Membranes prime the RapGEF EPAC1 to transduce cAMP signaling

doi:10.21203/rs.3.rs-2608569/v1

Download PDF

Article

Membranes prime the RapGEF EPAC1 to transduce cAMP signaling

https://doi.org/10.21203/rs.3.rs-2608569/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 12 Jul, 2023

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

EPAC1, a cAMP-activated GEF for Rap GTPases, is a major physiological transducer of cAMP signaling and a therapeutic target in cardiac diseases. The recent discovery that cAMP is compartmentalized in membrane-proximal nanodomains challenged the current model of EPAC1 activation in the cytosol. Here, we discover that anionic membranes are a major component of EPAC1 activation. We find that anionic membranes activate EPAC1 in the absence of cAMP, that they increase its affinity for cAMP by 2 orders of magnitude, and that they synergize with cAMP to yield maximal GEF activity. Thus, EPAC1 has four states that differ in their affinity for cAMP and GEF efficiency. In the cell cytosol, where cAMP is low, EPAC1 must thus be primed by membranes to bind cAMP. Examination of the cardiomyocyte-active chemical CE3F4 in this new framework further reveals that inhibition affects only membrane- and cAMP-activated EPAC1. Together, our findings reformulate previous concepts of cAMP signaling to include a hitherto overlooked role of membranes through EPAC proteins, with important implications for drug discovery.

Biological sciences/Biochemistry/Proteins/GTP-binding protein regulators

Biological sciences/Chemical biology/Membranes

The second messenger cyclic adenosine monophosphate (cAMP) transduces extracellular signals conveyed by hormones and neurotransmitters through G-protein coupled receptors and heterotrimeric G proteins. In cells, the opposite activities of G protein-regulated adenylyl cyclases, which synthesize cAMP, and phosphodiesterases, which degrade it, result in cAMP fluxes that regulate a broad array of major intracellular pathways in physiological and pathological conditions (reviewed in ^{1, 2}). While cAMP has long been viewed as a freely diffusible molecule, it is now widely recognized that it is instead segregated into membrane-proximal nanometer-sized compartments, thus maintaining its concentration low in the bulk cytosol and warranting the specificity of the cellular responses (^{3, 4, 5, 6}; reviewed in ⁷). When its concentration increases, cAMP binds to, and activates, specific targets, of which protein kinase A (PKA) is the best characterized (reviewed in ⁸). Another major transducer of cAMP signaling is the EPAC protein family, comprised of EPAC1 and EPAC2. EPAC proteins function as guanine nucleotide exchange factors (GEFs) to activate small GTPases of the Ras-related Rap family independently of PKA ^{9, 10}. In human, EPAC proteins regulate essential functions, such as integrin-mediated cell adhesion and cell–cell junction formation, and control important physiological processes such as insulin secretion, neurotransmitter release and cardiac functions (reviewed in ^{11, 12, 13}). Animal and clinical studies have implicated EPAC1 and/or EPAC2 in diverse human diseases, notably cancer (reviewed in ¹⁴), chronic inflammatory pain ¹⁵, pathological angiogenesis ¹⁶ and heart diseases (reviewed in ^{12, 17}). EPAC proteins are therefore highly studied as pharmacological targets for therapeutic intervention (reviewed in ^{13, 18}).

An important step towards targeting EPAC proteins by small molecules in diseases is to understand the molecular basis for their regulation. Seminal biochemical and structural studies uncovered how cAMP controls the activation of Rap GTPases by EPAC proteins ^{10, 19, 20}. In vitro, EPAC1 and EPAC2 are strongly autoinhibited in the absence of cAMP ¹⁰. The crystal structure of full-length, autoinhibited EPAC2 showed that its N-terminal regulatory region, which is comprised of a cyclic nucleotide-binding (CNB) domain, a Dishevelled / Egl-10 / Pleckstrin (DEP) domain and a second CNB domain, hinders access to the Rap-binding site located in the C-terminal GEF domain ¹⁹. No structure of an activated full-length EPAC protein is currently known. However, the crystal structure of a truncated EPAC2 construct bound to a cAMP analog and Rap1 showed that cAMP locks an alternative CNB-GEF intramolecular interface, thereby removing the CNB domain from GEF active site through a large rigid-body movement ²⁰. EPAC1 has the same domain organization as EPAC2, apart from the replacement of the first CNB domain by a 50-residue domain without recognizable structural homology, suggesting that its mechanism of autoinhibition release shares common features with that of EPAC2.

Biochemical analyses suggest however that the regulation of EPAC proteins by cAMP is more complex than a simple binary conformational switch. Notably, the dissociation constant of EPAC1 for cAMP is in the 3-4 mM range ^{21, 22, 23, 24} and half maximal activation is achieved in the 30-45 mM range ^{23, 25, 26}. Such concentrations are considerably higher than those reported for PKA ^{24, 25}, and they may not be attained in the bulk cytosol where inactive EPAC proteins are located. Furthermore, cAMP is unable to yield full activation of EPAC1 in vitro even at saturating concentration, as shown by its comparison to the EPAC-specific cAMP analog 8-(4-chlorophenylthio)-2’-O-methyl-cAMP (8CPT-2Me-cAMP, also called 007), which yields a maximal activity three times higher than that achieved by cAMP ^{25, 26, 27, 28}. Thus, EPAC proteins may have activation determinants other than cAMP, which remain to be identified.

The pivotal role of membranes in regulating the function of small GTPases (reviewed in ²⁹) and their regulators (reviewed in ^{30, 31}) is currently gaining center stage in small GTPase biology. Like most, if not all, small GTPase/GEF systems, EPAC and Rap proteins actuate their functions at the surface of membranes. Rap GTPases associate reversibly to membranes through their prenylated C-terminal extension ^{32, 33, 34, 35}. Likewise, ectopically expressed EPAC1 translocates from the cytosol to the plasma membrane upon increase in cAMP levels ^{36, 37}. How EPAC1 is recruited to the membranes has however remained poorly understood. Mutagenesis studies of EPAC1 in vitro and in cells suggested that it binds directly to the membrane through a polybasic loop of the DEP domain, a domain often involved in membrane binding, and binding was proposed to be specific for phosphatidic acid lipids (PA) ^{37, 38}. Hydrogen/deuterium exchange coupled to mass spectrometry (HDX-MS) of EPAC2 suggested that this loop becomes solvent-exposed upon activation by cAMP ³⁹, yet an isolated DEP domain, in which this loop is exposed, did not locate to the plasma membrane ³⁶. How activation by cAMP and membrane recruitment are coupled thus remains to be elucidated.

Small molecules that interfere with complex regulatory mechanisms can bring decisive insight into elusive functional determinants, as exemplified by Brefeldin A, a natural compound that highlighted the mechanism of activation of Arf GTPases by their GEFs at the interface with membranes ⁴⁰. Several agonists and inhibitors of EPAC1 and EPAC2 have been described (reviewed in ^{13, 18, 41, 42, 43}). In that regard, the EPAC1-specific inhibitor CE3F4, is especially interesting for its intriguing inhibitory properties. CE3F4 was discovered from a fluorescence-based GEF screen using a recombinant EPAC1 construct containing the CNB and GEF domains ⁴⁴. It efficiently inhibits the activation of the Rap1 GTPase in cardiomyocytes ⁴⁴ and prevents arrhythmia in mice ⁴⁵ and rat ⁴⁶, thus showing promising potential for drug development in cardiac pathologies. Surprisingly, CE3F4 does not inhibit the GEF domain of EPAC1 alone ⁴⁴, and NMR studies showed that it binds to the isolated CNB domain without competing with cAMP⁴⁷. Thus, CE3F4 appears to act through an allosteric mechanism with respect to both cAMP and Rap GTPases, yet to be elucidated, which makes it an appealing tool to interrogate the mechanism of EPAC1 activation.

