Skip to main content
NCBI home page
Science Advances logoLink to Science Advances
. 2021 Jul 21;7(30):eabg3147. doi: 10.1126/sciadv.abg3147

Cryo-EM structure of the human ELMO1-DOCK5-Rac1 complex

Mutsuko Kukimoto-Niino 1, Kazushige Katsura 1, Rahul Kaushik 1, Haruhiko Ehara 1, Takeshi Yokoyama 1,2, Tomomi Uchikubo-Kamo 1, Reiko Nakagawa 3, Chiemi Mishima-Tsumagari 1,†, Mayumi Yonemochi 1, Mariko Ikeda 1, Kazuharu Hanada 1, Kam Y J Zhang 1, Mikako Shirouzu 1,*
PMCID: PMC8294757  PMID: 34290093

Rac1-bound structure and mutagenesis of the ELMO1-DOCK5 complex provide insights into how ELMO modulates DOCK activity.

Abstract

The dedicator of cytokinesis (DOCK) family of guanine nucleotide exchange factors (GEFs) promotes cell motility, phagocytosis, and cancer metastasis through activation of Rho guanosine triphosphatases. Engulfment and cell motility (ELMO) proteins are binding partners of DOCK and regulate Rac activation. Here, we report the cryo–electron microscopy structure of the active ELMO1-DOCK5 complex bound to Rac1 at 3.8-Å resolution. The C-terminal region of ELMO1, including the pleckstrin homology (PH) domain, aids in the binding of the catalytic DOCK homology region 2 (DHR-2) domain of DOCK5 to Rac1 in its nucleotide-free state. A complex α-helical scaffold between ELMO1 and DOCK5 stabilizes the binding of Rac1. Mutagenesis studies revealed that the PH domain of ELMO1 enhances the GEF activity of DOCK5 through specific interactions with Rac1. The structure provides insights into how ELMO modulates the biochemical activity of DOCK and how Rac selectivity is achieved by ELMO.

INTRODUCTION

The dedicator of cytokinesis (DOCK) family of proteins is composed of evolutionally conserved, atypical guanine nucleotide exchange factors (GEFs) for the Rho family of guanosine triphosphatases (GTPases), Rac, and/or Cdc42 (1–3). DOCK proteins regulate the actin cytoskeleton, owing to their fundamental role in converting Rho GTPases from an inactive guanosine diphosphate (GDP)–bound state to an active guanosine 5′-triphosphate (GTP)–bound state. In mammals, the DOCK family of proteins constitutes 11 members (DOCK1 to DOCK11) and is divided into four subfamilies based on sequence similarity (2, 3): DOCK1 to DOCK5 (subfamilies A and B) exclusively activate Rac and DOCK6 to DOCK11 (subfamilies C and D) preferentially activate Cdc42, of which DOCK6, DOCK7, and DOCK10 also activate Rac (4–6). These DOCK proteins are involved in a variety of important biological processes, including cell migration, phagocytosis, cardiovascular development, myogenesis, axonal guidance, and neuronal differentiation (7). DOCK proteins are also associated with various diseases including cancer, immunodeficiency syndromes, and neuronal disorders (8).

The Dbl family of typical Rho GEFs contains tandem Dbl homology (DH) and pleckstrin homology (PH) domains, whereas the DOCK family has two DOCK homology regions (DHRs). DHR-1 is the phospholipid binding domain (9), and DHR-2 is the catalytic domain (1, 2). In addition, the DOCK-A/B subfamily has an Src homology 3 (SH3) domain at the N terminus, where it can bind to evolutionally conserved engulfment and cell motility (ELMO) scaffold proteins (10). ELMO proteins are essential for the physiological functions of DOCK, despite the fact that they are dispensable for the in vitro nucleotide exchange activity on Rac (1, 11–13).

DOCK1 is autoinhibited by the interaction between SH3 and DHR-2, and this inhibition is relieved by its binding to ELMO (13). The three mammalian ELMO proteins (ELMO1 to ELMO3) consist of Ras-binding domain (RBD), ELMO-inhibitory domain (EID), ELMO domain, PH domain, and proline-rich region (PXXP) (Fig. 1A). ELMO acts as an autoinhibitory switch with open/closed conformations (14). The autoinhibition of ELMO is mediated by the interaction between EID and the C-terminal region preceding the PXXP. The RBD and EID are involved in binding of ELMO to its upstream regulators, such as the small GTPase RhoG (15), brain-specific angiogenesis inhibitor (BAI) family of adhesion G protein–coupled receptors (16), and Giα (17) and Gβγ subunits (18) of the heterotrimeric G protein. The binding mode of BAI1 to the N-terminal domain (NTD), including the RBD, EID, and ELMO domain, of ELMO2 has been recently elucidated (19). The C-terminal domain (CTD) of ELMO, including the PH domain and PXXP, is essential for binding to DOCK1 (20). Biochemical studies suggest that the PH domain of ELMO stabilizes its ternary complex with DOCK1 and Rac (11) and that the GEF activity of DOCK1 increases twofold upon being complexed with ELMO.

Fig. 1. Cryo–electron microscopy structure of human ELMO1-DOCK5-Rac1.

Fig. 1

Open in a new tab

(A) The domain organization of human ELMO1, DOCK5, and Rac1 is shown above the cryo–electron microscopy (cryo-EM) density map of the ELMO1-DOCK5-Rac1 complex (two orthogonal views). The coloring of the bars matches the coloring of the density map. Regions not included in the plasmid construct are shown in white. Regions lacking ordered density are shown in gray. NTD, N-terminal domain; CTD, C-terminal domain. (B) The fold of the ELMO1-DOCK5-Rac1 monomer is shown in the ribbon diagram. The domain coloring is consistent with (A).

Previous crystallographic studies have elucidated the structures of three major domains of DOCK. First, the catalytic DHR-2 domain adopts a three-lobed structure in which lobe A forms a homodimer interface and lobes B and C form the GTPase-binding site (21–23). The nucleotide sensor, which contains an invariant valine residue essential for GEF activity, is provided by lobe C. Second, the DHR-1 domain adopts a C2-domain fold with the positively charged phosphatidyl inositol-(3,4,5)-triphosphate (PIP3)–binding pocket (24). Third, the SH3 domain, together with the following helices, forms a rigid assembly with ELMO1 CTD through SH3-PXXP interactions and intermolecular helix bundle formation (23). However, these partial structures provided limited structural insight on how ELMO contributes to DOCK-catalyzed Rac activation.

During the preparation of this manuscript, the cryo–electron microscopy (cryo-EM) structures of the ELMO1-DOCK2-Rac1 complex at 4.1-Å resolution and ELMO1-DOCK2 complex at 6.0-Å resolution were reported (25). The overall architecture of the ELMO1-DOCK2 complex in the open and closed conformations was clarified, providing a model for the regulation of DOCK2 activation. However, atomic models of the DOCK proteins remain unexplored, except for the known domains.

DOCK5 has the highest sequence similarity with DOCK1 or DOCK180, a prototype of the DOCK family (3). DOCK5 regulates osteoclast function and is a potential target for the development of antiosteoporotic drugs (26). DOCK5 also functions as a key regulator of epithelial invasion and metastasis (27). By focusing on the active state of the ELMO-DOCK complex, we have determined the cryo-EM structure of the ELMO1-DOCK5-Rac1 complex at 3.8-Å resolution. Our high-resolution structure and mutagenesis studies revealed the mechanism by which ELMO modulates DOCK GEF activity. By creating an ELMO-DOCK-Cdc42 model structure, we also propose a hypothesis on how Rac selectivity can be mediated by ELMO.

