Selective Methylation Changes on the Bacillus subtilis Chemotaxis Receptor McpB Promote Adaptation*

The Bacillus subtilis McpB is a class III chemotaxis receptor, from which methanol is released in response to all stimuli. McpB has four putative methylation sites based upon theEscherichia coli consensus sequence. To explore the nature of methanol release from a class III receptor, all combinations of putative methylation sites Gln371, Gln595, Glu630, and Glu637 were substituted with aspartate, a conservative substitution that effectively eliminates methylation. McpB(Q371D,E630D,E637D) in a Δ(mcpA mcpB tlpA tlpB)101::cat mcpC4::erm background failed to release methanol in response to either the addition or removal of the McpB-mediated attractant asparagine. In the same background, McpB(E630D,E637D) produced methanol only upon asparagine addition, whereas McpB(Q371D,E630D) produced methanol only upon asparagine removal. Thus methanol release from McpB was selective. Mutants unable to methylate site 637 but able to methylate site 630 had high prestimulus biases and were incapable of adapting to asparagine addition. Mutants unable to methylate site 630 but able to methylate site 637 had low prestimulus biases and were impaired in adaptation to asparagine removal. We propose that selective methylation of these two sites represents a method of adaptation novel from E. coliand present a model in which a charged residue rests between them. The placement of this charge would allow for opposing electrostatic effects (and hence opposing receptor conformational changes). We propose that CheC, a protein not found in enteric systems, has a role in regulating this selective methylation.

The Bacillus subtilis McpB is a class III chemotaxis receptor, from which methanol is released in response to all stimuli. McpB has four putative methylation sites based upon the Escherichia coli consensus sequence. To explore the nature of methanol release from a class III receptor, all combinations of putative methylation sites Gln 371 , Gln 595 , Glu 630 , and Glu 637 were substituted with aspartate, a conservative substitution that effectively eliminates methylation. McpB (Q371D,E630D,E637D) in a ⌬(mcpA mcpB tlpA tlpB)101::cat mcpC4::erm background failed to release methanol in response to either the addition or removal of the McpB-mediated attractant asparagine. In the same background, McpB (E630D,E637D) produced methanol only upon asparagine addition, whereas McpB (Q371D,E630D) produced methanol only upon asparagine removal. Thus methanol release from McpB was selective. Mutants unable to methylate site 637 but able to methylate site 630 had high prestimulus biases and were incapable of adapting to asparagine addition. Mutants unable to methylate site 630 but able to methylate site 637 had low prestimulus biases and were impaired in adaptation to asparagine removal. We propose that selective methylation of these two sites represents a method of adaptation novel from E. coli and present a model in which a charged residue rests between them. The placement of this charge would allow for opposing electrostatic effects (and hence opposing receptor conformational changes). We propose that CheC, a protein not found in enteric systems, has a role in regulating this selective methylation.
Chemotaxis is the process by which motile bacteria sense their environment and move toward more favorable surroundings. This is accomplished by altering the direction of flagellar rotation in response to changes in effector gradients. Information about the outside of the cell is transduced to the inside via transmembrane receptors called methyl-accepting chemotaxis proteins. Both ligand binding to the extracellular portion of the receptor and methyl modification of the intracellular portion induce conformational changes that modulate the activity of an associated histidine kinase, CheA (1). CheA activity in turn regulates phosphorylation of the response regulator CheY, the activated form of which binds to the motor switch to promote smooth swimming (1,2). However, bacteria respond only to changes in their environment. In the presence of unchanging stimuli, the transduced signal is shut off despite the fact that ligand is still bound. This ability to adapt is essential for chemotaxis and allows cells to respond to ever changing environmental conditions.
In Escherichia coli, adaptation depends largely upon the reversible methyl esterification of certain glutamate residues that serve to counteract the effects of bound ligand (3)(4)(5). The methylesterase CheB, which is also CheA-activated, demethylates the receptor to promote adaptation to attractant removal. Receptor demethylation results in the release of volatile methanol from the cell (6 -9). CheB also functions as a deamidase that converts the nascent glutamine residues of some sites to glutamates, which in turn serve as sites for CheR-mediated methylation (3). A dedicated methyltransferase, CheR, methylates the receptor using S-adenosylmethionine as the methyl donor to promote adaptation to attractant addition (10,11). A deletion in either of these genes results in cells that cannot adapt to stimuli (12).
