Research ArticleStructural Biology

Specificity of the CheR2 Methyltransferase in Pseudomonas aeruginosa Is Directed by a C-Terminal Pentapeptide in the McpB Chemoreceptor

See allHide authors and affiliations

Science Signaling  08 Apr 2014:
Vol. 7, Issue 320, pp. ra34
DOI: 10.1126/scisignal.2004849


Methyltransferases of the CheR family and methylesterases of the CheB family control chemoreceptor methylation, and this dynamic posttranslational modification is necessary for proper chemotaxis of bacteria. Studies with enterobacteria that contain a single CheR or CheB show that, in addition to binding at the methylation site, some chemoreceptors bind CheR or CheB through additional high-affinity sites at distinct pentapeptide sequences in the chemoreceptors. We investigated the recognition of chemoreceptors by CheR proteins in the human pathogen Pseudomonas aeruginosa PAO1. Of the four methyltransferases in PAO1, we detected an interaction only between CheR2 and the chemoreceptor methyl-accepting chemotaxis protein B (McpB), which contains the pentapeptide GWEEF at its carboxyl terminus. Furthermore, CheR2 was also the only paralog that methylated McpB in vitro, and deletion of the pentapeptide sequence abolished both the CheR2-McpB interaction and the methylation of McpB. When clustered according to protein sequence, bacterial CheR proteins form two distinct families—those that bind pentapeptide-containing chemoreceptors and those that do not. These two families are distinguished by an insertion of three amino acids in the β-subdomain of CheR. Deletion of this insertion in CheR2 prevented its interaction with and methylation of McpB. Pentapeptide-containing chemoreceptors are common to many bacteria species; thus, these short, distinct motifs may enable the specific assembly of signaling complexes that mediate different responses.


Bacteria constantly sense and adapt to changing environmental conditions to assure survival. This important function is primarily mediated by one-component systems, two-component systems, and chemosensory pathways (13). Chemosensory pathways are involved in mediating flagellum- and type IV pili–mediated taxis and also carry out alternative cellular functions (3). The proteins of chemosensory pathways have been classified as auxiliary proteins and core proteins based on the frequency of their occurrence (3). Core proteins are the CheA sensor kinase, CheW coupling protein, CheY response regulator, CheR methyltransferase, CheB methylesterase, and chemoreceptors (3). Pathway function involves the concerted action of the excitatory pathway and adaptational mechanism(s). The canonical excitatory pathway is initiated by signal recognition at the chemoreceptor, which in turn modulates CheA autophosphorylation and, subsequently, the transphosphorylation of CheY. When phosphorylated, CheY undergoes a conformational change, triggering an alteration of its activity (4). A number of adaptation mechanisms have evolved to restore the prestimulus behavior in the presence of the signal (5, 6). The canonical adaptation mechanism consists of the methylation and demethylation of chemoreceptors catalyzed by the CheR methyltransferase and CheB methylesterase, respectively (4).

Much of what we know in regard to chemosensory pathways is the result of studies of flagellum-mediated taxis in Escherichia coli and Salmonella typhimurium [reviewed in (4)]. E. coli has five chemoreceptors that feed stimuli into a single chemosensory pathway. However, genome analyses have shown that other bacteria have an elevated number of chemoreceptors and multiple copies of chemosensory signaling proteins that form different chemosensory pathways (3, 7). For such bacteria, the human pathogen Pseudomonas aeruginosa has become a model organism (8). This species has five gene clusters that encode chemosensory signaling proteins that assemble into four chemosensory pathways, termed Che, Che2, Wsp, and Chp. Two of the pathways mediate chemotaxis; whereas the Che pathway is essential for chemotaxis (9), the role of the Che2 pathway in chemotaxis is less clear (10). The Wsp and Chp pathways modulate diverse cellular processes by altering the abundance of cyclic diguanosine monophosphate (c-di-GMP) and cyclic adenosine monophosphate (cAMP), respectively (11, 12). P. aeruginosa has 26 chemoreceptor genes, of which only 4 are found in these gene clusters; the remaining chemoreceptor genes are scattered throughout the genome (8). Two chemoreceptors, encoded by mcpA and mcpB, are located in the che2 gene cluster (Fig. 1A). McpB (methyl-accepting chemotaxis protein B) has a PAS (Per-Arnt-Sim)–type sensor domain, lacks transmembrane regions, and is predicted to be of cytosolic location (10). Attempts to identify the function of McpB have been inconclusive. Initially, a study indicated that McpB was involved in aerotaxis (13), but a subsequent study showed that this was not the case (10). It was demonstrated that CheB2 of this pathway was essential for P. aeruginosa infection in a murine lung infection model, from which the authors concluded that CheB2 and the Che2 pathway are involved in a specific chemotactic response triggered during infection by a yet unknown signal (14).