RESULTS

Cryo-EM of the ELMO1-DOCK5-Rac1 complex

We coexpressed the first 1642 residues of human DOCK5 (of 1870 residues) (Fig. 1A) with human ELMO1 in human embryonic kidney–293 F cells to determine the structure of the ELMO1-DOCK5-Rac1 complex. The C terminus of DOCK5, which contains a proline-rich sequence (PXXP), is required for binding to the adaptor protein Crk (27). However, the C terminus was not included in the structural analysis because of its low sequence conservation between DOCK1 to DOCK5 and their orthologs in flies and worms (fig. S1). We purified the ELMO1-DOCK5-Rac1 complex using a nucleotide-deficient G15A mutant Rac1, which forms a high-affinity complex with Rho GEFs (28). The resulting ELMO1-DOCK5-Rac1 complex was dissociated during cryo-EM grid preparation. Subsequently, the complex was cross-linked and analyzed by cryo-EM.

We reconstructed a well-defined map of the ELMO1-DOCK5-Rac1 complex with dimensions of approximately 150 Å by 150 Å by 200 Å with a twofold (C2) symmetry (Fig. 1A). The final reconstruction of the complex was obtained at an overall resolution of 3.8 Å (figs. S2 and S3 and table S1). As flexibility in the periphery was observed, we performed focused refinement and improved the resolution to 4.2 Å (fig. S2). We generated all the atomic models of DOCK5 by homology modeling of the known domains (SH3, DHR-1, and DHR-2) (table S2) and de novo modeling of the remaining regions (detailed in the Supplementary Materials). We fitted the structures of nucleotide-free Rac1 and ELMO1 CTD. However, the density of the remaining region of ELMO1 could not be interpreted. The unresolved ELMO1 region corresponds to the NTD containing the RBD, EID, and ELMO domain (Fig. 1A).

Surface plasmon resonance (SPR) analysis showed that the ELMO1-DOCK5 complex used for structural analysis was specifically bound to the active RhoG (Q61L) mutant in the GTP-bound state (fig. S4). The equilibrium dissociation constant (KD) value of ELMO1-DOCK5 for the GTP-bound RhoG (Q61L) was 10.3 μM, which was comparable to the affinity of ELMO2 RBD for the GTP-bound RhoG (wild type, 7.8 μM), as reported by isothermal titration calorimetry (25). This suggested that the ELMO1 NTD was folded but disordered in the present cryo-EM structure.

Overall structure and biochemical activity of DOCK5

The ELMO1-DOCK5-Rac1 complex is a dimer of heterotrimers centered on a homodimer of the DOCK5 DHR-2 domain (Fig. 1A). Each heterotrimer forms a composite α-helical scaffold between ELMO1 CTD and DOCK5 (Fig. 1B). The CTD of ELMO1 interacts extensively with the DOCK5 SH3 and the subsequent helical region. The helical region interacts with the crook-shaped superhelix ARM (armadillo) in the center of DOCK5. The scaffold enables the ELMO1 PH domain, DOCK5 DHR-2 domain, and nucleotide-free Rac1 to interact with each other. A similar domain organization was recently reported in ELMO1-DOCK2-Rac1 (25). We analyzed the ELMO1-DOCK5-Rac1 complex by cross-linking mass spectrometry and identified intermolecular cross-links between Rac1, ELMO1, and DOCK5 (fig. S5 and table S3). In addition, interdomain cross-links were observed in ARM–SH3 and ARM–helical region in DOCK5 (fig. S5 and table S3). It is likely that these intercross-links stabilized the ELMO1-DOCK5-Rac1 complex and facilitated the cryo-EM analysis.

The DOCK5 fold identified here is similar to the reported DOCK2 fold in the ELMO1-DOCK2-Rac1 complex (25) [root mean square deviation (RMSD) of 3.4 Å over 1124 Cα atoms, z score of 15.9] in which the DOCK2 model is devoid of certain regions or is partially built as polyalanine. However, substantial differences were observed in the overall arrangement of the dimer between ELMO1-DOCK5-Rac1 and the reported ELMO1-DOCK2-Rac1 structures (fig. S6, A and B). In the side view with the DHR-2 domain at the bottom, the dimer of the ELMO1-DOCK5-Rac1 complex adopts a highly curved shape with each DHR-1 domain at the outer edge. In contrast, the dimer of the reported ELMO1-DOCK2-Rac1 complex assumes a relatively planar structure. In addition, the ELMO1-DOCK5-Rac1 complex was analyzed with a disordered ELMO1 NTD, whereas the reported ELMO1-DOCK2-Rac1 complex was analyzed with ELMO1 NTD in an “open” conformation (25).

Next, we examined the GEF activity of DOCK5. For this purpose, a full-length DOCK5 was prepared alone or in complex with ELMO1. The DOCK5 DHR-2 domain was also prepared alone. The GEF activity of DOCK5 DHR-2 was 80% lower than that of the intact DOCK5 (Fig. 2), suggesting that the isolated DOCK5 DHR-2 domain is unstable. The GEF activity of the ELMO1-DOCK5 complex was slightly higher (33%) than that of DOCK5 alone (Fig. 2). The GEF activity of the ELMO1ΔNTD-DOCK5 complex, which lacked the ELMO1 NTD, was much higher (290%) than that of DOCK5 alone (Fig. 2). This suggests that the NTD of ELMO1 exhibits an inhibitory role in the DOCK5 GEF activity. These results are in line with a recent report on DOCK2 (25). An important difference was that whereas the GEF activity of the ELMO1ΔNTD-DOCK5 complex was 220% higher than that of the intact ELMO1-DOCK5 complex (Fig. 2), the corresponding ELMO1ΔNTD-DOCK2 complex showed only an approximately 60% increase in GEF activity relative to the intact ELMO1-DOCK2 complex (25). Thus, we deduce that the ELMO1 NTD exhibits a distinct influence on the GEF activity of DOCK5 and DOCK2. This may be a reflection of the fact that ELMO1 NTD exists in different conformations upon being complexed with DOCK2 or DOCK5 (fig. S6A).

Fig. 2. In vitro GEF activities of the ELMO1-DOCK5 complex for Rac1.

Fig. 2

Open in a new tab

(A) The relative fluorescence intensities with addition of DOCK5 (green), ELMO1-DOCK5 (blue), ELMO1ΔNTR-DOCK5 (red), and DOCK5 DHR-2 (orange) in equal quantity (25 nM). The fluorescence intensities without any additions is shown in gray (closed squares). (B) The observed rate constants (kobs) of each GEF reaction determined from (A), subtracting the intrinsic exchange rate of Rac1. Data are presented as means ± SD (n = 3). **P < 0.001 and ***P < 0.0001 (two-tailed unpaired Student’s t test).

The superimposition of the monomer of the reported ELMO1-DOCK2-Rac1 complex onto that of the ELMO1-DOCK5-Rac1 complex revealed that the ELMO1 NTD in ELMO1-DOCK2-Rac1 causes steric hindrance with the DOCK5 C2 domain in ELMO1-DOCK5-Rac1 (fig. S6C). This may be due to a unique insertion in the C2 domain of DOCK5 (residues 355 to 375) (fig. S1). Thus, it is unlikely that the ELMO1 NTD in the present ELMO1-DOCK5-Rac1 complex adopts the open conformation like the reported ELMO1-DOCK2-Rac1 complex.

Interaction of Rac1 with the ELMO1 PH domain

In the case of ELMO1-DOCK2-Rac1 (25), side chains were visible in the region around the DHR-2 domain, but that of other regions were indistinct. The present structure of ELMO1-DOCK5-Rac1 provides further information on this aspect. Consistent with the previous crystallographic studies, the nucleotide-free Rac1 binds to the DOCK5 DHR-2 domain primarily through two switch regions (switches 1 and 2) (Fig. 3A). This allows the catalytic Val1559 residue within the nucleotide sensor of DOCK5 to access the nucleotide-binding site of Rac1. The interface between Rac1 and the PH domain of ELMO1 lies away from the switch regions and involves the side chains of Arg120 and Arg163 of Rac1 (Fig. 3B). These two basic residues are in proximity to Asp647 and Asn649 in the β6-β7 loop of the PH domain of ELMO1 and may interact electrostatically. Asp647 and Asn649 of ELMO1 are conserved or conservatively replaced by acidic residues in other ELMO proteins. In Rac1, Arg120 is located on the Rho insert helix and is conserved among Rho GTPases, whereas Arg163 proximal to the G5 box is specific to Rac1 to Rac3. Thus, this structure suggests that ELMO1 stabilizes the DOCK5 and Rac1 transition state complex through specific interactions between ELMO PH domain and Rac. The result also explains the marked increase in GEF activity of DOCK5 mediated by ELMO1 CTD in the ELMO1ΔNTD-DOCK5-Rac1 complex (Fig. 2).