Mechanistically, the binding of attractant to the N-terminal sensing domain of the E. coli aspartate receptor Tar is believed to cause a 1.6 Å downward displacement with a 5°tilt in the second transmembrane spanning helix (TM2) of one monomer relative to the other transmembrane helices of the homodimeric receptor complex. This piston-like displacement may be propagated down the length of the receptor because of its predominantly ␣-helical nature and thus affect CheA activity (13)(14)(15)(16). The cytosolic C terminus is an antiparallel coiledcoil containing the methylation and signaling regions (Fig. 1). Methylation of the receptor presumably causes a compensatory shift in TM2 that restores basal CheA activity.
Reversible receptor methyl esterification is also implicated in adaptation in Bacillus subtilis. However, deletions of cheB and cheR merely impair adaptation, particularly at high receptor occupancy (12). This has led to speculation that B. subtilis has both methylation-dependent and methylation-independent adaptation systems (9). Unlike E. coli, methanol is released both upon the addition and the removal of all attractants tested, a result that suggests the mechanism of adaptation is different between the two organisms. Moreover, adaptation also involves CheC, a protein with no E. coli counterpart (18 -20).
Sequence alignments reveal three distinct classes of receptor (21). Class III receptors are typified by the presence of four 14-amino acid insertion/deletion (INDEL) 1 regions located within the methylation and signaling regions. Class II receptors have only INDELs 2 and 3 within the signaling region, * This work was supported by Public Health Service Grant GM54365 (to G. W. O.) from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work is dedicated in loving memory to Jane Tucker Zimmer (1940 -1996).
whereas class I receptors contain no INDELs (Fig. 1). These INDEL regions are postulated to give the antiparallel coiledcoil receptor helices extra length without altering their orientation (21). Enteric bacteria such as E. coli and Salmonella typhimurium have class I receptors that release methanol in response to negative stimuli while suppressing methanol release in response to positive stimuli (22,23). The class III receptors found in the Gram-positive B. subtilis and in the archaeon Halobacterium salinarum release methanol in response to all stimuli (24 -27). Moreover, the B. subtilis receptor McpB undergoes net demethylation followed by net remethylation in response to the addition of asparagine such that the level of receptor methylation in the prestimulus and adapted states is the same (20). For class I receptors, attractants cause a net increase in receptor methylation, whereas repellents cause a net decrease (28,29).
The restoration of net receptor methylation in B. subtilis following the addition of asparagine suggested that selective methylation changes might be taking place at a specific subset of sites on the unbound versus bound receptor. To explore this possibility, all combinations of glutamines/glutamates within the four putative consensus sites reported by Hanlon and Ordal (30) were substituted with aspartate. Similar substitutions in the E. coli receptor Tar were very poorly methylated (31). In this way, we ensure a fixed negative charge at a given site. If the selective methylation hypothesis is true, then some mutants should greatly affect methanol release and adaptation to attractant addition, whereas other mutants should affect methanol release and adaptation to attractant removal.
The pAIN700 was subjected to polymerase chain reaction using the above primers, which were designed to create a new restriction site at the ligation junction to make the mutant identifiable by restriction digest. The polymerase chain reaction product was then ligated, digested with Dpn1 to remove template plasmid, and then transformed into TG1 E. coli. Amp R candidate colonies were screened using the introduced restriction sites (BglII for Q371D and Q595D, NheI for E630D, and EcoRV for E637D). The pertinent region of each single substitution mutant was sequenced at the W. M. Keck Center for Comparative and Functional Genomics. Double, triple, and quadruple substitutions were created in an iterative process. The proximity of site Glu 630 to Glu 637 required the use of a special forward primer to create the E630D/E637D double mutant. 2.8-kb EcoRI/NotI fragments containing the mutant mcpB alleles were subcloned into pAIN750 digested with the same enzymes to generate pONG1 through pONG1234. pAIN750 is a modified version of the amyE integration vector pDG1730 (17).
