Cooperation of two distinct coupling proteins creates chemosensory network connections

Significance

Signal transduction systems are important pathways that organisms use to sense and respond to their environments. Chemotaxis is controlled by a signal transduction system that allows bacteria to coordinate their movement in response to their environment. This response requires proper assembly and localization of large multiprotein chemotaxis complexes, which are built by interactions between coupling proteins. The significance of having multiple types of coupling proteins in a single signal transduction system is poorly understood. Here, we show that multiple coupling proteins allow bacteria to build superprotein interaction networks and localize them to cell poles, a role that is required for optimal chemotaxis.

Abstract

Although it is appreciated that bacterial chemotaxis systems rely on coupling, also called scaffold, proteins to both connect input receptors with output kinases and build interkinase connections that allow signal amplification, it is not yet clear why many systems use more than one coupling protein. We examined the distinct functions for multiple coupling proteins in the bacterial chemotaxis system of Helicobacter pylori, which requires two nonredundant coupling proteins for chemotaxis: CheW and CheV1, a hybrid of a CheW and a phosphorylatable receiver domain. We report that CheV1 and CheW have largely redundant abilities to interact with chemoreceptors and the CheA kinase, and both similarly activated CheA’s kinase activity. We discovered, however, that they are not redundant for formation of the higher order chemoreceptor arrays that are known to form via CheA–CheW interactions. In support of this possibility, we found that CheW and CheV1 interact with each other and with CheA independent of the chemoreceptors. Therefore, it seems that some microbes have modified array formation to require CheW and CheV1. Our data suggest that multiple coupling proteins may be used to provide flexibility in the chemoreceptor array formation.

Coupling or scaffold proteins provide critical connections between input receptors and output kinases in many types of signal transduction pathways (1–3). These connections confer multiple advantages such as cooperativity, signaling complex assembly, and protein localization (2). Indeed, many cellular signaling systems use multiple coupling proteins to fine tune these advantages, a process that has been well studied in eukaryotic systems (2, 4, 5). Bacterial chemotaxis is an example of a prokaryotic system that relies on coupling proteins of the CheW family to connect chemoreceptors to the CheA kinase and build connections that create multiprotein chemoreceptor arrays. Many bacterial chemotaxis systems possess multiple coupling proteins (6, 7), but the functions and advantages of having more than one coupling protein for these systems are not well understood.

The core bacterial chemotaxis sensory unit is composed of a chemoreceptor, a CheW family coupling protein, and the CheA output kinase (6). The coupling protein allows the chemoreceptors to control the CheA kinase and promotes connections between core units (3, 6–8). Chemotaxis coupling proteins have two basic architectures, CheW or CheV. CheW is a single domain protein with two defined subdomains, and CheV proteins are hybrids that add a C-terminal response regulator-like domain (Rec) to an N-terminal CheW domain (3, 9).

Chemotaxis core sensory units exist in cells as large multiprotein arrays at the poles (8, 10–12). Connections that form these arrays have been elucidated in the single-coupling protein system of Escherichia coli and are driven by interactions between CheW and a subdomain of CheA called P5, which is a structural mimic of CheW. These connections are vital for array formation, kinase control, and positive cooperativity that allow small signals to be greatly amplified (11). There are two documented types of CheW–CheA P5 interactions. Interactions at interface 1 occur between CheA P5 subdomain 1 and CheW subdomain 2 and lead to control of CheA kinase activity (3, 8). Interactions at interface 2 occur between CheA P5 subdomain 2 and CheW subdomain 1 and lead to intercomplex connections that build chemoreceptor arrays and positive cooperativity (3, 8). Thus, coupling proteins participate in two types of CheA interactions that are vital for chemotaxis function.

Although these types of CheW–CheA interactions have been elucidated in the single coupling protein E. coli system, it is not yet known how and whether these interactions differ in multicoupling protein systems. To date, the best-studied multicoupling protein system is that of Bacillus subtilis, a microbe that uses one CheV and one CheW to both perform receptor–kinase coupling in a somewhat functionally redundant manner (13, 14). There is evidence suggesting that some chemoreceptors might operate with particular coupling proteins, displaying greater affinity for one coupling protein than for another (15, 16). CheW and CheV may also have some affinity bias toward different chemoreceptors (17).

The human pathogen Helicobacter pylori has a chemotaxis system with four chemoreceptors called Tlps (TlpA, TlpB, TlpC, and TlpD), a CheA kinase, a CheY response regulator, and—relevant to this work—multiple coupling proteins (18, 19). Two of the H. pylori coupling proteins, CheW and CheV1, are critical for wild-type chemotaxis, acting in a nonredundant manner. Mutants lacking cheW or cheV1 appear unable to activate the CheA kinase as they swim without changing direction and are either completely (cheW) or severely (cheV1) compromised in a soft agar chemotaxis assay (9, 20). H. pylori also possesses two other CheV-type coupling proteins, but these play only minor roles in chemotaxis (9, 20). Because H. pylori CheW and CheV1 were both essential for chemotaxis in a nonredundant manner, we thought it ideal to dissect how these contribute to chemotaxis.

We initiated our work analyzing the protein interaction network of CheW and CheV1 as well as their ability to activate and control CheA’s kinase function. We found that they had nearly identical abilities in these regards, which did not explain why both were needed. However, when we examined their roles in assembly of the polar chemosensory array, we found that both were required and fulfilled nonredundant roles in these interactions. Our data demonstrate that some microbes use multiple coupling proteins to build the polar chemoreceptor supercluster and thus suggest this aspect of chemotaxis may be fine-tunable by modulating the levels or activities of these coupling proteins.

Results

CheW and CheV1 Both Form Direct Interactions with CheA and Chemoreceptors.

