Drosophila melanogaster S2 cells transfected with constructs carrying 3 untranslated regions with miRNA target sites and a luciferase reporter gene fused to the 5 end. By examining the levels of luciferase (which are indicative of protein expression) and the level sof reporter mRNAs, the authors observed that translation repression occurred at least 2 hours before mRNA decay, which is consistent with the results of Bazzini et al. mRNA decay requires the removal of the stabilizing poly(A) tail, which is a process that has also been proposed to have a role in miRNA-mediate d translation repression. However, Bazzini et al. observed complete mRNA deadenylation only at 6 hours post-fertilization, after translation repression had begun. Moreover, a transcript in which de adenylation had been delayed showed comparable translational repression to a control transcript. Similarly, Djuranovicetal. did not observe any shortening of the poly(A) tail length in miRNA target transcripts during translation repression. Thus, inhibition of protein synthesis seems to occur before deadenylation. But which stage of translation do miRNAs specifically target? To answer this question, Bazzinietal. examined the distribution of ribosomes along the transcript. They observed that ribosomes were uniformly distributed but showed low occupancy on the repressed transcript, which is indicative of decreased translation initiation, rather than a block in
elongation or premature termination. Djuranovicetal. used a different approach, assaying the cleavage of mRNAs that occurs when ribosomes stall during elongation. They found that control mRNAs (those not targeted by the tested miRNAs) were cleaved and degraded, whereas target mRNAs were stable, which indicates that target mRNAs had not progressed toelongation. Together, these studies reveal the kinetics of miRNA gene silencing, corroborating previous in vitro results. As the findings were consistent in zebrafish and D. melanogaster, it is likely that the kinetics also apply to other systems.
Rachel David
ORIGINAL RESEARCH PAPERS Bazzini, A. A., Lee, M. T. & Giraldez, A. J. Ribosome profiling shows that miR-430 reduces translation before causing mRNA decay in zebrafish. Science 13, 233237 (2012) | Djuranovic, S., Nahvi, A. & Green, R. miRNA-mediated gene silencing by translational repression followed by mRNA deadenylation and decay. Science 13, 237240 (2012)
CORBIS
RESEARCH HIGHLIGHTS
BANANASTOCK
M E TA B O L I S M
WNT chews the fat with glucose uptake

Adipose tissue regulates whole body metabolism through the secretion of factors that act at both proximal and distal sites. A recent study published in Cell Metabolism identified a link between the activation of -catenin in adipose progenitor cells and glucose uptake in muscles. Canonical WNT signalling, which results in the stabilization of the transcription factor -catenin, regulates several developmental processes. WNT ligands have been previously shown to inhibit differentiation of adipose tissue. To identify the cell type in which WNT ligands act to mediate this inhibitory effect, Zeve et al. generated transgenic mice that express a constitutively active form of -catenin specifically in differentiated adipocytes (A-BCA mice) or in peroxisome proliferatoractivated receptor- (PPAR)-expressing adipose progenitor cells (P-BCA mice). Constitutive activation of -catenin in mature adipocytes had no effect on adipose tissue integrity and whole body metabolism. By contrast, -catenin activation in adipose progenitor cells resulted in adipose tissue degeneration (lipodystrophy), as defined by low fat content, almost undetectable adipokine (adiponektin and leptin) levels, near-complete loss of visceral fatdepots and unconventional morphology of subcutaneous fat depots. Notably, subcutaneous depots in P-BCA mice lacked adipocytes but contained collagen fibres and fibroblast-like cells. Moreover, fate-tracing experiments revealed that active -catenin promoted the differentiation of adipose progenitor cells into a fibroblastic cell lineage rather than an adipocytic cell lineage. But what are the metabolic effects of -catenin-driven fat tissue degeneration? As expected, loss of adipose tissue correlated with increased triglyceride levels in the blood. However, unlike adipose tissue-deficient lipodystrophic mice (such as mice lacking PPAR expression), P-BCA mice did not show any increase in blood glucose and insulin levels after feeding. Inaddition, P-BCA mice displayed low glucose levels (hypoglycaemia) in response to fasting. Thus,loss of fat as a consequence of constitutive activation of -catenin in adipose progenitor cells is associated with an unexpected increase in glucoseuptake. Next, injection of radioactively labelled 2-deoxyglucose revealed that skeletal and cardiac muscles were the main sites of glucose absorption in P-BCA mice. Interestingly, increased glucose uptake in these mice was not due to increased insulin levels or insulin sensitivity. Instead, the authors observed increased AMP-activated protein kinase (AMPK) activation and increased membrane localization of glucose transporters inmuscles of P-BCA mice, both before and after feeding. Furthermore, AMPK inhibition was sufficient to reverse hypoglycaemia in fasting mice as well as to increase blood glucose levels infed mice, suggesting that insulin-independent glucose uptake in muscles of P-BCAmice requiresAMPK activity. Finally, Zeve et al. investigated the link between constitutive activation of canonical WNTsignalling in adipose progenitor cells and AMPK-dependent glucose uptake in muscle. Asthey failed to detect cells that differentiated from PPAR-expressing progenitors and WNT-induced gene expression in muscle, they hypothesized that increased muscle glucose uptake in P-BCA mice was induced by a soluble muscle-extrinsic factor. Indeed, incubation of isolated wild-type muscle in serum from P-BCA mice or in supernatant from -catenin-expressing adipose progenitor cell cultures was sufficient to increase AMPK activation and glucose uptake. Thus, activation of canonical WNT signalling in adipose progenitor cells and the resulting adipose tissue degeneration seem to regulate distal glucose uptake in muscles through an undefined fatmuscle endocrine axis. Further elucidation of the components of this pathway might provide new insights into the regulation of glucose homeostasis.
ORIGINAL RESEARCH PAPER Zeve, D. et al. WNT signaling activation in adipose progenitors promotes insulin-independent muscle glucose uptake. Cell Metab. 15, 492504 (2012)
Maria Papatriantafyllou
RESEARCH HIGHLIGHTS
IN BRIEF
G E N O M E I N S TA B I L I T Y
An autophagy-independent role for UVRAG

Ultraviolet irradiation resistance-associated gene (UVRAG) is a tumour suppressor gene known to activate autophagy. Here, Zhao et al. identify an autophagy-independent role for UVRAG in chromosomal stability. First,UVRAG was shown to interact with the DNA-dependent protein kinase (DNAPK) complex, which has a central role in the non-homologous end joining (NHEJ) pathway of DNA repair. Interestingly, this interaction was found to promote double-strand break repair in an autophagy-independent manner. Second, the authors observed that UVRAG is targeted to the centrosome by centrosome protein 63. Association of UVRAG with the centrosome was required for centrosome stability and correct chromosome segregation. Thus, the authors propose that the ternary role of UVRAG in autophagy, DNA repair and centrosome stability may explain why it is frequently mutated in cancers.
ORIGINAL RESEARCH PAPER Zhao, Z. et al. A dual role for UVRAG in maintaining chromosomal stability independent of autophagy. Dev. Cell 26 Apr 2012 (doi:10.1016/j. devcel.2011.12.027)
STEM CELLS
Ready to die fast

Embryonic stem (ES) cells the integrity of which is crucial for embryonic development are highly sensitive to DNA damage, but the mechanisms underlying rapid death are unclear. Dumitru et al. now show that BAX, a pro-apoptotic member of the B cell lymphoma 2 (BCL-2) family, is maintained in its active form at the trans-Golgi network (TGN) of healthy ES cells and is transferred to mitochondria to induce apoptosis following DNA damage. By contrast, in healthy non-ES cells BAX is only found in the cytosol in an inactive conformation and needs to be both activated and transferred to mitochondria in response to apoptotic stimuli. The authors also show that apoptosis after DNA damage in ES cells depends on p53, which is required for the TGN-to-mitochondria translocation of active BAX. Thus, active BAX at the TGN keeps undifferentiated EScells primed for rapid death.
ORIGINAL RESEARCH PAPER Dumitru, R. et al. Human embryonic stem cells have constitutively active Bax at the Golgi and are primed to undergo rapid apoptosis. Mol. Cell 3 May 2012 (doi:10.1016/j.molcel.2012.04.002)
RESEARCH HIGHLIGHTS
IN BRIEF
TELOMERES
Shelterin fends off six repair pathways

Telomeres have to be protected from being recognized as DNA damage by the repair machinery (the so-called end-protection problem). This is mediated by the shelterin complex, and deletion of individual shelterin subunits has revealed that end-protection involves repression of signalling by ataxia telangiectasia mutated (ATM) and ATR, as well as inhibition of repair mediated by the non-homologous end-joining (NHEJ) and homology-directed repair (HDR) pathways. To definitively pin-point all pathways that are repressed by the shelterin complex, Sfeir and de Lange generated mouse telomeres lacking all shelterin proteins and associated factors. They identified two additional pathways: alternative NHEJ (which was activated when Ku70Ku80 was also absent) and nucleolytic degradation (which was activated when p53-binding protein 1 (53BP1) was also absent). So,telomeres are protected from six DNA repair pathways by the shelterin complex, which functions together with DNArepairproteins.
ORIGINAL RESEARCH PAPER Sfeir, A. & de Lange, T. Removal of shelterin reveals the telomere end-protection problem. Science 336, 593597 (2012)
N U C L E A R T R A N S P O RT
Hikeshi puts out the fire in the nucleus

During heat shock, the chaperone HSP70s translocates to the nucleus, although its exact nuclear functions are not well understood. Moreover, the mechanism of translocation was unknown, but Kose et al. now identify an evolutionarily conserved protein that acts as a protein carrier for HSP70s. This protein, which does not seem to belong to the importin- family, was shown to preferentially bind to the ATP-bound form of HSP70s and to interact with FG repeat-containing nucleoporins, thereby translocating HSP70s through the nuclear pore complex. Importantly, the authors find that depletion of the protein, which they term Hikeshi (Japanese for firefighter), inhibits HSP70s import and reduces cell viability following heat shock.
ORIGINAL RESEARCH PAPER Kose, S, Furuta, M. & Imamoto, N. Hikeshi, a nuclear import carrier for HSP70s, protects cells from heat shock-induced nuclear damage. Cell 149, 578589 (2012)
P O S T- T R A N S L AT I O N A L M O D I F I C AT I O N
Keeping cell cycle progression in check

Retinoblastoma protein (RB) prevents cell cycle progression by binding to and inhibiting E2F transcription factors. This interaction is regulated by cyclin-dependent kinase (CDK), which phosphorylates RB, thereby inactivating it. RB phosphorylation at Thr373 and Ser608 has previously been shown to inhibit the association of the RB pocket domain with the E2F transactivation domain (TD). This study elucidates the mechanism by which this occurs by delineating the structural changes induced by Ser608 and Thr373 phosphorylation. Although both phosphorylation events ultimately prevent the association of the RB pocket domain with E2F, they stimulate distinct structural changes: Ser608 phosphorylation induces a conformational change in the large loop within the RB pocket domain that blocks binding of E2FTD; and Thr373 phosphorylation triggers a conformational change that results in docking of the RB pocket and the RB amino-terminaldomain.
ORIGINAL RESEARCH PAPER Burke, J. R., Hura, G. L. & Rubin, S. M. Structures of inactive retinoblastoma protein reveal multiple mechanisms for cell cycle control. Genes Dev. 8 May 2012 (doi:10.1101/gad.189837.112)
PERSPECTIVES
OPINION
Axis of ageing: telomeres, p53 andmitochondria

Ergn Sahin and Ronald A. DePinho
Abstract | Progressive DNA damage and mitochondrial decline are both considered to be prime instigators of natural ageing. Traditionally, these two pathways have been viewed largely in isolation. However, recent studies have revealed a molecular circuit that directly links DNA damage to compromised mitochondrial biogenesis and function via p53. This axis of ageing may account for both organ decline and disease development associated with advanced age and could illuminate a path for the development of relevant therapeutics.
Ageing and its inevitable companions diseas e and death have fascinated mankind for millennia and have spurred the search for eternal youth. Long believed to be unchangeable, life expectancy and healthspan have increased dramatically over the past two centuries. This has been fuelled by advances in medical science, improved hygiene and nutrition, and significant declines in mortality rates among theyoung1. Now, the question is: can lifespan be further increased, or have we reached a biologically determined maximum lifespan? Scientists have long tried to identify pathways that are relevant for ageing, but it was not until the 1990s that the first genetic foothold was established in ageing: a loss-of-function mutation in daf-2 (which encodes an insuli n/insulin-like growth facto r1 (IGF1)like receptor) doubled the life span of worms2. Since then, the interest in the molecular pathways that control ageing has exploded, and many more mutations in meta bolic pathways have been shown to affect lifespan in different model systems, ranging from yeast to mice3. Studies suggest that these pathways are also relevant for human lifespan, but how they might extend lifespan is not entirely clear4,5. However, among others, maintenance of mitochondrial function has been suggested to be an important mechanism of extending lifespan, as decreased mitochondrial function, impaired ATP generation and increased reactive oxygen species (ROS)
energy maintenance could account for the decline seen in stem and progenitor cells as well as in post-mitotic tissues19. In this Opinion article we present a model of how different pathways DNA damage and metabolic pathways intersect and converge on mitochondria to compromise energy maintenance and drive ageing. This integrated view provides a better understanding of the mechanisms that control the fundamental process of ageing and may open new avenues for therapeutic interventions for ageing and age-associated diseases.
Metabolic pathways in ageing Several metabolic pathways and molecules regulate lifespan in response to nutrient availability and balance energy expenditure with mitochondrial function (BOX1; FIG.1). The insulin and IGF1 pathway was the first evolutionarily conserved pathway shown to regulate lifespan. Mammalian target of rapamycin (mTOR) signalling has also been implicated3,20 (BOX1). There is increasing recognition that these metabolic pathways are intimately interconnected. For instance, AMP-activated protein kinase (AMPK), which is a central sensor of energy homeostasis that modulates mTOR signalling, activates forkhead box O (FOXO) transcription factors, which are targets of insulin and IGF1 signalling. This increases the expression of genes that are involved in stress resistance and energy balance21. AMPK also activates PPAR co-activator 1 (PGC1), which is a central regulator of mitochondrial biogenesis and function. Similarly, AMPK induces sirtui n 1 (SIRT1), which activates PGC1 and FOXO transcription factors22. There is also a close relationship between FOXO proteins and PGC1, as FOXO1 and FOXO3 have been shown to increase PGC1 activity, and PGC1 itself can augment the transcriptional activity of FOXO3 (REFS 2325). SIRT1 also inactivates the guardian of the genome, p53 (REFS 26,27). p53 itself intersects with severa l different longevity pathways, including insulin and IGF1, mTOR and AMPK28. The activation of AMPK and the repression of the insulin and IGF1 pathway and the mTOR pathway by p53 demonstrate the ability of p53 to positively regulate pathways that are essentia l for cell integrity and longevity28.
VOLUME 13 | JUNE 2012 | 397
levels have been implicated in driving the ageing process6. In addition to the role of metabolic pathways, telomere maintenance has been shown to be linked to ageing7. Cultured human fibroblasts divide a finite number of times before entering a non-dividing state called senescence8. Subsequent work established that the ends of chromosomes, or telomeres, became shorter with each round of replication. These observations fuelled speculation that the loss of telomeres represents a type of molecular clock that drives ageing912. These studies underscored the general importance of DNA integrity, as dysfunctional telomeres are recognized as DNA damage and activate the DNA damage response pathway, which leads to the activation of p53 (REFS 13,14). p53, in turn, induces growth arrest, apoptosis and senescence in stem and progenitorcells1518.
a model of how different pathways ... intersect and converge on mitochondria

Recent studies have uncovered additional mechanistic insights into telomere-mediated ageing, in which telomere shortening and associated DNA damage responses promote mitochondrial dysfunction, diminished oxidative defence and compromised energygenerating processes. This general decline in
PERSPECTIVES
Box 1 | Metabolic pathways and genes implicated in lifespan regulation
Evolutionary conserved insulin and IGF1 signalling pathway The insulin and insulin-like growth factor 1 (IGF1) pathway was the first evolutionarily conserved pathway shown to regulate lifespan in Caenorhabditiselegans. Components of this pathway include phosphoinositide 3-kinase (PI3K; AGE-1 in C. elegans), 3phosphoinositide-dependent kinase 1 (PDK-1), AKT (also known as PKB) and forkhead box O (FOXO) family transcription factors (DAF-16 in C.elegans). In response to insulin and IGF-1, PI3K activates PDK-1, which in turn activates AKT. AKT phosphorylates multiple downstream targets to promote cell survival and cell growth (see the figure, part a). Decreased activity of the insulin and IGF-1 pathway owing to lossoffunction mutations in genes that encode key components of this pathway or proteins that regulate its activity, such as growth hormones, extends lifespan in many species. The lifespan extension effect in these mutants is mediated by several transcription factors and their transcriptional targets, including DAF16, heat shock factor 1 (HSF1) and SKN1 (a transcription factor involved in the oxidative stress response), which stimulate the expression of genes that increase stress resistance, oxidative defence and mitochondrial function3. mTOR signalling Mammalian target of rapamycin (mTOR) is an evolutionarily conserved protein kinase that existsin two complexes, mTORC1 (see the figure, part b) and mTORC2 (not shown). mTORC2 is involved in cytoskeletal remodelling, whereas mTORC1 is a potent regulator of cellular growth and lifespan. mTORC1 activates the ribosomal protein S6 kinase (S6K) and inhibits eukaryotic translation initiation factor 4Ebinding protein 1 (4EBP1; a negative regulator of translation) to stimulate translation and cell growth137. Inhibition of mTOR through caloric restriction, rapamycin treatment or genetic means extends lifespan in species ranging from yeast to mice29. Similarly, inhibiting S6K increases lifespan in worms and flies, and female mice lacking S6K have prolonged lifespan and are protected from age-related pathologies such as insulin resistance, immunological decline and motor dysfunction3,29. Lifespan extension in these S6Kdeficient mice seems to be mediated by AMP-activated protein kinase (AMPK), which is a central sensor of energy homeostasis36. AMPK is potently activated by an increased AMP/ATP ratio, turns off anabolic pathways through indirect inhibition of mTORC1 and a b switches on catabolic pathways to AMP/ATP Insulin, IGF-1 generate ATP under energy stress. AMPK increases ATP levels by AMPK PI3K stimulating mitochondrial biogenesis and function, as well as fatty acid oxidation, among PDK-1 others138. Activation of AMPK in mTORC1 mice with metformin (an antidiabetic drug) enhances AKT lifespan, and deletion of the AMPK orthologue aak-2 in C.elegans abrogates the lifespan extension S6K 4EBP1 DAF-16 induced in daf-2 mutants, which HSF-1 SKN-1 supports the notion that the regulation of energy pathways Translation, Autophagy Ribosomal Stress response and preservation of mitochondrial biogenesis cell growth function is integral to longevity3,29.
Nature Reviews | Molecular Cell Biology
and this, in turn, stimulates PGC1 and subsequently increases levels of OXPHOS and mitochondrial biogenesis37. Further evidence for an essential and direct role of the maintenance of mitochondrial function in lifespan comes from mice expressing a proofreading-deficient mitochondrial DNA polymerase- (Pol) variant that causes a premature ageing syndrome with shortened lifespan (the mutator mice)38,39. Overexpression of the antioxidant enzyme catalase specifically in mitochondria reduces ROS-induced damage and significantly improves age-related cardiac decline in both wild-type and mutator mice40,41. However, many other ROS-related studies have yielded conflicting results and question the role of increased ROS levels for the ageing process42,43. Subjecting mutator mice to continuous exercise potently rescues the premature ageing phenotype, which indicates that mitochondrial biogenesis and turnover (both of which are quality control mechanisms) are important for slowing down the ageing proces s in these mice44. Although these studies indicate that mitochondrial decline drives ageing, other studies show a more complex picture, as mild impairment of mitochondrial function can extend lifespan in yeast, worms and mice3,4549. This dichotomous role of mitochondria in lifespan has also been noted with other molecules that are involved in ageing, including p53 and AMPK5052. Thus, a more detailed view of how these key molecules are tightly regulated in the context of mitochondrial biology is required to optimally control the molecular circuitry of ageing.
Telomeres and ageing A separate line of extensive research has implicated telomere integrity as a major regulator of longevity53,54. Telomeres are repetitive TTAGGG sequences that cap chromosomes and prevent the ends from being recognized as DNA damage55. Most human cells lack adequate levels of telomerase to maintain telo meres, and this results in telomere shortening with each round of replication9,11,56. The importance of telomere length in ageing was initially inferred from seminal studies carried out in primary human fibroblasts in the early 1960s by Hayflick and Moorhead8. These cells divide a finite number of times invitro and undergo telomere shortening with continuous passaging and, eventually, senescence8. Telomerase reactivation elongates telomeres and allows fibroblasts to bypass senescence and grow indefinitely, which demonstrates the causal role of shortene d telomeres in cellul ar ageing57.
www.nature.com/reviews/molcellbio
How these pathways, alone or in combination, regulate lifespan is being extensively investigated, and many different mechanisms seem to be involved3,29. Among these, mitochondrial function and ROS defence are considered to be important modulators of lifespan. Along these lines, decreased activity of the insulin and IGF1 pathway is associated with improved mitochondrial function, as demonstrated in long-lived Ames mice (which have very low levels of IGF1) and in mice with decreased levels of insulin receptor substrate 2 (IRS2)30,31. Reduced insulin and IGF1 signalling results in the activation of FOXO transcription factors, which induce
398 | JUNE 2012 | VOLUME 13
the expression of antioxidants, such as manganese superoxide dismutase (MnSOD) and catalase; accordingly, FOXO-deficient mice display increased levels of ROS and stem cell depletion3234. Decreased mTOR activity during dietary restriction is also associated with improved mitochondrial function, and lifespan extension in this model depends on increased levels of respiration35. Furthermore, mice lacking S6 kinase (S6K), which is a downstream component of mTOR signalling, are long lived and show increased oxidative phosphorylation (OXPHOS) and oxygen consumption36. AMPK enhances SIRT1 activity (through increased NAD+ levels)
2012 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES
Indeed, telomere length has been shown to gradually decline with age in many human tissues, including proliferative compartments and more quiescent tissues11,5860. Interestingly, even cells that express telo merase undergo telomere shortening over time, which points towards a complex regulation of telomere length61. Moreover, many studies have found a positive correlation between telomere shortening in human peripheral leukocytes and the risk of typical age-associated diseases62. Further support for the role of telo meres in ageing comes from patients with loss-of-function mutations in genes that are crucial for telomere length maintenance, as these mutations predispose individuals to accelerated ageing. Mutations in TERC (the RNA component of telomerase) and TERT (the catalytic component of telo merase) are found in patients with the premature ageing syndrome dyskeratosis congenita63. Mutations in the genes encoding Werner syndrome ATP-dependent helicase (WRN) and ataxia telangiectasia mutated (ATM) cause Werner syndrome and the neurodegenerative disorder ataxia telangiectasia, respectively64. In addition to these multisystem disorders, TERC and TERT loss-of-function mutations are associated with the development of more organ-restricted diseases such as liver fibrosis, idiopathic pulmonary fibrosis and bone marrow failure syndromes63. The manifestation of degenerative phenotypes in telomere maintenance conditions depends on the degree of telomere dysfunction, as evidenced by the earlier and more severe occurrence of pathologies in subsequent generationsof patients with dyskeratosis congenita, who have shorter telomeres (known as the anticipation effect)65. Although these studies of telomere maintenance disorders have provided evidence for the importance of telomeres for organ integrity and lifespan regulation, some pathologies seen in these patients are not typically observed during normal ageing and caution against a simple extrapolation of these findings to normal ageing. The exacerbated phenotypes in these patients might relate to excessive telomere shortening beyond what is seen in many proliferative and more static tissues during normal ageing in humans and may also be driven by environmental factors. A link between telomeres and ageing has also been obtained from studies in mice. There is increasing recognition that telomere length and integrity are compromised during ageing in wild-type mice66,67. Interestingly, the presence of one or a few
DNA damage Increased p53 and p16 activity Severe mitochondrial dysfunction
Ageing mTOR and S6K inhibition Insulin and IGF1 signalling inhibition
Figure 1 | Proposed causes of ageing. Major cellular pathways are implicated in the ageing process. Nature Reviews | Molecular Cell Biology Increased DNA damage, p53 and p16 activity and mitochondrial dysfunction have been shown to promote functional decline and ageing. By contrast, decreased activity in the mammalian target of rapamycin (mTOR), S6 kinase (S6K) and the insulin and insulin-like growth factor 1 (IGF1) pathways increase lifespan in different organisms.
dysfunctional telomeres in cells seems to be sufficient to trigger a DNA damage response, and this could underline the development of pathologies68,69. Accordingly, overexpression of telomerase can delay some age-associated changes in mice that have been engineered to be resistant to cancer70. Furthermore, the role of telomeres for organismal fitness and lifespan has been substantiated in mice lacking telomerase activity, which develop numerous age-associated degenerative pheno types once their telomeres become short and die prematurely71,72. Moreover, mouse models of Werner syndrome and ataxia telangiectasia develop classical human-like pathologies only when their telomeres are short, which demonstrates the essential role of critically short telo meres in disease manifestation16,73,74. Finally, telomerase overexpression can reverse ageassociate d decline in multiple tissues in mice with established degenerative phenotypes75. These studies indicate that telomere dysfunction can drive the functional decline of tissues, promote ageing and shorten lifespan and, importantly, that the ageing process can be prevented or even be reversed by telo merase reactivation. However, it is clear that currently there is only an elemental understanding of the precise role of telomeres in natural ageing and how telomeres might influence age-associated pathologies.
Telomeremitochondrion connection How do waning telomeres precipitate such widespread degeneration? One clue comes from the observation that patients with dyskeratosis congenita,Werner syndrome and ataxia telangiectasia, and mice with dysfunctional telomeres, develop organ failure particularly in highly proliferative organs such as the intestines, skin and bone marrow7678. These organs rely on continuous regeneration, which is mediated by resident stem and progenitor cells. This observation has led to the hypothesis that telomere-based ageing is primarily a
stem cell defect caused by the activation of p53 and the induction of growth arrest, senescence and apoptosis in these cellular compartments79. Indeed, telomere shortening is accompanied by increased p53 activity in these cells and, consequently, high levels of apoptosis15,18. Mice lacking p53 or its downstream targets show functional rescue of stem and progenitor cells in the haemato poietic system, skin and gastrointestinal tract, as well as concomitant rescue of tissue pathologies15,18,80. Although the stem cell theory of ageing helps to rationalize the failure of highly regenerative tissues, it does not readily explain age-dependent changes in more quiescent tissues that dependent less on stem and progenitor cell activity for tissue homeostasis. Indeed, general metabolic disorders and functional decline in mostly post-mitotic tissues such as the heart, liver and pancreas are well-recognized features in the aged, in patients with telomere maintenance disorders and in telomere-dysfunctional mice65,81. For instance, individuals with dyskeratosis congenita,Werner syndrome and ataxia telangiectasia are prone to develop insulin resistance and diabetes65,82,83. Moreover, cardiomyopathy has been recognized inpatients with dyskeratosis congenita, in telomere-dysfunctional mice and in mouse models of ataxia telangiectasia81,84,85. Liver fibrosis and pulmonary fibrosis represent other pathophysiological manifestations in patients with dyskeratosis congenita. Hepatic toxicity is a major side effect in patients with dyskeratosis congenita, who receive cytotoxic chemotherapy for aplastic anaemia-related bone marrow transplantation65. Quiescent tissues such as the heart and liver have also been reported to undergo age-dependent telomere shortening, the basis for which is unclear58,60. Together, these observations suggest additional mechanisms of telomere-induced ageing beyond traditional p53-dependent checkpoint responses of apoptosis and senescence.
VOLUME 13 | JUNE 2012 | 399
PERSPECTIVES
Telomere dysfunction
p53 PGC1,
Mitochondrion
Mitochondrial biogenesis Mitochondrial function ROS Functional decline of post-mitotic tissues and stem cells
Fatty acid oxidation Gluconeogenesis Glucose utilization
Ageing
Figure 2 | Telomerep53PGC pathway. p53 induced by telomere dysfunction binds to the Nature Reviews | Molecular Cell Biology promoters of PPAR co-activator 1 (PGC1) and PGC1 and represses the expression of PGC1A and PGC1B. The repression of both coactivators impairs overall mitochondrial biogenesis and function and leads to defective ATP generation and increased levels of reactive oxygen species (ROS). PGCs are also involved in energy metabolism by regulating different biochemical pathways such as fatty acid oxidation, gluconeogenesis, glucose uptake and oxidation. The compromise in mitochondrial function and other biochemical pathways might equally lead to functional decline in tissue stem cells and post-mitotic tissues and drive ageing. Telomerase reactivation or PGC overexpression can reverse PGCassociated metabolic and mitochondrial changes in mice with established telomere dysfunction. Telomere dysfunction may also lead to compromised mitochondrial function and energy metabolism through other pathways (dashed arrows).
Some clues for additional mechanisms have surfaced from recent work in TERTdeficient mice with telomere dysfunction. This study reported a marked compromise in mitochondrial biogenesis and function in diverse tissues, including liver, heart and haematopoietic stem cells, which raises the possibility that a fundamental problem in energy maintenance might contribute to the premature ageing phenotypes in these mice19. These marked mitochondrial changes seem to be caused by the combined suppression of the transcriptional co-activators PGC1 and PGC1 and their downstream targets (FIG.2). This is mediated by direct binding of p53 to the promoters of PGC1 and PGC1; accordingly, telomere-dysfunctional mice lacking p53 have normal PGC expression, increased mitochondrial DNA (mtDNA) content, improved gluconeogenesis and blunted doxorubicin-induced cardiomyopathy. In line with the important role of PGCs in regulating diverse processes, TERT-deficient mice show reduced expression of genes that are essential for gluconeogenesis, -oxidation and ROS defence, and they have greatly compromised OXPHOS, with reduced ATP generation, impaired gluconeogenic capacity and age-dependent cardiomyopathy. Of note, these changes are more pronounced with increasing telomere
400 | JUNE 2012 | VOLUME 13
dysfunction. Importantly, overexpression of TERT or PGC1 in mice with telomere dysfunction improves mitochondrial respiration and gluconeogenesis, which confirms that the phenotype of TERT-deficient mice is caused by PGC suppression19. This telomer e mitochondrion link is also suggested by other studies, including those demonstrating increased ROS levels and mitochondrial dysfunction in cultured human fibroblasts, in fibroblasts overexpressing a mutant form of TERT and in heart tissues of TERT-deficient mice8688. The basis for these defects has been suggested to be secondary to the activation of the p21transforming growth factor- (TGF)p53 pathways and increased mtDNA damage8688. Moreover, a recent study found reduced mitochondrial membrane hyperpolarization and impaired Ca2+ influx in telomere-dysfunctional mice, which leads to reduced insulin release in -cells89. Previous studies have suggested that TERT has telomere elongation-independent functions9094. However, TERC-deficient mice, which lack telomerase activity but have intact TERT expression, showed similar robust changes centred on PGC and mitochondrial suppression, which indicates that telomere dysfunction is the determining factor driving these changes19. Moreover, Tert- and Terc-knockout mice have indistinguishable
phenotypes and similar transcriptomic profiles95. That said, these genetic studies were not designed to rigorously exclude more subtle telomere-independen t roles of TERT. Indeed, continued study is warranted, as TERT has been shown to localize to mitochondria, retain reverse transcriptase activity, carry out different mitochondrial functions (such as modulating mtDNA integrity, improving respiratory chain function and affecting ROS production) and to potentially activate other pathways such as the WNT pathway (although this has been questionedrecently) 88,96100. Furthermore, aged tissues often demonstrate concomitant telomere dysfunction, increased DNA damage and p53 activity as well as reduced PGC levels and mitochondrial function7. Importantly, studies in cells derived from patients and mouse models corroborate this link between telomeres and mitochondria101103. For example, cells derived from patients with Werner syndrome have compromised mitochondrial function and increased ROS levels103. Although these observations support the importance of the telomerep53mitochondrion axisofageing, more work is required to assess whether mitochondrial biogenesis andfunctio n are consistently impaired in human telomere-shortening conditions.
An integrated view of ageing The connection between telomeres and mitochondria supports a model for agein g whereby DNA damage-induced p53 activation leads to mitochondrial dysfunction through the suppression of the master regulators of mitochondrial biogenesis andfunctio n, PGC1 and PGC1. The telomerep53mitochondrion model of ageing integrates many factors that have been shown to be important in the ageing process. On the genomic level, it accounts for ageing driven by DNA damage. This damag e can stem from telomere shortening or from a decrease in the expression of genes that mediate DNA stability and DNA repair. Second, the model accounts for ageing syndromes documented in mice with hyper active Tp53 alleles and mouse models that show increased DNA damage owing to mutations in Terc, Tert, the DNA repair genes Ku80 (also known as Xrcc5) and breast cance r1 (Brca1), and Zmpste24, which encodes a metalloproteinase involved in lamin A maturation (lamin A is a crucial component of the nuclear envelope, and mutations of this metalloproteinase lead to premature ageing)104,105. Finally, the model accounts for ageing phenotypes that stem from mitochondrial dysfunction,
PERSPECTIVES
asmicelacking PGC1, PGC1, BMI (a negative regulator of p16 that is upregulated in many tissues) or FOXO develop accelerate d tissue degeneration and mitochondrial dysfunction106109. This model could also explain both the slow but progressive physiological decline and the precipitous nature of the ageing process (FIG.3). Ensuing mitochondrial dysfunction sustains a feed-forward cycle of DNA damage and further mitochondrial dysfunction through the generation of DNAdamaging ROS and possibly other mitochondrion-derived factors, such as ironsulphur (FeS) clusters and NADH/ NAD. This boost in ROS production fuels a detrimental cycle of increased genotoxic damage, particularly to the G-rich sequences of telomeres, followed by sustained activation of p53, further mitochondrial decline, more ROS generation, and so on110,111. Increased ROS levels also damage other cellular components, including mtDNA, which further sustains this feed-forward spiral of damage by suppressing the expression of mtDNA-encoded genes for OXPHOS. Under conditions of severe nuclear or mtDNA damage, however, this cycle could be bypassed, and the premature ageing phenotype could be driven by increased apoptosis across different tissues, as reported in mutator mice (which carry a mutant form of Pol; see above)38,39. The continuous, escalating nature of this cycle could explain the differential effect of p53 on ageing in response to varying degrees of DNA damage. Under low levels of genotoxic stress, p53 induces the expression of antioxidants, thereby favouring cell survival. By contrast, increasing levels ofDNA damage promote the expression of pro-oxidants, which further promote cell damage112. Similarly, p53 has been shown to promote mitochondrial function and biogenesis in wild-type mice or in cells with low levels ofp53, but under conditions of genotoxic stress p53 is associated with impaired mitochondrial function19,113115. There is also evidence that p53 either does not change lifespan (as demonstrated in mice with a hypomorphic Mdm2 allele and in transgenic mice carrying an extra copy of the wild-type Tp53 locus) or delays ageing (which was shown in mice carrying one extra copy of Tp53 and increased copies of the tumour suppressor Arf (also known as Cdkn2a))116,117. Furthermore, studies in worms have demonstrated that the p53 orthologue CEP-1 can extend or decrease lifespan depending on the degree of mitochondrial impairment, which indicates that the dichotomous role of p53 might be preserved during evolution118.
DNA damage
p53
p16
BMI
Insulin, IGF1 ROS FeS NADH/NAD mTOR
AMPK
SIRT1
PGC1, ?
Mitochondrial dysfunction
Figure 3 | A unified theory of ageing. In this model, increased DNA damage (for example, owing Nature Reviews | Molecular Cell Biology to telomere attrition, impaired DNA repair and increased reactive oxygen species (ROS) levels) activates p53, and increasing levels of p53 ultimately lead to compromised mitochondrial function through the repression of PPAR co-activator 1 (PGC1) and PGC1 (which promote mitochondrial biogenesis). This p53mediated mitochondrial dysfunction triggers a cycle of DNA damage (by affecting the production of ROS, ironsulphur (FeS) clusters and NADH/NAD), which in turn leads to further p53 activation and mitochondrial compromise. This feed-forward loop could also account for the divergent and opposite effects of many players in the ageing process. Under mild stress conditions, several components depicted here (p53, mitochondria and AMP-activated protein kinase (AMPK)) have been shown to preserve cellular function but to promote cellular ageing under more severe stress conditions (see text). The interplay between p53 and other pathways that have been implicated in ageing is also indicated. p53 represses the activity of the insulin and insulin-like growth factor1 (IGF1) pathway and the mammalian target of rapamycin (mTOR) pathway and activates AMPK. How the altered activity of these pathways modifies mitochondrial function and the ageing process in the setting of increased DNA damage is not clear. Other p53dependent and p53independent pathways might cooperate in inducing mitochondrial dysfunction. For example, BMI1 indirectly inhibits p53 activation, and BMI1 loss upregulates p16 expression, which increases p53 activity indirectly (dashed arrow) by interacting with MDM2, the negative regulator of p53 (not shown). BMI1 has also been shown to induce mitochondrial dysfunction (probably indirectly). In addition, loss of sirtuins may contribute to mitochondrial dysfunction, as active sirtuin 1 (SIRT1) decreases p53 activity, and loss of SIRT1 promotes p53 activation and its downstream functions. SIRT1 also activates PGC1 and thereby boosts mitochondrial biogenesis. The consequences of mitochondrial dysfunction are, when mild, functional impairment (for example, decreased ATP generation and oxidation) without cell loss. However, under increased stress conditions, mitochondrial dysfunction leads to functional impairment and concomitant loss of parenchymal mass owing to increased apoptosis and senescence.
In the presented model, mild DNA damage and mild or moderate levels of p53 activation would allow repair and maintenance of cellular function, whereas excessive DNA damage and p53 activation would eliminate cells that are not fit to survive or carry too much damage. These genetic findings also highlight the differential effects of mitochondrial function on lifespan observed in worms, flies and mice and indicate that the functional consequences of mitochondrial impairment could also depend on the activity levels of p53. The mild inhibition of mitochondrial respiration could activate longevity pathways (including those gover ned by p53), whereas more pronounced impairment of mitochondrial respiration
with additional defects in biochemical processes, such as -oxidation, could trigger a lifespan-shortening programme. How p53 switches from a pro-survival to a pro-ageing protein is largely unknown, but the intersection of p53 with the insulin and IGF1, mTOR and AMPK pathways might provide some clues28. p53 suppresses the insulin and IGF1 pathway and the mTOR pathway through direct transcriptional upregulation of the negative regulators phosphatase and tensin homologue (PTEN), IGF1-binding protein 3 (IGF1BP3) and tuberous sclerosis protein 2 (TSC2; also known as tuberin), and p53 activates AMPK through direct transcriptional upregulation of its -subunit28. Interestingly, prematurely
VOLUME 13 | JUNE 2012 | 401
PERSPECTIVES
aged mice with increased levels of DNA damage or with p53 hyperactivation show suppression of the insulin and IGF1 pathway and the mTOR pathway119,120. These paradoxical findings can be reconciled by the view that the activation of these pathways represents a compensatory mechanism to maintain and prolong lifespan in the setting of ongoing DNA damage120. However, it is not clear whether the activation of these longevity pathways in the context of DNA damage leads to the same transcriptional and cellular changes as seen in long-lived carriers of mutations in these pathways. For example, the prolonged activation of AMPK seen under energy stress leads to p53-dependent cellular senescence and apop tosis, which indicates that AMPK activation can accelerate cellular ageing under specific conditions52. Similarly, IGF1 concentration decreases with age, and this has been linked to functional decline in different stem cells121. These findings further support the notion that the effects of classical ageing pathways might be context dependent. The model presented here acknowledges the existence of other, yet to be identified pathways that are involved in mediating mitochondrial and metabolic compromise, as p53 deficiency in mice with dysfunctional telomeres only partially restores PGC levels and ameliorates mitochondrial defects. Other p53 family members are prime candidates, in particular p63, which has been shown to be involved in organismal ageing and cellular senescence122,123. Sirtuins may also have a role, as they associate with telo meres and regulate p53 and PGC1124. Indeed, SIRT1 activity has been found to decrease in aged tissues, and this might contribute to the increased p53 activity and suppressed PGC1 activity seen in aged mouse and human tissues7,106. However, recent reports have questioned the relevance of SIRT1 in lifespan regulation125, which emphasizes the need for studies that focus on the role of sirtuins and their linkage to this axis in ageing and age-related pathologies. Another potential candidate is the BMIp16 pathway, as it is highly implicated in ageing, and loss of BMI impairs mitochondrial function107. Finally, p21-dependent signalling has been suggested to induce mitochondrial dysfunction in cultured human fibroblasts with critically short telomeres, although it should be noted that fibroblasts depend less on mitochondria and OXPHOS for ATP production and mainly use glycolysis to generate ATP86. In this context, it is important to note here that telomeredysfunctional yeast display increased expression of OXPHOS genes and proliferation of
402 | JUNE 2012 | VOLUME 13 2012 Macmillan Publishers Limited. All rights reserved
mitochondria (although the functionality of the mitochondria was not tested), and senescent fibroblasts with dysfunctional telo meres have been reported to have increased mitochondrial biogenesis126,127. These studies highlight not only the cell-specific effects of telomere dysfunction on mitochondrial bio logy but also the differences between mice and yeast, which might be related to growth conditions and yeast-specific regulation after telomere dysfunction. In this model, the cellular phenotypes induced by mitochondrial dysfunction would range from functional impairment (for example, decreased ATP generation) to classical cellular phenotypes of growth arrest, senescence and apoptosis. The link between impaired mitochondrial function and cellular senescence has been demonstrated invitro: decreasing the expression of Rieske FeS protei n (RISP) of complex III or pharmacological inhibition of the electron transport chain and OXPHOS are sufficient to trigger senescence128. The combined effects of these non-exclusive phenotypes would include cellular and tissue compromise and functionalfailure.
Deciphering these ageing networks could advance the development of therapeutic strategies
Conclusion The integrated model of ageing presented in this Opinion article focuses on the intersection of DNA damage and metabolic pathways and how they might converge on a common effector, mitochondria, to drive ageing. Although this model centres on mitochondria, other important effectors of ageing, such as dysregulated autophagy, translation and protein folding, no doubt conspire to bring about and/or reinforce the ageing process129131. It remains to be established whether and how these different effectors are commonly used by DNA damage and metabolic pathways to drive ageing. Similarly, how these effectors are interconnected merits further studies in different ageing model systems. It is largely unclear how mitochondria, p53 and other players of the ageing process can both extend and shorten lifespan. It will be crucial to establish the molecular mechanisms that determine different outcomes, which may involve cell-type-specific actions of p53 and/or various isoforms of the p53 family members. In this regard, it will be essential to characterize what other mitochondrial
biochemical pathways (beyond ROS production and OXPHOS) are impaired and might contribute to cellular and organism al ageing. Deciphering these ageing networks could yield biomarkers of ageing and advance the development of therapeutic strategies designed to rejuvenate both proliferating and quiescent tissues in the aged. These therapeutic strategies might include: stabilizing telomeres through transient telomerase reactivation; attenuation of p53 activation or neutralization of specific p53 targets that are specifically involved in ageing; and enhancin g PGC activity to promote mitochondrial biogenesis and function. Along these lines, a small molecule activator of telomerase has been reported to prolong healthspan in female mice without increasing the risk of cancer132. Moreover, PGC1 overexpression in skeletal muscle ameliorates age-associated decline of muscular function in wild-type mice133. Similarly, other interventions known to improve mitochondrial biogenesis and function, such as physical activity or administration of resveratrol (which is a putative sirtuin activator), have been shown to improve age-associated decline134,135. It is intriguing that sirtuins can deacetylate p53 (which decreases its activity) and PGCs (which increases their activity), thereby differentially modulating two key components in the pathway that connects DNA damage signalling and mitochondrial decline. Beyond ageing, recent studies have also uncovered the importance of the telomerep53mitochondrion axis for cancer, which suggests that this pathway could be targeted for cancer therapy136. The discovery of the molecular circuitry of ageing has positioned the field to develop rational strategies for ageing, a disease with 100% penetrance and 100% mortalit y. Although it remains to be determined whether natural ageing can be blocked or even be reversed, as recently demonstrated in the premature ageing setting75, therapeutic manipulation of this pathway may hold promise for the reduction of age-related diseases, which have become more prevalent with marked increases in life expectancy worldwide.
Ergn Sahin is at the Huffington Center On Aging and the Department of Molecular Physiology and Biophysics, Baylor College of Medicine, One Baylor Plaza, Houston, Texas77030, USA. Ronald A. DePinho is at the Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, Houston, Texas77030, USA. e-mails: esahin@bcm.edu; rdepinho@mdanderson.org doi:10.1038/nrm3352 Published online 16 May 2012
PERSPECTIVES
1. 2. 3. 4. Christensen, K., Doblhammer, G., Rau, R. & Vaupel, J.W. Ageing populations: the challenges ahead. Lancet 374, 11961208 (2009). Kenyon, C., Chang, J., Gensch, E., Rudner, A. & Tabtiang, R.A. C. elegans mutant that lives twice as long as wild type. Nature 366, 461464 (1993). Kenyon, C.J. The genetics of ageing. Nature 464, 504512 (2010). Deelen, J. etal. Gene set analysis of GWAS data for human longevity highlights the relevance of the insulin/IGF-1 signaling and telomere maintenance pathways. Age (Dordr.) 24 Nov 2011(doi:10.1007/ s11357-011-9340-3). Ziv, E. & Hu, D. Genetic variation in insulin/IGF-1 signaling pathways and longevity. Ageing Res. Rev. 10, 201204 (2011). Guarente, L. Mitochondria a nexus for aging, calorie restriction, and sirtuins? Cell 132, 171176 (2008). Sahin, E. & Depinho, R.A. Linking functional decline of telomeres, mitochondria and stem cells during ageing. Nature 464, 520528 (2010). Hayflick, L. & Moorhead, P.S. The serial cultivation of human diploid cell strains. Exp. Cell Res. 25, 585621 (1961). Harley, C.B., Futcher, A.B. & Greider, C.W. Telomeres shorten during ageing of human fibroblasts. Nature 345, 458460 (1990). Allsopp, R.C. etal. Telomere length predicts replicative capacity of human fibroblasts. Proc. Natl Acad. Sci. USA 89, 1011410118 (1992). Hastie, N.D. etal. Telomere reduction in human colorectal carcinoma and with ageing. Nature 346, 866868 (1990). Harley, C.B. Telomere loss: mitotic clock or genetic time bomb? Mutat. Res. 256, 271282 (1991). Vaziri, H. & Benchimol, S. From telomere loss to p53 induction and activation of a DNA-damage pathway at senescence: the telomere loss/DNA damage model of cell aging. Exp. Gerontol. 31, 295301 (1996). Takai, H., Smogorzewska, A. & de Lange, T. DNA damage foci at dysfunctional telomeres. Curr. Biol. 13, 15491556 (2003). Chin, L. etal. p53 deficiency rescues the adverse effects of telomere loss and cooperates with telomere dysfunction to accelerate carcinogenesis. Cell 97, 527538 (1999). Wong, K.K. etal. Telomere dysfunction and Atm deficiency compromises organ homeostasis and accelerates ageing. Nature 421, 643648 (2003). Ferron, S. etal. Telomere shortening and chromosomal instability abrogates proliferation of adult but not embryonic neural stem cells. Development 131, 40594070 (2004). Flores, I. & Blasco, M.A. A p53-dependent response limits epidermal stem cell functionality and organismal size in mice with short telomeres. PLoS ONE 4, e4934 (2009). Sahin, E. etal. Telomere dysfunction induces metabolic and mitochondrial compromise. Nature 470, 359365 (2011). Harrison, D.E. etal. Rapamycin fed late in life extends lifespan in genetically heterogeneous mice. Nature 460, 392395 (2009). Chiacchiera, F. & Simone, C. The AMPKFoxO3A axis as a target for cancer treatment. Cell Cycle 9, 10911096 (2010). Canto, C. & Auwerx, J. PGC-1, SIRT1 and AMPK, an energy sensing network that controls energy expenditure. Curr. Opin. Lipidol. 20, 98105 (2009). Daitoku, H., Yamagata, K., Matsuzaki, H., Hatta, M. & Fukamizu, A. Regulation of PGC-1 promoter activity by protein kinase B and the forkhead transcription factor FKHR. Diabetes 52, 642649 (2003). Olmos, Y. etal. Mutual dependence of Foxo3a and PGC-1 in the induction of oxidative stress genes. J.Biol. Chem. 284, 1447614484 (2009). Puigserver, P. etal. Insulin-regulated hepatic gluconeogenesis through FOXO1-PGC-1 interaction. Nature 423, 550555 (2003). Luo, J. etal. Negative control of p53 by Sir2 promotes cell survival under stress. Cell 107, 137148 (2001). Vaziri, H. etal. hSIR2(SIRT1) functions as an NADdependent p53 deacetylase. Cell 107, 149159 (2001). Feng, Z. & Levine, A.J. The regulation of energy metabolism and the IGF-1/mTOR pathways by the p53 protein. Trends Cell Biol. 20, 427434 (2010). Haigis, M.C. & Yankner, B.A. The aging stress response. Mol. Cell 40, 333344 (2010). 30. Sadagurski, M. etal. IRS2 increases mitochondrial dysfunction and oxidative stress in a mouse model of Huntington disease. J. Clin. Invest. 121, 40704081 (2011). 31. Brown-Borg, H.M., Johnson, W.T. & Rakoczy, S.G. Expression of oxidative phosphorylation components in mitochondria of long-living Ames dwarf mice. Age(Dordr.) 34, 4357 (2011). 32. Tothova, Z. etal. FoxOs are critical mediators of hematopoietic stem cell resistance to physiologic oxidative stress. Cell 128, 325339 (2007). 33. Renault, V.M. etal. FoxO3 regulates neural stem cell homeostasis. Cell Stem Cell 5, 527539 (2009). 34. Paik, J.H. etal. FoxOs cooperatively regulate diverse pathways governing neural stem cell homeostasis. CellStem Cell 5, 540553 (2009). 35. Zid, B.M. etal. 4E-BP extends lifespan upon dietary restriction by enhancing mitochondrial activity in Drosophila. Cell 139, 149160 (2009). 36. Selman, C. etal. Ribosomal protein S6 kinase 1 signaling regulates mammalian life span. Science 326, 140144 (2009). 37. Canto, C. etal. AMPK regulates energy expenditure by modulating NAD+ metabolism and SIRT1 activity. Nature 458, 10561060 (2009). 38. Trifunovic, A. etal. Premature ageing in mice expressing defective mitochondrial DNA polymerase. Nature 429, 417423 (2004). 39. Kujoth, G.C. etal. Mitochondrial DNA mutations, oxidative stress, and apoptosis in mammalian aging. Science 309, 481484 (2005). 40. Dai, D.F. etal. Overexpression of catalase targeted to mitochondria attenuates murine cardiac aging. Circulation 119, 27892797 (2009). 41. Dai, D.F. etal. Age-dependent cardiomyopathy in mitochondrial mutator mice is attenuated by overexpression of catalase targeted to mitochondria. Aging Cell 9, 536544 (2010). 42. Perez, V.I. etal. Is the oxidative stress theory of aging dead? Biochim. Biophys. Acta 1790, 10051014 (2009). 43. Hekimi, S., Lapointe, J. & Wen, Y. Taking a good look at free radicals in the aging process. Trends Cell Biol. 21, 569576 (2011). 44. Safdar, A. etal. Endurance exercise rescues progeroid aging and induces systemic mitochondrial rejuvenation in mtDNA mutator mice. Proc. Natl Acad. Sci. USA 108, 41354140 (2011). 45. Dillin, A. etal. Rates of behavior and aging specified by mitochondrial function during development. Science 298, 23982401 (2002). 46. Lee, S.S. etal. A systematic RNAi screen identifies a critical role for mitochondria in C. elegans longevity. Nature Genet. 33, 4048 (2003). 47. Kirchman, P.A., Kim, S., Lai, C.Y. & Jazwinski, S.M. Interorganelle signaling is a determinant of longevity in Saccharomyces cerevisiae. Genetics 152, 179190 (1999). 48. Dellagnello, C. etal. Increased longevity and refractoriness to Ca2+-dependent neurodegeneration in Surf1 knockout mice. Hum. Mol. Genet. 16, 431444 (2007). 49. Liu, X. etal. Evolutionary conservation of the clk-1-dependent mechanism of longevity: loss of mclk1 increases cellular fitness and lifespan in mice. GenesDev. 19, 24242434 (2005). 50. Vigneron, A. & Vousden, K. H. p53, ROS and senescence in the control of aging. Aging (Albany NY) 2, 471474 (2010). 51. Sung, J.Y., Woo, C.H., Kang, Y.J., Lee, K.Y. & Choi, H.C. AMPK induces vascular smooth muscle cell senescence via LKB1 dependent pathway. Biochem. Biophys. Res. Commun. 413, 143148 (2011). 52. Jones, R.G. etal. AMP-activated protein kinase induces a p53-dependent metabolic checkpoint. Mol.Cell 18, 283293 (2005). 53. Blackburn, E.H. Switching and signaling at the telomere. Cell 106, 661673 (2001). 54. Shay, J.W. & Wright, W.E. Hallmarks of telomeres in ageing research. J. Pathol. 211, 114123 (2007). 55. Chan, S.R. & Blackburn, E.H. Telomeres and telomerase. Phil. Trans. R. Soc. Lond. B 359, 109121 (2004). 56. Kim, N.W. etal. Specific association of human telomerase activity with immortal cells and cancer. Science 266, 20112015 (1994). 57. Bodnar, A.G. etal. Extension of life-span by introduction of telomerase into normal human cells. Science 279, 349352 (1998). 58. Aikata, H. etal. Telomere reduction in human liver tissues with age and chronic inflammation. Exp. Cell Res. 256, 578582 (2000). 59. Aubert, G. & Lansdorp, P.M. Telomeres and aging. Physiol. Rev. 88, 557579 (2008). 60. Chimenti, C. etal. Senescence and death of primitive cells and myocytes lead to premature cardiac aging and heart failure. Circ. Res. 93, 604613 (2003). 61. Chen, J. Hematopoietic stem cell development, aging and functional failure. Int. J. Hematol. 94, 310 (2011). 62. Blasco, M.A. Telomeres and human disease: ageing, cancer and beyond. Nature Rev. Genet. 6, 611622 (2005). 63. Calado, R.T. & Young, N.S. Telomere diseases. N.Engl. J. Med. 361, 23532365 (2009). 64. Martin, G.M. Genetic modulation of senescent phenotypes in Homo sapiens. Cell 120, 523532 (2005). 65. Armanios, M. Syndromes of telomere shortening. Annu. Rev. Genom. Hum. Genet. 10, 4561 (2009). 66. Flores, I. etal. The longest telomeres: a general signature of adult stem cell compartments. Genes Dev. 22, 654667 (2008). 67. Hewitt, G. etal. Telomeres are favoured targets of a persistent DNA damage response in ageing and stress-induced senescence. Nature Commun. 3, 708 (2012). 68. Hemann, M.T., Strong, M.A., Hao, L.Y. & Greider,C.W. The shortest telomere, not average telomere length, is critical for cell viability and chromosome stability. Cell 107, 6777 (2001). 69. Hao, L.Y. etal. Short telomeres, even in the presence of telomerase, limit tissue renewal capacity. Cell 123, 11211131 (2005). 70. Tomas-Loba, A. etal. Telomerase reverse transcriptase delays aging in cancer-resistant mice. Cell 135, 609622 (2008). 71. Lee, H.W. etal. Essential role of mouse telomerase in highly proliferative organs. Nature 392, 569574 (1998). 72. Rudolph, K.L. etal. Longevity, stress response, and cancer in aging telomerase-deficient mice. Cell 96, 701712 (1999). 73. Du, X. etal. Telomere shortening exposes functions for the mouse Werner and Bloom syndrome genes. Mol.Cell. Biol. 24, 84378446 (2004). 74. Chang, S. etal. Essential role of limiting telomeres in the pathogenesis of Werner syndrome. Nature Genet. 36, 877882 (2004). 75. Jaskelioff, M. etal. Telomerase reactivation reverses tissue degeneration in aged telomerase-deficient mice. Nature 469, 102106 (2011). 76. Vulliamy, T.J. & Dokal, I. Dyskeratosis congenita: the diverse clinical presentation of mutations in the telomerase complex. Biochimie 90, 122130 (2008). 77. Muftuoglu, M. etal. The clinical characteristics of Werner syndrome: molecular and biochemical diagnosis. Hum. Genet. 124, 369377 (2008). 78. Perlman, S., Becker-Catania, S. & Gatti, R.A. Ataxia-telangiectasia: diagnosis and treatment. Semin.Pediatr. Neurol. 10, 173182 (2003). 79. Sharpless, N.E. & DePinho, R.A. Telomeres, stem cells, senescence, and cancer. J. Clin. Invest. 113, 160168 (2004). 80. Choudhury, A.R. etal. Cdkn1a deletion improves stem cell function and lifespan of mice with dysfunctional telomeres without accelerating cancer formation. Nature Genet. 39, 99105 (2007). 81. Leri, A. etal. Ablation of telomerase and telomere loss leads to cardiac dilatation and heart failure associated with p53 upregulation. EMBO J. 22, 131139 (2003). 82. Lee, J.W., Harrigan, J., Opresko, P.L. & Bohr, V.A. Pathways and functions of the Werner syndrome protein. Mech. Ageing Dev. 126, 7986 (2005). 83. Yang, D.Q., Halaby, M.J., Li, Y., Hibma, J.C. & Burn,P. Cytoplasmic ATM protein kinase: an emerging therapeutic target for diabetes, cancer and neuronal degeneration. Drug Discov. Today 16, 332338 (2011). 84. Basel-Vanagaite, L. etal. Expanding the clinical phenotype of autosomal dominant dyskeratosis congenita caused by TERT mutations. Haematologica93, 943944 (2008). 85. Mercer, J.R. etal. DNA damage links mitochondrial dysfunction to atherosclerosis and the metabolic syndrome. Circ. Res. 107, 10211031 (2010).
5. 6. 7. 8. 9. 10. 11. 12. 13.
14. 15.
16. 17.
18.
19. 20. 21. 22. 23.
24. 25. 26. 27. 28. 29.
VOLUME 13 | JUNE 2012 | 403
PERSPECTIVES
86. Passos, J.F. etal. Feedback between p21 and reactive oxygen production is necessary for cell senescence. Mol. Syst. Biol. 6, 347 (2010). 87. Kovalenko, O.A. etal. A mutant telomerase defective in nuclear-cytoplasmic shuttling fails to immortalize cells and is associated with mitochondrial dysfunction. Aging Cell 9, 203219 (2010). 88. Haendeler, J. etal. Mitochondrial telomerase reverse transcriptase binds to and protects mitochondrial DNA and function from damage. Arterioscler. Thromb. Vasc. Biol. 29, 929935 (2009). 89. Guo, N. etal. Short telomeres compromise -cell signaling and survival. PLoS ONE 6, e17858 (2011). 90. Sarin, K.Y. etal. Conditional telomerase induction causes proliferation of hair follicle stem cells. Nature436, 10481052 (2005). 91. Lee, J. etal. TERT promotes cellular and organismal survival independently of telomerase activity. Oncogene 27, 37543760 (2008). 92. Choi, J. etal. TERT promotes epithelial proliferation through transcriptional control of a Myc- and Wntrelated developmental program. PLoS Genet. 4, e10 (2008). 93. Cong, Y. & Shay, J.W. Actions of human telomerase beyond telomeres. Cell Res. 18, 725732 (2008). 94. Flores, I., Cayuela, M.L. & Blasco, M.A. Effects of telomerase and telomere length on epidermal stem cell behavior. Science 309, 12531256 (2005). 95. Vidal-Cardenas, S.L. & Greider, C.W. Comparing effects of mTR and mTERT deletion on gene expression and DNA damage response: a critical examination of telomere length maintenanceindependent roles of telomerase. Nucleic Acids Res. 38, 6071 (2010). 96. Santos, J.H., Meyer, J.N., Skorvaga, M., Annab, L.A. & Van Houten, B. Mitochondrial hTERT exacerbates free-radical-mediated mtDNA damage. Aging Cell 3, 399411 (2004). 97. Sharma, N.K. etal. Human telomerase acts as a hTR-independent reverse transcriptase in mitochondria. Nucleic Acids Res. 40, 712725 (2011). 98. Ahmed, S. etal. Telomerase does not counteract telomere shortening but protects mitochondrial function under oxidative stress. J. Cell Sci. 121, 10461053 (2008). 99. Park, J.I. etal. Telomerase modulates Wnt signalling by association with target gene chromatin. Nature460, 6672 (2009). 100. Strong, M.A. etal. Phenotypes in mTERT+/ and mTERT/ mice are due to short telomeres, not telomere-independent functions of telomerase reverse transcriptase. Mol. Cell. Biol. 31, 23692379 (2011). 101. Patel, A.Y., McDonald, T.M., Spears, L.D., Ching,J.K. & Fisher, J.S. Ataxia telangiectasia mutated influences cytochrome c oxidase activity. Biochem. Biophys. Res. Commun. 405, 599603 (2011). 102. Ambrose, M., Goldstine, J.V. & Gatti, R.A. Intrinsicmitochondrial dysfunction in ATM-deficient lymphoblastoid cells. Hum. Mol. Genet. 16, 21542164 (2007). 103. Pallardo, F.V. etal. Mitochondrial dysfunction in some oxidative stress-related genetic diseases: ataxiatelangiectasia, down syndrome, Fanconi anaemia and Werner syndrome. Biogerontology 11, 401419 (2010). 104. Varela, I. etal. Accelerated ageing in mice deficient in Zmpste24 protease is linked to p53 signalling activation. Nature 437, 564568 (2005). 105. Lombard, D.B. etal. DNA repair, genome stability, and aging. Cell 120, 497512 (2005). 106. Wenz, T. Mitochondria and PGC-1 in aging and ageassociated diseases. J. Aging Res. 2011, 810619 (2011). 107. Liu, J. etal. Bmi1 regulates mitochondrial function and the DNA damage response pathway. Nature 459, 387392 (2009). 108. Salih, D.A. & Brunet, A. FoxO transcription factors in the maintenance of cellular homeostasis during aging. Curr. Opin. Cell Biol. 20, 126136 (2008). 109. Kim, W.Y. & Sharpless, N.E. The regulation of INK4/ ARF in cancer and aging. Cell 127, 265275 (2006). 110. von Zglinicki, T. Oxidative stress shortens telomeres. Trends Biochem. Sci. 27, 339344 (2002). 111. Passos, J.F., Saretzki, G. & von Zglinicki, T. DNA damage in telomeres and mitochondria during cellular senescence: is there a connection? Nucleic Acids Res. 35, 75057513 (2007). 112. Sablina, A.A. etal. The antioxidant function of the p53 tumor suppressor. Nature Med. 11, 13061313 (2005). 113. Matoba, S. etal. p53 regulates mitochondrial respiration. Science 312, 16501653 (2006). 114. Bae, B.I. etal. p53 mediates cellular dysfunction and behavioral abnormalities in Huntingtons disease. Neuron 47, 2941 (2005). 115. Saleem, A., Adhihetty, P.J. & Hood, D.A. Role of p53 in mitochondrial biogenesis and apoptosis in skeletal muscle. Physiol. Genom. 37, 5866 (2009). 116. Mendrysa, S.M. etal. Tumor suppression and normal aging in mice with constitutively high p53 activity. Genes Dev. 20, 1621 (2006). 117. Matheu, A. etal. Delayed ageing through damage protection by the Arf/p53 pathway. Nature 448, 375379 (2007). 118. Ventura, N. etal. p53/CEP-1 increases or decreases lifespan, depending on level of mitochondrial bioenergetic stress. Aging Cell 8, 380393 (2009). 119. Niedernhofer, L.J. etal. A new progeroid syndrome reveals that genotoxic stress suppresses the somatotroph axis. Nature 444, 10381043 (2006). 120. Hinkal, G. & Donehower, L.A. How does suppression of IGF-1 signaling by DNA damage affect aging and longevity? Mech. Ageing Dev. 129, 243253 (2008). 121. Devin, J.K. & Young, P.P. The effects of growth hormone and insulin-like growth factor-1 on the aging cardiovascular system and its progenitor cells. Curr.Opin. Investig. Drugs 9, 983992 (2008). 122. Keyes, W. M. & Mills, A. A. p63: a new link between senescence and aging. Cell Cycle 5, 260265 (2006). 123. Su, X. & Flores, E.R. TAp63: the fountain of youth. Aging (Albany NY) 1, 866869 (2009). 124. Finkel, T., Deng, C.X. & Mostoslavsky, R. Recent progress in the biology and physiology of sirtuins. Nature 460, 587591 (2009). 125. Burnett, C. etal. Absence of effects of Sir2 overexpression on lifespan in C. elegans and Drosophila. Nature 477, 482485 (2011). 126. Passos, J.F. etal. Mitochondrial dysfunction accounts for the stochastic heterogeneity in telomeredependent senescence. PLoS Biol. 5, e110 (2007). 127. Nautiyal, S., DeRisi, J.L. & Blackburn, E.H. The genome-wide expression response to telomerase deletion in Saccharomyces cerevisiae. Proc. Natl Acad. Sci. USA 99, 93169321 (2002). 128. Moiseeva, O., Bourdeau, V., Roux, A., DeschenesSimard, X. & Ferbeyre, G. Mitochondrial dysfunction contributes to oncogene-induced senescence. Mol.Cell. Biol. 29, 44954507 (2009). 129. Rubinsztein, D.C., Marino, G. & Kroemer, G. Autophagy and aging. Cell 146, 682695 (2011). 130. McCormick, M.A., Tsai, S.Y. & Kennedy, B.K. TORand ageing: a complex pathway for a complex process. Phil. Trans. R. Soc. B 366, 1727 (2011). 131. Hartl, F.U., Bracher, A. & Hayer-Hartl, M. Molecular chaperones in protein folding and proteostasis. Nature475, 324332 (2011). 132. de Jesus, B.B. etal. The telomerase activator TA-65 elongates short telomeres and increases health span of adult/old mice without increasing cancer incidence. Aging Cell 10, 604621 (2011). 133. Wenz, T., Rossi, S.G., Rotundo, R.L., Spiegelman, B.M. & Moraes, C.T. Increased muscle PGC-1 expression protects from sarcopenia and metabolic disease during aging. Proc. Natl Acad. Sci. USA 106, 2040520410 (2009). 134. Little, J.P., Safdar, A., Bishop, D., Tarnopolsky, M.A. & Gibala, M.J. An acute bout of high-intensity interval training increases the nuclear abundance of PGC-1 and activates mitochondrial biogenesis in human skeletal muscle. Am. J. Physiol. Regul. Integr. Comp. Physiol. 300, R1303R1310 (2011). 135. Timmers, S. etal. Calorie restriction-like effects of 30 days of resveratrol supplementation on energy metabolism and metabolic profile in obese humans. Cell Metab. 14, 612622 (2011). 136. Hu, J. etal. Antitelomerase therapy provokes ALT and mitochondrial adaptive mechanisms in cancer. Cell148, 651663 (2012). 137. Zoncu, R. Efeyan, A. & Sabatini, D.M. mTOR: from growth signal integration to cancer, diabetes and ageing. Nature Rev. Mol. Cell Biol. 12, 2135 (2011). 138. Hardie, D.G. Signal transduction: How cells sense energy. Nature 472, 176177 (2011).
Acknowledgements
The authors apologize to all their colleagues whose work could not be cited. Part of the work described in this article was supported by a fellowship from the Deutsche Forschungsgemeinschaft (to E.S.) and by R01 and U01 grants from the US National Institutes of Health (NIH) National Cancer Institute and the Robert A. and Renee E. Belfer Foundation. R.A.D. was supported by an Ellison Foundation for Medical Research Senior Scholar and an American Cancer Society Research Professor award.
Competing interests statement
The authors declare no competing financial interests.
FURTHER INFORMATION
Ronald A. DePinhos homepage: http://www.mdanderson. org/about-us/president-ronald-depinho-m-d-/index.html
ALL LINKS ARE ACTIVE IN THE ONLINE PDF
RESEARCH HIGHLIGHTS
C E L L M I G R AT I O N
Switching to 3D
Cell migration on two-dimensional(2D) surfaces is well characterized; in the 2Dcell migration model, cells repeatedly extend lamellipodia (actin-rich protrusions) at the leading edge, adhere to the substrate and then retract the trailing edge. This migration mode relies on highly polarized signalling that involves the second messenger phosphatidylinositol-3,4,5-trisphosphate (PtdIns(3,4,5)P3) and the RHO family members RAC1, CDC42 and RHOA, which are all enriched at the leading edge to coordinate cell protrusion. Bycontrast, the mechanisms underlying 3D migration in vivo are not well understood; Petrie et al. now show that cells can switch between lamellipodiumbased and lobopodium-based migration in 3D matrices, and that this depends on both the degree of polarization of intracellular signalling and the physical properties of the extracellular matrix (ECM). The authors first observed that many human fibroblasts migrating in dermal tissue explants formed large blunt, cylindrical protrusions called lobopodia. They then went on to study cell migration in two in vitro models of the 3D ECM, namely cell-derived matrix (CDM) and collagen, to define the mechanistic basis of lobopodium-dependent cell motility. CDM has the same mechanical properties as dermal explants as it is stiff and linearly elastic (that is, it does not undergo strain-induced stiffening), whereas collagen is soft and nonlinearly elastic (with a strain-stiffening behaviour). Fibroblasts migrating in CDM formed lobopodia, whereas they formed multiple protrusions tipped with small lamellipodia when placed in collagen. Interestingly, lobopodium-dependent migration was specific to the 3D ECM, as cells displayed the characteristic lamellipodium-based migration when placed on 2D CDM surfaces. Importantly, when
the elastic properties of the matrix dictate the mode of 3D migration
Vicky Summersby/NPG
collagen or CDM were treated to modify their elastic properties from linear to nonlinear, and vice versa, cells switched their type of migration, which suggests that the elastic properties of the matrix dictate the mode of 3D migration. Next, the authors imaged the localization of PtdIns(3,4,5)P3, and of the active forms of RAC1 and CDC42, in fibroblasts that migrated using lobopodia in 3D CDM. They observed that these signalling molecules were not polarized towards the leading edge, but were instead distributed in patches around the cell perimeter. Bycontrast, PtdIns(3,4,5)P3, RAC1 and CDC42 were concentrated at the leading edge of the small lamellipodia during migration in 3D collagen. Thus, signalling polarization is not required for lobopodium-based migration. Finally, small interfering RNA (siRNA)-mediated knockdown of RAC1 or CDC42 did not affect lobopodiummediated cell motility, indicating that RAC1 and CDC42 are dispensable for lobopodia formation. By contrast, knockdown or chemical inhibition of RHOA, its effector ROCK or the ROCK target myosin II induced a switch from lobopodium-based to lamellipodiumbased migration in 3D CDM. Thus, RHOA, ROCK and myosin II are required for lobopodium-based migration. This work identifies two modes of integrin-dependent 3D migration invitro; whether migrating cells other than fibroblasts use lobopodia, and in which physiological conditions, awaits further confirmation. Kim Baumann
ORIGINAL RESEARCH PAPER Petrie, R. J. et al. Nonpolarized signaling reveals two distinct modes of 3D cell migration. J. Cell Biol. 197, 439455 (2012)
REVIEWS
TFIIH: when transcription met DNArepair
Emmanuel Compe and Jean-Marc Egly
Abstract | The transcription initiation factor TFIIH is a remarkable protein complex that has a fundamental role in the transcription of protein-coding genes as well as during the DNA nucleotide excision repair pathway. The detailed understanding of how TFIIH functions to coordinate these two processes is also providing an explanation for the phenotypes observed in patients who bear mutations in some of the TFIIH subunits. In this way, studies of TFIIH have revealed tight molecular connections between transcription and DNA repair and have helped to define the concept of transcription diseases.
Nucleotide excision repair
(NER). The repair pathway that is used to remove the vast majority of lesions that are located on a DNA single strand, including lesions caused by ultraviolet (UV) light and cisplatin damage.
Helicases
Enzymes that move directionally along a nucleic acid phosphodiester backbone and separate two annealed nucleic acid strands by using energy derived from ATP hydrolysis.
Institut de Gntique et de Biologie Molculaire et Cellulaire, CNRS/INSERM/UdS, BP 163, 67404 Illkirch Cedex, C. U., Strasbourg, France. e-mails: compe@igbmc.fr; egly@igbmc.fr doi:10.1038/nrm3350 Published online 10 May 2012
Following gene activation, a host of proteins including RNA polymerase II (Pol II), general transcription factors, cofactors and chromatin-remodelling factors are assembled around the promoter and contribute to the accurate expression of protein-coding genes. In para llel, to maintain genome integrity and to ensure the continuation of transcription, DNA lesions that are caused by genotoxic agents have to be eliminated. This implies that there must be connections between the seemingly disparate events of transcription and DNA repair. A link between DNA repair and transcription was first suspected when it was found that the repair of DNA damage (in particular, ultraviolet (UV) lightinduced cyclobutane pyrimidine dimers) is much faster and more efficient on the coding strand of active genes than on the other parts of the genome1,2. Several years later, the multi-protein complex transcription initiation factor TFIIH (BOX1) was found to be indispensable during basal transcription and during nucleotide excision repair (NER) of DNA damage, thus revealing the functional link between these two processes3,4. The more recent demonstration that NER factors localize to the promoters of activated genes 5,6 strengthens the idea of an interplay between NER and transcription and raises questions about how these factors might act in proximity to ongoing transcription in the absence of exogenous genotoxic attack. The story of TFIIH started in 1989, when a factor named general transcription factor- (purified from rat liver)7 or basic transcription factor 2 (BTF2, purified from HeLa cells)8 was characterized as an indispensable transcription factor invitro. This factor was also isolated from yeast (termed yeast Pol II transcription factor b)9 and was finally designated TFIIH in 1992 (REF. 10). Peptide microsequencing revealed that TFIIH contains ERCC3 (excision repair cross complementing 3; also known asXPB)
and ERCC2 (also known as XPD)3,11, which are two potential helicases involved in DNA repair12,13. Its funda mental role during NER has since been established, and TFIIH (or some of its subunits) has also been shown to affect other cellular processes. For example, the CAK (cyclin-dependent kinase (CDK)-activating kinase) subcomplex of TFIIH1419 has been implicated in cell cycle control during the transition from G2 to Mphase2022. Indeed, this subcomplex is responsible for the activating phosphorylation of several kinases, including CDK1, CDK2, CDK4 and CDK6 (REFS 2327), and these phosphorylation events are required for cell cycle progression (reviewed in REF. 28). Furthermore, the XPD subunit, which associates with the TFIIH core complex, also interacts with MMS19 in the MMXD complex, which is required for chromosome segregation29. The demonstration that TFIIH has various cellular functions was greatly facilitated by the fact that mutations in its XPB, XPD and p8 (also known as TF2H5 and TTDA) subunits cause autosomal recessive disorders, including trichothiodystrophy (TTD), xeroderma pigmentosum and, in rare cases, the combined symptoms of xeroderma pigmentosum and Cockayne syndrome (TABLE 1). Whereas these diseases were initially defined as DNA repair syndromes, it seems that some of their clinical features cannot be explained by DNA repair defects alone and might be due to deficiencies in transcription. With this in mind, it has been hypothesized that such mutations in TFIIH might disturb the molecular architecture of this complex and consequently might affect its positioning within intermediate complexes that alter transcription and/or NER. In this Review, we focus on the molecular roles of TFIIH and its functional partners in both DNA repair and transcription. We also discuss the insights that have
VOLUME 13 | JUNE 2012 | 343
REVIEWS
Box 1 | The structure of TFIIH
Mammalian TFIIH is a CAK multiprotein complex Cyclin H with ten subunits that Kinase CDK7 consists of two main functional subcomplexes: MAT1 the core complex, which 5 to 3 ATPis composed of six dependent XPD helicase subunits (xeroderma 3 to 5 ATPpigmentosum group B dependent complementing protein XPB helicase (XPB), p62, p52, p44, p34 p44 and p8); and the CAK p34 (cyclin-dependent kinase p8 (CDK)-activating kinase) Core complex, which is p52 p62 composed of CDK7, cyclin H and MAT1 (REFS15,18,19,192) (see the figure). The core and CAK subcomplexes are bridged by the XPD subunit, which Nature Reviews | Molecular Cell51,80 Biology interacts with p44 and MAT1 of the core or the CAK subcomplex, respectively . Electron microscopy and image processing of the TFIIH complex revealed that it formsa ring-like structure with a central cavity that possibly interacts with DNA193,194. TFIIHhas several intrinsic enzymatic activities: CDK7 is a cyclin-dependent kinase and XPB and XPD are thought to act as ATP-dependent helicases of opposite polarities. Interestingly, in yeast, the subunit Ssl1 (suppressor of stemloop protein 1, which is the homologue of p44) is an E3 ubiquitin ligase189, and TFIIH regulates cullin neddylation via its RNA polymerase II transcription factor b (Tfb3) subunit (which is the homologue of mammalian MAT1)195.
been gained into how different mutations in TFIIH might disrupt NER and transcription and thereby give rise to the phenotypes observed in different disorders.
repair-initiating complex is established. Structural studies of Rad4, which is the yeast homologue of XPC, have revealed that XPC covers only the 3 side of a DNA lesion and leaves the 5 side almost completely free36. XPC (or Rad4 in yeast) thermodynamically destabilizes and distorts the DNA double helix, and this distortion is crucial for XPC recognition and the recruitment of additional NER factors37,38. In some cases, UV lightdamaged DNAbinding protein (UV-DDB), which consists of DDB1 and DDB2 (REFS 39,40), promotes the recognition of lesions41,42. Interestingly, it has been proposed that preferential UVDDB accumulation on internucleosomal DNA leads to the ubiquitylation of XPC by cullin 4A (CUL4A) ligase43. In this model, DDB2 (also known as XPE) would bind damaged DNA, and CUL4A ligase, associated with DDB1, would target XPC, DDB2 and/or nearby histones for ubiquitylation. This process is thought to reposition XPC and to result in the recruitment of further NER factors44. Once XPC is bound to the lesion, other XPCbinding partners such as centrin 2 (REF. 45) and/or RAD23 homologue B (RAD23B)46 might stabilize XPC. The correct positioning of XPC is thought to be important for the subsequent recruitment of TFIIH and the excision of damaged DNA. However, removal of the damaged oligonucleotide can occur invitro in the absence of RAD23B, centrin 2 and the UVDDB complex47. This does not exclude the possibility that these factors might contribute to the optimal positioning of XPC in a cellular context. And, when the XPCdamaged DNA intermediate complex is not accurately positioned (as is observed with mutated forms of XPC), XPC is rapidly degraded by a proteasome-independent mechanism48. TFIIH opens damaged DNA for excision. When correctly bound to damaged DNA, the carboxyterminal domain of XPC might adopt a three-dimensional structure that enables the recruitment of TFIIH through the inter action of XPC with at least two subunits of TFIIH: p62, which interacts with both the Cterminal and the amino terminal regions of XPC; and XPB, which associates with the Cterminal region of XPC48. TFIIH then mediates the excision of the damaged DNA, and several studies have aimed to elucidate how the different sub units of TFIIH mediate NER. For example, the concomitant presence of XPB and XPD helicases within TFIIH raises the question of how these subunits regulate NER. A model was initially proposed in which each helicase acts on both sides of the lesion to unwind the damaged DNA49. In an alternative model, the two helicases were proposed to bind at the lesion and move on individual DNA strands, so that blockage of either helicase could discriminate between the damaged strand and the undamaged strand 50. Howeve r, although the helicase activity of XPD is clearly required for NER5153, the helicase activity of XPB turned out to be dispensable, which suggests that only one of the TFIIH helicase activities is required during NER. Nonetheless, the ATPase activity of XPB participates in the NER pathway by anchoring TFIIH to the damaged chromatin54. The crystal structure of Archaeoglobus fulgidus XPB was preponderant for understanding the role of the XPB ATPase activity in NER55. In addition to
TFIIH in repair The main role of TFIIH in NER (BOX 2) is to open the DNA around the lesion and thereby allow the excision of the damaged oligonucleotide and its replacement by a new DNA fragment. The advances in our understanding of how TFIIH affects NER have been driven by: first, invitro reconstitution assays that use damaged DNA (which contains either a cis-platin adduct or a thymine dimer) as a substrate together with crude cellu lar extracts30 or recombinant proteins31; and second, a sophisticated invivo local UV light irradiation technique coupled with immunofluorescence staining32. These technical advances have indeed helped to define how TFIIH functions after a DNA lesion has been recognized by either the global genome repair (GGR) pathway or the transcription-coupled repair (TCR) pathway of NER; we discuss the GGR pathway here and refer readers to a review of the TCR pathway (REF. 33).
XPC-mediated recognition of DNA lesions. In the GGR pathway (FIG. 1), XPC (xeroderma pigmentosum groupC complementing protein) initially recognizes the damaged site and prepares this site for TFIIH recruitment. XPC can rapidly detect various DNA lesions that do not share any common chemical structures34; it then promotes bending of the double helix35, thus forming a transient recognition intermediate before a more stable
344 | JUNE 2012 | VOLUME 13
REVIEWS
Table 1 | Composition of the human TFIIH complex
TFIIH subcomplex Human
Core XPB
Yeast
Ssl2
Function
3 to 5 ATP-dependent helicase
Human genetic disorders