In this study, we characterized the interplay between cAMP and membranes in the activation of EPAC1 by combining reconstitutions of purified proteins in artificial membranes, quantitative fluorescence kinetics, structural analysis by crystallography, small angle X-ray scattering (SAXS) and HDX-MS and inhibition studies using CE3F4. Our data reveal a major role of membranes in cAMP and EPAC signaling, with important implications for the activation of EPAC1 in cells, for cAMP signaling and for drug discovery.

Membranes activate EPAC1 in the absence and presence of cAMP

To determine whether membranes affect EPAC1 activity directly, we used purified full-length EPAC1 (EPAC1^FL hereafter) (Figures S1A and S1B) and liposomes of controlled lipid composition. First, we characterized the interaction of EPAC1^FL with membranes, using a liposome flotation assay (Figure 1A). EPAC1^FL did not bind to liposomes containing only PC and PE lipids (neutral liposomes hereafter). In contrast, strong binding to liposomes containing phosphatidylserine (PS) and phosphatidylinositol-4,5-bisphosphate (PIP₂) was observed. Binding was slightly increased by enrichment of PS-PIP₂ liposomes with cholesterol, while addition of cAMP in the flotation assay had no discernible effect. Next, we characterized the interaction of EPAC1 with phosphatidic acid (PA), which has been reported to drive the localization of EPAC1 to the plasma membrane ³⁷, and with cardiolipin (CL), which is enriched in cardiomyocyte mitochondria (reviewed in ⁴⁸), an organelle where a subpopulation of EPAC1 has been reported to carry out important functions ^{49, 50}, using PC-PE liposomes enriched with 10% PA (PA-liposomes hereafter) or 10% CL (CL-liposomes hereafter) (Figure 1B). Surprisingly, we observed no increase in EPAC1^FL binding to PA- or CL-liposomes compared to neutral liposomes. Likewise, addition of cAMP did not increase binding of EPAC1^FL to PA- or CL- liposomes compared to neutral liposomes.

Next, we analyzed whether binding of EPAC1 to membranes modulates its GEF activity towards the small GTPase Rap1, using liposomes containing PS, PIP₂ and cholesterol (plasma membrane (PM) liposomes hereafter) to which EPAC1^FL binds strongly. The efficiency of EPAC1-stimulated GDP/GTP exchange was monitored by following the kinetics of dissociation of the fluorescent nucleotide BODIPY-GDP from Rap1. We used non-lipidated full-length Rap1A (Rap1 hereafter), which binds to PM liposomes (Figure S1C) likely through its polybasic C-terminal region (¹⁶⁷RKTPVEKKKPKKKSCLLL¹⁸⁴), thus allowing to reconstitute the membrane-associated reaction. k_obs were measured over a range of EPAC1 concentrations and used to determine catalytic efficiencies (k_cat/K_M, expressed in s^-1.M^-1) (Figures 1C and S1D). In solution, EPAC1^FL had no GEF activity towards Rap1 in the absence of cAMP. Addition of cAMP supported modest GEF activity (k_cat/K_M = 1.5x10⁴ s^-1.M^-1), thus confirming that EPAC1^FL autoinhibition is released by cAMP. Unexpectedly, PM liposomes alone were equally efficient at activating EPAC1^FL as cAMP alone (k_cat/K_M = 1.2x10⁴ s^-1.M^-1). Remarkably, addition of both cAMP and PM liposomes resulted in a marked increase in GEF efficiency as compared to either cAMP or PM liposomes alone (k_cat/K_M = 8.8x10⁴ s^-1.M^-1). PA-liposomes (Figures 1D and S1E) and CL-liposomes (Figures 1E and S1F) showed no increase in activity compared to neutral liposomes whether with or without cAMP, consistent with the lack of interaction of EPAC1^FL with these liposomes.

We conclude from these observations that EPAC1 binds strongly to PM liposomes but does not recognize PA or CL lipids directly, and that cAMP and PM liposomes synergize to yield maximal EPAC1 activity.

The affinity of EPAC1 for cAMP is dramatically increased by membranes

The synergy between cAMP and membranes suggests that membranes could affect how cAMP binds to EPAC1. We determined the apparent dissociation constant (K_D) of cAMP in solution and in the presence of PM liposomes (Figures 2A and S2A), by monitoring the kinetics of Rap1 activation by EPAC1^FL as a function of cAMP concentration. The apparent K_D was determined as the concentration of cAMP yielding half maximum activity, as previously described for EPAC proteins in solution ^{26, 28}. In solution, the apparent K_D was 28 mM, a value that is even higher than the mediocre K_D determined previously for truncated EPAC1 constructs using this and other methods ^{21, 26, 28}. Remarkably, the apparent K_Ddecreasedto 74 nM in the presence of PM liposomes, representing a ~378-fold increase in affinity. PM liposomes also increased the maximal activity (k_max) by ~5-fold, from 0.24 10^-2s^-1 in solution to 1.26 10^-2s^-1 (Figures 2B and S2A).

The widely used cAMP super agonist 007 has been reported to increase k_max by 3-fold in solution compared to cAMP ²⁵, which is in the same range as the increase of k_max promoted by membranes in the presence of cAMP. We therefore asked whether activation of EPAC1^FL by 007 remains sensitive to membranes, using the same approach as above (Figures 2C, 2D and S2B). The apparent K_D of 007 in solution was ~18-fold lower than that of cAMP (1.5 mM), a value that is in the same range as previously reported ^{26, 28}. Surpringly, PM liposomes decreased the K_D of 007 to 14 nM, representing a ~110-fold increase, and an affinity about 5 times higher than that of cAMP. PM liposomes also increased k_max from 0.25 10^-2s^-1 in solution to 3.16 10^-2s^-1, representing a ~13-fold increase. Thus, activation of EPAC1 by 007 is strongly potentiated the presence of membrane.

We conclude that anionic liposomes strongly potentiate the capability of cAMP to activate EPAC1, and that both the K_D and k_max of 007 surpass those of cAMP in the context of membranes.

Structural determinants of EPAC1 binding to membranes

The DEP domain is the only canonical membrane-binding domain in EPAC1 (reviewed in ³⁸). To gain insight into the membrane-binding site of EPAC1, we first determined the crystal structure of a construct encompassing the DEP and CNB domains (EPAC1^DEP-CNB,residues 50-318) in complex with cAMP at 2.3 Å resolution (Figure 3A, crystallographic statistics in Table S1). The structure shows that the DEP and CNB domains are connected by a kinked helix (residues 154-181, with the kink located at P171), which is sandwiched between the two domains and mediates entirely their interaction. This organization is similar to that seen in the structure of unbound EPAC2^CNB-DEP-CNB ⁵¹ and of autoinhibited full-length EPAC2 ¹⁹, with only a small difference in the relative orientations of the DEP and CNB domains (Figure S3A). This similarity, together with the fact that the two copies of EPAC1^DEP-CNB in the crystallographic asymmetric unit are identical, suggest that the DEP and CNB domains are rigidly associated and behave as a single unit in the course of EPAC activation. The structure also shows that cAMP binds to the CNB domain of EPAC1^DEP-CNB in the same manner as in the EPAC2^CNB-GEF-Rap1 complex ²⁰, suggesting that interlocking of the CNB and GEF domains by cAMP does not require further rearrangement of the CNB domain. We note a single difference between EPAC1 and EPAC2, located at Gln 270 in EPAC1 which forms a hydrogen bond with the sugar hydroxyl of cAMP that is not seen with the equivalent Lys 405 in EPAC2 (Figure S3B). The potential to form hydrogen bonds at this residue likely contributes to the distinct selectivities of EPAC1 and EPAC2 for cAMP-derived agonists ²⁸.