Fig. 1 Specific binding of the McpB-derived pentapeptide GWEEF to CheR2.

(A) C-terminal segment of an alignment of P. aeruginosa PAO1 chemoreceptors. Potential pentapeptide CheR binding sites are in red, and linker regions are in green. Proline residues are shaded in yellow. Receptors BdlA and PA1930 are not included in this alignment because their signaling domains are shorter than those of the receptors shown. (B) Organization of the che2 chemotaxis gene cluster of P. aeruginosa PAO1. (C) Microcalorimetric titration of CheR paralogs with the peptide GWEEF. Upper panel: Raw data for the titration of 10 to 30 μM protein with 0.5 to 1 mM pentapeptide. Lower panel: Integrated, dilution-corrected, and concentration-normalized data for the titration of CheR2 with the pentapeptide. The resulting binding parameters are given in Table 2. Shown is a representative of two experiments.

As mentioned, the CheR methyltransferases are among the core proteins of chemosensory pathways. CheR methylates glutamyl residues at the chemoreceptor signaling domain. The extent of methylation, in turn, modulates the capacity of the receptor to control CheA autophosphorylation (15). This mechanism adapts CheA autophosphorylation activity to a given signal concentration. CheR uses S-adenosylmethionine (SAM) as a substrate, and the methylation reaction gives rise to S-adenosylhomocysteine (SAH) (16). CheR and CheB bind to the methylation site of the chemoreceptor. However, the high-abundance chemoreceptors in E. coli and S. typhimurium have an additional binding site for CheR and CheB formed by a pentapeptide tethered through a flexible linker to the C-terminal end of the signaling domain (1721). Truncation of this pentapeptide abolishes receptor binding (17, 22), markedly decreases methylation and demethylation amounts (17, 22, 23), and impairs tactic responses (19), whereas its addition greatly improves methylation (24). In contrast, the low abundance receptors of E. coli do not have these pentapeptides. E. coli has a single CheR that acts on receptors with and without a pentapeptide. The physiological relevance of this pentapeptide binding site is not fully understood but may reside in an increase in the concentration of CheR at the methylation site, thereby increasing the effective enzyme concentration (25). P. aeruginosa, the model organism of this study, has four CheR methyltransferases (CheR1, CheR2, CheR3, and WspC) that are encoded by four different chemosensory gene clusters. Receptors with a C-terminal pentapeptide are found in many different bacteria (26), and the relative abundance of such chemoreceptors varies considerably: whereas some bacteria lack pentapeptide-containing receptors, in other species, such receptors account for about half of the total number of receptors (26).

In general, the interaction between two proteins is mediated by a single interface. In this context, the dual binding sites of CheR and CheB on pentapeptide-containing chemoreceptors are remarkable exceptions. This phenomenon raises the question of what forces prompted their evolution. The interaction of CheR with pentapeptides has been studied in species that contain a single CheR. However, many bacterial genomes contain multiple copies of CheR (7), raising the question of whether all or only a subset of the CheR paralogs interact with the pentapeptide. We have addressed these questions using P. aeruginosa as a model organism.


