- Email: [email protected]
Rickettsia Sca2 Recruits Two Actin Subunits for Nucleation but Lacks WH2 Domains
Article
Rickettsia Sca2 Recruits Two Actin Subunits for Nucleation but Lacks WH2 Domains Saif S. Alqassim,1,2 In-Gyun Lee,1 and Roberto Dominguez1,* 1 Department of Physiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania and 2College of Medicine, Mohammed Bin Rashid University of Medicine and Health Sciences, Dubai, United Arab Emirates
ABSTRACT The Rickettsia 1800-amino-acid autotransporter protein surface cell antigen 2 (Sca2) promotes actin polymerization on the surface of the bacterium to drive its movement using an actin comet-tail mechanism. Sca2 mimics eukaryotic formins in that it promotes both actin filament nucleation and elongation and competes with capping protein to generate filaments that are long and unbranched. However, despite these functional similarities, Sca2 is structurally unrelated to eukaryotic formins and achieves these functions through an entirely different mechanism. Thus, while formins are dimeric, Sca2 functions as a monomer. However, Sca2 displays intramolecular interactions and functional cooperativity between its N- and C-terminal domains that are crucial for actin nucleation and elongation. Here, we map the interaction of N- and C- terminal fragments of Sca2 and their contribution to actin binding and nucleation. We find that both the N- and C-terminal regions of Sca2 interact with actin monomers but only weakly, whereas the full-length protein binds two actin monomers with high affinity. Moreover, deletions at both ends of the N- and C-terminal regions disrupt their ability to interact with each other, suggesting that they form a contiguous ring-like structure that wraps around two actin subunits, analogous to the formin homology-2 domain. The discovery of Sca2 as an actin nucleator followed the identification of what appeared to be a repeat of three Wiskott-Aldrich syndrome homology 2 (WH2) domains in the middle of the molecule, consistent with the presence of WH2 domains in most actin nucleators. However, we show here that contrary to previous assumptions, Sca2 does not contain WH2 domains. Instead, our analysis indicates that the region containing the putative WH2 domains is folded as a globular domain that cooperates with other parts of the Sca2 molecule for actin binding and nucleation.
INTRODUCTION Numerous pathogens, including several bacteria and some viruses, take advantage of the actin cytoskeleton of the host eukaryotic cell for motility and infection (1–3). A common mechanism relies on the activation of actin nucleation on the surface of pathogens to generate actin comet tails that propel their movement (4). Pathogens that use this mechanism typically nucleate actin by exposing proteins that mimic eukaryotic nucleation-promoting factors, which recruit and activate the host Arp2/3 complex (5). However, some bacteria encode their own actin nucleators, i.e., proteins capable of directly activating the polymerization of eukaryotic actin, including Vibrio VopL/F (6–9), Burkholderia BimA (10), and Rickettsia Sca2 (11–13). Such bacterial nucleators have evolved independently of their eukaryotic counterparts, and they tend to polymerize actin using
Submitted November 9, 2018, and accepted for publication December 12, 2018. *Correspondence: [email protected] Editor: Laurent Blanchoin. https://doi.org/10.1016/j.bpj.2018.12.009 Ó 2018 Biophysical Society.
540 Biophysical Journal 116, 540–550, February 5, 2019
different molecular mechanisms than the eukaryotic nucleators (4). This feature, along with the requirement of actin-based motility for infection, highlights the importance of studying bacterial nucleators for a better understanding of both pathogenicity and actin nucleation mechanisms. Here, we focus on surface cell antigen 2 (Sca2), a surface protein of spotted fever Rickettsiae. Rickettsiae are intracellular Gram-negative pathogenic bacteria that produce severe human diseases such as typhus and spotted fever (14). Upon infection of eukaryotic cells, they undergo actin comet-tailbased motility by using two different proteins expressed on the surface of the bacterium at different stages of infection. Early postinfection, RickA, a protein that mimics eukaryotic nucleation promoting factors, recruits the host Arp2/3 complex to activate the formation of actin comet tails that are short and curved (15–17). Later postinfection, Sca2, a protein that mimics eukaryotic formins, polymerizes actin directly, creating actin comet tails that are long and unbranched (11,12,17). Sca2 is an autotransporter protein. Like other autotransporters, Sca2 uses an N-terminal signal sequence (residues
Actin Assembly by Rickettsia Sca2
1–33 in R. conorii Sca2, the species studied here) to cross the inner bacterial membrane, after which it inserts a C-terminal translocator domain (residues 1516–1795) into the outer membrane to form a pore through which the rest of the protein (passenger domain, residues 34–1515) is translocated and exposed on the surface of the bacterium (18,19). The requirement to pass through the narrow translocation pore and fold on the outside of the bacterium in the absence of an energy source imposes substantial constraints on the folding of autotransporter proteins. As a consequence, most autotransporters display relatively simple folds, based on a repeating basic folding unit, which in most cases consists of b-strands that stack up to form a b-solenoid (20,21). The structure of Sca2 residues 34–400, which we previously determined (12), suggests that Sca2 constitutes an exception among autotransporters in that the basic folding unit consists of helix-loop-helix motifs, which stack upon one another to give rise to an all-helical structure. We refer to this fragment of Sca2 as the N-terminal repeat domain (NRD) because a structural superimposition of the various helix-loop-helix motifs revealed conserved amino acids at specific locations. There is no high-resolution structural information for the remaining 1115 amino acids that form the passenger domain of Sca2. However, helix-loop-helix repeats of the kind observed in the structure of the NRD are clearly discernable at the sequence level between residues 1181 and 1341, which we thus refer to as the C-terminal repeat domain. Moreover, the entire passenger domain is predicted to be primarily helical, as experimentally confirmed for a fragment comprising residues 868–1515 using circular dichroism (12). Based on these considerations, we have proposed that most of the passenger domain of Sca2 adopts the same basic fold observed in the structure of the NRD (12). This poses a problem because toward the middle of the passenger domain (residues 868–1023), Sca2 presents what is thought to be a repeat of three WiskottAldrich syndrome homology 2 (WH2) domains. Consistent with this idea, simultaneously mutating all three of the predicted WH2 domains within full-length Sca2 impairs the nucleation activity (12). However, repeats of WH2 domains do not form part of globular structures in any of the proteins analyzed thus far; instead, WH2 domains are typically connected by flexible linkers of 15 amino acids (22). In Sca2, the separation between the putative WH2 domains is much greater (45–55 amino acids), and circular dichroism (12) and sequence analyzes suggest that the intervening regions between WH2 domains are folded. Here, we show that contrary to previous assumptions, Sca2 does not have WH2 domains and that the 868–1023 region cooperates with other parts of the Sca2 protein for actin binding and nucleation. Sca2 is also unique among bacterial nucleators in that it functions by a formin-like mechanism (11,12). Like eukaryotic formins, Sca2 generates long, unbranched filaments by
promoting both actin nucleation and elongation, i.e., it remains bound to filament barbed ends after nucleation and drives the processive incorporation of profilin-actin while also competing with capping protein (11,12). More specifically, in the absence of profilin, Sca2 acts as a barbed end cap, inhibiting filament elongation, whereas in the presence of profilin, it increases filament elongation, albeit only to 1/3 of that of actin alone (12). Several formins, including mDia2, Cdc12, and Bni1, exhibit a similar behavior (23). Yet, despite these functional similarities, Sca2 is structurally unrelated to formins and performs these functions through an entirely different mechanism (12). Also unlike formins, which form dimers, Sca2 is monomeric. However, it displays intramolecular interactions and cooperativity between its N- and C- terminal regions that are crucial for actin nucleation and elongation (12). Here, we map the interaction of N- and C- terminal fragments of Sca2 and their contributions to actin binding and nucleation and conclude that the two fragments interact in a head-to-tail circular manner to recruit two actin subunits for nucleation, somewhat analogous to the formin homology-2 (FH2) domain. MATERIALS AND METHODS Protein expression and purification All the constructs used in this study were amplified from the R. conorii sca2 gene (GenBank: AAL02648). All of the constructs, except Sca1090–1355, were expressed and purified as described previously (12), with slight modifications. Briefly, constructs Sca34–1515, Sca34–670, Sca34–400, Sca421–670, Sca869–1060, Sca1090–1515, and Sca868–1515 were cloned into the pTYB11 vector (New England BioLabs, Ipswich, MA) containing a chitin affinity tag and an intein domain. Sca1090–1355 was cloned into the pRSFduet vector with an N-terminal His-tag and an engineered N-terminal TEV protease site. All the proteins were expressed in Escherichia coli Rosetta or Arctic Express cells grown in Terrific Broth media. Cells were resuspended in chitin buffer (20 mM HEPES (pH 7.5), 500 mM NaCl, 1 mM EDTA, 4 mM benzamidine) or, in the case of construct Sca1090–1355, nickel lysis buffer (20 mM Tris-HCl (pH 7.5), 200 mM NaCl, 20 mM imidazole, 4 mM benzamidine) and lysed using a microfluidizer (MicroFluidics Corporation, Newton, MA). After purification on a chitin affinity column, the chitin tag was removed by self-cleavage of the intein, induced by the addition of 50 mM dithiothreitol (DTT) for 1–2 days at 20 C, which leaves no extra residues on the Sca2 constructs. For Sca1090–1355, the N-terminal Histag was cleaved after elution from the nickel affinity column by the addition of TEV protease at 4 C overnight, followed by a second round of nickel affinity to remove the TEV protease and any uncut protein. All the proteins were additionally purified by either ion exchange or size exclusion chromatography. Point mutations in the WH2 domains of Sca34–1515 were introduced using the QuikChange II XL site-directed mutagenesis kit (Stratagene, San Diego, CA) as before (12) and expressed and purified as described above. The putative WH2 peptides of Sca2 (Sca871–890, Sca942–962, Sca1003–1022) and that of Wiskott-Aldrich syndrome protein family member 1 (WAVE1) were cloned as maltose-binding protein (MBP) fusions into a modified pMAL-c2x vector containing an N-terminal His-tag and an engineered N-terminal TEV protease site by primer extension to express proteins with the following overall organization: His-TEV-MBPlinker(AAANNNNNNNNNNLG)-WH2. These fusion proteins were expressed in E. coli Arctic cells and purified by nickel affinity followed by ion exchange chromatography.
Biophysical Journal 116, 540–550, February 5, 2019 541
Alqassim et al.
Analytical gel filtration chromatography To investigate the formation of complexes between the various N- and C-terminal domains of Sca2, proteins were mixed, incubated for 30 min, and run through a Superose 6 10/300 GL column (GE Healthcare, Waukesha, WI) in a buffer containing 20 mM Tris-HCl (pH 7.5), 200 mM NaCl, and 1 mM DTT. Peak fractions were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).
Actin polymerization assay Actin polymerization was measured as the fluorescence increase resulting from the incorporation of pyrene-labeled actin into filaments, using a Cary Eclipse Fluorescence Spectrophotometer (Varian, Agilent, Santa Clara, CA). Before data acquisition, 2 mM Mg-ATP-actin (6% pyrene labeled) was mixed with different concentrations of Sca2 constructs in F-buffer (10 mM Tris-HCl (pH 7.5), 1 mM MgCl2, 50 mM KCl, 1 mM EGTA, 0.1 mM NaN3, 0.2 mM ATP). Data acquisition started 5–10 s after mixing. All the measurements were done at 25 C. Control experiments were carried out with the addition of buffer alone.
