|Home | About | Journals | Submit | Contact Us | Français|
Adaptation in bacterial chemotaxis involves reversible methylation of specific glutamate residues within the cytoplasmic domains of methyl-accepting chemotaxis proteins. The specific sites of methylation in Salmonella enterica and Escherichia coli chemoreceptors, identified 2 decades ago, established a consensus sequence for methylation by methyltransferase CheR. Here we report the in vitro methylation of chemoreceptors from Thermotoga maritima, a hyperthermophile that has served as a useful source of chemotaxis proteins for structural analysis. Sites of methylation have been identified by liquid chromatography-mass spectrometry/mass spectrometry. Fifteen sites of methylation were identified within the cytoplasmic domains of four different T. maritima chemoreceptors. The results establish a consensus sequence for chemoreceptor methylation sites in T. maritima that is distinct from the previously identified consensus sequence for E. coli and S. enterica. These findings suggest that consensus sequences for posttranslational modifications in one organism may not be directly extrapolated to analogous modifications in other bacteria.
Sensory adaptation to persisting stimulation permits cells to respond with greater sensitivity to temporal changes in stimuli. In Escherichia coli and Salmonella enterica serovar Typhimurium, adaptation during bacterial chemotaxis is in part mediated by reversible covalent modifications of transmembrane chemoreceptors, also referred to as methyl-accepting chemotaxis proteins, in which specific glutamate residues within the cytoplasmic domains are methylated by methyltransferase CheR and demethylated by methylesterase/deamidase CheB (12, 27). While molecular mechanisms of chemotaxis signal transduction and adaptation have been most extensively studied in E. coli and S. enterica (3, 30), Thermotoga maritima has been a useful source of chemotaxis proteins for structural characterization in cases where crystallization of mesophilic orthologs has failed (2, 9, 22).
T. maritima is a thermophilic bacterium that thrives at high temperatures, with an optimal growth temperature of 80°C (25). This temperature optimum poses difficulties for laboratory studies of bacterial physiology and/or behavior. However, since proteins from this organism continue to be useful for structural characterization, it is advantageous to characterize the in vitro methylation of the T. maritima chemoreceptors as a basis for the generation, manipulation, and interpretation of receptor signaling complexes. The T. maritima genome encodes six different transmembrane chemoreceptors with diverse periplasmic domains and conserved cytoplasmic domains. The C-terminal signaling domains have from 34% to 100% sequence identity in pairwise alignments with each other and between 23% and 29% identity in alignments with the E. coli aspartate chemoreceptor Tar. No ligands for the T. maritima receptors have yet been identified. The T. maritima genome also contains genes encoding the methylating and demethylating enzymes CheR and CheB, as well as receptor deamidase CheD and associated proteins CheC and CheX. The enzyme activities for T. maritima methylesterase CheB with methylated S. enterica Tar as the substrate (1) and deamidase CheD with T. maritima chemoreceptors as the substrate (6) have previously been reported.
The cytoplasmic signaling domains of chemoreceptors have an extended coiled-coil three-dimensional structure (13) and a conserved primary structure (16, 32). The greatest sequence conservation occurs within a region designated the highly conserved domain (HCD) at the distal tip of the coiled-coil domain, the locus of protein-protein interactions between receptors and the signaling proteins CheA and CheW. The next most highly conserved segments correspond to two regions encompassing the methylation sites. A methylation consensus sequence, Glx-Glx-X-X-Ala-Ser/Thr (modified residue is in bold), was originally proposed (27) based on direct identification of methylation sites in S. enterica Tar (27) and the E. coli serine chemoreceptor Tsr (11). At some sites, the methylated residue is encoded as Gln and the side chain is deamidated by methylesterase CheB prior to participation in the reversible methylation cycle (11). Additional methylation sites identified in E. coli Trg, the chemoreceptor that mediates responses to ribose and galactose, conform to this consensus (19).