To gain an understanding of why H. pylori requires two coupling proteins, we first characterized the protein–protein interaction network of H. pylori CheV1 and CheW. We identified directly interacting proteins using the bacterial adenylate cyclase two-hybrid (BACTH) system (21). For this approach, CheV1 and CheW were fused to the N or C terminus of T25 fragments, and CheV1, CheW, CheA, CheV2, CheV3, and the chemoreceptors (TlpA, TlpB, and TlpD) were fused to the bait T18 fragments. We found that both CheV1 and CheW displayed interactions with themselves, the other coupling proteins, CheA, and at least one chemoreceptor (Fig. 1 A and C). We found that only one fusion orientation—with the T25 fragment at the C-terminal end—was functional (Table S1). We thus focused on the functional fusions and quantified all positive interactions using a β-galactosidase assay (Fig. 1). CheV1 and CheW both displayed typical coupling protein interactions with CheA and chemoreceptors. The interaction with CheA was quite strong, yielding β-galactosidase levels that were almost double that of the positive control (Fig. 1). Both proteins also interacted with the TlpA chemoreceptor, with CheV1 showing significant interactions additionally with TlpB and TlpD (Fig. 1). Based on the X-gal plate, CheW did not appear to interact with TlpB and TlpD, as they both appeared similar to the negative control (white colonies) shown in Fig. 1C and Table S1. Both CheV1 and CheW were able to interact with one another and with themselves. Lastly, CheV1 interacted with CheV2 and CheV3.

Fig. 1.

BACTH analysis of CheV1 and CheW interactions. H. pylori (A) CheV1 and (B) CheW were fused to the N or C termini of the T25 fragments and tested for interaction with CheV1, CheV2, CheV3, CheW, CheA, TlpA, TlpB, or TlpD, fused to the N or C termini of the T18 fragments in E. coli cya⁻ BTH101. Positive (+) control, pKT25-zip and pUT18C-zip; negative (–) control, CheV1 or CheW plasmids cotransformed with empty T18 plasmids. (A and C) Representative LB X-gal and IPTG plates (n = 3). (B and D) β-galactosidase levels from positive interactions. n = 3 and error bars indicate SDs. Strains were compared with a negative control using unpaired t test (*P = 0.0127, **P < 0.001, ***P < 0.0001).

Table S1.

Bacterial two-hybrid interactions

Protein fusion	CheV1 → T25	T25 → CheV1	CheW → T25	T25 → CheW
T18 → CheA	+	–	+	–
CheA → T18	+	–	+	–
T18 → TlpA	+	–	–	–
TlpA → T18	+	–	+/−	–
T18 → TlpB	+	–	–	–
TlpB → T18	+	–	–	–
T18 → TlpD	–	–	–	–
TlpD → T18	+/−	–	–	–
T18 → CheW	+	–	+	–
CheW → T18	–	–	–	–
T18 → CheV1	–	–	–	–
CheV1 → T18	+	–	+	–
T18 → CheV2	–	–	–	–
CheV2 → T18	+	–	–	–
T18 → CheV3	–	–	–	–
CheV3 → T18	+/–	–	–	–

The chemotaxis proteins were fused to the N and C terminus of the adenylate cyclase T25 and T18 fragments. + indicates protein pairs that showed blue colonies and thus indicate protein interaction; – indicates protein pairs that showed white colonies and thus no interaction; +/− indicates protein pairs that were hard to distinguish. The colors determined were confirmed independently on three different occasions.

The BACTH protein interactions were verified using coimmunoprecipitation with purified proteins. Consistent with the BACTH, both CheW and CheV1 interacted with CheA and H. pylori’s cytoplasmic chemoreceptor, TlpD (Fig. 2 A and B). It is interesting to note that CheW did interact with the TlpD chemoreceptor in the coimmunoprecipitation experiments but not in the BACTH experiments. Based on these data, it seems possible that the BACTH cloning could have affected CheW’s structure and thus its ability to interact with TlpD. All H. pylori chemoreceptors have highly similar signaling regions that conserve residues that interact with coupling proteins in other systems (Fig. S1). Because of the high degree of similarity and the fact that TlpD is soluble in its full-length form, we used TlpD to characterize chemoreceptor–coupling protein interactions. It is possible, however, that CheW and CheV1 interact with the other chemoreceptors with varying strength or preference (7, 17). Finally, we saw that both CheV1 and CheW interacted with each other (Fig. 2C). Thus, the BACTH and coimmunoprecipitation results suggest that the protein interaction networks of both CheV1 and CheW are largely similar (Fig. 2D). Both interact directly with CheA, chemoreceptors, each other, and themselves. CheV1 may have several additional interactions with other coupling proteins, but these were not pursued (Fig. 2D).

Fig. 2.

Coimmunoprecipitation with purified proteins confirm that CheV1 and CheW interact with CheA, TlpD, and each other. (A–C) Mixtures of purified CheA (A), TlpD, CheV1 (V1), and/or CheW (W) were preincubated and immunoprecipitated as indicated along the top, followed by detection of specific proteins as indicated at the left. Each gel is representative of three immunoprecipitations (n = 3). (D) Model of the CheV1 and CheW protein interaction network identified by BACTH and coimmunoprecipitation, with line thickness denoting interaction strength.

Fig. S1.

Alignment of the signaling domain of H. pylori chemoreceptors. Alignment of the signaling domain of the H. pylori chemoreceptors, TlpD (HPG27_559), TlpA (HPG27_91), TlpB (HPG27_95), and TlpC (HELPY_0078) was done using T-Coffee Software Version 11. Red highlighting corresponds to good alignment, and yellow highlighting corresponds to average alignment.

Both CheW and CheV1 Possess Receptor–CheA Coupling.

We next tested whether CheW and CheV1 could each allosterically function to activate CheA autophosphorylation and to couple CheA to a chemoreceptor. These studies were done using an in vitro CheA phosphorylation assay similar to one used in previous studies (22, 23). In this assay, purified CheA ± coupling protein and receptor are incubated in vitro with radioactive [γ^-32P]-ATP and the amount of phosphorylated CheA determined using phosphorimaging of SDS PAGE gels (Fig. 3 A and D).

Fig. 3.

CheV1 and CheW activate CheA phosphorylation on their own and with the addition of the chemoreceptor TlpD. Equimolar mixtures of purified proteins as indicated (2 µM of each) were incubated with [γ-³²P]-ATP and phosphorylated CheA detected using SDS/PAGE. (A and D) Representative blots. (B and E) The level of phosphorylated CheA was compared with phosphorylated CheA at time 0 (C and F) The amount of CheA was calculated using the area under the curve with the definite integral as opposed to the rate, as the rate was not linear. n = 3, with error bars representing SD. *P < 0.001 determined by an unpaired t test for all compared with CheA only.