P. aeruginosa PAO1 has three chemoreceptors with a C-terminal extension

To identify chemoreceptors with a C-terminal extension that may harbor pentapeptides for CheR binding, we aligned the sequences of the cytosolic fragments of the 26 chemoreceptors from P. aeruginosa. The C-terminal section of this alignment (Fig. 1A) shows that there are two receptors, McpA and McpB, that have a C-terminal extension of 30 amino acids, which is comparable to the length of enterobacterial receptors (18). Both receptors are encoded by part of the che2 gene cluster (Fig. 1B). In addition, the receptor PA0411 contains a six–amino acid extension. We hypothesized that these three extensions could constitute CheR docking sites. Therefore, the terminal pentapeptides GWEEF, EVELF, and GVEQA were synthesized for binding studies.

Affinities of CheR paralogs for SAM and SAH differ largely

The four CheR methyltransferases of P. aeruginosa were produced as purified recombinant proteins, and their functionality was validated by microcalorimetric titrations with SAM and SAH (fig. S1 and Table 1). The binding of SAM to CheR1, CheR3, and WspC had dissociation constants (KDs) between 15 and 47 μM, values that are similar to that of S. typhimurium CheR (16, 27). In contrast, SAM bound with markedly weaker affinity to CheR2 (KD of almost 200 μM). SAH bound to all four paralogs more tightly than to SAM, with KD values that ranged from 0.1 to 3.5 μM (fig. S1 and Table 1). That SAH bound more tightly than SAM is also observed for enterobacterial enzymes (16, 28) and is indicative of product feedback inhibition. However, the ratios of the KD values for SAM and SAH (Table 1), which determine the magnitude of product feedback inhibition, vary greatly among the CheR paralogs: 13 for WspC, indicating modest feedback inhibition, and 361 for CheR2, suggesting strong feedback inhibition.

Table 1 Thermodynamic parameters derived from the microcalorimetric titration of the four CheR paralogs of P. aeruginosa PAO1 with SAM and SAH.

The corresponding data are shown in fig. S1. Data are of two different measurements with the errors of curve fitting indicated.

View this table:

Of the four CheR paralogs, only CheR2 binds pentapeptides

To establish whether any of the four CheR paralogs bound to the pentapeptides identified, we titrated the purified CheR proteins with the pentapeptides GWEEF (corresponding to McpB), EVELF (corresponding to McpA), and GVEQA (corresponding to PA0411). As a control, protein was titrated after every experiment with SAH to verify its capacity to recognize the methylation product (as marker for the native state of the protein). None of the four CheR enzymes bound either the EVELF or the GVEQA pentapeptides. However, whereas CheR1, CheR3, and WspC did not bind the GWEEF pentapeptide, CheR2 bound this pentapeptide with high affinity (Fig. 1C and Table 2). We then investigated whether any of the proteins bound the enterobacterial NWETF pentapeptide. Shiomi et al. (29) reported that the W (Trp) and F (Phe) of this peptide are essential for CheR binding, whereas mutation of the remaining three amino acids had only a slight effect. We found that CheR2, but none of the other paralogs, bound this noncognate pentapeptide (fig. S2). With a KD of about 1.4 μM (Table 2), the affinity of NWETF for CheR2 was greater than that for its cognate CheR in E. coli (reportedly 10 μM) (28).

Table 2 Thermodynamic parameters derived from the titration of native/mutant methyltransferases with pentapeptides, SAH, or with native/mutant McpB chemoreceptor.

The corresponding data are shown in Figs. 1 and 2 and figs. S2, S5, and S7. Data are of two different measurements with the errors of curve fitting indicated.

View this table:

CheR2 specifically targets the McpB chemoreceptor

The results, thus far, show that, in P. aeruginosa, there is a single chemoreceptor-pentapeptide interaction, namely, that between McpB and CheR2. The genes encoding the proteins are located next to each other in the genome, and both form part of the che2 chemotaxis cluster (Fig. 1B) (30). Because McpB lacks transmembrane regions, we produced the full-length recombinant protein and performed microcalorimetric binding studies with purified CheR2 (Fig. 2A). McpB and CheR2 had a high-affinity interaction (KD of about 0.15 μM), which was about threefold greater than that of the McpB-associated pentapeptide, GWEEF, for CheR2 (Table 2). To determine the contribution of the pentapeptide in CheR2 binding, a shortened construct of McpB that lacked the GWEEF peptide was titrated with CheR2 (Fig. 2B). Deletion of the peptide resulted in complete loss of binding; heat changes observed were similar to those of titrations of CheR2 with buffer.

