|Home | About | Journals | Submit | Contact Us | Français|
Rickettsia prowazekii, the etiologic agent of epidemic typhus, is a potential biological threat agent. Its outer membrane protein B (OmpB) is an immunodominant antigen and plays roles as protective envelope and as adhesins. The observation of the correlation between methylation of lysine residues in rickettsial OmpB and bacterial virulence has suggested the importance of an enzymatic system for the methylation of OmpB. However, no rickettsial lysine methyltransferase has been characterized. Bioinformatic analysis of genomic DNA sequences of Rickettsia identified putative lysine methyltransferases. The genes of the potential methyltransferases were synthesized, cloned, and expressed in Escherichia coli, and expressed proteins were purified by nickel-nitrilotriacetic acid (Ni-NTA) affinity chromatography. The methyltransferase activities of the purified proteins were analyzed by methyl incorporation of radioactively labeled S-adenosylmethionine into recombinant fragments of OmpB. Two putative recombinant methyltransferases (rRP789 and rRP027-028) methylated recombinant OmpB fragments. The specific activity of rRP789 is 10- to 30-fold higher than that of rRP027-028. Western blot analysis using specific antibodies against trimethyl lysine showed that both rRP789 and rRP027-028 catalyzed trimethylation of recombinant OmpB fragments. Liquid chromatography-tandem mass spectrometry (LC/MS-MS) analysis showed that rRP789 catalyzed mono-, di-, and trimethylation of lysine, while rRP027-028 catalyzed exclusively trimethylation. To our knowledge, rRP789 and rRP027-028 are the first biochemically characterized lysine methyltransferases of outer membrane proteins from Gram-negative bacteria. The production and characterization of rickettsial lysine methyltransferases provide new tools to investigate the mechanism of methylation of OmpB, effects of methylation on the structure and function of OmpB, and development of methylated OmpB-based diagnostic assays and vaccine candidates.
Rickettsial OmpBs belong to the outer membrane protein autotransporter family and contain large passenger domains. Rickettsial OmpBs have been shown to participate in adhesion to mammalian cells in vitro (8, 36). Ectopically expressed rickettsial OmpB in Escherichia coli has been demonstrated to be sufficient in mediating bacterial attachment and invasion of cultured mammalian cells (8, 24). Unlike the α-helical transmembrane proteins in plasma membranes, the autotransporter domains of these outer membrane proteins have structural characteristics of β-barrel integral membrane proteins (3, 33). As the immunodominant antigenic surface protein, native OmpB induces strong humoral and cellular immune responses in animal models and in patients (6, 12, 15, 16, 23). Methylation of OmpA of Rickettsia canadensis and ompB of Rickettsia typhi is known, but whether this is important for any other pathogenic rickettsia is unknown.
One important but not well understood structural feature of rickettsial OmpB is the methylation at ε-amino groups of lysine residues (28). Posttranslational modification of proteins by methylation can potentially alter their structure and function. Biochemical and genetic analyses suggest that methylation of lysine residues of OmpB correlates with the virulence of Rickettsia prowazekii (30). OmpBs from the virulent strains Breinl and Evir are found hypermethylated, while that of the avirulent Madrid E strain is mostly mono-methylated (11, 29). In addition, hypomethylated OmpB from the Madrid E strain elicits less protective immunity than OmpB from virulent strains of R. prowazekii.
The genes of the avirulent R. prowazekii Madrid E strain have been compared with those of the virulent revertant Evir strain to identify the genes that are inactivated in the attenuated Madrid E strain but not in the virulent Breinl and Evir strains. Genomic sequence and transcription analyses showed that the mutation of RP027/RP028 in the Madrid E strain reverts back to the wild type (39). The gene encoding one of the putative methyltransferases RP027/RP028 was found to have a frameshift mutation in the avirulent strain, resulting in two split genes, the RP027 and RP028 genes, which are detectable only in avirulent strains (4). The putative methyltransferases and their native substrates in Rickettsia have not been identified or characterized, and the mechanism of action of protein lysine methyltransferases in the virulence and the pathogenesis of Rickettsia is not well understood.
