Professional Documents
Culture Documents
pubs.acs.org/biochemistry
Laboratoire de Biologie et Pharmacologie Appliquee CNRS, Ecole Normale Superieure Paris-Saclay, 61 avenue du president Wilson,
94235 Cachan, France
Institut Curie, PSL Research University, Chemical Cell Biology group, 26 rue dUlm, 75248 Paris Cedex 05, France
Division of Molecular Pharmacology, Cancer Chemotherapy Center, Japanese Foundation for Cancer Research, Tokyo 135-8550,
Japan
#
Department of Applied Chemistry, Faculty of Science, Tokyo University of Science, Tokyo 162-8601, Japan
*
S Supporting Information
Figure 1. Small molecules and proteins used in this study. (A) Chemical structures of the compounds. (B) Small GTPases and GEF constructs used
for the in vitro nucleotide exchange assay in solution (left) and in the presence of a membrane (right). Numbers refer to the rst and last residues of
each construct. (C) SDSPAGE analysis of the recombinant proteins. The associated Figure S2 shows control inhibition assays performed with the
related small GTPase Rac1 and its GEF Dock5 and DLS of liposome integrity in the presence of the inhibitors.
concurrent processes through the combination of related Legionella pneumophila and Rickettsia prowazekii to activate Arf
regulators,7 whereby even on-target inhibition can lead to GTPases in infected cells.14 In addition to GEFs, membranes
detrimental o-pathway eects. play a pivotal role in the activation of Arf GTPases. These
Arf GTPases constitute a subfamily of small GTPases with GTPases depart from canonical small GTPases by a built-in
regulatory functions in most major aspects of membrane trac structural mechanism whereby they couple their activation by
and organelle structure in the cell.8,9 Their roles in diseases GTP to their recruitment to membranes.19 This mechanism
have expanded over the past decade, notably in cancer,1012 uses a myristoylated N-terminal helix, which constitutes the
inammation,13 and infections.1416 In metazoans, they are major site of interaction of Arf GTPases with membranes, and
activated by 15 dierent GEFs (ArfGEF hereafter) that bear a it is controlled by ArfGEFs to ensure that Arf-GTP is associated
conserved catalytic Sec7 domain that stimulates GDP/GTP with membranes.20,21 ArfGEFs are themselves regulated by
exchange.17 The Sec7 domain is decorated with various other membranes through membrane-binding domains: members of
domains, the nature of which denes ArfGEF subfamilies.18 An the cytohesin, EFA6, and BRAG subfamilies carry a
additional ArfGEF, RalF, is secreted by the bacterial pathogens phospholipid-binding pleckstrin homology (PH) domain that
5126 DOI: 10.1021/acs.biochem.7b00706
Biochemistry 2017, 56, 51255133
Biochemistry Article
follows their Sec7 domain; large Golgi ArfGEFs have molecule inhibitors constitute important tools for developing
membrane-binding elements in the N- and C-termini of their drugs to manipulate them in disease-relevant settings. However,
Sec7 domain; and RalF proteins have a membrane-binding using these inhibitors for deciphering molecular mechanisms in
domain uniquely found in bacteria that is also involved in disease-related pathways faces several challenges. First, all
autoinhibition.22 ArfGEFs share a highly conserved Sec7 catalytic domain,17
The rst discovered GEF inhibitor is Brefeldin A (BFA), a which raises the issue drug specicity. Second, inhibition
natural compound that inhibits Arf functions at the Golgi.23,24 eciency in vitro has been mostly analyzed in solution, which
The biochemical and structural mechanism of BFA has been underestimates GEF eciencies by one to several orders of
deciphered, leading to the rst direct demonstration that small magnitude.3942 Quantitative in vitro analysis of inhibitors
GTPases and their GEFs are druggable targets despite their using puried proteins is an ecient approach to address these
complex structural landscape.21,25,26 BFA binds at the interface questions.43 In this study, we used uorescence kinetics-based
between Arf and its GEFs, thereby stalling the Arf-GDP/ assays with puried Arf GTPases and representative human and
ArfGEF intermediate that initiates the nucleotide exchange bacterial ArfGEFs to establish the inhibitory proles of four
reaction in an abortive conformation. This mode of action inhibitors of unrelated structure, BFA, SecinH3, M-COPA, and
allows it to target only the large Golgi ArfGEFs by recognizing NAV-2729 (Figure 1A). The eect of membranes was
a tyrosine in the Arf-binding site, which is substituted by a investigated for three of these inhibitors, by incorporating
phenylalanine in BFA-insensitive ArfGEFs, and to be inactive articial membranes in the kinetics assays and by using
toward the Arf6 isoform.27 Remarkably, various chemicals of lipidated GTPases and ArfGEF constructs that contain a
unrelated structure have since been discovered to inhibit membrane-binding segment. This study reveals that each small
ArfGEFs, using a variety of screening strategies. The interfacial molecule has a specic inhibitory prole. It also uncovers an
mechanism of BFA inspired an in silico screen that produced unexpected sensitivity of LpRalF to BFA that was detected only
LM11, a chemical compound that inhibits Arf1 and cytohesins in the presence of membranes. This analysis should enable an
in vitro and in cells.28 Another inhibitor of cytohesins, SecinH3, accurate interpretation of the eect of these inhibitors in
was discovered using an RNA aptamer displacement in vitro dissecting the biology of Arf GTPases and their GEFs and
screen.29 A uorometric GEF assay-based screen identied facilitate their further development.