Fig. 2 Binding studies of native and mutant McpB and CheR2.

(A to C) Upper panel: Microcalorimetric titrations of 4 to 6 μM wild-type or mutant McpB with 50 to 70 μM wild-type or mutant CheR2: (A) titration of McpB with CheR2, (B) titration of McpBΔGWEEF with CheR2, and (C) titration of McpB with CheR2ΔGPN. Lower panel: Integrated, dilution-corrected, and concentration-normalized peak areas of the titration of McpB with CheR2. The resulting binding parameters are given in Table 2. Shown are representatives of two experiments each.

No allostery exists between the SAM or SAH binding site and the pentapeptide binding site

Multiple binding sites on proteins often are indicative of allosteric effects. To explore this possibility for CheR2, a series of microcalorimetric titrations were conducted in which the protein was saturated with one ligand and then titrated with the remaining ligand. We saturated CheR2 with SAM or SAH and tested its titration with GWEEF, or we saturated CheR2 with GWEEF and tested its titration with SAM or SAH (fig. S3). The derived parameters (table S1) were then compared to those obtained from controls with ligand-free CheR2 (fig. S3). Our data showed that the addition of saturating concentrations of SAM or SAH did not alter binding of the GWEEF pentapeptide to CheR2 and vice versa, suggesting that there are no allosteric effects between the SAH/SAM and pentapeptide binding sites at CheR2.

CheR methyltransferases form two distinct protein families on the basis of pentapeptide binding

Only one of the four CheR paralogs in P. aeruginosa bound pentapeptides, raising the question whether structural or sequence features determine the binding of a CheR to pentapeptides. To address this question, we aligned the CheR sequences of proteins known to bind or not to bind pentapeptides (26) (fig. S4). The selection of sequences was based on the following rationale: The NWETF peptide binds to CheR of S. typhimurium (17, 20) and E. coli (18, 21). Erwinia carotovora, Burkholderia mallei, and Ralstonia solanacearum each has a single CheR and 19, 3, or 4 pentapeptide-containing chemoreceptors, respectively (26), indicating that the CheR binds pentapeptides. We found here that P. aeruginosa CheR2 binds the GWEEF pentapeptide, whereas the remaining three paralogs did not. CheR-mediated receptor methylation has also been observed in Bacillus subtilis (31) and Thermotoga maritima (26). However, both bacteria are devoid of chemoreceptors with pentapeptide extensions (26), which strongly suggests that the CheR does not bind pentapeptides. CheRs that bind pentapeptides and those that do not cluster into two distinct groups (Fig. 3A). We noticed that pentapeptide-binding CheR proteins have a three–amino acid insertion (GXX; GPN in P. aeruginosa) that is absent in CheR proteins that do not bind pentapeptides (Fig. 3B). The inspection of the CheR structure revealed that this insertion is part of a loop that links strands 2 and 3 of the CheR β-subdomain that contains the pentapeptide-binding site (20) (Fig. 3C). To determine whether this insertion is essential for pentapeptide binding, we generated a CheR2 mutant in which GPN (Fig. 3B) was deleted. The titration of CheR2ΔGPN with SAH (fig. S5) revealed that the CheR2 mutant had an affinity for SAH that was similar to that of the wild-type protein (Table 2), indicating that this mutation did not substantially alter protein structure. However, CheR2ΔGPN did not bind the GWEEF or NWETF peptides (fig. S5) or full-length McpB (Fig. 2C), indicating that this insertion is essential for pentapeptide recognition.

Fig. 3 Pentapeptide-dependent and pentapeptide-independent CheR methyltransferases form two families.