Pulse-labeled Methylations-Experiments were performed as described (33). Briefly, 100 l of a 2-ml overnight culture was added to 10 ml of LB, incubated at 37°C, and grown to early stationary phase. Cells were washed twice in CB with 250 g ml Ϫ1 chloramphenicol, once in protoplast buffer with 250 g ml Ϫ1 chloramphenicol, and resuspended  (30), where the deviation of site 595 from the consensus is indicated in bold. Class I receptors release methanol only in response to negative stimuli, whereas class III receptors release methanol in response to all stimuli.
to A 600 ϭ 1.0 in 3 ml of protoplast buffer with 250 g ml Ϫ1 chloramphenicol and 4 mg ml Ϫ1 lysozyme. Cells were shaken at 250 rpm at 37°C for 20 min following the addition of 150 l of 1 M L-[methyl-3 H]methionine. 1-ml aliquots were removed and flash-frozen in dry ice/acetone baths. Samples were then thawed, and centrifuged at 3000 ϫ g for 45 min at 4°C. Pelleted membranes were resuspended in 100 l of 4ϫ SDS solubilizer, boiled for 7 min, and electrophoresed at 15 V cm Ϫ1 through 10% SDS-polyacrylamide gels, pH 8.8 (36). Gels were treated for fluorography as described (37).
Western Blot Analysis of McpB-Cells were treated as outlined for pulse label methylations except that no L-[methyl-3 H]methionine was added. Western analysis of mcpB mutants was performed as described (38). Antigen-antibody was detected using Supersignal (Pierce). Visualization required a 5-s to 1-min exposure of Kodak X-Omat AR film.
Continuous Flow Assay-The continuous flow assay has been described (39). Briefly, 100 l of a 2-ml overnight culture was added to 10    (40). Tethered Cell Assay-The tethered cell assay has been described (20). Strains were grown on tryptose blood agar base plates O/N in a 30°C wet incubator. Cells were taken straight off the plate and suspended in minimal medium. The suspension was then diluted to A 600 ϭ 0.014 in 25 ml of minimal medium and grown at 37°C with vigorous shaking (250 rpm) for 4.5 h. 15 min prior to harvesting, 200 l of a 5% glycerol, 0.5 M sodium lactate solution was added to the culture. The cells were blended at full speed in a Warring blender for 10 s to shear the flagella. A 250-l aliquot of blended cells was placed atop a glass coverslip preincubated with 10 l of anti-flagella antibody and incubated for 15 min. The coverslip was then inverted and placed within a laminar flow chamber to be observed by phase-contrast microscopy in a system similar to those described previously (9,41). CB with or without 504 M asparagine was pumped through the laminar flow chamber at a constant rate of 1 ml min Ϫ1 . Rotating cells were videotaped and analyzed by a Hobson Tracker, Bacterial Edition (Hobson Tracking Systems Ltd., Sheffield, U.K.), which generates text files containing continuous time rotational data. Programs written with Matlab software (The Mathworks, Inc.) processed the data. For each cell, the continuous time rotational data were converted to discrete time data with a step size of 0.1 s. All the data points within a 4-s window were then averaged to generate a data set with a step size of 4 s that contained the probability of counterclockwise rotation (smooth swimming behavior). The data for all cells in the sample population were then pooled and averaged, and the resulting data were smoothed over a five data point window to generate the final graph.

Effect of Triple and Quadruple Aspartate Substitutions on
Methanol Release-To test whether an aspartate substitution at a given site prevents its methylation, the quadruple aspartate mutant was tested in the continuous flow assay. In this experiment, the evolution of [ 3 H]methanol was followed by collecting eluant fractions as buffer or buffer plus asparagine was passed over cells caught on a 0.45-m filter (see "Experimental Procedures"). As expected, McpB (Q371D,Q595D,E630D,E637D) had no [ 3 H]methanol release upon either the addition or removal of asparagine (data not shown). McpB (Q371D,E630D,E637D) also failed to release [ 3 H]methanol (Fig. 2). Therefore, site 595 did not contribute to McpB methanol production and is not likely a site of methylation. Given the deviation of this site from the methylation consensus sequence, this was not too surprising ( Fig. 1). Moreover, no methylation was observed for McpB (Q371D,E630D,E637D) in pulse label methylation experiments (data not shown; see "Experimental Procedures"). Thus [ 3 H]methanol production from McpB was attributable only to sites 371, 630, and 637. Western blot analysis using anti-McpB antibody confirmed approximate wild-type expression for all mcpB mutants used in this study (data not shown). All experiments were done in a ⌬(mcpA mcpB tlpA tlpB)101::cat mcpC4::erm background to avoid the possible complication of receptor cooperativity. We conclude that sites 371, 630, and 637 are the only sites of methylation on McpB.
