Professional Documents
Culture Documents
history of humans
Alvaro Perdomo-Sabogal
1,2
, Sabina Kanton
1,2,a
,
Maria Beatriz C Walter
1,2,a
and Katja Nowick
1,2
Deciphering the molecular basis of how modern human
phenotypes have evolved is one of the most fascinating
challenges in biology. Here, we will focus on the roles of gene
regulatory factors (GRFs), in particular transcription factors
(TFs) and long non-coding RNAs (lncRNAs) during human
evolution. We will present examples of TFs and lncRNAs that
have changed or show signs of positive selection in humans
compared to chimpanzees, in modern humans compared to
archaic humans, or within modern human populations. On the
basis of current knowledge about the functions of these GRF
genes, we speculate that they have been involved in speciation
as well as in shaping phenotypes such as brain functions,
skeletal morphology, and metabolic processes.
Addresses
1
TFome Research Group, Bioinformatics Group, Interdisciplinary Center
of Bioinformatics, Department of Computer Science, University of
Leipzig, Ha rtelstrasse 16-18, D-04107 Leipzig, Germany
2
Paul-Flechsig-Institute for Brain Research, University of Leipzig,
Jahnallee 59, D-04109 Leipzig, Germany
Corresponding author: Nowick, Katja (nowick@bioinf.uni-leipzig.de)
Current Opinion in Genetics & Development 2014, 29:6067
This review comes from a themed issue on Genetics of human origin
Edited by Aida M Andre s and Katja Nowick
http://dx.doi.org/10.1016/j.gde.2014.08.007
0959-437/# 2014 Elsevier Ltd. All right reserved.
Introduction
The question of how relatively fewgenomic changes can
result in comparatively large phenotypic differences has
motivated much research. With the seminal paper by
King and Wilson [1], it became increasingly acknowl-
edged that gene expression differences are likely to
drive many phenotypic distinctions. Accordingly,
sequence changes in gene regulatory regions between
and within species have frequently been found and
linked to phenotypic alterations (Siepel and Arbiza,
2014, this issue). While it is expected that each single
change of this type mainly affects the expression of only
the respective gene, changes in the gene regulatory factors
(GRFs) could potentially modify the phenotype consider-
ably. This is because a GRF typically regulates the expres-
sion of several to many genes (Box 1). Whereas these
pleiotropic characteristics of GRFs suggest that many
GRFs should be conserved, non-deleterious evolutionary
changes in GRFs are prime candidates for driving pheno-
typic diversity. Such changes can encompass differences in
gene copy number, sequence, or expression.
In this review we will focus on two classes of GRFs,
transcription factors (TFs) and long non-coding RNAs
(lncRNAs). Both groups of genes primarily function at
transcriptional level to regulate gene expression by
directly or indirectly interacting with DNA. It is import-
ant to highlight that other types of non-coding RNA
molecules also regulate gene expression, but they mainly
operate at post-transcriptional or translational level, for
instance, microRNAs, tRNAs, or snoRNAs [2]. Those
readers interested in the evolution of e.g. microRNAs can
refer to some work on miRNAs in modern human popu-
lations [3,4], miRNA changes in anatomically modern
humans [5], and miRNA comparisons between humans
and other primates [610]. We will present here examples
of TF and lncRNA genes that have changed during
human evolution and discuss how they might be related
to human specic phenotypes or have medical con-
sequences. To complement this review, we explored
GRF genes within candidate regions that have been
found to be under positive selection in recent genome-
wide studies [1113,14
].
Evolutionary changes after the split from
chimpanzees
The appearance of new genes can have a strong impact on
the way phenotypes evolve. Twenty-three (0.7%) of the
3315 human TFs are human specic (Perdomo-Sabogal
et al., in preparation) with no ortholog in other species
including chimpanzees [30,31
], forkhead
box D4-like 5 (FOXD4L5), and synovial sarcoma X break-
point 1 (SSX1). Genetic disease associations and medical
a
They both equally contributed to this work.