Fig. 3. The ELMO1 PH domain assists the DOCK5 DHR-2 domain in stabilizing the nucleotide-free transition state of Rac1.

Fig. 3

Open in a new tab

(A) View of Rac1 in its nucleotide-free state between the ELMO1 PH domain (orange) and the DOCK5 DHR-2 domain (green). Rac1 is yellow except for switch 1 (red) and switch 2 (magenta). Lobe A of the DHR-2 domain is omitted for clarity. (B) Close-up view of the ternary interface between the ELMO1 PH domain (orange), DOCK5 DHR-2 domain (green), and Rac1 (yellow), superimposed on the crystal structure of the ELMO1-DOCK2 binding domain complex [Protein Data Bank (PDB) code 3A98] (gray). A red arrow indicates a conformational change in the β3-β4 loop of the ELMO1 PH domain. Residues of ELMO1 that were mutagenized are underlined. (C and D) The GEF activity of the wild-type (WT) and the mutated (C) ELMO1ΔNTD-DOCK5 complex and (D) full-length ELMO1-DOCK5 complex. Data are presented as means ± SD (n = 3). *P = 0.954, **P = 0.0143, and ***P < 0.001 (two-tailed unpaired Student’s t test).

To evaluate the effect of Rac1-ELMO1 PH domain interactions on the GEF activity of DOCK5, we generated a mutant ELMO1ΔNTD-DOCK5 complex in which Asp647 and Asn649 of ELMO1 are simultaneously replaced by alanine. The D647A/D649A mutation caused 28% reduction in the GEF activity of the ELMO1ΔNTD-DOCK5 complex (Fig. 3C). Similarly, in the context of full-length ELMO1-DOCK5 complex, 29% reduction of GEF activity was observed by the D647A/D649A mutation in ELMO1 (Fig. 3D). These results suggest the direct involvement of ELMO1 PH domain in the GEF activity of DOCK5.

A previous study showed that the W665A mutation of ELMO1 reduced its tendency to form a ternary complex with DOCK1 and Rac and also caused cell migration defects (12). Our structure shows that the Trp665 is located on Eα2 within the PH domain and forms a hydrophobic core in the domain (Fig. 3). Therefore, the W665A mutation is proposed to disrupt the ternary complex by destabilizing the folding of the PH domain.

Interaction of the ELMO1 PH domain with the DOCK5 DHR-2 domain

Similar to the interactions in the crystal structure of the complex formed by the binding domains of DOCK2 and ELMO1 (23), the 160-residue region at the N terminus of DOCK5 extensively interacts with the CTD of ELMO1 (Fig. 1B) and contains two contact areas: The first lies between the DOCK5 SH3 domain and the ELMO1 proline-rich region (residues 704 to 727), and the second lies between the DOCK5 helical region and the ELMO1 helices (Eα1 and Eα3) flanking the PH domain. Furthermore, the ELMO1-DOCK5-Rac1 complex structure reveals the third contact area between the DOCK5 DHR-2 domain and the ELMO1 PH domain (Fig. 3A). The interface is composed of an α6 within lobe B of the DOCK5 DHR-2 domain and a β3-β4 loop in the PH domain of ELMO1. In the ELMO1-DOCK5-Rac1 complex, the β3-β4 loop of the PH domain is closer to the DHR-2 domain as compared to that observed in the crystal structure of the ELMO1-DOCK2 binding domain complex (Fig. 3B). Because of this change, direct interactions between Asp602/Asp606 of ELMO1 and Arg1397 of DOCK5 are possible. In addition, Lys584 within the β2-β3 loop of the ELMO1 PH domain is in contact with Pro1403 in the C-terminal loop of the α6 of DOCK5.

Lys584 of ELMO1 and Pro1403 of DOCK5 are conserved among the ELMO and DOCK proteins, respectively (fig. S1), and thus, the corresponding interaction may occur between the ELMO1 to ELMO3 and DOCK1 to DOCK5 proteins. Conserved replacements of Asp602/Asp606 by Glu in the β3-β4 loop of ELMO1 occur in ELMO2 and ELMO3, and these residues may interact with Arg1397 of DOCK5 in a similar way. However, simultaneous mutations of ELMO1 Asp602 and Asp606 to alanine had no impact on the GEF activity of the ELMO1ΔNTD-DOCK5 complex (Fig. 3C). Therefore, the contribution of these interactions toward stabilizing the ELMO1-DOCK5-Rac1 complex may not be significant.

ARM and helical region

The region of approximately 600 residues (residue 628 to 1215) preceding the DHR-2 domain in DOCK5, termed ARM in DOCK2 (25), forms a low-curved superhelix with 10 helical repeats of mixed armadillo-like and HEAT-like repeats and helical insertions (Fig. 4A). A DALI search (29) revealed that the structure most similar to this superhelix is a domain termed HOOK occurring in telomere-associated protein Rif1 [Protein Data Bank (PDB) code 5nw5, z score of 12.3, RMSD of 7.7 Å over 334 Cα atoms]. Superposition of the two structures revealed that the fold of the first 10 of 12 repeats of Rif1 HOOK resembles that of DOCK5 ARM (fig. S7A). The ARM and lobe A (the DHR-2 domain) of DOCK5 constitute 13 consecutive helical repeats (Fig. 1B). The curved N terminus of ARM linked to the DHR-1 domain displayed considerable flexibility, as suggested by the lower resolution (fig. S3C).

Fig. 4. The α-helical scaffold of DOCK5 and ELMO1.

Fig. 4

Open in a new tab

(A) The ELMO1-DOCK5 scaffold is mediated by interaction of the DOCK5 helical region (dark green) with the DOCK5 ARM repeats (cyan) and the ELMO1 PH domain (orange). (B) The interface between the ELMO1 PH domain and the helical region of DOCK5, rotated 180° from (A). Residues of ELMO1 that were mutagenized (Fig. 3C) are underlined.

The fold of the DOCK5 ARM is in agreement with that reported for the DOCK2 ARM (PDB code 6tgb, z score of 13.5, RMSD of 2.6 Å over 517 Cα atoms) in which 335 residues (repeats 1 to 6) are built as polyalanine (25). Superposition of ELMO1-DOCK5-Rac1 with ELMO1-DOCK2-Rac1, based on lobe A, revealed that the relative arrangement of the ARM and lobe A differs substantially between DOCK2 and DOCK5 (fig. S7B). This results in an altered dimeric architecture between ELMO1-DOCK2-Rac1 and ELMO1-DOCK5-Rac1 (fig. S6, A and B).

The DOCK5 ARM uses its concave surface, the B helices of repeats 5 to 7, and the insert helices of repeat 8 for intramolecular interactions with the N-terminal helical region (Fig. 4A). Leu547 and Ile548, located at the middle of Eα1 helix in the ELMO1 PH domain (Fig. 4B), hydrophobically interact with the DOCK5 helical region. Mutation of these hydrophobic residues to alanine (L547A/I548A) abrogates the binding of DOCK1 (20). In the ELMO1ΔNTD-DOCK5 complex, the L547A/I548A mutation of ELMO1 reduced the GEF activity by 15% (Fig. 3C). This result suggested that the destabilization of the ELMO1 PH domain–DOCK5 helical region interactions affects the optimal interaction of the PH domain with Rac1.