(A) Clustering of sequences that are pentapeptide-dependent or pentapeptide-independent. The corresponding accession codes are provided in the legend of fig. S4. The figure was produced using the server. (B) Segment of the sequence alignment comprising the β-subdomain of pentapeptide-dependent and pentapeptide-independent CheR. The full alignment is shown in fig. S4. Secondary structure elements are indicated: h, helix; s, strand. Residues in yellow and pink interact with the pentapeptide. The tripeptide that is absent in the pentapeptide-independent methyltransferases is shown in pink. (C) CheR structure of S. typhimurium (20) with bound pentapeptide shown in ball-and-stick mode. The structural elements shown in yellow and pink correspond to the sequence fragments of the same color in (B).

To assess the effect of the tripeptide on the CheR sequence clustering, the cluster analysis shown in Fig. 3A was repeated using the sequences of pentapeptide-binding CheR proteins from which the tripeptide was deleted. The resulting sequence clustering (fig. S6) was very similar to that obtained for the native sequences (Fig. 3A), suggesting that this tripeptide is only one of several sequence characteristics that define these two distinct protein families. To verify this conclusion, we constructed a CheR1 mutant into which the GPN sequence was inserted at the appropriate position (Fig. 3B). The resulting CheR1+GPN protein bound SAH tightly but did not acquire the capacity to bind pentapeptides (fig. S7 and Table 2), confirming that the tripeptide insert is only one of several features that determine pentapeptide binding.

McpB is exclusively methylated by CheR2

To determine which of the four CheRs methylates McpB, we conducted methylation assays using purified proteins. All samples, including the controls, were composed of an extract of soluble P. aeruginosa proteins that contains the proteins necessary to degrade SAH and avoid product feedback inhibition. Thus, the activity in the control samples corresponds to the methylation caused by components of this cellular extract (Fig. 4A). The amount of McpB methylation in the presence of CheR1, CheR3, or WspC was similar to the control, which implies that McpB is not methylated by these enzymes. In contrast, in the presence of CheR2, the amount of methylated McpB was about 12-fold greater than that in the control sample, demonstrating that CheR2 is the only paralog that methylates McpB. To further characterize the CheR2-mediated methylation, samples were run on SDS–polyacrylamide gel electrophoresis (SDS-PAGE). The methylation of other chemoreceptors was found to increase their electrophoretic mobility on SDS-PAGE gels (3236). Exposure of McpB to CheR2 resulted in four distinct bands with increased electrophoretic mobility that likely corresponds to different methylated forms of McpB (Fig. 4B).

Fig. 4 Methylation of McpB by the four CheR paralogs.

(A) Methyltransferase assays of purified McpB and CheR paralogs. Data are means ± SD from three experiments. (B) SDS-PAGE gel stained with Coomassie detecting the shift in McpB mobility because of its methylation by CheR2. Lanes 1 and 2: purified McpB and CheR2, respectively; lane 3: soluble P. aeruginosa protein extract; lanes 4 to 6: methylation reaction of purified McpB and CheR (in the presence of P. aeruginosa protein extract) after 30, 60, and 90 min.

Disabling the CheR2 pentapeptide interaction abolishes McpB methylation

Having established that McpB was exclusively methylated by CheR2, we investigated the impact of disabling the pentapeptide-McpB interaction on the methylation of McpB by performing methylation assays with McpB, McpBΔGWEEF, CheR2, and CheR2ΔGPN. As previously shown, the interaction was detected exclusively between McpB and CheR2. Deletion of the terminal pentapeptide from McpB or the deletion of the GPN tripeptide, both of which abolished an interaction, also reduced the methylation of McpB to amounts similar to controls (Fig. 5). We can therefore conclude that the interaction of CheR2 with the McpB pentapeptide is essential for methylation under normal conditions.

Fig. 5 Pentapeptide-mediated CheR2-McpB interaction is essential for protein methylation.

Methylation assays using McpB and its mutant lacking the pentapeptide (McpBΔGWEEF), and CheR2 and its mutant lacking the central GPN tripeptide (CheR2ΔGPN). Data are means ± SD from three independent experiments.


In general, protein-protein interactions are mediated by a single interface. CheR and CheB appear to be exceptions because both proteins bind to two distinct sites. What is the reason for this unusual binding mode? From studies using E. coli and S. typhimurium, which have a single CheR and CheB and chemoreceptors with and without C-terminal pentapeptides, it was concluded that the physiological reason for the high-affinity binding of CheR to the pentapeptide is to increase the local concentration of CheR with the flexible linker, enabling methylation of neighboring receptors (17, 28, 37, 38).