Lysine methylation is a relatively stable posttranslational modification and regulates various functions of diverse proteins (13, 18). The regulatory potential of lysine methylation is greatly expanded because multiple methyl groups can be added to a single lysine residue, leading to mono-, di-, and trimethylation. Methylation of specific residues in histones plays a pivotal role in chromatin remodeling that leads to alteration of gene expression and repression (27). Methylation of membrane proteins has attracted little attention. OmpB methylation at lysine residues could conceivably affect bacterium-host cell interactions and inhibit alternative posttranslational modifications, such as acetylation, ubiquitination, or sumoylation. Characterization of methyltransferases is essential to better understand the structure and function of native OmpB, the effects of hypermethylation, and their role in the host immune response. In practical clinical applications, enzymatic methylation of OmpB in vitro may provide methylated OmpB for the detection of patient antibodies against virulent rickettsiae and for developing improved vaccine candidates. As a first step toward these goals, we searched for methyltransferases from R. prowazekii that can enzymatically methylate OmpB (10). Genes encoding hypothetical proteins and putative methyltransferases in the genome of R. prowazekii were examined by bioinformatics. Recombinant hypothetical proteins and putative methyltransferases were cloned, purified, and analyzed for methyltransferase activity. Two rickettsial protein lysine methyltransferases that enzymatically methylate recombinant OmpB were found in this study.
Recombinant OmpB fragments, including rOmpB(A) (33 to 273), rOmpB(AN) (33 to 744), and OmpB(K) (745 to 1353) based on the genomic sequence of R. typhi were prepared as previously described (10). Briefly, the bacterially expressed proteins in the inclusion body were dissolved in 8 M urea and 1% deoxycholate, purified by ion-exchange chromatography on DEAE-cellulose in 6 M urea, and refolded. Proteins were obtained by slow dialysis at stepwise, decreasing concentrations of urea. SET7 methyltransferase and the histone peptide substrate H3(1–17) were from New England Biolabs. N,N′-diallyltartardiamide (DATD) was from Bio-Rad.
Annotated genomic sequences of R. prowazekii Madrid E and R. prowazekii RP22 in NCBI and annotated encoded proteins in PIR (Protein Information Resources) were searched and examined for potential protein methyltransferases. Those methyltransferases targeting DNA, rRNA, tRNA, mRNA, and small molecule metabolites were eliminated. The remaining genes and encoded proteins that are annotated as hypothetical proteins with potential S-adenosylmethionine (SAM) binding domains were selected for further analysis. Structural bioinformatic analysis of the putative methyltransferases was carried out using the software programs HHpred (34) and 3DLigandSite (37). The structural templates were selected by the software at respective Web servers. Successful structural modeling was obtained only with the N-terminal half of the putative methyltransferases. No known three-dimensional structures with significant homology to the C-terminal half are available. The structural models were displayed using the software program Chimera. The secondary structural prediction of OmpB was carried out using the Web servers Psipred (19), HHpred (34), and Phyre (20).
Table 1 lists the names and accession numbers of genes and encoded proteins of putative methyltransferases. Codon-optimized single-stranded cDNAs were synthesized by Bioclone Inc. (San Diego, CA) and subcloned at NdeI/XhoI restriction sites into the pET28a expression vector (Novagen) containing a 6×His tag. Standard cloning procedures were performed, and the inserts were verified by restriction enzyme digestion and sequencing. The sequences of the cDNAs for the expression in E. coli are shown in Table S1 in the supplemental material.
BL21(DE3) competent cells (Agilent Technologies) were transformed with pET28a carrying one of the genes encoding hypothetical proteins and putative methyltransferases. Protein expression was carried out by growing the bacteria from single colonies at 37°C in 2 ml of LB broth supplemented with 100 μg/ml kanamycin overnight. The cell culture was transferred to fresh LB broth with kanamycin and incubated with shaking until the optical density at 600 nm (OD600) reached 0.3 to 0.5. The cultures were induced with 0.04 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and incubated at 22°C and at 250 rpm overnight. Cells were harvested by centrifugation and lysed, and the expressed proteins were purified using nickel-nitrilotriacetic acid (Ni-NTA) column chromatography according to the manufacturer's instructions (GoldBio, St. Louis, MO). Protein concentrations were determined using the Bio-Rad protein assay and bovine serum albumin as a standard (7).