NAV-2729, which blocks spontaneous activation of Arf6 and its
activation by cytohesins and BRAG.12 Other inhibitors have
been discovered by phenotypic screens. Golgicide A was
MATERIALS AND METHODS
Chemicals. GDP, GTP, and N-methylanthraniloyl GTP
identied from a high-throughput luciferase-based screen for (mGTP) were from Jena Bioscience. BFA was from Sigma and
small molecules that inhibit intracellular toxin transport in host SecinH3 from Tocris Bioscience. M-COPA was synthesized as
cells; it causes Golgi and TGN dispersal by targeting GBF1, an described in ref 33. NAV-2729 was synthesized as described in
ArfGEF that functions at the Golgi.30 Another inhibitor of Supporting Information (spectra shown in Figure S1).
GBF1 is LG186;31 LG186 is a derivative of Exo2, which was Theoretical solubilities were predicted using the U.S. Environ-
originally identied through an image-based phenotypic screen mental Protection Agencys EPISuite [compiled on Chem-
for perturbation of exocytosis.32 M-COPA (also called AMF- spider (http://www.chemspider.com)].
26) was identied on the basis of the similarity of its growth Protein Expression and Purication. Protein constructs
inhibitory prole toward human cancer cell lines with that of used in this study are shown in Figure 1B. N-Terminally
BFA; it dispersed GBF1 in a manner reminiscent of that of BFA truncated bovine 17Arf1 (identical to the human form) and
and was proposed to inhibit GBF1 specically.33 human 13Arf6 were expressed and puried and loaded with
ArfGEF inhibitors have already a long history in the GDP prior to use as described in ref 27. Myristoylated full-
elucidation of molecular pathways in the cell.34 For instance, length Arf1 (myrArf1 hereafter) was obtained by co-expression
BFA was instrumental in deciphering the molecular basis of with yeast N-myristoyl transferase in Escherichia coli and
membrane dynamics at the Golgi.35 Golgicide A identied a puried as described in ref 44. Human BRAG2Sec7, ARNOSec7,
pivotal role for GBF1 in maintaining bidirectional transport in BIG1Sec7, and EFA6Sec7 were produced in E. coli and puried as
the cell.30 SecinH3 revealed the role of cytohesins and Arfs in described in refs 39, 45, 27, and 41, respectively. Human
the response to insulin.29,36 LM11 pointed to a role of Arf1 in BRAG2Sec7PH, ARNOSec7PH, and EFA6Sec7PH were expressed in
breast cancer cell invasion through formation of focal E. coli and puried as described in refs 39, 41, and 46,
adhesions.37 ArfGEF inhibitors also contributed to identifying respectively. RalF proteins from L. pneumophila (LpRalF) and
potential therapeutic pathways that could be manipulated to R. prowazekii (RpRalF) and their isolated Sec7 domains were
combat diseases. For example, mouse models of inammation produced and puried as described in refs 42 and 47. Human
exposed to SecinH3 had reduced vascular permeability, a Rac1 and the GEF domain of mouse Dock5 were obtained as
process leading to inammation, which suggested that described in ref 48. Human BIG1DHSec7 was expressed in insect
cytohesins and the Arf6 GTPase are valid targets for the cells and puried as described in ref 49. All proteins were highly
treatment of inammatory conditions.13 Likewise, several pure as assessed by SDSPAGE (Figure 1C). Protein
ArfGEF inhibitors were shown to have antitumoral activities. concentrations were measured using their respective absorption
NAV-2729 reduced the level of uveal melanoma cell coecient at 280 nm after centrifugation to remove any
proliferation and tumorigenesis in a mouse model.12 M- precipitate.
COPA led to tumor regression in mice xenografts.33,38 BFA, Liposome Preparation. All lipids were purchased from
Golgicide A, SecinH3, and LM11 markedly reduced the Avanti Polar Lipids. Experiments carried out with human
incidence of stem cell tumors in Drosophila by blocking lipid ArfGEFs were performed with liposomes containing 48%
droplet usage.11 phosphatidylcholine (PC), 20% phosphatidylethanolamine
These studies highlighted Arf GTPases and their regulators (PE), 30% phosphatidylserine (PS), and 2% phosphatidylino-
as attractive candidates for therapeutic intervention. Small sitol 4,5-diphosphate [PI(4,5)P2]. Experiments with LpRalF
5127 DOI: 10.1021/acs.biochem.7b00706
Biochemistry 2017, 56, 51255133
Biochemistry Article
and RpRalF were performed with liposomes containing 40% performed on membranes, we used lipidated full-length Arf1,
PC, 19% PE, 25% PS, 1% PI(4,5)P2, and 15% cholesterol. All which carries a myristate lipid at its N-terminus. Membrane-
liposomes were prepared and extruded through 200 nm binding elements in ArfGEF constructs used in liposome-based
polycarbonate lters as described previously.39 Liposomes assays were the PH domains for ARNO, BRAG2, and EFA6,
were stored at room temperature and used within 2 days. the DCB-HUS domain for BIG1, and the C-terminal capping
Nucleotide Exchange Assays. Nucleotide exchange domain for RalF proteins. The PH domain of ARNO50 and the
kinetics were monitored by the change in tryptophan capping domain of RalF proteins42 are autoinhibitory, while the
uorescence that follows the conformational change from Arf- PH domains of BRAG239,40 and EFA641 and the DCD-HUS
GDP to Arf-GTP (excitation and emission wavelengths of 292 domains of BIG122 are not. The concentrations of GEFs were
and 340 nm, respectively). In some experiments with BFA and selected to support eciencies at which inhibitory eects could
liposomes, kinetics were monitered using Forster resonance be accurately measured. Arf1 was used as a reference in all
energy transfer (FRET) between tryptophan and mGDP (292 experiments, except in the case of EFA6, which activates Arf1
and 440 nm, respectively), which gives equivalent kinetics.43 only in the presence of membranes,41 in which case 13Arf6
Fluorescence measurements were performed with a Cary was used for the solution assays. Inhibitors were used at xed
Eclipse uorimeter (Varian) at 37 C, under continuous concentrations at which signicant inhibition of at least one
stirring in a total volume of 800 L. Experiments in solution ArfGEF could be measured. All inhibitors were assayed for
used 1 M N-terminally truncated Arf-GDP and 100 nM nonspecic eects in two control experiments. First, eects
ArfGEF, except for RpRalF and LpRalF that were used at 400 outside the Arf/ArfGEF targets were assayed on the small
nM. Protein and inhibitors (or DMSO for controls) were GTPase Rac1 and its specic GEF Dock5 as described in ref 48.