Genome analyses indicate that many species have multiple copies of signaling proteins. Here, we investigated this issue using an organism that has multiple CheR proteins. We demonstrate that in P. aeruginosa, the only interaction between a chemoreceptor pentapeptide and CheR is that between McpB and CheR2. The genes encoding these two proteins are next to each other and form part of gene cluster II, which encodes proteins of the Che2 chemotaxis pathway. Therefore, we conclude that the physiological relevance of the pentapeptide docking site is to target a specific chemoreceptor to a specific methyltransferase. Many of the bacteria that have pentapeptide-containing chemoreceptors have multiple CheR paralogs (26), which suggests that this specific targeting mechanism described here for CheR2 and McpB may also be applicable to other CheR-chemoreceptor interactions.

The physiological role of the Che2 chemotaxis pathway is still unclear (10). However, it has been suggested that it mediates chemotactic responses to an as yet unidentified signal molecule during infection (14). This prediction is supported by the demonstration that phosphate starvation increases general virulence-associated phenotypes and induces expression of the che2 gene cluster (39). Güvener et al. (10) have shown that proteins of the Che2 pathway do not colocalize with those of the Che pathway, the principal pathway for chemotaxis. Instead, proteins of the che2 pathway form their own signaling complexes. Because the formation of Che2 signaling complexes depended on McpB, Güvener et al. (10) concluded that among the 26 chemoreceptors, McpB is the only one that participates in the formation of Che2 signaling complexes. Together, these observations suggest that the formation of the Che2 signaling complexes, which consist of CheW2, McpB, CheA2, and CheY2, is highly specific. Our results are in full agreement with the conclusions drawn by Güvener et al. (10) because we demonstrated that pentapeptide-mediated CheR binding guarantees specific binding of CheR2 to McpB. Apart from the specificity of assembly of the Che2 signaling complex, the pentapeptide-mediated interaction enables a specific interaction of this complex with the cognate protein involved in pathway adaptation. We also showed that the pentapeptide of McpA, also encoded by the che2 cluster, is not recognized by any of the CheR paralogs, which is in agreement with the demonstration that the McpA receptor forms signaling complexes with proteins of the Che, but not the Che2, pathway (10). Our results therefore reveal one molecular mechanism that enables the specific interaction of proteins in the Che2 chemoreceptor pathway. In P. aeruginosa, McpB appears to be the only receptor that feeds into the Che2 pathway (10), and CheR2 is the only methyltransferase that methylates McpB. In E. coli, the single CheR acts on chemoreceptors with and without a pentapeptide, and the output of the single pathway reflects the activity of all receptors. Therefore, currently, it is difficult to tell whether the high specificity of McpB in pathway assembly is a more general feature of chemosensory pathways.

A major difference of CheR compared with other SAM-dependent methyltransferases was the insertion of the β-subdomain into the seven-stranded core domain, corresponding to the CheR C-terminal domain (20, 40). Because this subdomain contains the pentapeptide binding site, it could be that pentapeptide recognition was the evolutionary driving force leading to insertion of this subdomain. However, all four CheR paralogs of P. aeruginosa contain this β-subdomain, whereas only CheR2 binds pentapeptides; thus, the mere presence of the β-subdomain in methyltransferases did not direct receptor-specific recognition. Sequence clustering identified that CheR methyltransferases form two families distinguished by their ability to bind pentapeptide-containing receptors. We found that a three–amino acid insertion in the loop linking strands 2 and 3 of the subdomain is a distinct feature of pentapeptide-binding CheR proteins. Using site-directed mutagenesis, we also showed that this tripeptide is essential for pentapeptide binding. Inspection of the CheR-pentapeptide structure illustrates that these three amino acids form few interactions with the bound pentapeptides (20). However, the presence of this tripeptide may be essential to orient strands 2 and 3 of the β-subdomain to interact correctly with the pentapeptide. The presence of this insertion can potentially be used to make sequence-based predictions of pentapeptide-binding CheR proteins.