Methyltransferase activity was monitored by the transfer of 3H-methyl (3H-Me) from S-[3H-Me]adenosylmethionine (10 Ci/mmol; Perkin Elmer) to recombinant protein substrates. The purified recombinant methyltransferases were then examined for protein methyltransferase activity using recombinant OmpB fragments as substrates. We first developed a methyltransferase assay modified from the SET7 methyltransferase assay. The amount of methyl transferred from S-[3H-Me]adenosylmethionine to histone 3 peptide was determined by Whatman DE81 immobilized radioactivity after extensive washing with sodium bicarbonate (38). Unlike histones, the OmpB fragments are acidic proteins, so Whatman 3MM and 5% trichloroacetic acid (TCA) replacing DE81 and sodium bicarbonate were used. The rate of methyl transfer from S-[3H-Me]adenosylmethionine to histone 3 peptide was found to be the same as that using the conventional SET7 assay.
The 50-μl reaction mixtures contained 8.3 mM sodium phosphate, pH 8.0, 160 μM [3H-Me]SAM (34 mCi/mmol), and 2 μM rOmpB(AN) or rOmpB(K). The methylation reaction was initiated by adding limiting amounts of protein methyltransferase and incubated at 37°C. Aliquots of the reaction mixture were spotted onto Whatman 3MM cellulose filter paper discs (Fisher Scientific) at various time points. Reactions were stopped by the addition of 5% TCA. The paper discs were washed with 5% ice-cold TCA three times, followed by ethanol-ether (1:1 [vol/vol]). The amounts of acid-precipitable radioactivity were determined by liquid scintillation counting using a Perkin Elmer Wallace 1410 counter. One unit of methyltransferase is defined as the amount of enzyme required to catalyze the transfer of 1 pmol of methyl group to the substrate in 10 min.
Radioactivity associated with methylated OmpB fragments separated by SDS gel electrophoresis was determined using the method previously described with minor modifications (35). To quantitate the number of methyl groups transferred to each substrate molecule, the radioactivity associated with the OmpB fragments AN and K were measured using the following method. Protein substrate (3 μM) was incubated with rRP789 or rRP027-028 (6 μM) in the presence of radioactive SAM (0.16 mM with specific activity of 68 mCi/mmol) in 8.3 mM phosphate buffer (pH 8.0) at 37°C overnight. The reaction was stopped by adding SDS-sample buffer. Polyacrylamide gels were polymerized with reagents obtained from Bio-Rad Laboratories (Richmond, CA) except that N,N′-methylenebisacrylamide was replaced with DATD at a ratio of 1 part of DATD to 10 parts of acrylamide (35). After electrophoresis, the SDS gels were stained with Coomassie blue, followed by extensive destaining (Bio-Rad). Appropriately sized slices of the gel were placed in glass scintillation vials, and 0.5 ml of 2% (wt/vol) sodium metaperiodate was added. The vials were shaken for 30 min to dissolve the gel. Ecoscint A (6 ml; National Diagnostics, Atlanta, GA) was added to the vial, and the vial was cooled on ice for 20 min and then counted by liquid scintillation counting.
The enzymatic methylation reaction mixtures as described above were incubated at 37°C for 4 h. Protein samples were denatured in standard sample denaturing buffer containing 1% SDS and 10 mM mercaptoethanol. The reaction products were separated by SDS-PAGE and followed by Western blot analysis with 250 μg/ml rabbit anti-trimethyl lysine (Trend Pharma & Tech Inc., Surrey, Canada) at a 1:2,000 dilution in Odyssey blocking buffer (phosphate-buffered saline and 0.1% sodium azide) containing 0.1% Tween 20 at 4°C overnight, followed by incubation with IRDye (800CW)-conjugated anti-rabbit IgG(H+L) (donkey) secondary antibody (Rockland Immunochemicals, Inc., Gilbertsville, PA) at a 1:10,000 dilution in blocking buffer for 1 h at room temperature. The signal was detected using an Odyssey infrared imaging system (Li-Cor) after washing 3 times with 1× PBS containing 0.1% Tween 20.
The Michaelis-Menten constants and catalytic constants were obtained by direct fitting of the Michaelis-Menten equation to the initial rate dependence of the substrate concentrations using the software program Kaleidagraph.