It has been widely assumed that the consensus sequence derived from E. coli and S. enterica receptors can be used to identify potential sites of methylation in chemoreceptors of other bacteria. In this study, partially purified T. maritima methyltransferase CheR has been used to methylate the four different T. maritima chemoreceptor cytoplasmic domains representative of all six transmembrane chemoreceptors, and the sites of methylation have been identified by using liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS). Our results establish a distinct consensus methylation sequence for T. maritima chemoreceptors and thus demonstrate that the previously defined consensus methylation sequence for E. coli and S. enterica receptors is not generally applicable to receptors of other bacteria.
T. maritima cheR (open reading frame TM0464) was amplified by PCR using genomic T. maritima DNA as the template and primers that introduced NdeI and HindIII restriction sites at the 5′ and 3′ ends, respectively. The fragment was inserted into complementary sites in pJES307 (26) to generate plasmid pEP01. Plasmid DNA for pET28b (Novagen)-derived expression vectors encoding full-length T. maritima receptors (named according to open reading frame numbers) (18) TM0429fl (pTM0429), TM1143fl (pTM1143), and TM1428fl (pTM1428) with His tags and linkers at their N termini generated by insertion of coding sequences into the SalI and NotI restriction sites of the vector were provided by Bryan Beel (California Institute of Technology). Plasmid DNA for pET28a (Novagen)-derived expression vectors encoding the C-terminal cytoplasmic domains of T. maritima receptors TM0429c (residues 348 to 656, pTM0429c), TM1143c (residues 225 to 530, pTM1143c), and TM1428c (residues 261 to 566, pTM1428c) with His tags and linkers at their N termini generated by insertion of coding sequences into the NdeI and BamHI restriction sites of the vector were provided by Brian Crane (Cornell University). DNA encoding the C-terminal cytoplasmic domain of receptor TM1146 was amplified by PCR using genomic T. maritima DNA as the template and primers that introduced NdeI and BamHI restriction sites at the 5′ and 3′ ends, respectively. The fragment was inserted into complementary sites in pET28a to generate plasmid pTM1146c, which expresses TM1146c (residues 234 to 539) with a His tag and linker at the N terminus. All plasmids were transformed into E. coli strain BL21(DE3) (Novagen). Cells were grown at 37°C to mid-log phase in Luria-Bertani medium supplemented with 100 μg/ml ampicillin for pJES307-derived plasmids and 30 μg/ml kanamycin for pET28-derived plasmids. Expression of proteins was induced with 1.0 mM isopropyl-β-d-thiogalactopyranoside, and incubation was continued for 3 h at 37°C.
Expression vectors for mutant proteins were constructed using a QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer's protocol. Briefly, using pTM1143c and pTM1428c plasmid DNA as the template, plasmids pTM1143c(Q274EQ498E) and pTM1428c(Q499E), which encode deamidated receptors, were generated by converting Gln codons (CAG or CAA) to Glu codons (GAG or GAA). Constructs were confirmed by DNA sequencing and transformed into BL21(DE3) for expression.
For preparation of partially purified T. maritima CheR, cells were harvested by centrifugation for 10 min at 3,000 × g, resuspended in 100 mM potassium phosphate-1 mM EDTA-1 mM β-mercaptoethanol, pH 7.0 (buffer A), harvested by centrifugation, and then resuspended in 3 ml buffer A per gram (wet weight) of cells. A cell-free lysate was generated by sonication followed by centrifugation for 60 min at 100,000 × g. The supernatant was heated for 20 min at 80°C and centrifuged for 60 min at 100,000 × g. The supernatant, enriched in T. maritima CheR (~70% pure), was stored in aliquots at −20°C. Salt-washed membrane fractions containing full-length chemoreceptors were prepared as described previously (24). Cell-free lysates containing cytoplasmic domains of T. maritima receptors, prepared as described above, were incubated at 80°C for 20 min and clarified by centrifugation for 60 min at 100,000 × g. The supernatant was dialyzed into 20 mM sodium phosphate-500 mM NaCl-20 mM imidazole, pH 7.4, and was applied to a 5-ml HisTrap column (Amersham Biosciences) equilibrated in the same buffer. Receptors were eluted with 500 mM imidazole, dialyzed into buffer A, and diluted to a final concentration of 6 μM. Receptor proteins were estimated to be ~90% pure. The concentrations of CheR and receptor cytoplasmic domains were estimated by comparison to protein standards on Coomassie blue-stained sodium dodecyl sulfate-polyacrylamide gel electrophoretograms.