incubated in HKM buer [50 mM HEPES (pH 7.4), 120 mM None exhibited a signicant eect on the Dock5-mediated Rac1
potassium acetate, 1 mM MgCl2, and 1 mM DTT] for 2 min at activation (Figure S2A). Second, nonspecic eects on
37 C before initiation of the reaction with 100 M GTP. liposomes were monitored by using DLS, which indicated
Exchange assays with liposomes were performed as described that none of the compounds induced aggregation or disruption
above using 0.4 M myrArf1 in the presence of 100 M of liposomes (Figure S2B).
liposomes. ArfGEF concentrations were 5 nM for Inhibition of Human and Bacterial ArfGEFs by BFA in
BRAG2Sec7PH, 10 nM for ARNOSec7PH, EFA6Sec7PH, and LpRalF, Solution and on Membranes. The natural fungal toxin BFA
and 100 nM for RpRalF and BIG1DHSec7. Exchange rates (kobs) is the best described ArfGEF inhibitor to date. BFA inhibits
were determined from monoexponential ts over the entire ArfGEFs by an uncompetitive mechanism,25 in which it
kinetics and expressed as a percentage of control activity. This stabilizes the Arf-GDP/ArfGEF complex in an abortive
approach was preferred to initial rate velocity analysis, which conformation that cannot proceed to nucleotide dissocia-
can be misestimated if compounds absorb light or have intrinsic tion.21,26 Its eciency toward eukaryotic ArfGEFs has been
uorescence. All exchange reactions were performed in at least quantied in solution assays, showing that it inhibits specically
triplicate, and means are given the standard deviation (SD). large Golgi ArfGEFs such as BIG1, while PH domain-
Dynamic Light Scattering Analysis (DLS). The distribu- containing ArfGEFs are resistant (for example, see ref 27).
tion of liposome sizes was analyzed by DLS using a DynaPro We extended this analysis to the Sec7 domains of Legionella and
NanoStarTM instrument (Wyatt Technology). Sets of 10 Rickettsia RalF (Figure 2A and Figure S3A). We observed that
autocorrelation curves were acquired at 37 C at the end of Rickettsia RalF is inhibited by BFA, whereas Legionella RalF is
each exchange assay in a disposable cuvette (Eppendorf). The resistant, which is consistent with the presence of a Tyr residue
mean radius was determined with DYNAMICS (Wyatt in the active site of the former and a Phe residue in the latter.
Technology) assuming that the size distribution is a simple Next, we determined the eciency of BFA toward human and
Gaussian function. bacterial ArfGEFs in the presence of liposomes (Figure 2B and
RESULTS
Strategy. Two well-established (BFA and SecinH3) and
Figure S3B). BIG1DHSec7 and full-length Rickettsia RalF were
strongly inhibited, consistent with the sensitivities of their Sec7
domains in solution, while BRAG2Sec7PH and ARNOSec7PH,
two recently discovered (M-COPA and NAV-2729) structurally whose Sec7 domains are resistant in solution, were only slightly
unrelated inhibitors were selected for this study (Figure 1A). inhibited. Activation of myrArf1 by EFA6Sec7PH was also
To analyze the eciency of the inhibitors, we used highly insensitive to BFA, extending previous observations that it
puried small GTPases and GEFs (Figure 1B,C) and measured does not inhibit the activation of myrARF6 by EFA6.51
their inhibition by uorescence kinetics, which reports Surprisingly, full-length LpRalF was strongly inhibited in this
accurately and quantitatively on nucleotide exchange inhib- assay, at odds with the resistance of its Sec7 domain. This was
ition.43 We selected representative guanine nucleotide exchange not an indirect eect due to autoinhibition because ARNO,
factors from all ArfGEF subfamilies, including a large Golgi which is autoinhibited by its PH domain, remained resistant to
ArfGEF, BIG1, three endocytic ArfGEFs, ARNO (a cytohesin), BFA on membranes. Thus, the presence of the membrane-
EFA6, and BRAG2, and two ArfGEFs from pathogenic bacteria, binding domain in LpRalF creates a sensitivity that is not
L. pneumophila and R. prowazekii RalF. We used dierent Arf encoded by its Sec7 domain alone. Together, this analysis
and ArfGEF constructs for the solution and membrane shows that BFA is a strong inhibitor of Golgi ArfGEFs and
experiments, because membrane-binding elements are inhib- Rickettsia RalF in solution and on membranes and uncovers an
itory in Arf and in several ArfGEFs. For experiments performed overlooked inhibition of Legionella RalF that could not be
in solution, we used the isolated Sec7 domain of the ArfGEFs detected with the Sec7 domain alone in solution.