Ten micrograms each of the fragments rOmpB(AN) and rOmpB(K) were methylated using 30 μg of methyltransferase–3.2 mM S-adenosylmethionine (Perkin Elmer) in 50 μl of reaction mixture containing 8.3 mM sodium phosphate buffer, pH 8.0. The methylation reaction was carried out at 37°C overnight. The reaction samples were evaporated down to 20 μl using a Speedvac and mixed with SDS sample buffer. The reaction mixture components were separated by SDS-PAGE, and the corresponding AN and K protein bands were cut off from the gel and subjected to in-gel digestion as described previously (21), with modifications. Briefly, the samples were washed using 50% methanol and 5% acetic acid, followed by the reduction using 10 mM dithiothreitol (DTT). The protein samples were alkylated with 100 mM iodoacetamide in the dark. The gel pieces were dehydrated using acetonitrile and rehydrated in 100 mM (NH4)HCO3 twice. The protein digestion was carried out using sequencing-grade chymotrypsin (Roche, Madison, WI). One microgram of chymotrypsin in 50 mM (NH4)HCO3 was mixed with 10 μg of protein sample and digested overnight at 25°C. The digested peptides were extracted in 50% (vol/vol) acetonitrile and 5% (vol/vol) formic acid or acetic acid. The volume was reduced to less than 20 μl by evaporation, and the final volume was adjusted to 20 μl using 1% formic acid. The samples were purified using Zip-Tip pipette tips with C18 resin (Millipore, Billercia, MA) according to the manufacturer's protocol before liquid chromatography-tandem mass spectrometry (LC/MS-MS) analysis. LC/MS-MS was carried out at the NHLBI Proteomics Core Facility (Bethesda, MD) using an Eksigent nanoLC-Ultra 2D system (Dublin, CA) coupled to an LTQ Orbitrap Velos mass spectrometer (Thermo Scientific, San Jose, CA). Raw data files generated by the LTQ Orbitrap Velos instrument were analyzed using the Proteome Discoverer v1.3 software program (Thermo Scientific) and the Mascot search engine running on a six-processor cluster at NIH (http://biospec.nih.gov; version 2.3).
In order to identify the protein methyltransferases of OmpB, we examined the annotations of the Rickettsia prowazekii Madrid E genomic sequence at National Center for Biotechnology Information (NCBI). Sixty-two genes encoding known and potential methyltransferases were found (2). Those genes that encode methyltransferases of small molecule metabolites, tRNA, mRNA, rRNA, and DNA were excluded. All remaining genes that encode unknown and putative methyltransferases were selected. Table 1 shows the genes that encode unknown methyltransferases from R. prowazekii Madrid E (RP789, RP527, RP545, RP027, and RP028).
In order to identify the protein methyltransferase of OmpB in virulent strains of Rickettsia, we examined the genome of R. prowazekii Rp22, which is a virulent strain closely related to P. prowazekii Madrid E. The sequence of another virulent strain, Breinl, is not currently available. A putative methyltransferase with the protein locus ADE29537 was found in the R. prowazekii Rp22 genome and is included in the study. The protein ADE29537 is referred to as RP027-028 here. The amino acid sequence of RP027-028 aligns with those of RP028 and RP027 with 100% identity at its N terminus (1 to 243) and C terminus (283 to 535), respectively. RP028 and RP027 in Madrid E appear to be split fragments resulted from a frameshift mutation of RP027-028 (39). Table 1 lists the accession number, the protein locus, and the number of amino acid residues in the encoded protein. None of these putative methyltransferases have been cloned or biochemically characterized for any of the rickettsial species.
Structural bioinformatics analyses were carried out. Amino acid sequence alignment of RP789 and RP027-028 showed 45% identity and 65% homology. Molecular modeling of the N-terminal domains of RP789 and RP027-028 using 3DLigandSite (37) gave similar three-dimensional models of the SAM binding domains of RP789 and RP027-028 with bound SAM (Fig. 1), suggesting that both are SAM binding proteins. The C termini of RP789 and RP027-028 showed 38% identity and do not have substantial homology to any proteins in the RCSB Protein Database suitable for homology modeling.
Proteins homologous to RP789 and RP027-028 were found in the NCBI database. More than 40 proteins from various rickettsial species and strains were compiled. The accession numbers and identities of these homologous proteins to RP789 and RP027-028 are shown in Table S2 in the supplemental material.