Effect of Single Aspartate Substitutions on Methanol Release-To determine the contribution of individual sites on methanol production, continuous flow assays were performed on each of the single substitution mutants. All mutants released [ 3 H]methanol upon both the addition and removal of asparagine (Fig. 3). McpB (Q595D) had wild-type [ 3 H]methanol production, a result consistent with it not being a site of methylation (Fig. 3B). However, the release of [ 3 H]methanol was markedly diminished for the remaining mutants. We conclude that no site is solely responsible for [ 3 H]methanol release upon either the addition or removal of asparagine.
Effect of Double Aspartate Substitutions on Methanol Release-Greater insight into the B. subtilis methylation system has come from the double aspartate substitution mutants. McpB (Q371D,E630D) failed to release [ 3 H]methanol upon asparagine addition, whereas leaving [ 3 H]methanol release upon asparagine removal intact (Fig. 4A). Thus sites 371 and 630 were responsible for [ 3 H]methanol release upon asparagine addition,  It has been demonstrated for the E. coli receptor Tar that aspartate substitutions have polar effects on nearby sites. Specifically, methylation was nearly eliminated at sites seven residues N-terminal to a mutated site (31). We do not know the extent to which an aspartate at site 637 effects methylation of site 630 or vice versa. However, we note that because [ 3 H]methanol can be detected in both McpB (Q371D,E637D) and McpB (Q371D,E630D) , the unaltered sites must be at least partially susceptible to methylation changes. We conclude that methanol evolution from McpB is selective. Only sites 371 and 630 are subject to methyl turnover upon asparagine addition, whereas only sites 630 and 637 are subject to methyl turnover upon removal.
Delayed Methanol Release upon Asparagine Removal-Methanol release upon asparagine removal for some mutants lagged behind that of the native receptor.  (Table III).
Behavioral Effects of the McpB (Q637D) and McpB (Q371D,E637D) Mutants-To explore the behavioral consequences of a fixed negative charge at the various sites, we tested the aspartate substitution mutants in the tethered cell assay. In this assay, cells are tethered to a class coverslip using anti-flagella antibody. The coverslip is then placed over a laminar flow chamber, and rotational data are collected as buffer or buffer plus asparagine is passed over the cells (see "Experimental Procedures"). McpB (E637D) and McpB (Q371D,E637D) both had prestimulus biases of ϳ70% and both failed to adapt to asparagine addition. De-excitation and adaptation to asparagine removal was largely unaffected. However, McpB (Q371D,E637D) had a more pronounced removal response (Fig. 5A). We conclude that methylation of site 637 is involved in adaptation to the addition of asparagine. (E630D) and McpB (Q371D,E630D) mutants both had prestimulus biases of ϳ40%. McpB (E630D) adapted normally to the addition of asparagine but failed to adapt to asparagine removal. McpB (Q371D,E630D) initially adapted to asparagine addition but failed to maintain adaptation. This mutant was also impaired in adaptation to the removal of asparagine (Fig. 5B). We conclude that methylation of site 630 is involved in adaptation to asparagine removal.

Behavioral Effects of the McpB (E630D) and McpB (Q371D,E630D) Mutants-McpB
Behavioral Effects of Aspartate Substitution at Both Sites 630 and 637 or Neither-All other mutants had prestimulus biases closer to that of the native receptor (Table IV). Additionally, they all adapted to both the addition and removal of asparagine. The most remarkable phenotype of these mutants was their delay in removal response (Table V). McpB (E630D,E637D) and McpB (Q371D,E630D,E637D) were particularly striking with a delay of ϳ2 and 1.6 min, respectively. McpB (Q371D) appeared almost wild-type (Fig. 5C). We note that the effect of a fixed negative charge at site 637 (high bias and failure to adapt to asparagine addition) is only manifest when site 630 can be methylated. Conversely, the effect of a fixed negative charge at site 630 (low bias and failure to adapt to asparagine removal) is only manifest when site 637 can be methylated.