Available online at www.sciencedirect.com
ScienceDirect
Current Opinion in Genetics & Development 2014, 29:6067 www.sciencedirect.com
GRF evolution Perdomo-Sabogal et al. 61
Box 1 The molecular biology of GRFs.
Throughout this paper we will refer to GRFs as molecules that directly
or indirectly bind to DNA in order to regulate the expression of target
genes. These molecules comprise two major classes, transcription
factors (TFs) and long non-coding RNAs (lncRNAs). Different classifi-
cation schemes for TFs [15,16,17
,28
] or 53 000 [29
] but this
number is likely to change as more tissues are sequenced to higher
depth.
Figure B1
Repressive histone marks
(a) (c)
(b) (d)
Nucleosomes
DNA
Pre-mRNA
Active histone
marks
Pol II
TFs
(Chromatin modifiers)
Chromatin
remodelling
complex
IncRNA DNA
triplex
IncRNA interacting with
chromatin remodeling
complex
IncRNA
Pol II
TFs - cofactors
DNA binding TFs
Upstream
promoter region
TATA box
DNA
TSS
mRNA decay
Pre-mRNA
processing
Cytoplasm
Nucleus
Pol II
inhibition
Current Opinion in Genetics & Development
(a) Schematic representation of chromatin structure and chromatin-mediated gene regulation. TFs acting as chromatin modifiers and chromatin
remodeling complexes dynamically modify chromatin architecture and allow the access of the TF machinery to the DNA, thus regulating gene
expression. (b) TFs bind DNA promoter regions and recruit other TF co-factors to regulate transcription. Upon TF binding, chromatin changes
conformation for the RNA polymerase to initiate transcription. Similarly, TFs can bind and interact with other TFs to repress gene expression. (c)
LncRNAs can sequester TFs or participate in chromatin remodeling. Similarly to TFs, they can activate or repress transcription. In addition, lncRNAs
can form lncRNA-DNA triplex structures and regulate the assembling of the pre-initiation complex. (d) By binding to polymerase II, lncRNAs can also
repress transcription. TSS: transcription start site, Pol II: polymerase II.
www.sciencedirect.com Current Opinion in Genetics & Development 2014, 29:6067
consequences have been used as evidence to understand
the most likely functions genes are involved in. Interest-
ingly, some human specic TFs have been associated
with human diseases such as cancer (see also Long and
Zhang, 2014, this issue). For instance, a fusion of SSX1 with
the co-activator SYT has been found in synovial sarcoma
[32] and BAGE2 seems to be exclusively expressed in 22%
of melanomas [33]. Additional associations include variants
of ZNF286B, a gene that seems to be connected with
longevity [34], which might be also related to the longer
lifespan of humans compared to other primates.
LncRNAs have a higher fraction of human specic genes
than TFs. In one study, around 3.2% of the
14 682 reported lncRNAs have been suggested to be
human specic. Moreover, about 8% of the lncRNAs
expressed in humans were not found to be expressed
in chimpanzees or bonobos [27
]. However, it is still
possible that these and more lncRNAs are functional in
other species under different conditions. Interestingly,
hominid specic lncRNAs were enriched for expression
in testis, which could be related to sexual selection or
testis-specic functions [28
].
Among the rst TFs discovered to have human specic
sequence changes with a likely impact on human evolu-
tion is forkhead box P2 (FOXP2) (Figure 1). It is charac-
terized by two human specic amino acids [35]. From an
evolutionary perspective, FOXP2 draws attention because
of its role in the development of speech and language [36].
It has thus been speculated that either one or both human
specic amino acids found in FOXP2 are involved in
language evolution. This conclusion has been supported
by experimental studies showing that the human and
chimpanzee versions of FOXP2 regulate different sets
of target genes [37]. In addition, mice carrying the human
version of FOXP2 displayed changes in the central ner-
vous system, including changes in synaptic plasticity and
dendrites of the cortico-basal ganglia [38].
A second example of TFs with potential importance for
human evolution includes the family of ZNF proteins
with a Kruppel-associated box (KRAB-ZNFs). This family
has experienced remarkable evolutionary changes in
humans compared to non-human primates [31
,39,40].