and Arf constructs lacking their myristoylated N-terminal helix Inhibition of Human and Bacterial ArfGEFs by
(17Arf1 and 13Arf6, respectively, hereafter), which is SecinH3 in Solution and on Membranes. SecinH3 inhibits
autoinhibitory in solution and requires direct interaction with cytohesins in vitro29 and has been widely used as a proxy for
membranes for inhibition release.19,22 For experiments the involvement of cytohesins in cellular processes.13,36,52
5128 DOI: 10.1021/acs.biochem.7b00706
Biochemistry 2017, 56, 51255133
Biochemistry Article
Figure 2. Analysis of the inhibitory eciency of BFA (50 M) on Figure 3. Inhibition analysis of SecinH3 (5 M) in solution and on
human and bacterial GEFs. (A) Sensitivity of bacterial RalF proteins to membranes. (A) Sensitivity of human and bacterial ArfGEFs to
BFA in solution. kobs rates are expressed as a percentage of the rate SecinH3 in solution. All experiments were performed with 17Arf1,
obtained with DMSO. The right panel shows a representative except for the EFA6Sec7 experiment in which 13Arf6 was used. The
nucleotide kinetics curve. (B) Sensitivity of human and bacterial right panel shows a representative nucleotide kinetics curve. (B)
ArfGEFs to BFA with liposomes. All assays were performed with Sensitivity of human and bacterial ArfGEFs to SecinH3 with
myr
Arf1. The right panel shows a representative nucleotide kinetics liposomes. All assays were performed with myrArf1. The right panel
curve. kobs values are expressed as a percentage of the rate obtained shows a representative nucleotide kinetics curve. kobs values are
with DMSO. Values are the mean of triplicates SD. The associated expressed as described in the legend of Figure 2. The associated Figure
Figure S3 shows representative nucleotide kinetics curves for the S4 shows representative nucleotide kinetics curves for the experiments
experiments shown here. shown here.
Currently, its specicity toward other human ArfGEFs has been puried Arf GTPases and ArfGEFs has not been reported. A
only investigated for EFA6Sec7, which showed that this GEF is concentration of 16 M provided ecient inhibition of a
not a target of the drug in vitro.29 We analyzed its eciency on related Golgi ArfGEF, BIG1, in the solution GEF assay. M-
human and bacterial ArfGEFs in solution and with liposomes. COPA inhibited all Sec7 domains in solution at this
SecinH3 appeared to be insoluble at concentrations as low as concentration, with eciencies ranging from 20 to 60%
15 M, as judged by a noisy autouorescence signal of the inhibition (Figure 4A and Figure S5A). All ArfGEF constructs
compound alone that stabilized after prolonged stirring (not containing a membrane-binding domain were also inhibited in
shown); therefore, all experiments were performed at 5 M, a the presence of liposomes, with slightly higher eciencies (30
concentration close to its theoretical solubility where 80% inhibition) (Figure 4B and Figure S5B). There was no
autouorescence was no longer observed. At this concentration, clear correlation between the eciency of inhibition in solution
SecinH3 had a moderate inhibitory eect on ARNO in and with liposomes. For example, the eect of M-COPA was
solution, which was slightly higher with liposomes (Figure 3A strongest for BIG1 in solution and for LpRalF with liposomes
and Figure S4A). No inhibition was observed for other human and weakest for EFA6 in solution and for BRAG2 with
and bacterial GEFs, whether measurements were performed in liposomes. Our data thus indicate that M-COPA is a pan-
solution or in the presence of membranes (Figure 3B and ArfGEF inhibitor and that its inhibitory prole is dierent from
Figure S4B). Because ARNO activates all Arf isoforms, that of BFA, despite their similar patterns of inhibition of
including Arf1 and Arf6,46 we determined the eciency of cancer cell growth and dispersal of the Golgi.33,54 The activity
SecinH3 at blocking activation of Arf6. We observed that of M-COPA toward GBF1 could not be assessed directly,
SecinH3 inhibited 13Arf6 with an eciency slightly higher because this ArfGEF has resisted so far our attempts to produce
than that of 17Arf1 (Figure 3A). This analysis conrms that an active construct. However, a direct and potent inhibition of
SecinH3 inhibits cytohesins specically, and this eect does not GBF1 can be predicted from the inhibitory activity of M-COPA
depend on which Arf GTPase is the substrate, with the caveat on all ArfGEFs tested in our assay.
that the maximal inhibitory eect that could be measured (30% Inhibition of Human and Bacterial ArfGEFs by NAV-
of inhibition of ARNO toward 13Arf6 in solution or myrArf1 2729. NAV-2729 was recently described as an inhibitor of the
on membranes) was limited by its poor solubility. spontaneous activation of Arf6, and it also inhibited the
Inhibition of Human and Bacterial ArfGEFs by M- activation of Arf6 by its GEFs, ARNO and BRAG2, in
COPA in Solution and on Membranes. M-COPA inhibits solution.12 NAV-2729 was synthesized as described in
the growth of cancer cell lines with a prole that resembles that Supporting Information. We analyzed the inhibitory prole of
of BFA, and it causes the dispersal of the large Golgi ArfGEF NAV-2729 at a concentration of 25 M, which was reported to
GBF1, suggesting that this ArfGEF is a relevant mechanistic result in almost total inhibition of spontaneous and GEF-
target in cells.33,53,54 Direct analysis of this inhibitor toward stimulated Arf6 activation in vitro.12 Under the conditions used
5129 DOI: 10.1021/acs.biochem.7b00706
Biochemistry 2017, 56, 51255133
Biochemistry Article
Table 1. Family-wide Inhibition Proles of ArfGEF analysis indicates that inhibitors that target the Sec7 domain
Inhibitorsa can in general be assayed for specicity by using a simple setup
with soluble versions of the small GTPases and the isolated
BFA SecinH3 M-COPA NAV-2729
Sec7 domain of their GEFs. However, the inhibition by BFA of
sol. mb. sol. mb. sol. mb. sol. full-length LpRalF, but not its Sec7 domain alone, underlines
BRAG2 0b + 0 0 ++ ++ +++ the potential of small molecules to target the exchange
ARNO 0b + + ++ ++ +++ + mechanism as a whole by incorporating elements provided by
EFA6 0b 0 0 0 ++ + ++ appended regulatory domains.