DISCUSSION
The observation that B. subtilis releases methanol in response to all stimuli is significant because it suggests that the role of receptor methylation is different from that of enteric organisms. This may underscore a general difference between class I and class III receptors. Such a difference is particularly interesting because B. subtilis CheR and CheB methylate and demethylate E. coli receptors in vitro (11,42), and the methylesterase functionally complements its respective E. coli null mutant in vivo (9). E. coli receptors are known to migrate at different rates through SDS-polyacrylamide gel electrophoresis depending on their degree of methylation (43). McpB, however, runs as a single band that splits into a doublet upon asparagine addition and returns to a single band following asparagine removal (20). This suggested that McpB might predominantly have two methylated states and that the receptor is methylated selectively upon asparagine addition and removal. If this hypothesis is true, then methanol release upon the addition and removal of attractant each represent demethylation from a certain subset of sites. Our finding is that selective demethylation does occur; methyl groups come off McpB from sites 371 Methanol release, as measured in the flow assay, does not represent a single demethylation event but rather an increased level of methyl turnover at susceptible sites (38). No site was entirely responsible for methanol release upon either asparagine addition or removal (Fig. 3). However, the double substitution mutants reveal that site 371 was subject to demethylation only upon asparagine addition, whereas site 637 was subject to demethylation only upon asparagine removal. Site 630 appeared indiscriminate and was demethylated in both cases. It is important to point out that being subject to methyl turnover does not necessarily mean that a site was methylated to begin with. These results imply that, in the transition period between the addition of asparagine and the return to basal methanol production, the susceptible sites are 371 and 630. In the transition period between the removal of asparagine and the return to basal methanol production, the susceptible sites are 630 and 637. This suggests that the receptor conformational change following attractant addition to the adapted state is different from the conformational change following attractant removal.
Class III receptors are thought to closely resemble their class I counterparts with only the addition of four 14-amino acid INDEL regions, which serve to extend the length of the antiparallel coiled-coil structure without affecting the relative orientations of the helices (21). This places site 630 and site 637 11 Å apart on the same face of the coiled-coil helix (Fig. 1). The fact that these two sites were selectively demethylated is most striking and strongly supports a model involving selective receptor methylation in B. subtilis.
All McpB mutants exhibiting a methylation phenotype were tested in the tethered cell assay to determine the in vivo consequences of forced negative charges at the various sites. It was most interesting to see that site 630 and site 637 had an "opposite" global effect on chemotaxis. Mutants in site 637 (when site 630 was unchanged) had higher than normal (ϳ70%) prestimulus biases and failed to adapt to the addition of asparagine. Mutants in site 630 (when site 637 was unchanged) had low (ϳ40%) biases, and the McpB (E630D) mutant in particular had a marked inability to adapt to asparagine removal. These phenotypes were only manifest when the adja-cent site was unchanged. Mutants having both or neither sites 630 and 637 altered had more nearly wild-type biases and were capable of adaptation. Because a methylated glutamate residue is neutral, adaptation appears to involve switching the negative charge between these two sites.
These observations lead to the following electrostatic model for B. subtilis adaptation (Fig. 6). We propose that there is a fixed charged residue on one monomer of the functional homodimer that is strategically placed between sites 630 and 637 of the other monomer. This charge can be either positive or negative. Here the model is presented with a fixed negative charge promoting selective electrostatic repulsions. However, an analogous construction can be made using a fixed positive charge that promotes selective electrostatic attractions. We believe that attractant binding results in a conformational change in McpB that shifts one monomer up relative to the other to activate CheA (as opposed to E. coli, where the addition of attractant results in the TM2 region of one monomer shifting down to inhibit CheA) (13)(14)(15)(16). This results in demethylation of site 630. Adaptation to the addition of attractant is achieved by methylating site 637 but not site 630. This would cause a compensatory conformational change that balances the effect of bound ligand. In the model, we visualize this as a shifting of the monomer back down because of the elimination of electrostatic repulsion between the glutamate at site 637 and the hypothesized negative charge. Removal of attractant would cause the monomer to shift down relative to the other to inhibit CheA. This results in demethylation of site 637. Adaptation to attractant removal is achieved by methylating site 630 but not site 637. This causes a compensatory conformational change back to the original state. In the model, we visualize this as a restoring upward shift of the monomer because of the elimination of electrostatic repulsion between the glutamate at site 630 and the putative negative charge.
Single aspartate substitutions at sites 630 or 637 have opposing phenotypes because of opposing electrostatic effects from the hypothesized charge. Aspartate substitution at both sites would fail to generate a net force in either direction and would thus have a phenotype somewhere in between. The model is thus consistent with earlier findings in B. subtilis that show that cheB and cheR mutants are capable of adaptation by a methylation-independent mechanism (9). Because sites 630 and 637 are glutamates, cheB and cheR mutants would result in both of these sites being either methylated or unmethylated, respectively. The system is apparently designed to be robust, and the loss of either enzyme results in a receptor that defaults into a conformation that can at least partially adapt through the methylation-independent mechanism. The methylation-independent mechanism appears sufficient for adaptation to low concentrations of asparagine, as taxis in a cheB mutant in the capillary assay is nearly normal at low concentrations (38).