We first examined how addition of CheW or CheV1 to purified CheA protein would allosterically affect CheA phosphorylation, by measuring total phosphorylated CheA (Fig. 3B). CheV1 or CheW protein each significantly increased the CheA activity by 1.8- or 2.8-fold, respectively (Fig. 3C). These differences were not changed upon addition of a greater amount of CheW or CheV1 protein, suggesting the CheA was largely saturated. These results suggest that both CheV1 and CheW cause CheA to be more active, with CheW triggering greater activation. Previous work has shown that CheA activation is maximal with coupling proteins (24, 25).

We then examined CheV1 and CheW for each one’s ability to couple CheA to a chemoreceptor. In other systems, CheA becomes substantially more active when coupled to a chemoreceptor (26, 27). We thus incubated purified CheA with TlpD and either CheV1 or CheW (Fig. 3D). CheA phosphorylation was significantly increased with the addition of the chemoreceptor TlpD alone by twofold (Fig. 3F). The addition of CheV1 or CheW to the chemoreceptor–TlpD reaction significantly increased CheA phosphorylation by threefold or 4.8-fold, respectively, compared with CheA alone (Fig. 3F). Although both coupling proteins significantly increased CheA phosphorylation, CheW was marginally better able to activate CheA than CheV1 (P value= 0.0186). Overall, these data suggest that both CheV1 and CheW can connect chemoreceptors to CheA and lead to its activation, with CheW having a greater activity in this respect.

We then determined how a combination of both CheV1 and CheW coupling proteins would affect CheA phosphorylation. Addition of both CheV1 and CheW resulted in CheA activation that was in between that of CheV1 and CheW (Fig. 3). This outcome was true whether there was a chemoreceptor or not. This finding suggests that both proteins act independently on CheA but do not appear to synergize.

CheV1 and CheW Independently Promote CheA–Chemoreceptor Interactions in Vivo.

Our results above showed that CheV1 and CheW have overlapping interaction networks and both activate CheA, although to somewhat different levels. Given their similarities, one would predict that either could function in chemotaxis, but this is not the case (9, 20). We therefore hypothesized that CheV1 and CheW have in vivo activities in addition to protein–protein interactions and kinase activation. We thus sought to gain more insight into these putative in vivo roles. Our first step was to analyze CheW and CheV1’s roles in promoting the formation of the chemoreceptor–CheA complex. Both transmembrane and cytoplasmic chemoreceptors are associated with the membrane and retain CheA, which is normally cytoplasmic, at the membrane via coupling protein interactions (28–31). We therefore measured the amount of CheA associated with the membrane as an indicator of overall complex formation. We isolated membrane and cytoplasmic fractions from multiple strains using high-speed centrifugation and membrane washing as previously described (28, 30). Equal amounts of total protein from each fraction were separated by SDS/PAGE, followed by Western blotting (Fig. S2A). Control blots confirmed that the membrane and cytoplasmic fractions were substantially free of cross-contamination (Fig. S2B) (30). We then quantified and compared the amount of CheA found in the membrane relative to the cytoplasm (Fig. 4A). In wild type, the membrane had twice as much CheA as the cytoplasm. In a mutant lacking all chemoreceptors, CheA was almost fully cytoplasmic, consistent with the idea that CheA membrane interactions occur via chemoreceptors. CheA at the membrane was partially but significantly decreased in strains lacking either cheV1 or cheW, compared with wild type, and more substantially decreased when both proteins were lacking (Fig. 4B). These findings suggest that CheV1 and CheW each promote CheA interactions with the chemoreceptor complex in vivo. Furthermore, our data show they function in an additive way, suggesting each acts independent of the other.

Fig. S2.

Cellular membrane fractionation. (A) Coomassie Blue-stained SDS/PAGE gel with equal amounts of cytoplasmic “C” and membrane “M” cellular fractions (25 µg protein per lane) of various H. pylori strains. (B) Western blot controls from cytoplasmic “C” and membrane “M” fractions of various strains shown from the SDS/PAGE gel. (Top) “Membrane proteins” was probed with anti-TlpA22, which detects the membrane chemoreceptors TlpA and TlpB. (Bottom) “Cytoplasmic protein” was probed with anti-CheV1, which detects a cytoplasmic protein and has been previously used as a cytoplasmic control (30). Δtlp, strain lacking all chemoreceptors; ΔA, ΔcheA; ΔV1, ΔcheV1; ΔW, ΔcheW; ΔV1W, ΔcheV1 ΔcheW.

Fig. 4.

CheV1 and CheW are necessary to retain CheA at the cell membrane. (A) Western blots from cytoplasmic “C” and membrane “M” fractions of indicated strains probed with anti-CheA. (B) Quantification of the amount of CheA found in membrane fractions relative to CheA amount found in cytoplasmic fractions. Each image is representative of three independent cultures (n = 3). *P < 0.01 determined by an unpaired Student’s t test for strains compared with WT. Error bars represent the SD. Δtlp, strain lacking all chemoreceptors; ΔA, ΔcheA; ΔV1, ΔcheV1; ΔW, ΔcheW; ΔV1W, ΔcheV1 ΔcheW.

CheV1 and CheW Alter Each Other’s Interaction with CheA and Chemoreceptors in Vivo.

The membrane fractionation results suggested that CheV1 and CheW each promote CheA–chemoreceptor interactions. We next confirmed these interactions and examined what proteins were critical for forming them. To this end CheA, CheV1, or CheW were immunoprecipitated from whole cell lysates and examined for the presence of the other proteins.

In wild-type cells, all three proteins interacted with each other, with each being able to immunoprecipitate the other two, although the CheA bands were quite weak in the cheV1 and cheW mutants (Fig. 5 B–D). Similar results were obtained in immunoprecipitation experiments with mutants lacking all chemoreceptors (Fig. 5B). This outcome suggested that CheA forms complexes with each coupling protein independent of chemoreceptors.

Fig. 5.

CheV1 and CheW interact between themselves and with CheA in whole cell lysates. (A) CheA, CheV1, and CheW were detected by immunoblotting (IB) from whole-cell protein extracts of wild-type, ΔcheA, ∆tlpABCD, ΔcheV1, and ΔcheW H. pylori strains. Identical amounts of whole cell extracts were immunoprecipitated (IP) using anti-CheV1, anti-CheW, and anti-CheA antibodies and probed by IB using the respective antibody (B–D). Each gel is representative of three immunoprecipitations (n = 3).