Trichothiodystrophy and combined xeroderma pigmentosum and Cockayne syndrome
p62
Tfb1
Structural function and interacts with transcription factors and NER factors Regulates the XBP ATPase activity E3 ubiquitin ligase (in yeast) Structural function and strong interaction with p44 Regulates the XBP ATPase activity 3 to 5 ATP-dependent helicase and forms a bridge between the CAK and the core Kinase Modulates the CDK7 kinase activity CAK stabilization and regulates cullin neddylation (in yeast) Trichothiodystrophy Trichothiodystrophy, xeroderma pigmentosum and combined xeroderma pigmentosum and Cockayne syndrome
p52 p44 p34 p8 XPD XPD
Tfb2 Ssl1 Tfb4 Tfb5 Rad3
CAK
CDK7 Cyclin H MAT1
Kin28 Ccl1 Tfb3
CAK, cyclin-dependent kinase-activating kinase subcomplex; Ccl1, cyclin C like 1; CDK7, cyclin-dependent kinase7; MAT1, mnage trois 1; NER, nucleotide excision repair; Ssl, suppressor of stemloop protein; Tfb, RNA polymerase II transcription factor b; XPB, xeroderma pigmentosum group B complementing protein.
an ATP-binding site (which is located within the helicase motif Ia), there are two other structural motifs in XPB, which are termed RED and Thumb, the latter of which binds to DNA in a sequence-independent manner. It has been suggested that a large conformational change, which is driven by ATP hydrolysis, brings the RED and Thumb domains of XPB in close proximity to each other, and that this conformation favours the anchoring of TFIIH to DNA56. The XPB ATPase activity is regulated by XPC, as well as by the p8 and p52 subunits of TFIIH. The contribution of p52 was revealed by studies of the marionette (mrn) gene (which encodes p52) in Drosophila melanogaster, as in these studies mutations that destabilize the interaction between p52 and XPB consequently reduced the ATPase activity of XPB57. The study of XPD helicase function during NER was facilitated by the numerous XPD mutations that are found in patients with xeroderma pigmentosum and
Box 2 | The nucleotide excision repair pathway

The nucleotide excision repair (NER) pathway removes bulky DNA adducts that are caused by ultraviolet (UV) light irradiation, genotoxic chemicals and reactive metabolic by-products, each of which induce a strong distortion of the DNA double helix through covalent modification of one DNA strand. NER occurs by the sequential assembly of repair proteins at the site of DNA damage32,196. There are two NER subpathways: the global genome repair (GGR) pathway, which is initiated by xeroderma pigmentosum group C complementing protein (XPC) and removes DNA lesions from anywhere in the genome; and the transcription-coupled repair (TCR) pathway, whereby DNA damage that is located in the transcribed strand of active genes is recognized by elongating RNA polymerase II complexes33. After a lesion has been recognized, the subsequent steps in NER are identical for GGR and TCR. The ability of both pathways to detect DNA damage is crucial for reducing the risk of mutations that can result from incorrect or incomplete replication.
TTD and that diminish NER activity. Mutations in the helicase motifs of XPD (which abrogate the unwinding activity of XPD) or in its Cterminal end (which weaken the interaction with the p44 subunit) result in a decreased ability of TFIIH to open damaged DNA, which is a crucial step in NER51,52. Crystal structure data of XPD homologues from Sulfolobus acidocaldarius, Sulfolobus tokodaii and Thermoplasma acidophilum5860 revealed the presence of an arch domain and a 4FeS cluster in the helicase domain HD1. These structural results suggest that the channel under the arch, which is formed by HD1, the arch domain and the 4FeS domain, forms a passageway for the translocation of single-stranded DNA (ssDNA) during XPD-mediated unwindin g of the damaged DNA. Interestingly, in addition to its role in DNA opening, it has been proposed that XPD participates in DNA damage recognition and verification61. As XPB function in NER requires its ATPase activity but not its helicas e activity, it has been suggested that XPB might act as a wedge55 that uses ATP to keep the two strands of DNA around the lesion apart, thereby allowing XPD to unwind the DNA56. The putative role of XPD in damage sensing has been supported by the characterization of a hairpin structural motif in UvrB (the prokaryotic homologue of XPD), and this motif is essential for DNA binding and damage processing62,63. In the current model, XPB-mediated opening of the damaged DNA would allow correct binding of XPD to the DNA, and XPD would then utilize its helicase activity to verif y the DNA damage and ensure that the backbone distortion is not the result of an unusual DNA sequence. This process has been termed enzymatic proofreading and supports a bipartite damage recognition model in
VOLUME 13 | JUNE 2012 | 345
REVIEWS
which the function of XPCRAD23B is limited to the recognition of a DNA backbone distortion, whereas XPD is required to verify the presence of DNA damage through its helicase activity64. Further structural
DNA resynthesis Genomic DNA Nucleosome
analysis of the complex that is formed between XPD and damaged DNA should be undertaken to clearly demonstrate that human XPD participates in damage recognition.
Dual incision Antitumoural drugs, UV light
Damaged DNA CAF1 DDB1 XPE Post-incision repair synthesis and ligation Recognition of DNA lesion TFIIH CAK XPD Core XPC RAD23B
DNA ligase I DNA ligase III XRCC1
FEN1
XPB p8 p52 ATP DNA polymerase Release of damaged oligonucleotide 30 nt ATP RFC PCNA ERCC1 XPF Removal of core TFIIH XPG Opening of damaged DNA by TFIIH ADP
XPA RPA
Removal of CAK from TFIIH CAK
Figure 1 | TFIIH opens DNA to allow the incision and excision of damaged oligonucleotides. During global genome repair (GGR), which occurs following exposure to genotoxic agents such as ultraviolet (UV) light or antitumoural Nature Reviews |pigmentosum Molecular Cell Biology drugs, a lesion on the DNA (yellow star) is initially recognized through the binding of xeroderma group C complementing protein (XPC)RAD23B and/or the lesion sensor UV lightdamaged DNA-binding protein (UVDDB) complex (which contains DDB1 and XPE). The XPCRAD23B complex mediates the recruitment of the transcription initiation factor TFIIH to the damaged DNA, and this promotes opening of the DNA in an ATP-dependent manner. The unwinding of the DNA requires the helicase activity of XPD and the ATPase activity of XPB, the latter of which is regulated by XPC, as well as by the p52 and p8 subunits of TFIIH. The next step involves the recruitment of XPA, which promotes the release of the CAK (cyclin-dependent kinase (CDK)-activating kinase) subcomplex, and the association of replication protein A (RPA) with the single-stranded damaged DNA. The dissociation of CAK is a prerequisite for the enlargement of the DNA opening that is required to promote the recruitment of the XPFexcision repair cross complementing 1 (ERCC1) complex and XPG and the release of XPCRAD23B. XPFERCC1 makes an incision in the damaged DNA strand at the 5side of the bubble. This allows the concomitant incision at the 3 side by XPG and the release of the core TFIIH. DNAresynthesis begins when replication factor C (RFC) loads proliferating cell nuclear antigen (PCNA) onto the damagedDNA to accommodate DNA polymerase (Pol ), Pol and/or Pol . The final ligation step can be carried out by the DNAligaseIflap endonuclease 1 (FEN 1) complex or by the DNA ligase IIIX-ray repair cross-complementing protein1 (XRCC1) complex, depending on the cell cycle stage. Finally, chromatin assembly factor 1 (CAF1) mediates the arrival of histones H3 and H4 to allow nucleosome reassembly on the repaired DNA. nt, nucleotide.
REVIEWS
Transfer RNAs
(tRNAs). The ribonucleic acids that transport specific amino acids to the ribosome for incorporation into the growing polypeptide chain.
In addition to the function of XPB and XPD in NER, special attention has recently focused on the p8 subunit of TFIIH. After 10 years of investigations into the human TTDA disorder, which is a particular form of TTD65 that results from defective TFIIH, researchers were unable to identify the mutations that cause this disorder. However, studies in yeast66 helped to characterize the tenth sub unit of TFIIH, termed Tfb5, which is the homologue of p8. Mutations in this subunit turned out to result in the TTDA disorder67. Strikingly, although p8 is dispensable for transcription (at least invitro), the role of this subunit is crucial in NER, during which p8 stimulates the ATPase activity of XPB in a DNA-dependent manner and promotes the recruitment of XPA68. XPB, with the help of p8, is thereby thought to act as an ATPdriven motor that supplies the energy that is required to reorganize the intermediate DNA repair complex and thereby supports the repositioning of XPCRAD23B and the unwinding of DNA by XPD47. It seems that p8 also acts as a stabilizer of TFIIH, as the cellular concentration of TFIIH decreases considerably when p8 is mutated6769. Live-cell imaging studies have indicated that there are two distinct kinetic pools of p8: one with slow mobility that is bound to TFIIH and a free fraction that is homodimeric70 and shuttles between the nucleus and the cytoplasm. Following UV light irradiation, the free dimeric p8 fraction might shift towards a more stabl e complex with TFIIH through interaction with the Cterminal end of p52, which shares a common fold with p8 (REF. 71). When TFIIH is correctly bound to the XPCdamaged DNA complex, replication protein A (RPA) is recruited, which protects the ssDNA from enzymatic hydrolysi s and prepares it for DNA resynthesis31. RPA arrives together with XPA, which seems to stabilize the TFIIH XPCdamaged DNA ternary complex7275 through an unclear mechanism. XPA associates with the Nterminal moiety of XPC that has been repositioned by TFIIH; this results in the expansion of the DNA bubble around the damaged site76. In addition to the potential function of XPA in maintaining this open DNA structure, the Cterminal region of XPA mediates the release of CAK from the TFIIH core subcomplex and the arrival of other NER-specific factors such as XPG (also known
as ERCC5) and XPFERCC1 (REF. 77). Experiments have indicated that CAK is not required for DNA repair and even identified CAK as an inhibitor of NER activity7780. Whether post-translational modifications of XPA (such as its deacetylation by sirtuin 1 (SIRT1))81 are involved in CAK removal and/or in other NER steps remains to be established. An alternative and/or complementary step of this model proposes that XPC would be positioned at a region with disrupted base pairing upstream from the lesion and would be further translocated together with XPA by XPD helicase at the damaged site82. This suggests that XPD might have an additional role in translocation during the damage recognition process invivo. Although not without shortcomings83, a mathematical model has validated that these roles of XPC and XPD are possible by comparing a model of random order assembly and kinetic proofreading with a sequential assembly model84. The release of the damaged oligonucleotide. The generation of an open structure on the damaged DNA complex provides a platform for the arrival of the structure-specific endonucleases XPFERCC1 and XPG that are responsible for carrying out incisions 5 and 3 of the damaged site, respectively76. Concomitantly to an incision being made 5 of the damaged site, XPC, followed by XPA and TFIIH, are released from the DNA template and are recycled. XPG and XPFERCC1 endonucleases remain on the gapped DNA intermediate together with RPA, which coats the ssDNA region. Next, proliferating cell nuclear antigen (PCNA) and replication factorC (RFC) are positioned at the 3 primer template, and this forces the release of XPF endonuclease in an ATP-dependent manner85. After the recruitment of RFC, PCNA and DNA polymerase, repair synthesis is initiated, which precedes 3 DNA cleavage by XPG86. Following the release of XPG, RPA and the excised fragment of approximately 30nucleo tides87,88, the post-incision stage of NER consists of gapfilling DNA synthesis, ligation and the restoration of chromatin structure. Recent findings have revealed the complexity of the repair synthesis and ligation, which seem to depend at least partly on the growth state of cells89. Indeed, it seems that in non-proliferating cells, the preference is for DNA polymerase (Pol ) and Pol to synthesize the new DNA, whereas in cycling cells Pol is also required90. The nick that is formed when DNA has been resynthesized is sealed by either DNA ligase III Xray repair cross-complementing protein 1 (XRCC1) in quiescent cells or by both DNA ligase IIIXRCC1 and DNA ligase Iflap endonuclease1 (FEN1)79,85 in dividing cells91. After ligation has occurred, chromatin assembly factor1 (CAF1) mediates the arrival of histones H3 and H4 (REFS 92,93). This suggests that factors such as CAF1 ensure that nucleosome reassembly and/or repositioning on the naked DNA starts only after DNA repair hasended. Importantly, there are several close connections between TFIIH and the different partners that are required to eliminate DNA damage. As an example, XPG strengthens the interaction between core TFIIH and the CAK subcomplexes, and XPG forms a stable
VOLUME 13 | JUNE 2012 | 347
Box 3 | The transcription of eukaryotic protein-coding genes

Transcription is the molecular process by which a complementary RNA copy of a DNA sequence is made. In eukaryotes, three different types of RNA polymerases activate the transcription of a distinct class of RNAs: RNA polymerase I (Pol I) transcribes ribosomal RNAs (rRNAs); Pol II transcribes small regulatory RNAs and RNAs that will become mRNAs; and Pol III transcribes small RNAs such as transfer RNAs (tRNAs). Transcription by Pol II proceeds through sequential steps, which include chromatin remodelling, assembly of the pre-initiation complex (PIC), opening of the promoter, formation of the first phosphodiester bond, promoter clearance, elongation and termination. Transcription initiation requires at least six general transcription factors (TFIIA, TFIIB, TFIID, TFIIE, TFIIF and TFIIH) that allow RNA synthesis and can be further regulated by activators and repressors. Although emerging evidence indicates that the composition of PIC is not universal but promoter-dependent, invitro experiments have led to a model in which basal transcription factors are sequentially assembled with Pol II to generate the PIC8,197202.
REVIEWS
complex with TFIIH94. Furthermore, it was found that mutation of the Cterminal end of XPB prevents the 5 incision that is triggered by the XPFERCC1 endo nuclease complex95. Similarly, XPC mutations delay the recruitment of TFIIH to damaged DNA48. Thus, there seems to be an intimate relationship among the sensing of damaged DNA, the recruitment of TFIIH and the preparation of DNA for repair before the enzymes that mediate resynthesi s are recruited.
Nucleosome NR Co-repressor complex H3K9me3 TSS Mediator complex Co-activator complex Ligand Ubiquitylation machinery (E1, E2 and E3 ligases) Co-repressor degradation H3K4me3 H3K9ac H4K16ac IIF CTD TFIIH ATP Phosphorylation IIB IIF IIA IID Pol II IIE XPA RPA XPG ERCC1 XPF Histone H1
TFIIH in transcription Initially identified as an essential factor for Pol II-mediated transcription79, TFIIH also participates in the transcription of ribosomal RNA (rRNA) by Pol I9698 and probably also in the synthesis of 5S rRNA, tRNA andother small RNAs that are transcribed by Pol III99,100. Most of the transcription studies have focused on PolII (BOX 3), which transcribes protein-encodinggenes, most small nuclear RNAs (snRNAs) and microRNAs (miRNAs)101,102. Although it was suggested that a pre assembled holoenzyme complex that contains all the transcription components might be recruited to promoters, the initiation of transcription by Pol II actually seems to result from the sequential recruitment of general transcription factors103107. Typically, TFIID binds to the promoter, followed by TFIIA and TFIIB, which stabilize promoter-bound TFIID. After the recruitment of Pol II and TFIIF, a stable complex forms at the promoter together with TFIIDTFIIATFIIB, and this drives the association of TFIIE and the subsequent entry of TFIIH (FIG. 2).
Cooperation between XPB and CDK7 in basal transcription. When the pre-initiation complex (PIC) has been established, the ATP-dependent helicase activity of XPB108 is required for promoter opening109,110 and promoter escape111113. This role of XPB can be regulated by transcription factors, as illustrated by the actions of FBP (FUSE-binding protein) and FIR (FBP-interactin g repressor; also known as PUF60), which stimulate or inhibit, respectively, XPB helicase activity during the regulation of c-MYC gene transcription114. After the establishment of this open complex, transcription is initiated; this intimately depends on the phosphorylation status of the C-terminal domain (CTD) of the largest subunit of Pol II15,115119. The human Pol II CTD contains 52 heptad repeats that can be phosphorylated on Ser2, Ser5 and Ser7. Phosphorylation of Ser5 by the CDK7 kinase subunit of TFIIH controls the initiation of transcription120 and enhances the association of the Pol II CTD with the 7-methylguanosin e (m7G) RNA capping machinery 121,122. This TFIIH kinase activity towards the CTD of Pol II can be modulated by different factors, including MAT1 (mnage trois 1) and cyclin H, which are two binding partners of CDK7 within the CAK subcomplex18,19. In particular, it has been shown that although MAT1 deletion reduces Pol II Ser5 phosphorylation123, cyclin H that has been phosphorylated by CDK8 of the Mediator complex can repress CDK7 activity124. Pol II CTD phosphorylation by TFIIH also requires the contribution of
Co-activator complex Mediator complex IID Sequential recruitment of GTFs IIB IIE IIA Pol II CDK7
Sequential recruitment of NER factors
TFIIH: Phosphorylates CTD, XPC NRs and other TFs Opens DNA CSB
CSB
ERCC1 XPF XPA XPC XPG RPA IIE IIA IIB XPC Promoter escape
Ribosomal RNA
(rRNA). The ribonucleic acid element of the ribosome, which orchestrates protein synthesis.
Small nuclear RNAs

(snRNAs). Small ribonucleic acids, which are located in the nucleus and are involved in different molecular processes such as transcriptional regulation and RNA splicing.
Ubiquitylation machinery and proteasome
RNA Degradation of NRs and other TFs
pTEFb
MicroRNAs
(miRNAs). Short ribonucleic acids that are posttranscriptional regulators able to recognize complementary sequences on target mRNA transcripts.
pTEFb