The DEP domain does not bind to the plasma membrane on its own ³⁶and the structure confirms that it lacks a positively charged pocket that could accommodate PS or PIP₂ headgroups, such as the phosphoinositide-binding pocket of PH domains. Alternatively, EPAC1 could use more than just the DEP domain to interact with membranes, possibly through positively charged residues distributed throughout its surface. To address this question, we compared the binding of various purified EPAC1 constructs (EPAC1^DEP, EPAC1^DEP-CNB, EPAC1^Nt-DEP, EPAC1^CNB-GEF, see Figures S1A and S1B) to PM liposomes (Figure 3B and S3C). EPAC1^DEP did not bind to liposomes on its own, consistent with previous cellular assays ³⁶. Importantly, none of the truncated constructs bound as strongly as EPAC1^FL, suggesting that all domains contribute to optimal EPAC1 binding to anionic membranes. Combining our EPAC1^DEP-CNB structure to a model of EPAC1^CNB-GEF/Rap1, we then built a composite model of cAMP-activated EPAC1^DEP-CNB-GEF bound to Rap1 and used it to examine how EPAC1 binds to anionic membranes. We identified an extended cationic tract at the surface in the EPAC1^DEP-CNB-GEF model, comprised of the polybasic loop in the DEP domain (residues ⁷⁵RDRKYHLRLYRQ⁸⁶), and numerous lysines and arginines in the CNB and GEF domains (Figure 3C). Accordingly, docking of the EPAC1 model onto a membrane using the OPM/PPM server ⁵² predicted that EPAC1 uses this entire surface to bind to the membrane (Figure S3D). This suggests that EPAC1 uses an extended cationic surface to bind to anionic membranes through multiple non-specific electrostatic interactions contributed by both its regulatory and catalytic regions. In this model EPAC1 can accommodate Rap1 in a position where its polybasic, lipidated C-terminus points towards the membrane, indicating that this orientation is competent for Rap1 activation on the membrane.

To gain further insight into the organization of EPAC1 on membranes, we used HDX-MS, which allows to map simultaneously conformational changes associated with activation and interactions with membranes (⁵³, reviewed in ^{54, 55}), to compare inactive EPAC1 in solution to fully active cAMP- and PM liposome-bound EPAC1 (Figures 3D and S4A-C). Most protected and deprotected peptides are consistent with the model of active EPAC1 bound to the membrane. Strong interlocking of the CNB and GEF domains by cAMP is readily observed by protection of the cAMP binding site (peptide 251-282 in the CNB domain and peptide 323-347 in the switchboard), and by deprotection of the hinge between the CNB and GEF domain (peptide 301-308), which becomes more flexible following activation. Likewise, deprotection of several peptides is consistent with the release of autoinhibitory interactions. This is notably the case for the conserved kinked helix (peptide 166-176) between the DEP and CNB domains, which is a major autoinhibitory element (see spontaneous activity of EPAC1^DEP-CNB lacking this helix in Figures 4D and S5D), and for peptide 777-782 in the GEF domain, the equivalent of which faces this helix in the structure of autoinhibited EPAC2. Finally three peptides are protected within the predicted membrane-interacting surface: peptide 99-114 in the DEP domain, peptide 232-241 in the CNB domain, and peptides covering the region from 845 to 862 in the GEF domain. In addition, protection of peptide 323-347 from the switchboard, which projects a long loop within the predicted membrane-facing region, may also reflect protein-membrane interactions. Besides, HDX-MS reveals a maked protection of residues 26-46 in EPAC1^Nt, together with nearby residues 140-156 which are located in the DEP domain opposite to the predicted membrane-binding region. This suggests that EPAC1^Nt folds back onto the DEP domain in cAMP-, membrane-activated EPAC1. Overall, HDX-MS thus supports our model of fully active EPAC1 bound to the membrane and reveals that the N-terminal domain forms previously unknown intramolecular interactions in this active conformation.

Diversion of EPAC1 dynamics by the chemical inhibitor CE3F4

The above analysis indicates that EPAC1 visits multiple structural intermediates during the course of its dual activation by membranes and cAMP, which can be distinguished by their different GEF efficiencies and affinities for cAMP. Chemical inhibitors that divert specific intermediates can inform on their nature. An appealing compound in that regard is the allosteric inhibitor CE3F4, which inhibits cellular EPAC1 functions and Rap1 activation and the in vitro activation of Rap1 by EPAC1^a^-CNB-GEF, a truncated construct that contains the kinked a-helix and the CNB and GEF domains ⁴⁴. CE3F4 does not compete with cAMP ⁴⁴, and it also does not compete with the Rap1 GTPase directly as shown by the fact that it does not inhibit the isolated GEF domain (Figure S5A, ⁴⁴). Using purified proteins, we confirmed that CE3F4 inhibits the GEF activity of cAMP-activated full-length EPAC1 in solution (Figures 4A and S5B).

To determine whether CE3F4 disrupts the conformational landscape of EPAC1, we analyzed its effect on the structures of autoinhibited and cAMP-activated EPAC1^a^-CNB-GEF in solution, using synchrotron SAXS coupled to size exclusion chromatography (SEC-SAXS). SEC-SAXS informs on the ensemble of conformations in a given state, including the largest protein dimension (D_max). Data acquisition and analyses are summarized in Table S2 and Figure S6. The conformational change between EPAC1^a^-CNB-GEF and cAMP-EPAC1^a^-CNB-GEF was readily detected as an increase in D_max from 107 Å to 130 Å. CE3F4 decreased the D_max of cAMP-EPAC1^a^-CNB-GEF to 114 Å (Figure 4B), indicating that it induces an alternative conformational ensemble that is distinct from either autoinhibited or cAMP-activated EPAC1^a^-CNB-GEF.

To get further insight into how CE3F4 disrupts the conformational landscape of EPAC1, we compared how it affects EPAC1 properties in solution and in the presence of membranes. First, we asked whether CE3F4 affects the interaction of EPAC1 with the small GTPase Rap1, using size exclusion chromatography (Figures 4C and S5C). Formation of the Rap1-EPAC1^a^-CNB-GEF complex was readily observed in the presence of cAMP. In striking contrast, CE3F4 impaired this interaction. Since CE3F4 does not inhibit the GEF domain directly (Figure S5A), this indicates that it induces inhibitory rearrangements in the Rap1-binding site of the GEF domain in an allosteric manner. Next, we asked whether cAMP is necessary for this allosteric mechanism. For that, we assayed the effect of CE3F4 under conditions where partial activity is observed in the absence of cAMP. In solution, we used EPAC1^CNB-GEF, a construct that displays intrinsic GEF activity in the absence of cAMP, which is increased by cAMP (Figure S5D, ²²). CE3F4 did not inhibit EPAC1^CNB-GEF in the absence of cAMP, but inhibition was recovered in the presence of cAMP (Figure 4D). We then took advantage of the fact that EPAC1^FL is partially active on membranes in the absence of cAMP to investigate whether CE3F4 also requires cAMP in the context of the membrane. CE3F4 did not inhibit liposome-activated EPAC1^FL in the absence of cAMP, while inhibition was recovered upon addition of cAMP (Figures 4E and S5E). These data indicate that CE3F4 inhibits EPAC1 only in the presence of cAMP, whether in solution or on the membrane. Finally, we analyzed whether the allosteric structural changes induced by CE3F4 interfere with the interaction of EPAC1^FL with membranes. Consistently, CE3F4 reduced binding of EPAC1^FL to PM-liposomes only in the presence of cAMP (Figure 4F), suggesting that the structural elements that block binding of Rap1 are located close to the membrane interface.

Together, our findings indicate that CE3F4 recognizes specifically cAMP-activated EPAC1 intermediates, and that it acts by remodeling structural elements that respond to cAMP in a manner that blocks access to the Rap-binding site in the vicinity of the membrane.

Previous studies identified membranes as important determinants of EPAC1-Rap1 signaling by cAMP in cells ^{36, 37}, yet how membranes shape the behavior of EPAC1 has remained poorly understood. Here, we discovered that anionic membranes are pivotal regulators of the response of EPAC1 to cAMP by lowering the affinity of EPAC1 to cAMP by two orders of magnitude. Our data reveal that EPAC1 has a complex structural and dynamical landscape, which culminates with a highly active form activated by both cAMP and membranes acting in synergy. Remarkably, the different states can be distinguished by the cardiomyocyte-active EPAC1 inhibitor CE3F4, which uncovers important aspects of the inhibitory mechanism. We discuss below the implications of these findings for EPAC1 regulation and cAMP signaling in cells and for drug discovery.