Isothermal titration calorimetry ITC measurements were carried out on a VP-isothermal titration calorimetry instrument (MicroCal, Northhampton, MA). Protein samples were dialyzed for 2 days against ITC buffer (20 mM HEPES (pH 7.5), 200 mM NaCl, 1 mM DTT, and 35 mM LatB when monomeric actin is present). Titrations consisted of 10 mL injections, lasting for 25 s, with an interval of 300 s between injections. The heat of binding was corrected for the heat of injection, determined by injecting proteins into buffer. Data were analyzed using the program Origin (OriginLab, Northampton, MA). The temperature and parameters of the fit (stoichiometry and affinity) of each experiment are given in the figures.
RESULTS Intramolecular interaction of the N- and C-terminal regions of Sca2 Previously, we had shown that the N- and C-terminal regions of Sca2, fragments 34–670 (Sca34–670; see Fig. 1 A for construct definition) and 670–1515 (Sca670–1515), interact with each other—i.e., they copurify when coexpressed—and together, they replicate the activity of the full-length passenger domain (Sca34–1515) (12). Furthermore, deleting 200 amino acids near the middle of the protein (residues 671–867) did not significantly affect this interaction or the nucleation activity; coexpressed constructs His-Sca34–670 and GST-Sca868–1515 elute together during purification, and their complex can reproduce most of the nucleation activity of Sca34–1515 (12). To better understand this intramolecular interaction and the overall organization of Sca2 domains for nucleation, we tested the effect of N- and C-terminal truncations of constructs Sca34–670 and Sca868–1515 on the interaction. The two interacting constructs Sca34–670 and Sca868–1515 were used as a control in this analysis. Because previously we had tested their interaction only through copurification using two different affinity tags (12), here, we expressed the Sca34–670 and Sca868–1515 fragments independently and free of tags (see Materials
542 Biophysical Journal 116, 540–550, February 5, 2019
and Methods) and characterized their interaction quantitatively using isothermal titration calorimetry (ITC). Fitting of the ITC titration of Sca34–670 (in the syringe) into Sca868–1515 (in the ITC cell) showed that the two fragments interact with relatively high affinity (KD ¼ 7.0 5 0.3 mM) and the reaction was exothermic in character (Fig. 1 B). Furthermore, when mixed together, the two independently purified Sca2 fragments run as a single peak containing approximately equal amounts of both proteins by analytical gel filtration and SDS-PAGE analysis (Fig. 1 C), consistent with the formation of a relatively tight complex. To identify the minimal regions necessary for this intramolecular interaction, we tested binding of the intact N-terminal fragment (Sca34–670) with independently expressed N- and C-terminal truncations of the C-terminal fragment (Sca868–1515) and vice versa (Fig. 1, D–H). Our results show that truncations at either end of the N-terminal (Sca34–400 and Sca421–670) or C-terminal (Sca869–1060, Sca1090–1355, and Sca1090–1515) regions of Sca2 abrogate the interaction because the premixed proteins emerge separately by analytical gel filtration and SDS-PAGE analysis of the peak fractions. For one construct combination, Sca34–670 and Sca1090–1355, the gel filtration peak of the individually purified proteins partially overlaps with that of their mixture, but SDS-PAGE analysis shows that the two constructs are largely segregated toward the ends of the gel filtration peak, consistent with lack of interaction (Fig. 1 F). Therefore, Sca34–670 and Sca868–1515 are the smallest fragments tested here that can still account for the intramolecular interaction of Sca2. Moreover, analogous to the FH2 domain, these two fragments seem to interact in a circular, head-totail manner, since deletions at either end of both constructs interfere with their ability to form a complex. Sca2 binds two actin subunits with high affinity The FH2 domain of eukaryotic formins also interacts in a circular, head-to-tail manner to form a dimeric ring-like structure, which encircles in the middle two actin subunits at the barbed end of the actin filament (24,25). Each subunit of the FH2 dimer contributes to the overall binding affinity, which is typically higher for the dimer than the hemidimer (i.e., a single FH2), although reported values diverge widely for different formins (26–31). We hypothesized that the N- and C-terminal fragments of Sca2 may interact with actin in a similar manner. To test this idea, we used ITC to assess the interaction of the full-length passenger domain of Sca2 (construct Sca34–1515) with monomeric actin. Latrunculin B (LatB), a marine toxin that binds in the catalytic cleft near the nucleotide in actin, was used in these experiments to prevent actin polymerization. The titration of LatB-actin into Sca34–1515 produced an exothermic reaction (Fig. 2 A), whose binding isotherm fitted best to a two-site binding model (i.e., two actin subunits per Sca34–1515 molecule), and both sites had high affinity (KD1 ¼ 0.19 mM and
Actin Assembly by Rickettsia Sca2
A
B
C
D
E
F
G
H
FIGURE 1 Intramolecular interaction of the N- and C-terminal regions of Sca2. (A) Domain organization of Sca2 and constructs used in this study (SS, signaling sequence; NRD, N-terminal repeat domain; PRD1,2, proline-rich domains; WH2a–c, putative WH2 domains; CRD, C-terminal repeat domain; AC, predicted autochaperone domain; TD, translocator domain). (B) ITC titration of 200 mM Sca34–670 (in the syringe) into 20 mM Sca868–1515 (in the cell). The experiment was performed at 25 C. The dissociation constant (KD) and binding stoichiometry (N) derived from fitting of the binding isotherm are listed. Errors correspond to the SD of the fits. Open symbols correspond to titrations into buffer. (C–H) Analytical size exclusion chromatography (SEC) and SDSPAGE analyzes of mixtures of untagged N- and C-terminal Sca2 constructs as indicated. The gels shown as insets correspond to the SEC fractions of the peak, (legend continued on next page)
Biophysical Journal 116, 540–550, February 5, 2019 543
Alqassim et al.
A
C
B
D
KD2 ¼ 0.02 mM) (Fig. 2 A, blue curve). Fitting to a singlesite model, shown for comparison (Fig. 2 A, red curve), produced a much worse c2 value (109.5 compared to 8.5) and unrealistic stoichiometry (N ¼ 0.6). Note that the c2 values reported here are reduced, i.e., divided by degrees of freedom, and although these values provide a measure of the goodness of the fit, they cannot be compared across different experiments because the data are not weighted and the c2 depends on the magnitude of the scale.
FIGURE 2 Interaction of Sca2 constructs with monomeric actin. (A) ITC titration of 89 mM LatBactin (in the syringe) into 8.3 mM Sca34–1515 (in the cell). The experiment was performed at 20 C. A two-site binding model produced a better fit of the data to a binding isotherm than a one-site binding model, as indicated by the c2 values and realistic stoichiometry. The c2 values reported here are reduced, i.e., divided by degrees of freedom. Note that although the c2 value clearly defines the best fitting model for a particular titration, they are not generally comparable across different titrations because the data are not weighted and the c2 depends on the magnitude of the scale. The dissociation constant (KD) and binding stoichiometry (N) derived from the fits are listed. Errors correspond to the SD of the fits. (B) ITC titration of 220 mM Sca34–670 (in the syringe) into 22 mM LatB-actin (in the cell). The experiment was performed at 12.7 C. A two-site binding model produced an almost equally good fit of the data to a binding isotherm than a one-site binding model, but is prefered based on a more realistic stoichiometry of the interaction. A two-site binding model is also supported by the fact that this fragment of Sca2 has nucleation activity (12). The dissociation constant (KD) and binding stoichiometry (N) derived from the fits are listed. Errors correspond to the SD of the fits. (C) ITC titration of 444 mM Sca869–1060 (in the syringe) into 16.6 mM LatB-actin (in the cell). The experiment was performed at 20 C. The dissociation constant (KD) and binding stoichiometry (N) derived from fitting of the binding isotherm are listed. Errors correspond to the SD of the fits. (D) ITC titration of 396 mM Sca1090–1355 (in the syringe) into 19 mM LatB-actin (in the cell). The experiment was performed at 20 C. No binding is reported because the experiment could not be fitted to a binding isotherm. The different fits are colorcoded as indicated. To see this figure in color, go online.
We then asked whether the N- and C-terminal fragments of Sca2 independently contributed to this interaction. Curiously, the titration of the N-terminal half of Sca2 (Sca34–670) into LatB-actin produced an endothermic reaction. The titration was first fitted to a one-site binding model, but the stoichiometry of the interaction was unrealistically low (N ¼ 0.33), suggesting that Sca34–670 interacts with two actin subunits (Fig. 2 B). Fitting to a two-site binding model yielded two binding sites of lower affinity than those of the
indicated by a dashed box. (C) shows that the premixed constructs Sca34–670 and Sca868–1515 run together by SEC, consistent with the formation of a complex. (D)–(F) show that the N-terminal fragment Sca34–670 runs separately from N- and C-terminal deletions of construct Sca868–1515 when premixed, consistent with lack of interaction. (G) and (H) show that the C-terminal fragment Sca868–1515 runs separately from N- and C-terminal deletions of constructs Sca34–670 when premixed, consistent with lack of interaction. The SEC traces of the Sca2 constructs are color-coded as indicated. To see this figure in color, go online.