348 | JUNE 2012 | VOLUME 13 2012 Macmillan Publishers Limited. All rights reserved www.nature.com/reviews/molcellbio
REVIEWS
Spliceosomal snRNAs
Small ribonucleic acids that participate in the removal of introns from pre-mRNA.
Nuclear receptors
Ligand-dependent and -independent transcription factors that are highly conserved evolutionarily from invertebrates to higher organisms. The nuclear receptor superfamily includesreceptors for thyroidand steroid hormones, retinoids and vitaminD, as well as orphan receptors of unknownligands.
Ubiquitinproteasome machinery
A selective system of protein degradation. This first requires the ubiquitin conjugation of the target protein via three types of enzymes: E1 (ubiquitinactivation enzyme), E2 (ubiquitin-conjugation enzyme) and E3 (ubiquitin ligase). Polyubiquitylated substrates are then recognized anddegraded by the 26Sproteasome in an ATP-dependent manner.
non-coding RNAs, such as B2 RNA (which specifically represses PolII CTD phosphorylation by TFIIH)125 and the U1snRNA (which is a core splicing component that stimulates the Pol II CTD kinase activity of CDK7 )126. CDK7 also phosphorylates Ser7 of the PolII CTD127129, and this seems to be functionally important for the processing of spliceosomal snRNAs130. Phosphorylation of Ser2 by CDK9 (REFS131,132), which is a cyclin-dependen t kinase associated with positive transcription elongation factorb (pTEFb), results in transcription-coupled recruitment of 3-end processing factors122,133. In addition to these PolII CTD kinases and others (such as CDK2 (REF. 134), CDK8 (REF. 135), extracellular signal-regulated kinase1 (ERK1) and ERK2 (REF.136) or DNA-dependent protein kinase137,138), the Pol II CTD phosphorylation status requires the action of CTDphosphatases, such as TFIIF-associated CTD phosphatase 1 (FCP1; also known as CTDP1) which targets Ser2 (REFS131,139,140) and PolII-associated protein 2 (RPAP2), which targets Ser5 (REF.141). These different Pol II CTD-modification states, which include bivalent marks, are postulated to define a spatiotemporal code that instructs ordered engagement of PolII with functional complexes at various stages of the transcription cycle119,142144 and are probabl y initiated by TFIIH. TFIIH and nuclear receptors. In addition to its basal functions, TFIIH can modulate the activity of several transcriptional regulators, including p53 (REF. 145), herpe s simplex virion protein VP16 (REF. 146), Epstein Barr nuclear antigen 2 (EBNA2)147, hepatitis B virus (HBV) X protein (HBX)148, FIR149 or its D. melanogaster orthologue Half pint (HFP)150, as well as nuclear receptors. The interaction between TFIIH and each nuclear receptor is specific151 and either occurs in the absence of ligand as observed for retinoic acid receptor-1 (RAR1) and RAR 152,153, androgen receptor 151,154,
Figure 2 | TFIIH is an essential factor of transcription initiation. Gene activation
requires a large number of co-regulatory complexes with various functions and enzymatic activities203. Typically, genes are maintained in a latent or silenced state by co-repressor complexes that promote chromatin condensation, by histone H1 and histone post-translational modifications such as trimethylation of histone H3 at Lys9 (H3K9me3) and histone acetylation (H3K9ac) and methylation of the CpG islands near to the potential transcription start site (TSS). In the case of nuclear receptor (NR) responsive genes, gene activation is initiated following ligand induction and this promotes the removal and degradation of co-repressor complexes and the recruitment of co-activator complexes, including factors with chromatin remodelling activity, the Mediator complex and factors that are involved in RNA processing. The general transcription factors (GTFs) TFIID (which contains TATA box-binding protein), TFIIA, TFIIB (which stabilizes TFIID), RNA polymerase II (PolII), TFIIF (which anchors Pol II to the pre-initiation complex (PIC)), TFIIE and finally TFIIH are sequentially recruited. The ATPase-dependent helicase activity of XPB within TFIIH allows promoter opening. The cyclin dependent kinase 7 (CDK7) subunit of TFIIH phosphorylates the carboxy-terminal domain (CTD) of the largestsubunit of Pol II, nuclear receptors and possibly other transcription factors. Thenucleotide excision repair (NER) factors (pigmentosum group C complementing protein (XPC), cockayne syndrome B protein (CSB), XPA, XPG and XPF excision repair cross complementing 1 (ERCC1)) are then sequentially recruited to the promoter. This is followed by the promoter escape of Pol II in absence of TFIIA, TFIIB, TFIIE and XPC. Productive elongation starts after phosphorylation of the CTD of Pol II by positive transcription elongation factor b (pTEFb) and the ubiquitinproteasome machinery contributes to the turnover of nuclear receptors and other co-activators.
peroxisome proliferator activated receptor- (PPAR), PPAR (also known as PPAR), PPAR1 and PPAR2 (REF.155) and thyroid hormone receptor-1 (REF.156) or the interaction occurs in response to hormone as found for oestroge n receptor-157 and thyroid hormone 158 receptor- . All nuclear receptors so far studied are phosphorylated within their A/B domain by the CDK7 subunit of TFIIH, except vitaminD receptors (VDRs), which lack a functional A/B domain159. However, TFIIH can still regulate some VDR-responsive genes such as cytochrome P450 family 24 (CYP24) by phosphorylating the VDR DNA-binding partner ETS1, which might fulfil the role of an A/B domain. Understanding the effects of nuclear receptor phosphorylation by TFIIH has been complicated by the fact that CDK7 influences diverse molecular processes160163 and that nuclear receptors are subjected to several posttranslational modifications, including phosphorylation by other kinases164. Nonetheless, phosphorylation by CDK7 is necessary for optimal nuclear receptormediated transactivation153,155,157,165. Whereas the stabilization of nuclear receptor binding to their response elements might require a phosphorylation-independent effect of TFIIH (as observed for thyroid hormone receptor1)156, phosphorylation might specifically influence the interaction of nuclear receptors with other factors. As an example, vinexin-, which is a nuclear protein that participates in actin cytoskeletal organization166, interacts with the non-phosphorylated form of RARto repress gene transcription but not with other RAR isoforms167. CDK7-mediated phosphorylation of RAR disrupts the interaction of this receptor with vinexin- and restores RAR transactivation. Furthermore, TFIIH-mediated phosphorylation can regulate the turnover of nuclear receptors by triggering their ubiquitin-mediated degradation by the proteasome (that is, the ubiquitinproteasome machinery). Indeed, it has been shown that phosphorylation of androgen receptor by CDK7 directs the specific recruitment of the mouse homologue of the E3 ubiquitin ligase MDM2 and the proteasome to the promoter of androgen receptor specific target genes151. However, impairing CDK7mediate d androgen receptor phosphorylation can also preferentially promote the recruitment of CHIP (Cterminus of heat shock protein 70 (HSP70)-interactin g protein), which is another E3 ligase, and this results in dysregulated androgen receptor turnover and thereby the disrupted transactivation of target genes. In addition to a direct relationship between TFIIH and nuclear receptors, an even higher level of complexity emerges if one considers that TFIIH might cooperate with partners of nuclear receptors. This was observed for PPAR co-activator 1 (PGC1)168, which is a meta bolic regulator and a transcriptional co-activator for several nuclear receptors169, and also for the metastasisassociated protein 1 (MTA1), which represses oestrogen receptor-driven transcription170. Interestingly, MTA1 might disrupt CAK-induced transactivation control of oestrogen receptors, which suggests that TFIIH activity can, in turn, be modulated by these factors. Such a control of TFIIH function is also well illustrated by
VOLUME 13 | JUNE 2012 | 349
REVIEWS
the influence of the transcriptional Mediator complex. This complex is a key regulator of gene expression171 that functions as an adaptor and conveys essential information from transcription factors that are bound at upstream responsive elements to the basal Pol II transcription machinery. The Mediator complex promotes the positioning of TFIIH into the PIC172 and regulates the kinase activity of TFIIH through its CDK8 kinase subunit124. Other factors also participate in nuclear receptor-mediated transactivation by targeting TFIIH, as illustrated by the action of the cochaperone Ydj1 inyeast173. Many have tried to provide explanations for the broad range of clinical features that are observed in patients with xeroderma pigmentosum and TTD. Such studies have considered the various cellular functions of TFIIH and the position of the mutations found in the different subunits of TFIIH. Although few patients have been described with XPB mutations (and this reflects the essential role of this subunit in transcription110), XPD mutations have been associated with xeroderma pigmentosum and TTD. The heterogeneity of the pheno types that arise from these XPD mutations suggests that each mutation differently affects the biochemical properties of TFIIH and consequently that each mutation might disrupt distinct steps of transcription185. Depending on the cellular context, TFIIH might establish distinct interactions with various transcription factors. Therefore, each TFIIH mutation would specifically affect some factors and not others and thus influence the expression of distinct genes185. Interestingly, the TTD mutations that are found in XPB, XPD and GTF2H5 reduce the cellular concentration of TFIIH69,186. These findings support the hypothesis that reduced levels of TFIIH in individuals with TTD might have a significant effect on genes that are highly expressed in differentiated tissues. Nevertheless, the reduced levels of TFIIH that were observed in patients do not correlate with theheterogeneity of the phenotypes, suggesting that the clinical features in TTD are not just a result of altered TFIIH stability186. Such heterogeneity might be due to the combined effects of reduced TFIIH levels and the specific effects that each mutation has on component s of the transcriptionmachinery. Most of the XPD mutations found in patients suffering from either xeroderma pigmentosum, TTD or xeroderma pigmentosum and Cockayne syndrome weaken the binding between XPD and p44, which consequently reduces XPD helicase activity during NER52,58. These mutations also disturb the architecture of TFIIH and its ability to accurately interfere with nuclear receptors. Studies have thus tried to determine whether dysfunctions in hormone-dependent transcription might contribute to the phenotypes observed in patients that bear these mutations. As an example, patients with TTD have facial features that appear prematurely aged owing to the lack of subcutaneous fatty tissue, and female patients lack breast tissue182. By using an XPD-mutated TTD mouse model that also develops hypoplasia of adipose tissue187, it was demonstrated that this mutation disrupts CDK7-mediated phosphorylation of PPARs (which are implicated in lipid metabolism and adipogenesis) and that this mutation consequently impairs the expression of PPAR target genes. This explains, at least partially, the adipose defect that is observed in patients155. Patients with TTD also develop neurological features, such as microcephaly and hypomyelination188. In this case, as TFIIH stabilizes the binding of thyroid hormone receptors to their DNA-responsive elements, the limiting amount of TFIIH in TTD cells contributes to the dysregulation of thyroid hormone receptorresponsive genes that encode particular major myelin structural components156.
www.nature.com/reviews/molcellbio 2012 Macmillan Publishers Limited. All rights reserved
The TFIIH disorders TFIIH is directly involved in the aetiology of various diseases. Genetic polymorphisms of genes that encode subunits of TFIIH (such as XPD and XPB) seem to be associated with increased cancer susceptibility in many tissues, including skin tissue, breast tissue and lung tissue174,175. In addition to genetic variations, virusencoded proteins also target TFIIH. For example, the non-structura l proteins of the Rift Valley haemorrhagic fever (RVHF) virus outcompete XPD for p44 binding in the core complex, sequester p44 along with XPB in filamentous structures and prevent the formation of TFIIH176. This leads to a rapid and drastic suppression of host cellular RNA synthesis that parallels a decrease in TFIIH cellular concentration. HBV, which is a risk factor for hepatocellular carcinoma, also acts by disrupting the expression of TFIIH components and might decrease NER activity, thus increasing genomic instability177. Mutations in GTF2H5 (which encodes p8), XPB and XPD cause the human autosomal recessive disorders xeroderma pigmentosum, TTD and the combined symptoms of xeroderma pigmentosum and Cockayne syndrome. Patients with xeroderma pigmentosum have a 1,000-fold increased frequency of skin cancer178,179. In addition, patients with xeroderma pigmentosum can develop progressive neurological degeneration, immature sexual development and dwarfism180. Several mutations in the XPD and XPB helicases of TFIIH can cause xeroderma pigmentosum, sometimes combined with Cockayne syndrome, which associates with severe cachectic dwarfism, mental retardation, microcephaly and retinal and skeletal abnormalities181. Patients with TTD typically have dry and easily brittle hair and develop sterility, short stature and various neurological defects, including mental retardation, spasticity, tremors and ataxia182. TTD is caused by mutations in XPB, XPD and p8 (REF.183). For many years, these diseases were attributed to impaired NER, as this process is reduced or even absent in cells that have been isolated from patients. However, the clinical complexity of these syndromes cannot be solely explained on the basis of a DNA repair defect and may also involve transcription deficiencies. For instance, the skin photosensitivity observed in some patients can be correlated with an NER defect, whereas other clinical features, such as sterility or lipodystrophy, could be explained by dysfunctions in hormonedependent transcriptional regulation49,184.
350 | JUNE 2012 | VOLUME 13
REVIEWS
Crosstalk between transcription and repair Given that TFIIH has various cellular functions, it has been arduous to define which phenotypes are exclusively related to deficiencies in transcription and which are caused by defects in other cellular processes such as DNA repair. This is made more difficult by the fact that other molecular connections exist between transcription and DNA repair. For a while, the relationship between these processes was discerned only by the common TFIIH subunits that are involved, which could result from a distinct requirement for XPB and XPD in each process. However, it was then shown in yeast that a ubiquitin ligase activity of TFIIH, via its p44 subunit, mediates the transcriptional response to DNA damage189. We now know that other factors that are involved in transcription are also implicated in DNA repair, and vice versa. For example, transcriptional activators (such as Gal4VP16 and RAR) can stimulate DNA repair190. Conversely, a DNA repair complex (that contains XPC) seems to also function as a co-activator for octamer-binding protein4 (OCT4; also known as POU5F1) and SOX2 in embryonic stem cells191. Finally, NER factors are recruited to active promoters and facilitate chromatin modification to regulate transcription in the absence of exogenous genotoxic attack5,6. Indeed, NER factors seem to be associated with the transcription machinery at the promoters of several activated nuclear receptordependent genes. The recruitment occurs in a sequential order following PIC assembly and is distinct from the order that is required for a repair complex6. Although NER factors are not essential for PIC formation, it is likely that NER components optimize the efficiency of transcription, as patient cell lines that are mutated in different NER factors (such as XPC, XPA or XPG) dysregulate nuclear receptordependent
1. Bohr, V.A., Smith, C.A., Okumoto, D.S. & Hanawalt,P.C. DNA repair in an active gene: removal of pyrimidine dimers from the DHFR gene of CHO cells is much more efficient than in the genome overall. Cell40, 359369 (1985). Mellon, I., Spivak, G. & Hanawalt, P.C. Selective removal of transcription-blocking DNA damage from the transcribed strand of the mammalian DHFR gene. Cell 51, 241249 (1987). Schaeffer, L. etal. DNA repair helicase: a component of BTF2 (TFIIH) basic transcription factor. Science260, 5863 (1993). Demonstrated that the helicase XPB, which is involved in NER, is closely associated with the TFIIH transcription complex, suggesting that DNA repair and transcription are functionally related. Feaver, W.J. etal. Dual roles of a multiprotein complex from S. cerevisiae in transcription and DNArepair. Cell 75, 13791387 (1993). Schmitz, K.M. etal. TAF12 recruits Gadd45a and the nucleotide excision repair complex to the promoter of rRNA genes leading to active DNA demethylation. Mol. Cell 33, 344353 (2009). Revealed that the DNA repair machinery is recruited to the promoter of active genes, keeping these promoters in a hypomethylated state. Le May, N. etal. NER factors are recruited to active promoters and facilitate chromatin modification for transcription in the absence of exogenous genotoxic attack. Mol. Cell 38, 5466 (2010). Showed the sequential recruitment ofthe NER factors XPC, XPA, RPA, XPG and XPFERCC1 to the promoters of inducible genes in the absence of exogenous genotoxic attack. These NER factors (except cockayne syndrome B protein (CSB; also known as ERCC6)) are required to allow histone
genes owing to impaired association of NER factors with thetranscription machinery. Such observations raise the question of the potential role or roles of the NER factors at the promoters of active genes. Although poorly understood, it seems that NER factors might influence chromatin remodelling5,6. Nonetheless, this involvement of repair factors during transcription is forcing a reconsideration of the broad clinical features that are described for the so-called xeroderma pigmentosu m, TTD and Cockayne syndrome repair syndromes.
Conclusions and perspectives Studies of TFIIH have demonstrated the tight connections between factors that ensure accurate initiation of RNA synthesis, including the basal transcription machinery, transcription cofactors, the Mediator complex and NER factors. In retrospect, one can appreciat e how the methodical and sophisticated dissection of TFIIH functions during transcription and NER has also shed light on the biological machines that regulate these two processes and the medical consequences of errors in these processes. Last but not least, studies of TFIIH have demonstrated the close links between transcription and DNA repair and have also given rise to a new concept of transcription diseases. Further characterization of TFIIH binding partners will improve our understanding of the complexity of the mechanisms that specifically regulate the expression of protein coding genes at the right time and in the right amount in each cell. More importantly, such studies should provide explanations for the phenotypes that are observed in patients with mutations in the components of the transcription machinery and might allow therapeutic strategies to be designed that can modulate gene expression patterns.
13. Weeda, G., Ma, L., van der Ham, R., van der Eb, A.J. & Hoeijmakers, J.H.J. Structure and expression of the human XPBC/ERCC3 gene involved in DNA repair disorders xeroderma pigmentosum and Cockaynes syndrome. Nucleic Acids Res. 19, 63016308 (1991). 14. Feaver, W.J., Gileadi, O., Li, Y. & Kornberg, R.D. CTDkinase associated with yeast RNA polymerase II initiation factor b. Cell 67, 12231230 (1991). 15. Roy, R. etal. The MO15 cell cycle kinase is associated with the TFIIH transcription-DNA repair factor. Cell79, 10931101 (1994). Revealed that MO15 (also known as CDK7) kinase is present in the TFIIH transcription complex. 16. Fisher, R.P. & Morgan, D.O. A novel cyclin associates with MO15/CDK7 to form the CDK-activating kinase. Cell 78, 713724 (1994). 17. Mkel, T.P. etal. A cyclin associated with the CDKassociated kinase MO15. Nature 371, 254257 (1994). 18. Adamczewski, J.P. etal. MAT1, cdk7 and cyclin H form a kinase complex which is UV light-sensitive upon association with TFIIH. EMBO J. 15, 18771884 (1996). 19. Rossignol, M., Kolb-Cheynel, I. & Egly, J.M. Substratespecificity of the cdk-activating kinase (CAK) is altered upon association with TFIIH. EMBO J. 16, 16281637 (1997). 20. Larochelle, S., Pandur, J., Fisher, R.P., Salz, H.K. & Suter, B. Cdk7 is essential for mitosis and for invivo Cdk-activating kinase activity. Genes Dev. 12, 370381 (1998). 21. Fesquet, D., Morin, N., Doree, M. & Devault, A. IsCdk7/cyclin H/MAT1 the genuine cdk activating kinase in cycling Xenopus egg extracts? Oncogene 15, 13031307 (1997).
2.
3.
4. 5.
6.
modifications and active DNA demethylation that are necessary for efficient transcription. Conaway, R.C. & Conaway, J.W. An RNA polymerase II transcription factor has an associated DNAdependent ATPase (dATPase) activity strongly stimulated by TATA region of promoters. Proc. Natl Acad. Sci. USA 86, 73567360 (1989). Reported the purification of a transcription factor from rat liver that was designated transcription factor-, which has an associated DNA-dependent ATPase activity. 8. Gerard, M. etal. Purification and interaction properties of the human RNA polymerase B(II) general transcription factor BTF2. J. Biol. Chem. 266, 2094020945 (1991). Described the purification of the human cell transcription factor BTF2, which is required for the transcription of class II genes. 9. Feaver, W.J., Gileadi, O. & Kornberg, R. Purification and characterization of yeast RNA polymerase II transcription factor b. J.Biol. Chem. 266, 1900019005 (1991). Reported the purification of a transcription factor from yeast that was designated Tfb, which is required for Pol II transcription. 10. Flores, O., Lu, H. & Reinberg, D. Factors involved in specific transcription by mammalian RNA polymerase II. Identification and characterization of factor IIH. J.Biol. Chem. 267, 27862793 (1992). 11. Schaeffer, L. etal. The ERCC2/DNA repair protein is associated with the class II BTF2/TFIIH transcription factor. EMBO J. 13, 23882392 (1994). 12. Weber, C.A., Salazar, E.P., Stewart, S.A. & Thompson, L.H. ERCC2: cDNA cloning and molecular characterization of human nucleotide excision repair gene with high homology to yeast RAD3. EMBO J. 9, 14371447 (1990). 7.
VOLUME 13 | JUNE 2012 | 351
REVIEWS
22. Wu, L. etal. RNA antisense abrogation of MAT1 induces G1 phase arrest and triggers apoptosis in aortic smooth muscle cells. J. Biol. Chem. 274, 55645572 (1999). 23. Aprelikova, O., Xiong, Y. & Liu, E.T. Both p16 and p21 families of cyclin-dependent kinase (CDK) inhibitors block the phosphorylation of cyclin-dependent kinases by the CDK-activating kinase. J. Biol. Chem. 270, 1819518197 (1995). 24. Fesquet, D. etal. The MO15 gene encodes the catalytic subunit of a protein kinase that activates cdc2 and other cyclin dependent kinases (CDKs) through phosphorylation of Thr161 and its homologues. EMBO J. 12, 31113121 (1993). 25. Matsuoka, M., Kato, J.Y., Fisher, R.P., Morgan, D.O. & Sherr, C.J. Activation of cyclin-dependent kinase 4 (cdk4) by mouse MO15-associated kinase. Mol. Cell. Biol. 14, 72657275 (1994). 26. Poon, R.Y.C., Yamashita, K., Adamczewski, J.P., Hunt,T. & Shuttleworth, J. The cdc2-related protein p40MO15 is the catalytic subunit of a protein kinase that can activate p33cdk2 and p34cdc2. EMBO J. 12, 31233132 (1993). 27. Solomon, M.J., Harper, W.J. & Shuttleworth, J. CAK,the p34cdc2 activating kinase contains a protein kinase identical to or closely related to p40MO15. EMBO J. 12, 31333142 (1993). 28. Fisher, R.P. Secrets of a double agent: CDK7 in cell-cycle control and transcription. J. Cell Sci. 118, 51715180 (2005). 29. Ito, S. etal. MMXD, a TFIIH-independent XPD MMS19 protein complex involved in chromosome segregation. Mol. Cell 39, 632640 (2010). 30. Sancar, A. Mechanisms of DNA excision repair. Science 266, 19541956 (1994). 31. Aboussekhra, A. etal. Mammalian DNA nucleotide excision repair reconstituted with purified protein components. Cell 80, 859868 (1995). 32. Volker, M. etal. Sequential assembly of the nucleotide excision repair factors invivo. Mol. Cell 8, 213224 (2001). 33. Hanawalt, P.C. & Spivak, G. Transcription-coupled DNA repair: two decades of progress and surprises. Nature Rev. Mol. Cell Biol. 9, 958970 (2008). 34. Clement, F.C. etal. Dynamic two-stage mechanism of versatile DNA damage recognition by xeroderma pigmentosum group C protein. Mutat. Res. 685, 2128 (2010). 35. Bunick, C.G., Miller, M.R., Fuller, B.E., Fanning, E. & Chazin, W.J. Biochemical and structural domain analysis of xeroderma pigmentosum complementation group C protein. Biochemistry 45, 1496514979 (2006). 36. Min., J.H. & Pavletich, N.P. Recognition of DNA damage by the Rad4 nucleotide excision repair protein. Nature 449, 570575 (2007). 37. Sugasawa, K. etal. A multistep damage recognition mechanism for global genomic nucleotide excision repair. Genes Dev. 15, 507521 (2001). 38. Mocquet, V. etal. The human DNA repair factor XPCHR23B distinguishes stereoisomeric benzo[a] pyrenyl-DNA lesions. EMBO J. 26, 29232932 (2007). 39. Keeney, S., Chang, G.J. & Linn, S. Characterization of a human DNA damage binding protein implicated in xeroderma pigmentosum E.J.Biol. Chem. 268, 2129321300 (1993). 40. Takao, M. etal. A 127 kDa component of a UVdamaged DNA-binding complex which is defective in some xeroderma pigmentosum group E patients is homologous to a slime mold protein. Nucleic Acids Res. 21, 41114118 (1993). 41. Fitch, M.E. etal. The DDB2 nucleotide excision repair gene product p48 enhances global genomic repair in p53 deficient human fibroblasts. DNA Repair (Amst.) 2, 819826 (2003). 42. Wang, Q.E., Zhu, Q., Wani, G., Chen, J. & Wani, A.A. UV radiation-induced XPC translocation within chromatin is mediated by damaged-DNA binding protein, DDB2. Carcinogenesis 25, 10331043 (2004). 43. Fei, J. etal. Regulation of nucleotide excision repair by UVDDB: prioritization of damage recognition to internucleosomal DNA. PLoS Biol. 9, e1001183 (2011). 44. Gillette, T.G. etal. Distinct functions of the ubiquitin proteasome pathway influence nucleotide excision repair. EMBO J. 25, 25292538 (2006). 45. Araki, M. etal. Centrosome protein centrin 2/ caltractin 1 is part of the xeroderma pigmentosum group C complex that initiates global genome nucleotide excision repair. J. Biol. Chem. 276, 1866518672 (2001). Xie, Z., Liu, S., Zhang, Y. & Wang, Z. Roles of Rad23 protein in yeast nucleotide excision repair. Nucleic Acids Res. 32, 59815990 (2004). Tapias, A. etal. Ordered conformational changes in damaged DNA induced by nucleotide excision repair factors. J. Biol. Chem. 279, 1907419083 (2004). Bernardes de Jesus, B.M., Bjoras, M., Coin, F. & Egly,J.M. Dissection of the molecular defects caused by pathogenic mutations in the DNA repair factor XPC. Mol. Cell. Biol. 28, 72257235 (2008). Bootsma, D. & Hoeijmakers, J.H.J. DNA repair. Engagement with transcription. Nature 363, 114115 (1993). Wood, R.D. DNA damage recognition during nucleotide excision repair in mammalian cells. Biochimie 81, 3944 (1999). Coin, F. etal. Mutations in the XPD helicase gene result in XP and TTD phenotypes, preventing interaction between XPD and the p44 subunit of TFIIH. Nature Genet. 20, 184188 (1998). Demonstrated that mutations in the XPD Cterminal domain that are found in most patients with xeroderma pigmentosum and TTD prevent the interaction with p44, thus explaining the observed decrease in XPD helicase activity and the NER defect. Dubaele, S. etal. Basal transcription defect discriminates between xeroderma pigmentosum and trichothiodystrophy in XPD patients. Mol. Cell 11, 16351646 (2003). Tirode, F., Busso, D., Coin, F. & Egly, J.M. Reconstitution of the transcription factor TFIIH: assignment of functions for the three enzymatic subunits, XPB, XPD, and cdk7. Mol. Cell 3, 8795 (1999). Oksenych, V., de Jesus, B.B., Zhovmer, A., Egly, J.M. & Coin, F. Molecular insights into the recruitment of TFIIH to sites of DNA damage. EMBO J. 28, 29712980 (2009). Fan, L. etal. Conserved XPB core structure and motifs for DNA unwinding: implications for pathway selection of transcription or excision repair. Mol. Cell 22, 2737 (2006). Coin, F., Oksenych, V. & Egly, J.M. Distinct roles for the XPB/p52 and XPD/p44 subcomplexes of TFIIH in damaged DNA opening during nucleotide excision repair. Mol. Cell 26, 245256 (2007). Revealed that the helicase activity of XPB is not used for damaged DNA opening, which is instead driven by the ATPase activity of XPB in combination with the helicase activity of XPD. Furthermore, TFIIH from patients with mutated XPB is unable toinduce DNA opening around the lesion owing to impaired XPBp52 interaction and ATPase stimulation. Fregoso, M. etal. DNA repair and transcriptional deficiencies caused by mutations in the Drosophila p52 subunit of TFIIH generate developmental defects and chromosome fragility. Mol. Cell. Biol. 27, 36403650 (2007). Fan, L. etal. XPD helicase structures and activities: insights into the cancer and aging phenotypes from XPD mutations. Cell 133, 789800 (2008). Liu, H. etal. Structure of the DNA repair helicase XPD. Cell 133, 801812 (2008). Wolski, S.C. etal. Crystal structure of the FeS clustercontaining nucleotide excision repair helicase XPD. PLoS Biol. 6, e149 (2008). Mathieu, N., Kaczmarek, N. & Naegeli, H. Strand- and site-specific DNA lesion demarcation by the xeroderma pigmentosum group D helicase. Proc. Natl Acad. Sci. USA 107, 1754517550 (2010). Theis, K., Chen, P.J., Skorvaga, M., Van Houten, B. & Kisker, C. Crystal structure of UvrB, a DNA helicase adapted for nucleotide excision repair. EMBO J. 18, 68996907 (1999). Skorvaga, M., Theis, K., Mandavilli, B.S., Kisker, C. & Van Houten, B. The -hairpin motif of UvrB is essential for DNA binding, damage processing, and UvrCmediated incisions. J. Biol. Chem. 277, 15531559 (2002). Reardon, J.T. & Sancar, A. Recognition and repair of the cyclobutane thymine dimer, a major cause of skin cancers, by the human excision nuclease. Genes Dev. 17, 25392551 (2003). Stefanini, M., Botta, E., Lanzafame, M. & Orioli, D. Trichothiodystrophy: from basic mechanisms to clinical implications. DNA Repair (Amst.) 9, 210 (2010). 66. Ranish, J.A. etal. Identification of TFB5, a new component of general transcription and DNA repair factor IIH. Nature Genet. 36, 707713 (2004). 67. Giglia-Mari, G. etal. A new, tenth subunit of TFIIH is responsible for the DNA repair syndrome trichothiodystrophy group A. Nature Genet. 36, 714719 (2004). Demonstrated that p8 is an evolutionarily conserved subunit of TFIIH and identified GTF2H5 as the gene that causes the NER defect in TTDA. 68. Coin, F. etal. p8/TTDA as a repair-specific TFIIH subunit. Mol. Cell 21, 215226 (2006). 69. Vermeulen, W. etal. Sublimiting concentration of TFIIH transcription/DNA repair factor causes TTDA trichothiodystrophy disorder. Nature Genet. 26, 307313 (2000). 70. Vitorino, M. etal. Solution structure and selfassociation properties of the p8 TFIIH subunit responsible for trichothiodystrophy. J. Mol. Biol. 368, 473480 (2007). 71. Kainov, D.E., Vitorino, M., Cavarelli, J., Poterszman,A. & Egly, J.M. Structural basis for group A trichothiodystrophy. Nature Struct. Mol. Biol. 15, 980984 (2008). 72. Park, C.H. & Sancar, A. Formation of a ternary complex by human XPA, ERCC1, and ERCC4(XPF) excision repair proteins. Proc. Natl Acad. Sci. USA 91, 50175021 (1994). 73. Krasikova, Y.S., Rechkunova, N.I., Maltseva, E.A., Petruseva, I.O. & Lavrik, O.I. Localization of xeroderma pigmentosum group A protein and replication protein A on damaged DNA in nucleotide excision repair. Nucleic Acids Res. 38, 80838094 (2010). 74. Ikegami, T. etal. Solution structure of the DNA- and RPA-binding domain of the human repair factor XPA. Nature Struct. Biol. 5, 701706 (1998). 75. Saijo, M., Kuraoka, I., Masutani, C., Hanaoka, F. & Tanaka, K. Sequential binding of DNA repair proteins RPA and ERCC1 to XPA invitro. Nucleic Acids Res. 24, 47194724 (1996). 76. Evans, E., Moggs, J.G., Hwang, J.R., Egly, J.M. & Wood, R.D. Mechanism of open complex and dual incision formation by human nucleotide excision repair factors. EMBO J. 16, 65596573 (1997). 77. Coin, F. etal. Nucleotide excision repair driven by the dissociation of CAK from TFIIH. Mol. Cell 31, 920 (2008). Showed the release of CAK from the core TFIIH during the engagement of this complex in DNA repair. Following repair, CAK reappears with the core subunit of TFIIH on chromatin, coincident with the resumption of transcription. 78. Svejstrup, J.Q. etal. Different forms of TFIIH for transcription and DNA repair: holo-TFIIH and a nucleotide excision repairosome. Cell 80, 2128 (1995). 79. Araujo, S.J. etal. Nucleotide excision repair of DNA with recombinant human proteins: definition of the minimal set of factors, active forms of TFIIH, and modulation by CAK. Genes Dev. 14, 349359 (2000). 80. Sandrock, B. & Egly, J.M. A yeast four-hybrid system identifies Cdk-activating kinase as a regulator of the XPD helicase, a subunit of transcription factor IIH. J.Biol. Chem. 276, 3532835333 (2001). 81. Fan, W. & Luo, J. SIRT1 regulates UVinduced DNA repair through deacetylating XPA. Mol. Cell 39, 247258 (2010). 82. Sugasawa, K., Akagi, J., Nishi, R., Iwai, S. & Hanaoka, F. Two-step recognition of DNA damage for mammalian nucleotide excision repair: directional binding of the XPC complex and DNA strand scanning. Mol. Cell 36, 642653 (2009). 83. Naegeli, H. & Sugasawa, K. The xeroderma pigmentosum pathway: decision tree analysis of DNA quality. DNA Repair (Amst.) 10, 673683 (2011). 84. Kesseler, K.J., Kaufmann, W.K., Reardon, J.T., Elston,T.C. & Sancar, A. A mathematical model for human nucleotide excision repair: damage recognition by random order assembly and kinetic proofreading. J. Theor. Biol. 249, 361275 (2007). 85. Mocquet, V. etal. Sequential recruitment of the repair factors during NER: the role of XPG in initiating the resynthesis step. EMBO J. 27, 155167 (2008). 86. Staresincic, L. etal. Coordination of dual incision and repair synthesis in human nucleotide excision repair. EMBO J. 28, 11111120 (2009).
46. 47.
48.
49. 50. 51.
52.
53.
54.
55.
56.
57.
58. 59. 60. 61.
62.
63.
64.
65.
REVIEWS
87. Moggs, J.G., Yarema, K.J., Essigmann, J.M. & Wood,R.D. Analysis of incision sites produced by human cell extracts and purified proteins during nucleotide excision repair of a 1,3-intrastrand d(GpTpG)-cisplatin adduct. J. Biol. Chem. 271, 71777186 (1996). 88. Gillet, L.C. & Scharer, O.D. Molecular mechanisms of mammalian global genome nucleotide excision repair. Chem. Rev. 106, 253276 (2006). 89. Lehmann, A.R. DNA polymerases and repair synthesis in NER in human cells. DNA Repair (Amst.) 10, 730733 (2011). 90. Ogi, T. etal. Three DNA polymerases, recruited by different mechanisms, carry out NER repair synthesis in human cells. Mol. Cell 37, 714727 (2010). 91. Moser, J. etal. Sealing of chromosomal DNA nicks during nucleotide excision repair requires XRCC1 and DNA ligase III in a cellcyclespecific manner. Mol. Cell 27, 311323 (2007). 92. Gaillard, P.H.L. etal. Chromatin assembly coupled to DNA repair: a new role for chromatin assembly factor I. Cell 86, 887896 (1996). 93. Polo, S.E., Roche, D. & Almouzni, G. New histone incorporation marks sites of UV repair in human cells. Cell 127, 481493 (2006). 94. Ito, S. etal. XPG stabilizes TFIIH, allowing transactivation of nuclear receptors: implications for Cockayne Syndrome in XPG/CS Patients. Mol. Cell 26, 231243 (2007). Showed that XPG forms a stable complex with TFIIH and functions in maintaining the architecture of TFIIH, which underlines the contribution of XPG to transcription. 95. Coin, F. etal. Phosphorylation of XPB helicase regulates TFIIH nucleotide excision repair activity. EMBO J. 23, 48354846 (2004). 96. Hoogstraten, D. etal. Rapid switching of TFIIH between RNA polymerase I and II transcription and DNA repair invivo. Mol. Cell 10, 11631174 (2002). 97. Iben, S. etal. TFIIH plays an essential role in RNA polymerase I transcription. Cell 109, 297306 (2002). Found that TFIIH serves a function in ribosomal gene transcription. TFIIH is required for productive but not abortive ribosomal DNA transcription, which implies a post-initiation role for TFIIH in transcription. 98. Assfalg, R. etal. TFIIH is an elongation factor of RNA polymerase I. Nucleic Acids Res. 40, 650659 (2011). 99. Barski, A. etal. Pol II and its associated epigenetic marks are present at Pol III-transcribed noncoding RNA genes. Nature Struct. Mol. Biol. 17, 629634 (2010). 100. Oler, A.J. etal. Human RNA polymerase III transcriptomes and relationships to Pol II promoter chromatin and enhancer-binding factors. Nature Struct. Mol. Biol. 17, 620628 (2010). 101. Kornberg, R.D. Eukaryotic transcriptional control. Trends Cell Biol. 9, M46M49 (1999). 102. Sims, R.J. 3rd, Mandal, S.S. & Reinberg, D. Recenthighlights of RNApolymeraseII-mediated transcription. Curr. Opin. Cell Biol. 16, 263271 (2004). 103. Cairns, B.R. etal. RSC, an essential, abundant chromatin-remodeling complex. Cell 87, 12491260 (1996). 104. Chao, D.M. etal. A mammalian SRB protein associated with an RNA polymerase II holoenzyme. Nature 380, 8285 (1996). 105. Kim, Y.J., Bjorklund, S., Li, Y., Sayre, M.H. & Kornberg, R.D. A multiprotein mediator of transcriptional activation and its interaction with the Cterminal repeat domain of RNA polymerase II. Cell77, 599608 (1994). 106. Ossipow, V., Tassan, J.P., Nigg, E.A. & Schibler, U. A mammalian RNA polymerase II holoenzyme containing all components required for promoter-specific transcription initiation. Cell 83, 137146 (1995). 107. Zawel, L. & Reinberg, D. Common themes in assembly and function of eukaryotic transcription complexes. Annu. Rev. Biochem. 64, 533561 (1995). 108. Douziech, M. etal. Mechanism of promoter melting by the xeroderma pigmentosum complementation group B helicase of transcription factor IIH revealed by protein-DNA photocrosslinking. Mol. Cell. Biol. 20, 81688177 (2000). 109. Holstege, F.C., van der Vliet, P.C. & Timmers, H.T. Opening of an RNA polymerase II promoter occurs in two distinct steps and requires the basal transcription factors IIE and IIH. EMBO J. 15, 16661677 (1996). Presented a model in which the crucial function of TFIIH-associated DNA helicases is to create an ssDNA region during transcription. 110. Coin, F., Bergmann, E., Tremeau-Bravard, A. & Egly,J.M. Mutations in XPB and XPD helicases found in xeroderma pigmentosum patients impair the transcription function of TFIIH. EMBO J. 18, 13571366 (1999). 111. Moreland, R.J. etal. A role for the TFIIH XPB DNA helicase in promoter escape by RNA polymerase II. J.Biol. Chem. 274, 2212722130 (1999). 112. Dvir, A. etal. A role for ATP and TFIIH in activation of the RNA polymerase II preinitiation complex prior to transcription initiation. J. Biol. Chem. 271, 72457248 (1996). 113. Dvir, A., Conaway, R.C. & Conaway, J.W. A role for TFIIH in controlling the activity of early RNA polymerase II elongation complexes. Proc. Natl Acad. Sci. USA 94, 90069010 (1997). 114. Liu, J. etal. Defective interplay of activators and repressors with TFIH in xeroderma pigmentosum. Cell104, 353363 (2001). 115. Lu, H., Zawel, L., Fisher, L., Egly, J.M. & Reinberg, D. Human general transcription factor IIH phosphorylates the Cterminal domain of RNA polymerase II. Nature358, 641645 (1992). Showed that the phosphorylation of the C-terminal domain of the largest subunit of Pol II by CDK7 contributes to the transition from transcription initiation to elongation. 116. Feaver, W.J., Svejstrup, J.Q., Henry, N.L. & Kornberg, R.D. Relationship of CDK-activating kinase and RNA polymerase II CTD kinase TFIIH/TFIIK. Cell79, 11031109 (1994). 117. Shiekhattar, R. etal. Cdk-activating kinase complex is a component of human transcription factor TFIIH. Nature 374, 283287 (1995). 118. Bensaude, O. etal. Regulated phosphorylation of the RNA polymerase II Cterminal domain (CTD). Biochem. Cell Biol. 77, 249255 (1999). 119. Buratowski, S. Progression through the RNA polymerase II CTD cycle. Mol. Cell 36, 541546 (2009). 120. Serizawa, H., Conaway, J.W. & Conaway, R.C. Phosphorylation of Cterminal domain of RNA polymerase II is not required in basal transcription. Nature 363, 371374 (1993). 121. Cho, E., Tagaki, T., Moore, C.R. & Buratowski, S. mRNA capping enzyme is recruted to the transcritption complex by phosphorylation of the RNA polymerase II carboxy-terminal domain. Genes Dev. 11, 33193326 (1997). 122. Komarnitsky, P., Cho, E.J. & Buratowski, S. Different phosphorylated forms of RNA polymerase II and associated mRNA processing factors during transcription. Genes Dev. 14, 24522460 (2000). 123. Helenius, K. etal. Requirement of TFIIH kinase subunit Mat1 for RNA Pol II Cterminal domain Ser5 phosphorylation, transcription and mRNA turnover. Nucleic Acids Res. 39, 50255035 (2011). 124. Akoulitchev, S., Chuikov, S. & Reinberg, D. TFIIH is negatively regulated by cdk8-containing mediator complexes. Nature 407, 102106 (2000). 125. Yakovchuk, P., Goodrich, J.A. & Kugel, J.F. B2 RNA and Alu RNA repress transcription by disrupting contacts between RNA polymerase II and promoter DNA within assembled complexes. Proc. Natl Acad. Sci. USA 106, 55695574 (2009). 126. OGorman, W., Thomas, B., Kwek, K.Y., Furger, A. & Akoulitchev, A. Analysis of U1 small nuclear RNA interaction with cyclin H. J. Biol. Chem. 280, 3692036925 (2005). 127. Akhtar, M.S. etal. TFIIH kinase places bivalent marks on the carboxy-terminal domain of RNA polymerase II. Mol. Cell 34, 387393 (2009). 128. Glover-Cutter, K. etal. TFIIH-associated Cdk7 kinase functions in phosphorylation of Cterminal domain Ser7 residues, promoter-proximal pausing, and termination by RNA polymerase II. Mol. Cell. Biol. 29, 54555464 (2009). 129. Kim, M., Suh, H., Cho, E.J. & Buratowski, S. Phosphorylation of the yeast Rpb1 Cterminal domain at serines 2, 5, and 7.J. Biol. Chem. 284, 2642126426 (2009). 130. Egloff, S. etal. Serine7 of the RNA polymerase II CTD is specifically required for snRNA gene expression. Science 318, 17771779 (2007). 131. Cho, E.J., Kobor, M.S., Kim, M., Greenblatt, J. & Buratowski, S. Opposing effects of Ctk1 kinase and Fcp1 phosphatase at Ser 2 of the RNA polymerase II Cterminal domain. Genes Dev. 15, 33193329 (2001). 132. Zhou, M. etal. Tat modifies the activity of CDK9 to phosphorylate serine 5 of the RNA polymerase II carboxyl-terminal domain during human immunodeficiency virus type1 transcription. Mol.Cell.Biol. 20, 50775086 (2000). 133. Ahn, S.H., Kim, M. & Buratowski, S. Phosphorylation of serine 2 within the RNA polymerase II Cterminal domain couples transcription and 3 end processing. Mol. Cell 13, 6776 (2004). 134. Gebara, M.M., Sayre, M.H. & Corden, J.L. Phosphorylation of the carboxy-terminal repeat domain in RNA polymerase II by cyclin-dependent kinases is sufficient to inhibit transcription. J. Cell. Biochem. 64, 390402 (1997). 135. Hengartner, C.J. etal. Temporal regulation of RNA polymerase II by Srb10 and Kin28 cyclin-dependent kinases. Mol. Cell 2, 4353 (1998). 136. Bonnet, F., Vigneron, M., Bensaude, O. & Dubois,M.F. Transcription-independent phosphorylation of the RNA polymerase II Cterminal domain (CTD) involves ERK kinases (MEK1/2). Nucleic Acids Res. 27, 43994404 (1999). 137. Dvir, A., Peterson, S.R., Knuth, M.W., Lu, H. & Dynan, W.S. Ku autoantigen is the regulatory component of a template-associated protein kinase that phosphorylates RNA polymerase II. Proc. Natl Acad. Sci. USA 89, 1192011924 (1992). 138. Trigon, S. etal. Characterization of the residues phosphorylated invitro by different Cterminal domain kinases. J. Biol. Chem. 273, 67696775 (1998). 139. Chambers, R.S. & Kane, C.M. Purification and characterization of an RNA polymerase II phosphatase from yeast. J.Biol. Chem. 271, 2440824504 (1996). 140. Lin, P.S., Dubois, M.F. & Dahmus, M.E. TFIIFassociating carboxyl-terminal domain phosphatase dephosphorylates phosphoserines 2 and 5 of RNA polymerase II. J. Biol. Chem. 277, 4594945956 (2002). 141. Mosley, A.L. etal. Rtr1 is a CTD phosphatase that regulates RNA polymerase II during the transition from serine 5 to serine 2 phosphorylation. Mol. Cell 34, 168178 (2009). 142. Phatnani, H.P. & Greenleaf, A.L. Phosphorylation and functions of the RNA polymerase II CTD. Genes Dev. 20, 29222936 (2006). 143. Corden, J.L. Transcription. Seven ups the code. Science 318, 17351736 (2007). 144. Egloff, S. & Murphy, S. Cracking the RNA polymerase II CTD code. Trends Genet. 24, 280288 (2008). 145. Lu, H., Fisher, R.P., Bailey, P. & Levine, A.J. TheCDK7cycHp36 complex of transcription factor IIH phosphorylates p53, enhancing its sequencespecific DNA binding activity invitro. Mol. Cell. Biol. 17, 59235934 (1997). 146. Xiao, H. etal. Binding of basal transcription factor TFIIH to the acidic activation domains of VP16 and p53. Mol. Cell. Biol. 14, 70137024 (1994). 147. Tong, X., Drapkin, R., Reinberg, D. & Kieff, E. The 62- and 80-kDa subunits of transcription factor IIH mediate the interaction with Epstein-Barr virus nuclear protein 2. Proc. Natl Acad. Sci. USA 92, 32593263 (1995). 148. Qadri, I., Conaway, J.W., Conaway, R.C., Schaack, J. & Siddiqui, A. Hepatitis B virus transactivator protein, HBx, associates with the components of TFIIH and stimulates the DNA helicase activity of TFIIH. Proc. Natl Acad. Sci. USA 93, 1057810583 (1996). 149. Liu, J. etal. The FBP interacting repressor targets TFIIH to inhibit activated transcription. Mol. Cell 5, 331341 (2000). 150. Mitchell, N.C. etal. Hfp inhibits Drosophila myc transcription and cell growth in a TFIIH/Haydependent manner. Development 137, 28752884 (2010). 151. Chymkowitch, P., Le May, N., Charneau, P., Compe, E. & Egly, J.M. The phosphorylation of the androgen receptor by TFIIH directs the ubiquitin/proteasome process. EMBO J. 30, 468479 (2011). 152. Bastien, J. etal. TFIIH interacts with the retinoic acid receptor- and phosphorylates its AF1activating domain through cdk7. J. Biol. Chem. 275, 2189621904 (2000). 153. Rochette-Egly, C., Adam, S., Rossignol, M., Egly, J.M. & Chambon, P. Stimulation of RAR activation function AF1 through binding to the general transcription factor TFIIH and phosphorylation by CDK7. Cell 90, 97107 (1997). Revealed that RAR is targeted by the CDK7 subunit of TFIIH, suggesting that the activity of a transactivator could be modulated through its interaction with a general transcription factor.
VOLUME 13 | JUNE 2012 | 353
REVIEWS
154. Lee, D.K., Duan, H.O. & Chang, C. From androgen receptor to the general transcription factor TFIIH. Identification of cdk activating kinase (CAK) as an androgen receptor NH2-terminal associated coactivator. J. Biol. Chem. 275, 93089313 (2000). 155. Compe, E. etal. Dysregulation of the peroxisome proliferator-activated receptor target genes by XPD mutations. Mol. Cell. Biol. 25, 60656076 (2005). 156. Compe, E. etal. Neurological defects in trichothiodystrophy reveal a coactivator function of TFIIH. Nature Neurosci. 10, 14141422 (2007). Reported hypomyelination in the central nervous system of mice with TTD, which is related to the dysregulation of various thyroid hormone target genes. Proposed that such a dysregulation is likely to result from the inability of the mutated TFIIH to fully participate in the recruitment of thyroid hormone receptors to their response elements. 157. Chen, D. etal. Activation of estrogen receptor- by S118 phosphorylation involves a ligand-dependent interaction with TFIIH and participation of CDK7. Mol.Cell 6, 127137 (2000). 158. Liu, Y. etal. p62, a TFIIH subunit, directly interacts with thyroid hormone receptor and enhances T3-mediated transcription. Mol. Endocrinol. 19, 879884 (2005). 159. Drane, P., Compe, E., Catez, P., Chymkowitch, P. & Egly, J.M. Selective regulation of vitaminD receptorresponsive genes by TFIIH. Mol. Cell 16, 187197 (2004). 160. Nigg, E.A. Cyclin-dependent kinase 7: at the crossroads of transcription, DNA repair and cell cycle control? Curr. Opin. Cell Biol. 8, 312317 (1996). 161. Morgan, D.O. Cyclin-dependent kinases: engines, clocks, and microprocessors. Annu. Rev. Cell Dev. Biol. 13, 261291 (1997). 162. Schwartz, B.E., Larochelle, S., Suter, B. & Lis, J.T. Cdk7 is required for full activation of Drosophila heat shock genes and RNA polymerase II phosphorylation invivo. Mol. Cell. Biol. 23, 68766886 (2003). 163. Larochelle, S. etal. Requirements for Cdk7 in the assembly of Cdk1/cyclin B and activation of Cdk2 revealed by chemical genetics in human cells. Mol. Cell 25, 839850 (2007). 164. Rochette-Egly, C. Nuclear receptors: integration of multiple signalling pathways through phosphorylation. Cell Signal. 15, 355366 (2003). 165. Keriel, A., Stary, A., Sarasin, A., Rochette-Egly, C. & Egly, J.M. XPD mutations prevent TFIIH-dependent transactivation by nuclear receptors and phosphorylation of RAR. Cell 109, 125135 (2002). Demonstrated that mutations in XPD result in the decreased ability of nuclear receptors to be phosphorylated by TFIIH and to stimulate expression of target genes. 166. Kioka, N. etal. Vinexin: a novel vinculin-binding protein with multiple SH3 domains enhances actin cytoskeletal organization. J. Cell Biol. 144, 5969 (1999). 167. Bour, G., Plassat, J.L., Bauer, A., Lalevee, S. & Rochette-Egly, C. Vinexin- interacts with the nonphosphorylated AF1 domain of retinoid receptor- (RAR) and represses RAR-mediated transcription. J.Biol. Chem. 280, 1702717037 (2005). 168. Sano, M. etal. Menageatrois 1 is critical for the transcriptional function of PPAR coactivator 1. Cell.Metab. 5, 129142 (2007). 169. Puigserver, P. & Spiegelman, B.M. Peroxisome proliferator-activated receptor- coactivator 1 (PGC1): transcriptional coactivator and metabolic regulator. Endocr. Rev. 24, 7890 (2003). 170. Talukder, A.H. etal. MTA1 interacts with MAT1, a cyclin-dependent kinase-activating kinase complex ring finger factor, and regulates estrogen receptor transactivation functions. J. Biol. Chem. 278, 1167611685 (2003). 171. Conaway, R.C. & Conaway, J.W. Function and regulation of the Mediator complex. Curr. Opin. Genet. Dev. 21, 225230 (2011). 172. Esnault, C. etal. Mediator-dependent recruitment of TFIIH modules in preinitiation complex. Mol. Cell 31, 337346 (2008). Reported a direct interaction between a Mediator head subunit and a TFIIH core subunit and concluded that the Mediator head module has a crucial role in TFIIH and TFIIE recruitment to the PIC. 173. Moriel-Carretero, M., Tous, C. & Aguilera, A. Control of the function of the transcription and repair factor TFIIH by the action of the cochaperone Ydj1. Proc. Natl Acad. Sci. USA 108, 1530015305 (2011). 174. Manuguerra, M. etal. XRCC3 and XPD/ERCC2 single nucleotide polymorphisms and the risk of cancer: a HuGE review. Am. J. Epidemiol. 164, 297302 (2006). 175. Zhang, J., Gu, S.Y., Zhang, P., Jia, Z. & Chang, J.H. ERCC2 Lys751Gln polymorphism is associated with lung cancer among Caucasians. Eur. J. Cancer 46, 24792484 (2010). 176. Le May, N. etal. TFIIH transcription factor, a target for the Rift Valley hemorrhagic fever virus. Cell 116, 541550 (2004). 177. Jaitovich-Groisman, I. etal. Transcriptional regulation of the TFIIH transcription repair components XPB and XPD by the hepatitis B virus x protein in liver cells andtransgenic liver tissue. J. Biol. Chem. 276, 1412414132 (2001). 178. Kraemer, K.H. etal. Xeroderma pigmentosum and related disorders: examining the linkage between defective DNA repair and cancer. J.Invest. Dermatol. 103, 96S101S (1994). 179. Cleaver, J.E. Cancer in xeroderma pigmentosum and related disorders of DNA repair. Nature Rev. Cancer 5, 564573 (2005). 180. de Boer, J. & Hoeijmakers, J.H. Nucleotide excision repair and human syndromes. Carcinogenesis 21, 453460 (2000). 181. Nance, M.A. & Berry, S.A. Cockayne syndrome: review of 140 cases. Am. J. Med. Genet. 84, 4268 (1992). 182. Itin, P.H., Sarasin, A. & Pittelkow, M.R. Trichothiodystrophy: update on the sulfur-deficient brittle hair syndromes. J. Am. Acad. Dermatol. 44, 891920 (2001). 183. Hashimoto, S. & Egly, J.M. Trichothiodystrophy view from the molecular basis of DNA repair/transcription factor TFIIH. Hum. Mol. Genet. 18, R224R230 (2009). 184. Berneburg, M. & Lehmann, A.R. Xeroderma pigmentosum and related disorders: defects in DNA repair and transcription. Adv. Genet. 43, 71102 (2001). 185. Ueda, T., Compe, E., Catez, P., Kraemer, K.H. & Egly,J.M. Both XPD alleles contribute to the phenotype of compound heterozygote xeroderma pigmentosum patients. J. Exp. Med. 206, 30313046 (2009). 186. Botta, E. etal. Reduced level of the repair/ transcription factor TFIIH in trichothiodystrophy. Hum.Mol. Genet. 11, 29192928 (2002). Showed that alterations in any of the gene products that result in the clinical phenotype of TTD specifically reduce the cellular content of the TFIIH complex. 187. de Boer, J. etal. A mouse model for the basal transcription/DNA repair syndrome trichothiodystrophy. Mol. Cell 1, 981990 (1998). 188. Bergmann, E. & Egly, J.M. Trichothiodystrophy, a transcription syndrome. Trends Genet. 17, 279286 (2001). 189. Takagi, Y. etal. Ubiquitin ligase activity of TFIIH and the transcriptional response to DNA damage. Mol. Cell 18, 237243 (2005). 190. Frit, P. etal. Transcriptional activators stimulate DNA repair. Mol. Cell 10, 13911401 (2002). 191. Fong, Y.W. etal. A DNA repair complex functions as an oct4/sox2 coactivator in embryonic stem cells. Cell147, 120131 (2011). Revealed a selective co-activator role of an NER complex in transcription in the context of embryonic stem cells. 192. Gibbons, B.J. etal. Subunit architecture of general transcription factor TFIIH. Proc. Natl Acad. Sci. USA 109, 19491954 (2012). 193. Chang, W.H. & Kornberg, R.D. Electron crystal structure of the transcription factor and DNA repair complex, core TFIIH. Cell 102, 609613 (2000). 194. Schultz, P. etal. Molecular structure of human TFIIH. Cell 102, 599607 (2000). References 193 and 194 showed the electron crystal structure of the yeast core TFIIH and the human TFIIH complex, respectively. 195. Rabut, G. etal. The TFIIH subunit Tfb3 regulates cullin neddylation. Mol. Cell 43, 488495 (2011). 196. Guzder, S.N., Sung, P., Prakash, L. & Prakash, S. Nucleotide excision repair in yeast is mediated by sequential assembly of repair factors and not by a pre-assembled repairosome. J. Biol. Chem. 271, 89038910 (1996). 197. Matsui, T., Segall, J., Weil, P. & Roeder, R. Multiple factors required for accurate initiation of transcription by purified RNA polymerase II.J.Biol. Chem. 255, 1199211996 (1980). 198. Samuels, M., Fire, A. & Sharp, P.A. Separation and characterization of factors mediating accurate transcription by RNA polymerase II. J. Biol. Chem. 257, 1441914427 (1982). 199. Sawadogo, M. & Roeder, R. Interaction of a genespecific transcription factor with the adenovirus major late promoter upstream of the TATA box region. Cell43, 165175 (1985). 200. Reinberg, D. & Roeder, R.G. Factors involved in specific transcription by mammalian RNA polymerase II. Purification and functional analysis of initiation factors IIB and IIE. J. Biol. Chem. 262, 33103321 (1987). 201. Buratowski, S., Hahn, S., Guarente, L. & Sharp, P.A. Five intermediate complexes in transcription initiation by RNA polymerase II. Cell 56, 549561 (1989). 202. Conaway, J.W., Hanley, J.P., Garrett, K.P. & Conaway,R.C. Transcription initiated by RNA polymerase II and transcription factors from liver. Structure and action of transcription factors epsilon and tau. J.Biol. Chem. 266, 78047811 (1991). 203. Perissi, V. & Rosenfeld, M.G. Controlling nuclear receptors: the circular logic of cofactor cycles. NatureRev. Mol. Cell Biol. 6, 542554 (2005).
Acknowledgements
The authors thank all of their past and present associates who have participated to the TFIIH adventure. The authors especially would like to thank R. Conaway and H. Naegeli for critical reading of the manuscript and F. Coin for helpful discussions. The authors apologize to all their colleagues whose important findings could not be included in this Review because of space limitations. This study was supported by a European Research Council advanced grant, the Agence Nationale de la Recherche (N#ANR08MIEN02203), the Association pour la Recherche sur le Cancer and the Institut National du Cancer (INCA2008041). E.C. and J.M.E. are supported by the Institut National de la Sant et de la Recherche Mdicale.
FURTHER INFORMATION
Teams homepage: http://www.igbmc.fr/research/2/team/20/
REVIEWS
Regulation of glucose transport by insulin: traffic control of GLUT4