We also found evidence that CheV1 and CheW interacted with each other in wild-type H. pylori and in mutants lacking CheA or all chemoreceptors. These experiments were somewhat challenging, as the expression of CheW in the ∆cheA background was substantially lessened (Fig. 5 C and D). These experiments are consistent with the idea that CheA, CheW, and CheV1 all form independent interactions.

We next examined whether the interactions of CheV1 or CheW with CheA were dependent on the other coupling protein. A strain lacking CheW resulted in less CheV1 immunoprecipitated with CheA compared with the wild-type strain (Fig. 5B). A strain lacking CheV1 also showed reduced levels of CheW pulled down from CheA immunoprecipitation, compared with wild type (Fig. 5B). Deletion of either cheV1 or cheW did not affect the expression of the other (Fig. 5A), consistent with the fact that both are in separate operons and under distinct transcriptional control (7). Taken together, these results and the CheA membrane association experiments suggest that CheV1, CheW, and CheA all interact in vivo in a chemoreceptor-independent manner. Furthermore, CheV1 and CheW interact with each other and enhance the interaction of the other with CheA.

Given the similarities in the interactions between CheV1 and CheW, we revisited whether addition of each protein would affect the interaction of the other with purified CheA protein in this same assy. We mixed CheA or chemoreceptor with a combination of CheW and CheV1. These samples were then immunoprecipitated with anti-CheA or antichemoreceptor antibodies and then compared using Western blotting to samples that had only CheW or CheV1. Using this approach, we found that neither CheW nor CheV1 changed the amount of the other pulled down in vitro (Fig. 2 A and B). These results suggest that CheW and CheV1 do not substantially affect the binding of the other and therefore may have distinct and independent binding interactions. Ultimately the combination of our in vivo and in vitro interaction experiments suggests that both coupling proteins work to enhance chemoreceptor signaling complex interaction.

Loss of CheV1, CheW, or Both Abrogates Chemoreceptor–CheA Complex Formation at Cell Poles.

The above experiments showed that CheW and CheV1 interact directly and with CheA. CheW is known to interact with another CheW-like domain in the form of CheA’s P5 domain (3). These interactions build the multiprotein chemoreceptor–CheA arrays at the cell pole and promote signal amplification (32, 33). Given that CheV1 has a CheW domain, we thus explored the role of CheW and CheV1 in the building of the H. pylori chemoreceptor array. We used immunofluorescence on whole cells, with all proteins expressed from the native loci in native forms. The polar chemotaxis complex was detected using anti-CheA, antichemoreceptors (Tlps), or anti-CheW. All antibodies detected a discrete locus at one or both cell poles in the wild-type strain, as reported previously (Fig. 6A) (30, 31, 34). Mutants lacking the chemoreceptors lost polar localization of the other complex members, consistent with the idea that these proteins are critical to build the chemotaxis arrays (Fig. 6C), as reported previously (30, 34). Published data have previously shown that CheV1 localizes to cell poles in wild type and loses this localization in mutants lacking chemoreceptors (30, 34), identical to the behavior of CheW shown here. cheA, cheV1, or cheW mutants also were unable to build a polar signaling complex (Fig. 6 D–F). In other words, none of these proteins was redundant with any other. These results suggest that polar chemosensory array creation requires two coupling proteins, CheV1 and CheW, in addition to CheA.

Fig. 6.

cheV1 and cheW mutants are defective in localizing CheA and chemoreceptors to polar chemosensory clusters. CheA, the chemoreceptors (Tlp), and CheW were visualized in H. pylori G27 and its isogenic mutants using immunofluorescence (A–F). All proteins were expressed from their native loci and are marked with green, whereas H. pylori cells were visualized with anti-H. pylori antibodies and red secondary antibodies. Dashed lines indicate bacteria from different fields of view. (Scale bar, 4 µm.)

Discussion

Coupling proteins are key components of many types of signal transduction systems, providing functions beyond simply holding proteins together. Here, we analyzed the roles of two different coupling proteins, CheW and CheV, in a single chemotaxis system. Altogether, our results suggest that some microbes use two coupling proteins to build a robust chemosensory complex.

The results presented here show that CheW and CheV1 are both required to create the functional polar chemosensory array. Mutants that lack either CheW, CheV1, or both have chemotaxis proteins that appear in punctate, nonpolar structures (Fig. 6) and have a severe decrease in retaining CheA at the membrane (Fig. 4). Because both cheW and cheV1 mutants are nonchemotactic, we surmise that these nonpolar chemotaxis units are not functional. This finding sheds light on why cheV1 and cheW mutants are nonchemotactic, whereas both proteins are capable of typical coupling protein functions as we show here. Large chemosensory arrays are critical for chemotaxis because they allow the high positive cooperativity (Hill coefficients of 15–20), a property that underlies the high sensitivity and ability to amplify the chemotaxis system’s response to ligands (3, 8, 35). Indeed, isolated core chemoreceptor–CheW–CheA signaling complexes are able to regulate CheA but with little positive cooperativity (36). Our studies suggest that cheW and cheV1 mutant arrays are defective because they lack the positive cooperativity necessary for wild-type chemotaxis.

We envision several mechanisms that might underlie the ability of CheV1 or CheW to form functional polar chemosensory arrays. Cryo-electron tomography revealed the existence of CheW rings symbolic of CheW–CheW interactions in E. coli chemoreceptor array structures (37). We suggest a similar chemotaxis array in H. pylori in which CheW–CheV1 rings form the core of the complex. Recent work has shown that in E. coli, which does not contain CheV, CheW interacts with other CheW and CheA molecules to form large chemosensory arrays that are critical for cooperativity of signal response (3, 8). Piñas et al. defined two CheW–CheA interaction faces: interface 1, which promotes CheW–CheA interactions required for kinase control, and interface 2, which creates the interactions that connect chemosensory arrays (3). Our work suggests that some microbes have altered CheW such that the array-forming function is split between two coupling proteins. Indeed, the residues that form interface 2 are conserved among enteric CheW but not among CheW as a whole (7, 17). Other support comes from models that suggested complexes lacking CheA and having only CheW—so-called CheW-only linkers—create high cooperativity in signal sensing (38). In Borrelia burgdorferi, two distinct CheW proteins are necessary for chemotaxis and formation of the chemosensory arrays (39). These two CheW, like CheV1 and CheW in H. pylori, are under distinct transcriptional control (7, 39–41). We speculate that bacteria that use two coupling proteins may thus modulate the expression of each to alter the cooperativity behavior of the chemoreceptor–CheA cluster array.