544 Biophysical Journal 116, 540–550, February 5, 2019
Actin Assembly by Rickettsia Sca2
full-length passenger domain (KD1 ¼ 33.1 mM and KD2 ¼ 8.8 mM), although the c2 value (1.2) did not improve. A two-site binding model is preferred in this case because it addresses the low stoichiometry obtained with the one-site binding model and explains why this fragment of Sca2 displays nucleation activity, albeit lower than that of the fulllength passenger domain (12). We could not test binding of the entire C-terminal region (Sca868–1515) to LatB-actin because of the very low yields of this construct when expressed in isolation. Instead, we tested binding of two fragments that make up this construct (Sca869–1060 and Sca1090–1355). Sca869–1060, which contains a predicted repeat of three WH2 domains (see below), bound only one LatB-actin monomer with a KD of 10.8 mM (Fig. 2 C), and the reaction was endothermic in character, like for Sca34–670. In contrast, Sca1090–1355 did not appear to bind LatB-actin (Fig. 2 D). Therefore, these results suggest that the two halves of Sca2 contribute to actin binding, analogous to the individual subunits of the FH2 dimer (26,27), but only when combined together can they achieve high-affinity binding of two actin monomers. The fact that the interactions change from endothermic for the individual fragments to exothermic for the full-length passenger domain and the affinities increase manyfold suggests that additional interactions, presumably including actinactin interactions, are involved in the formation of this ternary complex. Sca2 does not contain WH2 domains As established above, Sca2 binds only two actin monomers, whereas the region containing the putative repeat of three WH2 domains (Sca869–1060) only binds one actin monomer and with much weaker affinity than the full-length passenger domain. All the WH2 domains analyzed by ITC so far bind actin with higher affinity than Sca869–1060, and their binding is consistently exothermic in character (32–34). These observations raised questions about the authenticity of the WH2 domains of Sca2. This is important because it is still assumed that Sca2 contains WH2 domains (22), and it was the identification of putative WH2 domains based on sequence analysis that supported the initial hypothesis that Sca2 was a bacterial actin assembly factor, although curiously, two different laboratories had proposed different definitions of the WH2 domains (11,13). We use here the latest definition of the WH2 domains of Sca2 (11) (Fig. 3 A), which produces a better alignment with other known WH2 domains (Fig. 3 B). Yet, the initial disagreement is understandable when one compares the sequences of the WH2 domains of Sca2 with other canonical WH2 domains. The WH2 domain is 17 amino acids in length and consists of an N-terminal a-helix that binds in the hydrophobic or target-binding cleft of actin, a loop, and a C-terminal LKKV motif (22). All three of the WH2 domains of Sca2 lack at least one critical element of the consensus
sequence. Thus, WH2a (residues 772–888) has a proline residue within the LKKV motif (22). WH2b (residues 943–960) lacks a basic residue at the N-terminus and key hydrophobic residues in the N-terminal a-helix of the domain. WH2c (residues 1004–1020) lacks the two basic residues of the LKKV motif. Importantly, mutations in the LKKV motif of classical WH2 domains disrupt binding to monomeric actin (35,36). We then analyzed the sequences connecting the WH2 domains of Sca2. Here again, Sca2 is a clear outlier, with the putative WH2 domains spaced farther apart (45–55 amino acids) than in other proteins containing repeats of WH2 domains (15 amino acids) (Fig. 3 A). A flexible 15-amino-acid linker seems to be optimal for proteins in which successive WH2 domains connect actin subunits along the long-pitch helix of the actin filament (37). The one exception is Cobl, in which a 69-amino-acid linker connects the second and third WH2 domains. However, Sca2 differs in yet another way from canonical WH2 domaincontaining proteins; the sequences between WH2 domains are predicted to contain a series of a-helices, whereas in other proteins, these sequences are typically unstructured (see the secondary structure predictions in Fig. 3 A). Hydrophobic cluster analysis further supports this conclusion, revealing a series of clusters of hydrophobic residues in the inter-WH2 domain linkers of Sca2 but not other proteins (Fig. S1). Such clusters typically denote the presence of secondary structural elements, which, depending on the shape of the clusters, can be defined as either a-helices or b-strands (38). The hydrophobic cluster analysis plots of the WH2 domains themselves are also different between Sca2 and other WH2 domain-containing proteins (Fig. S1). The fact that the putative WH2 domains of Sca2 form part of a predicted globular domain is finally supported by far-ultraviolet circular dichroism analysis of an Sca2 fragment comprising amino acids 868–1515, revealing a typical a-helical profile with minima at 208 and 222 nm (12). Therefore, several lines of evidence suggest that Sca2 does not contain a typical WH2 domain repeat. However, we had previously found that a full-length passenger domain construct carrying four mutations in each of the WH2 domains (a total of 12 mutations targeting both the a-helix and LKKV motif of each WH2 domain) had no detectable nucleation or elongation activity (12), demonstrating the importance of this region for actin assembly. But, because this region binds only one actin monomer (Fig. 2 C), we asked whether only one of the putative WH2 domains was still functional despite their noncanonical sequences. For this, we generated three Sca34–1515 mutants individually targeting each of the WH2 domains. We used the same mutations as before (12), i.e., two conserved hydrophobic residues within the N-terminal a-helix and the first two residues of the LKKV motif were simultaneously mutated to aspartic acid (Fig. 3 C). Using the pyreneactin polymerization assay, we found that the mutants
Biophysical Journal 116, 540–550, February 5, 2019 545
Alqassim et al.
A
C
B
D
FIGURE 3 The region 868–1023 of Sca2 is critical for nucleation but lacks WH2 domains. (A) A cartoon representation of the WH2-domain-containing regions of various cytoskeletal proteins containing tandem repeats of WH2 domains. WH2 domains are colored red, with a cylinder representing the a-helix of the domain. Secondary structure predictions obtained with the program Jpred (45) are also shown for each protein (two-dimensional prediction), with blue cylinders representing predicted a-helices. (B) Alignment of the sequences of the WH2 domains represented in (A), showing the regions corresponding to the N-terminal a-helix and LKKV motif, as well as the consensus sequence at the bottom. (C) Representation of the three mutants of Sca34–1515 (full-length passenger domain) targeting the putative WH2 domains. Each mutant carries four aspartic acid mutations of highly conserved residues of the canonical sequence of the domain. The SDS-PAGE shown on the right illustrates the purity of one of the mutants (Sca34–1515-WH2a*). (D) A time course of the polymerization of 2 mM Mg-ATP-actin (6% pyrene labeled) alone and with the addition of 50 nM Sca34–1515 wild type and WH2 domain mutants depicted in (C) (color-coded as indicated). To see this figure in color, go online.
targeting WH2 domains a and b (Sca34–1515-WH2a* and Sca34–1515-WH2b*) inhibited the polymerization activity in a concentration-dependent manner, whereas mutant Sca34–1515-WH2c* had a negligible effect compared to the wild-type passenger domain (Fig. 3 D; Fig. S2). Therefore, at least two of the predicted WH2 domains of Sca2 are important for actin assembly within the full-length protein—is this because they bind actin directly the way WH2 domains do? To address this question, we generated four MBP fusion constructs with the WH2 domains of Sca2 and a control WH2 domain from WAVE1 fused C-terminally to MBP and separated from MBP by a noncleavable 15-amino-acid linker (Fig. 4 A). The proteins were highly pure, and the yields were sufficiently high to allow for quantitative ITC experiments to assess their binding to LatB-actin (Fig. 4 B). As anticipated, the first WH2 of WAVE1 (WAVE1 WH2a) bound LatB-actin with high affinity and 1:1 stoichiometry (KD ¼ 0.38 mM, N ¼ 1.07), and
546 Biophysical Journal 116, 540–550, February 5, 2019
the reaction was exothermic (Fig. 4 C). This interaction is therefore very similar to that of synthetic WH2 peptides from other cytoskeletal proteins we have measured before in G-buffer (i.e., without the addition of LatB) (33) or in F-buffer with the addition of LatB (32,34). In contrast, none of the putative WH2 domains of Sca2 bound LatBactin (Fig. 4, D–F), conclusively demonstrating that these are not functional WH2 domains. DISCUSSION Rickettsia Sca2 is a pathogenic bacterial actin assembly factor that mimics eukaryotic formins in that it generates long and unbranched actin filaments by combining both nucleation and processive barbed-end elongation activities (11). However, we have shown before that Sca2 is structurally unrelated to formins and uses a different mechanism for actin assembly (12). Specifically, Sca2 is monomeric and lacks
Actin Assembly by Rickettsia Sca2
A
B
D
C
E
F
FIGURE 4 The putative WH2 domains of Sca2 do not bind actin. (A) A representation of four WH2 domain constructs expressed here as fusion proteins with maltose-binding protein (MBP). A 15-amino-acid flexible linker separates the MBP moiety at the N-terminus from the WH2 domain of WAVE1 (control) or the three putative WH2 domains of Sca2. (B) Purity of proteins illustrated in (A) assessed by SDS-PAGE analysis. (C) ITC titration of 254 mM WAVE1 WH2a into 18.5 mM LatB-actin. The experiment was performed at 25 C. The dissociation constant (KD) and binding stoichiometry (N) derived from fitting of the binding isotherm are listed. Errors correspond to the SD of the fits. (D–F) ITC titrations of the three constructs of putative WH2 domains of Sca2 into LatB-actin. The concentrations of the proteins in the syringe and in the cell are indicated. The titrations could not be fitted to binding isotherms because none of the putative WH2 domains of Sca2 appeared to bind LatB-actin. To see this figure in color, go online.
the dimeric FH2 domain found in all eukaryotic formins, which is crucial for both nucleation and elongation (39). The one element of the Sca2 sequence that still appeared to resemble eukaryotic nucleators was the postulated repeat of three WH2 domains (22), whose identification first drew attention to this protein as a potential actin assembly factor (11,13). Of note, several eukaryotic formins contain WH2domain-related sequences or form complexes with proteins
that contain WH2 domains and synergize with formins for nucleation (40–44). Here, however, we have disproven the existence of WH2 domains in Sca2, suggesting that Rickettsia Sca2 has evolved completely independently of any eukaryotic actin assembly factor to acquire the various complex functions associated with eukaryotic formins: nucleation, elongation, and inhibition of barbed-end capping (11,12).
Biophysical Journal 116, 540–550, February 5, 2019 547
Alqassim et al.
How does Sca2 accomplish these functions using a different fold? In thinking about this question, a comparison with eukaryotic formins provides some clues. Formins are dimeric, and specifically the two FH2 domains of the dimer interact head to tail to form a ring-like structure (25), which encircles two actin subunits at the barbed end of the actin filament while loosely contacting a third incoming subunit during processive elongation (24). Sca2 cannot be a dimer because bacterial autotransporters are either monomeric or trimeric, depending on the fold of the translocator domain, which can be 300 (monomeric) or 75 (trimeric) amino acids long (18). The passenger domain of Sca2 is monomeric in solution (12) and, consistently, Sca2 has a 281amino-acid monomeric type of translocator domain. However, like formins, Sca2 must ‘‘hug’’ the barbed end of the filament while allowing access to additional actin subunits for processive elongation. This can only be accomplished if Sca2 also forms a ring-like structure around the barbed end, contacting primarily two actin subunits and possibly a third incoming subunit (Fig. 5). The data presented here are consistent with this model. The N- and C-terminal regions of Sca2 (Sca34–670 and Sca868–1515) seem to behave to a large extent like the FH2 domains of the formin dimer: 1) they bind actin individually but with weak affinity (Sca34–670 interacts with two actin subunits,
whereas Sca868–1515 appears to contact only one subunit); 2) together (i.e., the full-length passenger domain), they bind two actin subunits with high affinity and account for the nucleation and barbed-end elongation activities of Sca2, which the individual fragments mostly lack (12); 3) deletions at either end of these two fragments disrupt their ability to interact with each other, suggesting that they bind in head-to-tail fashion to form a ring-like structure analogous to the formin FH2 domain. One of these fragments, Sca869–1060, comprises what was previously thought to be a repeat of three WH2 domains. However, Sca869–1060 binds a single actin subunit with weak affinity and following an endothermic reaction, clearly distinct from the higher affinity and exothermic reactions of canonical WH2 domains (32–34). Moreover, we have shown here that none of the putative WH2 domains of Sca2 binds actin in isolation, suggesting that the adverse effect of mutations of two of the domains is due to either disruption of a contiguous actinbinding surface or global changes in the structure of this region, which we now call the ‘‘middle domain’’ because of its location within the full-length passenger domain (Fig. 5). Analogous to the proline-rich FH1 domain of formins, Sca2 also contains two proline-rich segments (672–699 and 1077–1087) that mediate the binding of profilin-actin for processive barbed end elongation (11,12). Both FIGURE 5 Proposed mechanism of nucleation and elongation by Rickettsia Sca2. By analogy with eukaryotic formins, Sca2 is proposed to form a ring-like structure around the barbed end of the actin filament, contacting primarily two actin subunits and possibly a third incoming subunit. The data presented here support this model and suggest that the N- and C-terminal regions of Sca2 (Sca34–670 and Sca868–1515) behave in a way analogous to the FH2 domain of formins. Thus, we have shown here that these two fragments bind actin individually but with weak affinity, whereas the full-length passenger domain binds two actin subunits with high affinity and has both nucleation and elongation activities, although the individual fragments are mostly inactive (11,12). Furthermore, deletions at either end of these two fragments disrupt their ability to interact with each other, suggesting that they interact head to tail to form a ring-like structure analogous to the formin FH2 domain. We have also shown that Sca2 does not contain WH2 domains. The region previously thought to contain the repeat of three WH2 domains (Sca869–1060) binds actin as a whole but only one actin subunit and in manner clearly distinct from that of canonical WH2 domains. Accordingly, this region is represented here as forming part of the ring-like structure of Sca2 and is renamed as the ‘‘middle domain’’ because of its location in the sequence. The two proline-rich domains, implicated in the recruitment of profilin-actin for barbed end elongation (12), may project out because they form part of loops that are predicted to be unstructured. To see this figure in color, go online.