Dara Leto1 and Alan R.Saltiel2
Abstract | Despite daily fasting and feeding, plasma glucose levels are normally maintained within a narrow range owing to the hormones insulin and glucagon. Insulin increases glucose uptake into fat and muscle cells through the regulated trafficking of vesicles that contain glucose transporter type 4 (GLUT4). New insights into insulin signalling reveal that phosphorylation events initiated by the insulin receptor regulate key GLUT4 trafficking proteins, including small GTPases, tethering complexes and the vesicle fusion machinery. These proteins, in turn, control GLUT4 movement through the endosomal system, formation and retention of specialized GLUT4 storage vesicles and targeted exocytosis of these vesicles. Understanding these processes may help to explain the development of insulin resistance in type2 diabetes and provide new potential therapeutic targets.
Gluconeogenesis
De novo synthesis of glucose from non-carbohydrate carbon sources.
Trans-Golgi network
(TGN). The terminal Golgi stack where proteins are sorted and packaged into vesicles for delivery to their cellular destination.
Insulin resistance
Physiological condition that is defined by a failure of tissues and organs to respond to normal concentrations ofinsulin.
Life Sciences Institute, University of Michigan, 210Washtenaw Avenue, AnnArbor, Michigan 48105, USA. 2 Departments of Molecular and Integrative Physiology and Internal Medicine, LifeSciences Institute, AnnArbor, Michigan 48105, USA. e-mails: letod@umich.edu; saltiel@umich.edu doi:10.1038/nrm3351
1
Vesicle exocytosis has a crucial role in the delivery of cellular material to the plasma membrane and extracellular space1. Whereas some Golgi-derived vesicles are constantly delivered to the plasma membrane through constitutive exocytosis, other populations of vesicles that are enriched in particular lipids and/or proteins undergo accelerated exocytosis in response to extracellular stimuli. This regulated exocytosis controls the release of soluble factors (for example, thrombin-stimulated secretion of von Willebrand factor from endothelial cells)2, the integration of membrane proteins into the plasma membrane (for example, insertion of aquaporin 2 into the apical membrane of polarized kidney duct cells in response to Ca2+)3 and the delivery of intracellular membranes for cell surface expansion (for example, insulinlike growth factor 1 (IGF1)stimulated vesicle exocytosis during neurite extension)4. The controlled exocytosis of these specialized vesicles depends on the coordinated regulation of the vesicle trafficking machinery by signalling pathways. Uncovering how signalling pathways communicate with proteins that are involved in vesicle formation, movement, targeting and fusion is crucial for understanding a variety of physiological processes. One well-studied example of regulated exocytosi s is that of the hexose transporter GLUT4 (glucose transporter type 4), which mediates insulin-stimulated glucos e transport in fat and muscle5. Following a meal, increased nutrients in the blood lead to secretion of insulin. This hormone in turn prevents gluconeogenesis in the liver and promotes glucose uptake into muscle and adipose tissue through regulated trafficking of GLUT4
from intracellular stores to the plasma membrane6. Glucose uptake is the rate limiting step in glucose utilization and/ or storage and as such has a key role in the maintenanc e of glucose homeostasis (BOX1). GLUT4 is one of 14 members of the GLUT family of facilitative transmembrane hexose transporters, each ofwhich has a distinct affinity and specificity for particular hexoses, as well as unique tissue distribution, sub cellular localization and physiological function 7. GLUT4 is a high-affinity glucose transporter that is predominantly expressed in muscle cells and adipocytes. In the absence of insulin, the majority of GLUT4 is distributed between endosomes, the trans-Golgi network (TGN) and heterogeneous tubulovesicular structures that consist of endosomal sorting intermediates and specialized GLUT4 storage vesicles (GSVs). In the absence of insulin, only ~5% of the total transporter pool is found on the cell surface8,9. Exclusion of GLUT4 from the cell surface depends on efficient sorting and sequestration into GSVs that do not readily cycle to the plasma membrane in the absence of stimulation10 but translocate there in response to insulin or exercise, which results in a tenfold increase in glucose uptake11. The failure of GLUT4 to translocate to the plasma membrane in response to insulin is an early step in the development of insulin resistance and type2 diabetes mellitus6. About 90% of insulin-stimulated glucose uptake occurs in skeletal muscle12. Although adipose tissue accounts for only 10% of insulin-stimulated glucose uptake, this process is important for controlling wholebody energy homeostasis, as adipocytes serve as a
VOLUME 13 | JUNE 2012 | 383
REVIEWS
Box 1 | Muscle and adipose tissue in energy homeostasis
Maintenance of energy homeostasis during changing energy demands and availability depends on the concerted action of multiple organs and tissues, including the digestive system, pancreas, brain, liver, muscle and adipose tissue. Together, these tissues sense energy and communicate fuel availability to other organs through the release of metabolites and hormones (see the figure). Defects in the sensing of theenergy status, and the ability to respond appropriately, result in metabolic diseasessuch as type2 diabetes mellitus. The anabolic hormone insulin is released from pancreatic -cells when dietary carbohydrates or amino acids are abundant. Insulin stimulates the conversion of simple energy units such as monosaccharides (including glucose) and amino acids into complex macromolecules such as proteins, lipids and glycogen. This is accomplished, in part, by increasing glucose uptake in muscle and adipose tissue. The majority of insulin-stimulated glucose uptake occurs in skeletal muscle12, where glucose is stored as glycogen, which is mobilized when fuel demands are high or glucose is not abundant. About 10% of insulin-stimulated glucose uptake occurs in adipose tissue, where energy is stored as triglycerides. Triglycerides are released from adipocytes as free fatty acids (FFAs) and utilized as an energy source by other tissues when fuel availability is low. Adipose tissue serves an additional endocrine function by secreting adipokines and lipids167. These factors communicate the energy status to other tissues in the body, including the liver, muscle and brain, to alter fuel usage and feeding patterns. Adipocyte-specific Glut4 (glucose transporter type 4; also known as Slc2a4)-knockout (AG4KO) mice display several markers of insulin resistance, including an increase in circulating insulin levels, a failure of insulin to suppress hepatic glucose production through gluconeogenesis and a reduction in insulin-stimulated glucose uptake in muscle14, which demonstrates that glucose uptake supports the endocrine function ofadipose tissue. How adipocytes translate changes in glucose uptake or storage into signals that affect global energy balance is not completely understood but is at least partially explained by the actions of adipokines such as leptin, adiponectin, resistin and retinol-binding protein 4 (RBP4)167,168. Finally, the brain has an additional role in relaying information about the energy status back to other organs including the liver and adipose tissue.
Nutrients
cellular rheostat that senses the energy status, and responds by secreting numerous hormones, such as cytokines and chemokines that regulate metabolism in muscle, the liver and the brain13,14. Thus, understanding glucose transport mechanisms in both skeletal muscle and adipocytes is crucial for elucidating the mechanisms that underlie the physiology and pathophysiology of energy metabolism. In this Review, we discuss the current understanding of how GLUT4 is compartmentalized, the roles of specific membrane trafficking proteins in maintaining this compartmentalization and the emerging points of intersection between insulin signalling and trafficking, with particular attention to the role of small GTPases.
Insulin signalling and GLUT4 traffic Insulin regulates energy metabolism by initiating several signalling cascades that control cell growth and survival, as well as protein, glycogen and lipid uptake, synthesis and hydrolysis6. Some of these pathways, including the RAS extracellular signal-regulated kinase (ERK) and mammalian target of rapamycin complex 1 (mTORC1) pathways, are not important for controlling glucose transport15,16. Instead, glucose transport in adipose tissue requires a phosphoinositide 3kinase (PI3K) and an APS (adaptor protein with pleckstrin homology (PH) and Src homology 2 (SH2) domains; also known as SH2B2) signalling cascade. Although the importance of the PI3K signalling pathway for glucose uptake in muscle is well established, the requirement for the APS signalling pathway remains untested in this tissue. Together, these pathways ensure the efficient delivery of GLUT4 to the cell surface by assembling signalling platforms at the plasma membrane that are comprised of lipids, protein kinases, lipid kinases, small GTPases and adaptor proteins. These signalling pathways engage the trafficking machinery to regulate GLUT4 cycling (FIG.1).
The targets of insulin signalling. GLUT4 delivery to the cell surface requires its mobilization from intracellular membrane compartments, recognition of GLUT4containing vesicles at the plasma membrane and finally fusion of these two membranes. Insulin signalling coordinates these steps by engaging a series of small GTPases that cycle between an active GTP-bound conformation, in which they mediate their biological effects, and an inactive GDP-bound state. Insulin alters the activation state of small GTPases by regulating guaninenucleotid e exchange factors (GEFs) and GTPaseactivating proteins (GAPs), both of which control this nucleotide cycling. Active GTPases interact with multiple components of the trafficking machinery to confer directionality and specificity in membrane flow17. In addition to small GTPases, insulin signalling directly targets motor proteins, membrane tethers and fusion-regulating proteins, which suggests that insulin acts at multiple steps in the GLUT4 trafficking itinerary to increase the concentration of the transporter on the surface of thecell. The insulin signalling pathways acutely upregulate GLUT4 surface levels largely by increasing exocytosis of GSVs. However, there is evidence that insulin might
Digestive system
Brain
Liver
Glucose Glucose Glycogen Muscle
Glucose Triglycerides Insulin Pancreas Nature Reviews | Molecular Cell Biology

384 | JUNE 2012 | VOLUME 13
Adipokines, FFAs
Adipose tissue
REVIEWS
a
Insulin Insulin receptor P TC10 TC10 APS GDP C3G GTP CAP CRK c-CBL P CIP4 GAPEX5 Vesicle retention GLUT4-mediated glucose uptake Glucose Exocyst Vesicle targeting GSV P PIP2 IRS PI3K P PIP3 P PDK1 P AKT GSV P RGC Vesicle targeting
GLUT4-mediated glucose uptake Glucose
mTORC2
GSV
GSV P AS160 Vesicle translocation
Type2 diabetes mellitus

A chronic metabolic disorder that is characterized by increased plasma glucose levels that result from an inability of tissues to respondto insulin.
Anabolic hormone
Secreted peptide that signals to cells to upregulate metabolic processes that convert simple energy sources into macromolecules.
Figure 1 | Insulin signalling regulates GLUT4 exocytosis by engaging the trafficking machinery. The APS (adaptor Nature Reviews | Molecular Cell Biology protein with pleckstrin homology (PH) and Src homology 2 (SH2) domains)insulin signalling cascade is initiated when the activated insulin receptor recruits the adaptor proteins APS, c-CBL and c-CBL-associated protein (CAP) (a). Insulin receptor-catalysed Tyr phosphorylation of cCBL in turn triggers the recruitment of the adaptor protein CRK and the guanine nucleotide exchange factor (GEF) C3G to the plasma membrane, where C3G activates the small GTPase TC10. GTP-bound TC10 interacts with the exocyst complex, thereby creating targeting sites for the glucose transporter type 4 (GLUT4) storage vesicle (GSV) at the plasma membrane. TC10 also interacts with CDC42interacting protein 4 (CIP4), which is associated with the RAB5 and RAB31 GEF GAPEX5. Translocation of GAPEX5 to the cell surface modulates the activation state of its target small GTPases, which are involved in GSV retention and translocation. The insulin receptor simultaneously initiates the phosphoinositide 3kinase (PI3K)-dependent signalling cascade by phosphorylating insulin receptor substrate(IRS) proteins, thus producing docking sites for the recruitment and activation of PI3K (b). This kinase converts phosphatidylinositol4,5diphosphate (PIP2) to phosphatidylinositol3,4,5trisphosphate (PIP3), which serves as a platform for the recruitment of phosphoinositide-dependent kinase 1 (PDK1) and AKT. When at the plasma membrane, AKT is phosphorylated by PDK1 and mammalian target of rapamycin complex 2 (mTORC2), which results in AKT activation. AKTpromotes GSV exocytosis by phosphorylating and inactivating two GTPase-activating proteins (GAPs), AS160 and the RALGAP complex (RGC, consisting of a regulatory subunit (RGC1) and a catalytic subunit (RGC2)), which regulate small GTPases that are involved in GSV retention and targeting, respectively.
Adipokines
Hormones and cytokines that are released by adipocytes and signal to other tissues to alter feeding behaviour and metabolism.
Small GTPases
20 35 kDa guanine nucleotide-binding proteins that switch between an inactiveGDP-bound conformation andan active GTP-bound conformation.
also affect other stages of GLUT4 trafficking, including endocytosis, sorting in the endosomal system and the formation of GSVs. The development of new tools for high-resolution analysis of the different intra cellular membrane compartments that contain GLUT4 has aided in pinpointing some of the steps that regulate GLUT4cycling. PI3K-dependent insulin signalling. The PI3Kdependent signalling pathway is initiated following insulin binding to its cognate Tyr kinase receptor on the cell surface, which leads to the recruitment and Tyr phosphorylation of the insulin receptor substrate (IRS) family of adaptor proteins6,1820.Tyr-phosphorylated IRS proteins serve as docking sites for the SH2 domain of the p85 regulatory subunit of class I PI3K, and the interaction of IRS proteins and PI3K results in PI3K activation and the subsequent synthesis of phosphatidylinositol3,4,5 trisphosphate (PtdIns(3,4,5)P3) from PtdIns(4,5)P2 at the plasma membrane. PtdIns(3,4,5)P3 in turn serves as a docking site for several PH domain-containing Ser/Thr kinases that are implicated in glucose uptake, including phosphoinositide-dependent kinase1 (PDK1) and AKT (also known as PKB) 21. PDK1 and mTORC2 activate AKT through dual Ser/ Thr phosphorylation. Knockout and knockdown studies have demonstrated that IRS adaptor proteins and AKT are absolutely necessary for insulin-stimulated glucose uptake2224, and
Guanine nucleotide exchange factors

(GEFs). A family of enzymes that activate GTPases by catalysing GDP release, thusallowing cytoplasmic GTPto bind to the GTPase.
GTPase-activating proteins
(GAPs). A family of enzymes that inactive GTPases by catalysing GTP hydrolysis.
SNARE regulatory proteins

(Soluble N-ethylmaleimidesensitive factor attachment protein receptor regulatory proteins). A family of small helical proteins that bridge twomembranes and drive membrane fusion events.
overexpression of constitutively active AKT can largely, but not completely, mimic the effect of insulin25,26. AKT is a central hub connecting insulin signalling with downstream regulators of GLUT4 trafficking. Microscopy studies that examined GLUT4 trafficking in the presence of AKT inhibitors or dominant-negativ e AKT constructs indicate that the kinase affects the exocytic arm of the GLUT4 trafficking itinerary, probably by regulating the translocation, targeting and fusion of GLUT4containing vesicles2729. The most extensively studied targets of AKT are the RAB GAP AS160 (also known as TBC1D4, which targets RAB8, RAB10 and RAB14) and the RALGAP complex (RGC, which is comprised of a regulatory subunit (RGC1) and a catalytic subunit (RGC2) and targets RALA). These small GTPases are involved in GLUT4 vesicle translocation and targeting to the plasma membrane28,30,31. Some SNARE regulatory proteins, including Synip (also known as STXBP4) and CDP138 (138 kDa C2 domaincontaining phosphoprotein), are also direct substrates for AKT and regulate GLUT4 vesicle fusion with the plasma membrane3234. Whether AKT regulates other steps in GLUT4 trafficking, including endocytosis, sorting and GSV formation, is currently unclear. The search for additional AKT targets in muscle and adipose cells continues and will probably give further insight into the specific steps of GLUT4 trafficking that are regulated by thiskinase.
VOLUME 13 | JUNE 2012 | 385
REVIEWS
Adipocytes Cholesterol-dependent endocytosis Cholesterol
GLUT4 Dynamin
Adipocytes and muscle Clathrin-mediated endocytosis Clathrin AP2
Caveolin
RAB5 Dynein Microtubule
RAB5
RAB5 RAB5 RAB5
Sorting endosome
Figure 2 | Molecular mechanisms of GLUT4 internalization. Glucose transporter type 4 (GLUT4) is internalized via clathrin-mediated andReviews cholesterol-dependent Nature | Molecular Cell Biology endocytosis. Internalization of GLUT4 through clathrin-mediated endocytosis requires the adaptor protein AP2. AP2 coordinates packaging of this glucose transporter into endocytic vesicles by recruiting clathrin to the plasma membrane and by binding to an aminoterminal F5QQI sequence in GLUT4. Vesicle scission from the plasma membrane requires the GTPase dynamin, which assembles at the neck of the invaginating vesicle and, following GTP hydrolysis, constricts and breaks the membrane. Cholesteroldependent endocytosis requires the membrane-deforming protein caveolin, which oligomerizes in lipid rafts and forms large cave-like structures, or caveolae. Cholesteroldependent endocytosis is likely to also require dynamin to liberate caveolae from the cellsurface. Vesicles carrying newly endocytosed cargo move towards the cell interior on dynein motors, which attach to the vesicles via RAB5. Endocytosed vesicles fuse withendosomes, where their cargo is sorted for degradation or recycling. Whereas clathrin-mediated endocytosis is the predominant mode of GLUT4 endocytosis in muscle cells, both clathrin-mediated and caveolin-dependent endocytosis have been implicated in GLUT4 endocytosis in adipocytes. In adipocytes, an insulin-stimulated switch from cholesterol-dependent to clathrin-mediated endocytosis may alter the rate of GLUT4internalization.
GLUT4 vesicle retention and translocation. The TC10 effector EXO70, which is a subunit of the exocyst tethering complex, has been implicated in GLUT4 vesicle targeting. Together, these molecular targets support a role for APS signalling in the release of intracellular GSV retention mechanisms and in GLUT4 vesicle targeting to the plasma membrane. Disrupting individual components of the PI3Kdependent or the APS signalling pathways in adipocytes, either with pharmacological inhibitors or by small interfering RNA (siRNA)-mediated knockdown, inhibits GLUT4 exocytosis and glucose uptake22,24,4144, which suggests that these pathways converge to control the GLUT4 trafficking machinery. In muscle cells, the PI3Kdependent signalling pathway is absolutely required for GLUT4 exocytosis; however, the necessity of the APS signalling pathway is unknown. Invivo the situation is less clear; knockout studies have shown that AKT isoforms seem to have different effects on cellular metabolism and growth21,45. Although knockout mice for the majority of components in the APS signalling pathway have not yet been analysed, CAP-deficient mice are paradoxically more sensitive to insulin when placed on a high-fat diet than control mice46. The basis of this is unclear, but CAP is also expressed in many other tissue s, including macrophages, and seems to have a crucial role in these cells46. Moreover, CAP is a member of a larger family of sorbin homology (SoHo) proteins, and its function may be partially redundant with other member s of this family47. Despite extensive cell biology data supporting the role of insulin in increasing surface GLUT4 levels, many questions still remain regarding the mechanisms controlling GLUT4 localization, both in the absence and presence of the hormone. Below, we focus on the steps in GLUT4 cycling, our current understanding of the cellular machineries regulating these steps, and finally if and how insulin signalling intervenes at different stages of the GLUT4 traffickin g itinerary.
Lipid rafts
Rigid regions of the plasma membrane that are enriched in cholesterol and glycosphingolipids.
Effector
A protein that preferentially binds to an activated small GTPase.
Exocyst
An evolutionarily conserved protein complex that consists of eight subunits and targets exocytic vesicles to sites of docking and fusion at the plasma membrane.
The APSinsulin signalling pathway. Insulin also initiate s a PI3Kindependent signalling pathway by recruiting the adaptor protein APS, which binds with high affinity to the activated insulin receptor 35. Following insulin receptor phosphorylation, APS recruits a complex that comprises the proto-oncogene c-CBL and c-CBL-associated protein (CAP)36,37. This triggers insulin receptor-catalysed Tyr phosphorylation of cCBL. Phosphorylated c-CBL then interacts with the adaptor protein CRK, which is in complex with the GEF C3G38. C3G in turn catalyses the activation of TC10, a member of the RHOfamily of small GTPases39,40 that is localized in lipid rafts in the plasma membrane. Active TC10 interacts with effector proteins that regulate GLUT4 vesicle exocytosis. The TC10binding protein CIP4 (CDC42interacting protein 4) forms a stable complex with the RAB GEF GAPEX5, which regulates the activity of RAB5 family GTPases that are involved in
GLUT4 endocytosis The amount of cell surface GLUT4 is determined by the net rate of GLUT4 endocytosis and exocytosis. In the absence of insulin, hypoglycaemia is avoided by rapid endocytosis and slow exocytosis of GLUT4 in adipose and muscle tissues48,49. After stimulation with insulin, surface GLUT4 levels increase, which may be a result of an increased exocytosis rate, a decreased endocytosis rate or the modulation of both. Although it is widely accepted that most of the effects of insulin occur through increased GLUT4 exocytosis48,49, whether and how GLUT4 endocytosi s is altered by insulin remains unclear. GLUT4 is endocytosed through at least two separate pathways: clathrin-mediated endocytosis and cholesteroldependent endocytosis5052, both of which internalize GLUT4 into a sorting endosome compartment (FIG.2). In muscle cells, clathrin-mediated endocytosis is mainly responsible for GLUT4 internalization; however, adipocytes use both the clathrin-mediated and the cholesteroldependent pathways53. Internalization of GLUT4 via clathrin-mediated endocytosis probably requires the
REVIEWS
adaptor protein AP2, which recognizes and interacts with an F5QQI sequence at the amino ter minus of GLUT4 (REFS 51,54,55). By contrast, cholesteroldependen t internalization of GLUT4 seems to require caveolin, a protein that localizes in lipid rafts50,52; however, it should be noted that several studies have found no enrichment of GLUT4 in caveolae in adipocytes or muscle cells56,57. Abscission of GLUT4containing clathrin-coated pits, and probably also caveolae, requires the GTPase dynamin58,59. Once formed, movement of endocytosed vesicles towards the cell interior is mediated by microtubule-associated dynein motors, which attach to endocytic vesicles through the small GTPase RAB5 in adipocytes60,61. The relative importance of these two independent endocytic pathways is unclear. Although some studies have suggested that the two routes may endocytose GLUT4 at different rates, it is also possible that these pathways sort GLUT4 into distinct vesicle pools in the endosomalsystem. Reconciling the role of insulin. Whereas insulin does not affect GLUT4 internalization rates in muscle cells62, theeffect of the hormone in adipocytes is less clear. Several studies have reported that insulin augments surface GLUT4 levels in adipocytes not only by increasing the rate of exocytosis but also by decreasing the rate of GLUT4 endocytosis by ~two- to threefold48,51,63,64. Treatment with the cholesterol-depleting drug nystatin blocks GLUT4 internalization in basal adipocytes, whereas knockdown of Ap2 blocks GLUT4 internalization in insulin-stimulate d adipocytes, which suggests that stimulation with the hormone results in a switch from cholesterol-dependent internalization to clathrin-mediated internalization51. A difference in the efficiency of GLUT4 internalization via caveolin and clathrin could result in a slower rate of GLUT4 internalization in the presence of insulin. Although decreased internalization of GLUT4 in response to insulin in adipocytes would seem to support the effect of the hormone in increasing glucose transport, the importance of a small reduction in the endocytic rate in modulating overall surface levels of GLUT4 is controversial. Furthermore, it was recently suggested that the methods used previously to measure the endocytic rate in adipocytes did not properly account for the rate of GLUT4 exocytosis, and that this led to a perceived decrease in endocytic rates65. After correcting for these factors, no change was observed in the rate of endocytosis after the addition of insulin. Other studies in adipocytes have found that insulin increases GLUT4 trafficking through the recycling system49,66, suggesting that endocytosis of the transporter may be important for maintaining a pool of endosomes that quickly return to the plasma membrane. Although the jury is still out on this point, it seems reasonable that GLUT4 endocytosis might be unchanged or even increased by insulin to create a readily releasable pool of vesicles. compartments, including recycling endosomes , late endosomes and the TGN, before returning to the plasma membrane67. This diversity of recycling pathways allows cells to regulate the protein and lipid composition at distinct locations within the cell. In addition to the recycling membrane system, some specialized cell types contain non-secretory storage compartments that retain specific membrane proteins in the cell until a stimulus signals for their exocytosis to the plasma membrane1. A single GLUT4 molecule undergoes multiple cycles of exocytosis and endocytosis during its lifetime. Studiesof GLUT4 trafficking indicate that in unstimulated adipo cytes, a pool of GLUT4 constantly traffics between the cell surface and the cell interior through endosomal pathways; however, at least 50% of the transporter pool is sequestered away from this recycling system in specialized post-endosomal GSVs that undergo exocytosis in response to insulin49,66,68 (FIG.3). Dynamic regulation of the GSV. The existence of a GLUT4 storage compartment was proposed on the basis of ablation studies that depleted adipocytes of endosomes that contain transferrin receptor (TfR), a protein that traffics through the recycling endosomal system and only exhibits a twofold change in translocation in response to insulin6870. These studies showed that, in the basal state, at least half of the GLUT4 population is found in a vesicle compartment that largely lacks TfR and other recycling proteins, including mannose-6phosphate receptor (M6PR) and GLUT1. Stimulation with insulin depletes a proportion of these GLUT4enriched vesicles, and this led to the hypothesis that adipocytes and muscle cells sort GLUT4 into a special ized pool of insulin-sensitive vesicles71. The isolation and biochemical characterization of these GSVs has been challenging, mainly because GLUT4 and other proteins that are enriched in these vesicles also traffic through the endosomal system, resulting in a lack of GSV-specific markers. For example, electron microscopy studies showed that GLUT4 is found in all endo somal compartments8,72. Efforts to purify and analyse GSVs have used a combination of techniques, including compartment ablation, cell fractionation, immuno-isolatio n using GLUT4specific antibodies and immunodepletion of other vesicle populations using antibodies against non-GSV proteins, followed by immunoanalysis or mass spectrometry analysis. However, it is important to note that, so far, the purity of the GLUT4containing vesicles used in mass spectrometry analysis studies has been hard to ascertain. Nevertheless, these studies have provided valuable insight into the protein composition of GSVs and have identified GLUT4, insulin-regulated aminopeptidase (IRAP), sortilin, vesicle-associated membrane protein 2 (VAMP2) and low-density lipoprotein receptor-related protein 1 (LRP1) as major components of these vesicles69,7375. Although the functions of these proteins in GSVs are not completely understood, sortilin, IRAP and LRP1 all positively affect GLUT4 sorting into GSVs75,76, whereas VAMP2 is the vSNARE required for fusion of these vesicles with the plasma membrane77,78.
VOLUME 13 | JUNE 2012 | 387 2012 Macmillan Publishers Limited. All rights reserved
Hypoglycaemia
Physiological condition that is defined by abnormally low blood glucose levels.
Clathrin-mediated endocytosis
A mechanism for internalizing extracellular molecules and portions of the plasma membrane. This pathway is dependent on the membrane curvature-inducing coat proteinclathrin.
Cholesterol-dependent endocytosis
A clathrin-independent mechanism for internalizing molecules. This mechanism is blocked by drugs that deplete cellular cholesterol and often requires the lipid raft protein caveolin.
Sorting endosome
A membrane compartment that is localized close to the cell surface where recently endocytosed proteins are delivered and sorted for degradation or recycling.
Recycling endosomes
Membrane compartments thatmany recycling proteins pass through before returning to thecell surface.
GLUT4 sorting and retention After endocytosis, recycled membrane proteins can return quickly to the plasma membrane from sortin g endosomes or they can sort through intracellular
NATURE REVIEWS | MOLECULAR CELL BIOLOGY
REVIEWS
a
Plasma membrane TC10 Exocyst Slow recycling
RAB5
RAB4 Fast recycling
GSVs
RALA
Sorting endosome Recycling endosome
GSVs ARF6 RAB11 Adaptor-driven protein sorting Insulin
RAB8, RAB10, RAB14
RAB31
Insulin Formation of GSV multimeric complexes
Trans-Golgi network
ARF6 Clathrin
Adaptor protein
GLUT4 GLUT1
IRAP LRP1
VAMP2
Transferrin receptor
Donor compartment Sortilin
Figure 3 | The GLUT4 trafficking itinerary. a | Glucose transporter type 4 (GLUT4) is constantly between Nature Reviews cycled | Molecular Cell the Biology plasma membrane and intracellular membrane-bound compartments. After internalization, GLUT4 can take several different paths back to the plasma membrane. GLUT4 can recycle back to the plasma membrane from sorting endosomes (termed fast recycling) or recycling endosomes (termed slow recycling), or this glucose transporter can be packaged into GLUT4 storage vesicles (GSVs) that bud from recycling endosomes and/or the trans-Golgi network (TGN). GSVs do not cycle to the plasma membrane in the absence of insulin owing to intracellular retention mechanisms that may involve futile cycling with recycling endosomes or the TGN, or physical anchoring to an intracellular structure. Insulin increases surface GLUT4 levels 10- to 40fold by increasing GLUT4 exocytosis from GSVs and endosomes. GLUT4 trafficking throughthe endosomal system and the formation of GSVs require the actions of multiple small GTPases, including RAB5, RAB4,RAB11, RAB31, ARF6 and RALA. b | Hypothetical model for GSV formation. The GSV resident proteins GLUT4, insulin-regulated aminopeptidase (IRAP), sortilin and low-density lipoprotein receptor-related protein 1 (LRP1) interact with each other through their lumenal domains to form oligomeric protein complexes in the GSV donor compartments. The small GTPase ARF6 drives vesicle formation on a subdomain of donor membranes by recruiting adaptor proteins, which interact with clathrin and GSV-resident proteins. In the absence of insulin, GSVs may be prevented from cycling tothe plasma membrane through a TGNGSV futile cycle that is maintained by active RAB31. Insulin may increase GSVrelease by inactivating RAB31 and by activating RAB10 in adipocytes and RAB8 and RAB14 in muscle cells.
GSVs do not cycle to the cell surface in the absence of insulin, or do so only very slowly, suggesting that adipocytes and muscle cells contain a specialized machinery that retains this vesicle population within the cell49,66,79. Although insulin causes rapid translocation of GSVs to the plasma membrane, prolonged stimulation with the hormone (~15minutes) triggers GLUT4 movement to the plasma membrane in endosomes, suggesting that when preformed GSVs have been depleted, GLUT4 recycling from endosomes is favoured rather than reforming of new GSVs49,80. Thus, insulin may increase surface GLUT4 levels by acting on at least two processes in GLUT4 trafficking: exocytosis of GSVs and recycling via endosomes (FIG.3). Small GTPases regulate GLUT4 sorting. Understanding GLUT4 sorting and retention has proven difficult for several technical reasons. In addition to the difficulty in isolating a pure population of GSVs for characterization, the organization of the perinuclear endomembrane system from which GSVs arise is not well defined, and so trafficking of GLUT4 through this system is poorly
388 | JUNE 2012 | VOLUME 13
understood. Furthermore, because GLUT4 trafficking between the interior of the cell and the cell surface is a circular process, disruption of one trafficking step necessarily affects all other steps. Thus, the identification of the key players for any particular step is difficult. After endocytosis into sorting endosomes, GLUT4 is sorted into recycling endosomes and/or the TGN, and one or both of these compartments probably serve as the donor membrane for GSVs81,82. GLUT4 sorting and delivery into GSVs relies on the actions of small GTPases of the RAB and ARF families, which assemble effectors that mediate vesicle budding, transport and fusion. In adipocytes and muscle cells, RAB4, RAB5, RAB8, RAB10, RAB11, RAB14 and RAB31 have been implicated in regulating different steps in GLUT4 sorting, although additional small GTPases (including ARF6 and RALA) have been found to associate with GLUT4containing vesicles, and some of these GTPases also affect GSV formation83. RAB5, which drives homo- and heterotypic early endosomal fusion, is activated at the plasma membrane by insulin through TC10, which recruits the RAB5 GEF GAPEX5 to the
REVIEWS
cell surface. RAB5 may regulate mobility and sorting of GLUT4containing vesicles after endocytosis61,84,85. The generation of PtdIns(3)P by RAB5 has also been shown to regulate maturation of endosomes86. The role of RAB4 in GLUT4 trafficking is poorly understood; however, given its known function in protein sorting in endosomes in other systems85, and the observation that manipulating RAB4 function in adipocytes alters GLUT4 levels at the plasma membrane87,88, RAB4 may promote fast recycling of the transporter from sorting endosomes. GSV formation. Disrupting RAB11 function in adipocytes by overexpressing the binding domain of its effector protein RAB11binding protein (RAB11BP) increases the amount of GLUT4 that colocalizes with TfR, inhibits basal TfR and GLUT4 recycling and decreases insulin-stimulated GLUT4 translocation79. Together, these data suggest a role for RAB11 in GSV formation and traffickin g in the endosomalsystem. The AKT substrate AS160 is a RAB GAP that targets RAB8 and RAB14 in muscle cells89 and RAB10 in adipocytes83,90. These RABs have a positive role in GLUT4 translocation, suggesting that they may regulate GSV formation and/or intracellular retention8992. As insulin stimulates AKT-catalysed phosphorylation of AS160, blocking the function of the protein, this should relieve the inhibitory effect of AS160 on its target RABs93,94 (FIG.1). Whereas the activation of RAB8 and RAB14by insulin has been demonstrated in muscle cells89, so far the activation of RAB10 in adipocytes has not been detected92. Nevertheless, RAB10 is necessary for maximal GLUT4 exocytosis in response to insulin, and several lines of evidence indicate that cycling of this small GTPase may increase glucose uptake90,95. Together, the activation of RAB4, RAB5, RAB8, RAB11 and RAB14 seems to be a key means by which insulin signalling pathways alter GLUT4 trafficking96,97. GSV budding from recycling endosomes and/or the TGN depends on vesicle coats that assemble on subdomains of the donor compartments (FIG.4). Although the molecular forces that drive the packaging of GSV components into budding vesicles are incompletely understood, the observation that GLUT4, IRAP, sortilin and LRP1 interact with one another through their lumena l domains has led to a mass action hypothesis of GSV formation. This hypothesis posits that GSV components, which are highly expressed in mature adipocytes and muscle cells compared with other recycling proteins, find and interact with one another due to their high concentration in donor compartments. The formation of large multimeric complexes that contain GSV components would then allow for efficient sorting of these proteins into newly forming vesicles75,76,98. Adaptor proteins, which are recruited to GSV donor membranes by the small GTPase ARF6, interact with GSV components and clathrin to facilitate protein sorting and vesicl e buddin g. The adaptor proteins GGA (Golgi-localized -ear-containing ARF-binding protein), ACAP1 (ARFGAP with coiled-coil ankyrin repeat and PH domain-containing protein 1), AP1 and AP3 have
all been implicated in GSV formation99102. Although the specific roles of these different adaptor proteins are not currently understood, it is possible that they act in a coordinated fashion to drive vesicle budding from one donor membrane. Alternatively, they may assemble ondifferent membrane-bound compartments and act at different steps in GLUT4 trafficking (for example, sorting at early endosomes and recycling endosomes) to ultimately drive GSV formation. GSV retention. GLUT4 is sequestered away from the recycling system in the basal state through intracellular retention of GSVs, which occurs through two mechanisms that may require a common machinery: futile cycling and retention by anchoring proteins. Afutile cycle in which GSVs, the TGN and/or recycling endosomes undergo repeated cycles of fusion and fission is thought to disfavour GLUT4 translocation to the plasma membrane in the absence of insulin82. Little is known about the proteins that are necessary for maintaining a futile cycle, but the opposing actions of at least two RAB proteins are probably involved. RAB31 is hypothesized to participate in futile cycling in the basal state; in support of this, RAB31 negatively regulates GLUT4 translocation in adipocytes, and it participates in TGN-to-endosome trafficking in other cell types103,104. Insulin may promote the release of GLUT4 from futile cycling by recruiting the RAB31 GEF GAPEX5 to the plasma membrane, thereby decreasing RAB31 activity on intracellular membranes103. Anchoring proteins are hypothesized to promote GSV retention by either physically tethering GSVs to an unidentified intracellular site or by constraining GSV cycling to minimize exchange with the plasma membrane105,106. For example, TUG (tether containing UBX domain for GLUT4) interacts with GLUT4 and may anchor GSVs within the cell107. Overexpression of TUG sequesters GLUT4 away from TfR-positive structures, whereas the disruption of TUG function enhances basal glucose uptake and GLUT4 levels at the cell surface and increases dispersion of GLUT4 throughout the endosomal system106,107. Although these data may indicate a role for TUG as a scaffold that physically tethers GSVs, these results could also be consistent with a role for TUG in intracellular sorting of GLUT4. Insulin abrogates the interaction between TUG and GLUT4, suggesting that TUG is an important target of insulin action107.
Exocytosis of GSVs Insulin increases the amount of GLUT4 at the cell surface mainly by promoting exocytosis of GSVs48,108 and to a lesser extent by increasing exocytosis from the recycling system70. GSV exocytosis can be separated into three distinct processes: translocation to the cell periphery; targeting; and fusion (FIG.4). The application of total internal reflection fluorescence micro scopy (TIRFM) (BOX2) to study GSV exocytosis has shown that insulin probably regulates each of these steps by recruiting GLUT4containing vesicles to the cell periphery and then modulates the rates of targeting and fusion27,28,109,110. These imaging studies are consistent with earlier
VOLUME 13 | JUNE 2012 | 389
REVIEWS
a Translocation
Plasma membrane AKT RGC AKT AS160 P RAB8 RAB10 RAB14 Insulin GSV GLUT4 TUG TUG Kinesin Microtubule RALA MYO1C PKC Actin lament TC10 EXO70 P EXO84
b Targeting and disengagement