CheR2 bound exclusively to the terminal pentapeptide of McpB with an affinity about 20-fold higher than the affinity of the NWETF peptide to E. coli CheR (28). Although the latter peptide is absent from P. aeruginosa receptors, CheR2 also recognized this peptide with high affinity. A bioinformatic analysis of pentapeptides from different species has identified the most frequent amino acids for each of the five positions as [D/N]-[W/F]-[E/Q]-[T/E]-[F] (26). This cross-species NWETF pentapeptide–CheR2 binding may be the consequence of the fact that the enterobacterial NWETF pentapeptide contains the most abundant amino acids at each position of the consensus motif.

SAM and SAH compete for binding to CheR. Studies of the enterobacterial enzymes show that SAH binds more tightly to CheR than does SAM (16, 28), which implies that the enzymatic activity of CheR is subject to product feedback inhibition. We made similar observations for the four CheR paralogs of P. aeruginosa, indicating that this is a general feature of CheR proteins. However, large differences between the ratios of the SAH/SAM affinities were observed for the different paralogs. The differences were particularly pronounced for CheR2, which bound SAH much more tightly than it did SAM. An increase in the cellular SAH concentration would thus selectively inhibit CheR2. Further studies will be necessary to verify whether this is of physiological relevance.

Methylation assays showed that McpB was exclusively methylated by CheR2. Disrupting the CheR2-pentapeptide interaction, by deleting either the pentapeptide from McpB or the three–amino acid insertion from CheR2, abolished receptor binding and reduced methylation activity to background amounts (Fig. 5). These results contrast with the corresponding data from enterobacterial systems, in which residual methylation activity is observed after removal of the pentapeptide (17, 22). This suggests that the CheR2-McpB interaction is a strict requirement for any methylation activity.

Chemosensory pathways have been primarily studied in enterobacteria. It is now apparent that these systems are comparatively simple compared with those of other species that contain a greater number of receptors and multiple copies of signaling proteins (7). To assure that multiple copies of sensory proteins assemble into the correct pathways, molecular mechanisms must exist to guarantee the specificity of protein interactions. Here, we identified one of these mechanisms. Pentapeptide-containing chemoreceptors are found in many different species, suggesting that this mechanism is widespread.


Strains and plasmids

The strains and plasmids used in this study are listed in table S2.

Generation of recombinant proteins

The DNA fragments encoding cheR1 (PA3348), cheR2 (PA0175), cheR3 (PA0412), wspC (PA3706), mcpB (PA0176), and mcpBΔGWEEF were amplified by polymerase chain reaction (PCR) (table S3) and genomic DNA of P. aeruginosa PAO1. For wspC, PCR products were digested with Nde I and Sac I; for the cheR constructions, with Nde I and Eco RI; and for mcpB and mcpB-GWEEF, with Nde I and Bam HI. The resulting products were cloned into pET28b(+) (Novagen) linearized with the corresponding enzymes. To generate the CheR2 derivative lacking amino acids 186 to 188 (CheR2ΔGPN) and a CheR1 derivative with a GPN insertion after amino acid 180 (CheR1+GPN), a modified version of the Hemsley method (41) was used. Pairs of overlapping mutagenic primers (table S3) were used to amplify the entire plasmid with PfuTurbo DNA polymerase (Agilent Technologies) followed by the elimination of template DNA by digestion with Dpn I. The resulting PCR products were transformed into E. coli DH5α, and colonies were selected on LB agar-coated plates supplemented with kanamycin (50 μg/ml). Inserts of all plasmids and flanking regions were sequenced. E. coli BL21(DE3) was transformed with the expression plasmids, and cells were grown at 30°C in LB suspension supplemented with kanamycin (50 μg/ml). At an OD660 (optical density at 660 nm) = 0.4, the temperature was decreased to 18°C, and protein expression was induced at an OD660 = 0.8 by the addition of 0.1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG). After overnight growth, cells were harvested by centrifugation at 10,000g. Pellets were resuspended in buffer A [20 mM tris-HCl, 0.1 mM EDTA, 500 mM NaCl, 10 mM imidazole, 5 mM β-mercaptoethanol, 5% (v/v) glycerol (pH 8.0)] and broken by French press. After centrifugation at 20,000g, supernatants were loaded onto a HisTrap HP column (Amersham Biosciences). Proteins were eluted by applying a linear gradient (for 30 min) to 500 mM imidazole in buffer A.