548 Biophysical Journal 116, 540–550, February 5, 2019
Actin Assembly by Rickettsia Sca2 10. Benanti, E. L., C. M. Nguyen, and M. D. Welch. 2015. Virulent Burkholderia species mimic host actin polymerases to drive actin-based motility. Cell. 161:348–360.
segments form part of loop regions (658–714 and 1066– 1091), which are predicted to be unstructured and could therefore be loosely attached to the other domains of the Sca2 molecule that form the ring-like structure around the barbed end. None of the Sca2 fragments studied here could be crystallized (except Sca34–400, which we studied previously (12)), either alone or in complex with actin, but structural studies testing some of the predictions of this model should soon become possible using cryo-electron microscopy.
13. Kleba, B., T. R. Clark, ., T. Hackstadt. 2010. Disruption of the Rickettsia rickettsii Sca2 autotransporter inhibits actin-based motility. Infect. Immun. 78:2240–2247.
SUPPORTING MATERIAL
14. Schroeder, C., I. Chowdhury, ., S. Sahni. 2016. Human rickettsioses: host response and molecular pathogenesis. In Rickettsiales. S. Thomas, ed. Springer, pp. 399–446.
Two figures are available at http://www.biophysj.org/biophysj/supplemental/ S0006-3495(18)34504-1.
AUTHOR CONTRIBUTIONS S.S.A. and R.D. designed the experiments. S.S.A. and I.-G.L. prepared the proteins, including cloning, expression and purification, and S.S.A. performed all of the experiments for this work. S.S.A. and R.D. wrote the manuscript. All authors analyzed the data and reviewed the figures and manuscript.
ACKNOWLEDGMENTS We thank Grzegorz Rebowski for the preparation of actin for this study and help with Fig. 5, and Malgorzata Boczkowska for guidance with the actin polymerization assay, ITC experiments, and data analysis. This work was supported by National Institutes of Health grant R01 GM073791.
11. Haglund, C. M., J. E. Choe, ., M. D. Welch. 2010. Rickettsia Sca2 is a bacterial formin-like mediator of actin-based motility. Nat. Cell Biol. 12:1057–1063. 12. Madasu, Y., C. Suarez, ., R. Dominguez. 2013. Rickettsia Sca2 has evolved formin-like activity through a different molecular mechanism. Proc. Natl. Acad. Sci. USA. 110:E2677–E2686.
15. Gouin, E., C. Egile, ., P. Cossart. 2004. The RickA protein of Rickettsia conorii activates the Arp2/3 complex. Nature. 427:457–461. 16. Jeng, R. L., E. D. Goley, ., M. D. Welch. 2004. A Rickettsia WASPlike protein activates the Arp2/3 complex and mediates actin-based motility. Cell. Microbiol. 6:761–769. 17. Reed, S. C. O., R. L. Lamason, ., M. D. Welch. 2014. Rickettsia actinbased motility occurs in distinct phases mediated by different actin nucleators. Curr. Biol. 24:98–103. 18. Dautin, N., and H. D. Bernstein. 2007. Protein secretion in gram-negative bacteria via the autotransporter pathway. Annu. Rev. Microbiol. 61:89–112. 19. Bernstein, H. D. 2015. Looks can be deceiving: recent insights into the mechanism of protein secretion by the autotransporter pathway. Mol. Microbiol. 97:205–215. 20. Junker, M., C. C. Schuster, ., P. L. Clark. 2006. Pertactin b-helix folding mechanism suggests common themes for the secretion and folding of autotransporter proteins. Proc. Natl. Acad. Sci. USA. 103:4918–4923. 21. Kajava, A. V., and A. C. Steven. 2006. The turn of the screw: variations of the abundant b-solenoid motif in passenger domains of Type V secretory proteins. J. Struct. Biol. 155:306–315. 22. Dominguez, R. 2016. The WH2 domain and actin nucleation: necessary but insufficient. Trends Biochem. Sci. 41:478–490.
REFERENCES 1. Bugalha˜o, J. N., L. J. Mota, and I. S. Franco. 2015. Bacterial nucleators: actin’ on actin. Pathog. Dis. 73:ftv078. 2. Jimenez, A., D. Chen, and N. M. Alto. 2016. How bacteria subvert animal cell structure and function. Annu. Rev. Cell Dev. Biol. 32:373–397. 3. Dominguez, R. 2015. Subversive bacteria reveal new tricks in their cytoskeleton-hijacking arsenal. Nat. Struct. Mol. Biol. 22:178–179. 4. Lamason, R. L., and M. D. Welch. 2017. Actin-based motility and cellto-cell spread of bacterial pathogens. Curr. Opin. Microbiol. 35:48–57. 5. Truong, D., J. W. Copeland, and J. H. Brumell. 2014. Bacterial subversion of host cytoskeletal machinery: hijacking formins and the Arp2/3 complex. BioEssays. 36:687–696. 6. Liverman, A. D., H. C. Cheng, ., K. Orth. 2007. Arp2/3-independent assembly of actin by Vibrio type III effector VopL. Proc. Natl. Acad. Sci. USA. 104:17117–17122. 7. Yu, B., H. C. Cheng, ., M. K. Rosen. 2011. Mechanism of actin filament nucleation by the bacterial effector VopL. Nat. Struct. Mol. Biol. 18:1068–1074.
23. Kovar, D. R., E. S. Harris, ., T. D. Pollard. 2006. Control of the assembly of ATP- and ADP-actin by formins and profilin. Cell. 124:423–435. 24. Otomo, T., D. R. Tomchick, ., M. K. Rosen. 2005. Structural basis of actin filament nucleation and processive capping by a formin homology 2 domain. Nature. 433:488–494. 25. Xu, Y., J. B. Moseley, ., M. J. Eck. 2004. Crystal structures of a Formin Homology-2 domain reveal a tethered dimer architecture. Cell. 116:711–723. 26. Shimada, A., M. Nyitrai, ., A. Wittinghofer. 2004. The core FH2 domain of diaphanous-related formins is an elongated actin binding protein that inhibits polymerization. Mol. Cell. 13:511–522. 27. Zigmond, S. H. 2004. Formin-induced nucleation of actin filaments. Curr. Opin. Cell Biol. 16:99–105. 28. Li, F., and H. N. Higgs. 2003. The mouse Formin mDia1 is a potent actin nucleation factor regulated by autoinhibition. Curr. Biol. 13:1335–1340. 29. Moseley, J. B., I. Sagot, ., B. L. Goode. 2004. A conserved mechanism for Bni1- and mDia1-induced actin assembly and dual regulation of Bni1 by Bud6 and profilin. Mol. Biol. Cell. 15:896–907. 30. Pring, M., M. Evangelista, ., S. H. Zigmond. 2003. Mechanism of formin-induced nucleation of actin filaments. Biochemistry. 42:486–496.
8. Namgoong, S., M. Boczkowska, ., R. Dominguez. 2011. Mechanism of actin filament nucleation by Vibrio VopL and implications for tandem W domain nucleation. Nat. Struct. Mol. Biol. 18:1060–1067.
31. Kovar, D. R., J. R. Kuhn, ., T. D. Pollard. 2003. The fission yeast cytokinesis formin Cdc12p is a barbed end actin filament capping protein gated by profilin. J. Cell Biol. 161:875–887.
9. Burke, T. A., A. J. Harker, ., D. R. Kovar. 2017. The bacterial virulence factors VopL and VopF nucleate actin from the pointed end. J Cell Biol. 216:1267–1276.
32. Lee, S. H., F. Kerff, ., R. Dominguez. 2007. Structural basis for the actin-binding function of missing-in-metastasis. Structure. 15: 145–155.
Biophysical Journal 116, 540–550, February 5, 2019 549
Alqassim et al. 33. Chereau, D., F. Kerff, ., R. Dominguez. 2005. Actin-bound structures of Wiskott-Aldrich syndrome protein (WASP)-homology domain 2 and the implications for filament assembly. Proc. Natl. Acad. Sci. USA. 102:16644–16649. 34. Boczkowska, M., Z. Yurtsever, ., R. Dominguez. 2017. Crystal structure of Leiomodin 2 in complex with actin: a structural and functional reexamination. Biophys. J. 113:889–899. 35. Mattila, P. K., M. Salminen, ., P. Lappalainen. 2003. Mouse MIM, a tissue-specific regulator of cytoskeletal dynamics, interacts with ATPactin monomers through its C-terminal WH2 domain. J. Biol. Chem. 278:8452–8459. 36. Quinlan, M. E., J. E. Heuser, ., R. D. Mullins. 2005. Drosophila Spire is an actin nucleation factor. Nature. 433:382–388. 37. Rebowski, G., S. Namgoong, ., R. Dominguez. 2010. Structure of a longitudinal actin dimer assembled by tandem w domains: implications for actin filament nucleation. J. Mol. Biol. 403:11–23.
40. Chhabra, E. S., and H. N. Higgs. 2006. INF2 Is a WASP homology 2 motif-containing formin that severs actin filaments and accelerates both polymerization and depolymerization. J. Biol. Chem. 281: 26754–26767. 41. Heimsath, E. G., Jr., and H. N. Higgs. 2012. The C terminus of formin FMNL3 accelerates actin polymerization and contains a WH2 domainlike sequence that binds both monomers and filament barbed ends. J. Biol. Chem. 287:3087–3098. 42. Gould, C. J., S. Maiti, ., B. L. Goode. 2011. The formin DAD domain plays dual roles in autoinhibition and actin nucleation. Curr. Biol. 21:384–390. 43. Quinlan, M. E., S. Hilgert, ., E. Kerkhoff. 2007. Regulatory interactions between two actin nucleators, Spire and Cappuccino. J. Cell Biol. 179:117–128.
38. Bitard-Feildel, T., A. Lamiable, ., I. Callebaut. 2018. Order in disorder as observed by the ‘‘hydrophobic cluster analysis’’ of protein sequences. Proteomics. 18:e1800054.
44. Pechlivanis, M., A. Samol, and E. Kerkhoff. 2009. Identification of a short Spir interaction sequence at the C-terminal end of formin subgroup proteins. J. Biol. Chem. 284:25324–25333.
39. Breitsprecher, D., and B. L. Goode. 2013. Formins at a glance. J. Cell Sci. 126:1–7.
45. Drozdetskiy, A., C. Cole, ., G. J. Barton. 2015. JPred4: a protein secondary structure prediction server. Nucleic Acids Res. 43:W389–W394.