Exocyst SAP97 SEC5
c Fusion
STX4 SNAP23 CDP138 P Synip P AKT MUNC18C DOC2 P VAMP2
Insulin Insulin receptor
RALA P
Figure 4 | Insulin targets several steps in GLUT4 storage vesicle exocytosis. a | Insulin stimulates glucose Nature Reviews | Molecular Cell Biology transporter type 4 (GLUT4) storage vesicle (GSV) translocation by inhibiting the interaction between TUG (tether containing UBX domain for GLUT4) and GLUT4 and by stimulating AKT-catalysed phosphorylation of AS160, which inhibits the GTPase-activating protein (GAP) activity of AS160 towards RAB8, RAB10 and RAB14. Microtubule-based kinesin motors drive GSV movement from the perinuclear region to the periphery of the cell. Close to the cell surface, GSVs move along a meshwork of cortical actin to targeting sites on the plasma membrane. The actin-based myosin motor MYO1C, which drives these movements, recognizes GSVs by interacting with the vesicle-localized small GTPase RALA. Insulin may regulate the movement of GSVs along cytoskeletal tracks by increasing the ATPase activity of motor proteins or by modulating recognition of vesicles by molecular motors via factors such as calmodulin. b | Insulin assembles GSVtargeting sites at the plasma membrane by activating the small GTPase TC10. GTP-loaded TC10 and lipid-raft localized synapse-associated protein 97 (SAP97) recruit and stabilize the exocyst complex at the plasma membrane. Simultaneously, AKT activates RALA by phosphorylating and inhibiting the RALGAP complex (RGC, consisting of a regulatory subunit (RGC1) and a catalytic subunit (RGC2)). GTP-loaded RALA guides GSVs to targeting sites on the plasma membrane by interacting with the exocyst subunits SEC5 and EXO84. After the arrival of GSVs at the plasma membrane, protein kinase C (PKC)-catalysed phosphorylation of SEC5 inhibits the interaction between the exocyst and RALA, thus allowing GSVs to disengage from the targeting machinery before vesicle fusion. c | GSV fusion with the plasma membrane is driven by the assembly of SNARE complexes comprised of vesicle-localized vesicle-associated membrane protein 2 (VAMP2), plasma membrane-localized syntaxin 4 (STX4) and synaptosomal-associated protein 23 (SNAP23). The activated insulin receptor phosphorylates the fusion regulatory protein MUNC18C, which then stimulates GSV fusion, perhaps by recruiting DOC2 (double C2like domain-containing protein-) to SNARE complexes. AKT catalyses phosphorylation of the syntaxin 4 binding protein Synip, which may release its inhibition of GSV fusion. AKT also phosphorylates and recruits CDP138 to the cell surface, where CDP138 acts as a positive regulator of GSV fusion.
evidence from biochemical and cell biological studies that identified molecular motors18,96,111, tethers112 and fusion proteins33,113115 as important components of the cell machinery that are necessary for insulin-stimulated GLUT4 exocytosis. Translocation. Insulin stimulates the accumulation of GLUT4containing vesicles at the periphery of the cell within 5minutes of exposure to the hormone116,117. High-resolution imaging of GLUT4 shows that GLUT4containing vesicles move along the cytoskeletal tracks in a linear fashion both in the presence and absence of insulin110,116119. Treatment with drugs that depolymerize either microtubules or actin partially blocks insulin-stimulated accumulation of GLUT4 at the plasma membrane and glucose uptake117,120122, and disruption of both microtubules and actin completely inhibit s the effects of insulin on GLUT4 relocalization 116,123.
390 | JUNE 2012 | VOLUME 13
Whereas microtubules participate in long-range movement of GSVs to the periphery of the cell, cortical actin is involved in short movements that are coupled to docking and/or fusion near the plasma membrane123126. Whether insulin regulates GSV movement on cytoskeletal tracks is still debated. Some studies have reported an increase in the speed of movement or an increase in the recognition of GSVs by the cyto skeleton following insulin stimulation29,96,127.By contrast, other groups report that GLUT4containing vesicles are constantly moving along cytoskeletal tracks in the basal state, suggesting that they continually sample the plasma membrane. In this model, insulin would regulate the recognition of GSVs at the plasma membrane, rather than the mobilization of GSVs themselves27,29,110,119. Several microtubule-based kinesin and actin-based myosin motors have been implicated in GSV transport along cytoskeletal tracks, including the kinesins
REVIEWS
Box 2 | Study of GLUT4 vesicle exocytosis by TIRFM
Over the past 10years, the understanding of how insulin affects glucose transporter type 4 (GLUT4) trafficking to the plasma membrane has been greatly aided by total internal reflection fluorescence microscopy (TIRFM). This technique allows the visualization of events near the cell surface by illuminating a region of the cytoplasm that is within ~100200nm of the plasma membrane. The first TIRFM experiments tracked the movements of hundreds of single GLUT4-containing vesicles in live cells and showed that insulin increases the amount of GLUT4 within the visualized region, thereby confirming previous biochemical and cell biological evidence that suggested that insulin predominantly targets GLUT4 exocytic trafficking28. Advances in TIRFM technology in the past few years have enabled researchers to study individual steps in GLUT4 vesicle exocytosis by using automated high-resolution capture of exocytic events27,109 and dual-colour TIRFM33,80.pHsensitive fluorophores, which are quenched in the acidic vesicle lumen but fluoresce after opening of the fusion pore at the plasma membrane, have also been used to functionally separate targeting of GLUT4 vesicles versus fusion. For example, TIRFM imaging of the vesicle-localized SNARE VAMP2 (vesicle-associated membrane protein 2) and GLUT4 showed that GLUT4 fluorescence can be observed during both targeting and fusion, whereas VAMP2 fluorescence is observed specifically during fusion (see the figure). Together, these studies have shown that insulin increases the number of vesicles in the TIRF zone, decreases the duration of vesicle targeting and increases the frequency of fusion events27,109. Advances in TIRFM technology have also allowed researchers to investigate the roles of specific proteins in GLUT4 exocytosis, including TUG (tether containing UBX domain for GLUT4) and CDP138 (138 kDa C2 domain-containing phosphoprotein), which affect vesicle retention and fusion, respectively80,33. However, there are still limitations of this technique; in particular, the study of GLUT4exocytosis is restricted by the inability to easily distinguish different types ofGLUT4containing vesicles. As a consequence, TIRFM studies examine a mixed population of vesicles that include both GLUT4 storage vesicles (GSVs) and endosomes. These vesicles are likely to approach and fuse with the plasma membrane with different kinetics, different dependencies on insulin action and different molecular machinery. The development and use of techniques that can reliably identify GSVs, including a new approach that separates vesicle carriers on the basis of size80, will be important for elucidating the functions of specific signalling and trafficking proteins in GSV exocytosis. Immunofluorescence images are reproduced, with permission, from REF.80 (2011) The Rockefeller University Press. Translocation Targeting Fusion STX4 SNAP23 Plasma membrane TIRF zone Exocyst
constructs or by disruption of RAB proteins that they interact with, inhibits insulinstimulate d GLUT4 translocation96,111,119,127. So far, the role of the motor protein MYO1C in GSV mobility has been most extensively investigated. MYO1C moves from the cytoplasm to the plasma membrane in response to insulin111 and targets GSVs to sites of fusion at the plasma membrane. MYO1C does not recognize GSVs directly; instead, the motor protein is coupled to these vesicles through its interaction with RALA on GSVs128. Calmodulin is proposed to modulate MYO1C function in adipocytes, but its precise role is not completely understood. The interaction between MYO1C and RALA has been shown to be calmodulin-dependent, suggesting that this protein might potentiate GSV recognition by the actin cytoskeleton128. However, insulin has also been seen to decrease the association between MYO1C and calmodulin, suggesting that calmodulin release might be necessary for the motor activity of MYO1C18. Thus, the role of calmodulin in GSV translocatio n remains to beelucidated. Targeting. The delivery of vesicles to the correct cellular compartment requires a targeting step in which the vesicles are initially recognized by the destination membrane. In adipocytes, GSV targeting to lipids rafts on the cell surface is important for efficient insertion of GLUT4 into the plasma membrane133. After stimulation of cells with insulin, the level of immunodetectable GLUT4 increases in the plasma membrane fraction several minutes before GLUT4 is exposed extracellularly, raising the possibility that insulin regulates GSV targeting and/or fusion at the plasma membrane48. TIRFM studie s carried out in adipocytes also indicate that insulin increases the rate and decreases the duration of tethering27,110. However, these studies are limited by the ability to clearly distinguish tethering from fusion. In this regard, the application of pHsensitive fluorophores that gain their fluorescence signal only after the fusion pore opens should more clearly define the initiation of fusion and thus provide insights into distinct events that occur at the plasma membrane (BOX2). A crucial component of the GLUT4 targeting machinery is the exocyst, which is an evolutionarily conserved octameric complex that assembles at sites of exocytosis and tethers exocytic vesicles to the plasma membrane134. The exocyst is thought to mediate theinitial flexible contact between exocytic vesicles and the plasma membrane from a relatively long distance and can thus concentrate GSVs before the final membrane fusion step. Inhibition of exocyst assembly in adipocytes disrupts fusion of GSVs without affecting their translocation, demonstratin g that this complex is necessary for vesicle targeting at the plasma membrane112. Insulin regulates exocyst-mediated targeting through three steps: exocyst assembly; engagement of the exocyst by GSVs; and GSV disengagement from the exocys t to enable fusion112,128,133,135. First, insulin directs the assembly of the exocyst at the plasma membrane by promoting an interaction between active TC10 and the exocyst scaffolding subunit EXO70 (REF.112).
VOLUME 13 | JUNE 2012 | 391
VAMP2 GSV Glut4 GLUT4: VAMP2: Merge:
0.3
0.6
0.9
Time (sec)
1.2 1.5
1.8
2.1
2.4
2.7
Targeting
Fusion
SNAP23, synaptosomal-associated protein 23; STX4, syntaxin 4. Nature Reviews | Molecular Cell Biology
KIF3(REFS96,123) and KIF5B119, myosin 5A (MYO5A)127, MYO5B91 and MYO1C18,111,123,128. In yeast and mammalian cells, small GTPases couple motor proteins to vesicles61,96,128132. In adipocytes, KIF3 (REF.96) interacts with RAB4, and MYO1C interacts with RALA128. Inhibiting any of these motor proteins with blocking peptides, by siRNAmediated knockdown, expression of dominant-negative
REVIEWS
EXO70is constitutively associated with other exocyst sub units and thereby assembles the complex at the plasma membrane133. Other interactions between exocyst subunits and the plasma membrane facilitate complex assembly at discrete locations. For example, an inter action between the exocyst subunit SEC8 (also known as EXOC4) and synapse-associated protein 97 (SAP97; also known as DLG1) in lipid rafts anchors the exocyst complex in this subdomain of the plasma membrane133. EXO70 and SEC3 (also known as EXOC1) interact with PtdIns(4,5)P2, and may thus target the exocyst to zones of the plasma membrane that are enriched in this phospholipid136138. PtdIns(4,5)P2 is a known constituent of lipid rafts, and disruption of exocyst assembly in raft subdomains inhibits GLUT4 insertion into the plasma membrane and glucose uptake, further implicating this membrane subdomain in GLUT4 targeting133,139. Thus, multiple interactions between plasma membrane constituents and exocyst subunits coordinate exocyst assembl y in lipid rafts and localized GLUT4 targeting. The exocyst is recognized by GSVs via the small GTPase RALA, which is present on GLUT4containing vesicles. Insulin controls RALA activity by inhibiting a complex of proteins with RAL-GAP activity. The RAL-GAP complex is composed of RGC1 and RGC2, which contains a GAP domain with specific activit y towards RAL-GTPases. AKT-catalysed phosphorylation ofRGC2 on at least three residues inhibits the complex and allows for GTP loading on RALA30. siRNAmediated knockdown of RGC1 or RGC2 increases RALA activit y and insulin-stimulated glucose uptake, demonstratin g the regulatory role of this complex30. Moreover, siRNAmediated knockdown of Rala or overexpression of a dominant-negative RALA mutant blocks insulinstimulate d glucose uptake and GLUT4 insertion into the plasma membrane. By contrast, constitutively active RALA mutants increase the effect of insulin, indicating that the activation of this small GTPase is required for insulin-regulated GLUT4 exocytosis128. Once activated, RALA interacts with exocyst subunits SEC5 (also known as EXOC2) and EXO84 (REFS128,140,141). Although the precise role of these two RALA-binding proteins remains uncertain, both are required for insulin-stimulated glucose uptake128. Large tethering complexes such as the exocyst are thought to disengage from vesicles and/or disassemble before fusion occurs142. Disengagement may allow for fusion of opposing membranes and recycling of the tethering machinery for additional rounds of vesicle targeting. Indeed, RALA dissociates from the exocyst through insulin-dependent, protein kinase C (PKC)-catalysed phosphorylation of Ser89 in the RALA-binding domain of SEC5, which triggers SEC5 release135. Mutation of Ser89 to either Ala (to block phosphorylation) or to Asp (to mimic phosphorylation) blocks GLUT4 insertion into the plasma membrane, suggesting that both engagement and disengagement from the targeting machinery are necessary steps preceding fusion135. This phosphorylatio n-dependent release of GSVs from the exocyst raises the possibility that the exocyst also serves a gatekeeper function in controlling GSVfusion.
Although RAB10 has been implicated in insulinstimulated GSV translocation to the cell surface in adipocytes, it may have an additional role at the plasma membrane. This function of RAB10 has largely emerged from TIRFM experiments performed in adipocytes by overexpressing AS1604P (in which the four phosphorylation sites have been mutated to Ala), a mutant form of the RAB10 GAP that is not regulated by insulin. These cells exhibit decreased tethering of GSVs at the plasma membrane, with little to no effect on fusion27,28,143. There are currently no known RAB10 effectors in adipocytes; however, in MadinDarby canine kidney (MDCK) cells, RAB10 interacts with the exocyst, and GTP loading on RAB10 may enhance this interaction144. It will be of interest to determine whether such an interaction is also regulated by insulin in adipocytes and, if so, how this small GTPase functions with RALA and TC10 to regulate exocyst-mediated targeting of GSVs to the plasma membrane. SNARE regulation during GSV fusion. Following vesicle tethering, SNARE proteins on both the GSV and the plasma membrane (the donor and acceptor compartments, respectively) form tight complexes that provide the driving force for membrane mixing and fusion145. GSV fusion with the plasma membrane is driven by SNARE ternary complex assembly between the GSVlocalized vSNARE VAMP2, the plasma membrane-localized tSNARE syntaxin 4 (STX4) and its plasma membrane-associated accessory protein synaptosomal-associated protein 23 (SNAP23)77,78,146148. TIRFM studies indicate that insulin increases the rate of GSV fusion, probably in an AKT-dependent manner80,143. Proteins that have been implicated as targets of insulin action include the SNARE regulatory proteins MUNC18C (also known as STXBP3), Synip (also known as STXBP4), tomosyn (also known as STXBP5) and CDP138, suggesting that SNARE complex assembly may be controlled by insulin149. The most extensively studied regulator of SNARE complex formation in GSV trafficking is the SEC1 MUNC18 family member MUNC18C, which binds to STX4 on the plasma membrane. SEC1MUNC18 proteins are thought to switch from negative regulators of SNARE assembly to indispensible components of the fusogenic machinery during the fusion process. This dual function may occur through a change in their mode of binding to syntaxins150. Studies in adipocytes support both a negative role151155 and a required role156158 for MUNC18C in GSV exocytosis. MUNC18C regulation by insulin is still not completely understood. Some studies have suggested that insulin alters the mode of binding between MUNC18C and STX4, thus enhancing its fusogenic function113,159. Alternatively, others have reported that the insulin recepto r catalyses phosphorylation of MUNC18C at Tyr219 and Tyr521, which results in its dissociation from STX4 (REF.115). Interestingly, in pancreatic -cells, Tyr phosphorylation of MUNC18C increases its binding to double C2like domain-containing protein- (DOC2) in a manner that precludes STX4 binding160.
REVIEWS
DOC2 belongs to a class of proteins that increase membrane curvature during fusion of synaptic vesicles with the plasma membrane161 and has been implicated as a positive regulator of GLUT4 exocytosis in adipocytes114. MUNC18C may promote fusion by recruiting DOC2 to sites of SNARE assembly at the plasma membrane. Two other STX4binding proteins, Synip and tomosy n, have been implicated as negative regulators of SNARE complex assembly in adipocytes. Stimulation with insulin decreases the binding affinity betweenSTX4 and Synip in adipocytes, thus presumably freeing STX4to interact with VAMP2 on the incoming GSV32. Insulin-mediated regulation of Synip is proposed to occur through AKT-catalysed phosphorylation at Ser99 (REFS 34,162), but the importance of this phosphorylation event has been disputed163. Although tomosyn can also compete with VAMP2 for binding to STX4 and SNAP23 in adipocytes, its regulation by insulin has not yet been investigated164. CDP138, a previously uncharacterized protein containing a Ca2+-binding C2 domain, has been identified as a novel AKT substrate that positively regulates GLUT4 insertion into the plasma membrane in adipocytes. Overexpression of a phosphorylation-deficient S197A mutant or a Ca2+- and lipid binding-deficient mutant of CDP138 blocked GLUT4 vesicle fusion but had no effect on vesicle accumulation at the plasma membrane33. The C2 domain of CDP138 shares homology with those found in synaptotagmin proteins, which promote vesicle fusion by binding to lipids in a Ca2+dependent manner and thereby induce membrane curvature165. Determining the mechanism by which CDP138 acts and elucidating the role of AKT-mediated phosphorylation in promoting its positive function in GLUT4 vesicle fusion are important areas for future investigations. transporter can follow through the complex cellular membrane network. New insights into how insulin signalling intersects with these trafficking intermediates have revealed that phosphorylation events initiated by the insulin receptor can lead to the regulation of small GTPases by modulating the activity or localization of GEFs and GAPs. These small GTPases in turn mediat e crucial steps and define: when and where GLUT4 cycles; the regulation of endocytosis; the entry into and the exit out of the recycling endosome system; the formation of specialized storage vesicles and the retention ofthese GSVs within the cell; and finally the release of the vesicles and the transport to discrete regions in the plasma membrane for tethering and fusion. Moreover, the generation of different phosphoinositide species downstream of these phosphorylation events seems to provide localization signals for these signalling events, thereby ensuring their efficiency. Despite this progress, many questions remain. How does GLUT4 move through the endomembrane system? Although it has been possible to study GLUT4 vesicle trafficking close to the plasma membrane with TIRFM, new microscopy techniques may permit the study of trafficking that occurs deeper within the cell. This should help to provide insights into how GSVs are formed and then retained in cells. For example, it is not yet clear whether GSVs are dynamically cycled back into the endosome network, or whether a physica l tether retains these vesicles in cells. It will also be important to understand how individual components of the exocyst complex contribute to GLUT4 targeting, and how they control accessibility of GSVs to the fusion machinery. All eight members of the exocyst complex are conserved from mammals to yeast, and in each species every component is required for secretion166. These findings suggest that the exocyst proteins are likely to have additional uncharacterized regulatory roles. How and why do these pathways differ between muscle and fat cells? It may take another 25years to answer these key questions, and hopefully new insights will be gained as to how these processes are altered in obesity and other states that give rise to insulin resistance.
Conclusions and future directions Nearly 25years after the discovery that insulin stimulates glucose uptake by altering the trafficking itinerary of GLUT4 in muscle and adipose cells, we are beginning to understand the varied paths that this hexose
1. 2. 3. 4. 5. 6. 7. 8.
Chieregatti, E. & Meldolesi, J. Regulated exocytosis: new organelles for non-secretory purposes. Nature Rev. Mol. Cell Biol. 6, 181187 (2005). Lowenstein, C.J., Morrell, C.N. & Yamakuchi, M. Regulation of WeibelPalade body exocytosis. Trends Cardiovasc. Med. 15, 302308 (2005). Brown, D. The ins and outs of aquaporin2 trafficking. Am. J.Physiol. Renal Physiol. 284, F893F901 (2003). Pfenninger, K.H. etal. Regulation of membrane expansion at the nerve growth cone. J.Cell Sci. 116, 12091217 (2003). Pessin, J.E. & Saltiel, A.R. Signaling pathways in insulin action: molecular targets of insulin resistance. J.Clin. Invest. 106, 165169 (2000). Saltiel, A.R. & Kahn, C.R. Insulin signalling and the regulation of glucose and lipid metabolism. Nature 414, 799806 (2001). Thorens, B. & Mueckler, M. Glucose transporters in the 21st century. Am. J.Physiol. Endocrinol. Metab. 298, e141e145 (2010). Slot, J.W., Geuze, H.J., Gigengack, S., Lienhard, G.E. & James, D.E. Immuno-localization of the insulin regulatable glucose transporter in brown adipose tissue of the rat. J.Cell Biol. 113, 123135 (1991).
9.
10.
11. 12.
13.
The first to examine the localization of GLUT4 in intact adipocytes by immunostaining, showing that the transporter is distributed between the TGN, endosomes and small tubulo-vesicular compartments. Martin, S. etal. Effects of insulin on intracellular GLUT4 vesicles in adipocytes: evidence for a secretory mode of regulation. J.Cell Sci. 113 , 34273438 (2000). Bogan, J.S. & Kandror, K.V. Biogenesis and regulation of insulin-responsive vesicles containing GLUT4. Curr. Opin. Cell Biol. 22, 506512 (2010). Bryant, N.J., Govers, R. & James, D.E. Regulated transport of the glucose transporter GLUT4. Nature Rev. Mol. Cell Biol. 3, 267277 (2002). Kraegen, E.W., James, D.E., Jenkins, A.B. & Chisholm, D.J. Dose-response curves for in vivo insulin sensitivity in individual tissues in rats. Am. J.Physiol. 248, e353e362 (1985). Zisman, A. etal. Targeted disruption of the glucose transporter 4 selectively in muscle causes insulin resistance and glucose intolerance. Nature Med. 6, 924928 (2000).
14. Abel, E.D. etal. Adipose-selective targeting of the GLUT4 gene impairs insulin action in muscle and liver. Nature 409, 729733 (2001). 15. Lazar, D.F. etal. Mitogen-activated protein kinase kinase inhibition does not block the stimulation of glucose utilization by insulin. J.Biol. Chem. 270, 2080120807 (1995). 16. Polak, P. etal. Adipose-specific knockout of raptor results in lean mice with enhanced mitochondrial respiration. Cell Metab. 8, 399410 (2008). 17. Cai, H., Reinisch, K. & Ferro-Novick, S. Coats, tethers, Rabs, and SNAREs work together to mediate the intracellular destination of a transport vesicle. Dev. Cell 12, 671682 (2007). 18. Yip, M.F. etal. CaMKII-mediated phosphorylation of the myosin motor Myo1c is required for insulinstimulated GLUT4 translocation in adipocytes. Cell Metab. 8, 384398 (2008). 19. Myers, M.G.Jr & White, M.F. Insulin signal transduction and the IRS proteins. Annu. Rev. Pharmacol. Toxicol. 36, 615658 (1996). 20. Lee, J. & Pilch, P.F. The insulin receptor: structure, function, and signaling. Am. J.Physiol. 266, C319C334 (1994).
VOLUME 13 | JUNE 2012 | 393
REVIEWS
21. Whiteman, E.L., Cho, H. & Birnbaum, M.J. Role of Akt/protein kinase B in metabolism. Trends Endocrinol. Metab. 13, 444451 (2002). 22. Katome, T. etal. Use of RNA interference-mediated gene silencing and adenoviral overexpression to elucidate the roles of AKT/protein kinase B isoforms in insulin actions. J.Biol. Chem. 278, 2831228323 (2003). 23. Clarke, J.F., Young, P.W., Yonezawa, K., Kasuga, M. & Holman, G.D. Inhibition of the translocation of GLUT1 and GLUT4 in 3T3L1 cells by the phosphatidylinositol 3kinase inhibitor, wortmannin. Biochem. J. 300, 631635 (1994). 24. Quon, M.J. etal. Insulin receptor substrate 1 mediates the stimulatory effect of insulin on GLUT4 translocation in transfected rat adipose cells. J.Biol. Chem. 269, 2792027924 (1994). 25. Kohn, A.D., Summers, S.A., Birnbaum, M.J. & Roth, R.A. Expression of a constitutively active Akt Ser/Thr kinase in 3T3L1 adipocytes stimulates glucose uptake and glucose transporter 4 translocation. J.Biol. Chem. 271, 3137231378 (1996). 26. Ng, Y., Ramm, G., Lopez, J.A. & James, D.E. Rapid activation of Akt2 is sufficient to stimulate GLUT4 translocation in 3T3L1 adipocytes. Cell Metab. 7, 348356 (2008). 27. Bai, L. etal. Dissecting multiple steps of GLUT4 trafficking and identifying the sites of insulin action. Cell Metab. 5, 4757 (2007). One of the earliest studies to visualize GLUT4 trafficking in real-time by TIRFM, showing that insulin modulates the rates of GLUT4 vesicle docking and fusion. 28. Gonzalez, E. & McGraw, T.E. Insulin signaling diverges into Akt-dependent and -independent signals to regulate the recruitment/docking and the fusion of GLUT4 vesicles to the plasma membrane. Mol. Biol. Cell 17, 44844493 (2006). 29. Fujita, H. etal. Identification of three distinct functional sites of insulin-mediated GLUT4 trafficking in adipocytes using quantitative single molecule imaging. Mol. Biol. Cell 21, 27212731 (2010). 30. Chen, X.W. etal. A Ral GAP complex links PI 3 kinase/Akt signaling to RalA activation in insulin action. Mol. Biol. Cell 22, 141152 (2010). 31. Zeigerer, A., McBrayer, M.K. & McGraw, T.E. Insulin stimulation of GLUT4 exocytosis, but not its inhibition of endocytosis, is dependent on RabGAP AS160. Mol. Biol. Cell 15, 44064415 (2004). 32. Min, J. etal. Synip: a novel insulin-regulated syntaxin 4binding protein mediating GLUT4 translocation in adipocytes. Mol. Cell 3, 751760 (1999). 33. Xie, X. etal. C2 domain-containing phosphoprotein CDP138 regulates GLUT4 insertion into the plasma membrane. Cell Metab. 14, 378389 (2011). 34. Yamada, E. etal. Akt2 phosphorylates Synip to regulate docking and fusion of GLUT4containing vesicles. J.Cell Biol. 168, 921928 (2005). 35. Hu, J., Liu, J., Ghirlando, R., Saltiel, A.R. & Hubbard, S.R. Structural basis for recruitment of the adaptor protein APS to the activated insulin receptor. Mol. Cell 12, 13791389 (2003). 36. Ribon, V., Printen, J.A., Hoffman, N.G., Kay, B.K. & Saltiel, A.R. A novel, multifuntional cCbl binding protein in insulin receptor signaling in 3T3L1 adipocytes. Mol. Cell. Biol. 18, 872879 (1998). 37. Liu, J., Kimura, A., Baumann, C.A. & Saltiel, A.R. APS facilitates cCbl tyrosine phosphorylation and GLUT4 translocation in response to insulin in 3T3L1 adipocytes. Mol. Cell. Biol. 22, 35993609 (2002). 38. Ribon, V., Hubbell, S., Herrera, R. & Saltiel, A.R. The product of the cbl oncogene forms stable complexes in vivo with endogenous Crk in a tyrosine phosphorylation-dependent manner. Mol. Cell. Biol. 16, 4552 (1996). 39. Knudsen, B.S., Feller, S.M. & Hanafusa, H. Four proline-rich sequences of the guanine-nucleotide exchange factor C3G bind with unique specificity to the first Src homology 3 domain of Crk. J.Biol. Chem. 269, 3278132787 (1994). 40. Chiang, S.H. etal. Insulin-stimulated GLUT4 translocation requires the CAP-dependent activation of TC10. Nature 410, 944948 (2001). 41. Okada, T., Kawano, Y., Sakakibara, T., Hazeki, O. & Ui, M. Essential role of phosphatidylinositol 3kinase in insulin-induced glucose transport and antilipolysis in rat adipocytes. Studies with a selective inhibitor wortmannin. J.Biol. Chem. 269, 35683573 (1994). 42. Ahn, M.Y., Katsanakis, K.D., Bheda, F. & Pillay, T.S. Primary and essential role of the adaptor protein APS for recruitment of both cCbl and its associated protein CAP in insulin signaling. J.Biol. Chem. 279, 2152621532 (2004). 43. Chang, L., Chiang, S.H. & Saltiel, A.R. TC10 is required for insulin-stimulated glucose uptake in adipocytes. Endocrinology 148, 2733 (2007). 44. Sharma, P.M. etal. Inhibition of phosphatidylinositol 3kinase activity by adenovirus-mediated gene transfer and its effect on insulin action. J.Biol. Chem. 273, 1852818537 (1998). 45. Schultze, S.M., Jensen, J., Hemmings, B.A., Tschopp, O. & Niessen, M. Promiscuous affairs of PKB/AKT isoforms in metabolism. Arch. Physiol. Biochem. 117, 7077 (2011). 46. Lesniewski, L.A. etal. Bone marrow-specific Cap gene deletion protects against high-fat diet-induced insulin resistance. Nature Med. 13, 455462 (2007). 47. Zhang, M., Kimura, A. & Saltiel, A.R. Cloning and characterization of Cbl-associated protein splicing isoforms. Mol. Med. 9, 1825 (2003). 48. Satoh, S. etal. Use of bismannose photolabel to elucidate insulin-regulated GLUT4 subcellular trafficking kinetics in rat adipose cells. Evidence that exocytosis is a critical site of hormone action. J.Biol. Chem. 268, 1782017829 (1993). 49. Muretta, J.M., Romenskaia, I. & Mastick, C.C. Insulin releases Glut4 from static storage compartments into cycling endosomes and increases the rate constant for Glut4 exocytosis. J.Biol. Chem. 283, 311323 (2008). 50. Ros-Baro, A. etal. Lipid rafts are required for GLUT4 internalization in adipose cells. Proc. Natl Acad. Sci. USA 98, 1205012055 (2001). 51. Blot, V. & McGraw, T.E. GLUT4 is internalized by a cholesterol-dependent nystatin-sensitive mechanism inhibited by insulin. EMBO J. 25, 56485658 (2006). 52. Shigematsu, S., Watson, R.T., Khan, A.H. & Pessin, J.E. The adipocyte plasma membrane caveolin functional/structural organization is necessary for the efficient endocytosis of GLUT4. J.Biol. Chem. 278, 1068310690 (2003). 53. Lajoie, P. & Nabi, I.R. Lipid rafts, caveolae, and their endocytosis. Int. Rev. Cell. Mol. Biol. 282, 135163 (2010). 54. AlHasani, H. etal. Roles of the N- and Ctermini of GLUT4 in endocytosis. J.Cell Sci. 115, 131140 (2002). 55. Owen, D.J., Collins, B.M. & Evans, P.R. Adaptors for clathrin coats: structure and function. Annu. Rev. Cell Dev. Biol. 20, 153191 (2004). 56. Kandror, K.V., Stephens, J.M. & Pilch, P.F. Expression and compartmentalization of caveolin in adipose cells: coordinate regulation with and structural segregation from GLUT4. J.Cell Biol. 129, 9991006 (1995). 57. Malide, D., Ramm, G., Cushman, S.W. & Slot, J.W. Immunoelectron microscopic evidence that GLUT4 translocation explains the stimulation of glucose transport in isolated rat white adipose cells. J.Cell Sci. 113(Pt. 23), 42034210 (2000). 58. Mettlen, M., Pucadyil, T., Ramachandran, R. & Schmid, S.L. Dissecting dynamins role in clathrinmediated endocytosis. Biochem. Soc. Trans. 37, 10221026 (2009). 59. Kao, A.W., Ceresa, B.P., Santeler, S.R. & Pessin, J.E. Expression of a dominant interfering dynamin mutant in 3T3L1 adipocytes inhibits GLUT4 endocytosis without affecting insulin signaling. J.Biol. Chem. 273, 2545025457 (1998). 60. Guilherme, A. etal. Perinuclear localization and insulin responsiveness of GLUT4 requires cytoskeletal integrity in 3T3L1 adipocytes. J.Biol. Chem. 275, 3815138159 (2000). 61. Huang, J., Imamura, T. & Olefsky, J.M. Insulin can regulate GLUT4 internalization by signaling to Rab5 and the motor protein dynein. Proc. Natl Acad. Sci. USA 98, 1308413089 (2001). 62. Foster, L.J., Li, D., Randhawa, V.K. & Klip, A. Insulin accelerates inter-endosomal GLUT4 traffic via phosphatidylinositol 3kinase and protein kinase B. J.Biol. Chem. 276, 4421244221 (2001). 63. Jhun, B.H., Rampal, A.L., Liu, H., Lachaal, M. & Jung, C.Y. Effects of insulin on steady state kinetics of GLUT4 subcellular distribution in rat adipocytes. Evidence of constitutive GLUT4 recycling. J.Biol. Chem. 267, 1771017715 (1992). 64. Czech, M.P. & Buxton, J.M. Insulin action on the internalization of the GLUT4 glucose transporter in isolated rat adipocytes. J.Biol. Chem. 268, 91879190 (1993). 65. Habtemichael, E.N., Brewer, P.D., Romenskaia, I. & Mastick, C.C. Kinetic evidence that Glut4 follows different endocytic pathways than the receptors for transferrin and 2macroglobulin. J.Biol. Chem. 286, 1011510125 (2011). 66. Govers, R., Coster, A.C. & James, D.E. Insulin increases cell surface GLUT4 levels by dose dependently discharging GLUT4 into a cell surface recycling pathway. Mol. Cell. Biol. 24, 64566466 (2004). The authors show that adipocytes contain a recycling pool of GLUT4 and a pool of GLUT4 that is sequestered in GSVs, which are mobilized to the plasma membrane only after stimulation with insulin. 67. Maxfield, F.R. & McGraw, T.E. Endocytic recycling. Nature Rev. Mol. Cell Biol. 5, 121132 (2004). 68. Livingstone, C., James, D.E., Rice, J.E., Hanpeter, D. & Gould, G.W. Compartment ablation analysis of the insulin-responsive glucose transporter (GLUT4) in 3T3L1 adipocytes. Biochem. J. 315, 487495 (1996). By chemically ablating compartments that contain TfR, this study demonstrates that a large portion of cellular GLUT4 is found in a vesicle population that is distinct from endosomes. 69. Martin, S. etal. The glucose transporter (GLUT4) and vesicle-associated membrane protein2 (VAMP2) are segregated from recycling endosomes in insulinsensitive cells. J.Cell Biol. 134, 625635 (1996). 70. Tanner, L.I. & Lienhard, G.E. Insulin elicits a redistribution of transferrin receptors in 3T3L1 adipocytes through an increase in the rate constant for receptor externalization. J.Biol. Chem. 262, 89758980 (1987). 71. Kupriyanova, T.A., Kandror, V. & Kandror, K.V. Isolation and characterization of the two major intracellular Glut4 storage compartments. J.Biol. Chem. 277, 91339138 (2002). 72. Slot, J.W., Geuze, H.J., Gigengack, S., James, D.E. & Lienhard, G.E. Translocation of the glucose transporter GLUT4 in cardiac myocytes of the rat. Proc. Natl Acad. Sci. USA 88, 78157819 (1991). 73. Martin, S. etal. The glucose transporter GLUT4 and the aminopeptidase vp165 colocalise in tubulovesicular elements in adipocytes and cardiomyocytes. J.Cell Sci. 110, 22812291 (1997). 74. Lin, B.Z., Pilch, P.F. & Kandror, K.V. Sortilin is a major protein component of Glut4containing vesicles. J.Biol. Chem. 272, 2414524147 (1997). 75. Jedrychowski, M.P. etal. Proteomic analysis of GLUT4 storage vesicles reveals LRP1 to be an important vesicle component and target of insulin signaling. J.Biol. Chem. 285, 104114 (2010). Uses proteomics to characterize the protein components of immuno-isolated GSVs and provides evidence supporting a mass action model of GSVformation. 76. Shi, J., Huang, G. & Kandror, K.V. Self-assembly of Glut4 storage vesicles during differentiation of 3T3L1 adipocytes. J.Biol. Chem. 283, 3031130321 (2008). 77. Martin, L.B., Shewan, A., Millar, C.A., Gould, G.W. & James, D.E. Vesicle-associated membrane protein 2 plays a specific role in the insulin-dependent trafficking of the facilitative glucose transporter GLUT4 in 3T3L1 adipocytes. J.Biol. Chem. 273, 14441452 (1998). 78. Rea, S. etal. Syndet, an adipocyte target SNARE involved in the insulin-induced translocation of GLUT4to the cell surface. J.Biol. Chem. 273, 1878418792 (1998). 79. Zeigerer, A. etal. GLUT4 retention in adipocytes requires two intracellular insulin-regulated transport steps. Mol. Biol. Cell 13, 24212435 (2002). 80. Xu, Y. etal. Dual-mode of insulin action controls GLUT4 vesicle exocytosis. J.Cell Biol. 193, 643653 (2011). The authors use a novel TIRFM approach to distinguish vesicles by size and show that immediately after insulin stimulation GLUT4 is mainly exocytosed from GSVs, but after prolonged insulin treatment GLUT4 arrives at the plasma membrane in endosomes. 81. Shewan, A.M. etal. GLUT4 recycles via a trans-Golgi network (TGN) subdomain enriched in Syntaxins 6 and 16 but not TGN38: involvement of an acidic targeting motif. Mol. Biol. Cell 14, 973986 (2003). 82. Karylowski, O., Zeigerer, A., Cohen, A. & McGraw, T.E. GLUT4 is retained by an intracellular cycle of vesicle formation and fusion with endosomes. Mol. Biol. Cell 15, 870882 (2004).
REVIEWS
Demonstrates that futile cycling between GLUT4 vesicles and endosomes is part of a mechanism that retains GLUT4 within non-stimulated cells. 83. Miinea, C.P. etal. AS160, the Akt substrate regulating GLUT4 translocation, has a functional Rab GTPase-activating protein domain. Biochem. J. 391, 8793 (2005). Biochemical characterization of the AS160 RAB GAP activity identifies RAB8, RAB10 and RAB14 as targets of AS160. 84. Lodhi, I.J. etal. Insulin stimulates phosphatidylinositol 3phosphate production via the activation of Rab5. Mol. Biol. Cell 19, 27182728 (2008). 85. Zerial, M. & McBride, H. Rab proteins as membrane organizers. Nature Rev. Mol. Cell Biol. 2, 107117 (2001). 86. Zoncu, R. etal. A phosphoinositide switch controls the maturation and signaling properties of APPL endosomes. Cell 136, 11101121 (2009). 87. Cormont, M. etal. Potential role of Rab4 in the regulation of subcellular localization of Glut4 in adipocytes. Mol. Cell. Biol. 16, 68796886 (1996). 88. Mari, M. etal. The Rab4 effector Rabip4 plays a role in the endocytotic trafficking of Glut 4 in 3T3L1 adipocytes. J.Cell Sci. 119, 12971306 (2006). 89. Sun, Y., Bilan, P.J., Liu, Z. & Klip, A. Rab8A and Rab13 are activated by insulin and regulate GLUT4 translocation in muscle cells. Proc. Natl Acad. Sci. USA 107, 1990919914 (2010). 90. Sano, H. etal. Rab10, a target of the AS160 Rab GAP, is required for insulin-stimulated translocation of GLUT4 to the adipocyte plasma membrane. Cell Metab. 5, 293303 (2007). 91. Ishikura, S. & Klip, A. Muscle cells engage Rab8A and myosin Vb in insulin-dependent GLUT4 translocation. Am. J.Physiol. Cell Physiol. 295, C1016C1025 (2008). 92. Sano, H., Roach, W.G., Peck, G.R., Fukuda, M. & Lienhard, G.E. Rab10 in insulin-stimulated GLUT4 translocation. Biochem. J. 411, 8995 (2008). 93. Kane, S. etal. A method to identify serine kinase substrates. Akt phosphorylates a novel adipocyte protein with a Rab GTPase-activating protein (GAP) domain. J.Biol. Chem. 277, 2211522118 (2002). 94. Sano, H. etal. Insulin-stimulated phosphorylation of a Rab GTPase-activating protein regulates GLUT4 translocation. J.Biol. Chem. 278, 1459914602 (2003). 95. Sano, H., Peck, G.R., Kettenbach, A.N., Gerber, S.A. & Lienhard, G.E. Insulin-stimulated GLUT4 protein translocation in adipocytes requires the Rab10 guanine nucleotide exchange factor Dennd4C. J.Biol. Chem. 286, 1654116545 (2011). 96. Imamura, T. etal. Insulin-induced GLUT4 translocation involves protein kinase C-mediated functional coupling between Rab4 and the motor mrotein kinesin. Mol. Cell. Biol. 23, 48924900 (2003). 97. Schwenk, R.W. & Eckel, J. A novel method to monitor insulin-stimulated GTP-loading of Rab11a in cardiomyocytes. Cell. Signal. 19, 825830 (2007). 98. Shi, J. & Kandror, K.V. The luminal Vps10p domain of sortilin plays the predominant role in targeting to insulin-responsive Glut4containing vesicles. J.Biol. Chem. 282, 90089016 (2007). 99. Li, J. etal. An ACAP1containing clathrin coat complex for endocytic recycling. J.Cell Biol. 178, 453464 (2007). 100. Gillingham, A.K., Koumanov, F., Pryor, P.R., Reaves, B.J. & Holman, G.D. Association of AP1 adaptor complexes with GLUT4 vesicles. J.Cell Sci. 112, 47934800 (1999). 101. Li, L.V. & Kandror, K.V. Golgi-localized, -earcontaining, Arf-binding protein adaptors mediate insulin-responsive trafficking of glucose transporter 4 in 3T3L1 adipocytes. Mol. Endocrinol. 19, 21452153 (2005). 102. Blot, V. & McGraw, T.E. Molecular mechanisms controlling GLUT4 intracellular retention. Mol. Biol. Cell 19, 34773487 (2008). 103. Lodhi, I.J. etal. Gapex5, a Rab31 guanine nucleotide exchange factor that regulates Glut4 trafficking in adipocytes. Cell Metab. 5, 5972 (2007). 104. Rodriguez-Gabin, A.G., Cammer, M., Almazan, G., Charron, M. & Larocca, J.N. Role of rRAB22b, an oligodendrocyte protein, in regulation of transport of vesicles from trans Golgi to endocytic compartments. J.Neurosci. Res. 66, 11491160 (2001). 105. Yu, C., Cresswell, J., Loffler, M.G. & Bogan, J.S. The glucose transporter 4regulating protein TUG is essential for highly insulin-responsive glucose uptake in 3T3L1 adipocytes. J.Biol. Chem. 282, 77107722 (2007). 106. Liu, L.B., Omata, W., Kojima, I. & Shibata, H. The SUMO conjugating enzyme Ubc9 is a regulator of GLUT4 turnover and targeting to the insulinresponsive storage compartment in 3T3L1 adipocytes. Diabetes 56, 19771985 (2007). 107. Bogan, J.S., Hendon, N., McKee, A.E., Tsao, T.S. & Lodish, H.F. Functional cloning of TUG as a regulator of GLUT4 glucose transporter trafficking. Nature 425, 727733 (2003). 108. Yang, J. & Holman, G.D. Comparison of GLUT4 and GLUT1 subcellular trafficking in basal and insulinstimulated 3T3L1 cells. J.Biol. Chem. 268, 46004603 (1993). 109. Huang, S. etal. Insulin stimulates membrane fusion and GLUT4 accumulation in clathrin coats on adipocyte plasma membranes. Mol. Cell. Biol. 27, 34563469 (2007). 110. Lizunov, V.A., Matsumoto, H., Zimmerberg, J., Cushman, S.W. & Frolov, V.A. Insulin stimulates the halting, tethering, and fusion of mobile GLUT4 vesicles in rat adipose cells. J.Cell Biol. 169, 481489 (2005). 111. Bose, A. etal. Glucose transporter recycling in response to insulin is facilitated by myosin Myo1c. Nature 420, 821824 (2002). Reports a role for the actin motor MYO1C in GLUT4 vesicle exocytosis. 112. Inoue, M., Chang, L., Hwang, J., Chiang, S.H. & Saltiel, A.R. The exocyst complex is required for targeting of Glut4 to the plasma membrane by insulin. Nature 422, 629633 (2003). Identifies the exocyst complex as a crucial component of the cellular trafficking machinery that is mobilized by insulin and facilitates GLUT4 targeting. 113. DAndrea-Merrins, M., Chang, L., Lam, A.D., Ernst, S.A. & Stuenkel, E.L. Munc18c interaction with syntaxin 4 monomers and SNARE complex intermediates in GLUT4 vesicle trafficking. J.Biol. Chem. 282, 1655316566 (2007). 114. Fukuda, N. etal. DOC2B: a novel syntaxin4 binding protein mediating insulin-regulated GLUT4 vesicle fusion in adipocytes. Diabetes 58, 377384 (2009). 115. Jewell, J.L. etal. Munc18c phosphorylation by the insulin receptor links cell signaling directly to SNARE exocytosis. J.Cell Biol. 193, 185199 (2011). 116. Patki, V. etal. Insulin action on GLUT4 traffic visualized in single 3T3l1 adipocytes by using ultra-fast microscopy. Mol. Biol. Cell 12, 129141 (2001). 117. Fletcher, L.M., Welsh, G.I., Oatey, P.B. & Tavare, J.M. Role for the microtubule cytoskeleton in GLUT4 vesicle trafficking and in the regulation of insulinstimulated glucose uptake. Biochem. J. 352(Pt. 2), 267276 (2000). 118. Oatey, P.B., Van Weering, D.H., Dobson, S.P., Gould, G.W. & Tavare, J.M. GLUT4 vesicle dynamics in living 3T3 L1 adipocytes visualized with greenfluorescent protein. Biochem. J. 327, 637642 (1997). 119. Semiz, S. etal. Conventional kinesin KIF5B mediates insulin-stimulated GLUT4 movements on microtubules. EMBO J. 22, 23872399 (2003). 120. Wang, Q., Bilan, P.J., Tsakiridis, T., Hinek, A. & Klip, A. Actin filaments participate in the relocalization of phosphatidylinositol3kinase to glucose transportercontaining compartments and in the stimulation of glucose uptake in 3T3L1 adipocytes. Biochem. J. 331, 917928 (1998). 121. Wang, Q., Khayat, Z., Kishi, K., Ebina, Y. & Klip, A. GLUT4 translocation by insulin in intact muscle cells: detection by a fast and quantitative assay. FEBS Lett. 427, 193197 (1998). 122. Omata, W., Shibata, H., Li, L., Takata, K. & Kojima, I. Actin filaments play a critical role in insulin-induced exocytotic recruitment but not in endocytosis of GLUT4 in isolated rat adipocytes. Biochem. J. 346 (Pt. 2), 321328 (2000). 123. Huang, J., Imamura, T., Babendure, J.L., Lu, J.C. & Olefsky, J.M. Disruption of microtubules ablates the specificity of insulin signaling to GLUT4 translocation in 3T3L1 adipocytes. J.Biol. Chem. 280, 4230042306 (2005). 124. Bose, A. etal. Unconventional myosin Myo1c promotes membrane fusion in a regulated exocytic pathway. Mol. Cell. Biol. 24, 54475458 (2004). 125. Lopez, J.A. etal. Identification of a distal GLUT4 trafficking event controlled by actin polymerization. Mol. Biol. Cell 20, 39183929 (2009). 126. Chen, Y., Wang, Y., Ji, W., Xu, P. & Xu, T. A pre-docking role for microtubules in insulin-stimulated glucose transporter 4 translocation. FEBS J. 275, 705712 (2008). 127. Yoshizaki, T. etal. Myosin 5a is an insulin-stimulated Akt2 (protein kinase B) substrate modulating GLUT4 vesicle translocation. Mol. Cell. Biol. 27, 51725183 (2007). 128. Chen, X.W., Leto, D., Chiang, S.H., Wang, Q. & Saltiel, A.R. Activation of RalA is required for insulinstimulated Glut4 trafficking to the plasma membrane via the exocyst and the motor protein Myo1c. Dev. Cell 13, 391404 (2007). 129. Lipatova, Z. etal. Direct interaction between a myosin V motor and the Rab GTPases Ypt31/32 is required for polarized secretion. Mol. Biol. Cell 19, 41774187 (2008). 130. Boldogh, I.R., Ramcharan, S.L., Yang, H.C. & Pon, L.A. A type V myosin (Myo2p) and a Rab-like Gprotein (Ypt11p) are required for retention of newly inherited mitochondria in yeast cells during cell division. Mol. Biol. Cell 15, 39944002 (2004). 131. Jordens, I. etal. The Rab7 effector protein RILP controls lysosomal transport by inducing the recruitment of dynein-dynactin motors. Curr. Biol. 11, 16801685 (2001). 132. Hales, C.M., Vaerman, J.P. & Goldenring, J.R. Rab11 family interacting protein 2 associates with Myosin Vb and regulates plasma membrane recycling. J.Biol. Chem. 277, 5041550421 (2002). 133. Inoue, M., Chiang, S.H., Chang, L., Chen, X.W. & Saltiel, A.R. Compartmentalization of the exocyst complex in lipid rafts controls Glut4 vesicle tethering. Mol. Biol. Cell 17, 23032311 (2006). 134. He, B. & Guo, W. The exocyst complex in polarized exocytosis. Curr. Opin. Cell Biol. 21, 537542 (2009). 135. Chen, X.W. etal. Exocyst function is regulated by effector phosphorylation. Nature Cell Biol. 13, 580588 (2011). 136. He, B., Xi, F., Zhang, X., Zhang, J. & Guo, W. Exo70 interacts with phospholipids and mediates the targeting of the exocyst to the plasma membrane. EMBO J. 26, 40534065 (2007). 137. Liu, J., Zuo, X., Yue, P. & Guo, W. Phosphatidylinositol 4,5bisphosphate mediates the targeting of the exocyst to the plasma membrane for exocytosis in mammalian cells. Mol. Biol. Cell 18, 44834492 (2007). 138. Zhang, X. etal. Membrane association and functional regulation of Sec3 by phospholipids and Cdc42. J.Cell Biol. 180, 145158 (2008). 139. Martin, T.F. PI(4,5)P2 regulation of surface membrane traffic. Curr. Opin. Cell Biol. 13, 493499 (2001). 140. Moskalenko, S. etal. The exocyst is a Ral effector complex. Nature Cell Biol. 4, 6672 (2002). 141. Moskalenko, S. etal. Ral GTPases regulate exocyst assembly through dual subunit interactions. J.Biol. Chem. 278, 5174351748 (2003). 142. Munson, M. & Novick, P. The exocyst defrocked, a framework of rods revealed. Nature Struct. Mol. Biol. 13, 577581 (2006). 143. Jiang, L. etal. Direct quantification of fusion rate reveals a distal role for AS160 in insulin-stimulated fusion of GLUT4 storage vesicles. J.Biol. Chem. 283, 85088516 (2008). 144. Babbey, C.M., Bacallao, R.L. & Dunn, K.W. Rab10 associates with primary cilia and the exocyst complex in renal epithelial cells. Am. J.Physiol. Renal Physiol. 299, F495F506 (2010). 145. Jahn, R. & Scheller, R.H. SNAREs engines for membrane fusion. Nature Rev. Mol. Cell Biol. 7, 631643 (2006). 146. Macaulay, S.L. etal. Functional studies in 3T3L1 cells support a role for SNARE proteins in insulin stimulation of GLUT4 translocation. Biochem. J. 324, 217224 (1997). 147. Cheatham, B. etal. Insulin-stimulated translocation of GLUT4 glucose transporters requires SNARE-complex proteins. Proc. Natl Acad. Sci. USA 93, 1516915173 (1996). 148. Volchuk, A. etal. Syntaxin 4 in 3T3L1 adipocytes: regulation by insulin and participation in insulindependent glucose transport. Mol. Biol. Cell 7, 10751082 (1996). 149. Bryant, N.J. & Gould, G.W. SNARE proteins underpin insulin-regulated GLUT4 traffic. Traffic 12, 657664 (2011). 150. Carr, C.M. & Rizo, J. At the junction of SNARE and SM protein function. Curr. Opin. Cell Biol. 22, 488495 (2010).
VOLUME 13 | JUNE 2012 | 395
REVIEWS
151. Thurmond, D.C. etal. Regulation of insulin-stimulated GLUT4 translocation by Munc18c in 3T3L1 adipocytes. J.Biol. Chem. 273, 3387633883 (1998). 152. Macaulay, S.L. etal. Cellular munc18c levels can modulate glucose transport rate and GLUT4 translocation in 3T3L1 cells. FEBS Lett. 528, 154160 (2002). 153. Tamori, Y. etal. Inhibition of insulin-induced GLUT4 translocation by Munc18c through interaction with syntaxin4 in 3T3L1 adipocytes. J.Biol. Chem. 273, 1974019746 (1998). Reports a regulatory role for MUNC18C in SNARE-mediated GLUT4 vesicle fusion. 154. Khan, A.H. etal. Munc18c regulates insulinstimulated glut4 translocation to the transverse tubules in skeletal muscle. J.Biol. Chem. 276, 40634069 (2001). 155. Kanda, H. etal. Adipocytes from Munc18cnull mice show increased sensitivity to insulin-stimulated GLUT4 externalization. J.Clin. Invest. 115, 291301 (2005). 156. Thurmond, D.C., Kanzaki, M., Khan, A.H. & Pessin, J.E. Munc18c function is required for insulinstimulated plasma membrane fusion of GLUT4 and insulin-responsive amino peptidase storage vesicles. Mol. Cell. Biol. 20, 379388 (2000). 157. Oh, E., Spurlin, B.A., Pessin, J.E. & Thurmond, D.C. Munc18c heterozygous knockout mice display increased susceptibility for severe glucose intolerance. Diabetes 54, 638647 (2005). 158. Latham, C.F. etal. Molecular dissection of the Munc18c/syntaxin4 interaction: implications for regulation of membrane trafficking. Traffic 7, 14081419 (2006). 159. Hu, S.H., Latham, C.F., Gee, C.L., James, D.E. & Martin, J.L. Structure of the Munc18c/Syntaxin4 Npeptide complex defines universal features of the Npeptide binding mode of Sec1/Munc18 proteins. Proc. Natl Acad. Sci. USA 104, 87738778 (2007). 160. Jewell, J.L., Oh, E., Bennett, S.M., Meroueh, S.O. & Thurmond, D.C. The tyrosine phosphorylation of Munc18c induces a switch in binding specificity from syntaxin 4 to Doc2. J.Biol. Chem. 283, 2173421746 (2008). 161. McMahon, H.T., Kozlov, M.M. & Martens, S. Membrane curvature in synaptic vesicle fusion and beyond. Cell 140, 601605 (2010). 162. Okada, S. etal. Synip phosphorylation is required for insulin-stimulated Glut4 translocation. Biochem. Biophys. Res. Commun. 356, 102106 (2007). 163. Sano, H., Kane, S., Sano, E. & Lienhard, G.E. Synip phosphorylation does not regulate insulinstimulated GLUT4 translocation. Biochem. Biophys. Res. Commun. 332, 880884 (2005). 164. Fujita, Y. etal. Tomosyn: a syntaxin-1binding protein that forms a novel complex in the neurotransmitter release process. Neuron 20, 905915 (1998). 165. Martens, S., Kozlov, M.M. & McMahon, H.T. How synaptotagmin promotes membrane fusion. Science 316, 12051208 (2007). 166. Lipschutz, J.H. & Mostov, K.E. Exocytosis: the many masters of the exocyst. Curr. Biol. 12, R212R214 (2002). 167. Rosen, E.D. & Spiegelman, B.M. Adipocytes as regulators of energy balance and glucose homeostasis. Nature 444, 847853 (2006). 168. Herman, M.A. & Kahn, B.B. Glucose transport and sensing in the maintenance of glucose homeostasis and metabolic harmony. J.Clin. Invest. 116, 17671775 (2006).
Acknowledgements
This work was supported by a US National Institutes of Health (NIH) grant R01DK076906. The authors thank M. Uhm and D. Bridges for their critical reading and discussions of the manuscript.
Competing statement FURTHER interests INFORMATION
The authors declare no competing financial interests. Alan R. Saltiels homepage : http://www.lsi.umich.edu/ facultyresearch/labs/saltiel
REVIEWS
Geometry and force behind kinetochore orientation: lessons from meiosis