BIG1 +++b +++ 0 0 +++ +++ + Caveats in the interpretation of the eciencies and
LpRalF 0 +++ + + ++ +++ + specicities of inhibitors can lead to misleading results.55
RpRalF +++ +++ 0 0 +++ +++ + First, most inhibitors except BFA are poorly soluble as
a
Percentages of inhibition were evaluated from kobs values determined predicted from their theoretical solubilities, which can lead to
without and with the inhibitor as indicated. Legend: sol., in solution; misinterpretation if the inhibitors are used at the concentrations
mb., in the presence of 100 M liposomes; 0, < 10% inhibition; +, 10 above full solubility. Direct measurement of the actual solubility
20% inhibition; ++, 2040% inhibition; +++, >40% inhibition. is dicult, and caution is required to avoid solubility issues.
b
Determined in ref 27. The percentages of inhibition can be found This was exemplied in our study of SecinH3, which exhibited
in Table S1. a bizarre autouorescence signal. This signal stabilized upon
prolonged stirring of the sample, pointing to a solubility issue
ArfGEFs with a higher eciency toward BRAG2. We note that that was resolved by decreasing its concentration from 15 to 5
the concentration at which we observed inhibition of ArfGEFs M, which is close to its predicted solubility. At this
by M-COPA is higher than those that resulted in growth concentration, specic interference with cytohesins but no
defects or dispersal of the Golgi, which were in the range of other ArfGEFs was observed, consistent with previous
0.11 M.33,54 A possible explanation is that M-COPA is more studies,29 although this was at the cost of maximal inhibition,
ecient on GBF1 than on any of the ArfGEFs included in this which did not exceed 30%. Another approach would be to
analysis. Alternatively, our observation that M-COPA is a monitor the inhibitory response at increasing inhibitor
potent inhibitor of other ArfGEFs raises the possibility that M- concentrations; a discontinuity with the appearance of a
COPA also aects the activity of ArfGEFs other than GBF1, plateau of inhibition can suggest that the solubility limit has
each with a moderate eciency but with eects that add up to been reached. Finally, analysis of the kinetics using initial
yield bioactivity. The observation that NAV-2729 also inhibits velocities or single points can incorrectly estimate the inhibitory
several ArfGEFS, with BRAG2 being the most aected, suggests eects for inhibitors that have intrinsic uorescence or
that it may also function by small and cumulative eects on absorbance at the wavelengths used to measure the kinetics.
several ArfGEFs to achieve bioactivity. Modeling the kinetics over the entire duration of the assay
All inhibitors were ecient at inhibiting the activation of (ideally to the plateau) alleviates such issues. With these
Arf1, which was used as a reference in most experiments. technical precautions in mind, it is desirable to measure the
Inhibitory eects were also measured on Arf6 for all inhibitors, eciencies and specicities of small molecules using puried
except for BFA. Notably, our family-wide analysis showed that proteins as an important initial step toward validating them as
NAV-2729 is a dual Arf1/Arf6 inhibitor and is more eective tools for cellular or chemical biology.
toward Arf1 than Arf6 is; this is in contrast to a previous report Mounting evidence indicates that Arf GTPases and their
that found that it targets only Arf6.12 On the basis of these GEFs are pivotal regulatory nodes in major physiological
observations, it is likely that the other Arf family members functions such as insulin signaling,29,36 vascular stability,13 or
(Arf3Arf5) not assayed in this study are also inhibited by lipolysis11 and in pathologies such as cancers1012 and
these small molecules with similar ArfGEF specicity proles. infections.1416 Small molecule inhibitors are therefore urgently
Because membranes are integral components of the needed to interrogate the biology of these pathways,34 and their
activation of Arf GTPases by their GEFs and productive accurate biochemical characterization is an important step
interaction between Arf GTPases and GEFs takes place on toward the identication of which target(s) they engage to yield
membranes,22 we also compared the eciencies of the bioactivity.56 Our family-wide inhibitory proles of ArfGEF
inhibitors using lipidated Arf GTPases and ArfGEF constructs inhibitors will provide increased accuracy for future analysis of
comprising a membrane-binding domain (Table 1 and Table molecular pathways involved in diseases and help attain
S1). Surprisingly, membranes had in general a moderate eect therapeutic goals. Ultimately, this knowledge will feed the
on the eciency and selectivity of inhibition. A notable development of drugs that modulate Arf and ArfGEF
exception was Legionella RalF. The Sec7 domain of Legionella interactions and of other biotechnology applications.57,58
RalF was resistant to BFA in solution, as is expected from the
presence of a discriminating Phe residue in its active site. In
contrast, the full-length protein, whose activity can be detected
*
ASSOCIATED CONTENT
S Supporting Information
only in the presence of membranes,42 was strongly inhibited by The Supporting Information is available free of charge on the
BFA. On the basis of the presence of aromatic residues in the ACS Publications website at DOI: 10.1021/acs.bio-
inhibitory capping domain that mimic equivalent residues of chem.7b00706.