Previous studies show that the G171E mutation in DOCK1 causes a significant reduction in GEF activity and ELMO binding (11, 13). The corresponding Gly170 of DOCK5 is located at the center of a 20-residue loop (Gly160-Ser179) connecting the three-helix bundle (α1-α3) and α4 (Fig. 4, A and B). Our cryo-EM structure suggests that the G170E mutation of DOCK5 causes steric hindrance with Pro86 in α1 and disrupts the α-helical scaffold of DOCK5, which is likely to be important for optimal ELMO binding.

DHR-1 and C2 domains

The interface of the DHR-1 domain with the preceding C2 domain is mediated by the edge of the β sandwich composed of two four-stranded β sheets (Fig. 5A). The DOCK5 equivalent Lys residues involved in PIP3 binding of DOCK1 create a positively charged pocket, which is expected to face the cell membrane (Fig. 5B).

Fig. 5. Two tandem C2-domain fold module of DOCK5.

Fig. 5

Open in a new tab

(A) Ribbon representation of the DOCK5 C2 and DHR-1 domains. Location of the nine-residue deletion (residues 506 to 514) in RLC mice are shown in yellow. (B) The surface electrostatic potentials of the DOCK5 C2 and DHR-1 domains in the same orientation as (A). (C and D) Superimposition of (C) Ca2+-bound synaptotagmin 1–C2B domain (PDB code 5CCH) and (D) PIP3-bound KIBRA C2 domain (PDB code 6FJC) onto the DOCK5 C2 domain. The DHR-1 domain adjacent to the DOCK5 C2 domain is shown in gray. (E) Two orthogonal views of the dimer of the ELMO1-DOCK5-Rac1 heterotrimer in ribbon representation. (F) The surface electrostatic potentials of the ELMO1-DOCK5-Rac1 complex. The molecular orientation is consistent with (E). Arrowheads indicate positively charged pockets on DOCK5 DHR-1.

Considering the connectivity to the DHR-1 domain, we proposed the structure of the preceding C2 domain to be an eight β-stranded “type I” C2 domain fold decorated by C-terminal insert α helices (Fig. 5A). According to the DALI search, the predicted C2 domain of DOCK5 closely resembles the C2B domain of synaptotagmin 1 (PDB code 5cch, z score of 17.6, RMSD of 1.4 Å over 127 Cα atoms) and C2 domain of KIBRA (kidney and brain expressed protein) (PDB code 6fjc, z score of 12.0, RMSD of 2.3 Å over 110 Cα atoms). By superposing these structures, we found that the Ca2+-binding site of synaptotagmin 1 corresponds to a negatively charged pocket in the C2 domain of DOCK5 (Fig. 5, B and C). However, the C2 domain of DOCK5 lacks the key aspartate residues conserved in the Ca2+-binding C2 domains and is therefore predicted to be incapable of Ca2+ binding. The PIP3-binding site of the KIBRA C2 domain corresponds to the buried surface of the DHR-1 domain interface of the C2 domain of DOCK5 (Fig. 5D). These findings suggested that the C2 domain of DOCK5 may partake in a structural role rather than in membrane binding. We observed the EM density derived from ELMO1 NTD in contact with the C2 domain at low thresholds (fig. S3D). The recent cryo-EM structure of the ELMO1-DOCK2 binary complex also showed that the C2 domain of DOCK2 was in contact with the ELMO1 NTD through the ELMO domain (25). It is therefore suggested that the C2 domain of DOCK plays a role in protein-protein interactions.

The dyad-symmetric dimer of the ELMO1-DOCK5-Rac1 complex allows each DHR-1 domain to simultaneously bind with the cell membrane and allows the DHR-2 domain to access the membrane-localized Rac1 (Fig. 5, E and F). As the DHR-1 and DHR-2 domains are widely separated by ARM repeats, the monomeric ELMO-DOCK heterodimer may not facilitate optimal binding of membrane-localized Rac to the DHR-2 domain. Mutations of DOCK2 residues at the homodimer interface lead to defects in cell migration, whereas GEF activity remains unaltered (30).

DISCUSSION

The cryo-EM structure of ELMO1-DOCK5-Rac1 revealed a dyad-symmetric dimer of heterotrimers of ELMO1 CTD, DOCK5, and Rac1 (Fig. 1). The structure shows that the PH domain of ELMO1 stabilizes the transition state of the DOCK5 (DHR-2)–Rac1 complex, providing the structural basis for ELMO1-mediated enhancement of the catalytic activity of DOCK5 (Fig. 2). The role of ELMO1 in the biochemical activity of DOCK5 was verified by structure-based mutagenesis (Fig. 3). The results showed that the PH domain of ELMO1 is involved in the modulation of the GEF activity of DOCK5 through specific interactions with Rac1.

ELMO binds only to Rac-specific DOCK GEFs (DOCK1 to DOCK5). Thus, we speculated that ELMO may also affect substrate specificity of DOCK GEFs. To test this possibility, we created a Cdc42-bound ELMO-DOCK model by replacing Cdc42–DHR-2 in DOCK9 (21) with Rac1–DHR-2 in the present cryo-EM structure (Fig. 6). As ELMO1 interacts with DHR-2 via lobe B (Fig. 3), we superposed Cdc42–DHR-2 and Rac1–DHR-2 based on lobe B. In this model, Cdc42 and lobe C shift closer to lobe B, and the PH domain of ELMO1 causes a collision with Cdc42. This is due to the different arrangement of lobes B and C among the structures of Cdc42-bound DHR-2 and Rac1-bound DHR-2 (22). These findings suggested that ELMO also contributes to the strict Rac specificity of DOCK1 to DOCK5.

Fig. 6. Differences in Rac1-bound and Cdc42-bound DHR-2 structures causes a collision of Cdc42 with ELMO1.

Fig. 6

Open in a new tab

(A) Ternary complex formation between the DOCK5 DHR-2 domain, Rac1, and the ELMO1 PH domain according to the cryo-EM structure. The color codes are as follows: orange, ELMO1; green, DOCK5; yellow, Rac1. (B) The crystal structure of the DOCK9 (DHR-2)–Cdc42 complex (PDB code 2WM9) is docked to ELMO1 by replacing DOCK5 (DHR-2)–Rac1 in (A). Blue, DOCK9; purple, Cdc42.

While the fold and assembly of the ELMO1-DOCK5-Rac1 monomer closely resembles that of the reported ELMO1-DOCK2-Rac1 monomer, the dimeric cryo-EM density observed for ELMO1-DOCK5-Rac1 substantially differs from the dimeric density of ELMO1-DOCK2-Rac1 (fig. S6). These variations may be attributed to the difference in the α-helical scaffold (fig. S7B). Thus, the architecture of the ELMO-DOCK complex may have variability even among the members of the closely related DOCK A/B subfamily. Further studies are required to determine the structural basis of ELMO-DOCK complexes involving other isoforms.

The deduced structure of ELMO1-DOCK5-Rac1 indicates that the ELMO1 NTD is intrinsically flexible and runs at the periphery of the DOCK5 C2 domain (fig. S3D). The existence of a poorly ordered, “closed” conformation of ELMO1 NTD is described in ELMO1-DOCK2-Rac1; however, only its open structure has been analyzed (25). The deletion of ELMO1 NTD resulted in a 220% increase in GEF activity of the ELMO1-DOCK5 complex (Fig. 2). This may be because ELMO1 NTD restricts Rac1 access to the DHR-2 domain of DOCK5. In DOCK2, ELMO1 NTD has been shown to directly interact with the DHR-2 domain in a binary complex with ELMO1 (25). In this closed structure of the ELMO1-DOCK2 binary complex, the DOCK2 linker sequence at the C terminus of α4 (referred to as the “phosphorylation linker”) forms an interface with the ELMO1 NTD (25). The phosphorylation linker sequence is less conserved in DOCK proteins of the A/B subfamilies; however, the Ser213 residue of three phosphorylation sites in DOCK2 is moderately conserved in DOCK2 to DOCK5 (fig. S1). It is therefore likely that the regulatory mechanisms of ELMO in these DOCK proteins are in common.