We report here that CheV1 and CheW possess largely overlapping but not identical interactions. Although their interaction strengths with various chemoreceptors may vary, as suggested for other systems (17), we envision that both proteins still interact with all chemoreceptors due to conserved residues in their signaling domains. TlpA and TlpB are integral membrane proteins in H. pylori, and due to the challenging nature of their purification, we were only able to test their interaction with CheV1 and CheW using the BACTH assay. However, it is possible that multiple coupling proteins in a single organism do provide some chemoreceptor specificity, as seen in Campylobacter jejuni, where the aspartate chemoreceptor prefers CheV over CheW (16, 42). Furthermore, our localization results suggest that both proteins are important to build large sensory arrays, but we do not yet know the structural and protein compositional difference between the small patches and large polar superarrays.

Coupling proteins play an important role in CheA kinase activity. Our results revealed that both CheV1 and CheW activate CheA and couple it to a chemoreceptor. CheW, however, had a greater ability to activate CheA than CheV1. This observation is consistent with the data presented by Ortega and Zhulin that CheV proteins limit CheA kinase output and thus can function with chemoreceptors that have a high intrinsic ability to activate CheA (17). These authors proposed that this limitation is due to CheV acting as a phosphate sink, a protein that can be phosphorylated by CheA to direct phosphates away from other targets (17). We did not detect phosphorylated CheV1 in our CheA kinase assay, a finding also reported by others, and so do not have direct support for this aspect of the model (22, 23). Another possibility, consistent with our data, is that CheV1 itself is less able to activate CheA via aspects of its structure. Our results may reflect the nature of the chemoreceptor–CheA complex as demonstrated in B. subtilis, where CheV increased CheA kinase activity depending on the methylation nature of chemoreceptors, which is reflective of ligand binding (43). CheV was also able to decrease CheA activity in B. subtilis when its coupling domain was mutated, leaving only its receiver domain active; this suggests an important interplay between CheV’s domains influencing its ultimate function (14). The idea that coupling proteins have different abilities to enhance or inhibit signal transduction is also seen in eukaryotic scaffold systems (44). We suggest that although both CheV1 and CheW appear to have similar functions, CheW may play a more critical role for CheA kinase activation than CheV1. Future research on overexpressing CheW in a cheV1 mutant may provide valuable knowledge on whether a greater amount of either CheW or CheV1 can make up for the loss of the other. These experiments are challenging, however, due to limited H. pylori genetic tools and the fact that overexpression of CheW, at least in E. coli, inhibits chemotaxis apparently by binding to chemoreceptors and CheA individually (45, 46).

CheV1 may function to connect the chemoreceptor complex to additional proteins and may be able to interact with more proteins than CheW, including CheV2 and CheV3. It is also possible that CheV1 or CheW have additional interactions beyond chemotaxis proteins, similar to the ParP protein from Vibrio parahaemolyticus (47). ParP, like CheV1, is a hybrid protein with a CheW-like domain. Like CheV1, ParP mutants have decreased soft agar migration and are more smooth swimming than wild type (9, 19, 40). ParP interacts with CheA and a membrane protein called ParC, and stabilizes chemotaxis complexes at the cell pole by preventing CheA dissociation (47). A common theme is the use of the CheW domain as a connection point for protein–protein interactions either between arrays or with other proteins.

In sum, our work suggests that division between two coupling proteins in a signaling complex may provide several advantages to bacteria similar to the advantages conferred by scaffolding proteins in mammalian cell signaling. The combination of two coupling proteins results in proper formation of the chemoreceptor–CheA chemotaxis complex and large polar arrays, localization to the cell membrane of the complex components, and finally stimulation of CheA kinase activity leading to optimal chemotaxis in H. pylori. Our results thus highlight the important functions of multiple coupling proteins in signal transduction systems in helping organisms efficiently respond to dynamic environments by rewiring key interactions among signal transduction proteins.

Materials and Methods

Detailed materials and methods are provided in SI Materials and Methods, as are all H. pylori and E. coli strains (Table S2) and primers (Table S3). H. pylori strains were grown under microaerobic conditions on either Columbia horse blood agar or Brucella broth with FBS. All antibodies were created to purified H. pylori proteins. Immunoprecipitations from whole cell extract and immunofluorescence were analyzed on H. pylori with natively expressed proteins.

Table S2.