Yoshinori Watanabe
Abstract | During mitosis, replicated chromosomes (sister chromatids) become attached at the kinetochore by spindle microtubules emanating from opposite poles and segregate equationally. In the first division of meiosis, however, sister chromatids become attached from the same pole and co-segregate, whereas homologous chromosomes connected by chiasmata segregate to opposite poles. Disorder in this specialized chromosome attachment in meiosis is the leading cause of miscarriage in humans. Recent studies have elucidated the molecular mechanisms determining chromosome orientation, and consequently segregation, in meiosis. Comparative studies of meiosis and mitosis have led to the general principle that kinetochore geometry and tension exerted by microtubules synergistically generate chromosome orientation.
Genetic information is packed into chromosomes, which have to be properly transmitted to replicated cells or offspring. In mitosis the cell division observed in somatic cells the replicated chromosomes (which consist of two sister chromatids) are split and equally divided between daughter cells through a process called equational division. For this process to be successful, the association of sister chromatids, which is known as cohesion and is established during DNA replication, must be preserved until chromosome alignment, the process by which kinetochores of sister chromatids are each captured by microtubules emanating from opposite spindle poles (BOX1). Any mistake in chromosome segregation may increase the chance of tumorigenesis in humans1,2. Eukaryotic cells have another cell division process called meiosis, which is required for sexual reproduction and reduces the chromosome number by half. Meiosis consists of two rounds of chromosome segregation following a single round of DNA replication. During the first division of meiosis (meiosisI), homologous chromosomes (homologues) rather than sister chromatids are pulled to opposite poles and segregated to the daughter cells (known as reductional division). Only at the second meiotic division (meiosisII) do sister chromatids segregate as they do in mitosis (BOX1). Thus, meiosis I has a unique form of chromosome segregation3,4. Much has been learnt about how sister chromatids bi-orient (that is, how they are pulled towards opposite poles) in mitosis. In particular, tension across the paired kinetochores that are attached by microtubules has an essential role in stabilizing the attachment; an absence of tension leads to the release and recapture of microtubules (reorientation)5. Although meiosis must have evolved from mitosis, it is elusive how the principle that determines chromosome orientation in mitosis was modified to generate the new principle observed in meiosis, in which sister chromatids mono-orient (that is, when they move towards the same pole) and paired homologues (known as bivalents) bi-orient 3,6. Recent studies of meiosis have shed new light on the importance of kinetochore geometry in addition to the tension-dependent reorientation machinery. In this Review, I discuss the molecular mechanisms governing chromosome behaviour during meiosis, including the establishment of meiosis-specific chromosome orientation. I further discuss parallels between meiosi s and mitosis and propose a common principle for determinin g chromosome orientation.
Kinetochores
Large protein complexes that assemble on chromosomes and mediate the attachment to spindle microtubules.
Laboratory of Chromosome Dynamics, Institute of Molecular and Cellular Biosciences, University of Tokyo, Yayoi, Tokyo 113-0032, Japan. e-mail: ywatanab@iam.utokyo.ac.jp doi:10.1038/nrm3349 Published online 16 May 2012
Kinetochore attachment to microtubules Kinetochores are proteinaceous complexes that assemble on each chromosome and mediate their interaction with microtubules7,8. During chromosome alignment and segregation in mitosis, microtubules capture the resolved
REVIEWS
Box 1 | Chromosome segregation in mitosis and meiosis
In mitosis (see the figure, part a), sister kinetochores are bi-oriented (that is, they move towards opposite poles) and all cohesin associated with the chromatin is cleaved by separase at the onset of anaphase, thereby causing sister chromatids to separate to opposite sides (equational segregation). In meiosis I (see the figure, part b), sister kinetochores are captured by microtubules that emanate from the same pole, whereas paired homologues are captured by opposite poles. Atthe onset of anaphase I, cohesin along the chromosome arms, but not at centromeres, is cleaved by separase, so that homologues separate to opposite sides (reductional segregation). Persistent centromeric cohesin is essential for the bi-orientation of sister chromatids in the following meiosis II, when sister chromatids separate as they do in mitosis.
a Mitosis
Rad21cohesin Kinetochore
b Meiosis
Rec8cohesin
Spindle pole Separase Microtubule
Meiosis I
Cleaved cohesin
Meiosis II
Centromeres
Specialized genomic regions where the kinetochore assembles.
(paired but separated) sister kinetochores from opposite poles in a process known as amphitelic attachment (chromosomes are bi-oriented), and this generates full tension across centromeres (FIG.1a). However, during the early stages of chromosome alignment, either or both sister kinetochores are transiently bound by microtubules from one pole, which is known as monotelic or syntelic attachment, respectively (chromosomes are mono-oriente d) (FIG.1a). These erroneous attachments are usually corrected into amphitelic attachment. A further complication arises from the fact that a single kinetochore bindsto multiple microtubules (although this does not apply to budding yeast) so that one kinetochore can be bound simultaneously by microtubules emanating from both poles (known as merotelic attachment), which leads to undesirable chromosome bi-orientation (FIG.1a). Because it is unlikely that merotelic attachment can be detected
by surveillance mechanisms that monitor tension, it is thought that this form of attachment is the main source of aneu ploidy in mammaliancells9. By contrast, during meiosis I, homologues are connected by reciprocal recombination (chiasmata), which produces paired chromosomes known as bi valents (FIG.1b). Importantly, sister kinetochores are fused or juxtaposed side-by-side at this stage, so they tend to attach to microtubules in a syntelic manner. As a result, the bivalent is captured from opposite poles, thereby generating full tension. As in mitosis, merotelic attachment has to be avoided to ensure homologue disjunction in meiosisI (FIG.1b). It is notable that if recombination does not occur or if chiasmata are lost (resulting in univalents), the kinetochores of sister chromatids are still fused (FIG.1b). Consequently, and in contrast to the mitotic chromosome, univalents tend to maintain syntelic attachment, although
VOLUME 13 | JUNE 2012 | 371
REVIEWS
a Mitosis
Sister chromatids Separated sister kinetochores Microtubule
for preserving the association of paired chromosomes, accumulating evidence suggests that the cohesin complex, which constitutes the sister chromatid cohesion axis, provides a structural and functional basis for other factors that act in meiotic chromosome differentiation. Meiosis-specific cohesin. During both mitosis and meiosis, sister chromatid cohesion is established usually during DNA replication by a multisubunit complex termed cohesin. The mitotic cohesin complex comprises two SMC (structural maintenance of chromosome) family proteins, which are termed Smc1 and Smc3, as well as two accessory subunits, which are named Rad21 (also known as kleisin subunit; Scc1 in Saccharomycescerevisiae) and Scc3. Smc1Smc3 heterodimers topologically embrace DNA strands through their long coiled-coil stretches. Rad21 interacts with the two ends of the cohesin ring, thereby closing it. It has been proposed that sister chromati d cohesion is established when the DNA strand is replicated while being entrapped in the cohesin ring, which thus glues the sister chromatids together 10,11. Cohesin complexes are modified in meiosis. The most common modification is the replacement of Rad21 by its meiotic counterpart, Rec8 (REF. 12), although mammals have another meiotic kleisin subunit, termed RAD21L, and Caenorhabditiselegans has COH-3 and COH-4 (REFS1316). Meiotic counterparts of other subunits are also reported in several organisms12,17. For example, the fission yeast protein Rec11, which is a meiosis-specific Scc3 subunit, is exclusively required for chromatid arm cohesion and recombination, whereas canonical (mitotic) Scc3 is required for centromeric cohesion in meiosis but not for recombination18. Thus, a combination of canonical and meiosis-specific cohesin subunits may diversify cohesin functions along meiotic chromosomes. Functions of meiotic cohesin. In addition to promoting sister chromatid cohesion, cohesin complexes that include Rec8 are thought to mediate numerous processes during meiosis, such as the formation of double-strand breaks (DSBs) and synaptonemal complexes, mono-orientation of sister kinetochores and persistent cohesion throughout anaphaseI17,1923. For example, cohesin complexes have been proposed to mediate homologue pairing following recombination in meiotic chromosomes (the mechanism of homologue pairing and recombination is reviewed extensively elsewhere; see REFS 2426). In most organisms, loose association (pairing) and subsequent tight association (synapsis) of homologues initiate through telomere clustering at the nuclear periphery, a chromosomal configuration known as a bouquet. Accompanying nuclear movement promotes the side-by-side configuration of chromosomes, thus facilitating homologue search27,28. In C.elegans, this process is promoted by specific proteins that localize at a defined chromosomal site (called the pairing centre) rather than at the telomeres29. Although DSBs and their repair process promote pairing through the recognition of homologous DNA sequences, accumulating evidence in several organisms suggests that weak interstitial pairing occurs between homologues before or independently
Amphitelic Monotelic Syntelic Merotelic (bi-orientation) (mono(mono(bi-orientation) orientation) orientation) Homologues Kinetochore
b Meiosis I
Bivalent
Half bivalent Chiasma No exchange
Bi-orientation of bivalent Fused sister kinetochores
Mono-orientation of bivalent
Univalent Univalent
Syntelic (mono-orientation)
Merotelic (bi-orientation)
Figure 1 | Types of chromosome structure, kinetochore geometry and Nature Reviews | Molecular Cell Biology kinetochoremicrotubule attachment. Homologous chromosomes (homologues) have no interaction in mitosis, whereas in meiosis they are connected by chiasmata, forming bivalents. a | In mitosis (or meiosis II), kinetochores on sister chromatids are separated and oriented back-to-back. As a result, sister kinetochores are captured by microtubules from opposite poles, and this results in amphitelic attachment (during which chromosomes are bi-oriented). However, mono-orientation can occur if either one or both sister kinetochores are bound by microtubules from one pole, which gives rise to monotelic or syntelic attachment, respectively. If a kinetochore is bound by microtubules from both poles (merotelic attachment), chromosomes will bi-orient, but this may lead to aneuploidy. b | By contrast, in meiosis I sister kinetochores are juxtaposed or fused and orient side-by-side, thus binding to microtubules in a syntelic manner. Although bivalents are often captured from one pole (mono-orientation) during the initial stage of alignment, this tensionless attachment is so unstable that they are shifted to bi-orientation in the presence of chiasmata, which then produce tension. Theunivalent, which can be produced by the accidental loss of chiasmata or a recombination defect in meiosis I, isnot identical to mitotic sister chromatids because sister kinetochores are fused. Suchfused kinetochores often biorient by unstable merotelic attachment, otherwise it would be impossible to produce tension.
they occasionally bi-orient by merotelic attachment. The principle kinetochore attachment to microtubules in meiosisII is thought to be virtually the same as what occurs in mitosis, as kinetochore geometry and orientation are similarly regulated in both divisions (BOX1).
Synaptonemal complexes
Ribbon-like protein structures between pachytene chromosomes that mediatesynapsis.
Cohesion in meiotic chromosomes Meiotic cell differentiation occurs in G1 phase of the cell cycle. In meiotic prophase, chromosomes are modified for meiosis-specific features, such as recombination and reductional division at meiosis I. Although sister chromatid cohesion mediated by cohesin is primarily required
REVIEWS
of DSB formation25,26. How this occurs is largely unclear, but recent cytological studies in mouse germ cells suggest a model in which cohesin complexes that distribute uniquely along the chromosome axis but identically between homologues may act to recognize one another 13. The two homologues become tightly connected through sister chromatid cohesion in the distal regions (FIG.1b) through chiasmata. This occurs through the process of reciprocal recombination between non-sister strands of two homologues. Chiasmata are essential for the capture of homologues by microtubules from opposite poles and thus for their disjunction (separation) at meiosisI (see below). In the absence of chiasmata, separated homologues undergo random segregation at metaphase I. Thus, although meiotic recombination contributes to the generation of genetic diversity, its primary aim might be the production of physical linkages between homologues, which is essential for proper segregation of homologues at meiosis I3. Some insect species have different mechanisms for establishing the association of homologues even in the absence of recombination and thereby allowing correct segregatio n at meiosisI30. Persisting centromeric cohesion by shugoshin. In mitosis, sister chromatid cohesion is established in S phase and maintained until metaphase, when the sister chromatids are aligned by the balance of polewards pulling forces and sister chromatid cohesion. For the onset of anaphase, the APC/C (anaphase-promoting complex; also known as the cyclosome) triggers the degradation of securin, an inhibitory chaperone for the specific endopeptidase separase (also known as separin), which cleaves cohesin that includes Rad21 and removes cohesion along the entirechromosome10,11 (BOX1). This removal of cohesin triggers chromosome movement to oppositepoles. In meiosis, Rec8 largely replaces Rad21 and is cleaved only along the chromosome arms at anaphaseI31,32. By contrast, centromeric Rec8 must be protected from cleavage because the residual cohesion at the centro meres is used to ensure equational division at meiosis II (see below) (BOX1). This protection of Rec8 at anaphase I is mediated by the conserved centromeric protein family shugoshin (SGO; which means guardian spirit in Japanese). Shugoshin was initially identified in the Drosophila melanogaster meiotic mutant MEI-S332, which shows precocious sister centromere separation in meiosis33,34 and was later rediscovered as a protector of Rec8 through functional screening in yeast3537. Shugoshin proteins are conserve d from yeast to plants and mammals3842. In yeast Sgo1 forms a complex with protein phosphatase 2A (PP2A), and the Rec8-mediated protection by Sgo1 depends on the recruitment of PP2A to the centromeres43,44. Recent studies in budding and fission yeast show that Rec8 phosphorylation by casein kinase1 (CK1) and by cell division control protein 7 (Cdc7) is a prerequisite for the cleavage of Rec8 by separase4547. Together, these findings have suggested that Sgo1PP2A antagonizes Rec8 phosphorylation and thus its cleavage at centromeres in meiosis I (FIG.2). Consistent with this, inactivation of Sgo1 or PP2A leads to the loss of sister chromatid cohesion during anaphaseI; however, sister chromatids still co-segregate, which indicates that monopolar attachment is completed before the onset of anaphase I (BOX1). Consequently, chromosome segregation defects in Sgo1-depleted cells become visible mainly in meiosis II, as the premature separation of sister chromatids results in random segregation in the ensuing anaphase II35,36. Mammals have two SGO-like proteins, SGOL1 and SGOL2. As centromeric cohesin is protected not only during meiosis I but also during prophase of mitosis, it is thought that SGOL1PP2A may solely interact with and dephosphorylate mitotic cohesin and thereby prevent cohesin dissociation during prophase4850, whereas SGOL2PP2A protects meiotic cohesin from cleavage during anaphase I40,41. In human cells, SGOL2 has an auxiliary role in mitotic protection, presumably by supplying sufficient PP2A to centromeres43,51. Moreover, SGOL2 has a specific role in recruiting the microtubule depolymerizing mitotic centromere-associated kinesin (MCAK; also known as KIF2C) to centromeres and thereby promo ting chromosome alignment at metaphase5153. Recent studies in yeast and humans haveshown that shugoshin proteins also act as a conserved centromeric adaptor of the AuroraB complex (see below).
Kinetochore geometry in meiosis The importance of kinetochore orientation during mitosis has been long recognized and studied in several organisms54. Electron microscopy analyses of prometaphase chromosomes show that in mitosis, sister kineto chores face opposite directions before attachment by microtubules55. Ideally, this back-to-back assembly of siste rkinetochores may facilitate the attachment of the two kinetochores to opposite poles by geometric restriction. In contrast to mitosis, sister kinetochores in meiosisI are fused or juxtaposed side-by-side, which may facilitate monopolar attachment54,56,57 (FIGS1,3a). The molecular mechanism that regulates kinetochore geometry has recently been elucidated mainly inyeasts.
Mono-orientation in budding yeast. Genetic studies in budding yeast have identified a protein complex called monopolin, which is required for mono-orientation of sister kinetochores in meiosis I5860. The monopolin complex is composed of Csm1 (chromosome segregation in meiosis protein 1), Lrs4 (loss of rDNA silencing protein 4), Mam1 (monopolar microtubule attachment during meiosis I protein 1) and CK1 and localizes to centromeres specifically in meiosis I. Spo13 (sporulationspecific protein 13), another factor that is required for mono-orientation, associates with the Polo-like kinase Cdc5 and acts to recruit or stabilize the monopolin complex at centromeres6163. Cdc7 kinase, which is known to be essential for the initiation of DNA replication, also acts to localize monopolin at centromeres63. Evidence suggests that the enrichment of CK1 activity at kinetochores, which depends on the presence of Csm1Lrs4, might be an ultimate requirement for the establishment of mono-orientation in budding yeast60. However, which CK1 substrates are required for mono-orientation and the mechanism by which this occurs remain elusive.
VOLUME 13 | JUNE 2012 | 373
APC/C
(Anaphase-promoting complex; also known as the cyclosome ). A multicomponent ubiquitin ligase that targets proteins for degradation by the proteasome.
REVIEWS
CK1 P P Cohesin (Rec8) Anaphase I Sister chromatids P P
Separase Sgo1PP2A Kinetochore P P P Pericentric heterochromatin (inner centromere) P P P
Figure 2 | Sgo1PP2A protects cohesin at centromeres. In meiosis I, centromeric Nature Reviews | Molecular Cell Biology cohesin carrying Rec8 is protected from separase cleavage. For Rec8 to be cleaved by separase, itmust be phosphorylated by casein kinase 1 (CK1) in yeast. Shugoshin 1 (Sgo1) recruits protein phosphatase 2A (PP2A) to the pericentric heterochromatin regions and antagonizes this phosphorylation, thus preventing the cleavage of pericentric cohesin.
Condensin
A major protein component of mitotic chromosomes that is required for chromosome condensation. The molecular architecture of condensin is similar to that of cohesin, in which two SMC (structural maintenance of chromosome) proteins are linked to each other at one end, and the other end is closed by a kleisinsubunit.
In fission yeast, homologues of Csm1 and Lrs4 (termed Pcs1 and Mde4, respectively) are required to prevent merotelic attachment of single kinetochores in mitosis rather than to prevent bi-orientation of sister kinetochores in meiosis64. Therefore, it was hypothesized that the Csm1Lrs4 complex has a conserved role in clamping together adjacent microtubule attachment sites to mediate mono-orientation. In budding yeast (in which paired sister centromeres assemble a single kinetochore and bind only one microtubule) this would involve conjoining two microtubule attachment sites on sister kinetochores, but in fission yeast there are multiple microtubule attachment sites on each kinetochore, so these multiple attachment sites would be joined rather than paired sister kinetochores. The structural analysis of Csm1Lrs4 suggests that this complex forms a V-shape with two pairs of kinetochore-binding domains, indicating that Csm1Lrs4 may bring kinetochores together, which favours the clamp model 65. In fission yeast, however, the Csm1Lrs4complex can be functionally replaced by the condensin complex, which can compact chromatin and is enriched to kinetochores in a manner that depends on the presence of Csm1Lrs4 (REF. 66). Thus, the V-shaped clamp model is debatable. Mono-orientation in fission yeast. The mechanisms that establish monopolar attachment of kinetochores during meiosis I seem to have diverged between budding yeast and fission yeast. For example, fission yeast Pcs1Mde4 is dispensable for sister kinetochore monoorientation (see above), and the fission yeast CK1 homologue is not required for the canonical mono-orientation pathway (T.Sakuno and Y.W., unpublished observations). Moreover, although cohesin has a crucial role in
sister kinetochore mono-orientation in fission yeast (see below), this might be not the case in budding yeast 60,67. A pioneering study in this field postulated that fission yeast cohesin complexes regulate kinetochore orientation, as rec8 mutation causes equational, rather than reductional, division in meiosis I68. Like metazoan centromeres, centromeres in fission yeast comprise two domains, a kinetochore-assembling core centromere and heterochromatic pericentric regions (also known as the inner centromere in metazoans), and a single kinetochore can attach to several microtubules69. Intriguingly, in mitosis the cohesin complex that includes Rad21 accumulates preferentially at the pericentric heterochromatin regions, whereas in meiosis the Rec8-containing cohesin complex accumulates at the core centromere in addition to the pericentric regions68 (FIG.3b). Removing Rec8 and replacing it withRad21 during meiosis I results in accumulation of the Rad21-containing cohesin complex at the pericentric regions, but much less at the central core region, leading to bi-orientation of sister chromatids, as occurs during mitosis70. Importantly, inactivating Rec8 specifically at the core centromere while preserving its other functions leads to kinetochore bi-orientation at meiosisI, which confirms the crucial role of Rec8 for sister kinetochore mono-orientation71 (FIG.3b). Genetic screening to search for factors that regulate mono-orientation has further identified a meiosisspecific kinetochore protein, termed Moa1 (monopolar attachment protein 1). Moa1 interacts with Rec8 and localizes exclusively at the core centromere from prophase I to metaphase I but disappears in anaphase I71. Moa1 may assist Rec8 after pre-meiotic DNA replication to maintain the core centromere cohesion72, although its precise molecular function remains elusive. Cohesion regulates mono-orientation in higher organisms. REC8 has been shown to have an essential role in sister kinetochore mono-orientation in meiosisI in plants and nematodes, as REC8 mutants in these organisms exhibit equational chromosome segregation at meiosis I, as observed in fission yeast 16,73,74. In D.melanogaster, the orientation disruptor (ord) mutant, which has impaired sister chromatid cohesion, also exhibits a mono-orientation defect in meiosis I33. It was recently shown that inactivation of REC8 only at the kinetochore in mouse oocytes leads to bi-orientation of univalents in meiosis I (K. Tachibana-Konwalski and K. Nasmyth, personal communication). Thus, it seems that cohesion at the kinetochore mediated by the REC8-containing cohesin complex has an essential and conserved role in sister chromatid mono-orientation in meiosis I in higher organisms, including mammals. Kinetochore geometry regulates chromosome orientation. To examine the role of cohesion at the centromere, and therefore establish the role of kinetochore geometry in chromosome orientation, the core centromere was visualized in fission yeast by looping out and excising this DNA region from the neighbouring chromosomal domains during prophase I, the stage before the attachment of kinetochores to the spindle microtubules.
Pericentric heterochromatin
Heterochromatin that is assembled at the pericentric region and is composed of repeated specific DNA sequences.
REVIEWS
a
Mitosis CENPC DNA Cohesin (Rad21) Cohesin (Rec8)
Mitosis
Meiosis I
Meiosis I Moa1 Kinetochore Kinetochore Pericentric heterochromatin (inner centromere)
Bi-orientation (back-to-back)
Mono-orientation (side-by-side)
Rec8 inactivation only at the kinetochore
Figure 3 | Kinetochore geometry is defined by cohesion within the centromere. a | The distinct kinetochore geometries Nature Reviews Molecular Cell Biology of mitotic and meiotic (bivalent) chromosomes in mice. Centromere protein C (CENPC) is labelled in|red, and chromosomes are shown in green. b | Studies in fission yeast reveal that in mitosis there is no cohesion at the core centromere, and Rad21-containing cohesin accumulates in pericentric heterochromatin. By contrast, cohesion at the core centromere is established in meiosis I, and this depends on meiotic cohesin Rec8 and the meiotic kinetochore protein Moa1 (monopolar attachment protein 1). It is unknown how Moa1 regulates Rec8. During anaphase I, core centromeric Rec8 is likely to be cleaved by separase because the core centromeric cohesion is disrupted by prophase II, and only pericentric cohesin is protected by shugoshin75,134 (FIG.2). The side panels indicate the orientation of microtubule attachment to kinetochores. Image in part a is courtesy of J. Kim, University of Tokyo, Japan.
Thisanalysis revealed that cohesion at the core centro mere is indeed established and maintained, particularly during meiosis I, whereas it is lost in mutants that lack moa1 or rec8 and that are defective in sister kinetochore mono-orientation75 (FIG.3b). Importantly, although pericentric cohesion is preserved throughout meiosis I until metaphase II in an Sgo1-dependent manner, cohesion at the central core region is resolved before prophaseII, which is likely to be dependent on Rec8 cleavage by separase (note that Sgo1 localizes only at the pericentric heterochromatic regions but not at the core centromere in meiosisI (FIG.2). Remarkably, cohesin is intrinsically prevented at the core centromere in normal mitotic cells, and this is still true even if Rec8-containing cohesion is ectopically expressed in mitosis and localized to the central core region75. This finding is consistent with the notion that the establishment of cohesion at the core centro mere requires not only Rec8 but also Moa1, even in meiosisI71. Finally, when an artificial tether is introduced at the core centromere, monopolar attachment and, therefore, co-segregation of sister kinetochores are restored in moa1-mutant or rec8-mutant meiotic cells and even established in mitotic cells, albeit abortively 75. Thus, it seems that sister kinetochore mono-orientation is achieved by joining together DNA duplexes at the core centromeres and not by the action of a kinetochore proteinitself.
Aurora B stabilizes chromosome orientation Although the orientation of a kinetochoremicrotubule attachment might be influenced by the geometry or the shape of kinetochores, erroneous attachment is commonly observed in the initial stage of chromosome alignment in mitosis and is corrected before anaphase. Tension has an important role in chromosome attachment correction, as only tension-exerted attachment is stabilized during chromosome alignment. Recent finding s suggest that this may also be true in meiosis.
Lessons from the mitotic chromosome. In mitosis, centromeric cohesion produces tension by counteracting the pulling force of microtubules that capture sister kinetochores from opposite poles, and this tension has a crucial role in stabilizing attachment; indeed, mono polar attachment is highly unstable owing to the lack of tension (FIG.1a). Studies in several organisms have established that the trial-and-error process of kinetochore microtubule attachment relies largely on the mitotic kinase Aurora B (also known as Ipl1 in S.cerevisiae and Ark1 in S.pombe)5,76,77. Aurora B forms a complex with the regulatory subunits borealin, survivin and inner centromere protein (INCENP). This complex is called chromosomal passenger complex (CPC) and localizes at the centre between two sister kinetochores (known as the inner centromere) from prometaphase until meta phase78. In the absence of tension across centromeres
VOLUME 13 | JUNE 2012 | 375
REVIEWS
a
Mitosis Kinetochore
MIS12 complex NDC80 complex Microtubule
Aurora B Unstable microtubule Aurora B
KNL1
Unstable attachment
Stable microtubule
PP1, PP2A
Stable attachment
Figure 4 | Tension-dependent reorientation by Aurora B. a | Aurora B localizes at the inner centromeres in mitotic Nature Reviews | Molecular Cell Biology chromosomes and produces a phosphorylation gradient (dark and light blue areas), thus phosphorylating kinetochore proteins. During prometaphase, kinetochores are captured by microtubules, often in an erroneous manner (monotelic attachment, syntelic attachment or merotelic attachment; FIG.1a). In either case, the attachment sites are in close proximity to Aurora B activity, which peaks at the inner centromeres, so attachment becomes unstable. However, once two kinetochores are captured from opposite poles, the tension that is generated across the centromeres moves the attachment sites away from Aurora B, thus selectively stabilizing the attachment. b | The KMN (kinetochore null protein 1 (KNL1)MIS12 complexNDC80 complex) network at kinetochores is conserved in eukaryotes and binds directly to microtubules7,135. A gradient of Aurora B phosphorylation that emanates from the inner centromere phosphorylates the KMN network, thereby destabilizing the attachment. When tension is present, the kinetochore is positioned away from Aurora B, and this facilitates the association of protein phophatase 1 (PP1) with the KMN component KNL1 (REF. 82). Tension also acts to redistribute PP2A from the inner centromere to the kinetochore side84. Thus, the increased distance from Aurora B and the enrichment of phosphatase activity may act cumulatively to dephosphorylate the KMN network and thereby stabilize microtubule attachment.
KMN network
The conserved kinetochore linker complex that directly binds to the microtubule plusend and is composed of KNL1 (kinetochore null protein1), MIS12 and NDC80 subcomplexes.
(during monopolar attachment), Aurora B localizes in close proximity to the kinetochores, whereas bipolar (amphitelic) attachment produces tension that brings about the spatial separation of kinetochores from Aurora B79 (FIG.4). Increasing evidence suggests that if Aurora B is close to kinetochores, Aurora B phosphorylates the microtubule-anchoring kinetochore proteins of the KMN network (comprising KNL1 (kineto chore null protein 1; also known as Spc105 in yeast), the MIS12 (also known as Mtw1 in yeast) complex and the NDC80 complex) and destabilizes the attachment 80,81. By contrast, when kinetochores are separated from Aurora B under tension, overall kineto chore phosphorylation is suppressed, which stabilizes the attachment (FIG.4). This suppression of phosphorylation is mediated not only by the separation of the substrates from AuroraB but also by the action of PP1, which is enriched at kineto chores when tension is exerted82. Indeed, either artificially tethering the Aurora B complex to the kineto chore or depleting PP1 from kinetochores dramatically destabilizes kinetochoremicrotubule attachment in prometaphase or metaphase, and, consistently, inhibiting Aurora B or tethering PP1 at kinetochores stabilizes the attachment 79,82,83. Moreover, a recent study revealed that centromeric PP2A, which redistributes in response to tension, also acts in silencing the phosphorylation of Aurora B substrates at kinetochores84.
Owing to the correction mechanism that occurs during kinetochoremicrotubule attachment, the position of Aurora B within the centromere is crucial for establishing proper chromosome orientation. Recent studies in fission yeast and humans indicated that shugoshin proteins which associate with nucleosomes that contain histone H2A phosphorylated at Ser121 (and H2A phosphorylated at Thr120 in humans)85, a modification mediated by BUB1 act as a centromeric adaptor of the CPC8688. Because shugoshin proteins associate directly with a regulatory subunit of the CPC, both BUB1 and shugoshin proteins are required for the full localization of Aurora B at the centromeres. It is possible that the original role of shugoshin proteins was to recruit Aurora B or to establish chromosome bi-orientation, because fission and budding yeast shugoshin proteins retain only this role during mitosis86,89. The function of cohesin protection might have co-evolved for the generation of robust tension by counteracting the spindle force, thus improvin g the fidelity of chromosome bi-orientation. Studies in fission yeast, Xenopus laevis and human cells have shown that, in addition to BUB1 and shugoshin proteins, haspin kinase is required for targeting AuroraB to centromeres9092. Haspin phosphorylates histone H3 at Thr3 (H3T3), and survivin binds directly to phosphorylated H3T3 (REF93). Thus, targeting of Aurora B to centromeres depends on two histone marks phosphorylated H3T3 and H2AS121 (H2AT120 in humans) (BOX2).
REVIEWS
Box 2 | Aurora B position is defined by two histone kinases
Proper targeting of Aurora B is crucial to establish correct a PP2A Coiled-coil chromosome orientation. Two histone kinases have been P-V-I SGO motif Shug found to, either directly or indirectly, mediate this Protection oshin localization. BUB1 (REFS 35,50,130132) is a multitasking CPC protein kinase that is localized at the kinetochore, is part BUB1 P of the spindle assembly checkpoint (SAC) and is required P Kinetochore H2A for proper chromosome segregation in eukaryotes. H3 BUB1phosphorylates histone H2A at Thr120 (equivalent Haspin to yeast H2AS121) (see the figure, part a) around Histone DNA centromeres, where shugoshin (SGO) family proteins are Cohesin 85 known to bind through their SGO motif . Fission yeast HP1 Inner centromere and human SGO1 associate through their P-V-I motif with HP1 (heterochromatin protein 1), and the DNA b localization of SGO1 at the pericentric heterochromatin Phosphorylated H3T3 133 Phosphorylated H2AT120 depends partly on this interaction . In addition to BUB1, Haspin kinase haspin kinase has been shown to be crucial for Aurora B (H3 phosphorylation) localization by phosphorylating histone H3 at Thr3 (H3T3; see the figure, part a) only in M phase. Phosphorylated H3T3 interacts directly with the conserved BIR domain of survivin, and this interaction is important for targeting Aurora B to the centromeric nucleosome9093. Although BUB1 and haspin kinases cooperate in targeting Aurora B (as part of the chromosomal passenger BUB1 kinase complex (CPC), which also contains survivin and borealin) (H2A to centromeres, the activities of these kinases are phosphorylation) distributed differently along the chromosome (see the figure, part b); BUB1 localizes at kinetochores, whereas haspin is found at cohesin sites along the chromosome92. Intersection Accordingly, shugoshin proteins, which bind to phosphorylated H2A at Thr120 (H2AT120), are enriched at Phosphorylated Aurora B the inner kinetochore and the adjacent pericentric region H3T3 Phosphorylated H2AT120 (the inter-kinetochore axis), whereas survivin, which binds to phosphorylated H3T3, distributes at the inter-sister axis Nature Reviews | Molecular Cell Biology around centromeres (note that cohesin is protected by shugoshin proteins only in these regions). The CPC culminates in a merged region of these histone phosphorylation marks, at the centre between two sister kinetochores92. This geometry allows erroneous microtubule attachment to be destabilized by Aurora B, with only amphitelic attachment being stabilized (FIG.4). In meiosis, kinetochore geometry is modified so that Aurora B position relative to kinetochores is changed accordingly (see FIG.5). Genetic analyses in fission yeast have shown that the two pathways mediated by Bub1 and haspin are redundant in recruiting the CPC to centromeres92. Because phosphorylation of survivin (or of borealin in humans) by CDK1 promotes the interaction with the centromeric adaptor shugoshin , localization of the fission yeast Aurora B homologue to the centromere is limited to M phase of the cell cycle88.
Because haspin localizes at the sister chromatid cohesio n sites and BUB1 localizes at kinetochores, the AuroraBshugoshin complex is most enriched at the intersection of the two histone marks, the site known as the inner centro mere92 (BOX2). This is defined by the network comprising the kinetochoreBUB1 phosphorylate d H2AS121 pathway and the cohesin haspinphosphorylated H3T3 pathway, which I propose to call the inner centromereshugoshin (ICS) network. In addition to the overall distance of the kinetochore from Aurora B (or inter-kinetochore distance), local tension exerted at the attachment sites, which is sensi tively reflected by the intra-kinetochore stretch94,95, may also contribute to the stabilization of attachment 96. Thus, lack of tension is translated to a loss of attachment, which can directly activate the spindle assembly checkpoint (SAC) a signalling system that blocks cell cycle progression into anaphase if incorrect attachment is present 9799.
Tension stabilizes biorientation of bivalents in meiosis. During meiotic prophase I, homologues are held together by chiasmata, which are produced by reciprocal recombination between maternal and paternal chromatids, and this results in paired homologues, known as a bivalent (FIG.1b). Because sister kinetochores are fused before chromosome alignment, tension is produced when a bivalent is captured by micro tubules from opposite poles. Seminal experiments in grasshopper spermatocytes suggest that, as in mitosis (see above), tension is a fundamental requirement for stabilizing the bi-orientation of bivalents100. By using a micro-needle, the author exquisitely manipulated the bivalent to observe the process of chromosome alignment in meiosis I. This study demonstrated that monopolar attachment of a bivalent is not stable, so the bivalent reorients, which gives rise to bipolar attachment. Strikingly, the induced monopolar attachment of a bivalent could be stabilized by manually pulling the
VOLUME 13 | JUNE 2012 | 377
REVIEWS
Meiosis I Aurora B
rec12
(Merotelic bi-orientation)
Kinetochore Univalent 6 univalents
Aurora B CENPC Microtubules
Kinetochore geometry
Chiasmata Wild type Aurora B CENPC 3 sister chromatids

Mitosis
Microtubule
Bivalent
3 bivalents
Figure 5 | Repositioning of Aurora B promoted by chiasmata establishes meiotic chromosome InmeiosisI, Nature Reviews orientation. | Molecular Cell Biology sister kinetochores are juxtaposed or fused because of the specialized sister chromatid cohesion that is established at the core centromeres (FIG.3). This geometry positions Aurora B (blue) next to paired kinetochores (red), rather than between them, as observed in mitosis. The initial attachment of fused kinetochores is often merotelic and unstable. This unstable status persists in univalents because the meiotic kinetochore geometry prevents the amphitelic attachment that is usually observed in mitosis and that can position attachment sites far from Aurora B. Only in the presence of chiasmata does the bi-orientation of bivalents change this unstable attachment of fused kinetochores into stable syntelic attachment. Immunofluorescence images show Aurora B (green), centromere protein C (CENPC; shown in magenta) and microtubules (blue) of fission yeast cells (three chromosomes in a haploid genome) in mitosis (left) and meiosis I (right), the latter in rec12 mutant and wild-type cells). Images are reproduced, with permission, from REF. 108 (2011) Elsevier.
bivalent in the opposite direction to generate tension101, which indicates that tension is the trigger to stabilize the attachment of bivalents. So how does tension stabilize bivalent orientation? Similarly to mitosis, genetic studies have shown that the Aurora B homologues in yeast, Ipl1 and Ark1, have an essential role in chromosome orientation in meiosis. Indeed, their inactivation stabilizes erroneous attachment and leads to SAC silencing 67,102. Recent studies using Aurora kinase inhibitors indicate that this is also true in mouse oocytes103107, although Aurora C (which is a meiosis-specific Aurora kinase) may largely replace the Aurora B function in mammalian meiosis107. Precise measurements of the position of Aurora B combined with inspection of kinetochore behaviour in fission yeast have provided mechanistic insights into how the monopolar attachment of fused sister kinetochores is stabilized in bivalents108. When bivalents bi-orient at metaphase I, the kinetochores localize at the very edge of the chromosome and oscillate along the spindl e co ordinately (but keeping kinetochore distance constant), whereas AuroraB always localizes inwards from the kinetochores. In bivalents, fused sister kinetochores can be separated from the Aurora B sites only when they are attached and pulled in the same direction and the other half-bivalent that is connected by chiasmata is pulled in the opposite direction (FIG.5, bivalent). Thus, the monopolar attachment of sister kinetochores is stabilize d by the bipolar attachment of the bivalent.
378 | JUNE 2012 | VOLUME 13
In both meiosis I and II, the Aurora B yeast homologues (or Aurora C in mammals) are localized at the inner centromeres as in mitosis108,109, and their localization partly depends on shugoshin proteins and Bub1, at least in fission yeast 86,102. Thus, the ICS network may have a crucial role in defining the localization of Aurora B in meiosis as well. The involvement of PP1 or PP2A in the stabilization of kinetochoremicrotubule attachment has not been well defined in meiosis. Behaviour of univalents. If homologues are not properly paired at prophase I through chiasmata, each chromatid (univalent) shows unstable attachment because the side-by-side configuration of sister kinetochores may force their mono-orientation (FIG.1b). In this case, because of the lack of chiasmata, Aurora B localizes close to the fused kinetochores of the univalent, which randomly move along the spindle, indicative of unstable attachment (FIG.5, univalent). Thus, the fused kinetochores of a univalent cannot stabilize any attachment in principle. However, merotelic attachment is somehow stabilized during prolonged metaphase, thereby leading to evasion of the SAC108. For example, in trisomic maize, the univalent that is excluded from the exchange with the other two homologues often undergoes equational separation of sister kinetochores in meiosis I110. Similarly, in XO oocytes, which carry only one Xchromosome, equational segregation of the X univalent at anaphase I reaches 32%111.
REVIEWS
Moreover, Mlh1 -deficient mice, in which meiotic recombination is initiated but not repaired properly, produce substantial numbers of univalents that largely exhibit unstable attachment, although some establish bi-orientatio n at metaphase I112. Thus, univalents often, but not always, establish bi-orientation. Analysis of the fission yeast rec12 (a homologue of budding yeast SPO11) mutant, which is defective in recombination, reveals that a predominant population of univalents exhibits unstable bi-orientation of fused kinetochores in a merotelic manner 108,113 (FIG.5). This merotelic bipolar attachment is unstable, but is somehow stabilized after extended metaphase and finally evades the SAC, leading to anaphase. Because the protection of cohesion at the centromeres by shugoshin proteins is robust in fission yeast, unlike in oocytes, most merotelic attachment of univalents ends with the co-segregation of sister chromatids at anaphaseI despite exhibiting transient splitting of sister kinetochores, whereas homologues segregate randomly 108,113. Although the precise mechanism of this late stabilization of merotelic attachment remains elusive, one plausible explanation is that repeated challenge by microtubules might impair kinetochore rigidity or geometry, and this leads to the excessive extrusion of the attachment sites from the Aurora B sites108. Univalents that are produced as a consequence of recombination defects often lack the synaptonemal complex, which is usually formed between paired homologues. However, mouse oocytes in which Mlh1 has been deleted achieve full synaptonemal complex formation between homologues in prophase I, but the resultant univalents still bi-orient 112. Thus, lack of the synaptonemal complex might not be the reason of the aberrant orientation of univalents. Moreover, in the fission yeast rec12 mutant, cohesion at the core centromere is preserved, which indicates that kinetochore geometry is kept intact 108 (FIG.1b). Thus, recombination components or the synaptonemal complex might not be directly involved in the regulation of kinetochore orientation of univalents. Crucially, bi-orientation of univalents in fission yeast rec12-mutant cells is completely suppressed by introducing an artificial reciprocal exchange of sister strands between two univalents108. Thus, monopolar attachment is indeed established, and this depends on the physical linkage between homologues. cohesion is an underlying cause of age-related aneuploidy 117,118. Mice that are heterozygous for Bub1 also show an age-dependent increase in aneuploidy 119. Recent studies of old mice have revealed that the localization of REC8-containing cohesin is much lower in the chromosomes of aged oocytes120,121, leading to weakened sister chromatid cohesion and an increase in the incidence of univalents. Notably, the distance between sister kinetochores increases even in bivalents, suggesting that dysfunction of kinetochore geometry also contributes to the abnormal chromosome segregation in meiosis I120. Aged oocytes do not only show decreased levels of cohesin but also of its protector SGOL2 (REF. 121), which, as in fission yeast, may cause a synergistic defect in chromosome segregation at meiosis I108. Recent studies suggest that cohesin turnover in oocytes is scarcely detectable during the growing phase and even several months after birth, and these results support the hypothesis that cohesin loaded onto chromosomes only before birth has to maintain the cohesion of bivalents and is lost in an age-related manner 122,123. In addition to weakened cohesion, the SAC might be a risk factor for aneuploidy in oocytes. Indeed, misalignment of one or a few chromosomes is occasionally neglected by the SAC in oocytes111,112, presumably because oocytes have a larger cell volume per chromosome than mitotic cells, in which SAC activation at a single kineto chore is sufficient to generate an activation signal that is transduced to the rest of the cell. Moreover, univalents, which usually show unstable attachment, occasionally establish merotelic bipolar attachment, which also contributes to the evasion of the SAC108,124. Thus, it is likely that age-related impairment of cohesion leads to defects in chiasmata, kinetochore geometry and centromeric protection. These defects and the intrinsic weakness of the SAC may contribute synergistically to aneuploidy in aged oocytes.
Aneuploidy in aged oocytes In mammals, oocytes start to undergo meiosis in the fetal ovary and then enter a prophase I arrest following the formation of bivalent chromosomes. Remarkably, this arrest lasts until ovulation, which is an interval of many months in mice but several decades in humans, and this potentially explains why maternal age-related miscarriage and birth defects originate predominantly from chromosome segregation errors during meiosis I114. The precise mechanism of this failure has been elusive, although age-related increases in univalents have been observed in both female mice and humans115,116. The first mouse genetic mutant with enhanced agerelated oocyte defects was obtained by deleting meiosisspecific cohesin Smc1b, which suggests that deficient
Conclusions and perspectives Kinetochore are captured by the spindle microtubules and have a fundamental role in chromosome segregation. In mitosis, sister chromatid cohesion is tightly established at the pericentric region but not at the core centromere, thus promoting the separation of assembled sister kinetochores and bi-orientation at metaphase. By contrast, in meiosis I there is cohesion at the core centromere, which promotes monopolar attachment of sister kinetochores. Therefore, attachment is stabilized only when the bivalent is captured from opposite poles, which exerts full tension at the kinetochores. This principle might have been modified from that in mitosis, in which tension is produced in sister kinetochores that are captured from opposite poles. It has become increasingly clear that Aurora B kinase plays a central part in monitoring the tension that is exerted on kinetochores, promoting the re-orientation of incorrect (tensionless) attachment in both mitosis and meiosis. Although kinetochore geometry might influence, at least in part, the initial access of microtubules to kinetochores, it has been shown that the mitotic kineto chores in rat kangaroo cells when attached in a syntelic
VOLUME 13 | JUNE 2012 | 379
REVIEWS
manner (this occurs frequently in normal prometaphase) tend to remain side-by-side; they return to a back-to-back orientation only by the external force of microtubules125. Thus, kinetochore structure might be plastic rather than elastic, suggesting that the access of microtubules is not always properly biased by kinetochore geometry. Moreover, in the early stage of prometaphase I, fused sister kinetochores frequently attain merotelic attachment, which is challenged over time and corrected depending on the presence of chiasmata103,108 (FIG.5). These observations indicate that kinetochore geometry plays a crucial part at the reorientation step rather than at the initial attachment step. The initial attachment would be relatively unstable, as the number of attached microtubules is not sufficient to produce full tension, but this premature tension may transiently define the relative position of Aurora B and microtubule attachment sites within the centromeres. Only favourable attachment that is located away from Aurora B would be stabilized, whereas unfavourable attachment that is close to Aurora B would be destabilized and re-oriented. This model seems applicable to both mitosis and meiosis (FIGS4,5) and supports the notion that chromosome orientation is determined by kinetochore geometry and tension exerted by microtubules. Recent developments in live-cell imaging enabled complete kinetochore tracking during chromosome congression in mitosis and meiosis103,126. Further progress of this technique will help to visualize the complete process of search and capture on the single microtubule level and will prove or improve the above-mentioned geometry and tension model for kinetochore orientation. Cohesion at the inner centromere (pericentric region) is integral for joining sister centromeres to ensure the back-to-back configuration of kinetochores and to promote bi-orientation. In fact, weakened centromeric cohesion triggers merotelic attachment and lagging of chromosomes at anaphase, which might be one of the major reasons for the chromosomal instability that is frequently observed in cancer cells9,127129. It is notable that haspin kinase, which directly regulates the centromeric localization of Aurora B, binds to the cohesin-associating protein PDS5 and mislocalizes in cohesin mutants92 (BOX2). Thus, cohesion defects may impair not only tension across centromeres or the basis of kinetochore geometry but also the ICS network, thereby predisposing cells to chromosomal instability in multiple ways. It is reasonable to consider that the ICS network, which has a fundamental role in linking sister chromatid cohesion to chromosome orientation (BOX2), substantially contributes to the prevention of tumorigenesis or birth defects in humans. This possibility should be examined in future studies. Further work will define the basic molecular mechanisms for the regulation of chromosome segregation and will contribute to our understanding of how this miraculous system sustains the succession oflife.
1. 2. 3. 4.
5. 6. 7. 8. 9. 10. 11. 12. 13.
14.
Ricke, R.M., van Ree, J.H. & van Deursen, J.M. Whole chromosome instability and cancer: a complex relationship. Trends Genet. 24, 457466 (2008). Gordon, D.J., Resio, B. & Pellman, D. Causes and consequences of aneuploidy in cancer. Nature Rev. Genet. 13, 189203 (2012). Moore, D.P. & Orr-Weaver, T.L. Chromosome segregation during meiosis: building an unambivalent bivalent. Curr. Top. Dev. Biol. 37, 263299 (1998). Petronczki, M., Siomos, M.F. & Nasmyth, K. Un menage a quatre: the molecular biology of chromosome segregation in meiosis. Cell 112, 423440 (2003). Tanaka, T.U. Kinetochore-microtubule interactions: steps towards biorientation. EMBO J. 29, 40704082 (2010). Hauf, S. & Watanabe, Y. Kinetochore orientation in mitosis and meiosis. Cell 119, 317327 (2004). Cheeseman, I.M. & Desai, A. Molecular architecture of the kinetochore-microtubule interface. Nature Rev. Mol. Cell Biol. 9, 3346 (2008). Santaguida, S. & Musacchio, A. The life and miracles of kinetochores. EMBO J. 28, 25112531 (2009). Cimini, D. etal. Merotelic kinetochore orientation is a major mechanism of aneuploidy in mitotic mammalian tissue cells. J.Cell Biol. 153, 517527 (2001). Uhlmann, F. A matter of choice: the establishment of sister chromatid cohesion. EMBO Rep. 10, 10951102 (2009). Nasmyth, K. Cohesin: a catenase with separate entry and exit gates? Nature Cell Biol. 13, 11701177 (2011). Watanabe, Y. Sister chromatid cohesion along arms and at centromeres. Trends Genet. 21, 405412 (2005). Ishiguro, K., Kim, J., Fujiyama-Nakamura, S., Kato, S. & Watanabe, Y. A new meiosis-specific cohesin complex implicated in the cohesin code for homologous pairing. EMBO Rep. 12, 267275 (2011). Lee, J. & Hirano, T. RAD21L, a novel cohesin subunit implicated in linking homologous chromosomes in mammalian meiosis. J.Cell Biol. 192, 263276 (2011).
15. Herran, Y. etal. The cohesin subunit RAD21L functions in meiotic synapsis and exhibits sexual dimorphism in fertility. EMBO J. 30, 30913105 (2011). 16. Severson, A.F., Ling, L., van Zuylen, V. & Meyer, B.J. The axial element protein HTP3 promotes cohesin loading and meiotic axis assembly in C.elegans to implement the meiotic program of chromosome segregation. Genes Dev. 23, 17631778 (2009). 17. Revenkova, E. & Jessberger, R. Keeping sister chromatids together: cohesins in meiosis. Reproduction 130, 783790 (2005). 18. Kitajima, T.S., Yokobayashi, S., Yamamoto, M. & Watanabe, Y. Distinct cohesin complexes organize meiotic chromosome domains. Science 300, 11521155 (2003). 19. Klein, F. etal. A central role for cohesins in sister chromatid cohesion, formation of axial elements, and recombination during yeast meiosis. Cell 98, 91103 (1999). The first study describing the central role of meiotic cohesin in chromosome differentiation. 20. Watanabe, Y. & Nurse, P. Cohesin Rec8 is required for reductional chromosome segregation at meiosis. Nature 400, 461464 (1999). The first study to show that cohesin Rec8 is required for establishing mono-orientation of sister kinetochores. 21. Xu, H., Beasley, M.D., Warren, W.D., van der Horst, G.T. & McKay, M.J. Absence of mouse REC8 cohesin promotes synapsis of sister chromatids in meiosis. Dev. Cell 8, 949961 (2005). 22. Kim, K.P. etal. Sister cohesion and structural axis components mediate homolog bias of meiotic recombination. Cell 143, 924937 (2010). 23. Ellermeier, C. & Smith, G.R. Cohesins are required for meiotic DNA breakage and recombination in Schizosaccharomyces pombe. Proc. Natl Acad. Sci. USA 102, 1095210957 (2005). 24. Neale, M.J. & Keeney, S. Clarifying the mechanics of DNA strand exchange in meiotic recombination. Nature 442, 153158 (2006). 25. Gerton, J.L. & Hawley, R.S. Homologous chromosome interactions in meiosis: diversity amidst conservation. Nature Rev. Genet. 6, 477487 (2005).
26. Bhalla, N. & Dernburg, A.F. Prelude to a division. Annu. Rev. Cell Dev. Biol. 24, 397424 (2008). 27. Scherthan, H. A bouquet makes ends meet. Nature Rev. Mol. Cell Biol. 2, 621627 (2001). 28. Yamamoto, A. & Hiraoka, Y. How do meiotic chromosomes meet their homologous partners?: lessons from fission yeast. Bioessays 23, 526533 (2001). 29. Sato, A. etal. Cytoskeletal forces span the nuclear envelope to coordinate meiotic chromosome pairing and synapsis. Cell 139, 907919 (2009). 30. Wolf, K.W. How meiotic cells deal with non-exchange chromosomes. Bioessays 16, 107114 (1994). 31. Buonomo, S.B. etal. Disjunction of homologous chromosomes in meiosis I depends on proteolytic cleavage of the meiotic cohesin Rec8 by separin. Cell 103, 387398 (2000). 32. Kitajima, T.S., Miyazaki, Y., Yamamoto, M. & Watanabe, Y. Rec8 cleavage by separase is required for meiotic nuclear divisions in fission yeast. EMBO J. 22, 56435653 (2003). 33. Goldstein, L.S. Mechanisms of chromosome orientation revealed by two meiotic mutants in Drosophila melanogaster. Chromosoma 78, 79111 (1980). 34. Miyazaki, W.Y. & Orr-Weaver, T.L. Sister-chromatid cohesion in mitosis and meiosis. Annu. Rev. Genet. 28, 167168 (1994). 35. Kitajima, T.S., Kawashima, S.A. & Watanabe, Y. The conserved kinetochore protein shugoshin protects centromeric cohesion during meiosis. Nature 427, 510517 (2004). The first study to identify shugoshins as protectors of cohesin at the centromeres. 36. Rabitsch, K.P. etal. Two fission yeast homologs of Drosophila MeiS332 are required for chromosome segregation during meiosis I and II. Curr. Biol. 14, 287301 (2004). 37. Marston, A.L., Tham, W.H., Shah, H. & Amon, A. A genome-wide screen identifies genes required for centromeric cohesion. Science 303, 13671370 (2004).
REVIEWS
38. Katis, V.L., Galova, M., Rabitsch, K.P., Gregan, J. & Nasmyth, K. Maintenance of cohesin at centromeres after meiosis I in budding yeast requires a kinetochoreassociated protein related to MEIS332. Curr. Biol. 14, 560572 (2004). 39. Hamant, O. etal. A REC8dependent plant shugoshin is required for maintenance of centromeric cohesion during meiosis and has no mitotic functions. Curr. Biol. 15, 948954 (2005). 40. Lee, J. etal. Unified mode of centromeric protection by shugoshin in mammalian oocytes and somatic cells. Nature Cell Biol. 10, 4252 (2008). 41. Llano, E. etal. Shugoshin2 is essential for the completion of meiosis but not for mitotic cell division in mice. Genes Dev. 22, 24002413 (2008). 42. Wang, M. etal. OsSGO1 maintains synaptonemal complex stabilization in addition to protecting centromeric cohesion during rice meiosis. Plant J. 67, 583594 (2011). 43. Kitajima, T.S. etal. Shugoshin collaborates with protein phosphatase 2A to protect cohesin. Nature 441, 4652 (2006). 44. Riedel, C.G. etal. Protein phosphatase 2A protects centromeric sister chromatid cohesion during meiosisI. Nature 441, 5361 (2006). 45. Brar, G.A. etal. Rec8 phosphorylation and recombination promote the step-wise loss of cohesins in meiosis. Nature 441, 532536 (2006). 46. Ishiguro, T., Tanaka, K., Sakuno, T. & Watanabe, Y. ShugoshinPP2A counteracts caseinkinase1dependent cleavage of Rec8 by separase. Nature Cell Biol. 12, 500506 (2010). 47. Katis, V.L. etal. Rec8 phosphorylation by casein kinase 1 and Cdc7Dbf4 kinase regulates cohesin cleavage by separase during meiosis. Dev. Cell 18, 397409 (2010). 48. Salic, A., Waters, J.C. & Mitchison, T.J. Vertebrate shugoshin links sister centromere cohesion and kinetochore microtubule stability in mitosis. Cell 118, 567578 (2004). The first demonstration that shugoshins protect centromeric cohesion in vertebrate mitotic cells. 49. McGuinness, B.E., Hirota, T., Kudo, N.R., Peters, J.M. & Nasmyth, K. Shugoshin prevents dissociation of cohesin from centromeres during mitosis in vertebrate cells. PLoS Biol. 3, e86 (2005). 50. Kitajima, T.S., Hauf, S., Ohsugi, M., Yamamoto, T. & Watanabe, Y. Human Bub1 defines the persistent cohesion site along the mitotic chromosome by affecting shugoshin localization. Curr. Biol. 15, 353359 (2005). 51. Tanno, Y. etal. Phosphorylation of mammalian Sgo2 by Aurora B recruits PP2A and MCAK to centromeres. Genes Dev. 24, 21692179 (2010). 52. Huang, H. etal. Tripin/hSgo2 recruits MCAK to the inner centromere to correct defective kinetochore attachments. J.Cell Biol. 177, 413424 (2007). 53. Rivera, T. etal. Xenopus Shugoshin 2 regulates the spindle assembly pathway mediated by the chromosomal passenger complex. EMBO J. 31,14671479 (2012). 54. stergren, G. The mechanism of coorientation in bivalents and multivalents. Hereditas 37, 85156 (1951). 55. Journey, L.J. & Whaley, A. Kinetochore ultrastructure in vincristine-treated mammalian cells. J.Cell Sci. 7, 4954 (1970). 56. Goldstein, L.S. Kinetochore structure and its role in chromosome orientation during the first meiotic division in male D.melanogaster. Cell 25, 591602 (1981). This paper outlines the morphological differences between mitotic and meiotic sister kinetochores. 57. Hodges, C.A. & Hunt, P.A. Simultaneous analysis of chromosomes and chromosome-associated proteins in mammalian oocytes and embryos. Chromosoma 111, 165169 (2002). 58. Toth, A. etal. Functional genomics identifies monopolin: a kinetochore protein required for segregation of homologs during meiosis I. Cell 103, 11551168 (2000). By using a genetic approach, this study identifies monopolin as a necessary complex for sister kinetochore mono-orientation in budding yeast. 59. Rabitsch, K.P. etal. Kinetochore recruitment of two nucleolar proteins is required for homolog segregation in meiosis I. Dev. Cell 4, 535548 (2003). 60. Petronczki, M. etal. Monopolar attachment of sister kinetochores at meiosis I requires casein kinase 1. Cell 126, 10491064 (2006). 61. Katis, V.L. etal. Spo13 facilitates monopolin recruitment to kinetochores and regulates maintenance of centromeric cohesion during yeast meiosis. Curr. Biol. 14, 21832196 (2004). 62. Lee, B.H., Kiburz, B.M. & Amon, A. Spo13 maintains centromeric cohesion and kinetochore coorientation during meiosis I. Curr. Biol. 14, 21682182 (2004). 63. Matos, J. etal. Dbf4dependent CDC7 kinase links DNA replication to the segregation of homologous chromosomes in meiosis I. Cell 135, 662678 (2008). 64. Gregan, J. etal. The kinetochore proteins Pcs1 and Mde4 and heterochromatin are required to prevent merotelic orientation. Curr. Biol. 17, 11901200 (2007). 65. Corbett, K.D. etal. The monopolin complex crosslinks kinetochore components to regulate chromosomemicrotubule attachments. Cell 142, 556567 (2010). 66. Tada, K., Susumu, H., Sakuno, T. & Watanabe, Y. Condensin association with histone H2A shapes mitotic chromosomes. Nature 474, 477483 (2011). 67. Monje-Casas, F., Prabhu, V.R., Lee, B.H., Boselli, M. & Amon, A. Kinetochore orientation during meiosis is controlled by Aurora B and the monopolin complex. Cell 128, 477490 (2007). 68. Watanabe, Y., Yokobayashi, S., Yamamoto, M. & Nurse, P. Pre-meiotic S phase is linked to reductional chromosome segregation and recombination. Nature 409, 359363 (2001). 69. Pidoux, A. & Allshire, R. Kinetochore and heterochromatin domains of the fission yeast centromere. Chromosome Res. 12, 521534 (2004). 70. Yokobayashi, S., Yamamoto, M. & Watanabe, Y. Cohesins determine the attachment manner of kinetochores to spindle microtubules at meiosis I in fission yeast. Mol. Cell. Biol. 23, 39653973 (2003). 71. Yokobayashi, S. & Watanabe, Y. The kinetochore protein Moa1 enables cohesion-mediated monopolar attachment at meiosis I. Cell 123, 803817 (2005). 72. Kagami, A. etal. Acetylation regulates monopolar attachment at multiple levels during meiosis I in fission yeast. EMBO Rep. 12, 11891195 (2011). 73. Yu, H.G. & Dawe, R.K. Functional redundancy in the maize meiotic kinetochore. J.Cell Biol. 151, 131141 (2000). 74. Chelysheva, L. etal. AtREC8 and AtSCC3 are essential to the monopolar orientation of the kinetochores during meiosis. J.Cell Sci. 118, 46214632 (2005). 75. Sakuno, T., Tada, K. & Watanabe, Y. Kinetochore geometry defined by cohesion within the centromere. Nature 458, 852858 (2009). This study establishes the causal link between kinetochore geometry and cohesion within the centromere. 76. Tanaka, T.U. etal. Evidence that the Ipl1Sli15 (Aurora kinase-INCENP) complex promotes chromosome biorientation by altering kinetochorespindle pole connections. Cell 108, 317329 (2002). This first study to show that Aurora B destabilizes microtubulekinetochore attachment and thereby promotes chromosome biorientation. 77. Hauf, S. etal. The small molecule Hesperadin reveals a role for Aurora B in correcting kinetochoremicrotubule attachment and in maintaining the spindle assembly checkpoint. J.Cell Biol. 161, 281294 (2003). 78. Ruchaud, S., Carmena, M. & Earnshaw, W.C. Chromosomal passengers: conducting cell division. Nature Rev. Mol. Cell Biol. 8, 798812 (2007). 79. Liu, D., Vader, G., Vromans, M.J., Lampson, M.A. & Lens, S.M. Sensing chromosome biorientation by spatial separation of aurora B kinase from kinetochore substrates. Science 323, 13501353 (2009). This study demonstrates the influence of the relative location of Aurora B and kinetochores on the stability of the microtubulekinetochore attachment. 80. DeLuca, J.G. etal. Kinetochore microtubule dynamics and attachment stability are regulated by Hec1. Cell 127, 969982 (2006). 81. Welburn, J.P. etal. Aurora B phosphorylates spatially distinct targets to differentially regulate the kinetochore-microtubule interface. Mol. Cell 38, 383392 (2010). 82. Liu, D. etal. Regulated targeting of protein phosphatase 1 to the outer kinetochore by KNL1 opposes Aurora B kinase. J.Cell Biol. 188, 809820 (2010). 83. Lampson, M.A. & Cheeseman, I.M. Sensing centromere tension: Aurora B and the regulation of kinetochore function. Trends Cell Biol. 21, 133140 (2011). 84. Foley, E.A., Maldonado, M. & Kapoor, T.M. Formation of stable attachments between kinetochores and microtubules depends on the B56PP2A phosphatase. Nature Cell Biol. 13, 12651271 (2011). 85. Kawashima, S.A., Yamagishi, Y., Honda, T., Ishiguro, K. & Watanabe, Y. Phosphorylation of H2A by Bub1 prevents chromosomal instability through localizing shugoshin. Science 327, 172177 (2010). 86. Kawashima, S.A. etal. Shugoshin enables tensiongenerating attachment of kinetochores by loading Aurora to centromeres. Genes Dev. 21, 420435 (2007). 87. Vanoosthuyse, V., Prykhozhij, S. & Hardwick, K.G. Shugoshin 2 regulates localization of the chromosomal passenger proteins in fission yeast mitosis. Mol. Biol. Cell 18, 16571669 (2007). 88. Tsukahara, T., Tanno, Y. & Watanabe, Y. Phosphorylation of the CPC by Cdk1 promotes chromosome biorientation. Nature 467, 719723 (2010). 89. Indjeian, V.B., Stern, B.M. & Murray, A.W. The centromeric protein Sgo1 is required to sense lack of tension on mitotic chromosomes. Science 307, 130133 (2005). 90. Kelly, A.E. etal. Survivin reads phosphorylated histone H3 threonine 3 to activate the mitotic kinase Aurora B. Science 330, 235239 (2010). 91. Wang, F. etal. Histone H3 Thr3 phosphorylation by Haspin positions Aurora B at centromeres in mitosis. Science 330, 231235 (2010). 92. Yamagishi, Y., Honda, T., Tanno, Y. & Watanabe, Y. Two histone marks establish the inner centromere and chromosome biorientation. Science 330, 239243 (2010). This study, together with references 85, 90 and 91, establishes that the inner centromere (or Aurora B position) is defined by the ICS network, which is composed of two histone phosphorylation pathways. 93. Jeyaprakash, A.A., Basquin, C., Jayachandran, U. & Conti, E. Structural basis for the recognition of phosphorylated histone h3 by the survivin subunit of the chromosomal passenger complex. Structure 19, 16251634 (2011). 94. Maresca, T.J. & Salmon, E.D. Intrakinetochore stretch is associated with changes in kinetochore phosphorylation and spindle assembly checkpoint activity. J.Cell Biol. 184, 373381 (2009). 95. Uchida, K.S. etal. Kinetochore stretching inactivates the spindle assembly checkpoint. J.Cell Biol. 184, 383390 (2009). 96. Akiyoshi, B. etal. Tension directly stabilizes reconstituted kinetochoremicrotubule attachments. Nature 468, 576579 (2010). 97. Pinsky, B.A. & Biggins, S. The spindle checkpoint: tension versus attachment. Trends Cell Biol. 15, 486493 (2005). 98. Nezi, L. & Musacchio, A. Sister chromatid tension and the spindle assembly checkpoint. Curr. Opin. Cell Biol. 21, 785795 (2009). 99. Murray, A.W. A brief history of error. Nature Cell Biol. 13, 11781182 (2011). 100. Nicklas, R.B. How cells get the right chromosomes. Science 275, 632637 (1997). 101. Nicklas, R.B. & Koch, C.A. Chromosome micromanipulation. 3. Spindle fiber tension and the reorientation of mal-oriented chromosomes. J.Cell Biol. 43, 4050 (1969). Micromanipulation experiments in meiotic cells demonstrate that the physical appliance of tension can lead to a stabilization of attachment. 102. Hauf, S. etal. Aurora controls sister kinetochore mono-orientation and homolog biorientation in meiosisI. EMBO J. 26, 44754486 (2007). 103. Kitajima, T.S., Ohsugi, M. & Ellenberg, J. Complete kinetochore tracking reveals error-prone homologous chromosome biorientation in mammalian oocytes. Cell 146, 568581 (2011). 104. Lane, S.I., Chang, H.Y., Jennings, P.C. & Jones, K.T. The Aurora kinase inhibitor ZM447439 accelerates first meiosis in mouse oocytes by overriding the spindle assembly checkpoint. Reproduction 140, 521530 (2010). 105. Shuda, K., Schindler, K., Ma, J., Schultz, R.M. & Donovan, P.J. Aurora kinase B modulates chromosome alignment in mouse oocytes. Mol. Reprod. Dev. 76, 10941105 (2009). 106. Sharif, B. etal. The chromosome passenger complex is required for fidelity of chromosome transmission and cytokinesis in meiosis of mouse oocytes. J.Cell Sci. 123, 42924300 (2010).
VOLUME 13 | JUNE 2012 | 381
REVIEWS
107. Yang, K.T. etal. AuroraC kinase deficiency causes cytokinesis failure in meiosis I and production of large polyploid oocytes in mice. Mol. Biol. Cell 21, 23712383 (2010). 108. Sakuno, T., Tanaka, K., Hauf, S. & Watanabe, Y. Repositioning of Aurora B promoted by chiasmata ensures sister chromatid mono-orientation in meiosisI. Dev. Cell 21, 534545 (2011). This study identifies a molecular mechanism that explains how chiasmata are more likely to cause the bi-orientation of bivalents than the bi-orientation of univalents at meiosis I. 109. Parra, M.T. etal. A perikinetochoric ring defined by MCAK and AuroraB as a novel centromere domain. PLoS Genet. 2, e84 (2006). 110. Maguire, M.P. A possible role for the synaptonemal complex in chiasma maintenance. Exp. Cell Res. 112, 297308 (1978). 111. LeMaire-Adkins, R., Radke, K. & Hunt, P.A. Lack of checkpoint control at the metaphase/anaphase transition: a mechanism of meiotic nondisjunction in mammalian females. J.Cell Biol. 139, 16111619 (1997). 112. Nagaoka, S.I., Hodges, C.A., Albertini, D.F. & Hunt, P.A. Oocyte-specific differences in cell-cycle control create an innate susceptibility to meiotic errors. Curr. Biol. 21, 651657 (2011). 113. Hirose, Y. etal. Chiasmata promote monopolar attachment of sister chromatids and their cosegregation toward the proper pole during meiosisI. PLoS Genet. 7, e1001329 (2011). 114. Hassold, T. & Hunt, P. To err (meiotically) is human: the genesis of human aneuploidy. Nature Rev. Genet. 2, 280291 (2001). 115. Henderson, S.A. & Edwards, R.G. Chiasma frequency and maternal age in mammals. Nature 218, 2228 (1968). 116. Angell, R.R., Xian, J., Keith, J., Ledger, W. & Baird, D.T. First meiotic division abnormalities in human oocytes: mechanism of trisomy formation. Cytogenet. Cell Genet. 65, 194202 (1994). 117. Revenkova, E. etal. Cohesin SMC1 is required for meiotic chromosome dynamics, sister chromatid cohesion and DNA recombination. Nature Cell Biol. 6, 555562 (2004). 118. Hodges, C.A., Revenkova, E., Jessberger, R., Hassold, T.J. & Hunt, P.A. SMC1-deficient female mice provide evidence that cohesins are a missing link in age-related nondisjunction. Nature Genet. 37, 13511355 (2005). 119. Leland, S. etal. Heterozygosity for a Bub1 mutation causes female-specific germ cell aneuploidy in mice. Proc. Natl Acad. Sci. USA 106, 1277612781 (2009). 120. Chiang, T., Duncan, F.E., Schindler, K., Schultz, R.M. & Lampson, M.A. Evidence that weakened centromere cohesion is a leading cause of age-related aneuploidy in oocytes. Curr. Biol. 20, 15221528 (2010). 121. Lister, L.M. etal. Age-related meiotic segregation errors in mammalian oocytes are preceded by depletion of cohesin and Sgo2. Curr. Biol. 20, 15111521 (2010). 122. Revenkova, E., Herrmann, K., Adelfalk, C. & Jessberger, R. Oocyte cohesin expression restricted to predictyate stages provides full fertility and prevents aneuploidy. Curr. Biol. 20, 15291533 (2010). 123. Tachibana-Konwalski, K. etal. Rec8containing cohesin maintains bivalents without turnover during the growing phase of mouse oocytes. Genes Dev. 24, 25052516 (2010). 124. Kouznetsova, A., Lister, L., Nordenskjold, M., Herbert, M. & Hoog, C. Biorientation of achiasmatic chromosomes in meiosis I oocytes contributes to aneuploidy in mice. Nature Genet. 39, 966968 (2007). The first study to suggest that biorientated univalents evade the SAC, thus raising the risk of aneuploidy. 125. Loncarek, J. etal. The centromere geometry essential for keeping mitosis error free is controlled by spindle forces. Nature 450, 745749 (2007). 126. Magidson, V. etal. The spatial arrangement of chromosomes during prometaphase facilitates spindle assembly. Cell 146, 555567 (2011). 127. Barber, T.D. etal. Chromatid cohesion defects may underlie chromosome instability in human colorectal cancers. Proc. Natl Acad. Sci. USA 105, 34433448 (2008). 128. Solomon, D.A. etal. Mutational inactivation of STAG2 causes aneuploidy in human cancer. Science 333, 10391043 (2011). 129. Holland, A.J. & Cleveland, D.W. Boveri revisited: chromosomal instability, aneuploidy and tumorigenesis. Nature Rev. Mol. Cell Biol. 10, 478487 (2009). 130. Kiburz, B.M., Amon, A. & Marston, A.L. Shugoshin promotes sister kinetochore biorientation in Saccharomyces cerevisiae. Mol. Biol. Cell 19, 11991209 (2008). 131. Tang, Z., Sun, Y., Harley, S.E., Zou, H. & Yu, H. Human Bub1 protects centromeric sister-chromatid cohesion through Shugoshin during mitosis. Proc. Natl Acad. Sci. USA 101, 1801218017 (2004). 132. Boyarchuk, Y., Salic, A., Dasso, M. & Arnaoutov, A. Bub1 is essential for assembly of the functional inner centromere. J.Cell Biol. 176, 919928 (2007). 133. Yamagishi, Y., Sakuno, T., Shimura, M. & Watanabe, Y. Heterochromatin links to centromeric protection by recruiting shugoshin. Nature 455, 251255 (2008). 134. Kiburz, B.M. etal. The core centromere and Sgo1 establish a 50kb cohesin-protected domain around centromeres during meiosis I. Genes Dev. 19, 30173030 (2005). 135. Lampert, F. & Westermann, S. A blueprint for kinetochores new insights into the molecular mechanics of cell division. Nature Rev. Mol. Cell Biol. 12, 407412 (2011).
Acknowledgements
The author thanks S. Hauf for critically reading the manuscript and his current and previous laboratory members for discussions. The author apologizes to authors whose work was not discussed in this Review owing to space limitations. Work in the Y.W. laboratory was supported by a GrantinAid for Specially Promoted Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
The author declares no competing financial interests.
FURTHER INFORMATION
Yoshinori Watanabes homepage: http://www.iam.u-tokyo.ac.jp/watanabe-lab/eindex.html
REVIEWS
Specialized ribosomes: a new frontier in gene regulation and organismal biology