Receptor methylation assays to determine initial rates of methylation were performed as described previously (24). Each 100-μl reaction mixture contained 40 to 50 μl of either salt-washed membranes or receptor cytoplasmic domains (each containing ~6 μM receptor) and CheR protein preparation in 100 mM potassium phosphate-100 μM [3H]S-adenosylmethionine at 162 Ci/mol (specific activity, 15 Ci/mmol; NEN Life Science Products, Inc.), pH 7.0. Samples were preequilibrated at temperatures ranging from 30 to 70°C for 10 min, and reactions were initiated by addition of 10 μl of CheR (ranging from 0.39 to 2.92 pmol to achieve linear rates). For typical assays, five to six time points were taken at 3- to 6-min intervals, and initial rates were estimated using linear regression analysis. Each CheR-receptor pair was assayed at two different CheR concentrations to confirm linearity with respect to CheR, and all methylation rates reported were derived from assays repeated three times. Assays to determine steady-state levels of methylation were performed using a final concentration of 3 μM receptor cytoplasmic domains, 0.05 μM CheR in 100 mM potassium phosphate, pH 7.0. Samples were preequilibrated for 10 min at 50°C, and methylation was initiated by the addition of [3H]S-adenosylmethionine to a final concentration of 100 μM at 162 Ci/mol.
Methylation of receptors to steady-state levels was performed as described above, except that the reaction volume was increased to 200 μl. Reactions were performed in the presence and absence of S-adenosylmethionine without addition of radiolabel, and reactions were carried out for 90 min. Protein samples were denatured and reduced with 7 mM guanidine HCl and 10 mM dithiothreitol at 50°C for 30 min. Reduced cysteine residues were alkylated with 20 mM iodoacetamide at 25°C for 1 h in the dark. Samples were then buffer exchanged into 50 mM NH4HCO3 using a Microcon 10,000 molecular weight cutoff (Millipore). Sequencing-grade trypsin (Promega) was then added to samples at a 1:50 (wt/wt) ratio, and samples were digested overnight at 37°C. In cases where peptides were too large for exact methylation site determination, sequencing-grade Glu-C (Roche) was added to tryptic peptides and incubated overnight at 25°C.
Peptides were injected onto a reverse-phase column (75 μm by 15 cm self-packed with Magic C18 AQ; 3 μm; 200 Å; Michrom Bioresources Inc.) using an ultimate nano-liquid chromatography system (Dionex/LC Packings). The column was equilibrated in 0.1% formic acid (solvent A) and eluted with a linear gradient of 2 to 45% solvent B (0.1% formic acid in acetonitrile) at a flow rate of 200 nl/min over 30 min, and peptides were analyzed by electrospray ionization-MS/MS using an LTQ ion trap mass spectrometer (ThermoFinnigan) equipped with a nanospray source (Proxeon Biosystems). Each MS scan was followed by subsequent zoom scans and MS/MS scans of the four most abundant, multiply charged ions, with a dynamic exclusion of 1 min. Data files (.dta) for MS/MS spectra were generated by Bioworks software (ThermoFinnigan) and were searched against a database containing the sequences of T. maritima receptors and methyltransferase CheR, using Sequest (8) and/or the general purpose m. use (GPM) search engine (7). Search results were further analyzed by manual inspection.