550 Biophysical Journal 116, 540–550, February 5, 2019
Rickettsia Sca2 Recruits Two Actin Subunits for Nucleation but Lacks WH2 Domains Saif S. Alqassim,1,2 In-Gyun Lee,1 and Roberto Dominguez1,* 1 Department of Physiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania and 2College of Medicine, Mohammed Bin Rashid University of Medicine and Health Sciences, Dubai, United Arab Emirates
ABSTRACT The Rickettsia 1800-amino-acid autotransporter protein surface cell antigen 2 (Sca2) promotes actin polymerization on the surface of the bacterium to drive its movement using an actin comet-tail mechanism. Sca2 mimics eukaryotic formins in that it promotes both actin filament nucleation and elongation and competes with capping protein to generate filaments that are long and unbranched. However, despite these functional similarities, Sca2 is structurally unrelated to eukaryotic formins and achieves these functions through an entirely different mechanism. Thus, while formins are dimeric, Sca2 functions as a monomer. However, Sca2 displays intramolecular interactions and functional cooperativity between its N- and C-terminal domains that are crucial for actin nucleation and elongation. Here, we map the interaction of N- and C- terminal fragments of Sca2 and their contribution to actin binding and nucleation. We find that both the N- and C-terminal regions of Sca2 interact with actin monomers but only weakly, whereas the full-length protein binds two actin monomers with high affinity. Moreover, deletions at both ends of the N- and C-terminal regions disrupt their ability to interact with each other, suggesting that they form a contiguous ring-like structure that wraps around two actin subunits, analogous to the formin homology-2 domain. The discovery of Sca2 as an actin nucleator followed the identification of what appeared to be a repeat of three Wiskott-Aldrich syndrome homology 2 (WH2) domains in the middle of the molecule, consistent with the presence of WH2 domains in most actin nucleators. However, we show here that contrary to previous assumptions, Sca2 does not contain WH2 domains. Instead, our analysis indicates that the region containing the putative WH2 domains is folded as a globular domain that cooperates with other parts of the Sca2 molecule for actin binding and nucleation.
INTRODUCTION Numerous pathogens, including several bacteria and some viruses, take advantage of the actin cytoskeleton of the host eukaryotic cell for motility and infection (1–3). A common mechanism relies on the activation of actin nucleation on the surface of pathogens to generate actin comet tails that propel their movement (4). Pathogens that use this mechanism typically nucleate actin by exposing proteins that mimic eukaryotic nucleation-promoting factors, which recruit and activate the host Arp2/3 complex (5). However, some bacteria encode their own actin nucleators, i.e., proteins capable of directly activating the polymerization of eukaryotic actin, including Vibrio VopL/F (6–9), Burkholderia BimA (10), and Rickettsia Sca2 (11–13). Such bacterial nucleators have evolved independently of their eukaryotic counterparts, and they tend to polymerize actin using
Submitted November 9, 2018, and accepted for publication December 12, 2018. *Correspondence: [email protected] Editor: Laurent Blanchoin. https://doi.org/10.1016/j.bpj.2018.12.009 Ó 2018 Biophysical Society.
540 Biophysical Journal 116, 540–550, February 5, 2019
different molecular mechanisms than the eukaryotic nucleators (4). This feature, along with the requirement of actin-based motility for infection, highlights the importance of studying bacterial nucleators for a better understanding of both pathogenicity and actin nucleation mechanisms. Here, we focus on surface cell antigen 2 (Sca2), a surface protein of spotted fever Rickettsiae. Rickettsiae are intracellular Gram-negative pathogenic bacteria that produce severe human diseases such as typhus and spotted fever (14). Upon infection of eukaryotic cells, they undergo actin comet-tailbased motility by using two different proteins expressed on the surface of the bacterium at different stages of infection. Early postinfection, RickA, a protein that mimics eukaryotic nucleation promoting factors, recruits the host Arp2/3 complex to activate the formation of actin comet tails that are short and curved (15–17). Later postinfection, Sca2, a protein that mimics eukaryotic formins, polymerizes actin directly, creating actin comet tails that are long and unbranched (11,12,17). Sca2 is an autotransporter protein. Like other autotransporters, Sca2 uses an N-terminal signal sequence (residues
Actin Assembly by Rickettsia Sca2
1–33 in R. conorii Sca2, the species studied here) to cross the inner bacterial membrane, after which it inserts a C-terminal translocator domain (residues 1516–1795) into the outer membrane to form a pore through which the rest of the protein (passenger domain, residues 34–1515) is translocated and exposed on the surface of the bacterium (18,19). The requirement to pass through the narrow translocation pore and fold on the outside of the bacterium in the absence of an energy source imposes substantial constraints on the folding of autotransporter proteins. As a consequence, most autotransporters display relatively simple folds, based on a repeating basic folding unit, which in most cases consists of b-strands that stack up to form a b-solenoid (20,21). The structure of Sca2 residues 34–400, which we previously determined (12), suggests that Sca2 constitutes an exception among autotransporters in that the basic folding unit consists of helix-loop-helix motifs, which stack upon one another to give rise to an all-helical structure. We refer to this fragment of Sca2 as the N-terminal repeat domain (NRD) because a structural superimposition of the various helix-loop-helix motifs revealed conserved amino acids at specific locations. There is no high-resolution structural information for the remaining 1115 amino acids that form the passenger domain of Sca2. However, helix-loop-helix repeats of the kind observed in the structure of the NRD are clearly discernable at the sequence level between residues 1181 and 1341, which we thus refer to as the C-terminal repeat domain. Moreover, the entire passenger domain is predicted to be primarily helical, as experimentally confirmed for a fragment comprising residues 868–1515 using circular dichroism (12). Based on these considerations, we have proposed that most of the passenger domain of Sca2 adopts the same basic fold observed in the structure of the NRD (12). This poses a problem because toward the middle of the passenger domain (residues 868–1023), Sca2 presents what is thought to be a repeat of three WiskottAldrich syndrome homology 2 (WH2) domains. Consistent with this idea, simultaneously mutating all three of the predicted WH2 domains within full-length Sca2 impairs the nucleation activity (12). However, repeats of WH2 domains do not form part of globular structures in any of the proteins analyzed thus far; instead, WH2 domains are typically connected by flexible linkers of 15 amino acids (22). In Sca2, the separation between the putative WH2 domains is much greater (45–55 amino acids), and circular dichroism (12) and sequence analyzes suggest that the intervening regions between WH2 domains are folded. Here, we show that contrary to previous assumptions, Sca2 does not have WH2 domains and that the 868–1023 region cooperates with other parts of the Sca2 protein for actin binding and nucleation. Sca2 is also unique among bacterial nucleators in that it functions by a formin-like mechanism (11,12). Like eukaryotic formins, Sca2 generates long, unbranched filaments by
promoting both actin nucleation and elongation, i.e., it remains bound to filament barbed ends after nucleation and drives the processive incorporation of profilin-actin while also competing with capping protein (11,12). More specifically, in the absence of profilin, Sca2 acts as a barbed end cap, inhibiting filament elongation, whereas in the presence of profilin, it increases filament elongation, albeit only to 1/3 of that of actin alone (12). Several formins, including mDia2, Cdc12, and Bni1, exhibit a similar behavior (23). Yet, despite these functional similarities, Sca2 is structurally unrelated to formins and performs these functions through an entirely different mechanism (12). Also unlike formins, which form dimers, Sca2 is monomeric. However, it displays intramolecular interactions and cooperativity between its N- and C- terminal regions that are crucial for actin nucleation and elongation (12). Here, we map the interaction of N- and C- terminal fragments of Sca2 and their contributions to actin binding and nucleation and conclude that the two fragments interact in a head-to-tail circular manner to recruit two actin subunits for nucleation, somewhat analogous to the formin homology-2 (FH2) domain. MATERIALS AND METHODS Protein expression and purification All the constructs used in this study were amplified from the R. conorii sca2 gene (GenBank: AAL02648). All of the constructs, except Sca1090–1355, were expressed and purified as described previously (12), with slight modifications. Briefly, constructs Sca34–1515, Sca34–670, Sca34–400, Sca421–670, Sca869–1060, Sca1090–1515, and Sca868–1515 were cloned into the pTYB11 vector (New England BioLabs, Ipswich, MA) containing a chitin affinity tag and an intein domain. Sca1090–1355 was cloned into the pRSFduet vector with an N-terminal His-tag and an engineered N-terminal TEV protease site. All the proteins were expressed in Escherichia coli Rosetta or Arctic Express cells grown in Terrific Broth media. Cells were resuspended in chitin buffer (20 mM HEPES (pH 7.5), 500 mM NaCl, 1 mM EDTA, 4 mM benzamidine) or, in the case of construct Sca1090–1355, nickel lysis buffer (20 mM Tris-HCl (pH 7.5), 200 mM NaCl, 20 mM imidazole, 4 mM benzamidine) and lysed using a microfluidizer (MicroFluidics Corporation, Newton, MA). After purification on a chitin affinity column, the chitin tag was removed by self-cleavage of the intein, induced by the addition of 50 mM dithiothreitol (DTT) for 1–2 days at 20 C, which leaves no extra residues on the Sca2 constructs. For Sca1090–1355, the N-terminal Histag was cleaved after elution from the nickel affinity column by the addition of TEV protease at 4 C overnight, followed by a second round of nickel affinity to remove the TEV protease and any uncut protein. All the proteins were additionally purified by either ion exchange or size exclusion chromatography. Point mutations in the WH2 domains of Sca34–1515 were introduced using the QuikChange II XL site-directed mutagenesis kit (Stratagene, San Diego, CA) as before (12) and expressed and purified as described above. The putative WH2 peptides of Sca2 (Sca871–890, Sca942–962, Sca1003–1022) and that of Wiskott-Aldrich syndrome protein family member 1 (WAVE1) were cloned as maltose-binding protein (MBP) fusions into a modified pMAL-c2x vector containing an N-terminal His-tag and an engineered N-terminal TEV protease site by primer extension to express proteins with the following overall organization: His-TEV-MBPlinker(AAANNNNNNNNNNLG)-WH2. These fusion proteins were expressed in E. coli Arctic cells and purified by nickel affinity followed by ion exchange chromatography.
Biophysical Journal 116, 540–550, February 5, 2019 541
Alqassim et al.
Analytical gel filtration chromatography To investigate the formation of complexes between the various N- and C-terminal domains of Sca2, proteins were mixed, incubated for 30 min, and run through a Superose 6 10/300 GL column (GE Healthcare, Waukesha, WI) in a buffer containing 20 mM Tris-HCl (pH 7.5), 200 mM NaCl, and 1 mM DTT. Peak fractions were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).
Actin polymerization assay Actin polymerization was measured as the fluorescence increase resulting from the incorporation of pyrene-labeled actin into filaments, using a Cary Eclipse Fluorescence Spectrophotometer (Varian, Agilent, Santa Clara, CA). Before data acquisition, 2 mM Mg-ATP-actin (6% pyrene labeled) was mixed with different concentrations of Sca2 constructs in F-buffer (10 mM Tris-HCl (pH 7.5), 1 mM MgCl2, 50 mM KCl, 1 mM EGTA, 0.1 mM NaN3, 0.2 mM ATP). Data acquisition started 5–10 s after mixing. All the measurements were done at 25 C. Control experiments were carried out with the addition of buffer alone.