Noticeably, KRAB-ZNF genes are enriched among the
genes with differential expression between the prefrontal
cortex of humans and chimpanzees, suggesting a role for
KRAB-ZNFs in establishing some specic features of the
human brain [39]. It has been proposed that especially
evolutionary young KRAB-ZNF genes are involved in
early brain development in humans [30]. Furthermore,
the existence of human specic changes in splicing pat-
terns, DNA binding residues and number of zinc nger
domains in KRAB-ZNFs indicate accelerated evolution
with functional divergence for this TF family in the
human lineage [40]. Despite we can only speculate which
impacts these changes have had at the phenotypic level, it
is interesting to note that several KRAB-ZNF genes are
involved in embryonic development, spermatogenesis,
bone formation, metabolism, and other functions [41]
and that mutations in various KRAB-ZNF genes have
been implicated in cognitive disorders, such as intellec-
tual disability [42] and autism spectrum disorders [43,44].
A particular example of a fast evolving KRAB-ZNF is PR
domain-containing protein 9 (PRDM9). The amino acids
that determine the binding site specicity of PRDM9 are
highly diverged between humans and chimpanzees and
evolve under positive selection in primates [45]. This
probably indicates that human PRDM9 is likely to bind to
DNA sites distinct from chimpanzee PRDM9. The TF
PRDM9 determines recombination hotspots [46], which
might explain why recombination hotspots in humans and
chimpanzees do not usually overlap [47]. However, in
modern humans, most hotspots are shared between indi-
viduals [48], making PRDM9 a good candidate gene for
facilitating speciation events that occurred after the split
of humans and chimpanzees.
Sequence changes in lncRNAs occur more rapidly than in
TFs. In particular, changes that alter the structure of the
lncRNA are of major interest for studying new functions.
One way to alter the structure of a lncRNA is through gain
or loss of exons. This can for instance occur through
changes in splice sites. A recent study discovered
434 human specic splice sites (among a total of
15 870 investigated), which might produce a human
specic transcript structure for 388 lncRNA genes
(Nitsche et al., unpublished data). Exons can also be
gained by the incorporation of transposable elements,
as has been seen for the lncRNA ANRIL during primate
evolution [49]. Given that the incorporation of transpo-
sons into lncRNAs seems common [50] and that many
transposable elements are primate or even human specic
[51], transposons might have had an important impact on
modifying structure and function of lncRNAs during
human evolution.
Besides changes in exon number, single nucleotide
changes at particular positions can also drastically alter
the lncRNA structure. A good example of this is the
HAR1 region, which is the fastest evolving region in
the human genome, while being highly conserved across
vertebrates (see also Pollard and Hubisz, in this issue)
[54]. It has 118 bp that contain 18 xed human specic
changes and forms a stable secondary RNA structure,
which differs from the less stable ancestral HAR1 struc-
ture [52,53] (Figure 1). The HAR1 region is part of a pair
of lncRNA genes, the HAR1F and HAR1R, which overlap
with each other and are divergently transcribed [54].
HAR1 is very specically expressed in Cajal-Retzius cells,
neurons that are important for the organization of the
developing brain. These neurons have been connected
62 Genetics of human origin
Current Opinion in Genetics & Development 2014, 29:6067 www.sciencedirect.com
with the occurrence of Alzheimers disease, autism,
schizophrenia and other mental disorders [54]. Thus,
HAR1 seems essential for a healthy development and
functioning of the human brain. Another example of a
lncRNA with fast sequence evolution in primates is HOX
transcript antisense RNA (HOTAIR). Especially its exons
one and six have experienced many sequence changes
[55]. HOTAIR displays some structural differences be-
tween humans and chimpanzees [55] that could poten-
tially result in functional dissimilarities between the
orthologs. It orchestrates the transcriptional silencing of
around 40 kb within the HOXD locus [56] and may thus
play a role during development [5759].