Arf, a plausible model is one in which BFA binds at the
interface between the Sec7 and capping domains to stabilize Chemical synthesis of NAV-2729; analysis of nonspecic
RalF in an autoinhibited conformation. Thus, inhibition of eects; representative kinetics curves for the experiments
LpRalF by BFA would be an intramolecular variation of the with BFA, SecinH3, M-COPA, and NAV-2729; and a
interfacial mechanism whereby BFA inhibits Golgi ArfGEFs by table of percentages of inhibition (PDF)
trapping an abortive Arf/Sec7 complex.21 Together, our SMILES molecular formula strings (CSV)
5131 DOI: 10.1021/acs.biochem.7b00706
Biochemistry 2017, 56, 51255133
Biochemistry Article
AUTHOR INFORMATION
Corresponding Authors
(9) Donaldson, J. G., and Jackson, C. L. (2011) ARF family G
proteins and their regulators: roles in membrane transport, develop-
ment and disease. Nat. Rev. Mol. Cell Biol. 12, 362375.
*E-mail: jacqueline.cherls@ens-paris-saclay.fr. (10) Morishige, M., Hashimoto, S., Ogawa, E., Toda, Y., Kotani, H.,
*E-mail: mahel.zeghouf@ens-paris-saclay.fr. Hirose, M., Wei, S., Hashimoto, A., Yamada, A., Yano, H., Mazaki, Y.,
ORCID Kodama, H., Nio, Y., Manabe, T., Wada, H., Kobayashi, H., and Sabe,
Tatiana Caneque: 0000-0002-1110-0643 H. (2008) GEP100 links epidermal growth factor receptor signalling
to Arf6 activation to induce breast cancer invasion. Nat. Cell Biol. 10,
Isamu Shiina: 0000-0002-8749-2540 8592.
Mahel Zeghouf: 0000-0003-3686-2510 (11) Singh, S. R., Zeng, X., Zhao, J., Liu, Y., Hou, G., Liu, H., and
Present Address Hou, S. X. (2016) The lipolysis pathway sustains normal and
@ transformed stem cells in adult Drosophila. Nature 538, 109113.
T.Y.: Pharmaceuticals and Medical Devices Agency, Shin-
Kasumigaseki Building, Tokyo 100-0013, Japan. (12) Yoo, J. H., Shi, D. S., Grossmann, A. H., Sorensen, L. K., Tong,
Z., Mleynek, T. M., Rogers, A., Zhu, W., Richards, J. R., Winter, J. M.,
Author Contributions Zhu, J., Dunn, C., Bajji, A., Shenderovich, M., Mueller, A. L.,
S.B. and F.P. contributed equally to this work. Woodman, S. E., Harbour, J. W., Thomas, K. R., Odelberg, S. J.,
Funding Ostanin, K., and Li, D. Y. (2016) ARF6 Is an Actionable Node that
This work was supported by grants from the Fondation pour la Orchestrates Oncogenic GNAQ Signaling in Uveal Melanoma. Cancer
Recherche Medicale, the Institut National du Cancer, and the Cell 29, 889904.
(13) Zhu, W., London, N. R., Gibson, C. C., Davis, C. T., Tong, Z.,
CNRS to J. Cherls and from the Institut National du Cancer
Sorensen, L. K., Shi, D. S., Guo, J., Smith, M. C., Grossmann, A. H.,
to R.R. Thomas, K. R., and Li, D. Y. (2012) Interleukin receptor activates a
Notes MYD88-ARNO-ARF6 cascade to disrupt vascular stability. Nature 492,
The authors declare no competing nancial interest. 252255.
(14) Nagai, H., Kagan, J. C., Zhu, X., Kahn, R. A., and Roy, C. R.
ACKNOWLEDGMENTS (2002) A bacterial guanine nucleotide exchange factor activates ARF
on Legionella phagosomes. Science 295, 679682.
The authors are grateful to Lionel Duarte, Yann Ferrandez, (15) Selyunin, A. S., Sutton, S. E., Weigele, B. A., Reddick, L. E.,
Alexandra Joubert, and Lurlene Akendengue (ENS Paris- Orchard, R. C., Bresson, S. M., Tomchick, D. R., and Alto, N. M.
Saclay) for protein purications. The authors thank TOCRIS (2011) The assembly of a GTPase-kinase signalling complex by a
Bioscience for performing quality controls on SecinH3. bacterial catalytic scaffold. Nature 469, 107111.
ABBREVIATIONS
BFA, Brefeldin A; DCB-HUS, dimerization and cyclophilin-
(16) Wesolowski, J., Weber, M. M., Nawrotek, A., Dooley, C. A.,
Calderon, M., St, St. Croix, C. M., Hackstadt, T., Cherfils, J., and
Paumet, F. (2017) Chlamydia Hijacks ARF GTPases To Coordinate
Microtubule Posttranslational Modifications and Golgi Complex
binding homology upstream of sec7 domain; DLS, dynamic Positioning. mBio 8, e02280-16.
light scattering; DMSO, dimethyl sulfoxide; EDTA, ethyl- (17) Cox, R., Mason-Gamer, R. J., Jackson, C. L., and Segev, N.
enediaminetetraacetic acid; FRET, Forster resonance energy (2004) Phylogenetic analysis of Sec7-domain-containing Arf nucleo-
tranfer; GAP, GTPase-activating protein; GDP, guanosine tide exchangers. Mol. Biol. Cell 15, 14871505.
diphosphate; GEF, guanine nucleotide exchange factor; GTP, (18) Casanova, J. E. (2007) Regulation of Arf activation: the Sec7
guanosine triphosphate; IC50, half-maximal inhibitory concen- family of guanine nucleotide exchange factors. Traffic 8, 14761485.
tration; LpRalF, L. pneumophila RalF protein; PC, phosphati- (19) Pasqualato, S., Renault, L., and Cherfils, J. (2002) Arf, Arl, Arp
dylcholine; PE, phosphatidylethanolamine; PH, pleckstrin and Sar proteins: a family of GTP-binding proteins with a structural
homology; PI(4,5)P2, phosphatidylinositol 4,5-bisphosphate; device for front-back communication. EMBO Rep. 3, 10351041.