Isothermal titration calorimetry

Isothermal titration calorimetry (ITC) was performed on a VP microcalorimeter (MicroCal) at 25°C. Proteins were dialyzed into 20 mM tris-HCl, 150 mM NaCl, 2 mM MgCl2, 0.1 mM EDTA, 1 mM dithiothreitol (DTT) (pH 7.5) for titrations of CheR paralogs and mutants with small ligands, and into 20 mM Pipes, 150 mM NaCl, and 1 mM DTT (pH 7.5) for titrations of McpB and McpB-GWEEF. For protein-protein interaction, both ligands were dialyzed into the latter buffer. The solutions of small molecules were prepared in the dialysis buffer and placed into the injector syringe. In all cases, control experiments involved the injection of syringe ligand into buffer. Raw data were integrated, corrected for dilution effects, and concentration-normalized before curve fitting using the “one binding site model” of ORIGIN.

Methylation assays

For the assay using tritium-labeled SAM, a modified version of the assays described by Stock et al. (42) was used. Reactions were carried out in 20 mM Pipes, 150 mM NaCl (pH 7.5). Purified McpB or McpBΔGWEEF (1.7 μM) was incubated in the presence or absence of equimolar concentrations of CheR with 100 μM [3H]SAM (0.83 μCi per sample; PerkinElmer) and a crude extract of P. aeruginosa PAO1 (final protein concentration, 5 mg/ml). The resulting mixtures were incubated at 30°C for 30 min, and the reaction was stopped by adding 25 μl of ice-cold 10% (v/v) acetic acid. The amount of methyl ester groups transferred to McpB was quantified as described previously (43). For the SDS-PAGE assay, the reaction mixtures in the above buffer contained 1.7 μM McpB or McpBΔGWEEF, 1.6 μM CheR2 or CheR2ΔGPN, and 100 μM SAM. Samples were incubated at 30°C, and 18 μl of aliquots was removed at intervals. A volume of 6 μl of 4× sample buffer was added to each aliquot, which was resolved on 10% (w/v) SDS-PAGE gels.


Fig. S1. Microcalorimetric titrations of the four CheR paralogs of P. aeruginosa PAO1 with S-adenosylmethionine (SAM) and S-adenosylhomocysteine (SAH).

Fig. S2. Microcalorimetric titration of CheR2 with the NWETF pentapeptide.

Fig. S3. Lack of an allosteric interaction between the pentapeptide-binding site and the SAM/SAH-binding site in CheR2.

Fig. S4. Sequence alignment of different CheR sequences.

Fig. S5. SAH binding by CheR2ΔGPN.

Fig. S6. Clustering of CheR sequences after deletion of the GXX tripeptide.

Fig. S7. Microcalorimetric titrations of CheR1+GPN with the GWEEF pentapeptide or SAH.

Table S1. Lack of allostery between the pentapeptide-binding site and the SAM/SAH-binding site in CheR2.

Table S2. Strains and plasmids used in this study.

Table S3. Oligonucleotides used in this study.

References (4446)


Acknowledgments: We thank J. L. Ramos for the critical reading of the manuscript and B. Pakuts for correcting the English. Funding: We acknowledge financial support from FEDER funds and Fondo Social Europeo through grants from the Junta de Andalucía (grants P09-RNM-4509 and CVI-7335 to T.K.) and the Spanish Ministry for Economy and Competitiveness (grant Bio2010-16937 to T.K.). Author contributions: C.G.-F. and A.C.L. carried out the research and analyzed the data. T.K. analyzed the data and wrote the manuscript. Competing interests: The authors declare they have no competing interests.
View Abstract

Navigate This Article