Shifeng Xue and Maria Barna
Abstract | Historically, the ribosome has been viewed as a complex ribozyme with constitutive rather than intrinsic regulatory capacity in mRNA translation. However, emerging studies reveal that ribosome activity may be highly regulated. Heterogeneity in ribosome composition resulting from differential expression and post-translational modifications of ribosomal proteins, ribosomal RNA (rRNA) diversity and the activity of ribosome-associated factors may generate specialized ribosomes that have a substantial impact on how the genomic template is translated into functional proteins. Moreover, constitutive components of the ribosome may also exert more specialized activities by virtue of their interactions with specific mRNA regulatory elements such as internal ribosome entry sites (IRESs) or upstream open reading frames (uORFs). Here we discuss the hypothesis that intrinsic regulation by the ribosome acts to selectively translate subsets of mRNAs harbouring unique cis-regulatory elements, thereby introducing an additional levelof regulation in gene expression and the life of an organism.
A major challenge in biology is to understand how protein expression is regulated with exquisite temporal and spatial precision to control unique cell behaviours and to give rise to the remarkable diversity of cell types that is characteristic of metazoan development. Decades of research have demonstrated numerous layers of regulation that orchestrate this process. Indeed, almost every step in the control of gene expression has evolved many levels of regulatory specificity: from exquisitely modified histone tails, to the countless small RNAs that fine-tune the abundance of transcripts, as well as the remarkable repertoire of post-translational modifications that steer protein behaviour. On the contrary, such regulatory control has not been fully realized at a crucial step in gene expression, that of protein production. In particular, although the ribosome is well known to be an immensely complex molecular machine, it is believed to function as somewhat of an automaton in translatin g the geneticcode. Recent findings are now, however, shifting the view of ribosome function, revealing unique and unexpected roles for the translation machinery itself in directing fundamental aspects of cell and organismal biology. Here we discuss the hypothesis that specialized ribosomes, which have a unique composition or specialized activity, confer regulatory control in gene expression. Examples of heterogeneity in ribosomes can include diversity in the composition and post-translational modifications of subsets of ribosomal proteins, variations in ribosomal RNA (rRNA) sequences or binding to distinct ribosome-associated factors, all of which may contribute to the occurrence of specialized ribosomes in different cell types. In addition, even core ribosome components that show little variation may exert more specialized activity by virtue of their interactions with specific cis-acting regulatory elements that are present within subsets of mRNAs. In a multicellular organism, greater regulation in ribosome activity may provide an important new layer for the spatiotemporal control of gene expression. In this Review, we examine the evidence for hetero geneity in ribosomes, how regulatory elements in mRNAs may interface with ribosomes to confer transcrip t-specific regulation of gene expression, and how such mechanisms may bestow greater specificity in decoding the genomic template into unique cellular and molecular behaviours as well as morphology.
Department of Biochemistry and Biophysics, Cardiovascular Research Institute, University of California, San Francisco, California94158, USA. Correspondence to M.B. e-mail: maria.barna@ucsf.edu doi:10.1038/nrm3359
REVIEWS
Box 1 | Ribosome biogenesis
Ribosome biogenesis is a highly orchestrated process involving all three RNA polymerases and over 150 non-ribosomal factors that take part in the maturation process129. Three of the four ribosomal RNA (rRNA) species, namely 18S, 5.8S and 28S, are transcribed as a single 47S transcript (pre-rRNA) by RNA polymerase I (Pol I) in the nucleolus (see the figure). The fourth rRNA, 5S, is transcribed by Pol III in the nucleus. Strikingly, pre-rRNAs are extensively modified by pseudouridylation () and methylation (M). This process is orchestrated by small nucleolar RNAs (snoRNAs) that guide site-specific modifications of ~200 nucleotides in the pre-rRNA. Although the roleof rRNA modifications is not fully understood, they are important for ribosome function130. For example, recent findings demonstrate that loss of rRNA pseudouridylation decreases translational fidelity, transfer RNA (tRNA) binding and binding to structured RNA elements within mRNAs in an evolutionarily conserved manner from yeast to human cells87. All ribosomal proteins are transcribed by Pol II and translated in the cytoplasm (not shown). They are then imported into the nucleus and are assembled on the pre-rRNA while it is being transcribed. Although less well studied, some ribosomal proteins have also been shown to be involved in rRNA processing, subunit assembly and export131135. As the ribosomal proteins assemble, the pre-rRNA folds and undergoes a series of cleavages to generate the 18S, 5.8S and 28S rRNAs. The 18S rRNA along with 32 ribosomal proteins assemble into the small ribosomal 40S subunit. The 28S, 5.8S and 5SrRNA along with additional 47 ribosomal proteins assemble into the large 60S subunit. Both ribosomal subunits are exported into the cytoplasm in a coordinated process136,137. RPL, large subunit ribosomal protein; RPS, small subunit ribosomal protein.
rDNA locus Pol I 18S 5.8S 28S 18S 5.8S 28S
Transcription Pol III 5S Large subunit components 5S

RPS 18S 5.8S
RPL 28S
singular, unilateral function. The basic mechanisms of how the genetic code is translated into proteins fits well with this concept. For example, the ribosome has a universally conserved role in catalysing protein synthesis. Moreover, protein synthesis is remarkably fast. Elongation of the polypeptide chain by one amino acid occurs in approximately 60ms and in a highly accurate fashion4. This requires a sophisticated level of precision and automation. However, the powerful metaphor of a machine has also perhaps stereotyped the view of the ribosome as a passive participant in translation. A crucial question is whether certain parts of the ribosome are changeable and/or activated by additional layers of regulation that depend on the context. To understand this principle, the molecular machine has to be examined at its very core, as ribosome biogenesis is a highly coordinated process (BOX1). Indeed, ribosomes make up much of the cells mass, and ribosome biogenesis uses a significant proportion of the cells energy 5. The eukaryotic ribosome contains 4 RNA species and 79ribosomal proteins. Despite the constitutive nature of the ribosome, several components could also give rise to divergent regulation of the molecular machine. In understanding the construction of the ribosome (BOX1), possible points of variation can also be envisioned. We will discuss the evidence, or the lack thereof, for individual components exerting unique and more regulatory control.
Cleavage and modications 28S

M
5.8S
M
RPS
M
18S
5.8S
M
28S
RPL
RPL
RPL Small subunit components
Nucleolus Nucleus RPS RPL
Nuclear export RPL
18S
M
RPS RPS Cytoplasm 60S RPS 40S RPS RPL
Ribosomal proteins imported into the nucleolus for assembly Nature Reviews | Molecular Cell Biology
Nucleolus
A conserved organelle that assembles around ribosomal DNA genes. The nucleolus is the site of ribosomal RNA transcription and the site of ribosome subunit assembly.
The ribosome as a molecular machine The ribosome has been described as a molecular machine par excellence1,2. This metaphor can be traced back to an influential article written by Bruce Alberts, in which the analogy of the molecular machine was first used to describe the highly coordinated behaviour of large complexes such as the DNA replication machinery or the ribosome3. The machine analogy influenced the manner in which the molecular complex was conceptualized; that although it may be composed of a number of distinct elements, a high degree of coordination and precision of these parts serves a
Heterogeneity in ribosome function and activity Francis Crick first proposed, in relation to the RNA world hypothesis, that an ancestral proto-ribosome may have been comprised only of functional rRNAs, which shared homology with present-day rRNAs6. Biochemical and genetic experiments have revealed a direct involvement of conserved rRNA nucleotides, highlighting their central role in ribosome function7,8. Remarkably, early attempts in delineating the minimal components necessary for ribosome activity demonstrated that peptidyl transferase, the ribosomal activity responsible for the catalysis of peptide-bond formation, might only require rRNAs in the absence of the majority of ribo somal proteins9. This raises the question of the function of the ribosomal proteins that associate with the ribosome. Ribosomal proteins are likely to have coevolved important roles in rRNA folding and function10,11 (BOX1). However, we speculate that the emergence of ribosomal proteins might, in addition, imbue greater specificity to the RNA-based translation machinery to control protein synthesis. It is also important to consider whether ribosomal proteins may be acting at multiple times and/or locations to exert unique activities. For example, ribosomal proteins may have important roles in rRNA processing within the nucleolus and also exert more specialized functions in translational control as part of the mature ribosome within the cytoplasm or even as components of extraribosomal complexes (BOX2). In this section, we examin e the possibility of further regulation in ribosome composition and activity with respect to the proteins and rRNAs that constitute the ribosome.
REVIEWS
Box 2 | Extraribosomal functions of ribosomal proteins in translational control
Some ribosomal proteins have been shown to have additional 3 UTR AAAAAA functions besides their roles in ribosomes (for a review, see P PABP GAIT REF.138). Most ribosomal proteins are RNA-binding proteins and RPL13A complex eIF4G some of these proteins can potentially bind to additional RNAs or ORF 5 UTR 7 mG protein complexes outside of the ribosome. For example, many Met ribosomal proteins regulate their own synthesis. In Saccharomyces eIF4E eIF2 cerevisiae, ribosomal protein Rpl30 inhibits splicing of its own eIF3 43S ribosomal transcript139. In human cells, RPS13 also inhibits splicing of its complex 40S owntranscript by binding to its intron140. Ribosomal proteins canalso function in other protein complexes141. Nature Reviews | Molecular Cell Biology The most interesting extraribosomal protein that functions in translational control is RPL13A. RPL13A regulates thetranslation of specific mRNAs as part of a non-ribosomal complex (see the figure). Following interferon- (IFN) treatment, monocytes show a marked increase in transcription of ceruloplasmin mRNA, which is an acute phase inflammatory protein. Ceruloplasmin protein synthesis is terminated after 16hours even though ceruloplasmin mRNA isstill present142. The IFN-activated inhibitor of translation (GAIT) complex was found to bind to an RNA structure in the3untranslated region (UTR) of ceruloplasmin mRNA to inhibit translation. RPL13A is one of the four proteins that constitute the GAIT complex. RPL13A is phosphorylated and released from the ribosome 24hours after IFN stimulation143. Phosphorylated RPL13A binds strongly to cap-bound eukaryotic translation initiation factor eIF4G and blocks the recruitment of the 43S ribosomal complex144. It also has the same repressive effect on a number of other inflammatory response genes that have the same 3 UTR RNA structure, such as many chemokines and chemokine receptors145. Following this paradigm, certain ribosomal proteins that are not essential for the catalytic activity of theribosome may be coopted for additional, specialized functions in translational control acting in complexes that function apart from the ribosome itself. m7G, 7methylguanosine; Met, initiator tRNA; ORF, open reading frame; PABP,poly(A)-binding protein.
Paralogues
Homologous genes that are separated by a duplication event and that have evolved new functions.
ASH1
(Asymmetric synthesis of homothallic switching endonuclease (HO)). A gene encoding a repressor that inhibits the transcription of HO an endonuclease that causes mating-type switching in Saccharomyces cerevisiae. ASH1 mRNA is transported to the bud before translation. In the bud, Ash1 prevents the daughter cell from switching its mating type following celldivision.
Ribosomal protein paralogues. In mammals, although there are ~2,000 sequences encoding ribosomal proteins, most of them are predicted to be pseudogenes12, and most functional ribosomal proteins are encoded by a single gene. By contrast, in yeast, plants and flies many ribosomal proteins are encoded by more than one gene. Saccharomyces cerevisiae, for example, arose from an ancient genome duplication, and 59 of the 78 ribosomal proteins retained two genomic copies13. Remarkably, despite having high sequence homology, the two ribo somal protein gene copies do not always seem to be functionally redundant. Deletions of individual paralogues can give rise to completely unique phenotypes showing differences in cell size, bud size selection 14, sporulation15, susceptibility to virus16,17 and drug sensitivity 18,19, and these phenotypes are as dissimilar as deletions in completely unrelated ribosomal proteins. For example, deletions of at least six ribosomal proteins, but not their paralogues, affect the localization of ASH1 mRNA to the bud tip of the emerging daughter cell, a process that requires tight translational regulation18. It is, however, important to note that distinct phenotypes observed among some ribosomal protein paralogues18 may be confounded by additional suppressor mutations that arise during the propagation of ribosomal proteindeletion strains20,21, and hence a more careful examination is needed to confirm these phenotypes. Despite these caveats, the many examples of unique functional roles for individual ribosomal protein paralogues suggest a differential requirement for individual paralogues in specific molecular and cellularevents. Recently, it has been shown that splicing of yeast ribosomal protein mRNAs can regulate the expression of both ribosomal protein paralogue genes at the transcript level, often in a non-reciprocal manner. For example, the
removal of the intron in the gene encoding ribosomal protein S29A (Rps29a) reduces not only its own expression but also the expression of RPS29B19 (FIG.1a). This is surprising, as a change in the expression of any single copy of the duplicated ribosomal protein gene pair would be expected to increase the expression of the other if it was simply required to maintain a stoichiometric level of expression. On the contrary, in wild-type cells ~70%of all duplicated ribosomal protein genes are asymmetrically expressed19. This suggests that in yeast cells the ratio of paralogues is regulated and that paralogues are not merely equivalent, functional substitutes. It is tempting to speculate that variation in ribosome composition, for example by incorporating different paralogues, may be a mechanism to provide unique regulatory activity to the ribosome. Although lossoffunction phenotypes in distinct ribosomal protein gene copies would support this possibility, little is known to date about the functional roles of distinct ribosomal protein paralogues in translational control. Plants have even more gene copies encoding a single ribosomal protein than yeast. Each ribosomal protein in Arabidopsis thaliana is encoded by two to seven genes22. Rapeseed, Brassica napus, has 966 genes encoding 79ribosomal proteins23. Some of the ribosomal protein gene copies are identical in sequence, but many of the paralogues display sequence variations and are differentially expressed during development 24,25. For example, in A.thaliana, RPS5A is strongly expressed in dividing cells, whereas RPS5B is expressed in cells undergoing differ entiation24 (FIG.1b). The gene expression of ribosomal protein L11B (RPL11B; previously known as RPL16B) is also high in all actively dividing cells, whereas the expression pattern of RPL11A is much more tissue-specific26. Moreover, the two A.thaliana RPL23A paralogues, which
VOLUME 13 | JUNE 2012 | 357
REVIEWS
a S. cerevisiae
Exon Intron Exon RPS29A pre-mRNA RPS29B pre-mRNA RPS29 paralogue pair
b A. thaliana
RPS5A RPS5B RPS5A + RPS5B
Exon
Intron
Exon
Exon Exon
Intron Intron
Exon Exon
RPS7A pre-mRNA RPS7 paralogue RPS7B pair pre-mRNA
Four-cell stage
Eight-cell stage
Globular stage
Transition stage
Ribosomal protein intron deletion: stress sensitivity
d Human
Somatic cells RPS4Y1 RPS4Y2 RPS4Y1 mRNA Y chromosome
c D. melanogaster
Somatic cells Polysome RPL22 Testis cells
RPS4Y1 RPS4Y1
RPL22
Testis and prostate cells RPS4Y1 RPS4Y1 mRNA RPS4Y1 RPS20

P
RPL22-like (37% identity to RPL22)
RPS4Y2 RPS4Y2 mRNA RPS4Y2 RPS4Y1 RPS4Y2
e D. discoideum
P
RPS20 RPS19 RPL2 RPL18 Spores Amoebae RPL18
RPS19 RPL2
f Mouse
mRNA expression 1 Low expression 5 1 2 3 4 Tissues 5 High expression 3 2 4 Ribosomal proteins
Ribosome switch Fruiting body
Aggregation
Mound Slug
Figure 1 | Heterogeneity in the ribosome at the level of ribosomal protein composition and post-translational Nature Reviews | Molecular Cell Biology modifications across different species. a | Introns of genes encoding ribosomal proteins in Saccharomycescerevisiae regulate the expression of both the intron-containing ribosomal protein genes and their paralogues, and this has important outcomes in cells: for example, increased fitness under stress. 70% of ribosomal protein paralogue pairs show differential expression. b | In plants, ribosomal protein paralogues have different functions and different expression patterns. For example, in Arabidopsis thaliana, RPS5A is expressed in rapidly dividing cells early in embryonic development, whereas RPS5B is expressed in cells undergoing differentiation. c | In Drosophila melanogaster, ribosomal protein paralogues show different expression patterns in the adult testes. For example, RPL22 is expressed ubiquitously, but RPL22like protein levels are specifically increased in the testes. Both proteins are incorporated into translationally active ribosomes (called the polysomes). d | In humans, only some ribosomal protein paralogues have been identified; however, notable examples exist. RPS4Y1 is expressed ubiquitously, whereas RPS4Y2 is restricted to the testis and prostate. e | In the life cycle of the social amoebae Dictyostelium discoideum, ribosome switches are characterized by changes in ribosome composition at the level of ribosomal protein expression and post-translational modifications. Forexample, phosphorylation (P) on RPS19 and methylation (M) on RPL2 are lost as D.discoideum aggregates from a single cell amoebae to a multicellular fruiting body, while RPL20 gains phosphorylation marks. RPL18 is exclusively found in ribosomes of developing cells and not in those of the amoebae stage. f | In mice, the mRNA expression pattern of ribosomal proteins varies dramatically among tissues. Thismay potentially translate into unique ribosomal protein compositions of ribosomes in distinct cell types. Part a is modified, with permission, from REF. 19 (2011) Elsevier. Part b is modified, with permission, from REF. 24 (2001) TheCompany of Biologists. Part f is modified, with permission, from REF. 37 (2011) Elsevier.
REVIEWS
possess high sequence homology, are not of equivalent importance for normal plant development 27,28. Deletion of RPL23AA but not of RPL23AB impeded growth and led to morphological abnormalities, revealing specialized functions for specific ribosomal protein paralogues. A genome-wide study of ribosomal proteins in B.napus showed that different paralogues of the same ribosomal protein are not regulated in the same pattern among tissues. Interestingly, more than two-thirds of ribosomal proteins are encoded by paralogues that have opposite expression patterns in embryos and seedlings, suggesting that a different set of ribosome paralogues might be required for early development23. Heterogeneity in ribosomal protein paralogue expression was also found in Drosophila melanogaster. For example, some paralogues show differential expression levels in the adult testes29. RpL22 is transcribed ubiquitously, but RpL22-like is enriched in the testes (FIG.1c). Both genes are essential in the fly, which suggests that they are not functionally redundant. Moreover, both proteins are incorporated into ribosomes as they are detected in polysomes (which are translationally active ribosomes) and 80S fractions but not in pre-ribosomal fractions30. In addition, RpS5b, RpS19a, RpL10Aa and RpL37b show enhanced expression in the testes compared to their paralogues29. Such heterogeneity in ribosomal protein expression in the gonads suggests that the development of germ cells may require tissue-specific variations in the translational machinery. In mammals, most ribosomal proteins are encoded by only a single gene copy. However, some notable exceptions exist. For example, RPS4 is encoded by three genes (namely RPS4X, RPS4Y1 and RPSY2) that are located on the X chromosome and the Ychromosome31,32. In human males, RPS4X and RPS4Y1 are expressed ubiquitously, but RPSY2 expression is restricted to the testis and prostate, suggesting a male-specific role for the ribosomal protein encoded by RPSY2 and the possibility of testisspecific ribosomes31 (FIG.1d). Although the sequences of the three RPS4 proteins are very similar, distinct amino acids in the carboxyl terminus of RPS4Y2 possibly facilitate unique interactions with distinct, potentially testisspecific, ribosomal proteins or extraribosomal factors31. In addition, mice carry paralogues for RPL10, RPL22 and RPL39, which are also differentially expressed 33. Ribosomes from rodent liver, mammary gland and testis were isolated and analysed by two-dimensional (2D) gel electrophoresis and mass spectrometry. Surprisingly, RPL10like and RPL39like were only found in ribosomes from the testis but not in the liver or the mammary gland. Interestingly, RPL39like localized to the nucleoli and was found in 80S and polysome fractions, which is consistent with RPL39like being a ribosomal protein33. By contrast, RPL22like was absent from the testes but is specifically expressed in the liver and mammary gland. Therefore, this suggests that ribosomes expressed in different tissues contain unique ribosomal protein paralogues, and similarly to the findings in D.melanogaster, variations in the expression patterns of ribosomal protein paralogues seem to be evolutionarily conserved and partic ularly eviden t in germcells. Ribosomal protein expression. In addition to the fact that some ribosomal proteins are encoded by paralogue genes, the expression of core ribosomal proteins may not be ubiquitous, at least under certain circumstances. It has previously been suggested that the equimolar production of ribosomal proteins is important for proper ribosome biogenesis and that ribosomal proteins are transcribed at the same rate and have similar lifespans34,35. However, most of these studies focused on only a handful of ribosomal proteins in a single cell type. Recent studies that examined ribosomal proteins more globally at an organismal level have shown that different ribosomal proteins are expressed at distinct levels in unique cell types36,37. In the social amoebae Dictyostelium discoideum, ribosomes are composed of different ribosomal proteins at two different states of its life cycle, the vegetative stage and the spore stage (FIG.1e). Many ribosomal proteins are expressed at different levels in one state compared to the other state, suggesting that these proteins may be developmentally regulated during cell differentiation38,39. In mice, the levels of Rpl38 transcripts exhibit a tissuespecific expression pattern37. For example, during embryonic development, Rpl38 expression is substantially increased in developing somites and in a specific subset of motor neurons within the spinal cord. Strikingly, this expression pattern mirrors, to a large extent, the tissue s that are affected by the loss of function of RPL38 in mutant embryos37. Moreover, Rpl38 is expressed only insome adult tissues40. RPL10 expression in mouse embryos is also highly tissue-specific as this ribosomal protein is enriched in the developing epidermis and limb buds41 and shows a very specific expression pattern in fetal bone42. Notably, RPL10 is one of the few mammalian ribosomal proteins that has a paralogue; however, this cannot solely account for its restricted expression pattern as the paralogue encoding RPL10like is also differentially expressed33. RPL10like has only been found to be expressed in the testes but not other tissues such as theliver. More globally, a large-scale quantitative expressionprofiling screen examined the expression of 72 ribosomal proteins in 14 tissues and cell types in the developing vertebrate embryo (FIG.1f). This study revealed dramatic differences in the expression levels of ribosomal proteins among different tissues and cell types, and revealed coregulated expression patterns among specific groups of ribosomal proteins37. Additional studies also point to the possibility of greater heterogeneity in ribosomal protein expression in adult tissues36. Although most of these studies only examine the mRNA levels of ribosomal proteins, the huge range of variation in expression among tissues suggests that ribosomes may be heterogeneous in distinct cell and tissue types. These findings also raise the possibility that some ribosomes may contain more than one copy of certain, highly expressed ribosomal proteins. This is analogous to the dimerization of the acidic ribosomal proteins in the ribosomes of Escherichia coli, which contain two protein copies of RpL7 and RpL12 per ribosome4345. Importantly, as ribosomal proteins may have extra-ribosomal functions, can be degraded at the protein level46 or are translationally controlled146, further
Polysomes
Or polyribosomes; clusters of two or more ribosomes attached at different sites on the same strand of mRNA. mRNAs bound to polysomes are being actively translated.
Somites
Blocks of mesoderm on either side of the neural tube of a developing vertebrate embryo. Somites will develop into structures including the vertebrae.
REVIEWS
proteomics studies will be key in confirming the contribution of heterogenous ribosomal protein expression in specialized ribosomal activity. Post-translational modifications. In addition to the differences in ribosomal protein expression levels and patterns, post-translational modifications on ribosomal proteins can add another layer of specificity. In many organisms, ribosomal proteins are post-translationally modified. In yeast, at least 16 ribosomal proteins are either methylated or acetylated47,48. In humans, at least 11 large subunit ribosomal proteins and most of the small subunit proteins are also post-translationally modified49,50. In A.thaliana, 23 of the 80 ribosomal protein families contain at least one covalent modi fication48. Modifications of ribosomal proteins seem to be highly regulated. For example, ribosomes from different phases of the D.discoideum lifecycle show specific changes in ribosomal protein phosphorylation and methylation patterns that may facilitate unique translational needs of the cell51 (FIG.1e). Historically, the post-translational modifications on RPS6 are the best characterized. For example, in mammals, RPS6 is phosphorylated on five residues in response to mitogen and growth factor signalling 52. It has long been believed that RPS6 phosphorylation has an important function in the translational control of a subclass of mRNAs that harbour a 5 tract oligo pyrimidine (5 TOP) sequence, and that this level of regulation may imbue the ribosome with greater specificity. However, unexpectedly, a knock-in mouse, in which all five phosphorylated residues of RPS6 were mutated to Ala, did not show a decrease in general or 5 TOP mRNA translation but paradoxically resulted in an increase in protein synthesis rates. Moreover, the invivo physiological role of RPS6 phosphorylation seems to be rather limited as mice with defectiv e RPS6 phosphorylation only display decreased cell size and impaired glucose homeostasis 53. Hence, the physio logical significance of ribosomal protein posttranslationa l modifications needs to be carefully evaluate d in thefuture. A recent study showed that in addition to acetylation, methylation and phosphorylation, many ribosomal proteins can be modified at Ser and Thr residues by the addition of Olinked -dNacetylglucosamine (OGlcNAc). Overexpression of OGlcNAc transferase (OGT), the enzyme involved in OGlcNAc modification, resulted in an increase in the 60S and 80S peaks in the polysome profile, suggesting that OGlcNAc modificatio n might be important in ribosome biogenesis54. The same ribosomal proteins exhibit different dynamics of specific post-translational modifications, suggesting an even more complex regulation of the ribosome activity 54. In addition to being phosphorylated and O-GlcNAcylated, ribosomal proteins are regulated by ubiquitylation. For example, Rpl28 in S.cerevisiae is heavily ubiquitylated during S phase of the cell cycle, whereas ubiquitylation is reduced in G1 phase. Rpl28 is located at the peptidyl transferase centre, and its ubiquitylation has a stimulatory effect
on translation. Ribosomes that carry polyubiquitylated Rpl28 are able to translate a reporter gene at a faster rate than ribosomes with monoubiquitylated Rpl28 (REF.55). In addition, mass spectrometry studies have shown that ribosomal proteins can be methylated under different conditions50. Together, these studies identify unique post-translational modifications on specific ribosomal proteins that may affect the function of the ribosome. Ribosome-associated factors. Proteomics studies of S.cerevisiae ribosomes identified 77 ribosome-associated proteins, which suggests that apart from ribosomal protein s, many additional factors may also modulate ribosome activity. Cells lacking some of these proteins show defects in the rate and fidelity of protein synthesis as well as ribosome biogenesis56. Interestingly, ribosomeassociated proteins may also regulate the translation of specific subsets of mRNAs. For example, in D.melanogaster, the pro-apoptotic protein Reaper binds directly and specifically to the 40S subunit and disrupts AUG recognition by the scanning 48S complex, thereby inhibiting cap-dependent translation (FIG.2a). Interestingly, Reaper may facilitate the selective translation of specific pro-apoptotic mRNAs that harbour internal ribosome entry site (IRES) elements (see below), as these mRNAs can bypass scanning, thereby augmenting the apoptotic programme57. Another ribosome-associated protein is glycogen synthase 1 (GYS1). In HeLa cells, GYS1 associates with actively translating ribosomes. Knockdown of GYS1 causes a decrease in polysomes and a change in the subsets of mRNAs in polysomes58. It is therefore clear that the translation of mRNAs may be enhanced or inhibited by distinct ribosome-associated factors, adding a further level of complexity to the translation machinery. RACK1, the receptor for activated protein kinase C (PKC), is a scaffold protein that associates with the 40S subunit in a 1:1 ratio (reviewed in REFS59,60). Many signalling molecules interact with RACK1 on the ribosome and affect protein synthesis. For example, activated PKCII binds to ribosome-bound RACK1 and phosphorylates the translation initiation factor eIF6, releasing it from the 60S subunit and allowing subunit joining 61. Recent work has also suggested that RACK1 can recruit the microRNA (miRNA)-induced silencing complex (miRISC) to the ribosome, where this complex facilitates miRNA-mediated gene repression62. Other binding partners of RACK1 include integrin receptors60, which can potentially localize ribosomes to the cell membrane, and the yeast RNA-binding protein Scp160p63,64 (FIG.2b). Scp160p associates with specific mRNAs, and loss of Scp160p reduces the association of those mRNAs with polysomes65. As Scp160p interacts with the ribosome through RACK1 (REFS63,64), these results suggest that Scp160p brings specific mRNAs to the ribosome to enhance their translation. Yeast contain many ribosome-associated proteins 56, and it is intriguing to hypothesize that additional inter mediaries could associate with the ribosome and may contribute towards transcript-specific translational regulation.
REVIEWS
a
Apoptotic cell
Integrin
Reaper Inhibition of cap-dependent translation Reaper Translation of IRES-contaning mRNA
RACK1 Localized translation?
RACK1
miRISC miRNA miRNA-mediated translational inhibition
NH2
RACK1 Scp160p RNA binding protein
Selective translation of specic mRNAs recruited by RNA-binding proteins
Figure 2 | Ribosome-associated factors affect ribosome function. a | In Drosophila melanogaster, Reaper protein associates with the 40S small ribosomal unit, disrupts the cap-scanning mechanism of translational initiation and Nature Reviews | Molecular Cell Biology promotes the translation of mRNAs containing internal ribosome entry site (IRES) elements that may be required for cell death. b | Ribosome-bound RACK1 can potentially localize ribosomes to the cell membrane by interacting with transmembrane receptors. RACK1 can recruit the microRNA (miRNA)-induced silencing complex (miRISC) to the ribosome to facilitate miRNA-mediated repression. RACK1 also binds to RNA-binding proteins such as yeast Scp160p, and this interaction may facilitate the recruitment of specific mRNAs to the ribosome.
Mammalian target of rapamycin complex 2

(mTORC2). A protein kinase complex that includes mTOR, RICTOR and other proteins. mTORC2 regulates cell growth, metabolism and survival in response to environmental cues such as nutrients and growth factors.
Ribosome-associated factors can also function in processes that are unrelated to translation. For example, mammalian target of rapamycin complex 2 (mTORC2) is a highly conserved protein kinase complex that integrates cell growth and metabolism in response to nutrients and growth factors66. Following insulin signalling mTORC2 binds to ribosomes, and this inter action activates mTORC2 independently of translation 67. Moreover, it seems that ribosome-bound mTORC2 may phosphorylate its substrate AKT during or immediately after AKT mRNA translation68. Importantly, these findings suggest that ribosome-associated factors such as mTORC2 can also operate to control specific molecular events independently from a direct role in translational control. Heterogeneity in rRNA. Heterogeneity in rRNA may also contribute to specialized ribosomal activity. The malaria-causing Plasmodium parasites carry two classes of ribosomal DNA genes, which allow the use of different types of ribosomes during their life cycle. One form is predominantly found when the parasite is inthe sporozoite form (type S) in the mosquito, whereas theother form is found when it is in the gametocyte form (type A) in the bloodstream of the host after infection69 (FIG.3). The two types of ribosomes seem to possess different functions, as substitution of the 25SrRNA of S.cerevisiae with the rRNA of type A P.falciparum produces no phenotype, whereas substitution with therRNA of type S is lethal70. The differences betweenthe two classes of rRNA in Plasmodiumfalciparum are mainly found in the variable regions of the rRNA, and only few differences are evident in areas that are conserved among all eukaryotes71.
Sporozoite
The cellular form of Plasmodium parasites when they infect a new host.
Expansion segments
A region of ribosomal RNA (rRNA) that has dramatically increased in length from prokaryotes to eukaryotes during evolution.
Small nucleolar RNAs

(snoRNAs). Small RNA molecules that function in ribosome biogenesis in the nucleolus by guiding the assembly of macromolecular complexes on the target RNAto allow site-specific modifications or processing reactions to occur.
During evolution, from prokaryotes to eukaryotes, the ribosome has increased in size and the rRNA has changed substantially. For example, D.melanogaster has a 30 nucleotide 2S rRNA in addition to the 18S, 5.8S, 28S and 5S rRNAs72. The most drastic variations in the rRNA are found in the variable regions, which are also known as expansion segments. Expansion segments are generally located on the surface of the ribosome, are very flexible and may have more subtle roles in fine-tuning translational control. For example, the largest expansion segment, ES7L (also known as D2), has expanded from 53 nucleotides in E.coli to a dramatic 689 nucleotides in mice73. In yeast, ES7L (which consists of 216 nucleotides) was shown to interact with multiple eukaryotic specific ribosomal proteins such as P0 in vitro, and this inter action might result in the recruitment of G proteins tothe ribosome74. It will therefore be important to delineate the functional roles of expansion segments and whether changes in these sequences influence ribosome function and translational control specificity. Ribosomal RNA is heavily modified by methylation on the 2-hydroxyl group of ribose, conversion of uridine to pseudouridine and methylation of bases. Most of these modifications occur in important functional areas of the ribosome75. The 2-hydroxyl methylation and conversion of uridine to pseudouridine are facilitated by C/Dbox and H/ACA box small nucleolar RNAs (snoRNAs ), respectively. There is evidence for the heterogeneity in the expression levels of both C/D box and H/ ACA box snoRNAs in different tissues76. The expression of some C/D box snoRNAs has also been shown to oscillate in a circadian manner 77. However, a careful examination of how tissue-specific differences in snoRNA expression directly affect rRNA modifications has not been
VOLUME 13 | JUNE 2012 | 361
REVIEWS
Gametocytes Gametocyte Zygote Sporozoites Oocyst Fertilization and maturation Liver Sporozoites Merozoites Asexual blood cycle
ribosome to interface with specific subsets of mRNAs to enhance translation79. However, the precise functions of only a handful of ribosomal proteins have been identified to date. Here we discuss the nature of regulatory elements in mRNAs that may imbue regulation specific to the ribosome in the recognition and the translation of specific subsets ofmRNAs. IRES elements. In eukaryotic translational initiation, the 40S ribosomal subunit is recruited to the mRNA by translation initiation factors2. The 40S ribosome complex, together with associated factors, scans along the 5 UTR for the initiation codon. Translation initiation factors are released at the initiation codon, and the 60Sribosomal subunit joins to form a translationally active 80S ribosome. Most eukaryotic mRNAs are translated via this cap-dependent mechanism, but it is approximated that more than 10% of cellular mRNAs rely on cap-independent mechanisms of translational initiation involving IRES elements80. The IRES is an RNA structured element positioned at the 5 UTR of specific mRNAs that can recruit the ribosome directly to the initiation region of mRNAs with a reduced requirement for canonical initiation factors2. IRES elements were first discovered in RNA viruses, which produce uncapped RNAs and require these elements for efficient translation of viral mRNAs81,82. Later, many additional cellular mRNAs were discovered to contain IRES elements in their 5 UTR. This class of mRNAs is often thought to encompass stress response genes, such as hypoxia-inducible factor 1A (HIF1A), TP53 and X-linked inhibitor of apoptosis (XIAP), and the IRES elements within their 5 UTRs are thought to be active during cellular stress conditions when cap-dependent translation is inhibited. Modifications in rRNAs are important for IRESmediated translation. In particular, alterations in rRNA pseudouridylation cause a specific defect in the translation of some IRES-containing mRNAs. In mice that are hypomorphic for dyskerin (which is the enzyme that converts rRNA uridine to pseudouridine) general cap-dependent translation is not impaired. However, translation of IRES-containing mRNAs including the tumour suppressors p27 and p53 is perturbed8386. As a result, these mice have a higher propensity to develop cancer. Importantly, ribosomes that lack pseudouridine modifications show a direct impairment in binding to IRES elements87. Emerging evidence suggests that ribosomal proteins themselves may function to recognize specific IRES elements. For example, RPS25 is specifically required for IRES-dependent translation of viral mRNAs88. In yeast harbouring a deletion in RPS25, cap-dependent translational initiation and ribosome fidelity are not affected. However, the translation of two distinct viral IRES elements from the hepatitis C virus (HCV) and cricket paralysis virus (CrPV) is defective (FIG. 4a). Moreover, Rps25deficient 40S subunits are unable to bind to the CrPV IRES element. Structural studies have placed Rps25 in the head of the 40S subunit, where it can interact with two stemloops in the CrPV IRES89.
Type S rRNA in sporozoites in mosquito
Type A rRNA in gametocytes in the bloodstream
Chimeric yeast ribosome with type S rRNA Non-functional
rRNA variable region sequence heterogeneity
Chimeric yeast ribosome with type A rRNA Functional
Figure 3 | rRNA heterogeneity. Plasmodium falciparum expresses differentCell forms of Nature Reviews | Molecular Biology ribosomal RNAs (rRNAs) when it is in sporozoite form (type S) in the mosquito and when it is in the asexual form (type A) in the bloodstream. The two forms of rRNA differ mainly in the variable regions . Chimeric yeast ribosomes in which part of the 25S rRNA was replaced with type S P. falciparum rRNA are non-functional, whereas replacement with type A P. falciparum rRNA does not produce any defects, suggesting the two types of rRNA are functionally distinct.
carried out. Knockdown studies using morpholino oligomers showed that the deletion of different snoRNAs in zebrafish causes specific developmental phenotypes. For example, U26 snoRNA-deficient embryos have a decreased body size and specific defects in brain and eye development 78. These results suggest that snoRNAs and snoRNA-mediated modifications on rRNA have very specific effects in an organism.
Transcript-specific control of ribosome function Several key questions need to be addressed in order to determine how changes in ribosome composition, structure and/or activity affect translational control. A central unresolved question is the nature of specific cisregulatory elements within mRNAs that may interface with specialized ribosomes to provide regulatory control and whether additional trans-acting factors such as RNA-binding proteins have a role in this process. In addition, even core components of the ribosome that remain constitutively associated with the translation machinery may exert regulative effects by interfacing with specific regulatory elements that are present on subsets of mRNAs. Therefore, a greater understanding of the cis-acting elements within mRNAs that promote translational control is an important avenue to explore. An intriguing hypothesis is that ribosomal proteins could themselves act as specificity filters, allowing the
362 | JUNE 2012 | VOLUME 13
REVIEWS
The function of RPS25 in IRES-mediated translation of viral mRNAs is conserved in mammalian cells88. However, it remains unknown whether RPS25 can regulate the translation of mRNAs that contain specific cellular IRES elements. Consistent with this idea, RPS19 and RPL11, two ribosomal proteins that are mutated in patients with DiamondBlackfan anaemia (DBA), regulate IRES-dependent translation of at least two mRNAs that are important for erythroid differentiation, namely BCL2associated athanogene (BAG1) and cold shock domain containing protein E1 (CSDE1)90. In addition, mass spectrometry studies have shown that ribosomes bound to the HCV IRES have a different methylation pattern from ribosomes bound to host mRNAs, indicating a role of methylated ribosomal proteins in IRESmediated translational control 50. These studies may reflect an emerging theme: that ribosomal proteins may confer more specialized regulation of translation by interfacing with specific IRES elements in subsets ofmRNAs. Upstream open reading frames. Almost half of all human and mouse transcripts contain at least one upstream open reading frame (uORF)91. Moreover, ribosomal profiling has recently uncovered many near-cognate uORFs that do not start with an AUG codon and, therefore, have been difficult to predict computationally92,93. uORFs can regulate the translation of the downstream main ORF. In cap-dependent translation, the ribosome scans along the 5 UTR to start translation at the first start codon. As it is difficult to reinitiate translation after translating an uORF, most uORFs repress to the translation of the downstream main ORF. However, ribosome occupancy on some uORFs with near-cognate, non-AUG start codons correlates positively with the translation of the main ORF. For example, many mRNAs needed for sporulation in yeast contain near-cognate uORFs. Ribosome occupancy at such uORFs is higher in meiotic cells than in vegetative cells, suggesting that there is a shift in the translation machinery during meiosis to translate uORFs94. This transition can change the types of mRNAs that are translated, promoting the translation of mRNAswith non-AUG uORFs over those with cognateones. In plants, RPL24 is implicated in translational reinitiation of polycistronic mRNAs and mRNAs containing uORFs (FIG.4b). A.thaliana carrying a mutation in RPL24B have developmental defects in vascular patterning and gynoecium structure owing to a reduction in translation of auxin response genes. Interestingly, many genes of the auxin response factor E class contain uORFs that inhibit the translation of the main ORF. The removal of the uORFs in these genes results in partial suppression the RPL24B phenotype95. RPL24 is located at the interface between 40S and 60S subunits and may help in subunit joining during translational re initiation74. The cauliflower mosaic virus (CaMV) takes advantage of this RPL24 function to promote infection. RPL24 binds to the CaMV transactivator (TAV), and this binding is essential for the translation of the second ORF from the polycistronic CaMV RNA. TAV also binds eIF3, which is the initiation factor that stimulates the binding of
a
CrPV IRES
RPS25
ORF
m7G
uORF
uORF
uORF 5 UTR
ORF RPL24
Figure 4 | Cis-regulatory elements within mRNAs that Nature Reviews | Molecular Cell interface with ribosomes or ribosomal proteins toBiology confer transcript-specific regulation of gene expression. a | Internal ribosome entry site (IRES) elements: the ribosomal protein RPS25 is necessary for the recruitment of the small ribosomal subunit 40S to the cricket paralysis virus (CrPV) intergenic region IRES to initiate translational initiation by a cap-independent mechanism in yeast and mammals. Ribosomal RNA (rRNA) modifications (such as pseudouridylation ()) are important for ribosome binding to IRES elements. b|Upstream open reading frames (uORFs): RPL24 promotes the reinitiation of translation following an uORF, thereby allowing the translation of the downstream main ORF. m7G,7methylguanosine, UTR,untranslated region.
the ternary complex to the 40S. TAV can remain associated with the 40S subunit after termination at the first ORF and potentially help in recruiting eIF3 and RPL24containing 60S subunits for reinitiation96. It will be interesting to see whether RPL24 also has a role in overcoming uORF-mediated repression in mammalian cells as a mechanism to promote translational control of specific subsets of uORF-containing mRNAs. Moreover, it will be important to determine whether additional ribosomal proteins modulate uORF-mediated regulation of translational control.
Ribosome specificity in cell biology As translation rates are a much better predictor of protein levels than mRNA levels, translational control is an important mechanism to regulate the expression of many genes97. The possibility of greater specificity in translation through a more regulatory function of the ribosome opens an important avenue of exploration into the cell ular and developmental processes that are regulated by this new level of control. This is a very rapidly emerging field, and although much remains to be addressed, clear examples of specialized ribosome functions are becoming evident across all species.
Ribosome specificity in bacteria. In bacteria, a new paradigm is emerging for heterogeneity in ribosomes as a mechanism for stress adaptation. Remarkably, the formation of stress-ribosomes that contain rRNAs with sequence-specific differences allows for the selective
VOLUME 13 | JUNE 2012 | 363
Upstream open reading frame