Implications for EPAC regulation by membranes

Membranes are increasingly recognized as pivotal determinants of small GTPase functions (reviewed in ^{30, 31}), notably by regulating the activity of their regulators (e.g. ^{56, 57}). To understand the inner workings of small GTPase signaling pathways of biomedical interest, it is thus essential to analyze and quantify the contribution of membranes. Our study provides important insights into this aspect of EPAC1 functions.

Our results demonstrate that EPAC1 has at least four states in equilibrium (Figure 5A): a soluble apo form which is fully autoinhibited, two modestly active forms bound to either cAMP or membranes, and an optimally active form that is both cAMP- and membrane-bound. We propose that membranes relieve autoinhibition by repositioning the regulatory domains away from the catalytic site, and that this open conformation remains highly dynamic until it is interlocked by binding of cAMP to yield full activation. Our study further suggests that EPAC1 binds to membranes by multiple non-specific electrostatic interactions with PS and phosphoinositide lipids, which are are major constituents of this membrane (reviewed in ⁵⁸). These protein/membrane interactions likely compensate the loss of autoinhibitory intramolecular interactions in the cytosol, providing an energetic rationale for the ability of the membrane to act as a moderate activator of EPAC1 on its own. Given the potent effect measured for PS and PIP₂lipids, it is surprising that the PA-specific recruitment of EPAC1 to the plasma membrane ³⁷ was not not recapitulated by PA-containing liposomes. Since PA is an essential precursor in the synthesis of major plasma membrane phospholipids, including PS and phosphoinositides (reviewed in ⁵⁹), a tentative explanation is that removal of PA in cells impairs EPAC1 recruitment indirectly by affecting the content of the plasma membrane in these lipids.

The orientation of GEFs with respect to the membrane has been shown to be a critical determinant for their activity ^{60, 61}. Here, the predicted electrostatic interface of EPAC1 with the membrane suggests that EPAC1 is closely apposed to the membrane and precisely oriented to activate membrane-associated Rap1. We propose that this allows electrostatic coincidence detection of EPAC1 and Rap1 at the plasma membrane through their concomitant interactions with anionic lipids (Figure 5B). Such a mechanism may in turn be important for the specificity of EPAC1 for Rap GTPase isoforms. Like Rap1A, the Rap1B isoform carries numerous positively charged residues in its C-terminus (RKTPVPGKARKKSSCQLL). In contrast, the related Rap2 isoforms have only one or no such residue (Rap2A: YAAQPDKDDPCCSACNIQ, Rap2B: YAAQPNGDEGCCSACVIL, Rap2C: YSSLPEKQDQCCTTCVVQ). Thus, Rap1 isoforms have determinants for electrostatic coincidence detection with EPAC1 at the plasma membrane, while Rap2 isoforms lack such determinants. Alternatively, Rap2 isoforms may be activated in other compartments where EPAC1 has been observed, such as the mitochondria or the nuclear pore (reviewed in ⁶²), or solely by RapGEF subfamilies other than EPAC (reviewed in ⁶³). Because of its shared domain organization, it is likely that the related EPAC2 protein follows a similar membrane-binding and activation scenario. Future studies are needed to identify the specific determinants for alternative subcellular EPAC1 targeting and for the specificity of regulation of EPAC2 by membranes.

Implications for cAMP signaling by EPAC1 in cells

The actual concentration of cAMP at any moment and subcellular location depends on a complex balance between synthesis, degradation, buffering and diffusion processes, which ultimately should match the localization and the cAMP affinity of specific cAMP signal transducers (reviewed in ^{1, 2, 7}). Seminal studies recently established that cAMP is restricted within receptor- and membrane-proximal nanodomains, thus does not diffuse freely in the cytosol where its concentration remains low ^{3, 4, 5, 6}. By identifying membranes as a previously overlooked component of cAMP signaling through EPAC1, our findings bring important insights into this issue.

First, they resolve the conundrum of the intriguingly low affinity of EPAC1 for cAMP observed in solution, which makes it unlikely that it can be activated in the bulk cytosol given the low cAMP concentration and variations in this compartment ⁶. Hence, soluble cAMP-bound EPAC1 is likely a negligible species in the cell. In contrast with soluble EPAC1, membrane-attached EPAC1 has a high affinity for cAMP, suggesting that in cells EPAC1 must be membrane attached to be activated by cAMP. We propose that membranes prime EPAC1 for subsequent activation by a cAMP influx. Accordingly, activation of EPAC1 by cAMP likely takes place almost exclusively on the membrane, with cAMP securing the localization of EPAC1 on the membrane rather than driving it. In this new schema, alternation between soluble and membrane-attached EPAC1 is likely spontaneous, with EPAC1 falling off the membrane if it does not encounter a cAMP signal.

Recent studies showed that nanometer-sized cAMP compartments associated with GPCRs translate into localized PKA activation to ensure receptor-specific signaling ⁶. Specific activation of EPAC1 by different b-adrenergic receptors has been reported, which resulted in the stimulation of distinct signaling arms in cardiac myocytes ⁶⁴, raising the question of whether EPAC1 may also respond to specific receptors through cAMP nanodomains. Our data reveal that EPAC1 has biochemical determinants to function in membrane-proximal cAMP nanodomains, which may explain the mechanism of these receptor-specific responses (Figure S7). Notably, while soluble EPAC1 is not activated at the low cAMP concentration of the cytosol, membrane-attached EPAC1 is highly sensitive to cAMP concentrations in the range of those measured in receptor-associated nanodomains. In addition, our data resolve the intriguing discrepancy between the sensitivities of PKA and EPAC1 to cAMP fluctuations, by showing that upon membrane binding, the response of EPAC1 to cAMP (affinity and concentration resulting in half maximal activation) is similar to that of PKA ^{24, 25}. Thus, cAMP nanodomains may control both PKA and EPAC pathways.

Together, our study revisits the molecular mechanisms that enable EPAC1 to transduce cAMP signaling in cells. Important questions for future studies will be to document how membrane-restricted activation of EPAC1 articulates with compartmentalized cAMP.

Implications for EPAC1 drug discovery

The growing evidence that EPAC proteins play pivotal roles in physiological and pathogenic conditions has motivated intense research to discover specific agonists and antagonists (reviewed in ⁴³). Our findings provide a new framework for their discovery and improvement.

First, our study identifies previously unknown conformations of EPAC1, each featuring distinct intramolecular, protein/membrane and protein/ligand interactions (Figure 5A). Each of these states thus offers unique structural characteristics that can be exploited to discover activating and inhibitory drugs. Importantly, soluble cAMP-activated EPAC1 appears an irrelevant target for inhibition, as this intermediate is likely in negligible amount in cells. Alternatively, partially activated, membrane-bound EPAC1 may constitute an interesting target for chemicals that bind at the EPAC1-membrane interface, as recently shown by a proof-of-concept GEF inhibitor, Bragsin, which acts by binding at the interface between an ArfGEF and the membrane to alter its orientation ⁶¹.

Second, our findings provide new insights into the cellular mechanism of action of agonists that compete with cAMP. Notably, we find that the affinity of cAMP for membrane-bound EPAC1 is only slightly higher than that of the widely used super agonist 007 for soluble EPAC1 ^{25, 26, 27, 28}. Thus, 007 may by-pass the requirement that EPAC1 is bound to the membrane to be activated. Accordingly, it may yield non-physiological activation of EPAC1 in the cytosol and/or on alternative membranes, hence miswire EPAC1 signaling. Such side effects should be considered when using this and other cAMP analogs to interrogate the biology of EPAC proteins in the cell.