This structure also explains the molecular mechanism of disease-associated mutations in DOCK proteins. First, deletion of residues 506 to 514 of DOCK5 causes rupture of lens cataract (RLC) in mice (31). These DOCK5 residues lie within β4 of the DHR-1 domain and are located at the interface with the preceding C2 domain (Fig. 5A). Therefore, the deletion of these residues is expected to severely affect the structure and function of the two-tandem C2-domain fold module that facilitates membrane binding. Second, the R1104W mutation of DOCK2 has been identified in patients with early-onset invasive infections (32). Arg1104 of DOCK2 is conserved among DOCK family proteins (fig. S1), and the corresponding Arg1121 of DOCK5 is involved in the formation of ARM repeats (Fig. 4A). The mutation of this Arg residue to Trp is expected to destabilize the ARM repeats and the α-helical scaffold formed with ELMO1.

In summary, the high-resolution structure of the ELMO1-DOCK5-Rac1 complex and structure-based mutagenesis provided insights into the role of ELMO CTD in DOCK GEF activity and Rac specificity. The NTD of ELMO1 was found to be highly flexible in the present structure, and its regulatory role will be investigated in future structural studies.

MATERIALS AND METHODS

Expression and purification of the ELMO1-DOCK5-Rac1 complex

The gene encoding human DOCK5 (residues 1 to 1642) was cloned from Human Brain Whole Marathon-Ready complementary DNA (Takara) into the mammalian expression vector pOriP (33) harboring N-terminal FLAG and streptavidin-binding peptide (SBP) tags, followed by a tobacco etch virus (TEV) protease cleavage site. A gene encoding human ELMO1 (residues 1 to 727) (OriGene) was cloned in the same way into the pOriP vector but without the SBP tag. Nucleotide-deficient G15A mutant of human Rac1 (residues 1 to 177) was synthesized by the Escherichia coli cell–free system and purified using a previously described protocol for the wild-type protein (23).

The DOCK5 and ELMO1 expression vectors were cotransfected into FreeStyle 293-F cells (Life Technologies) using 293fectin transfection reagent (Life Technologies) according to the manufacturer’s protocol. The cells were grown in a FreeStyle 293 Expression medium (Life Technologies) using an incubation shaker equipped with a CO2 controller (TAITEC) and were harvested 48 hours after transfection. The cells were resuspended in lysis buffer [100 mM tris-HCl buffer at pH 8.0, 300 mM NaCl, 100 mM arginine-HCl, 100 mM glutamic acid–Na, 500 mM sucrose, 1 mM EDTA, and 5 mM dithiothreitol (DTT)] containing cOmplete protease inhibitor cocktail (Roche) and then disrupted by sonication. The lysate was cleared by centrifugation, and the supernatant was loaded onto streptavidin Sepharose beads (GE Healthcare Life Sciences), which were pre-equilibrated with wash buffer [100 mM tris-HCl at pH 8.0, 300 mM NaCl, 100 mM arginine-HCl, 100 mM glutamic acid–Na, 500 mM sucrose, 1 mM EDTA, 5 mM DTT, and 2 mM tris(2-carboxyethyl)phosphine (TCEP)]. Each product was eluted with the same buffer containing 4 mM desthiobiotin, and then the affinity tags were cleaved by TEV protease at 4°C overnight. To prepare the ELMO1-DOCK5-Rac1 complex, an excess amount of purified Rac1 (G15A) was added and incubated at the same time. The cleaved tags, protease, and unbound Rac1 were removed by size exclusion chromatography on a HiLoad 16/600 Superose 6 pg column (GE Healthcare Life Sciences), pre-equilibrated with 20 mM Hepes-NaOH buffer (pH 8.0), 300 mM NaCl, and 1 mM TCEP. The purified ELMO1-DOCK5-Rac1 complex was incubated with 1 mM bis(sulfosuccinimidyl)suberate (BS3) for 15 min at room temperature, quenched for 10 min using 50 mM tris-HCl (pH 8.0), and subjected to the second size exclusion chromatography on a HiLoad 16/600 Superose 6 pg column. Peak fractions were concentrated to 5 to 6 mg/ml and used for cryo-EM analysis.

Cryo-EM grid preparation and data collection

Quantifoil holey carbon grids (R1.2/1.3, 300 mesh copper, Quantifoil Micro Tools GmbH) were covered with a thin layer of continuous carbon film, prepared using a JEE-420 vacuum evaporator (JEOL). Immediately before use, the carbon-coated grids were glow-discharged at 5 mA for 9 s with a PIB-10 plasma ion bombarder (Vacuum Device). Protein sample (3 μl of 200 nM protein containing 0.06% digitonin) was absorbed onto the carbon film for 30 s at 4°C in 100% humidity, blotted for 1 s, and plunge-frozen in liquid ethane using Vitrobot Mark IV (Thermo Fisher Scientific).

Cryo-EM imaging of the complex was performed on a 300-kV Titan Krios G3i microscope (Thermo Fisher Scientific) equipped with a GIF Quantum LS energy filter (Gatan) and a K3 direct electron detector (Gatan) operating in the electron counting mode. Micrograph movies were acquired at a nominal magnification of ×105,000, corresponding to a calibrated pixel size of 0.83 Å per pixel. Each movie was recorded for 3 s and subdivided into 60 frames. The electron flux rate was set to 14 e− per pixel per second at the detector, resulting in an accumulated exposure of 60 e−/Å2 at the specimen. The data were automatically acquired by the image shift method using SerialEM (34), with a defocus range of −0.5 to −3.0 μm. A total of 4689 movies were recorded.

Cryo-EM image processing

Movie micrographs were motion corrected by MotionCor2 (35) with dose weighting. The contrast transfer function (CTF) parameters of the motion-corrected micrographs were estimated using CTFFIND4 (36), followed by two-dimensional (2D)/3D classification and 3D refinement using RELION-3 (37) and 3.1 (fig. S2). Initially, the 2D references for particle picking were generated by automatic particle picking and 2D classification. A total of 2,145,777 particles were subsequently picked and extracted with down-sampled pixel size of 3.32 Å per pixel. After 2D classification, 371,792 particles were selected and subjected to 3D classification to select 149,846 homogeneous particles. These particles were re-extracted with an original pixel size of 0.83 Å per pixel and used for 3D refinement by imposing the C2 symmetry, followed by Bayesian polishing and CTF refinement. This resulted in a final map with an overall resolution of 3.8 Å according to fourier shell correlation (FSC) = 0.143 criterion. Density modification of the refined reconstructions was performed with PHENIX (38). To further improve the map of the DHR-1 domain region, the particles were expanded to 299,692 subparticles of monomeric form, and subsequent 3D classification was performed using a mask around the DHR-1 domain region. This resulted in the selection of 62,455 particles with a better density at the DHR-1 domain region. The 3D refinement of these particles resulted in a map with a resolution of 4.2 Å, which was used to interpret the DHR-1 domain region.

Model building

The DOCK5 protein was divided into different regions based on domain identification and availability of homolog protein structures in viz. DOCK5(N-terminal), DOCK5(UR1), DOCK5(DHR-1), DOCK5(UR2), and DOCK5(DHR-2).