Bacterial strains and plasmids

Strain/plasmid	Purpose	Description/genotype	Origin/reference
H. pylori mG27	BACTH	PCR template;	(52)
		Mouse adapted G27
H. pylori G27	Soft-Agar, Co-IP		(53)/N. Salama, Fred Hutchison Cancer Research Center, Seattle
H. pylori G27 ΔcheV1	Soft-Agar, Co-IP	ΔcheV1::cat	(23)
H. pylori G27 ΔcheW	Soft-Agar, Co-IP	ΔcheW::aphA3	(54)
H. pylori G27 ΔcheV1cheW	Soft-Agar, Co-IP	ΔcheW::aphA3	This work
		ΔcheV1::cat
H. pylori G27 ΔcheV2	Soft-Agar	ΔcheV2::cat	(23)
H. pylori G27 ΔcheV3	Soft-Agar	ΔcheV3::cat	(23)
H. pylori G27 ΔcheA	Soft-Agar, Co-IP	ΔcheA::cat	(23)
H. pylori mG27 ΔtlpABCD	Soft-Agar, Co-IP	ΔtlpC::aphA3	(23)
E. coli BL21 CheAY	Protein purification	pTrc-cheAY	(22)/D. Beier, University of Würzburg, Wuerzburg, Germany
E. coli BL21 TlpD	Protein purification	pGEX-6P-2-tlpD	(49)
E. coli BL21 CheW	Protein purification	pGEX-6P-2-cheW	(30)
E. coli BL21 CheV1	Protein purification	pGEX-6P-2-cheV1	(23)
E. coli DH5α	BACTH	Cloning strain; endA1 hsdR17 (rK-mK+) supE44 thi-1 recA1 gyrA (NalR) relA1 (lacIZYA-argF)U169 deoR (80dlac(lacZ)M15)	Invitrogen
E. coli BTH101	BACTH	Reporter strain for BACTH; F-, gal E15, gal K16 mcrA1, mcrB1, ara D139, rpsL1(Strr), hsdR2, cya-99 (adenylate cyclase deficient) REC+	(21)
pKT25 (T25-X)	BACTH	BACTH plasmid encoding the T25 fragment allowing for in-frame protein fusion at the C-terminal end of T25; Kmr	(21)/J. Gober, University of California, Los Angeles
pUT18 (Y-T18)	BACTH	BACTH plasmid encoding the T18 fragment allowing for in-frame protein fusion at the N-terminal end of T18; Apr	(21)/J. Gober, University of California, Los Angeles
pKT25-zip (T25-zip)	BACTH	BACTH control plasmid with the yeast GCN4 leucine zipper fused to the T25 fragment at the C-terminal end; Kmr	(21)/J. Gober, University of California, Los Angeles
pUT18C-zip (T18-zip)	BACTH	BACTH control plasmid with the yeast GCN4 leucine zipper fused to the T18 fragment at the C-terminal end; Apr	(21)/J. Gober, University of California, Los Angeles, CA
pKT25-cheV1	BACTH	Fusion, Kmr	This work
cheV1-pKNT25	BACTH	Fusion, Kmr	This work
pKT25-cheW	BACTH	Fusion, Kmr	This work
cheW-pKNT25	BACTH	Fusion, Kmr	This work
cheV1-pUT18	BACTH	Fusion, Apr	This work
pUT18C-cheV1	BACTH	Fusion, Apr	This work
cheV2-pUT18	BACTH	Fusion, Apr	This work
pUT18C-cheV2	BACTH	Fusion, Apr	This work
cheV3-pUT18	BACTH	Fusion, Apr	This work
pUT18C-cheV3	BACTH	Fusion, Apr	This work
cheW-pUT18	BACTH	Fusion, Apr	This work
pUT18C-cheW	BACTH	Fusion, Apr	This work
cheA-pUT18	BACTH	Fusion, Apr	This work
pUT18C-cheA	BACTH	Fusion, Apr	This work
tlpA-pUT18aa330-675	BACTH	Fusion, Apr	This work
pUT18C-tlpAaa330-675	BACTH	Fusion, Apr	This work
tlpB-pUT18aa240-565	BACTH	Fusion, Apr	This work
pUT18C-tlpBaa240-565	BACTH	Fusion, Apr	This work
tlpD-pUT18	BACTH	Fusion, Apr	This work
pUT18C-tlpD	BACTH	Fusion, Apr	This work

Ap^r, ampicillin resistant; Co-IP, coimmunoprecipitation; Km^r, kanamycin resistant.

Table S3.

Primers used in this study

Primer	Oligonucleotide sequence	Restriction site
cheV1 pKT25 F	atatatCTGCAGaaATGGCTGATAGTTTAGCGGG	PstI
cheV1 pKT25 R	atatatGGATCCtcTGCTAATTCCAAAAATTGCTTAAC	BamHI
cheV1 pUT18 F	atatatCTGCAGaATGGCTGATAGTTTAGCGGG	PstI
cheV1 pUT18 R	atatatGGATCCtcTGCTAATTCCAAAAATTGCTTAAC	BamHI
cheV2 pUT18 F	atatatCTGCAGaGTGGTAAGAGATATTGACAAAACGA	PstI
cheV2 pUT18 R	atatatGGATCCtcTGAAAGCGTTTTTTTAAGCATTTCA	BamHI
cheV3 pUT18 F	atatatCTGCAGaATGGCAGAAAAAACAGCTAACG	PstI
cheV3 pUT18 R	atatatGGATCCtcCGCATTCTTGTCTAAAATCTTAGAAATT	BamHI
cheW pUT18 F	atatatCTGCAGaGTGAGCAATCAATTAAAAGATTTATTTGAAA	PstI
cheW pUT18 R	atatatGGATCCtcGAAGTCTTTTTTTAAGATTTCTTCCACTC	BamHI
cheA pUT18 F	atatatGTCGACaATGGATGATTTGCAAGAAATAATG	SalI
cheA pUT18 R	atatatGGATCCtcCGATTCGCCTCCTTCTAATTT	BamHI
tlpApUT18 F (cytoplasmic signaling domain: nt 988–2028, aa 330–675)	atatatCTGCAGaCGTTTGGAAGTCGTTTCTAG	PstI
tlpApUT18 R (cytoplasmic signaling domain: nt 988–2028, aa 330–675))	atatatTCTAGAgcAAACTGCTTTTTATTCACAT	XbaI
tlpB pUT18 F (cytoplasmic signaling domain: nt 718–1695, aa 240–565)	atatatCTGCAGaGATGAACTGGTCCTTAAAAT	PstI
tlpB pUT18 R (cytoplasmic signaling domain: nt 718–1695, aa 240–565)	atatatGGATCCtcAGTTTTAAACAAATTCACTT	BamHI
tlpD pUT18 F	atatatCTGCAGaATGTTTGGGAATAAGCAGTT	PstI
tlpD pUT18 R	atatatGGATCCtcTTCGCCTTTTTGAATTTTTTCA	BamHI

Restriction sites are underlined. F, forward primer; R, reverse primer.

SI Materials and Methods

Bacterial Strains and Growth Conditions.

All H. pylori and E. coli strains used in this study are listed in Table S1. H. pylori was grown on Columbia horse blood agar [5% (vol/vol) defibrinated horse blood, 50 µg/mL cycloheximide, 10 µg/mL vancomycin, 5 µg/mL 130 cefsulodin, 2.5 units per mL polymyxin B, and 0.2% (wt/vol) β-cyclodextrin] at 37 °C in microaerobic conditions [5–10% (vol/vol) O₂, 10% (vol/vol) CO₂, and 80–85% (vol/vol) N₂]. For liquid cultures, H. pylori was grown in Brucella broth with 10% (vol/vol) heat-inactivated FBS (BB10) under the same conditions stated above. Antibiotic concentrations used for mutant selection were 15 µg/mL kanamycin or 13 µg/mL chloramphenicol. E. coli was grown on LB media with 100 µg/mL ampicillin or 60 µg/mL kanamycin.