Isothermal titration calorimetry ITC measurements were carried out on a VP-isothermal titration calorimetry instrument (MicroCal, Northhampton, MA). Protein samples were dialyzed for 2 days against ITC buffer (20 mM HEPES (pH 7.5), 200 mM NaCl, 1 mM DTT, and 35 mM LatB when monomeric actin is present). Titrations consisted of 10 mL injections, lasting for 25 s, with an interval of 300 s between injections. The heat of binding was corrected for the heat of injection, determined by injecting proteins into buffer. Data were analyzed using the program Origin (OriginLab, Northampton, MA). The temperature and parameters of the fit (stoichiometry and affinity) of each experiment are given in the figures.
RESULTS Intramolecular interaction of the N- and C-terminal regions of Sca2 Previously, we had shown that the N- and C-terminal regions of Sca2, fragments 34–670 (Sca34–670; see Fig. 1 A for construct definition) and 670–1515 (Sca670–1515), interact with each other—i.e., they copurify when coexpressed—and together, they replicate the activity of the full-length passenger domain (Sca34–1515) (12). Furthermore, deleting 200 amino acids near the middle of the protein (residues 671–867) did not significantly affect this interaction or the nucleation activity; coexpressed constructs His-Sca34–670 and GST-Sca868–1515 elute together during purification, and their complex can reproduce most of the nucleation activity of Sca34–1515 (12). To better understand this intramolecular interaction and the overall organization of Sca2 domains for nucleation, we tested the effect of N- and C-terminal truncations of constructs Sca34–670 and Sca868–1515 on the interaction. The two interacting constructs Sca34–670 and Sca868–1515 were used as a control in this analysis. Because previously we had tested their interaction only through copurification using two different affinity tags (12), here, we expressed the Sca34–670 and Sca868–1515 fragments independently and free of tags (see Materials
542 Biophysical Journal 116, 540–550, February 5, 2019
and Methods) and characterized their interaction quantitatively using isothermal titration calorimetry (ITC). Fitting of the ITC titration of Sca34–670 (in the syringe) into Sca868–1515 (in the ITC cell) showed that the two fragments interact with relatively high affinity (KD ¼ 7.0 5 0.3 mM) and the reaction was exothermic in character (Fig. 1 B). Furthermore, when mixed together, the two independently purified Sca2 fragments run as a single peak containing approximately equal amounts of both proteins by analytical gel filtration and SDS-PAGE analysis (Fig. 1 C), consistent with the formation of a relatively tight complex. To identify the minimal regions necessary for this intramolecular interaction, we tested binding of the intact N-terminal fragment (Sca34–670) with independently expressed N- and C-terminal truncations of the C-terminal fragment (Sca868–1515) and vice versa (Fig. 1, D–H). Our results show that truncations at either end of the N-terminal (Sca34–400 and Sca421–670) or C-terminal (Sca869–1060, Sca1090–1355, and Sca1090–1515) regions of Sca2 abrogate the interaction because the premixed proteins emerge separately by analytical gel filtration and SDS-PAGE analysis of the peak fractions. For one construct combination, Sca34–670 and Sca1090–1355, the gel filtration peak of the individually purified proteins partially overlaps with that of their mixture, but SDS-PAGE analysis shows that the two constructs are largely segregated toward the ends of the gel filtration peak, consistent with lack of interaction (Fig. 1 F). Therefore, Sca34–670 and Sca868–1515 are the smallest fragments tested here that can still account for the intramolecular interaction of Sca2. Moreover, analogous to the FH2 domain, these two fragments seem to interact in a circular, head-totail manner, since deletions at either end of both constructs interfere with their ability to form a complex. Sca2 binds two actin subunits with high affinity The FH2 domain of eukaryotic formins also interacts in a circular, head-to-tail manner to form a dimeric ring-like structure, which encircles in the middle two actin subunits at the barbed end of the actin filament (24,25). Each subunit of the FH2 dimer contributes to the overall binding affinity, which is typically higher for the dimer than the hemidimer (i.e., a single FH2), although reported values diverge widely for different formins (26–31). We hypothesized that the N- and C-terminal fragments of Sca2 may interact with actin in a similar manner. To test this idea, we used ITC to assess the interaction of the full-length passenger domain of Sca2 (construct Sca34–1515) with monomeric actin. Latrunculin B (LatB), a marine toxin that binds in the catalytic cleft near the nucleotide in actin, was used in these experiments to prevent actin polymerization. The titration of LatB-actin into Sca34–1515 produced an exothermic reaction (Fig. 2 A), whose binding isotherm fitted best to a two-site binding model (i.e., two actin subunits per Sca34–1515 molecule), and both sites had high affinity (KD1 ¼ 0.19 mM and
Actin Assembly by Rickettsia Sca2
A
B
C
D
E
F
G
H
FIGURE 1 Intramolecular interaction of the N- and C-terminal regions of Sca2. (A) Domain organization of Sca2 and constructs used in this study (SS, signaling sequence; NRD, N-terminal repeat domain; PRD1,2, proline-rich domains; WH2a–c, putative WH2 domains; CRD, C-terminal repeat domain; AC, predicted autochaperone domain; TD, translocator domain). (B) ITC titration of 200 mM Sca34–670 (in the syringe) into 20 mM Sca868–1515 (in the cell). The experiment was performed at 25 C. The dissociation constant (KD) and binding stoichiometry (N) derived from fitting of the binding isotherm are listed. Errors correspond to the SD of the fits. Open symbols correspond to titrations into buffer. (C–H) Analytical size exclusion chromatography (SEC) and SDSPAGE analyzes of mixtures of untagged N- and C-terminal Sca2 constructs as indicated. The gels shown as insets correspond to the SEC fractions of the peak, (legend continued on next page)
Biophysical Journal 116, 540–550, February 5, 2019 543
Alqassim et al.
A
C
B
D
KD2 ¼ 0.02 mM) (Fig. 2 A, blue curve). Fitting to a singlesite model, shown for comparison (Fig. 2 A, red curve), produced a much worse c2 value (109.5 compared to 8.5) and unrealistic stoichiometry (N ¼ 0.6). Note that the c2 values reported here are reduced, i.e., divided by degrees of freedom, and although these values provide a measure of the goodness of the fit, they cannot be compared across different experiments because the data are not weighted and the c2 depends on the magnitude of the scale.
FIGURE 2 Interaction of Sca2 constructs with monomeric actin. (A) ITC titration of 89 mM LatBactin (in the syringe) into 8.3 mM Sca34–1515 (in the cell). The experiment was performed at 20 C. A two-site binding model produced a better fit of the data to a binding isotherm than a one-site binding model, as indicated by the c2 values and realistic stoichiometry. The c2 values reported here are reduced, i.e., divided by degrees of freedom. Note that although the c2 value clearly defines the best fitting model for a particular titration, they are not generally comparable across different titrations because the data are not weighted and the c2 depends on the magnitude of the scale. The dissociation constant (KD) and binding stoichiometry (N) derived from the fits are listed. Errors correspond to the SD of the fits. (B) ITC titration of 220 mM Sca34–670 (in the syringe) into 22 mM LatB-actin (in the cell). The experiment was performed at 12.7 C. A two-site binding model produced an almost equally good fit of the data to a binding isotherm than a one-site binding model, but is prefered based on a more realistic stoichiometry of the interaction. A two-site binding model is also supported by the fact that this fragment of Sca2 has nucleation activity (12). The dissociation constant (KD) and binding stoichiometry (N) derived from the fits are listed. Errors correspond to the SD of the fits. (C) ITC titration of 444 mM Sca869–1060 (in the syringe) into 16.6 mM LatB-actin (in the cell). The experiment was performed at 20 C. The dissociation constant (KD) and binding stoichiometry (N) derived from fitting of the binding isotherm are listed. Errors correspond to the SD of the fits. (D) ITC titration of 396 mM Sca1090–1355 (in the syringe) into 19 mM LatB-actin (in the cell). The experiment was performed at 20 C. No binding is reported because the experiment could not be fitted to a binding isotherm. The different fits are colorcoded as indicated. To see this figure in color, go online.
We then asked whether the N- and C-terminal fragments of Sca2 independently contributed to this interaction. Curiously, the titration of the N-terminal half of Sca2 (Sca34–670) into LatB-actin produced an endothermic reaction. The titration was first fitted to a one-site binding model, but the stoichiometry of the interaction was unrealistically low (N ¼ 0.33), suggesting that Sca34–670 interacts with two actin subunits (Fig. 2 B). Fitting to a two-site binding model yielded two binding sites of lower affinity than those of the
indicated by a dashed box. (C) shows that the premixed constructs Sca34–670 and Sca868–1515 run together by SEC, consistent with the formation of a complex. (D)–(F) show that the N-terminal fragment Sca34–670 runs separately from N- and C-terminal deletions of construct Sca868–1515 when premixed, consistent with lack of interaction. (G) and (H) show that the C-terminal fragment Sca868–1515 runs separately from N- and C-terminal deletions of constructs Sca34–670 when premixed, consistent with lack of interaction. The SEC traces of the Sca2 constructs are color-coded as indicated. To see this figure in color, go online.