Receptor cytoplasmic domains have several advantages over full-length transmembrane receptors for biochemical analyses. The cytoplasmic domains do not require solubilization from membranes, and their smaller size simplifies the identification of methylation sites within proteolytic peptides. Furthermore, kinetic analyses of CheR-receptor interactions would be facilitated by a soluble substrate that could be assayed at substantially higher concentrations than those achievable with receptor-enriched membrane fractions. However, previous studies on the methylation of cytoplasmic domains of E. coli and S. enterica receptors showed C-terminal fragments to be poorly methylated compared to full-length receptors (17, 31). Nonetheless, we examined the methylation rates of both the C-terminal cytoplasmic domains (TM0429c, TM1143c, and TM1428c) and full-length transmembrane receptors (TM0429fl, TM1143fl, and TM1428fl) of T. maritima. Initial rates of methylation were determined with saturating concentrations of receptors using concentrations of T. maritima CheR at which rates were linearly proportional to enzyme concentration (Table (Table1).1). The data indicate that T. maritima receptors can be efficiently methylated in vitro at rates comparable to that previously reported for S. enterica receptor methylation at 30°C (1.1 mol CH3 · mol CheR−1 · min−1 ). As expected, for all substrates, the methylation rates for both full-length receptors and cytoplasmic domains increased as the temperature was increased up to 70°C. Interestingly, the methylation rates with the cytoplasmic domains (TM0429c, TM1143c, and TM1428c) were ~2-fold higher than the rates with the corresponding full-length receptors (TM0429fl, TM1143fl, and TM1428fl). We conclude that unlike the cytoplasmic fragments of E. coli and S. enterica receptors, T. maritima receptor cytoplasmic domains are methylated similarly to their full-length counterparts in in vitro methylation reactions.
The T. maritima genome encodes six different transmembrane chemoreceptors (TM0023, TM0429, TM0918, TM1143, TM1146, and TM1428), with three of these (TM0023, TM0918, and TM0429) possessing identical cytoplasmic domains. Studies were performed using TM0429c, TM1143c, TM1146c, and TM1428c, representative of the entire set of receptors in the T. maritima genome. In order to facilitate the identification of all methylation sites in these receptors, the cytoplasmic domains were methylated to saturation. At the steady-state levels of methylation achieved under the assay conditions, the stoichiometries of methylation were estimated to be 1.7, 2.8, 2.9, and 3.7 mol CH3/mol receptor for TM1143c, TM1428c, TM1146c, and TM0429c, respectively (Fig. (Fig.1).1). The calculated stoichiometries, though subject to significant error due to the estimation of receptor concentration, suggest that TM1143c, TM1146c, TM1428c, and TM0429c contain at least two, three, three, and four methylation sites, respectively, capable of in vitro methylation. It should be noted that the sites methylated in these in vitro assays are likely to represent a subset of the full array of sites that can be methylated in vivo. Additional sites might be exposed by the presence of chemoeffectors or generated by posttranslational deamidation of specific Gln residues. The T. maritima receptors used in these studies were expressed in the absence of CheB and CheD, enzymes that in other bacteria are known to deamidate specific Gln side chains, generating Glu residues which can then be methylated by CheR (6, 11, 15).