Finally, our results provide unprecedented insight into the molecular and cellular mechanism of action of the cardiomyocyte-active CE3F4 inhibitor. First, the fact that the binding site for CE3F4 exists only in cAMP-bound EPAC1 intermediates, of which soluble cAMP-activated EPAC1 is a negligeable species at physiological cAMP concentrations, makes it likely that, in cells, the actual target of CE3F4 is membrane- and cAMP-bound EPAC1 (Figure 5A). Second, our data suggest that CE3F4 binds to a cAMP-responsive region outside the GEF domain in a manner that projects elements into the Rap1-binding site. We propose that this cAMP-responsive region is the switchboard, a region between the CNB and GEF domains that functions as a hinge during activation by cAMP ^{19, 20}. A candidate inhibitory element in this region is a flexible b-hairpin of the GEF domain located between the switchboard and the Rap1-binding site (EPAC2: ⁹¹⁹GNKTFIDNLVN⁹²⁹, EPAC1: ⁷⁸⁶GNHTLVENLIN⁷⁹⁶; Figure 5C). In SOS, a Ras GTPase exchange factor whose GEF domain is related to that of EPAC, the equivalent b-hairpin (⁹⁴³GNPEVLKRHGKELIN⁹⁵⁷) hinders access to the Ras-binding site, from which it is removed upon allosteric activation of SOS by Ras-GTP ⁶⁵ (Figure 5D). We propose that CE3F4 acts in a similar manner, by projecting the b-hairpin into the Rap-binding site, using the flexible Gly/Asn residues at both ends of the b-hairpin to support this movement. Together, our study reveals that CE3F4 has a remarkable allosteric mechanism which requires that EPAC1 is fully activated to be inhibited in cells. By targeting EPAC1 “in action”, such a mechanism makes CE3F4 highly specific to a membrane-localized cAMP-EPAC1-Rap1 signaling axis in cardiovascular diseases.

In conclusion, our findings provide a robust framework to understand the molecular basis of EPAC1 activation by cAMP at the surface of membranes in cells and in diseases and the action of activators and inhibitors. Future studies are now needed to establish how EPAC1 and EPAC2 respond to compartmentalized cAMP signaling within this framework and for mechanism-based improvement of the CE3F4 inhibitor. More generally, our work highlights the importance to design drug discovery strategies that take fully advantage of elusive Achille’s heels, such as structural malleability or protein-membrane interactions, which are critical in regulating small GTPase signaling in physiological functions and in diseases.

Protein cloning, expression and purification

Human EPAC1^DEP (residues 50-184), EPAC1^Nt-DEP (residues 1-184), EPAC1^DEP-CNB (residues 50-318), EPAC1^Nt-DEP-CNB (residues 1-318), EPAC1^CNB-GEF (residues 184-881), EPAC1^a^-CNB-GEF (residues 149-881), and EPAC1^DEP-CNB-GEF (residues 184-881) were cloned into the Gateway destination vector pHMGWA carrying a 6xHis-MBP tag in N-terminus. Human EPAC1^GEF (residues 321-881)was cloned into the Gateway destination vector pETG20a carrying a 6xHis-TRX tag in N-terminus. Human full-length EPAC1 (EPAC1^FL) was cloned into pET28a vector (6xHis in N-terminus). A TEV protease cleavage site was inserted during cloning after the 6xHis-MBP, 6xHis-TRX or 6xHis N-terminal tags of each vector. Full-length human Rap1A was cloned into the pET3a vector carrying a 6xHis tag in C-terminus. All clones were confirmed by sequencing (GATC Biotech). All EPAC1 constructs and Rap1 were produced in the E. coli BL21 (DE3) pG-KJE8 chaperone-expressing strain ⁶⁶, in LB medium with 0.5 g/L L-arabinose and 2.5 mg/L tetracycline. Induction of proteins was done with 0.5 mM IPTG overnight at 20°C. Bacterial pellets were resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 300 mM NaCl, 2 mM b-mercaptoethanol, 2 mM MgCl₂, 0.2% n-Dodecyl b-D-maltoside, 0,5 mg/mL lysozyme, 7.5 U/mL benzonase, and anti-protease cocktail) and cells were broken using a French press. After centrifugation at 15000g during 30 min, the cleared lysate supernatant was filtered over a 0.22 mm filter and loaded onto a Ni-NTA affinity chromatography column (HisTrap FF, GE Healthcare) and His-tagged proteins were eluted with 250 mM imidazole. For all proteins except EPAC1^FL and Rap1, the 6xHis tag was cleaved by the TEV protease (1/10 w/w ratio) during overnight dialysis in storage buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 2 mM β-mercaptoethanol, 2 mM MgCl₂, 2,5% glycerol, 0.02% DDM) at 4°C. The cleaved tag was removed by a second Ni-NTA affinity chromatography step. The purity of all proteins was polished on either a Superdex75 or a Superdex200 16/600 size exclusion column equilibrated with storage buffer. Protein purity was assessed by SDS-PAGE.

Liposomes preparation and binding assay

Lipids (natural origin) were from Avanti Polar Lipids, except NBD-PE (Sigma). Liposomes were prepared as described in ⁶¹ in a buffer containing 50 mM HEPES pH 7.4, 120 mM potassium acetate. All liposomes contain phosphatidylethanolamine (PE) at 20% and other lipids as indicated, and are completed to 100% by phosphatidylcholine (PC). PM liposomes contain 48% PC, 20% PE, 10% phosphatidylserine (PS), 20% cholesterol, 2% phosphatidylinositol-4,5-bisphosphate (PIP₂). All liposomes were extruded through a 0,2 mm pore size filter before use, and size distributions were checked by dynamic light scattering (DLS) on a DynaPro instrument (Wyatt). Binding of proteins to liposomes was assayed by liposome flotation as described in ⁵⁷. When indicated, 1 mM cAMP was added to the sucrose gradient. 1/8 of the bottom and top fractions were analyzed by SDS-PAGE using Coomassie staining. Quantification of the intensity of the bands was done using Image Lab^TM v6.0 on a ChemiDoc MP Imaging System (Biorad).

Nucleotide exchange kinetics assay

For kinetics assays, the Rap1 GTPase was loaded with BODIPY-FL-GDP (Jena Bioscience) prior to nucleotide exchange by incubating 50 mM Rap1 with 250 mM BODIPY-GDP and 10 mM EDTA for 30 min at room temperature. Nucleotide exchange was stopped by the addition of 75 mM MgCl₂. Removal of excess nucleotides and buffer exchange was done on a PD SpinTrap G-25 (GE HealthCare Life Science). Nucleotide exchange kinetics were monitored by following the decay of BODIPY-GDP fluorescence upon replacement by GTP, using excitation and emission wavelengths of 485 and 530 nm, respectively. Measurements were carried out at 30°C on a Cary Eclipse fluorimeter (Varian) under stirring in 800mL cuvettes, or in microplates wells (150mL) on a FLEXstation (Molecular Devices). In all experiments, proteins and liposomes were pre-incubated in HKM buffer (HEPES 50mM pH 7.4, 120 mM potassium acetate, 1 mM MgCl₂) for 1 min under stirring before the exchange reaction was initiated by addition of 100 µM GTP. cAMP was added as indicated. Liposomes were added at 100 mM (total lipids). For kinetics in solution, k_obs values were determined from a mono-exponential fit. We noted that kinetics in the presence of membranes were best fit by one fast and one slow exponentials. These two components likely correspond to the activation rates of membrane-associated and soluble Rap1, respectively. Pseudo first order k_cat/K_M were determined by linear regression from k_obs values measured over a range of GEF concentrations following the Michaelis-Menten formalism as described previously ⁶⁷. In the case of experiments carried out with liposomes, k_obs values corresponding to the fast component were used. Values are mean ±sd of of the indicated number of independent experiments.