Homology modeling of DOCK5(N-terminal), DOCK5(DHR-1), DOCK5(DHR-2), RAC1, and ELMO1

The N terminus of DOCK5 protein (1 to 176) was directly modeled through a homology-based approach for protein structure prediction by using the structural information of the DOCK2 protein (3A98_A) in the crystal structure of the DOCK2-ELMO1 protein complex. On the basis of the cryo-EM map for the region corresponding to DOCK5(N-terminal) and domain analysis, an additional helix was modeled to extend the model structure to 215 residues (1 to 215). Similarly, DOCK5(DHR-1) region (436 to 640) was modeled using the structural information of DOCK1(DHR-1) protein (3L4C_A), and the DOCK5(DHR-2) region (1212 to 1642) was modeled using the structural details of DOCK2(DHR-2) in the DOCK2-Rac1 protein complex (3B13_A). Further, the structural information of Rac1 protein (3B13_B) in the crystal structure of the DOCK2-Rac1 protein complex was used to model the Rac1 protein, and the structural information of ELMO1 protein (3A98_B) in the crystal structure of the DOCK2-ELMO1 protein complex was used to model ELMO1 protein in the ELMO1-DOCK5-Rac1 complex. The model building of different regions of the ELMO1-DOCK5-Rac1 protein complex through a homology-based approach is summarized in table S2.

De novo modeling for DOCK5(UR1) and DOCK5(UR2)

The subsequent regions of DOCK5 protein (221 to 430 and 646 to 1210) were named as unsolved region 1 (UR1) and UR2, respectively. For UR1 and UR2, no suitable homolog structure could be identified in PDB, and the de novo approach for protein structure prediction was implemented to model these regions. The protein sequences of DOCK5(UR1) and DOCK5(UR2) were screened against various sequence databases to retrieve some additional information for DOCK5(UR1) and DOCK5(UR2). The sequence analysis of DOCK5(UR2) using InterPro Classification and Superfamily Database screening revealed that it is an armadillo-type fold. The same was further supported by the presence of helical repeats in the predicted secondary structure of DOCK5(UR2). The predicted model structures without the expected armadillo-type fold were rejected. However, the sequence database screening failed to provide any additional insights about DOCK5(UR1). An evolutionary coupling-directed contact map–based de novo approach was implemented to predict the model structure for DOCK5(UR1). Different de novo approaches used for modeling DOCK5(UR1) and DOCK5(UR2) are further discussed in Supplementary Materials and Methods.

Surface plasmon resonance

For the preparation of SPR samples, the gene encoding human RhoG (residues 1 to 184) was cloned into the pCR2.1 vector (Invitrogen) with an N-terminal His tag and TEV protease cleavage site. A constitutively activated Q61L mutation was introduced into the RhoG sequence with a QuikChange site-directed mutagenesis kit (Agilent Technologies). Wild-type and mutant proteins were synthesized using the large-scale dialysis mode of the E. coli cell–free reaction (39, 40). The proteins were purified using His-tag affinity column chromatography. After digestion with TEV protease, proteins were further purified by size exclusion chromatography on a HiLoad 16/60 Superdex 75 pg column (GE Healthcare Life Sciences). For the purification of the Q61L mutant protein, 1 mM MgCl2 and 10 μM GTP were added during the process.

SPR experiments were performed on a Biacore T200 instrument (GE Healthcare Life Sciences). ELMO1-DOCK5 (1 to 1642) complex was immobilized on a CM5 sensor chip using the Amine Coupling Kit (GE Healthcare Life Sciences). The buffer contained 10 mM Hepes (pH 7.5), 150 mM NaCl, 3 mM MgCl2, and 0.005% surfactant P-20. Five different concentrations (1.8 to 28.8 μM) of RhoG (wild-type or Q61L mutant) were injected continuously, and the response was measured. Data were processed using the manufacturer’s software by steady-state affinity analysis with five sample concentrations. The KD value for the RhoG (Q61L) mutant was determined using the average of the values obtained in three experiments.

In vitro GEF assays

Samples for the in vitro GEF assays were prepared by cloning the full-length DOCK5 (residues 1 to 1870) and the DOCK5 DHR-2 domain (residues 1216 to 1642) into the pOriP vector in the same way as that done for structural analysis. Measurement of the exchange reaction was performed using a previously reported protocol (25) with slight modifications. GEF proteins (25 nM) (DOCK5 alone or ELMO1-DOCK5 complex) or control buffer was incubated at 25°C with 1.6 μM fluorescent boron-dipyrromethene-fluor (BODIPY-FL)–GDP–loaded Rac1 and 100 μM GTP in a reaction mixture containing 20 mM tris-HCl (pH 7.0), 150 mM NaCl, 10 mM MgCl2, and bovine serum albumin (0.2 mg/ml). Release of BODIPY-FL-GDP by Rac1 was measured by monitoring the decrease in fluorescence at excitation/emission wavelengths of 485/535 nm on an ARVO X3 spectrofluorimeter (PerkinElmer). The observed rate constants (kobs) of each reaction were determined by nonlinear least-squares fitting of the data with a single exponential decay model using KaleidaGraph software (Synergy Software).

Acknowledgments

We acknowledge RIKEN ACCC for the supercomputing resources at the Hokusai BigWaterfall. We also thank the staff scientists at the University of Tokyo’s cryo-EM facility, especially A. Tsutsumi and M. Kikkawa. Funding: This work was supported by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science by the Japan Science and Technology Agency (nos. 15K06987 and 19K06575 to M.K.-N. and no. 18H02395 to K.Y.J.Z.) and by the Platform Project for Supporting Drug Discovery and Life Science Research [Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)] from the Japan Agency for Medical Research and Development (AMED) under grant no. JP20am0101082 (to M.S.) and JP20am0101115 (support no. 1507). We also acknowledge the RIKEN Dynamic Structural Biology Project for support. Author contributions: M.K.-N. and M.S. conceived and designed the study. K.K. and T.U.-K. performed cryo-EM experiments. M.K.-N., K.K., H.E., and T.Y. performed image processing. R.N. performed cross-linking mass spectrometry. C.M.-T., M.Y., M.I., and K.H. expressed and purified proteins. R.K. performed model building under the guidance of K.Y.J.Z. M.K.-N. performed structure determination, SPR, and GEF assays and analyzed the data. M.K.-N. wrote the manuscript with input from others. M.K.-N., K.Y.J.Z., and M.S. secured funding. M.S. supervised the study. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The electron density maps and model of the ELMO1-DOCK5-Rac1 complex have been deposited in the Electron Microscopy Data Bank with ID EMD-30802 and in the PDB with 7DPA, respectively. Cross-linking mass spectrometry data have been deposited to ProteomeXchange (accession no. PXD026097) via jPOST (accession no. JPST001178).

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/7/30/eabg3147/DC1

View/request a protocol for this paper from Bio-protocol.