Construction of Bacterial Two-Hybrid Plasmids and Bacterial Two-Hybrid Analysis.

Genomic DNA from H. pylori strain mG27 was extracted using the Wizard genomic DNA kit (Promega). The full-length tlpD, cheW, cheV1, cheV2, cheV3, and cheA genes were amplified from this genomic DNA using the primers listed in Table S2 using Phusion DNA polymerase with the following thermocycler conditions: 98 °C for 5 min, 98 °C for 10 s, 55 °C for 30 s, 72 °C for 2.5 min, repeat 39 times, and 72 °C for 10 min. Cytoplasmic signaling domains of tlpA (nucleotide 988–2028, amino acid 330–675) and tlpB (nucleotide 718–1695, amino acid 240–565) genes were amplified as above. PCR products of all genes were digested with the appropriate restriction enzyme (Table S2). BACTH plasmids pKT25, pKNT25, pUT18, and pUT18C were also digested using the appropriate restriction enzymes. Digested plasmids were phosphatase treated using TSAP phosphatase following the manufacturer’s suggested protocol (Promega). PCR products and plasmids were gel purified using the illustra GFX PCR DNA and gel band purification kit (GE Healthcare). The purified digested PCR gene products and plasmids were ligated and used to transform E. coli DH5a to appropriate antibiotic resistance. Colony PCR screening was used to determine which colonies contained correct plasmids. All cloned plasmids were verified by sequencing and contained no errors.

Recombinant BACTH plasmids at 40 ng each were cotransformed into competent E. coli BTH101 cells (Table S2). For positive controls, pKT25-zip and pUT18C-zip were cotransformed in BTH101. These plasmids express leucine zipper motifs that result in a strong dimer interaction (21). For negative controls, empty BACTH plasmids were cotransformed with each other and with each recombinant plasmid. Transformed cells were plated on LB plates containing 100 µg/mL ampicillin, 30 µg/mL kanamycin, 40 µg/mL X-gal, and 0.5 mM IPTG and incubated at 30 °C for 40–48 h using the positive and negative controls as reference for recording colony colors.

β-Galactosidase Assay.

The efficiency of BACTH interactions between plasmids was quantified using a β-galactosidase assay as described previously (21). Transformants were grown in LB broth with 0.5 mM IPTG, 100 µg/mL ampicillin, and 30 µg/mL at 30 °C overnight. Liquid cultures were then diluted 1:100 in LB broth and grown again until OD₆₀₀ was 0.3–0.7. The OD₆₀₀ was recorded, and 100 µL of each culture was mixed with 0.9 mL of cold Z buffer (0.06 M Na₂HPO₄0.7H₂O, 0.04 M NaH₂PO₄.H₂O, 0.01 M KCl, 0.001 M MgSO₄0.7H₂O, 0.056 M 2-mercaptoethanol, pH 7.0). We added 100 µL of chloroform followed by 50 µL of 0.1% SDS, followed by vortexing to permeabilize the cells. The samples were placed at 30 °C for 5 min with the tops off. We added 0.2 mL of ONPG substrate solution [4 mg/mL ortho-Nitrophenyl-β-galactoside (ONPG) in 0.1 M phosphate buffer, pH 7.0] to the sample and observed the color until the sample turned yellow. At this point, the reaction was stopped by adding 0.5 mL 1 M Na₂CO₃ and the time recorded. The OD₄₂₀ of the samples was measured. Miller units (β-galactosidase activity) were calculated using the following formula: 1,000 × ([OD₄₂₀ / (time [min] × volume of culture used × OD₆₀₀)] (48).

Protein Purification and Whole Cell Extract Preparation.

TlpD, CheV1, CheW, and CheA were purified as previously described (23, 30, 49). In brief, all fusion plasmids were expressed in E. coli BL21 at 37 °C and induced with 0.5 mM IPTG. The glutathione S-transferase (GST) fusion expression plasmids (TlpD, CheV1, and CheW) were applied to a GST Prep column for purification (GE Healthcare). GST tags were cleaved using Precission protease. His-tagged CheA was applied to a His Prep column (GE Healthcare). All purified proteins were dialyzed in storage buffer [50 mM Hepes, pH 7.6, 50 mM KCl, 20% (vol/vol) glycerol] and stored in –80 °C. Protein concentrations were determined by measuring absorption at 280 nm on a nanodrop and using the coextinction coefficient of each protein based on its amino acid composition.

H. pylori whole cell extracts for use in coimmunoprecipitation were prepared by resuspending H. pylori cell pellets grown from liquid cultures in BB10 to an OD₆₀₀ of 1 in B-PER reagent (ThermoFisher Scientific) and half a tablet of protease inhibitor (Roche) or 0.5 mM phenylmethylsulfonyl fluoride (PMSF) (Gold Biotechnology).

Coimmunoprecipitation.

Coimmunoprecepitations were performed using Dynabeads Protein A for Immunoprecipitation following the manufacturer’s suggested protocol (ThermoFisher Scientific). In brief, antibodies [10 µL of anti-CheA (34), 40 µL anti-CheV1 (34), 10 µL anti-TlpD (50), or 40 µL anti-CheW (30)] diluted in a total of 200 µL washing buffer (PBS, 0.02% tween, pH 7.4) were first bound to Dynabeads by incubating with rotation at room temperature for 25 min followed by removal of supernatant from the beads using a magnetic rack. Dynabeads were cross-linked to antibodies using bis(sulfosuccinimidyl)suberate (BS3) following the manufacturer’s instructions (ThermoFisher Scientific). Either purified protein mixtures or whole-cell lysates were used in these experiments as the antigen and mixed with the antibody beads for 30–60 min at room temperature in PBS containing 0.5 mM of PMSF. For purified protein, mixtures containing 100 µM of each protein were preincubated together before adding beads for 30–60 min at room temperature in PBS containing 0.5 mM of PMSF. For whole cell lysate, 1,000 µL at OD₆₀₀ 1.6–1.8 was added to the Dynabeads–antibody complex and gently resuspended by pipetting. The protein–antibody complexes were washed three times using washing buffer from above and then eluted with 50 mM glycine, pH 2.8. Samples were mixed with Laemmli sample buffer [60 mM Tris·HCl, pH 6.8, 2% (wt/vol) SDS, 5% (vol/vol) glycerol, 1% β-mercaptoethanol, 0.02% bromophenol blue] and stored at –20 °C. Before running on an SDS/PAGE gel, samples were heated for 10 min at 95 °C.