Determination of dissociation constants and maximal activities

The apparent dissociation constant K_D and maximal activity k_max of EPAC1^FL for cAMP and 007 in solution and on membranes were determined by monitoring the kinetics of Rap1 activation as a function of cAMP or 007 concentration. K_Dwas identified as the concentration of cAMP or 007 yielding half maximal activity as previously described for EPAC proteins in solution ^{26, 28}.

Crystallographic analysis

Crystallization screens were performed using the sitting-drop vapor diffusion method at 18 °C with a Mosquito robot (TTP LabTech) in 96-well crystallization plates by mixing 100 nL of EPAC1^DEP-CNB at 5 mg/mL with 100 nL of precipitant solution. Crystals were obtained in 20% w/v polyethylene glycol 3350, 100 mM BIS-TRIS propane pH 6.5, 100 mM sodium phosphate, 100 mM potassium phosphate. Crystals were cryo-protected using the reservoir solution supplemented with 15% glycerol prior to flash freezing. A complete diffraction dataset was collected on ID30B beamline (ESRF synchrotron, France) and was integrated with the program AutoProc ⁶⁸. The structure was solved with the automated molecular replacement pipeline BALBES (CCP4 suite) ⁶⁹ using the CNB domain of activated EPAC2 as a model (PDB entry 3CF6, ²⁰). Refinement was carried out with the program Phenix ⁷⁰ using TLS parameters, in alternation with graphical building using Coot41 ⁷¹. The quality of the structure was assessed using MolProbity42. Data collection and refinement statistics are reported in Table S1. Coordinates and structure factors have been deposited in the Protein Data Bank with entry code 6H7E.

Structural modeling

The GEF domain of EPAC1 was modeled using the protein fold recognition Phyre2 server ⁷². The composite model of activated EPAC1^DEP-CNB-GEF bound to Rap1 was generated by superposing our EPAC1^DEP-CNB-cAMP crystal structure and the model of EPAC1^GEF onto the crystal structure of EPAC2^CNB-GEF-Rap1-cAMP (²⁰, PDB entry 3CF6). Docking of the EPAC1 model onto an anionic lipid membrane was done with the OMP server using default parameters ⁵².

HDX-MS

EPAC1^FL (15µM) was deuterated at room temperature alone or in complex with cAMP (4.5 mM) and freshly prepared PM liposomes (3.125 mM total lipid). To optimize protein stability over time, a new mixture was freshly prepared every 3 replicates. Deuteration was carried out by incubating 3µL of the EPAC1 solutions into 57µL of deuterium buffer (50 mM HEPES pD 8.0, 150 mM NaCl, 2 mM β-mercaptoethanol, 2 mM MgCl₂and D₂O) for defined durations (0.5, 1, 2, 5 and 10 min), corresponding to a 95% deuterated final solution. All conditions were analyzed in triplicate. 55 µL of each deuterated preparation were quenched with 55 µL of quench buffer (150 mM Glycine pH 1.5 and 4M guanidine-HCl) at 1 °C during 30 sec with a final pH of 2.4. Deuteration, quenching and injection of the samples were automatically performed, three by three, using a Leap HDX Automation manager (Waters). Digestion and chromatography were done on an Acquity UPLC system with HDX technology (Waters, Manchester, UK). Digestion was performed on a pepsin-immobilized cartridge (Enzymate pepsin column, 300 Å, 5 µm, 2.1 mm I.D. × 30 mm, Waters, Manchester, UK) at a 100µL/min flow rate of 0.1% acid formic solution at 20 °C. Peptides were then trapped on a UPLC pre-column (ACQUITY UPLC BEH C18 Van-Guard pre-column, 2.1 mm I.D. x 5 mm, 1.7 µM particle diameter, Waters) and separated on UPLC column (ACQUITY UPLC BEH C18, 1.0 mm I.D. x 100 mm, 1.7 µM particle diameter, Waters) at 0.1 °C. Peptide separation was performed at a flow rate of 40µL/min, with an elution gradient of solvent A (0.1% formic acid, water) and solvent B (0.1% formic acid, acetonitrile) from 2-40% B over 7 min followed by a 0.5 min ramp to 85% B. Peptides were then analyzed with a Synapt G2Si HDMS (Waters, Wilmslow, UK) with Glu-fibrino peptide infused for calibration and lock mass correction in positive ion and resolution mode. MS/MS acquisition were done using data independent acquisition (MS^E) on a range of 50-2000 m/z. Protein Lyse Global Server 2.5.3 (PLGS, Waters) searches were done to identify unlabeled peptides using a homemade database with the protein of interest and pepsin sequences with methionine oxidation as variable modification sites. DynamX 3.0 (Waters) was next employed to filter identified peptides with a minimum intensity of 1000, a minimum fragment of 0.3 per amino acid and their presence in at least 2 out of 3 replicates. Isotopic profiles for all identified peptides were manually checked and validated without back-exchange correction. Deuteration mean values were reported as relative. Differences of deuterium uptake were statistically validated with a p-value of 0.01 with the MEMHDX software ⁷³ with statistical significance thresholds set to 0.01. HDX results were mapped onto the model of fully active EPAC1^FL. The chromatography, acquisition and processing methods are summarized in Supplementary information. HDX-MS datas have been deposited to the PRIDE repository ⁷⁴ with accession code PXD040227.

Inhibition studies

All inhibition experiments were done with either the active R-CE3F4 enantiomer or with the CE3F4 racemic mix, which contains >90% of the active enantiomer, synthesized as described ⁷⁵. CE3F4 was prepared at 50 mM in DMSO and used at a final concentration of 50 mM except when indicated otherwise. DMSO was used as a control in kinetic experiments.

SAXS data acquisition and analysis

All small angle X-ray scattering (SAXS) data were obtained from size exclusion chromatography-SAXS (SEC-SAXS) experiments on the SWING beamline (SOLEIL synchrotron, France), using the EPAC1^a^-CNB-GEF construct. EPAC1 was used at a concentration of 60 μM to obtain high SAXS intensities. Samples were injected on a high-performance liquid chromatography (HPLC) size exclusion column (BioSEC-3 300 Å, Agilent Technologies, Inc.) equilibrated with the elution buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 0.5 mM EDTA). For SEC-SAXS experiments with CE3F4, EPAC1 was pre-incubated with 100 μM CE3F4 (racemic mix) and the elution buffer was supplemented with 100 μM CE3F4. For SEC-SAXS experiments with cAMP, EPAC1 was activated with 1 mM cAMP, and the elution buffer was supplemented with 200 μM cAMP. Data reduction to absolute units, frame averaging, and subtraction were done using the FOXTROT program suite (SOLEIL synchrotron). Frames corresponding to the high-intensity fractions of the peak and having constant radius of gyration (R_g) within error were averaged. All the data analyses were performed with programs from the ATSAS package36 suite ⁷⁶ and are summarized in Table S2. R_g values were evaluated by Guinier Wizard using the data within the range of Guinier approximation sR_g < 1.3 and by Distance Distribution Wizard, both of which are modules of the PRIMUS program. The maximum distance Dmax was estimated with PRIMUS and refined by trial and error with GNOM. The distance distribution functions were calculated with GNOM. SAXS profiles were compared with DATCMP (ATSAS suite). SAXS data sets have been deposited in the small angle scattering biological data bank SASBDB (https://www.sasbdb.org/) (accession codes: SASDQA8: unbound EPAC1 ; SASDQB8: EPAC1-cAMP; SASDQC8: EPAC1-CE3F4 ; SASDQD8: EPAC1-cAMP-CE3F4).

Analysis of complex formation by size exclusion chromatography

All experiments were carried out in elution buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 0.5 mM EDTA) containing DMSO or 100 mM CE3F4 (racemic mix). EPAC1^a-CNB-GEF was used at 25 mM in all experiments. EPAC1 was pre-incubated with DMSO or 100 mM C3EF4 for 5 min, before addition of Rap1 (75 mM), cAMP (1 mM) and EDTA (5 mM). Proteins samples were injected onto a Superdex 200 25 mL gel filtration column (GE-Healthcare) at 0.5 mL/min. Fractions were analyzed by SDS-PAGE.