544 Biophysical Journal 116, 540–550, February 5, 2019
Actin Assembly by Rickettsia Sca2
full-length passenger domain (KD1 ¼ 33.1 mM and KD2 ¼ 8.8 mM), although the c2 value (1.2) did not improve. A two-site binding model is preferred in this case because it addresses the low stoichiometry obtained with the one-site binding model and explains why this fragment of Sca2 displays nucleation activity, albeit lower than that of the fulllength passenger domain (12). We could not test binding of the entire C-terminal region (Sca868–1515) to LatB-actin because of the very low yields of this construct when expressed in isolation. Instead, we tested binding of two fragments that make up this construct (Sca869–1060 and Sca1090–1355). Sca869–1060, which contains a predicted repeat of three WH2 domains (see below), bound only one LatB-actin monomer with a KD of 10.8 mM (Fig. 2 C), and the reaction was endothermic in character, like for Sca34–670. In contrast, Sca1090–1355 did not appear to bind LatB-actin (Fig. 2 D). Therefore, these results suggest that the two halves of Sca2 contribute to actin binding, analogous to the individual subunits of the FH2 dimer (26,27), but only when combined together can they achieve high-affinity binding of two actin monomers. The fact that the interactions change from endothermic for the individual fragments to exothermic for the full-length passenger domain and the affinities increase manyfold suggests that additional interactions, presumably including actinactin interactions, are involved in the formation of this ternary complex. Sca2 does not contain WH2 domains As established above, Sca2 binds only two actin monomers, whereas the region containing the putative repeat of three WH2 domains (Sca869–1060) only binds one actin monomer and with much weaker affinity than the full-length passenger domain. All the WH2 domains analyzed by ITC so far bind actin with higher affinity than Sca869–1060, and their binding is consistently exothermic in character (32–34). These observations raised questions about the authenticity of the WH2 domains of Sca2. This is important because it is still assumed that Sca2 contains WH2 domains (22), and it was the identification of putative WH2 domains based on sequence analysis that supported the initial hypothesis that Sca2 was a bacterial actin assembly factor, although curiously, two different laboratories had proposed different definitions of the WH2 domains (11,13). We use here the latest definition of the WH2 domains of Sca2 (11) (Fig. 3 A), which produces a better alignment with other known WH2 domains (Fig. 3 B). Yet, the initial disagreement is understandable when one compares the sequences of the WH2 domains of Sca2 with other canonical WH2 domains. The WH2 domain is 17 amino acids in length and consists of an N-terminal a-helix that binds in the hydrophobic or target-binding cleft of actin, a loop, and a C-terminal LKKV motif (22). All three of the WH2 domains of Sca2 lack at least one critical element of the consensus
sequence. Thus, WH2a (residues 772–888) has a proline residue within the LKKV motif (22). WH2b (residues 943–960) lacks a basic residue at the N-terminus and key hydrophobic residues in the N-terminal a-helix of the domain. WH2c (residues 1004–1020) lacks the two basic residues of the LKKV motif. Importantly, mutations in the LKKV motif of classical WH2 domains disrupt binding to monomeric actin (35,36). We then analyzed the sequences connecting the WH2 domains of Sca2. Here again, Sca2 is a clear outlier, with the putative WH2 domains spaced farther apart (45–55 amino acids) than in other proteins containing repeats of WH2 domains (15 amino acids) (Fig. 3 A). A flexible 15-amino-acid linker seems to be optimal for proteins in which successive WH2 domains connect actin subunits along the long-pitch helix of the actin filament (37). The one exception is Cobl, in which a 69-amino-acid linker connects the second and third WH2 domains. However, Sca2 differs in yet another way from canonical WH2 domaincontaining proteins; the sequences between WH2 domains are predicted to contain a series of a-helices, whereas in other proteins, these sequences are typically unstructured (see the secondary structure predictions in Fig. 3 A). Hydrophobic cluster analysis further supports this conclusion, revealing a series of clusters of hydrophobic residues in the inter-WH2 domain linkers of Sca2 but not other proteins (Fig. S1). Such clusters typically denote the presence of secondary structural elements, which, depending on the shape of the clusters, can be defined as either a-helices or b-strands (38). The hydrophobic cluster analysis plots of the WH2 domains themselves are also different between Sca2 and other WH2 domain-containing proteins (Fig. S1). The fact that the putative WH2 domains of Sca2 form part of a predicted globular domain is finally supported by far-ultraviolet circular dichroism analysis of an Sca2 fragment comprising amino acids 868–1515, revealing a typical a-helical profile with minima at 208 and 222 nm (12). Therefore, several lines of evidence suggest that Sca2 does not contain a typical WH2 domain repeat. However, we had previously found that a full-length passenger domain construct carrying four mutations in each of the WH2 domains (a total of 12 mutations targeting both the a-helix and LKKV motif of each WH2 domain) had no detectable nucleation or elongation activity (12), demonstrating the importance of this region for actin assembly. But, because this region binds only one actin monomer (Fig. 2 C), we asked whether only one of the putative WH2 domains was still functional despite their noncanonical sequences. For this, we generated three Sca34–1515 mutants individually targeting each of the WH2 domains. We used the same mutations as before (12), i.e., two conserved hydrophobic residues within the N-terminal a-helix and the first two residues of the LKKV motif were simultaneously mutated to aspartic acid (Fig. 3 C). Using the pyreneactin polymerization assay, we found that the mutants
Biophysical Journal 116, 540–550, February 5, 2019 545
Alqassim et al.
A
C
B
D
FIGURE 3 The region 868–1023 of Sca2 is critical for nucleation but lacks WH2 domains. (A) A cartoon representation of the WH2-domain-containing regions of various cytoskeletal proteins containing tandem repeats of WH2 domains. WH2 domains are colored red, with a cylinder representing the a-helix of the domain. Secondary structure predictions obtained with the program Jpred (45) are also shown for each protein (two-dimensional prediction), with blue cylinders representing predicted a-helices. (B) Alignment of the sequences of the WH2 domains represented in (A), showing the regions corresponding to the N-terminal a-helix and LKKV motif, as well as the consensus sequence at the bottom. (C) Representation of the three mutants of Sca34–1515 (full-length passenger domain) targeting the putative WH2 domains. Each mutant carries four aspartic acid mutations of highly conserved residues of the canonical sequence of the domain. The SDS-PAGE shown on the right illustrates the purity of one of the mutants (Sca34–1515-WH2a*). (D) A time course of the polymerization of 2 mM Mg-ATP-actin (6% pyrene labeled) alone and with the addition of 50 nM Sca34–1515 wild type and WH2 domain mutants depicted in (C) (color-coded as indicated). To see this figure in color, go online.
targeting WH2 domains a and b (Sca34–1515-WH2a* and Sca34–1515-WH2b*) inhibited the polymerization activity in a concentration-dependent manner, whereas mutant Sca34–1515-WH2c* had a negligible effect compared to the wild-type passenger domain (Fig. 3 D; Fig. S2). Therefore, at least two of the predicted WH2 domains of Sca2 are important for actin assembly within the full-length protein—is this because they bind actin directly the way WH2 domains do? To address this question, we generated four MBP fusion constructs with the WH2 domains of Sca2 and a control WH2 domain from WAVE1 fused C-terminally to MBP and separated from MBP by a noncleavable 15-amino-acid linker (Fig. 4 A). The proteins were highly pure, and the yields were sufficiently high to allow for quantitative ITC experiments to assess their binding to LatB-actin (Fig. 4 B). As anticipated, the first WH2 of WAVE1 (WAVE1 WH2a) bound LatB-actin with high affinity and 1:1 stoichiometry (KD ¼ 0.38 mM, N ¼ 1.07), and
546 Biophysical Journal 116, 540–550, February 5, 2019
the reaction was exothermic (Fig. 4 C). This interaction is therefore very similar to that of synthetic WH2 peptides from other cytoskeletal proteins we have measured before in G-buffer (i.e., without the addition of LatB) (33) or in F-buffer with the addition of LatB (32,34). In contrast, none of the putative WH2 domains of Sca2 bound LatBactin (Fig. 4, D–F), conclusively demonstrating that these are not functional WH2 domains. DISCUSSION Rickettsia Sca2 is a pathogenic bacterial actin assembly factor that mimics eukaryotic formins in that it generates long and unbranched actin filaments by combining both nucleation and processive barbed-end elongation activities (11). However, we have shown before that Sca2 is structurally unrelated to formins and uses a different mechanism for actin assembly (12). Specifically, Sca2 is monomeric and lacks
Actin Assembly by Rickettsia Sca2
A
B
D
C
E
F
FIGURE 4 The putative WH2 domains of Sca2 do not bind actin. (A) A representation of four WH2 domain constructs expressed here as fusion proteins with maltose-binding protein (MBP). A 15-amino-acid flexible linker separates the MBP moiety at the N-terminus from the WH2 domain of WAVE1 (control) or the three putative WH2 domains of Sca2. (B) Purity of proteins illustrated in (A) assessed by SDS-PAGE analysis. (C) ITC titration of 254 mM WAVE1 WH2a into 18.5 mM LatB-actin. The experiment was performed at 25 C. The dissociation constant (KD) and binding stoichiometry (N) derived from fitting of the binding isotherm are listed. Errors correspond to the SD of the fits. (D–F) ITC titrations of the three constructs of putative WH2 domains of Sca2 into LatB-actin. The concentrations of the proteins in the syringe and in the cell are indicated. The titrations could not be fitted to binding isotherms because none of the putative WH2 domains of Sca2 appeared to bind LatB-actin. To see this figure in color, go online.
the dimeric FH2 domain found in all eukaryotic formins, which is crucial for both nucleation and elongation (39). The one element of the Sca2 sequence that still appeared to resemble eukaryotic nucleators was the postulated repeat of three WH2 domains (22), whose identification first drew attention to this protein as a potential actin assembly factor (11,13). Of note, several eukaryotic formins contain WH2domain-related sequences or form complexes with proteins
that contain WH2 domains and synergize with formins for nucleation (40–44). Here, however, we have disproven the existence of WH2 domains in Sca2, suggesting that Rickettsia Sca2 has evolved completely independently of any eukaryotic actin assembly factor to acquire the various complex functions associated with eukaryotic formins: nucleation, elongation, and inhibition of barbed-end capping (11,12).
Biophysical Journal 116, 540–550, February 5, 2019 547
Alqassim et al.
How does Sca2 accomplish these functions using a different fold? In thinking about this question, a comparison with eukaryotic formins provides some clues. Formins are dimeric, and specifically the two FH2 domains of the dimer interact head to tail to form a ring-like structure (25), which encircles two actin subunits at the barbed end of the actin filament while loosely contacting a third incoming subunit during processive elongation (24). Sca2 cannot be a dimer because bacterial autotransporters are either monomeric or trimeric, depending on the fold of the translocator domain, which can be 300 (monomeric) or 75 (trimeric) amino acids long (18). The passenger domain of Sca2 is monomeric in solution (12) and, consistently, Sca2 has a 281amino-acid monomeric type of translocator domain. However, like formins, Sca2 must ‘‘hug’’ the barbed end of the filament while allowing access to additional actin subunits for processive elongation. This can only be accomplished if Sca2 also forms a ring-like structure around the barbed end, contacting primarily two actin subunits and possibly a third incoming subunit (Fig. 5). The data presented here are consistent with this model. The N- and C-terminal regions of Sca2 (Sca34–670 and Sca868–1515) seem to behave to a large extent like the FH2 domains of the formin dimer: 1) they bind actin individually but with weak affinity (Sca34–670 interacts with two actin subunits,
whereas Sca868–1515 appears to contact only one subunit); 2) together (i.e., the full-length passenger domain), they bind two actin subunits with high affinity and account for the nucleation and barbed-end elongation activities of Sca2, which the individual fragments mostly lack (12); 3) deletions at either end of these two fragments disrupt their ability to interact with each other, suggesting that they bind in head-to-tail fashion to form a ring-like structure analogous to the formin FH2 domain. One of these fragments, Sca869–1060, comprises what was previously thought to be a repeat of three WH2 domains. However, Sca869–1060 binds a single actin subunit with weak affinity and following an endothermic reaction, clearly distinct from the higher affinity and exothermic reactions of canonical WH2 domains (32–34). Moreover, we have shown here that none of the putative WH2 domains of Sca2 binds actin in isolation, suggesting that the adverse effect of mutations of two of the domains is due to either disruption of a contiguous actinbinding surface or global changes in the structure of this region, which we now call the ‘‘middle domain’’ because of its location within the full-length passenger domain (Fig. 5). Analogous to the proline-rich FH1 domain of formins, Sca2 also contains two proline-rich segments (672–699 and 1077–1087) that mediate the binding of profilin-actin for processive barbed end elongation (11,12). Both FIGURE 5 Proposed mechanism of nucleation and elongation by Rickettsia Sca2. By analogy with eukaryotic formins, Sca2 is proposed to form a ring-like structure around the barbed end of the actin filament, contacting primarily two actin subunits and possibly a third incoming subunit. The data presented here support this model and suggest that the N- and C-terminal regions of Sca2 (Sca34–670 and Sca868–1515) behave in a way analogous to the FH2 domain of formins. Thus, we have shown here that these two fragments bind actin individually but with weak affinity, whereas the full-length passenger domain binds two actin subunits with high affinity and has both nucleation and elongation activities, although the individual fragments are mostly inactive (11,12). Furthermore, deletions at either end of these two fragments disrupt their ability to interact with each other, suggesting that they interact head to tail to form a ring-like structure analogous to the formin FH2 domain. We have also shown that Sca2 does not contain WH2 domains. The region previously thought to contain the repeat of three WH2 domains (Sca869–1060) binds actin as a whole but only one actin subunit and in manner clearly distinct from that of canonical WH2 domains. Accordingly, this region is represented here as forming part of the ring-like structure of Sca2 and is renamed as the ‘‘middle domain’’ because of its location in the sequence. The two proline-rich domains, implicated in the recruitment of profilin-actin for barbed end elongation (12), may project out because they form part of loops that are predicted to be unstructured. To see this figure in color, go online.