Evolutionary changes after the split from
archaic humans
The availability of genomes from two archaic humans,
Neanderthals and Denisovans, allows us to better under-
stand how and when evolution has acted on particular
GRF genes. For instance, for the aforementioned
FOXP2, it was shown that the human specic amino
acids already existed in Neanderthals [6062]. However,
a variant in intron eight of FOXP2, which corresponds to
a position in a putative binding site of the TF POU3F2,
seems to be specic to modern humans [63]. This variant
may play a role in the regulation of FOXP2 expression,
suggesting that functional evolutionary changes in
FOXP2 continued on the lineage to modern humans.
This putative functional change could be another likely
reason to explain the signatures of a selective sweep that
may have occurred in the last 50 000 years within the
genomic region of FOXP2 [63,64]. It is possible that
these ongoing changes within the FOXP2 locus resulted
in differences in the timing, location or amplitude of
FOXP2 expression and contributed to improved speech
capabilities.
Our comparisons between modern humans, Neanderthals
and Denisovan identied a repertoire of TF genes that
has been duplicated during recent human evolution
(Supplementary Table S1). Such changes in copy number
could have an impact on human phenotypes, for instance,
by causing expression changes due to altered gene dosage
or by the disruption of genes from their regulatory
sequence through the structural rearrangement accom-
panying the duplication [65]. While most of these differ-
ences in copy number exist as copy number variants in
healthy modern human individuals [65], our analysis
revealed that some of these duplicated genes have been
associated with developmental, immune, cardiovascular,
neurodegenerative, psychological, and cancerous diseases
in modern humans (Supplementary Table S1). For
example, BolA-Like 2 protein (BOLA2) is a TF gene
with two to ve copies that have been introduced into the
modern human genome after the split from Neanderthals
and Denisovans [66,67]. Interestingly, these copies reside
in a chromosomal region (16p11.2) associated with devel-
opmental constraints and intellectual disability [68,69].
Signatures of positive selection have been found within
the genomic regions of the Runt-related transcription
factor 2 (RUNX2) and the Homeobox D gene cluster
(HOXD) in modern humans [61,70,71]. While the protein
sequence of RUNX2 is conserved between modern
humans, Neanderthals and Denisovans, its promoter
region presents two derived alleles in modern humans.
This might have caused gene regulatory changes with
implications for cranial, skeletal and bone development in
modern humans [70]. The HOXDcluster is also important
during development, in particular during the develop-
ment of the axial skeleton and morphogenesis of fore-
limbs [72]. Furthermore, the HOXD cluster displays
differences in methylation between archaic and modern
humans [73
] and a recent
selective sweep [11] in Utah residents with northern
and western European ancestry (CEU). Single nucleotide
variants in this gene have been connected to changes in
high-density lipoprotein (HDL) cholesterol, triglyceride
levels and cardiovascular disease risk in humans [77,78].
Mutations in Wwox caused symptoms similar to epilepsy,
mental retardation, and ataxia in rats and upper motor
64 Genetics of human origin
Table 1
List of TF genes overlapping human genomic regions showing signatures either of positive selection [12,13,14
] or a recent selective
sweep [11]
Gene symbol Number
of genes
Source
WWOX 1 [11,12,14
]
MYEF2, FBN1 2 [11,12]
ZMYM6 1 [11,14
]
PPARA, KCNH5 2 [12,13]
HIF1A, SNAPC1, DPF1 3 [12,14
]
KCNH7 1 [13,14
]
CTNND2, BMI1, AFF2, BBX, NFE2L2 5 [11]
RGS9, ERBB4, ATF6, PHF19, DUSP12, RFX3, CIITA, NCOA7, APC, TRIM14, SETBP1, POLR2K, FOXE1,
HSF2, YTHDC1, HEY2
16 [12]
CLOCK, MSTN, LIN28B, ISX 4 [13]
ANKRD45, RRN3, SFPQ, SIN3A, SLC30A9, CCDC71, RNF135, NCOA1, PCGF1, HIRA, MCM6, ASXL2,
FOXP1, RHOA, TERF2IP, TAX1BP3, HIPK1, KCNIP4, RFX5, ADNP2, ZBTB41, PAPOLA, POGZ, FMNL2,