For identification of methylation sites, fully methylated and unmethylated cytoplasmic domain samples were prepared by performing in vitro methylation reactions in the presence and absence of S-adenosylmethionine. Parallel reactions with radiolabeled S-adenosylmethionine confirmed that maximum levels of methylation with stoichiometries similar to those in Table Table11 had been achieved in the samples prepared for analysis. Initial analysis of the MS/MS data indicated that each receptor contained several methylation sites. Manual interpretation of the MS/MS spectra from trypsin-digested samples prepared in the presence and absence of S-adenosylmethionine was sufficient to identify the exact residue methylated in most cases (Table (Table2).2). For instance, the peptide LQEISASTEEVTSR from TM1143c had a high Sequest Xcorr score (a measure of the number of peaks of common mass between observed and expected spectra) for methylation taking place at either Glu505 or Glu506 (the bold residues in the peptide). However, manual inspection of the MS/MS spectra revealed the presence of two peaks, m/z 591.43 and m/z 973.5, that can be attributed to methylation occurring at Glu505 and not Glu506 (Fig. (Fig.2).2). In some cases, however, peptides containing multiple methylation sites were too large for exact site determination. To resolve this matter, additional digestion of TM0429c and TM1428c tryptic peptides was performed with Glu-C, a protease that cleaves specifically at the C termini of unmethylated glutamate residues. The Glu-C digests allowed identification of three additional methylation sites, Glu401 and Glu596 in TM0429c and Glu310 in TM1428c (Table (Table22).
In total, 13 methylation sites were identified: 2 sites in TM1143c, 3 sites in TM1428c, and 4 sites in both TM1146c and TM0429c, corresponding well with the stoichiometries estimated from methylation assays, except for TM1146c, which had an additional site identified (Fig. (Fig.1).1). Integration of peaks of methylated and unmethylated peptides from assays performed in parallel indicated that the identified sites all exhibited significant levels of methylation ranging from ~30 to 70% (data not shown). Furthermore, high sequence coverage was obtained from the trypsin and trypsin/Glu-C digests for all receptors (TM0429c, 87%; TM1143c, 88%; TM1146c, 90%; and TM1428c, 85%) (Fig. (Fig.3A).3A). Notably, there was complete coverage of all EE, QE, and EQ pairs, suggesting that all potential methylation sites were observed in the MS/MS analysis.
Some methylation sites within E. coli and S. enterica chemoreceptors are encoded as glutamine residues. Deamidation of the side chains by CheB and CheD (6, 11, 15) converts these residues to glutamates, yielding additional sites for methylation. Based on similarities among the initially identified methylation sites in TM0429c, TM1143c, and TM1428c and inspection of the corresponding T. maritima receptor sequences, we predicted Gln274 and Gln498 of TM1143c and Gln499 of TM1428c to be sites of deamidation. Subsequently, we learned that B. R. Crane and colleagues (personal communication) had identified the same two Gln residues in TM1143c as sites of CheD-mediated deamidation. To determine if these confirmed and predicted sites of deamidation could be methylated once deamidated, we constructed deamidated receptors by substituting these glutamine residues with glutamates, yielding TM1143c(Q274EQ498E) and TM1428c(Q499E). These deamidated receptors were then methylated to saturation (data not shown) and analyzed by LC-MS/MS. Methylation was observed at E499 in TM1428c and at E274 but not E498 in TM1143c (Table (Table2).2). Lack of observed methylation at E498 does not preclude the possibility that this site may be methylated under some conditions in vivo. In total, 15 methylation sites were identified in the four T. maritima chemoreceptors. Additional deamidation sites not examined in this study have recently been reported (6).
In bacterial chemotaxis, the signaling effects of ligand binding are counterbalanced by reversible methylation of specific glutamate residues within the cytoplasmic domains of chemoreceptors, resetting the receptor signaling output back to the prestimulus level. While signaling and adaptation have been most extensively characterized in E. coli and S. enterica, T. maritima has emerged as a valuable organism for structural analysis of chemotaxis proteins (2, 4-6, 9, 21, 22, 29). In this study we have taken initial steps towards characterizing in vitro methylation of T. maritima receptors and have identified specific sites within the cytoplasmic domains that can be methylated by methyltransferase CheR in the absence of stimuli.