REFERENCES AND NOTES

  • 1.Brugnera E., Haney L., Grimsley C., Lu M., Walk S. F., Tosello-Trampont A. C., Macara I. G., Madhani H., Fink G. R., Ravichandran K. S., Unconventional Rac-GEF activity is mediated through the Dock180-ELMO complex. Nat. Cell Biol. 4, 574–582 (2002). [DOI] [PubMed] [Google Scholar]
  • 2.Côté J.-F., Vuori K., Identification of an evolutionarily conserved superfamily of DOCK180-related proteins with guanine nucleotide exchange activity. J. Cell Sci. 115, 4901–4913 (2002). [DOI] [PubMed] [Google Scholar]
  • 3.Meller N., Merlot S., Guda C., CZH proteins: A new family of Rho-GEFs. J. Cell Sci. 118, 4937–4946 (2005). [DOI] [PubMed] [Google Scholar]
  • 4.Miyamoto Y., Yamauchi J., Sanbe A., Tanoue A., Dock6, a Dock-C subfamily guanine nucleotide exchanger, has the dual specificity for Rac1 and Cdc42 and regulates neurite outgrowth. Exp. Cell Res. 313, 791–804 (2007). [DOI] [PubMed] [Google Scholar]
  • 5.Kukimoto-Niino M., Tsuda K., Ihara K., Mishima-Tsumagari C., Honda K., Ohsawa N., Shirouzu M., Structural basis for the dual substrate specificity of DOCK7 guanine nucleotide exchange factor. Structure 27, 741–748.e3 (2019). [DOI] [PubMed] [Google Scholar]
  • 6.Ruiz-Lafuente N., Alcaraz-García M. J., García-Serna A. M., Sebastián-Ruiz S., Moya-Quiles M. R., García-Alonso A. M., Parrado A., Dock10, a Cdc42 and Rac1 GEF, induces loss of elongation, filopodia, and ruffles in cervical cancer epithelial HeLa cells. Biol. Open 4, 627–635 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Laurin M., Côté J. F., Insights into the biological functions of Dock family guanine nucleotide exchange factors. Genes Dev. 28, 533–547 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Gadea G., Blangy A., Dock-family exchange factors in cell migration and disease. Eur. J. Cell Biol. 93, 466–477 (2014). [DOI] [PubMed] [Google Scholar]
  • 9.Côté J. F., Motoyama A. B., Bush J. A., Vuori K., A novel and evolutionarily conserved PtdIns(3,4,5)P3-binding domain is necessary for DOCK180 signalling. Nat. Cell Biol. 7, 797–807 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Gumienny T. L., Brugnera E., Tosello-Trampont A. C., Kinchen J. M., Haney L. B., Nishiwaki K., Walk S. F., Nemergut M. E., Macara I. G., Francis R., Schedl T., Qin Y., Van Aelst L., Hengartner M. O., Ravichandran K. S., CED-12/ELMO, a novel member of the CrkII/Dock180/Rac pathway, is required for phagocytosis and cell migration. Cell 107, 27–41 (2001). [DOI] [PubMed] [Google Scholar]
  • 11.Lu M., Kinchen J. M., Rossman K. L., Grimsley C., deBakker C., Brugnera E., Tosello-Trampont A. C., Haney L. B., Klingele D., Sondek J., Hengartner M. O., Ravichandran K. S., PH domain of ELMO functions in trans to regulate Rac activation via Dock180. Nat. Struct. Mol. Biol. 11, 756–762 (2004). [DOI] [PubMed] [Google Scholar]
  • 12.Grimsley C. M., Kinchen J. M., Tosello-Trampont A. C., Brugnera E., Haney L. B., Lu M., Chen Q., Klingele D., Hengartner M. O., Ravichandran K. S., Dock180 and ELMO1 proteins cooperate to promote evolutionarily conserved Rac-dependent cell migration. J. Biol. Chem. 279, 6087–6097 (2004). [DOI] [PubMed] [Google Scholar]
  • 13.Lu M., Kinchen J. M., Rossman K. L., Grimsley C., Hall M., Sondek J., Hengartner M. O., Yajnik V., Ravichandran K. S., A steric-inhibition model for regulation of nucleotide exchange via the Dock180 family of GEFs. Curr. Biol. 15, 371–377 (2005). [DOI] [PubMed] [Google Scholar]
  • 14.Patel M., Margaron Y., Fradet N., Yang Q., Wilkes B., Bouvier M., Hofmann K., Côté J. F., An evolutionarily conserved autoinhibitory molecular switch in ELMO proteins regulates Rac signaling. Curr. Biol. 20, 2021–2027 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Katoh H., Negishi M., RhoG activates Rac1 by direct interaction with the Dock180-binding protein Elmo. Nature 424, 461–464 (2003). [DOI] [PubMed] [Google Scholar]
  • 16.Park D., Tosello-Trampont A. C., Elliott M. R., Lu M., Haney L. B., Ma Z., Klibanov A. L., Mandell J. W., Ravichandran K. S., BAI1 is an engulfment receptor for apoptotic cells upstream of the ELMO/Dock180/Rac module. Nature 450, 430–434 (2007). [DOI] [PubMed] [Google Scholar]
  • 17.Li H., Yang L., Fu H., Yan J., Wang Y., Guo H., Hao X., Xu X., Jin T., Zhang N., Association between Gαi2 and ELMO1/Dock180 connects chemokine signalling with Rac activation and metastasis. Nat. Commun. 4, 1706 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Wang Y., Xu X., Pan M., Jin T., ELMO1 directly interacts with Gβγ subunit to transduce GPCR signaling to Rac1 activation in chemotaxis. J. Cancer 7, 973–983 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Weng Z., Situ C., Lin L., Wu Z., Zhu J., Zhang R., Structure of BAI1/ELMO2 complex reveals an action mechanism of adhesion GPCRs via ELMO family scaffolds. Nat. Commun. 10, 51 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Komander D., Patel M., Laurin M., Fradet N., Pelletier A., Barford D., Côté J. F., An alpha-helical extension of the ELMO1 pleckstrin homology domain mediates direct interaction to DOCK180 and is critical in Rac signaling. Mol. Biol. Cell 19, 4837–4851 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Yang J., Zhang Z., Roe S. M., Marshall C. J., Barford D., Activation of Rho GTPases by DOCK exchange factors is mediated by a nucleotide sensor. Science 325, 1398–1402 (2009). [DOI] [PubMed] [Google Scholar]
  • 22.Kulkarni K., Yang J., Zhang Z., Barford D., Multiple factors confer specific Cdc42 and Rac protein activation by dedicator of cytokinesis (DOCK) nucleotide exchange factors. J. Biol. Chem. 286, 25341–25351 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Hanawa-Suetsugu K., Kukimoto-Niino M., Mishima-Tsumagari C., Akasaka R., Ohsawa N., Sekine S., Ito T., Tochio N., Koshiba S., Kigawa T., Terada T., Shirouzu M., Nishikimi A., Uruno T., Katakai T., Kinashi T., Kohda D., Fukui Y., Yokoyama S., Structural basis for mutual relief of the Rac guanine nucleotide exchange factor DOCK2 and its partner ELMO1 from their autoinhibited forms. Proc. Natl. Acad. Sci. U.S.A. 109, 3305–3310 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Premkumar L., Bobkov A. A., Patel M., Jaroszewski L., Bankston L. A., Stec B., Vuori K., Côté J. F., Liddington R. C., Structural basis of membrane targeting by the Dock180 family of Rho family guanine exchange factors (Rho-GEFs). J. Biol. Chem. 285, 13211–13222 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Chang L., Yang J., Jo C. H., Boland A., Zhang Z., McLaughlin S. H., Abu-Thuraia A., Killoran R. C., Smith M. J., Côté J. F., Barford D., Structure of the DOCK2−ELMO1 complex provides insights into regulation of the auto-inhibited state. Nat. Commun. 11, 3464 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Vives V., Laurin M., Cres G. G., Larrousse P., Morichaud Z., Noel D., Côté J. F., Blangy A., The Rac1 exchange factor Dock5 is essential for bone resorption by osteoclasts. J. Bone Miner. Res. 26, 1099–1110 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Frank S. R., Köllmann C. P., van Lidth de Jeude J. F., Thiagarajah J. R., Engelholm L. H., Frödin M., Hansen S. H., The focal adhesion-associated proteins DOCK5 and GIT2 comprise a rheostat in control of epithelial invasion. Oncogene 36, 1816–1828 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.García-Mata R., Wennerberg K., Arthur W. T., Noren N. K., Ellerbroek S. M., Burridge K., Analysis of activated GAPs and GEFs in cell lysates. Methods Enzymol. 406, 425–437 (2006). [DOI] [PubMed] [Google Scholar]