ACTR5, PAWR, LHX8, USF1, EBF1, LBX2, CHD2, ARIH2, PHTF1
33 [14
]
Figure 1
legend
TF IncRNA
changes
after the split
from chimpanzees
changes
after the split
from archaic
humans
two
amino acid
substitutions
intron 8
variant
several
duplications
differences within
modern humans
positively
selected
additional and
suspected
functions
implicated in
cardiovascular
diseases,
may be
involved in
brain
development
0.7 % human specific TFs
Gene regulatory factors
3.2 % human specific IncRNAs
WWOX FOXP2 BOLA2 HAR1F/R KRAB-ZNFs PRDM9 HOXD HOTAIR RUNX2
18 human
specific
changes,
structural
differences
differential expression in
prefrontal cortex,
human specific
KRAB-ZNF genes,
changes in
splicing patterns,
DNA binding,
ZNF domain number
high
divergence in
DNA binding
residues
non-synonymous
allelic variations
positive
selection
within gene
cluster,
differences in
methylation
positive
selection
within
genomic
region,
promotor
region with
two derived
alleles
structural
differences
prominent function metabolism brain recombination skeleton
involved in
language
evolution
might be
involved in
intellectual
disability
specific
expression in
Cajal-Retzius
neurons
during brain
development
potentially involved in,
intellectual disability,
autism,
Alzheimers disease;
metabolism, bone formation,
embryonic development
might have
impacts on
fertility and
speciation
development
of axial
skeleton and
forelimb
morphogenesis
maybe involved
In human
specific skeletal
morphology
via regulation
of HOXD
implications in
cranial,
skeletal and
bone
development
Current Opinion in Genetics & Development
Functional roles and evolutionary changes of a selection of TFs and lncRNAs potentially involved in human evolution. TFs (given in rectangles) and
lncRNAs (given in hexagons) are grouped in boxes based on their different functions. Evolutionary changes after the split from chimpanzees and
archaic humans are given in yellow and red boxes, respectively, and differences found within modern humans are given in blue boxes. The arrow
between KRAB-ZNFs and PRDM9 indicates PRDM9 belonging to KRAB-ZNFs, whereas the arrow between HOTAIR and HOXD indicates regulation of
HOXD by HOTAIR. All references concerning this summarizing figure can be found within the text.
Current Opinion in Genetics & Development 2014, 29:6067 www.sciencedirect.com
neuron disease phenotypes in mice [79]. We speculate
that evolutionary changes in WWOX might have been
involved in functional changes in metabolism and brain in
modern humans. Also associations between long inter-
genic ncRNAs (lincRNAs) and regions under positive
selection have been reported. In particular, 48 lincRNAs
and three putative regulatory variants of lincRNAs are
located within these regions [14
].
Conclusions
In this review, we explored and illustrated the putative
roles of TFs and lncRNAs during human evolution
(Figure 1). Owing to their function in the regulation of
many genes, we expect evolutionary changes in GRF
genes to be of considerable inuence on phenotypes.
Although previous research efforts have shed light on
the evolution of some prominent GRFs, for instance the
TFs FOXP2, PRDM9, RUNX2, and HOXD and the
lncRNAs HAR1F/R, HOTAIR, and ANRIL, in general,
evolutionary changes in GRFs are still largely unex-
plored. Moreover, a major drawback for interpreting
phenotypic consequences of GRF changes is that many
GRFs have not yet been completely characterized at
functional level. Nevertheless, a few themes seem to
emerge. First, TF and lncRNA genes have continuously
been altered during human evolution. Second, as a class,
lncRNAs are evolving faster than TFs. Third, at least
some of the GRFs that changed during human evolution
seem to be involved in brain functions, metabolism, and
skeletal morphology, phenotypes that have indeed chan-
ged during recent human evolution. It seems possible
that TFs and lncRNAs together have shaped gene regu-
latory networks affecting human phenotypes in complex
ways. We hope that our review stimulates further research
on evolutionary and functional changes of GRFs to gain a
better understanding of the evolution of our own species.