Including the previously determined deamidation sites (6) and the methylation and deamidation sites identified here, a total of 19 sites have been identified. Interestingly, the identified methylation and deamidation sites clustered to four regions within the receptors and in many cases were found to be arranged in tandem pairs spaced six to seven residues apart. Sequence alignments of the receptor cytoplasmic domains demonstrated that the first tandem pair (methylation region 2) mapped to a similar location for three of the four receptors (Fig. (Fig.3).3). However, a second tandem pair mapped to two different locations. Sites in TM0429 and TM1428 aligned with one another (methylation region 3), but the sites in TM1143 did not (methylation region 4). Mapping the methylation regions onto the crystal structure of the TM1143 cytoplasmic domain (20) demonstrated that methylation regions 2 and 3 lie spatially near one another, while methylation region 4 of TM1143 is positioned eight helical turns away in the C-terminal direction compared to methylation region 3 (Fig. (Fig.3B3B).
There were, however, several instances where identified sites were not found in tandem, with the majority of those cases limited to TM1146. In this receptor, all four identified methylation sites (Glu255, Glu284, Glu479, and Glu515) did not have a corresponding site pair. However, for methylation sites Glu479 and Glu515, this finding was similar to what was observed for methylation site Glu281 in TM1143 and Glu506 in TM1428. In those instances, only after alteration of specific glutamines to glutamates (Q274E in TM1143c and Q499E in TM1428c) were the paired deamidation/methylation sites identified (Table (Table2).2). Analysis of the TM1146 sequence shows two glutamine residues, Gln472 and Gln508, which align perfectly with previously determined deamidation sites from the other T. maritima receptors (Fig. (Fig.3A),3A), making both Gln472 and Gln508 possible deamidation sites which could potentially form tandem pairs with identified methylation sites Glu479 and Glu515. Chao et al. (6) identified consecutive deamidation sites (Gln282 and Gln283) in TM1428 that mapped upstream of methylation region 2 (Fig. (Fig.3).3). Similarly, methylation site Glu255 in TM1146 was also found to map to this region (methylation region 1). Despite the presence of glutamate pairs (Glu275, 276 and Glu289, 290 in TM1428 and Glu248, 249 in TM1146) flanking these methylation and deamidation sites, none were identified as methylation sites. It should be noted, however, that similar to methylation regions 2 and 3, mapping methylation regions 1 and 4 onto the TM1143c structure showed methylation region 1 to lie spatially near methylation region 4 on the opposite strand (Fig. (Fig.3B).3B). Clustering of the identified sites located within these four methylation regions to two distinct areas of the receptors likely facilitates the accessibility and efficiency with which the modifying proteins (CheB, CheD, and CheR) alter these sites.
Previous studies directly identifying sites of methylation in E. coli (11, 19) and S. enterica (27) receptors yielded the methylation consensus sequence Glx-Glx-X-X-Ala-Ser/Thr (modified residue is in bold). Since then, this consensus sequence has been used to predict putative methylation sites in chemoreceptors from other organisms (16) and as the basis for identification of methylation sites through indirect methods (10, 23, 33). The lack of direct chemical analyses has precluded assessment of the universality of this consensus sequence.
In this study, methylation sites in T. maritima receptors were directly identified by utilizing LC-MS/MS. A sequence alignment of the identified methylation and deamidation sites from this study and those recently reported (6) revealed a novel consensus methylation sequence, Ala/Ser-sm-X-Glx-Glu-X-sm-Ala/Ser (modified residues are in bold; sm represents a small amino acid [Gly, Ala, Ser, or Thr]), that appears as a tandem heptad repeat centered around the Glx-Glu pair and overlaps at the Ala/Ser residue (Fig. (Fig.4).4). A previous study analyzing receptor cytoplasmic sequences proposed that the original consensus methylation sequence be extended to include residues to the N-terminal side of the Glx-Glx pair (16), and this extended conservation appears to be a feature of the T. maritima consensus as well. Remarkably, only 2 (E284 and E515 in TM1146) of the 19 methylation and/or deamidation sites that were identified in T. maritima match the E. coli/S. enterica consensus sequence. Moreover, while methylation of E. coli/S. enterica receptors occurs strictly at the second Glx residue within the consensus sequence, methylation of T. maritima receptors occurs primarily at the first Glx within the consensus sequence (Fig. (Fig.4).4). Methylation of the first Glx residue within the consensus sequence has also been observed in photoreceptor HtrI of Halobacterium salinarum (23). At some T. maritima methylation and/or deamidation sites, modification occurs at the second Glu residue, and in two cases, both residues of the Glx-Glu pair are modified. No distinguishing features within the local primary sequence of the sites appear to correlate with the choice of the first, second, or both as the site(s) of modification.