(20) Goldberg, J. (1998) Structural basis for activation of ARF
PS, phosphatidylserine; RpRalF, R. prowazekii RalF protein;
GTPase: mechanisms of guanine nucleotide exchange and GTP-
TGN, trans-Golgi network; SDSPAGE, sodium dodecyl myristoyl switching. Cell 95, 237248.
sulfatepolyacrylamide gel electrophoresis.
inhibitor brefeldin a for arf proteins and sec7 domains. J. Biol. Chem. (45) Beraud-Dufour, S., Robineau, S., Chardin, P., Paris, S., Chabre,
281, 1180511814. M., Cherfils, J., and Antonny, B. (1998) A glutamic finger in the
(28) Viaud, J., Zeghouf, M., Barelli, H., Zeeh, J. C., Padilla, A., guanine nucleotide exchange factor ARNO displaces Mg2+ and the
Guibert, B., Chardin, P., Royer, C. A., Cherfils, J., and Chavanieu, A. beta-phosphate to destabilize GDP on ARF1. EMBO J. 17, 3651
(2007) Structure-based discovery of an inhibitor of Arf activation by 3659.
Sec7 domains through targeting of protein-protein complexes. Proc. (46) Peurois, F., Veyron, S., Ferrandez, Y., Ladid, I., Benabdi, S.,
Natl. Acad. Sci. U. S. A. 104, 1037010375. Zeghouf, M., Peyroche, G., and Cherfils, J. (2017) Characterization of
(29) Hafner, M., Schmitz, A., Grune, I., Srivatsan, S. G., Paul, B., the activation of small GTPases by their GEFs on membranes using
Kolanus, W., Quast, T., Kremmer, E., Bauer, I., and Famulok, M. artificial membrane tethering. Biochem. J. 474, 12591272.
(2006) Inhibition of cytohesins by SecinH3 leads to hepatic insulin (47) Folly-Klan, M., Sancerne, B., Alix, E., Roy, C. R., Cherfils, J., and
resistance. Nature 444, 941944. Campanacci, V. (2015) On the use of Legionella/Rickettsia chimeras
(30) Saenz, J. B., Sun, W. J., Chang, J. W., Li, J., Bursulaya, B., Gray, to investigate the structure and regulation of Rickettsia effector RalF. J.
N. S., and Haslam, D. B. (2009) Golgicide A reveals essential roles for Struct. Biol. 189, 98104.
GBF1 in Golgi assembly and function. Nat. Chem. Biol. 5, 157165. (48) Vives, V., Cres, G., Richard, C., Busson, M., Ferrandez, Y.,
(31) Boal, F., Guetzoyan, L., Sessions, R. B., Zeghouf, M., Spooner, Planson, A. G., Zeghouf, M., Cherfils, J., Malaval, L., and Blangy, A.
R. A., Lord, J. M., Cherfils, J., Clarkson, G. J., Roberts, L. M., and (2015) Pharmacological inhibition of Dock5 prevents osteolysis by
Stephens, D. J. (2010) LG186: An inhibitor of GBF1 function that affecting osteoclast podosome organization while preserving bone
causes Golgi disassembly in human and canine cells. Traffic 11, 1537 formation. Nat. Commun. 6, 6218.
1551. (49) Ramaen, O., Joubert, A., Simister, P., Belgareh-Touze, N.,
(32) Feng, Y., Jadhav, A. P., Rodighiero, C., Fujinaga, Y., Olivares-Sanchez, M. C., Zeeh, J. C., Chantalat, S., Golinelli-Cohen, M.
Kirchhausen, T., and Lencer, W. I. (2004) Retrograde transport of P., Jackson, C. L., Biou, V., and Cherfils, J. (2007) Interactions
cholera toxin from the plasma membrane to the endoplasmic between conserved domains within homodimers in the BIG1, BIG2,
reticulum requires the trans-Golgi network but not the Golgi apparatus and GBF1 Arf guanine nucleotide exchange factors. J. Biol. Chem. 282,
in Exo2-treated cells. EMBO Rep. 5, 596601. 2883428842.
(33) Ohashi, Y., Iijima, H., Yamaotsu, N., Yamazaki, K., Sato, S., (50) DiNitto, J. P., Delprato, A., Gabe Lee, M. T., Cronin, T. C.,
Okamura, M., Sugimoto, K., Dan, S., Hirono, S., and Yamori, T. Huang, S., Guilherme, A., Czech, M. P., and Lambright, D. G. (2007)
(2012) AMF-26, a novel inhibitor of the Golgi system, targeting ADP- Structural basis and mechanism of autoregulation in 3-phosphoinosi-
ribosylation factor 1 (Arf1) with potential for cancer therapy. J. Biol. tide-dependent Grp1 family Arf GTPase exchange factors. Mol. Cell 28,
Chem. 287, 38853897. 569583.