Acknowledgements.

This study was supported by grants to J.C. from the Laboratory of Excellence LERMIT supported by the French National Research Agency (ANR-10-LABX-33) under the program “Investissements d'Avenir” ANR-11-IDEX-0003-01, the Fondation pour la Recherche Médicale (grants DEQ20150331694 and EQU202003010344) and the Institut National du Cancer (grant 2014-160). F.P. was supported by a PhD grant from the Ecole Normale Supérieure Paris-Saclay, C.S. by a PhD grant from the LERMIT program. The authors acknowledge the support and the use of resources of the French Proteomic Infrastructure ProFI ANR-10-INBS-08-03, the GIS IBiSA and Région Alsace for financial support in purchasing a Synapt G2SI HDMS instrument.We are grateful to the scientific staff at the SAXS beamline SWING at synchrotron SOLEIL (Gif-sur-Yvette, France) and the X-ray crystallography beamline ID30 at synchrotron ESRF (Grenoble, France) for their expert help and advice. We thank Olivier Duclos and Magali Mathieu (SANOFI) for providing SAXS beamline time at synchrotron SOLEIL and Ilham Ladhid and Ellyn Renou (Cherfils lab, CNRS and ENS Paris-Saclay) for their help with cloning.

Beavo JA, Brunton LL. Cyclic nucleotide research -- still expanding after half a century. Nat Rev Mol Cell Biol 3, 710-718 (2002).
Musheshe N, Schmidt M, Zaccolo M. cAMP: From Long-Range Second Messenger to Nanodomain Signalling. Trends Pharmacol Sci 39, 209-222 (2018).
Bock A, et al. Optical Mapping of cAMP Signaling at the Nanometer Scale. Cell 182, 1519-1530 e1517 (2020).
Zhang JZ, et al. Phase Separation of a PKA Regulatory Subunit Controls cAMP Compartmentation and Oncogenic Signaling. Cell 182, 1531-1544 e1515 (2020).
Tenner B, et al. Spatially compartmentalized phase regulation of a Ca(2+)-cAMP-PKA oscillatory circuit. Elife 9, (2020).
Anton SE, et al. Receptor-associated independent cAMP nanodomains mediate spatiotemporal specificity of GPCR signaling. Cell 185, 1130-1142 e1111 (2022).
Zaccolo M, Zerio A, Lobo MJ. Subcellular Organization of the cAMP Signaling Pathway. Pharmacol Rev 73, 278-309 (2021).
Taylor SS, Ilouz R, Zhang P, Kornev AP. Assembly of allosteric macromolecular switches: lessons from PKA. Nat Rev Mol Cell Biol 13, 646-658 (2012).
Kawasaki H, et al. A family of cAMP-binding proteins that directly activate Rap1. Science 282, 2275-2279 (1998).
de Rooij J, et al. Epac is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP. Nature 396, 474-477 (1998).
Bos JL. Epac proteins: multi-purpose cAMP targets. Trends Biochem Sci 31, 680-686 (2006).
Lezoualc'h F, Fazal L, Laudette M, Conte C. Cyclic AMP Sensor EPAC Proteins and Their Role in Cardiovascular Function and Disease. Circ Res 118, 881-897 (2016).
Robichaux WG, 3rd, Cheng X. Intracellular cAMP Sensor EPAC: Physiology, Pathophysiology, and Therapeutics Development. Physiol Rev 98, 919-1053 (2018).
Kumar N, et al. Insights into exchange factor directly activated by cAMP (EPAC) as potential target for cancer treatment. Mol Cell Biochem 447, 77-92 (2018).
Singhmar P, et al. Critical role for Epac1 in inflammatory pain controlled by GRK2-mediated phosphorylation of Epac1. Proc Natl Acad Sci U S A 113, 3036-3041 (2016).
Liu H, et al. Epac1 inhibition ameliorates pathological angiogenesis through coordinated activation of Notch and suppression of VEGF signaling. Sci Adv 6, eaay3566 (2020).
Laudette M, Zuo H, Lezoualc'h F, Schmidt M. Epac Function and cAMP Scaffolds in the Heart and Lung. J Cardiovasc Dev Dis 5, (2018).
Parnell E, Palmer TM, Yarwood SJ. The future of EPAC-targeted therapies: agonism versus antagonism. Trends Pharmacol Sci 36, 203-214 (2015).
Rehmann H, Das J, Knipscheer P, Wittinghofer A, Bos JL. Structure of the cyclic-AMP-responsive exchange factor Epac2 in its auto-inhibited state. Nature 439, 625-628 (2006).
Rehmann H, Arias-Palomo E, Hadders MA, Schwede F, Llorca O, Bos JL. Structure of Epac2 in complex with a cyclic AMP analogue and RAP1B. Nature 455, 124-127 (2008).
de Rooij J, Rehmann H, van Triest M, Cool RH, Wittinghofer A, Bos JL. Mechanism of regulation of the Epac family of cAMP-dependent RapGEFs. J Biol Chem 275, 20829-20836 (2000).
Kraemer A, et al. Dynamic interaction of cAMP with the Rap guanine-nucleotide exchange factor Epac1. J Mol Biol 306, 1167-1177 (2001).
Rehmann H, Schwede F, Doskeland SO, Wittinghofer A, Bos JL. Ligand-mediated activation of the cAMP-responsive guanine nucleotide exchange factor Epac. J Biol Chem 278, 38548-38556 (2003).
Dao KK, et al. Epac1 and cAMP-dependent protein kinase holoenzyme have similar cAMP affinity, but their cAMP domains have distinct structural features and cyclic nucleotide recognition. J Biol Chem 281, 21500-21511 (2006).
Enserink JM, et al. A novel Epac-specific cAMP analogue demonstrates independent regulation of Rap1 and ERK. Nat Cell Biol 4, 901-906 (2002).
Rehmann H. Characterization of the activation of the Rap-specific exchange factor Epac by cyclic nucleotides. Methods Enzymol 407, 159-173 (2006).
Brown LM, Rogers KE, McCammon JA, Insel PA. Identification and validation of modulators of exchange protein activated by cAMP (Epac) activity: structure-function implications for Epac activation and inhibition. J Biol Chem 289, 8217-8230 (2014).
Schwede F, et al. Structure-guided design of selective Epac1 and Epac2 agonists. PLoS Biol 13, e1002038 (2015).
Peurois F, Peyroche G, Cherfils J. Small GTPase peripheral binding to membranes: molecular determinants and supramolecular organization. Biochem Soc Trans 47, 13-22 (2019).
Cherfils J, Zeghouf M. Regulation of small GTPases by GEFs, GAPs, and GDIs. Physiol Rev 93, 269-309 (2013).
Nawrotek A, Zeghouf M, Cherfils J. Protein-membrane interactions in small GTPase signalling and pharmacology: perspectives from Arf GTPases studies. Biochem Soc Trans, (2020).
Beranger F, Tavitian A, de Gunzburg J. Post-translational processing and subcellular localization of the Ras-related Rap2 protein. Oncogene 6, 1835-1842 (1991).
Nancy V, Callebaut I, El Marjou A, de Gunzburg J. The delta subunit of retinal rod cGMP phosphodiesterase regulates the membrane association of Ras and Rap GTPases. J Biol Chem 277, 15076-15084 (2002).
Wilson JM, Prokop JW, Lorimer E, Ntantie E, Williams CL. Differences in the Phosphorylation-Dependent Regulation of Prenylation of Rap1A and Rap1B. J Mol Biol 428, 4929-4945 (2016).
Zhang X, et al. Cyclase-associated protein 1 (CAP1) is a prenyl-binding partner of Rap1 GTPase. J Biol Chem 293, 7659-7673 (2018).
Ponsioen B, Gloerich M, Ritsma L, Rehmann H, Bos JL, Jalink K. Direct spatial control of Epac1 by cyclic AMP. Mol Cell Biol 29, 2521-2531 (2009).
Consonni SV, Gloerich M, Spanjaard E, Bos JL. cAMP regulates DEP domain-mediated binding of the guanine nucleotide exchange factor Epac1 to phosphatidic acid at the plasma membrane. Proc Natl Acad Sci U S A 109, 3814-3819 (2012).