548 Biophysical Journal 116, 540–550, February 5, 2019
Actin Assembly by Rickettsia Sca2 10. Benanti, E. L., C. M. Nguyen, and M. D. Welch. 2015. Virulent Burkholderia species mimic host actin polymerases to drive actin-based motility. Cell. 161:348–360.
segments form part of loop regions (658–714 and 1066– 1091), which are predicted to be unstructured and could therefore be loosely attached to the other domains of the Sca2 molecule that form the ring-like structure around the barbed end. None of the Sca2 fragments studied here could be crystallized (except Sca34–400, which we studied previously (12)), either alone or in complex with actin, but structural studies testing some of the predictions of this model should soon become possible using cryo-electron microscopy.
13. Kleba, B., T. R. Clark, ., T. Hackstadt. 2010. Disruption of the Rickettsia rickettsii Sca2 autotransporter inhibits actin-based motility. Infect. Immun. 78:2240–2247.
SUPPORTING MATERIAL
14. Schroeder, C., I. Chowdhury, ., S. Sahni. 2016. Human rickettsioses: host response and molecular pathogenesis. In Rickettsiales. S. Thomas, ed. Springer, pp. 399–446.
Two figures are available at http://www.biophysj.org/biophysj/supplemental/ S0006-3495(18)34504-1.
AUTHOR CONTRIBUTIONS S.S.A. and R.D. designed the experiments. S.S.A. and I.-G.L. prepared the proteins, including cloning, expression and purification, and S.S.A. performed all of the experiments for this work. S.S.A. and R.D. wrote the manuscript. All authors analyzed the data and reviewed the figures and manuscript.
ACKNOWLEDGMENTS We thank Grzegorz Rebowski for the preparation of actin for this study and help with Fig. 5, and Malgorzata Boczkowska for guidance with the actin polymerization assay, ITC experiments, and data analysis. This work was supported by National Institutes of Health grant R01 GM073791.
11. Haglund, C. M., J. E. Choe, ., M. D. Welch. 2010. Rickettsia Sca2 is a bacterial formin-like mediator of actin-based motility. Nat. Cell Biol. 12:1057–1063. 12. Madasu, Y., C. Suarez, ., R. Dominguez. 2013. Rickettsia Sca2 has evolved formin-like activity through a different molecular mechanism. Proc. Natl. Acad. Sci. USA. 110:E2677–E2686.
15. Gouin, E., C. Egile, ., P. Cossart. 2004. The RickA protein of Rickettsia conorii activates the Arp2/3 complex. Nature. 427:457–461. 16. Jeng, R. L., E. D. Goley, ., M. D. Welch. 2004. A Rickettsia WASPlike protein activates the Arp2/3 complex and mediates actin-based motility. Cell. Microbiol. 6:761–769. 17. Reed, S. C. O., R. L. Lamason, ., M. D. Welch. 2014. Rickettsia actinbased motility occurs in distinct phases mediated by different actin nucleators. Curr. Biol. 24:98–103. 18. Dautin, N., and H. D. Bernstein. 2007. Protein secretion in gram-negative bacteria via the autotransporter pathway. Annu. Rev. Microbiol. 61:89–112. 19. Bernstein, H. D. 2015. Looks can be deceiving: recent insights into the mechanism of protein secretion by the autotransporter pathway. Mol. Microbiol. 97:205–215. 20. Junker, M., C. C. Schuster, ., P. L. Clark. 2006. Pertactin b-helix folding mechanism suggests common themes for the secretion and folding of autotransporter proteins. Proc. Natl. Acad. Sci. USA. 103:4918–4923. 21. Kajava, A. V., and A. C. Steven. 2006. The turn of the screw: variations of the abundant b-solenoid motif in passenger domains of Type V secretory proteins. J. Struct. Biol. 155:306–315. 22. Dominguez, R. 2016. The WH2 domain and actin nucleation: necessary but insufficient. Trends Biochem. Sci. 41:478–490.
REFERENCES 1. Bugalha˜o, J. N., L. J. Mota, and I. S. Franco. 2015. Bacterial nucleators: actin’ on actin. Pathog. Dis. 73:ftv078. 2. Jimenez, A., D. Chen, and N. M. Alto. 2016. How bacteria subvert animal cell structure and function. Annu. Rev. Cell Dev. Biol. 32:373–397. 3. Dominguez, R. 2015. Subversive bacteria reveal new tricks in their cytoskeleton-hijacking arsenal. Nat. Struct. Mol. Biol. 22:178–179. 4. Lamason, R. L., and M. D. Welch. 2017. Actin-based motility and cellto-cell spread of bacterial pathogens. Curr. Opin. Microbiol. 35:48–57. 5. Truong, D., J. W. Copeland, and J. H. Brumell. 2014. Bacterial subversion of host cytoskeletal machinery: hijacking formins and the Arp2/3 complex. BioEssays. 36:687–696. 6. Liverman, A. D., H. C. Cheng, ., K. Orth. 2007. Arp2/3-independent assembly of actin by Vibrio type III effector VopL. Proc. Natl. Acad. Sci. USA. 104:17117–17122. 7. Yu, B., H. C. Cheng, ., M. K. Rosen. 2011. Mechanism of actin filament nucleation by the bacterial effector VopL. Nat. Struct. Mol. Biol. 18:1068–1074.
23. Kovar, D. R., E. S. Harris, ., T. D. Pollard. 2006. Control of the assembly of ATP- and ADP-actin by formins and profilin. Cell. 124:423–435. 24. Otomo, T., D. R. Tomchick, ., M. K. Rosen. 2005. Structural basis of actin filament nucleation and processive capping by a formin homology 2 domain. Nature. 433:488–494. 25. Xu, Y., J. B. Moseley, ., M. J. Eck. 2004. Crystal structures of a Formin Homology-2 domain reveal a tethered dimer architecture. Cell. 116:711–723. 26. Shimada, A., M. Nyitrai, ., A. Wittinghofer. 2004. The core FH2 domain of diaphanous-related formins is an elongated actin binding protein that inhibits polymerization. Mol. Cell. 13:511–522. 27. Zigmond, S. H. 2004. Formin-induced nucleation of actin filaments. Curr. Opin. Cell Biol. 16:99–105. 28. Li, F., and H. N. Higgs. 2003. The mouse Formin mDia1 is a potent actin nucleation factor regulated by autoinhibition. Curr. Biol. 13:1335–1340. 29. Moseley, J. B., I. Sagot, ., B. L. Goode. 2004. A conserved mechanism for Bni1- and mDia1-induced actin assembly and dual regulation of Bni1 by Bud6 and profilin. Mol. Biol. Cell. 15:896–907. 30. Pring, M., M. Evangelista, ., S. H. Zigmond. 2003. Mechanism of formin-induced nucleation of actin filaments. Biochemistry. 42:486–496.
8. Namgoong, S., M. Boczkowska, ., R. Dominguez. 2011. Mechanism of actin filament nucleation by Vibrio VopL and implications for tandem W domain nucleation. Nat. Struct. Mol. Biol. 18:1060–1067.
31. Kovar, D. R., J. R. Kuhn, ., T. D. Pollard. 2003. The fission yeast cytokinesis formin Cdc12p is a barbed end actin filament capping protein gated by profilin. J. Cell Biol. 161:875–887.
9. Burke, T. A., A. J. Harker, ., D. R. Kovar. 2017. The bacterial virulence factors VopL and VopF nucleate actin from the pointed end. J Cell Biol. 216:1267–1276.
32. Lee, S. H., F. Kerff, ., R. Dominguez. 2007. Structural basis for the actin-binding function of missing-in-metastasis. Structure. 15: 145–155.
Biophysical Journal 116, 540–550, February 5, 2019 549
Alqassim et al. 33. Chereau, D., F. Kerff, ., R. Dominguez. 2005. Actin-bound structures of Wiskott-Aldrich syndrome protein (WASP)-homology domain 2 and the implications for filament assembly. Proc. Natl. Acad. Sci. USA. 102:16644–16649. 34. Boczkowska, M., Z. Yurtsever, ., R. Dominguez. 2017. Crystal structure of Leiomodin 2 in complex with actin: a structural and functional reexamination. Biophys. J. 113:889–899. 35. Mattila, P. K., M. Salminen, ., P. Lappalainen. 2003. Mouse MIM, a tissue-specific regulator of cytoskeletal dynamics, interacts with ATPactin monomers through its C-terminal WH2 domain. J. Biol. Chem. 278:8452–8459. 36. Quinlan, M. E., J. E. Heuser, ., R. D. Mullins. 2005. Drosophila Spire is an actin nucleation factor. Nature. 433:382–388. 37. Rebowski, G., S. Namgoong, ., R. Dominguez. 2010. Structure of a longitudinal actin dimer assembled by tandem w domains: implications for actin filament nucleation. J. Mol. Biol. 403:11–23.
40. Chhabra, E. S., and H. N. Higgs. 2006. INF2 Is a WASP homology 2 motif-containing formin that severs actin filaments and accelerates both polymerization and depolymerization. J. Biol. Chem. 281: 26754–26767. 41. Heimsath, E. G., Jr., and H. N. Higgs. 2012. The C terminus of formin FMNL3 accelerates actin polymerization and contains a WH2 domainlike sequence that binds both monomers and filament barbed ends. J. Biol. Chem. 287:3087–3098. 42. Gould, C. J., S. Maiti, ., B. L. Goode. 2011. The formin DAD domain plays dual roles in autoinhibition and actin nucleation. Curr. Biol. 21:384–390. 43. Quinlan, M. E., S. Hilgert, ., E. Kerkhoff. 2007. Regulatory interactions between two actin nucleators, Spire and Cappuccino. J. Cell Biol. 179:117–128.
38. Bitard-Feildel, T., A. Lamiable, ., I. Callebaut. 2018. Order in disorder as observed by the ‘‘hydrophobic cluster analysis’’ of protein sequences. Proteomics. 18:e1800054.
44. Pechlivanis, M., A. Samol, and E. Kerkhoff. 2009. Identification of a short Spir interaction sequence at the C-terminal end of formin subgroup proteins. J. Biol. Chem. 284:25324–25333.
39. Breitsprecher, D., and B. L. Goode. 2013. Formins at a glance. J. Cell Sci. 126:1–7.
45. Drozdetskiy, A., C. Cole, ., G. J. Barton. 2015. JPred4: a protein secondary structure prediction server. Nucleic Acids Res. 43:W389–W394.
550 Biophysical Journal 116, 540–550, February 5, 2019