(34) Mishev, K., Dejonghe, W., and Russinova, E. (2013) Small (51) Franco, M., Peters, P. J., Boretto, J., van Donselaar, E., Neri, A.,
molecules for dissecting endomembrane trafficking: a cross-systems DSouza-Schorey, C., and Chavrier, P. (1999) EFA6, a sec7 domain-
view. Chem. Biol. 20, 475486. containing exchange factor for ARF6, coordinates membrane recycling
(35) Jackson, C. L. (2002) Brefeldin A revealing the fundamental and actin cytoskeleton organization. EMBO J. 18, 14801491.
principles governing membrane dynamics and protein transport. (52) Li, J., Malaby, A. W., Famulok, M., Sabe, H., Lambright, D. G.,
Subcell. Biochem. 34, 233272. and Hsu, V. W. (2012) Grp1 plays a key role in linking insulin
(36) Fuss, B., Becker, T., Zinke, I., and Hoch, M. (2006) The signaling to glut4 recycling. Dev. Cell 22, 12861298.
cytohesin Steppke is essential for insulin signalling in Drosophila. (53) Shiina, I., Umezaki, Y., Ohashi, Y., Yamazaki, Y., Dan, S., and
Nature 444, 945948. Yamori, T. (2013) Total synthesis of AMF-26, an antitumor agent for
(37) Schlienger, S., Ramirez, R. A., and Claing, A. (2015) ARF1 inhibition of the Golgi system, targeting ADP-ribosylation factor 1. J.
regulates adhesion of MDA-MB-231 invasive breast cancer cells Med. Chem. 56, 150159.
through formation of focal adhesions. Cell. Signalling 27, 403415. (54) Ignashkova, T. I., Gendarme, M., Peschk, K., Eggenweiler, H.
(38) Ohashi, Y., Okamura, M., Hirosawa, A., Tamaki, N., Akatsuka, M., Lindemann, R. K., and Reiling, J. H. (2017) Cell survival and
A., Wu, K. M., Choi, H. W., Yoshimatsu, K., Shiina, I., Yamori, T., and protein secretion associated with Golgi integrity in response to Golgi
Dan, S. (2016) M-COPA, a Golgi Disruptor, Inhibits Cell Surface stress-inducing agents. Traffic 18, 530544.
Expression of MET Protein and Exhibits Antitumor Activity against (55) Arrowsmith, C. H., Audia, J. E., Austin, C., Baell, J., Bennett, J.,
MET-Addicted Gastric Cancers. Cancer Res. 76, 38953903. Blagg, J., Bountra, C., Brennan, P. E., Brown, P. J., Bunnage, M. E.,
(39) Aizel, K., Biou, V., Navaza, J., Duarte, L. V., Campanacci, V., Buser-Doepner, C., Campbell, R. M., Carter, A. J., Cohen, P.,
Cherfils, J., and Zeghouf, M. (2013) Integrated conformational and Copeland, R. A., Cravatt, B., Dahlin, J. L., Dhanak, D., Edwards, A. M.,
lipid-sensing regulation of endosomal ArfGEF BRAG2. PLoS Biol. 11, Frederiksen, M., Frye, S. V., Gray, N., Grimshaw, C. E., Hepworth, D.,
e1001652. Howe, T., Huber, K. V., Jin, J., Knapp, S., Kotz, J. D., Kruger, R. G.,
(40) Jian, X., Gruschus, J. M., Sztul, E., and Randazzo, P. A. (2012) Lowe, D., Mader, M. M., Marsden, B., Mueller-Fahrnow, A., Muller, S.,
The pleckstrin homology (PH) domain of the Arf exchange factor OHagan, R. C., Overington, J. P., Owen, D. R., Rosenberg, S. H.,
Roth, B., Ross, R., Schapira, M., Schreiber, S. L., Shoichet, B.,
Brag2 is an allosteric binding site. J. Biol. Chem. 287, 2427324283.
Sundstrom, M., Superti-Furga, G., Taunton, J., Toledo-Sherman, L.,
(41) Padovani, D., Folly-Klan, M., Labarde, A., Boulakirba, S.,
Walpole, C., Walters, M. A., Willson, T. M., Workman, P., Young, R.
Campanacci, V., Franco, M., Zeghouf, M., and Cherfils, J. (2014)
N., and Zuercher, W. J. (2015) The promise and peril of chemical
EFA6 controls Arf1 and Arf6 activation through a negative feedback
probes. Nat. Chem. Biol. 11, 536541.
loop. Proc. Natl. Acad. Sci. U. S. A. 111, 1237812383.
(56) Schurmann, M., Janning, P., Ziegler, S., and Waldmann, H.
(42) Folly-Klan, M., Alix, E., Stalder, D., Ray, P., Duarte, L. V.,
(2016) Small-Molecule Target Engagement in Cells. Cell Chem. Biol.
Delprato, A., Zeghouf, M., Antonny, B., Campanacci, V., Roy, C. R.,
23, 435441.
and Cherfils, J. (2013) A novel membrane sensor controls the
(57) Zhu, Y., Xiao, T., Lei, S., Zhou, F., and Wang, M. W. (2015)
localization and ArfGEF activity of bacterial RalF. PLoS Pathog. 9,
Application of chemical biology in target identification and drug
e1003747.
discovery. Arch. Pharmacal Res. 38, 16421650.
(43) Zeeh, J. C., Antonny, B., Cherfils, J., and Zeghouf, M. (2008) In
(58) Milroy, L. G., Grossmann, T. N., Hennig, S., Brunsveld, L., and
vitro assays to characterize inhibitors of the activation of small G Ottmann, C. (2014) Modulators of protein-protein interactions. Chem.
proteins by their guanine nucleotide exchange factors. Methods Rev. 114, 46954748.
Enzymol. 438, 4156.
(44) Franco, M., Chardin, P., Chabre, M., and Paris, S. (1995)
Myristoylation of ADP-ribosylation factor 1 facilitates nucleotide
exchange at physiological Mg2+ levels. J. Biol. Chem. 270, 13371341.