  • 29.Holm L., DALI and the persistence of protein shape. Protein Sci. 29, 128–140 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Terasawa M., Uruno T., Mori S., Kukimoto-Niino M., Nishikimi A., Sanematsu F., Tanaka Y., Yokoyama S., Fukui Y., Dimerization of DOCK2 is essential for DOCK2-mediated Rac activation and lymphocyte migration. PLOS ONE 7, e46277 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Omi N., Kiyokawa E., Matsuda M., Kinoshita K., Yamada S., Yamada K., Matsushima Y., Wang Y., Kawai J., Suzuki M., Hayashizaki Y., Hiai H., Mutation of Dock5, a member of the guanine exchange factor Dock180 superfamily, in the rupture of lens cataract mouse. Exp. Eye Res. 86, 828–834 (2008). [DOI] [PubMed] [Google Scholar]
  • 32.Dobbs K., Domínguez Conde C., Zhang S. Y., Parolini S., Audry M., Chou J., Haapaniemi E., Keles S., Bilic I., Okada S., Massaad M. J., Rounioja S., Alwahadneh A. M., Serwas N. K., Capuder K., Çiftçi E., Felgentreff K., Ohsumi T. K., Pedergnana V., Boisson B., Haskoloğlu Ş., Ensari A., Schuster M., Moretta A., Itan Y., Patrizi O., Rozenberg F., Lebon P., Saarela J., Knip M., Petrovski S., Goldstein D. B., Parrott R. E., Savas B., Schambach A., Tabellini G., Bock C., Chatila T. A., Comeau A. M., Geha R. S., Abel L., Buckley R. H., İkincioğulları A., Al-Herz W., Helminen M., Doğu F., Casanova J. L., Boztuğ K., Notarangelo L. D., Inherited DOCK2 deficiency in patients with early-onset invasive infections. N. Engl. J. Med. 372, 2409–2422 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Mukai T., Kobayashi T., Hino N., Yanagisawa T., Sakamoto K., Yokoyama S., Adding l-lysine derivatives to the genetic code of mammalian cells with engineered pyrrolysyl-tRNA synthetases. Biochem. Biophys. Res. Commun. 371, 818–822 (2008). [DOI] [PubMed] [Google Scholar]
  • 34.Mastronarde D. N., Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol. 152, 36–51 (2005). [DOI] [PubMed] [Google Scholar]
  • 35.Zheng S. Q., Palovcak E., Armache J. P., Verba K. A., Cheng Y., Agard D. A., MotionCor2: Anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods 14, 331–332 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Rohou A., Grigorieff N., CTFFIND4: Fast and accurate defocus estimation from electron micrographs. J. Struct. Biol. 192, 216–221 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Zivanov J., Nakane T., Forsberg B. O., Kimanius D., Hagen W. J., Lindahl E., Scheres S. H., New tools for automated high-resolution cryo-EM structure determination in RELION-3. eLife 7, e42166 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Terwilliger T. C., Ludtke S. J., Read R. J., Adams P. D., Afonine P. V., Improvement of cryo-EM maps by density modification. Nat. Methods 17, 923–927 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Terada T., Murata T., Shirouzu M., Yokoyama S., Cell-free expression of protein complexes for structural biology. Methods Mol. Biol. 1091, 151–159 (2014). [DOI] [PubMed] [Google Scholar]
  • 40.Katsura K., Matsuda T., Tomabechi Y., Yonemochi M., Hanada K., Ohsawa N., Sakamoto K., Takemoto C., Shirouzu M., A reproducible and scalable procedure for preparing bacterial extracts for cell-free protein synthesis. J. Biochem. 162, 357–369 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Song Y., DiMaio F., Wang R.-Y., Kim D., Miles C., Brunette T., Thompson J., Baker D., High-resolution comparative modeling with RosettaCM. Structure 21, 1735–1742 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Yang J., Yan R., Roy A., Xu D., Poisson J., Zhang Y., The I-TASSER Suite: Protein structure and function prediction. Nat. Methods 12, 7–8 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Xu D., Zhang Y., Ab initio protein structure assembly using continuous structure fragments and optimized knowledge-based force field. Proteins 80, 1715–1735 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Li J., Deng X., Eickholt J., Cheng J., Designing and benchmarking the MULTICOM protein structure prediction system. BMC Struct. Biol. 13, 2 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Källberg M., Wang H., Wang S., Peng J., Wang Z., Lu H., Xu J., Template-based protein structure modeling using the RaptorX web server. Nat. Protoc. 7, 1511–1522 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Simoncini D., Schiex T., Zhang K. Y., Balancing exploration and exploitation in population-based sampling improves fragment-based de novo protein structure prediction. Proteins 85, 852–858 (2017). [DOI] [PubMed] [Google Scholar]
  • 47.Colovos C., Yeates T. O., Verification of protein structures: Patterns of nonbonded atomic interactions. Protein Sci. 2, 1511–1519 (1993). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Eisenberg D., Lüthy R., Bowie J. U., VERIFY3D: Assessment of protein models with three-dimensional profiles. Methods Enzymol. 277, 396–404 (1997). [DOI] [PubMed] [Google Scholar]
  • 49.Maghrabi A. H. A., McGuffin L. J., ModFOLD6: An accurate web server for the global and local quality estimation of 3D protein models. Nucleic Acids Res. 45, W416–W421 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Singh A., Kaushik R., Mishra A., Shanker A., Jayaram B., ProTSAV: A protein tertiary structure analysis and validation server. Biochim. Biophys. Acta 1864, 11–19 (2016). [DOI] [PubMed] [Google Scholar]
  • 51.Siew N., Elofsson A., Rychlewski L., Fischer D., MaxSub: An automated measure for the assessment of protein structure prediction quality. Bioinformatics 16, 776–785 (2000). [DOI] [PubMed] [Google Scholar]
  • 52.Berenger F., Shrestha R., Zhou Y., Simoncini D., Zhang K. Y. J., Durandal: Fast exact clustering of protein decoys. J. Comput. Chem. 33, 471–474 (2012). [DOI] [PubMed] [Google Scholar]
  • 53.Berenger F., Zhou Y., Shrestha R., Zhang K. Y. J., Entropy-accelerated exact clustering of protein decoys. Bioinformatics 27, 939–945 (2011). [DOI] [PubMed] [Google Scholar]
  • 54.Heo L., Park H., Seok C., GalaxyRefine: Protein structure refinement driven by side-chain repacking. Nucleic Acids Res. 41, W384–W388 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Pettersen E. F., Goddard T. D., Huang C. C., Couch G. S., Greenblatt D. M., Meng E. C., Ferrin T. E., UCSF Chimera—A visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004). [DOI] [PubMed] [Google Scholar]
  • 56.Afonine P. V., Poon B. K., Read R. J., Sobolev O. V., Terwilliger T. C., Urzhumtsev A., Adams P. D., Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr. D Struct. Biol 74, 531–544 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Wang S., Sun S., Li Z., Zhang R., Xu J., Accurate de novo prediction of protein contact map by ultra-deep learning model. PLOS Comp. Biol. 13, e1005324 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Emsley P., Lohkamp B., Scott W., Cowtan K., Features and development of Coot. Acta Crystallogr. D 66, 486–501 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Williams C. J., Headd J. J., Moriarty N. W., Prisant M. G., Videau L. L., Deis L. N., Verma V., Keedy D. A., Hintze B. J., Chen V. B., Jain S., Lewis S. M., Arendall W. B. III, Snoeyink J., Adams P. D., Lovell S. C., Richardson J. S., Richardson D. C., MolProbity: More and better reference data for improved all-atom structure validation. Protein Sci. 27, 293–315 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Liu F., Rijkers D. T., Post H., Heck A. J., Proteome-wide profiling of protein assemblies by cross-linking mass spectrometry. Nat. Methods 12, 1179–1184 (2015). [DOI] [PubMed] [Google Scholar]
  • 61.Schmidt R., Sinz A., Improved single-step enrichment methods of cross-linked products for protein structure analysis and protein interaction mapping. Anal. Bioanal. Chem. 409, 2393–2400 (2017). [DOI] [PubMed] [Google Scholar]
  • 62.Klykov O., Steigenberger B., Pektaş S., Fasci D., Heck A. J. R., Scheltema R. A., Efficient and robust proteome-wide approaches for cross-linking mass spectrometry. Nat. Protoc. 13, 2964–2990 (2018). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/7/30/eabg3147/DC1

Articles from Science Advances are provided here courtesy of American Association for the Advancement of Science

RESOURCES