Consonni SV, Maurice MM, Bos JL. DEP domains: structurally similar but functionally different. Nat Rev Mol Cell Biol 15, 357-362 (2014).
Li S, et al. Mechanism of intracellular cAMP sensor Epac2 activation: cAMP-induced conformational changes identified by amide hydrogen/deuterium exchange mass spectrometry (DXMS). J Biol Chem 286, 17889-17897 (2011).
Renault L, Guibert B, Cherfils J. Structural snapshots of the mechanism and inhibition of a guanine nucleotide exchange factor. Nature 426, 525-530 (2003).
Rehmann H. Epac-inhibitors: facts and artefacts. Sci Rep 3, 3032 (2013).
Ahmed A, et al. Recent Advances in EPAC-Targeted Therapies: A Biophysical Perspective. Cells 8, (2019).
Bouvet M, Blondeau JP, Lezoualc'h F. The Epac1 Protein: Pharmacological Modulators, Cardiac Signalosome and Pathophysiology. Cells 8, (2019).
Courilleau D, et al. Identification of a tetrahydroquinoline analog as a pharmacological inhibitor of the cAMP-binding protein Epac. J Biol Chem 287, 44192-44202 (2012).
Prajapati R, et al. Usefulness of Exchanged Protein Directly Activated by cAMP (Epac)1-Inhibiting Therapy for Prevention of Atrial and Ventricular Arrhythmias in Mice. Circ J 83, 295-303 (2019).
Dominguez-Rodriguez A, Ruiz-Hurtado G, Sabourin J, Gomez AM, Alvarez JL, Benitah JP. Proarrhythmic effect of sustained EPAC activation on TRPC3/4 in rat ventricular cardiomyocytes. J Mol Cell Cardiol 87, 74-78 (2015).
Boulton S, Selvaratnam R, Blondeau JP, Lezoualc'h F, Melacini G. Mechanism of Selective Enzyme Inhibition through Uncompetitive Regulation of an Allosteric Agonist. J Am Chem Soc, (2018).
Dudek J, Hartmann M, Rehling P. The role of mitochondrial cardiolipin in heart function and its implication in cardiac disease. Biochim Biophys Acta Mol Basis Dis 1865, 810-821 (2019).
Fazal L, et al. Multifunctional Mitochondrial Epac1 Controls Myocardial Cell Death. Circ Res 120, 645-657 (2017).
Wang Z, et al. A cardiac mitochondrial cAMP signaling pathway regulates calcium accumulation, permeability transition and cell death. Cell Death Dis 7, e2198 (2016).
Rehmann H, et al. Structure and regulation of the cAMP-binding domains of Epac2. Nat Struct Biol 10, 26-32 (2003).
Lomize MA, Pogozheva ID, Joo H, Mosberg HI, Lomize AL. OPM database and PPM web server: resources for positioning of proteins in membranes. Nucleic Acids Res 40, D370-376 (2012).
Ohashi Y, et al. Membrane characteristics tune activities of endosomal and autophagic human VPS34 complexes. Elife 9, (2020).
Burke JE. Dynamic structural biology at the protein membrane interface. J Biol Chem 294, 3872-3880 (2019).
Lau AM, Jia R, Bradshaw RT, Politis A. Structural predictions of the functions of membrane proteins from HDX-MS. Biochem Soc Trans 48, 971-979 (2020).
Aizel K, et al. Integrated conformational and lipid-sensing regulation of endosomal ArfGEF BRAG2. PLoS Biol 11, e1001652 (2013).
Peurois F, et al. Characterization of the activation of small GTPases by their GEFs on membranes using artificial membrane tethering. Biochem J 474, 1259-1272 (2017).
van Meer G, Voelker DR, Feigenson GW. Membrane lipids: where they are and how they behave. Nat Rev Mol Cell Biol 9, 112-124 (2008).
Vance JE. Phospholipid synthesis and transport in mammalian cells. Traffic 16, 1-18 (2015).
Karandur D, Nawrotek A, Kuriyan J, Cherfils J. Multiple interactions between an Arf/GEF complex and charged lipids determine activation kinetics on the membrane. Proc Natl Acad Sci U S A 114, 11416-11421 (2017).
Nawrotek A, et al. PH-domain-binding inhibitors of nucleotide exchange factor BRAG2 disrupt Arf GTPase signaling. Nat Chem Biol 15, 358-366 (2019).
Formoso K, Lezoualc'h F, Mialet-Perez J. Role of EPAC1 Signalosomes in Cell Fate: Friends or Foes? Cells 9, (2020).
Gloerich M, Bos JL. Regulating Rap small G-proteins in time and space. Trends Cell Biol 21, 615-623 (2011).
Berthouze-Duquesnes M, et al. Specific interactions between Epac1, beta-arrestin2 and PDE4D5 regulate beta-adrenergic receptor subtype differential effects on cardiac hypertrophic signaling. Cell Signal 25, 970-980 (2013).
Freedman TS, et al. A Ras-induced conformational switch in the Ras activator Son of sevenless. Proc Natl Acad Sci U S A 103, 16692-16697 (2006).
Nishihara K, Kanemori M, Kitagawa M, Yanagi H, Yura T. Chaperone coexpression plasmids: differential and synergistic roles of DnaK-DnaJ-GrpE and GroEL-GroES in assisting folding of an allergen of Japanese cedar pollen, Cryj2, in Escherichia coli. Appl Environ Microbiol 64, 1694-1699 (1998).
Beraud-Dufour S, et al. A glutamic finger in the guanine nucleotide exchange factor ARNO displaces Mg2+ and the beta-phosphate to destabilize GDP on ARF1. EMBO J 17, 3651-3659 (1998).
Vonrhein C, et al. Data processing and analysis with the autoPROC toolbox. Acta Crystallogr D Biol Crystallogr 67, 293-302 (2011).
Keegan RM, Long F, Fazio VJ, Winn MD, Murshudov GN, Vagin AA. Evaluating the solution from MrBUMP and BALBES. Acta Crystallogr D Biol Crystallogr 67, 313-323 (2011).
Adams PD, et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta crystallographica Section D, Biological crystallography 66, 213-221 (2010).
Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of Coot. Acta Crystallographica Section D: Biological Crystallography 66, 486-501 (2010).
Kelley LA, Mezulis S, Yates CM, Wass MN, Sternberg MJ. The Phyre2 web portal for protein modeling, prediction and analysis. Nat Protoc 10, 845-858 (2015).
Hourdel V, et al. MEMHDX: an interactive tool to expedite the statistical validation and visualization of large HDX-MS datasets. Bioinformatics 32, 3413-3419 (2016).
Perez-Riverol Y, et al. The PRIDE database and related tools and resources in 2019: improving support for quantification data. Nucleic Acids Res 47, D442-D450 (2019).
Courilleau D, Bouyssou P, Fischmeister R, Lezoualc'h F, Blondeau JP. The (R)-enantiomer of CE3F4 is a preferential inhibitor of human exchange protein directly activated by cyclic AMP isoform 1 (Epac1). Biochem Biophys Res Commun 440, 443-448 (2013).
Petoukhov MV, et al. New developments in the ATSAS program package for small-angle scattering data analysis. J Appl Crystallogr 45, 342-350 (2012).

There is NO Competing Interest.

SartreSupplementaryTablesS1S2Submitted.docx
Supplementary Tables S1 S2
SartreSupplementaryFiguresS1S7Submitted.pptx
SartreSupplementarydatasubmitted.xlsx
SupplementaryFigureLegends.docx

Download PDF

Journal Publication

published 12 Jul, 2023

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Membranes prime the RapGEF EPAC1 to transduce cAMP signaling

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Material And Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1