In the past decade the determination and publication of sequences of entire genomes have increased exponentially. Currently, 339 genomes have been completed, and a staggering 1,502 genome projects are still in progress (http://www.genomesonline.org). As more and more genome sequences become available, comparative genomic analyses will serve as an important tool in identifying consensus sequences and mapping putative sites of posttranslational modification in homologous proteins within and among different organisms. Our results comparing the methylation sites in T. maritima chemoreceptors with previously identified sites in E. coli and S. enterica suggest that direct chemical analysis will be required for conclusive identification of sites of modification. The consensus sequence for methylation sites in T. maritima receptors is distinct from that for E. coli and S. enterica receptors. Even within the T. maritima receptors there are unexpected deviations from the well-conserved T. maritima consensus sequence. For example, methylation site 1 in TM0429 (Fig. (Fig.3A3A and and4)4) contains a Thr residue in position 1 of the consensus sequence instead of the Glu/Gln residue that is found in all other methylation sites. Because this site does not match the consensus sequence, its identification as a site of methylation would likely have been missed by methods other than direct chemical analysis of the entire receptor domain. When targeted mutagenesis of predicted sites is used to identify sites of posttranslational modification, sites that deviate significantly from the consensus may be missed. Importantly, the subset of sites identified by this route serves to reinforce the starting sequence and perpetuates a consensus sequence that may not be fully inclusive.
We expected that glutamine residues targeted by deamidation enzymes, once deamidated, would subsequently be methylated. Based on predictions and subsequent confirmation (6) of three sites of deamidation, receptors containing Gln-to-Glu substitutions were constructed and analyzed for methylation. Q499E of TM1428c and Q274E of TM1143c were observed to be methylated, but Q498E of TM1143c was not methylated under in vitro methylation conditions. The lack of methylation at a confirmed deamidation site emphasizes the limitation of using in vitro reactions for identification of modification sites. It is likely that specific in vivo conditions and/or environmental cues may be required for utilization of all modification sites. For example, in Bacillus subtilis, methylation site selection is coordinated by the interaction of CheY with receptors (14, 33). The absence of CheY as well as the rest of the chemotaxis proteins, full-length receptors, and chemoeffectors in our assays might explain the lack of observed methylation at Q498E. It is important to note that the methylation sites identified by in vitro modification are likely to represent a minimum subset of sites available for methylation in vivo.
In summary, the methylation sites identified within T. maritima receptors have established a consensus sequence that differs from the previously identified consensus sequence for E. coli and S. enterica receptors. Knowledge of the methylation sites in T. maritima receptors provides an important foundation for structural characterization of different receptor signaling states. Additionally, these findings provide cautionary anecdotal evidence that consensus sequences for posttranslational modifications in one bacterium may not necessarily be applicable to analogous modifications in other species.
This work was supported by grants from the National Institutes of Health to A.M.S. (R37GM047958) and to Peter Lobel for shared instrumentation (RR017992). E.P. was supported by an NIH training grant (T32 GM08360). A.M.S. is an investigator of the Howard Hughes Medical Institute.
We thank Peter Lobel for valuable discussions of and contributions to the mass spectrometry analysis and for critical reading of the manuscript. We thank Bryan Beel for plasmid constructs and Brian Crane for plasmid constructs, coordinates of TM1143c, and communication of data prior to publication.