(uORF). An uORF is defined by a start codon and an inframe stop codon in the 5 untranslated region of anmRNA.
REVIEWS
a E.coli
MazF AUG MazF Anti-Shine b D. melanogaster germ plasm Dalgarno Germ plasm + AUG Cytoplasmic ribosome AUG Mitochondrial ribosomes may help the translation of cytoplasmic germ cellspecic mRNAs Mitochondrial ribosomes Mitochondrion D. melanogaster embryo Pole cells
Stress Endonuclease cleavage by MazF MazF
Selective translation of leaderless mRNAs
c Vertebrate embryo
RPL38 expression Neural tube
Vertebra
d Neuron
Ribosomal protein mRNAs Ribosomal protein mRNAs localized in dendrites may form dendrite-specic ribosomes
Lateral motor columns Hox mRNAs
Somite
Wild type
RPL38 mutant
Pea3 positive motor neurons
Wild type
RPL38 mutant
Wild type
RPL38 mutant Localized translation at synapses
RPL38 mediates transcript-specic translational control
Figure 5 | Ribosome specificity in cell and developmental biology. a | In Escherichiacoli, MazF cleaves the 3 end of Nature Reviews | Molecular Cell Biology the 16S ribosomal RNA (rRNA) and releases the anti-ShineDalgarno sequence, thereby generating specialized ribosomes that translate leaderless mRNAs that are involved in the stress response. b | Mitochondrial ribosomes (green) in the germ plasm (yellow area) of Drosophila melanogaster embryos are found outside of mitochondria and may translate certain cytoplasmic germ cell-specific mRNAs. c | Gene expression of Rpl38 (pink) is highly increased in somites and the neural tube of developing mouse embryos and is responsible for translating specific homeobox (Hox) mRNAs that are necessary for axial skeletal patterning and the specification of PEA3-expressing motor neurons. d | Localized translation takes placein dendrites, far away from the cell body. mRNAs for ribosomal proteins are found in dendrites and can potentially form dendrite-specific ribosomes.
ShineDalgarno sequence
A ribosomal binding site of approximately eight nucleotides in the mRNA of bacteria, located upstream of the initiation codon. Helps to recruit the small ribosomal subunit to the mRNA to initiate protein synthesis.
translation of specific mRNAs (FIG.5a). In E. coli, the sequence-specific endonuclease MazF, which preferentially cleaves single-stranded mRNAs at ACA sequences, is activated following stress. Although MazF induction blocks the synthesis of most proteins, the translation of ~10% of all transcripts remains active, and some of these mRNAs encode proteins that promote cell death. Strikingly, MazF is able to selectively promote trans lation of these transcripts by cleaving ACA sites at, or closely upstream of, the AUG start codon to gen erate leaderless mRNAs. MazF also changes the type of ribosomes that are formed to translate these leaderless mRNAs by cleaving 16S rRNAs near the 3 end, thereby releasing theanti-ShineDalgarno sequence. In bacteri a, translational initiation depends heavily on base pairing between the ShineDalgarno sequence in the mRNA and the anti-ShineDalgarno sequence in the 16S rRNA98. The removal of the anti-ShineDalgarno sequence by MazF renders a subpopulation of specialized ribosomes that are less efficient in translating most mRNAs but improves translation of leaderless mRNAs that no longer possess the ShineDalgarno sequence99.
The role of ribosome heterogeneity in plants. In plants, mutations in ribosomal proteins result in various phenotypes. Loss-of-function mutations in a number of ribo somal proteins are embryonic lethal, whereas others have more varied patterning and developmental defects100,101. For example, mutations in RPL10A, RPL9 and RPL5 give a piggyback phenotype, in which mutants have pointed leaves and pronounced marginal serrations. These genes are crucial in dorsalventral patterning in leaves, possibly through translational regulation of dorsal genes of the class III homeodomain-Leu zipper (HDZIPIII) family and ventral genes of the KANADI family 102. A.thaliana with mutantions in RPL24B have developmental defects in vascular patterning and gynoecium structure owing to a reduction in translation of auxin response genes95. As discussed, the effect of RPL24 may be specifically at the level of translation reinitiation of mRNAs that harbour uORFs (see above). Although a clear mechanism for translational regulation by these ribosomal proteins is not fully understood, it is clear that ribosomal proteins have an unexpected and important role in regulating specific developmental processes and pattern formation100.
REVIEWS
Ribosome specificity in invertebrate embryonic develop ment. In D.melanogaster, the loss of a copy of several ribosomal proteins leads to a Minute phenotype. Minute flies are smaller than their wild-type counterparts, have delayed larval development and short, thin bristles29. This phenotype has been attributed to a general decrease in protein synthesis. However, some ribosomal protein mutants show additional patterning defects. For example, some Minute flies have a reduction in wing size, whereas at least two additional ribosomal protein mutants display an increase in wing size and wing vein patterning defects103,104. This suggests that some ribosomal proteins are specifically required for embryonic development and patterning. Germline formation in D.melanogaster may require specific ribosomes that are only produced in the mitochondria (FIG.5b). Mitochondrial ribosomes are typically smaller than cytoplasmic ribosomes, resemble prokaryotic ribosomes and are required to translate the mitochondrial genome. Strikingly, however, these ribosomes can be localized outside of mitochondria in polar granules, which are large ribonucleoprotein complexes found at the posterior end of D.melanogaster embryos and are required for germ cell formation105,106. Polysomes on polar granules contain both mitochondrial and cytoplasmic ribosomes. Blocking mitochondrial ribosomes with the prokaryotic translation inhibitors kasugamycin and chloramphenicol, both of which do not affect cytoplasmic ribosomes, causes a decrease in the expression levels of the germ cellspecific protein GCL in polar granules and a defect in germ cell formation. The translation of another germ cell mRNA, namely Nanos (Nos), is not affected by these drugs, suggesting that mitochondrial ribosomes may only be responsible for the proper translation of a subset of mRNAs 107. Importantly, further studies on the codon usage of the GCL mRNA are urgently required to determine whether mitochondrial ribosomes are indeed participating in this level of translational control. Although it is not entirely understood why mitochondrial ribosomes that are located in the cytoplasm may possess unique properties that endow them with the ability to translate specific cytoplasmic mRNAs, these studies suggest a specificity in translational regulation governed at the level of heterogeneous ribosomes. Ribosome specificity in vertebrate embryonic development. Ribosomal proteins are important regulators of vertebrate embryonic development. For example, morpholino knockdown of ribosomal proteins in zebrafish produces phenotypes that can be both unique and specific, resulting in abnormalities in the development of the brain, body trunk, eye and ear 108. How can mutations in ribosomal proteins cause tissue-specific effects? One important paradigm that provided insight into the role of ribosomal proteins in controlling specialized ribosome activity stems from the analysis of RPL38 function, as the results linked this component ofthe ribosome to the formation of the mammalian body plan (FIG.5c). In a forward genetic screen, Rpl38
was identified as the gene that is deleted in tail short (Ts) mice. These mice have highly specific defects in the formation of the axial skeleton37. These defects include homeotic transformations and a change in the identity of one skeletal element from another, which suggests an unexpected regulatory role for ribosomes during tissue patterning. Homeobox (Hox) genes have been shown to be key regulators of vertebrae segment identity, and the loss of function of specific Hox genes causes homeotic transformations109,110. Ts/+ mouse embryos also show a defect in the specification of PEA3positive motor neurons, which also requires HOX-directed patterning. Strikingly, Ts/+ embryos show no defects in general protein synthesis, however, the translation of 8 of the 39 Hox mRNAs is specifically reduced. At themolecular level, RPL38 regulates the formation ofthe 80S complex during translation initiation on these mRNAs, while being part of the ribosome. The effects of RPL38 are highly specific as other ribosomal protein deficiencies in mice do not give rise to phenotypes similar to the phenotypes observed in RPL38deficient mice, even when the general rate of protein synthesis is decreased. Interestingly, Ts/+ phenotypes are inherited in a dominant form, as haploinsufficiency in RPL38 expression is sufficient for patterning defects to manifest. This suggests that RPL38 is rate-limiting for the translation of selective transcripts such as HOX mRNAs. Moreover, this is consistent with the fact that HOX mutants also display phenotypes in a heterozygous condition, thereby revealing that precise levels of HOX proteins are necessary for establishing the mammalian body plan. It will be important to determine whether RPL38 recognizes unique sequences or structures within subsets of mRNAs either directly or through interactions with other RNA-binding proteins. The tissue specific phenotypes associated with RPL38 loss of function are also, at least in part, explained by the observation that Rpl38 expression is highly regulated during embryonic development. A striking enrichment of Rpl38 transcripts is evident in specific regions of the embryo such as developing somites and within the neural tube. These regions mirror where the tissue specific phenotypes associated with loss of RPL38 expression are observed37. These findings suggest that the increased expression of specific ribo somal proteins may produce heterogeneous ribosomes in distinct cell and tissue types with unique specificities in translating specific classes ofmRNAs. In addition to Ts/+ mice, the few mouse models with mutated ribosomal proteins that have been generated or characterized to date display specific phenotypes. For example, mice carrying a dominant mutation in Rpl24 display a white ventral midline spot, white hind feet and a kinked tail. These mice also have reduced numbers of retinal ganglion cells and extra digits in the limbs 111. On the contrary, Rpl29 / mice are smaller in size and possess specific defects in osteogenesis associated with increased bone fragility 112. These phenotypes suggest that additional ribosomal proteins may have more specialized roles during embryonic development.
VOLUME 13 | JUNE 2012 | 365
REVIEWS
Linking ribosome specificity to disease. Loss of some ribosomal proteins has been associated with certain morphological abnormalities, and these phenotypes may rely on the activation of a p53dependent stress response. A subset of ribosomal proteins are able to interact with MDM2, an inhibitor of p53, and thereby affect p53 activity 113. Morpholino knockdown of rpl11, rps7 and rps19 in zebrafish causes an upregulation of p53, and certain phenotypes can be partially rescued by knockdown of tp53 (REFS114116). Importantly, increased p53 protein levels cannot fully explain the phenotypes observed in ribosomal protein mutants. For example, the tissuespecific defects associated with Rpl38 loss of function were not reversed following the reduction of p53 protein levels37. Moreover, the mouse embryos carrying Rps6+/ and p53/ mutations have tissue-specific phenotypes, including defects in placenta and liver development 117. In humans, accumulating evidence links mutations in the genes of several ribosomal proteins, including RPS19, RPS24, RPS17, RPL5, RPL11 and RPL35A, to DiamondBlackfan anaemia (DBA)118. Notably, patients with DBA carrying ribosomal protein mutations also exhibit tissue-specific defects including limb defects, cleft palate, abnormalities in heart development, growth failure and a predisposition for cancer. Mutations in different ribosomal proteins also give rise to different types of birth defects. For instance, most patients with RPL5 mutations have a cleft palate, whereas most individual s with RPL11 mutations do not have any craniofacial defects119. In addition, mutations in RPL21 have been identified in hereditary hypotrichosis simplex (HHS), a disorder that is characterized by progressive hair loss early in childhood120. Together, patients with genetic disorders that are linked to mutations in ribosomal proteins show remarkably specific phenotypes, which suggests that ribosomal proteins have unique functions in different tissues. Intracellular localization of specialized ribosomes. A clue for the possible roles of specialized ribosomal protein function in specific cell types comes from research in neurons. Neurons are extremely polarized cells in which axons and dendrites perform very different functions from the cell body. Local protein synthesis in neuronal dendrites is important in remodelling synaptic connections and has a role in learning and memory 121 (FIG.5d). Long-term potentiation (LTP), a cellular measure of memory formation, can be induced by localized translation in isolated hippocampal dendrites, which have been removed from the cell bodies122. Polysomes are found in dendrites, and the number of polysomes in dendritic spines increases after the initiation of LTP123. Thereby, mRNAs are translated in an extremely localized manner, although the mechanism for this regulation is poorly understood. Strikingly, transcripts for specific subsets of ribosomal proteins are enriched up to tenfold at these sites compared to the cell body 124. As most ribosomal proteins are assembled onto the ribosome in the nucleus, it is surprising to find pools of specific ribosomal protein transcripts in neuronal processes. Additional components of the translational machinery have also been found in neuronal processes and are transported as part of large ribonucleoprotein particles that contain mRNA, ribosomes and other RNA binding proteins125,126. It has been suggested that ribosomal subunits are stored as inactive forms in RNA granules and are activated by the addition of locally synthesized ribosomal proteins126. Thus, ribosomes at these neuronal processes may be distinct from ribosomes in the cell body owing to the presence of newly, locally synthesized ribosomal proteins that may influence the translation of mRNAs that are localized in the dendrites. In support of this hypothesis, recent studie s have shown that the transmembrane receptor DCC, which participates in axon guidance, associates with specific ribosomal proteins and spatially regulates translation in neuronal axons and dendrites 127. Interestingly, some mRNAs present in dendrites are known to contain IRES elements128. By extension, similar localized mechanisms for translational control may provide a unique regulation in multiple cellular contexts, including translational regulation in polarized cells, at the leading edge of migrating cells or during specific steps in cell differentiation.
Cleft palate
A craniofacial abnormality that results from a failure to fuse the left and right palatal shelves at the midline during embryogenesis. It can be caused by several environmental and genetic factors, including defects in sonic hedgehog signalling.
Long-term potentiation
(LTP). A long-lasting increase in the size of the postsynaptic response to synaptic transmissions. LTP is thought to be a key mechanism for learning and long-term memory formation in the brain.
Conclusions and future perspectives Several parallels can be drawn between what has been termed a DNA code and a possible ribosome code. Like DNA, rRNA is extensively modified. Histones, once considered as boring housekeeping proteins, are now clearly recognized as active participants in chromatin remodelling and transcriptional control through exquisite post-translational modifications identified within histone tails. Likewise, the view of ribosomal proteins as only carrying out rote-like functions is undergoing a paradigm shift. Post-translational modifications and differences in ribosomal protein composition may add greater regulatory specificity to the RNA-based translation machinery, as do histones to DNA. However, several outstanding questions need to be addressed before a ribosome code for translational control could be accepted. First, as discussed in this Review, many examples now exist across all species suggesting that great hetero geneity characterizes the ribosome, both at the level of rRNA and ribosomal proteins. However, despite these examples, this hypothesis has not been explored in a systematic manner. Most ribosomes are purified from tissues that contain an amalgamation of multiple cell types, and any differences in ribosome composition in a particular cell may be lost in such an ensemble. Studies in cell lines have the advantage of being able to identify heterogeneity in a single cell type; however, this approach lacks the dynamics of an invivo system in which more variations may be present.A systems approach is necessary to examine the extent of variation in ribosomes more globally across different cell types, perhaps during early stages of organogenesis or during specific steps in cell differentiation. It is also tempting to speculate that a heterogeneous pool of ribosomes may already exist in a single cell, for example in germ cells and neurons. If so, how is the stoichiometry and the
REVIEWS
formation of such heterogeneous pools of ribosomes regulated? What regulates the intracellular localization of heterologous ribosomes in a single cell? Second, although examples of heterogeneity are evident, little is known on how heterogeneity affects translational control. For example, to date classes of mRNAs that may respond to specialized ribosomes remain elusive. The discovery of a specific requirement for RPL38containing ribosomes in facilitating translational regulation of subsets of HOX mRNAs hints at exciting possible future work into this problem. Third, virtually nothing is known about the upstream regulatory signals that might produce ribosome heterogeneity in the first place. The realization that components of the ribosome may be more highly regulated opens a new avenue of exploration in identifying upstream signalling pathways and trans-acting factors that control their expression. Lastly, it will be crucial to identify cis-acting translational regulons within mRNAs that interface with specialized ribosomes to confer translational specificity. As discussed, examples of these important elements may include IRES elements and uORFs. Moreover, it is equally important to determine whether even constitutive components of the ribosome that show little variation may also confer more specialized activity by virtue of their ability to interact with such specific translation regulons. Emerging evidence suggests that there is a great disconnect between the expression of the genome at the level of the transcriptome versus the proteome, thereby leaving ample room for regulation at the level of translational control97. Thus, variations in gene expression at the level of translational control have tremendous importance to cell biology. Therefore, we hypothesize that it will be important to conceptualize translational control in the same light as transcriptional control as a process in which enhancers or attenuators may fine-tune protein abundance that culminates in unique readout s with important biological significance. Ultimately, a key question is how ribosome-mediated gene regulation may shape our view of organismal biolog y. It has long been known that a small handful of signalling molecules (such as sonic hedgehog (SHH), fibroblast growth factors (FGFs), WNT ligands and bone morphogenetic proteins (BMPs)) act many times during embryonic development and in distinct locations to produce entirely different tissues or structures (for example, limbs versus motor neurons). Transcriptional profiling has not provided a complete explanation for these different mechanisms of action. The hypothesis that developmentally regulated signals directly converge on ribosome specificity opens an entirely new avenue of exploration into whether rapid changes in the production of proteins from pre-existing pools of mRNAs may instead influence fundamental aspects of cell and embryonic development.
1. 2.
3. 4. 5. 6. 7. 8. 9. 10.
11. 12. 13.
14. 15.
Frank, J. The ribosome a macromolecular machine par excellence. Chem. Biol. 7, R133R141 (2000). Sonenberg, N. & Hinnebusch, A.G. Regulation of translation initiation in eukaryotes: mechanisms and biological targets. Cell 136, 731745 (2009). Reviews the mechanism of translation initiation and its regulation. Alberts, B. The cell as a collection of protein machines: preparing the next generation of molecular biologists. Cell 92, 291294 (1998). Zaher, H.S. & Green, R. Fidelity at the molecular level: lessons from protein synthesis. Cell 136, 746762 (2009). Warner, J.R. The economics of ribosome biosynthesis in yeast. Trends Biochem. Sci. 24, 437440 (1999). Crick, F.H.C. The origin of the genetic code. J.Mol. Biol. 38, 367379 (1968). Nissen, P. The structural basis of ribosome activity in peptide bond synthesis. Science 289, 920930 (2000). Noller, H.F., Hoang, L. & Fredrick, K. The 30S ribosomal P site: a function of 16S rRNA. FEBS Lett. 579, 855858 (2005). Noller, H.F., Hoffarth, V. & Zimniak, L. Unusual resistance of peptidyl extraction transferase to protein procedures. Science 256, 14161419 (1992). Held, W.A., Mizushima, S. & Nomura, M. Reconstitution of Escherichia coli 30S ribosomal subunits from purified molecular components. J.Biol. Chem. 245, 57205730 (1973). Rohl, R. & Nierhaus, K.H. Assembly map of the large subunit (50S) of Escherichia coli ribosomes. Proc. Natl Acad. Sci. USA 79, 729733 (1982). Balasubramanian, S. etal. Comparative analysis of processed ribosomal protein pseudogenes in four mammalian genomes. Genome Biol. 10, R2 (2009). Kellis, M., Birren, B.W. & Lander, E.S. Proof and evolutionary analysis of ancient genome duplication in the yeast Saccharomyces cerevisiae. Nature 428, 617624 (2004). Ni, L. & Snyder, M. A genomic study of the bipolar bud site selection pattern in Saccharomyces cerevisiae. Mol. Biol. Cell 12, 21472170 (2001). Enyenihi, A.H. & Saunders, W.S. Large-scale functional genomic analysis of sporulation and meiosis
16.
17.
18.
19.
20. 21. 22. 23.
24.
25.
in Saccharomyces cerevisiae. Genetics 54, 4754 (2003). Ohtake, Y. & Wickner, R.B. Yeast virus propagation depends critically on free 60S ribsomal subunit concentration. Mol. Cell. Biol. 15, 27722781 (1995). Carroll, K. & Wickner, R.B. Translation and M1 double-stranded RNA propagation: MAK18=RPL41B and cycloheximide curing. J.Bacteriol. 177, 28872891 (1995). Komili, S., Farny, N.G., Roth, F.P. & Silver, P.A. Functional specificity among ribosomal proteins regulates gene expression. Cell 131, 557571 (2007). Parenteau, J. etal. Introns within ribosomal protein genes regulate the production and function of yeast ribosomes. Cell 147, 320331 (2011). Shows that introns in ribosomal protein genes regulate the expression of both the intron-containing genes and their paralogues. Hughes, T.R. etal. Widespread aneuploidy revealed by DNA microarray expression profiling. Nature Genet. 25, 333337 (2000). Steffen, K. etal. Ribosome deficiency protects against ER stress in Saccharomyces cerevisiae. Genetics 191, 107118 (2012). Barakat, A. etal. The organization of cytoplasmic ribosomal protein genes in the arabidopsis genome. Plant Physiol. 127, 398415 (2001). Whittle, C.A. & Krochko, J.E. Transcript profiling provides evidence of functional divergence and expression networks among ribosomal protein gene paralogs in Brassica napus. Plant Cell 21, 22032219 (2009). Weijers, D. etal. An. Arabidopsis Minute-like phenotype caused by a semi-dominant mutation in a ribosomal protein S5 gene. Development 128, 42894299 (2001). Shows differential expression of RPS5 paralogues in A. thaliana development. Falcone Ferreyra, M.L., Pezza, A., Biarc, J., Burlingame, A.L. & Casati, P. Plant L10 ribosomal proteins have different roles during development and translation under ultravioletB stress. Plant Physiol. 153, 18781894 (2010).
26. Williams, M.E. & Sussex, I.M. Developmental regulation of ribosomal protein L16 genes in Arabidopsis thaliana. Plant J. 8, 6576 (1995). 27. Degenhardt, R.F. & Bonham-Smith, P.C. Arabidopsis ribosomal proteins RPL23aA and RPL23aB are differentially targeted to the nucleolus and are disparately required for normal development. Plant Physiol. 147, 128142 (2008). 28. Degenhardt, R.F. & Bonham-Smith, P.C. Transcript profiling demonstrates absence of dosage compensation in Arabidopsis following loss of a single RPL23a paralog. Planta 228, 627640 (2008). 29. Marygold, S.J. etal. The ribosomal protein genes and Minute loci of Drosophila melanogaster. Genome Biol. 8, R216 (2007). 30. Kearse, M.G., Chen, A.S. & Ware, V.C. Expression of ribosomal protein L22e family members in Drosophila melanogaster: rpL22like is differentially expressed and alternatively spliced. Nucleic Acids Res. 39, 27012716 (2011). 31. Lopes, A.M. etal. The human RPS4 paralogue on Yq11.223 encodes a structurally conserved ribosomal protein and is preferentially expressed during spermatogenesis. BMC Mol. Biol. 11, 33 (2010). Shows the differential expression of human RPS4 paralogues and predicts the differences in protein structures between the paralogues. 32. Fisher, E.M. etal. Homologous ribosomal protein genes on the human X and Y chromosomes: escape from X inactivation and possible implications for Turner syndrome. Cell 63, 12051218 (1990). 33. Sugihara, Y. etal. Proteomic analysis of rodent ribosomes revealed heterogeneity including ribosomal proteins L10like, L22like 1, and L39like. J.Proteome Res. 9, 13511366 (2010). 34. Hariharan, N., Kelley, D.E. & Perry, R.P. Equipotent mouse ribosomal protein promoters have a similar architecture that includes internal sequence elements. Genes Dev. 3, 17891800 (1989). 35. Kim, C.H. & Warner, J.R. Messenger RNA for ribosomal proteins in yeast. J.Mol. Biol. 165, 7989 (1983).
VOLUME 13 | JUNE 2012 | 367
REVIEWS
36. Bortoluzzi, S., DAlessi, F., Romualdi, C. & Danieli, G.A. Differential expression of genes coding for ribosomal proteins in different human tissues. Bioinformatics 17, 11521157 (2001). 37. Kondrashov, N. etal. Ribosome-mediated specificity in Hox mRNA translation and vertebrate tissue patterning. Cell 145, 383397 (2011). First paper that shows heterogeneity in ribosomal protein expression in different tissues of a developing vertebrate embryo. It also shows that loss of a single ribosomal protein can have profound but specific effects on translational regulation invivo. 38. Ramagopal, S. & Ennis, H.L. Regulation of synthesis of cell-specific ribosomal proteins during differentiation of Dictyostelium discoideum. Proc. Natl Acad. Sci. USA 78, 30833087 (1981). 39. Ramagopal, S. Induction of cell-specific ribosomal proteins in aggregation-competent nonmorphogenetic Dictyostelium discoideum. Biochem. Cell Biol. 68, 12811287 (1990). Demonstrates dramatic changes in ribosome composition as D. discoideum shifts from vegetatively growing to aggregation-competent. 40. Sahin, F. etal. RPL38, FOSL1, and UPP2 are predominantly expressed in the pancreatic ductal epithelium. Pancreas 30, 158167 (2005). 41. Mills, A.A., Mills, M.J., Gardiner, D.M., Bryant, S.V. & Stanbridge, E.J. Analysis of the pattern of QM expression during mouse development. Differentiation 64, 161171 (1999). 42. Green, H. etal. The ribosomal protein QM is expressed differentially during vertebrate endochondral bone development. J.Bone Miner. Res. 15, 10661075 (2000). 43. Subramanian, A.R. Copies of proteins L7 and L12 and heterogeneity of the large subunit of Escherichia coli ribosome. J.Mol. Biol. 95, 18 (1975). 44. Hardy, S.J.S. The stoichiometry of the ribosomal proteins of Escherichia coli. Mol. General Genet. 140, 253274 (1975). 45. Oleinikov, A.V., Jokhadze, G.G. & Traut, R.R. Asingle-headed dimer of Escherichia coli ribosomal protein L7/L12 supports protein synthesis. Proc. Natl Acad. Sci. USA 95, 42154218 (1998). 46. Lam, Y.W., Lamond, A.I., Mann, M. & Andersen, J.S. Analysis of nucleolar protein dynamics reveals the nuclear degradation of ribosomal proteins. Curr. Biol. 17, 749760 (2007). 47. Lee, S.W. etal. Direct mass spectrometric analysis of intact proteins of the yeast large ribosomal subunit using capillary LC/FTICR. Proc. Natl Acad. Sci. USA 99, 59425947 (2002). 48. Carroll, A.J. Heazlewood, J. L., Ito, J. & Millar, A. H. Analysis of the Arabidopsis cytosolic ribosome proteome provides detailed insights into its components and their post-translational modification. Mol. Cell. Proteomics. 7, 347369 (2008). 49. Odintsova, T.I. etal. Characterization and analysis of posttranslational modifications of the human large cytoplasmic ribosomal subunit proteins by mass spectrometry and Edman sequencing. J.Protein Chem. 22, 249258 (2003). 50. Yu, Y., Ji, H., Doudna, J.A. & Leary, J.A. Mass spectrometric analysis of the human 40S ribosomal subunit: native and HCV IRES-bound complexes. Protein Sci. 14, 14381446 (2005). 51. Ramagopal, S. Covalent modifications of ribosomal proteins in growing and aggregation-competent Dictyostelium discoideum: phosphorylation and methylation. Biochem. Cell Biol. 69, 263268 (1991). 52. Krieg, J., Hofsteenge, J. & Thomas, G. Identification of the 40S ribosomal protein S6 phosphorylation sites induced by cycloheximide. J.Biol. Chem. 263, 1147311477 (1988). 53. Ruvinsky, I. etal. Ribosomal protein S6 phosphorylation is a determinant of cell size and glucose homeostasis. Genes Dev. 19, 21992211 (2005). 54. Zeidan, Q., Wang, Z., De Maio, A. & Hart, G.W. OGlcNAc cycling enzymes associate with the translational machinery and modify core ribosomal proteins. Mol. Biol. Cell 21, 19221936 (2010). 55. Spence, J. etal. Cell cycle-regulated modification of the ribosome by a variant multiubiquitin chain. Cell 102, 6776 (2000). 56. Fleischer, T.C., Weaver, C.M., McAfee, K.J., Jennings,J.L. & Link, A.J. Systematic identification and functional screens of uncharacterized proteins associated with eukaryotic ribosomal complexes. Genes Dev. 20, 12941307 (2006). 57. Colon-Ramos, D.A. etal. Direct ribosomal binding by a cellular inhibitor of translation. Nature Struct. Mol. Biol. 13, 103111 (2006). Demonstrates that the translation regulator Reaper can bind directly to the 40S ribosome subunit to inhibit cap-dependent translational initiation. 58. Fuchs, G., Diges, C., Kohlstaedt, L.A., Wehner, K.A. & Sarnow, P. Proteomic analysis of ribosomes: translational control of mRNA populations by glycogen synthase GYS1. J.Mol. Biol. 410, 118130 (2011). 59. Adams, D.R., Ron, D. & Kiely, P.A. RACK1, a multifaceted scaffolding protein: structure and function. Cell Commun. Signal. 9, 22 (2011). 60. Nilsson, J., Sengupta, J., Frank, J. & Nissen, P. Regulation of eukaryotic translation by the RACK1 protein: a platform for signalling molecules on the ribosome. EMBO Rep. 5, 11371141 (2004). Reviews, together with reference 59, the multiple functions of RACK1 as a protein that associates with the ribosome. 61. Ceci, M. etal. Release of eIF6 (p27BBP) from the 60S subunit allows 80S ribosome assembly. Nature 426, 574579 (2003). 62. Jannot, G. etal. The ribosomal protein RACK1 is required for microRNA function in both C.elegans and humans. EMBO Rep. 12, 581586 (2011). 63. Baum, S., Bittins, M., Frey, S. & Seedorf, M. Asc1p, a WD40domain containing adaptor protein, is required for the interaction of the RNA-binding protein Scp160p with polysomes. Biochem. J. 380, 823830 (2004). 64. Coyle, S.M., Gilbert, W.V. & Doudna, J.A. Direct link between RACK1 function and localization at the ribosome in vivo. Mol. Cell. Biol. 29, 16261634 (2009). 65. Li, A.M., Watson, A. & Fridovich-Keil, J.L. Scp160p associates with specific mRNAs in yeast. Nucleic Acids Res. 31, 18301837 (2003). 66. Zoncu, R. Efeyan, A. & Sabatini, D. M. mTOR: from growth signal integration to cancer, diabetes and ageing. Nature Rev. Mol. Cell Biol. 12, 2135 (2011). 67. Zinzalla, V., Stracka, D., Oppliger, W. & Hall, M.N. Activation of mTORC2 by association with the ribosome. Cell 144, 757768 (2011). Shows that mTORC2 interacts with the ribosome and this interaction activates mTORC2 independently of translation. 68. Oh, W.J. etal. mTORC2 can associate with ribosomes to promote cotranslational phosphorylation and stability of nascent Akt polypeptide. EMBO J. 29, 39393951 (2010). 69. Gunderson, J. etal. Structurally distinct, stage-specific ribosomes occur in Plasmodium. Science 238, 933937 (1987). Shows that different forms of rRNAs are expressed in different stages in the lifecycle of Plasmodium berghei. 70. Velichutina, I.V., Rogers, M.J., McCutchan, T.F. & Liebman, S.W. Chimeric rRNAs containing the GTPase centers of the developmentally regulated ribosomal rRNAs of Plasmodium falciparum are functionally distinct. RNA 4, 594602 (1998). 71. Rogers, M.J. etal. Structural features of the large subunit rRNA expressed in Plasmodium falciparum sprozoites that distinguish it from the asexually expressed large subunit rRNA. RNA 2, 134145 (1996). 72. Pavlakis, G.N., Jordan, B.R., Wurst, R.M. & Vournakis, J.N. Sequence and secondary structure of Drosophila melanogaster 5.8S and 2S rRNAs and of the processing site between them. Nucleic Acids Res. 7, 22132238 (1979). 73. Hassouna, N. Michot, B. & Bachellerie, J. P. The complete nucleotide sequence of mouse 28S rRNA gene. Implications for the process of size increase of the large subunit rRNA in higher eukaryotes. Nucleic Acids Res. 12, 35633583 (1984). 74. Ben-Shem, A. etal. The structure of the eukaryotic ribosome at 3.0 resolution. Science 334, 15241529 (2011). 75. Decatur, W. A. & Fournier, M. J. rRNA modifications and ribosome function. Trends Biochem. Sci. 27, 344351 (2002). 76. Castle, J.C. etal. Digital genome-wide ncRNA expression, including SnoRNAs, across 11 human tissues using polyA-neutral amplification. PLoS ONE 5, e11779 (2010). 77. Hughes, M.E., Grant, G.R., Paquin, C., Qian, J. & Nitabach, M.N. Deep sequencing the circadian and diurnal transcriptome of Drosophila brain. Genome Res. 3Apr 2012 (doi:10.1101/gr.128876.111). Higa-Nakamine, S. etal. Loss of ribosomal RNA modification causes developmental defects in zebrafish. Nucleic Acids Res. 40, 391398 (2012). Mauro, V.P. & Edelman, G.M. The ribosome filter hypothesis. Proc. Natl Acad. Sci. USA 99, 1203112036 (2002). Spriggs, K.A., Stoneley, M., Bushell, M. & Willis, A.E. Reprogramming of translation following cell stress allows IRES-mediated translation to predominate. Biol. Cell 100, 2738 (2008). Pelletier, J. & Sonenberg, N. Internal initiation of translation of eukaryotic mRNA directed by a sequence derived from poliovirus RNA. Nature 334, 320325 (1988). Jang, S.K. etal. A segment of the 5 nontranslated region of encephalomyocarditis virus RNA directs internal entry of ribosomes during in vitro translation. J.Virol. 62, 26362643 (1988). Yoon, A. etal. Impaired control of IRES-mediated translation in Xlinked dyskeratosis congenita. Science 312, 902906 (2006). First paper that shows the importance of rRNA pseudouridylation in viral and cellular IRES-mediated translation. Bellodi, C., Kopmar, N. & Ruggero, D. Deregulation of oncogene-induced senescence and p53 translational control in Xlinked dyskeratosis congenita. EMBO J. 29, 18651876 (2010). Bellodi, C. etal. Loss of function of the tumor suppressor DKC1 perturbs p27 translation control and contributes to pituitary tumorigenesis. Cancer Res. 70, 60266035 (2010). Montanaro, L. etal. Novel dyskerin-mediated mechanism of p53 inactivation through defective mRNA translation. Cancer Res. 70, 47674777 (2010). Jack, K. etal. rRNA pseudouridylation defects affect ribosomal ligand binding and translational fidelity from yeast to human cells. Mol. Cell 44, 660666 (2011). Shows that rRNA pseudouridylation affects IRES-dependent translational initiation and translational fidelity. Landry, D.M., Hertz, M.I. & Thompson, S.R. RPS25 is essential for translation initiation by the Dicistroviridae and hepatitis C viral IRESs. Genes Dev. 23, 27532764 (2009). Shows that a ribosomal protein interacts directly with an IRES element and regulates translation mediated by the IRES. Muhs, M. etal. Structural basis for the binding of IRES RNAs to the head of the ribosomal 40S subunit. Nucleic Acids Res. 39, 52645275 (2011). Horos, R. etal. Ribosomal deficiencies in Diamond Blackfan anemia impair translation of transcripts essential for differentiation of murine and human erythroblasts. Blood 119, 262272 (2012). Calvo, S.E., Pagliarini, D.J. & Mootha, V.K. Upstream open reading frames cause widespread reduction of protein expression and are polymorphic among humans. Proc. Natl Acad. Sci. USA 106, 75077512 (2009). Ingolia, N.T., Ghaemmaghami, S., Newman, J.R.S. & Weissman, J.S. Genome-wide analysis in vivo of translation with nucleotide resolution using ribosome profiling. Science 324, 218223 (2009). Ingolia, N.T., Lareau, L.F. & Weissman, J.S. Ribosome profiling of mouse embryonic stem cells reveals the complexity and dynamics of mammalian proteomes. Cell 147, 789802 (2011). Brar, G.A. etal. High-resolution view of the yeast meiotic program revealed by ribosome profiling. Science 335, 552557 (2012). Nishimura, T., Wada, T., Yamamoto, K.T. & Okada, K. The Arabidopsis STV1 protein, responsible for translation reinitiation, is required for auxin-mediated gynoecium patterning. Plant Cell 17, 29402953 (2005). Park, H. S. etal. A plant viral reinitiation factor interacts with the host translational machinery. Cell 106, 723733 (2001). Shows, together with reference 95, that RPL24 is important for the translational reinitiation of uORF containing mRNAs in plants. Schwanhausser, B. etal. Global quantification of mammalian gene expression control. Nature 473, 337342 (2011). Kozak, M. Initiation of translation in prokaryotes and eukaryotes. Gene 234, 187208 (1999).
78. 79. 80.
81.
82.
83.
84.
85.
86.
87.
88.
89. 90.
91.
92.
93.
94. 95.
96.
97. 98.
REVIEWS
99. Vesper, O. etal. Selective translation of leaderless mRNAs by specialized ribosomes generated by MazF in Escherichia coli. Cell 147, 147157 (2011). Shows that MazF in E.coli cleaves both rRNA and specific mRNAs, generating specialized ribosomes without the anti-ShineDalgarno sequence. These specialized ribosomes are competent to translate the leaderless mRNAs generated by MazF. 100. Byrne, M.E. A role for the ribosome in development. Trends Plant Sci. 14, 512519 (2009). Reviews the myriad of ribosomal protein mutants in plants. 101. Szakonyi, D. & Byrne, M.E. Ribosomal protein L27a is required for growth and patterning in Arabidopsis thaliana. Plant J. 65, 269281 (2011). 102. Pinon, V. etal. Three piggyback genes that specifically influence leaf patterning encode ribosomal proteins. Development 135, 13151324 (2008). 103. Brehme, K.S. Development of the Minute phenotype in Drosophila melanogaster. A comparative sudy of the growth of three Minute mutants. J.Exp. Zool. 88, 135160 (1941). 104. Marygold, S.J., Coelho, C.M. & Leevers, S.J. Genetic analysis of RpL38 and RpL5, two Minute genes located in the centric heterochromatin of chromosome 2 of Drosophila melanogaster. Genetics 169, 683695 (2005). 105. Kobayashi, S., Amikura, R. & Okada, M. Presence of mitochondrial large ribosomal RNA outside mitochondria in germ plasm of Drosophila melanogaster. Science 260, 15211524 (1993). 106. Amikura, R., Kashikawa, M., Nakamura, A. & Kobayashi, S. Presence of mitochondria-type ribosomes outside mitochondria in germ plasm of Drosophila embryos. Proc. Natl Acad. Sci. USA 98, 91339138 (2001). 107. Amikura, R., Sato, K. & Kobayashi, S. Role of mitochondrial ribosome-dependent translation in germline formation in Drosophila embryos. Mech. Dev. 122, 10871093 (2005). 108. Uechi, T. etal. Ribosomal protein gene knockdown causes developmental defects in zebrafish. PLoS ONE 1, e37 (2006). 109. Alexander, T., Nolte, C. & Krumlauf, R. Hox genes and segmentation of the hindbrain and axial skeleton. Annu. Rev. Cell Dev. Biol. 25, 431456 (2009). 110. Wellik, D.M. Hox patterning of the vertebrate axial skeleton. Dev. Dyn. 236, 24542463 (2007). 111. Oliver, E.R., Saunders, T.L., Tarle, S.A. & Glaser, T. Ribosomal protein L24 defect in belly spot and tail (Bst), a mouse Minute. Development 131, 39073920 (2004). 112. Oristian, D.S. etal. Ribosomal protein L29/HIP deficiency delays osteogenesis and increases fragility of adult bone in mice. J.Orthop. Res. 27, 2835 (2009). 113. Zhang, Y. & Lu, H. Signaling to p53: ribosomal proteins find their way. Cancer Cell 16, 369377 (2009). 114. Chakraborty, A., Uechi, T., Higa, S., Torihara, H. & Kenmochi, N. Loss of ribosomal protein L11 affects zebrafish embryonic development through a p53dependent apoptotic response. PLoS ONE 4, e4152 (2009). 115. Duan, J. etal. Knockdown of ribosomal protein S7 causes developmental abnormalities via p53 dependent and independent pathways in zebrafish. Int. J.Biochem. Cell Biol. 43, 12181227 (2011). 116. Danilova, N., Sakamoto, K.M. & Lin, S. Ribosomal protein S19 deficiency in zebrafish leads to developmental abnormalities and defective erythropoiesis through activation of p53 protein family. Blood 112, 52285237 (2008). 117. Panic, L. etal. Ribosomal protein S6 gene haploinsufficiency is associated with activation of a p53dependent checkpoint during gastrulation. Mol. Cell. Biol. 26, 88808891 (2006). 118. Boria, I. etal. The ribosomal basis of Diamond Blackfan anemia: mutation and database update. Hum. Mutat. 31, 12691279 (2010). 119. Gazda, H.T. etal. Ribosomal protein L5 and L11 mutations are associated with cleft palate and abnormal thumbs in DiamondBlackfan anemia patients. Am. J.Hum. Genet. 83, 769780 (2008). 120. Zhou, C. etal. Mutation in ribosomal protein L21 underlies hereditary hypotrichosis simplex. Hum. Mutat. 32, 710714 (2011). 121. Bramham, C.R. & Wells, D.G. Dendritic mRNA: transport, translation and function. Nature Rev. Neurosci. 8, 776789 (2007). 122. Vickers, C.A., Dickson, K.S. & Wyllie, D.J. Induction and maintenance of late-phase long-term potentiation in isolated dendrites of rat hippocampal CA1 pyramidal neurones. J.Physiol. 568, 803813 (2005). 123. Ostroff, L.E., Fiala, J.C., Allwardt, B. & Harris, K.M. Polyribosomes redistribute from dendritic shafts into spines with enlarged synapses during LTP in developing rat hippocampal slices. Neuron 35, 535545 (2002). 124. Moroz, L.L. etal. Neuronal transcriptome of Aplysia: neuronal compartments and circuitry. Cell 127, 14531467 (2006). Characterizes the mRNAs found in neuronal processes and finds an enrichment of ribosomal protein mRNAs in processes as compared to the soma. 125. Sossin, W.S. & DesGroseillers, L. Intracellular trafficking of RNA in neurons. Traffic 7, 15811589 (2006). 126. Moccia, R. etal. An unbiased cDNA library prepared from isolated Aplysia sensory neuron processes is enriched for cytoskeletal and translational mRNAs. J.Neurosci. 23, 94099417 (2003). 127. Tcherkezian, J., Brittis, P.A., Thomas, F., Roux, P.P. & Flanagan, J.G. Transmembrane receptor DCC associates with protein synthesis machinery and regulates translation. Cell 141, 632644 (2010). 128. Pinkstaff, J.K., Chappell, S.A., Mauro, V.P., Edelman, G.M. & Krushel, L.A. Internal initiation of translation of five dendritically localized neuronal mRNAs. Proc. Natl Acad. Sci. USA 98, 27702775 (2001). 129. Kressler, D., Hurt, E. & Bassler, J. Driving ribosome assembly. Biochim. Biophys. Acta 1803, 673683 (2010). 130. Baxter-Roshek, J.L., Petrov, A.N. & Dinman, J.D. Optimization of ribosome structure and function by rRNA base modification. PLoS ONE 2, e174 (2007). 131. Schafer, T. etal. Hrr25dependent phosphorylation state regulates organization of the pre40S subunit. Nature 441, 651655 (2006). 132. Martin-Marcos, P., Hinnebusch, A.G. & Tamame, M. Ribosomal protein L33 is required for ribosome biogenesis, subunit joining, and repression of GCN4 translation. Mol. Cell. Biol. 27, 59685985 (2007). 133. Ferreira-Cerca, S. etal. Analysis of the in vivo assembly pathway of eukaryotic 40S ribosomal proteins. Mol. Cell 28, 446457 (2007). 134. Peltz, S.W. etal. Ribosomal protein L3 mutants alter translational fidelity and promote rapid loss of the yeast killer virus. Mol. Cell. Biol. 19, 384391 (1999). 135. Rhodin, M.H., Rakauskaite, R. & Dinman, J.D. Thecentral core region of yeast ribosomal protein L11is important for subunit joining and translational fidelity. Mol. Genet. Genomics 285, 505516 (2011). 136. Henras, A.K. etal. The post-transcriptional steps of eukaryotic ribosome biogenesis. Cell. Mol. Life Sci. 65, 23342359 (2008). 137. Johnson, A.W., Lund, E. & Dahlberg, J. Nuclear export of ribosomal subunits. Trends Biochem. Sci. 27, 580585 (2002). 138. Warner, J.R. & McIntosh, K.B. How common are extraribosomal functions of ribosomal proteins? Mol. Cell 34, 311 (2009). Examines the extraribosomal function of many ribosomal proteins. 139. Dabeva, M.D. & Warner, J.R. Ribosomal protein L32 of Saccharomyces cerevisiae regulates both splicing and translation of its own transcript. J.Biol. Chem. 268, 1966919674 (1993). 140. Malygin, A.A., Parakhnevitch, N.M., Ivanov, A.V., Eperon, I.C. & Karpova, G.G. Human ribosomal protein S13 regulates expression of its own gene at the splicing step by a feedback mechanism. Nucleic Acids Res. 35, 64146423 (2007). 141. Wan, F. etal. Ribosomal protein S3: a KH domain subunit in NFB complexes that mediates selective gene regulation. Cell 131, 927939 (2007). 142. Mazumder, B. & Fox, P.L. Delayed translational silencing of ceruloplasmin transcript in -interferonactivated U937 monocytic cells: role of the 3untranslated region. Mol. Cell. Biol. 19, 68986905 (1999). 143. Mazumder, B. etal. Regulated release of L13a from the 60S ribosomal subunit as a mechanism of transcript-specific translational control. Cell 115, 187198 (2003). 144. Kapasi, P. etal. L13a blocks 48S assembly: role of a general initiation factor in mRNA-specific translational control. Mol. Cell 25, 113126 (2007). 145. Vyas, K. etal. Genome-wide polysome profiling reveals an inflammation-responsive posttranscriptional operon in -interferon-activated monocytes. Mol. Cell. Biol. 29, 458470 (2009). 146. Hsieh, A. C. et al. The translational landscape of mTOR signalling steers cancer initiation and metastasis. Nature 485, 5561 (2012).
Acknowledgements
The authors would like to thank D. Ruggero, C. Bellodi, M. McMahon, C. Stumpf and members of the Barna laboratory for discussion and critical reading of the manuscript. S.X. is supported by the Agency of Science, Technology and Research of Singapore. This work was supported by the Program for Breakthrough Biomedical Research, UCSF (to M.B.) and the National Institutes of Health (NIH) Directors New Innovator Award,1DP2OD008509 (to M.B.).
FURTHER INFORMATION
Maria Barnas homepage: http://biochemistry.ucsf.edu/labs/ barna/Site/Welcome.html
VOLUME 13 | JUNE 2012 | 369

Nature

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nature

Uploaded by

Copyright:

Available Formats

RESEARCH HIGHLIGHTS

Sequestration at the IPOD stops division

Spc42 was sequestered by Rnq1 when in its unstructured state

VOLUME 13 | JUNE 2012

MACMILLAN SOUTH AFRICA

When two become one

VOLUME 13 | JUNE 2012

Preventing WNT signalling

this ZNRF3 function is evolutionarily conserved

MACMILLAN SOUTH AFRICA

VOLUME 13 | JUNE 2012

miRNAs strict schedule

VOLUME 13 | JUNE 2012

WNT chews the fat with glucose uptake

VOLUME 13 | JUNE 2012

An autophagy-independent role for UVRAG

Ready to die fast

VOLUME 13 | JUNE 2012

Shelterin fends off six repair pathways

Hikeshi puts out the fire in the nucleus

Keeping cell cycle progression in check

VOLUME 13 | JUNE 2012

Axis of ageing: telomeres, p53 andmitochondria

a model of how different pathways ... intersect and converge on mitochondria

2012 Macmillan Publishers Limited. All rights reserved

Fatty acid oxidation Gluconeogenesis Glucose utilization

2012 Macmillan Publishers Limited. All rights reserved

Insulin, IGF1 ROS FeS NADH/NAD mTOR

5. 6. 7. 8. 9. 10. 11. 12. 13.

19. 20. 21. 22. 23.

24. 25. 26. 27. 28. 29.

VOLUME 13 | JUNE 2012 | 403

Competing interests statement

The authors declare no competing financial interests.

the elastic properties of the matrix dictate the mode of 3D migration

VOLUME 13 | JUNE 2012

2012 Macmillan Publishers Limited. All rights reserved

Human genetic disorders

p52 p44 p34 p8 XPD XPD

Tfb2 Ssl1 Tfb4 Tfb5 Rad3

CDK7 Cyclin H MAT1

Kin28 Ccl1 Tfb3

Box 2 | The nucleotide excision repair pathway

DNA ligase I DNA ligase III XRCC1

Removal of CAK from TFIIH CAK

Box 3 | The transcription of eukaryotic protein-coding genes

Sequential recruitment of NER factors

Small nuclear RNAs

Ubiquitylation machinery and proteasome

RNA Degradation of NRs and other TFs

Nature Reviews | Molecular Cell Biology

Figure 2 | TFIIH is an essential factor of transcription initiation. Gene activation

VOLUME 13 | JUNE 2012 | 351

49. 50. 51.

58. 59. 60. 61.

VOLUME 13 | JUNE 2012 | 353

Competing interests statement

The authors declare no competing financial interests.

Regulation of glucose transport by insulin: traffic control of GLUT4

Glucose Glucose Glycogen Muscle

Glucose Triglycerides Insulin Pancreas Nature Reviews | Molecular Cell Biology

2012 Macmillan Publishers Limited. All rights reserved

GLUT4-mediated glucose uptake Glucose

GSV P AS160 Vesicle translocation