Acknowledgements
This work was supported by the Departamento Administrativo de Ciencia,
Tecnologa e Innovacio n Colciencias from Colombia, call Francisco Jose de
Caldas 497/2009 (APS), Science without Borders Brazil (MBCW) and the
Volkswagen Foundation within the initiative Evolutionary Biology (KN).
Appendix A. Supplementary data
Supplementary material related to this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.gde.2014.08.007.
References and recommended reading
Papers of particular interest, published within the period of review,
have been highlighted as:
of special interest
of outstanding interest
1. King MC, Wilson AC: Evolution at two levels in humans and
chimpanzees. Science (New York, NY) 1975, 188:107-116.
2. Qureshi IA, Mehler MF: Emerging roles of non-coding RNAs in
brain evolution, development, plasticity and disease. Nat Rev
Neurosci 2012, 13:528.
3. Quach H, Barreiro LB, Laval G, Zidane N, Patin E, Kidd KK,
Kidd JR, Bouchier C, Veuille M, Antoniewski Cet al.: Signatures of
purifying and local positive selection in human miRNAs. Am J
Hum Genet 2009, 84:316-327.
4. Li J, Liu Y, Xin X, Kim TS, Cabeza EA, Ren J, Nielsen R, Wrana JL,
Zhang Z: Evidence for positive selection on a number of
microRNA regulatory interactions during recent human
evolution. PLoS Genet 2012, 8:e1002578.
5. Lopez-Valenzuela M, Ramrez O, Rosas A, Garc a-Vargas S, de la
Rasilla M, Lalueza-Fox C, Espinosa-Parrilla Y: An ancestral miR-
1304 allele present in Neanderthals regulates genes involved
in enamel formation and could explain dental differences with
modern humans. Mol Biol Evol 2012, 29:1797.
6. Brameier M: Genome-wide comparative analysis of
microRNAs in three non-human primates. BMC Res Notes
2010, 3:64.
7. Somel M, Xie G, Li N, Hu Y, Chen W, Pa a bo S, Khaitovich P, Liu X,
Tang L, Yan Z et al.: MicroRNA-driven developmental
remodeling in the brain distinguishes humans from other
primates. PLoS Biol 2011, 9:e1001214.
8. Hu HY, Zhao M, Zeng R, Chen W, Pa a bo S, Khaitovich P, Guo S, Xi J,
Yan Z, Fu Net al.: MicroRNA expression and regulation in human,
chimpanzee, and macaque brains. PLoS Genet 2011, 7:e1002327.
9. Dannemann M, Pru fer K, Lizano E, Nickel B, Burbano HA, Kelso J:
Transcription factors are targeted by differentially expressed
miRNAs in primates. Genome Biol Evol 2012, 4:552-564.
10. McLoughlin HS, Wan J, Spengler RM, Xing Y, Davidson BL:
Human-specic microRNA regulation of FOXO1: implications
for microRNA recognition element evolution. Hum Mol Genet
2014, 23:2593-2603.
11. Sabeti PC, Varilly P, Fry B, Lohmueller J, Hostetter E, Cotsapas C,
Xie X, Byrne EH, McCarroll SA, Gaudet R et al.: Genome-wide
detection and characterization of positive selection in human
populations. Nature 2007, 449:913-918.
12. Pickrell JK, Feldman MW, Pritchard JK, Coop G, Novembre J,
Kudaravalli S, Li JZ, Absher D, Srinivasan BS, Barsh GS et al.:
Signals of recent positive selection in a worldwide sample of
human populations. Genome Res 2009, 19:826-837.
13. Metspalu M, Remm M, Pitchappan R, Singh L, Thangaraj K,
Villems R, Kivisild T, Romero IG, Yunusbayev B, Chaubey G et al.:
Shared and unique components of human population
structure and genome-wide signals of positive selection in
South Asia. Am J Hum Genet 2011, 89:731-744.
14.