Immunoblotting.

Samples were separated on 12% (wt/vol) acrylamide SDS/PAGE gels. SDS/PAGE gels were then either stained using Coomassie Brilliant Blue R-250 Dye (ThermoFisher Scientific) or used for immunoblotting. For immunoblotting, gels were soaked in transfer buffer [48 mM Tris-base, 39 mM glycine, 1.3 mM SDS, 20% (vol/vol) methanol] for 25 min and then transferred to an immunoblot polyvinylidene diflouride (PDVF) membrane (Biorad) by semidry transfer for 45 min at 12 V. The membrane was blocked for 1–2 h with blocking buffer (PBS with 1% milk plus 0.2% Tween-20) at room temperature. Primary antibody was added and incubated for 16 h at 4 °C using the following antibody dilutions: 1:1,500 for anti-TlpA22, 1:60 for CheV1, 1:150 for CheW, and 1:1,000 for CheA. After incubation, the membranes were washed and HRP-conjugated secondary antibodies (Santa Cruz Biotechnology) added at 1:1,500. For anti-TlpA22, anti-CheV1, and anti-CheA, goat anti-rabbit IgG was used, and for anti-CheW, goat anti-guinea pig was used. After incubation, the membranes were washed and treated with luminol, p-coumaric acid, and hydrogen peroxide. Blots were then visualized using a Chemidoc imaging system (Biorad).

Cellular Fractionation.

Bacteria were grown as above in BB10 and collected at an OD of 1–1.5 by centrifugation and stored at –20 °C. Bacterial pellets were resuspended in lysis buffer [50 mM Tris⋅HCl, pH 7, 10% (vol/vol) glycerol, 1 mM 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride (AEBSF), and 10 mM DTT]. Cells were further lysed by sonication in 30-s bursts for 3–5 min while kept on ice. Unlysed cells were removed by centrifugation at 4,000 g, 4 °C, 15 min. The supernatant was then pelleted using an ultracentrifuge at 240,000 g for 30 min at 4 °C (Beckman TLA 100.3). The supernatant was collected and considered to be the “cytoplasmic” fraction. The remaining pellet was rinsed three times with high salt buffer (2 M KCl in lysis buffer). The pellet was resuspended in high salt buffer using light sonication (20 Amp in 30-s bursts). The resuspended pellet was spun again at 240,000 g for 20 min at 4 °C. The pellet was then washed and resuspended in high salt buffer as above, followed by centrifugation at 240,000 g as above for 20 min at 4 °C. The pellet was then washed three times with lysis buffer and resuspended in lysis buffer by light sonication. The final resuspended pellet was spun for 240,000 g for 20 min. The final pellet was resuspended in a small volume of lysis buffer. Total protein concentration from cytoplasmic and membrane fractions were quantified by measuring absorption at 280 nm and analyzed using Coomassie Blue staining of SDS/PAGE gels as described above.

Phosphorylation Assay.

In vitro phosphorylation assays were performed using purified proteins as described previously (23). In brief, 2 µM each of CheA, CheV1, and CheW were incubated in a final volume of 30 µL reaction buffer (20 mM MgCl₂, 50 mM KCl, 50 mM Tris⋅HCl, pH 7.5) for 30–45 min at room temperature. A radioactive ATP mixture [11 µM [γ-³²P]-ATP (Perkin-Elmer LAS) and 2 mM unlabeled ATP] was used to start the reactions, by adding to a final concentration of 0.2 mM. The reactions were stopped at the indicated times by mixing with 2× Laemmli sample buffer. CheA collected right after adding radioactive ATP (time 0) was used as a reference control for all reactions. The reactions were electrophoresed on a 12% SDS/PAGE gel (wt/vol). Gels were dried and exposed to a phosphoimager cassette (Biorad) overnight. Autoradiography images were obtained by scanning the cassette on a Personal Molecular Imager (Biorad). The intensity of the CheA bands was determined using ImageJ software and plotted. The area under the curve was determined by calculating definite integrals—area between each point = (x₂ – x₁) [f(x₂) + f(x₁)/2].

Image Analysis.

Western blots and phosphorylation assay images were analyzed using Image Lab Software (Biorad) or ImageJ Software (NIH) (51).

Immunofluorescence.

Bacterial cultures were grown to an OD₆₀₀ of 0.5 in BB10 and concentrated using centrifugation to an OD₆₀₀ of 1. Cells were prepared and stained as described in Lertsethtakarn et al. (34). Briefly, 50 µL of each culture was placed on poly-l-lysine–coated slides and fixed with PLP [75 mM NaPO₄, pH 7.4, 2.5 mM NaCl, 2% (wt/vol) paraformaldehyde]. Cells were then washed, blocked [3% (vol/vol) BSA, 0.1% Triton X-100 in PBS], and then primary antibodies added in 3% (vol/vol) BSA, 1% saponin, 0.1% triton X-100, and 0.02% Na azide in PBS. All protein antibodies were preabsorbed as previously described and used at the following dilutions: 1:200 anti-CheA (34), 1:200 anti-Tlps (50),1:50 anti-CheW (30), or 1:500 chicken anti-H. pylori (AgriSera AB). After incubation, the cells were washed, and fluorescent secondary antibodies that recognize each of the primary protein and H. pylori antibodies were incubated with the cells. Goat anti-rabbit Alexa Fluor 594 (1:300 dilution), goat anti-guinea pig Alexa Fluor 594 (1:300 dilution), or goat anti-chicken 238 Alexa fluor 488 (1:500 dilution) (Abcam) were used as secondary antibodies. After incubation, the cells were washed four times with blocking buffer and aspirated off. Finally, a drop of mounting media was placed on the bacteria before covering with a coverslip to prevent photobleaching (Vectashield). The cells were then visualized using a Nikon Eclipse E600 fluorescent microscope; the Texas Red filter was used for the Alexa Fluor 594 channel and the GFP filter for the Alexa Fluor 488 channel. Images were merged and analyzed using Adobe Photoshop C2.