Because B. subtilis and E. coli CheB and CheR appear to be interchangeable, it stands to reason that something is regulating selective methylation at a level above these enzymes. Obvious candidates for this are CheC and CheD, two proteins found in B. subtilis and the archaea, but not found in E. coli. Both CheC and CheD appear to bind the cytosolic portion of McpB 2 and are known to affect receptor methylation (18). Interestingly, a cheC mutant has a phenotype that is almost identical to that of the aspartate substitution at site 637 in that both have markedly high prestimulus biases and fail to adapt to the addition of attractant (Fig. 5A, 18). It is therefore possible that CheC is required to facilitate methylation of this site.
Methylation sites 371 and 637 are immediately adjacent to 2 C. J. Kristich and G. W. Ordal, unpublished data.  INDEL regions 1 and 4 ( Fig. 1), and there is a correlation between the presence of these INDELs and the presence of both CheC and CheD. This suggests two possible mechanisms for regulating selective methylation. The first involves exposing sites because of induced changes in receptor conformation. We imagine that the selective binding of CheC or CheD to these INDELs could determine whether a particular site is subject to methyl modification by causing it to be either buried at the interface between the two monomers (thus protected from modification) or exposed because of a twisting motion. Another mechanism might involve steric interference. The binding CheC or CheD to the INDELs could cause them to overlap the consensus sequences of sites 371 and 637 and thus deny the methylesterase and/or the methyltransferase access to the sites. Such a binding event could be dependent upon the methylated state of the receptor. Lastly, it is tempting to speculate on the source of the putative charge. Conceivably, it could come from a residue on one monomer of the homodimer, from a protein that binds to one monomer and thus fixes a negative charge between sites 630 and 637 of the other, or even be the result of a post-translational modification. Kim et al. (44) has recently crystallized the FIG. 6. Electrostatic model for B. subtilis adaptation. This model is depicted with a fixed negative charge between sites 630 and 637. However, an analogous construction can be made proposing a fixed positive charge between the two sites. Phosphorylation of one CheA in the homodimeric complex is meant to depict basal CheA activity; phosphorylation of both CheAs, an increase in activity, and no phosphorylation, CheA inhibition. CheA is associated with the receptor via either of two scaffolding proteins, CheW or CheV. A, homodimeric conformational changes that take place upon attractant addition. To the left is the homodimer in the absence of attractant. Sites 630 and 637 are shown on one monomer relative to the hypothesized negative charge on the other. In the absence of attractant, we show site 630 methylated and site 637 unmethylated. Binding of attractant (middle) results in an upward shift in the second transmembrane helix, and correspondingly, the entire cytosolic portion of one monomer of the homodimer such that CheA activity is increased. Adaptation to attractant addition (right) results from demethylation of site 630 and methylation of site 637. Methylation of site 637 removes electrostatic repulsion between site 637 and the putative negative charge. This results in a compensatory downward shift that restores CheA activity to basal levels. B, homodimeric conformational changes that take place upon attractant removal. Shown to the left is the adapted receptor. Removal of attractant (middle) results in a downward shift in the second transmembrane helix and corresponding cytosolic portion of one monomer such that CheA activity is decreased below basal levels. Adaptation to the removal of attractant (right) results from demethylation of site 637 and methylation of site 630. Methylation of site 630 removes electrostatic repulsion between site 630 and the hypothesized negative charge. This results in a compensatory upward shift that re-establishes basal CheA activity.
C terminus of an E. coli serine receptor Tsr in which all sites of methylation were changed to glutamine residues. They report possible hydrogen bonding contacts between the glutamines representing methylation sites Gln 297 , Glu 304 , and Gln 311 on the first methylated helix of one monomer of the homodimer with residues Glu 479 , Gln 472 , and Glu 465 on the second methylated helix of the other monomer (44). Assuming these hydrogen bonding contacts lie on the dimeric interface, the multiple sequence alignment of Le Moual and Koshland (21) would implicate the lysine residue at position 381. It is also possible that the putative charge comes from an auxiliary protein bound to one monomer of the homodimer. The proximity of INDEL regions 1 and 4 to the sites of methylation make CheC and CheD attractive candidates.