Based on the gene sequences of the above putative methyltransferases, we produced the corresponding proteins in E. coli. E. coli codon-optimized genes that encode potential methyltransferases were synthesized and inserted into the expression vector pET28a. The corresponding proteins were expressed in E. coli BL21(DE3) and purified using Ni-NTA affinity chromatography. The level of expression and the purity of the proteins were analyzed by SDS-polyacrylamide gel electrophoresis. rRP789 and rRP027-028 were highly expressed as soluble proteins and successfully purified as shown in Fig. S1 in the supplemental material. Purified rRP789 and rRP027-028 had molecular masses of 63.5 and 61.5 kDa, respectively, as expected. The overall yields of purified rRP789 and rRP027-028 were 0.25 and 0.75 mg per 100 ml of culture, respectively. rRP527 was expressed at a lower level than rRP789 or rRP027-028. Purified rRP527 had the expected molecular mass of 50 kDa and had an overall yield of 0.15 mg per 100 ml of culture (see Fig. S1). rRP545 had a very low expression level. The expression level of rRP545 was not significantly improved by increasing the concentrations of IPTG to 0.5 or 1 mM, varying lengths of induction time, or changing the host to BL21(DE3) Gold pLysS. rRP028 and rRP027 were expressed at high levels but were insoluble. Purification of rRP028 and rRP027 in the presence of 8 M urea, followed by refolding by dialyzing against stepwise-decreasing concentrations of urea, did not yield soluble protein.
The radioactivity assay was used to examine the enzymatic methylation catalyzed by purified recombinant, potential methyltransferases using recombinant OmpB fragments. Three rOmpB fragments, AN, K, and A, were used and are referred to as rOmpB(AN), rOmpB(K), and rOmpB(A), respectively. As shown in Fig. 2A, rRP789 catalyzed 3H-Me transfer from S-[3H-Me]adenosylmethionine to rOmpB(AN) or rOmpB(K). rOmpB(AN) was methylated by rRP789 at a higher rate and to a greater extent than rOmpB(K). No methyl transfer was detected in the absence of either rRP789 or the rOmpB fragment. Similarly, as shown in Fig. 2B, rRP027-028 also catalyzed 3H-Me transfer from S-[3H-Me]adenosylmethionine to rOmpB(AN) and (K). No methyl transfer was detected in the absence of either rRP027-028 or the recombinant OmpB fragment. The rate of methyl transfer catalyzed by rRP027-028 to rOmpB(AN) and rOmpB(K), however, was appreciably slower than those catalyzed by rRP789. rOmpB(A) was not methylated by either rRP789 or rRP027-028 under the same assay conditions (data not shown). The specific activity of methyltransferase rRP789 was determined from the initial rate of methylation in the presence of excess of rOmpB(AN) and rOmpB(K). As shown in Table 2, the specific activities of methyltransferase rRP789 were 56,120 U/mg and 24,500 U/mg toward rOmpB(AN) and rOmpB(K), respectively. The rate of rRP027-028-catalyzed methylation of rOmpB fragments was very slow. The specific activity was determined at comparable enzyme and substrate concentrations, and the apparent specific activities toward rOmpB(AN) and rOmpB(K) thus determined were 1,628 and 1,944 U/mg, respectively. The methylation of rOmpB fragments by rRP789 and rRP027-028 clearly established their methyltransferase activity.
Purified rRP527 at 13 μM did not catalyze detectable methylation of either rOmpB(AN) or rOmpB(K) (data not shown). The methyltransferase activity of rRP527 toward rOmpB(AN) or rOmpB(K) was undetectable and remained undetectable at various pHs from 5 through 9. The very small amounts of pure soluble rRP545, rRP027, and rRP028 precluded meaningful determination of methyltransferase activity. No further studies of rRP527, rRP545, rRP027, and rRP028 were carried out.
Protein methyltransferases transfer methyl groups to lysine and arginine residues in proteins. In native OmpB from R. prowazekii, only lysine methylation was reported. For lysine residues, up to three methyl groups may be transferred to each lysine. Western blot analyses using antibodies specific to trimethyl lysine were carried out to determine the types of methylation in OmpB(AN) and OmpB(K) catalyzed by rRP789 and rRP027-028. Due to the closeness of the molecular weight of OmpB K to that of the methyltransferases, methyltransferases were removed from the reaction mixtures before Western blot analysis. Both rRP789 and rRP027-028 contained 6×His tags, while recombinant rOmpB(AN) and rOmpB(K) did not. rRP789 and rRP027-028 in the reaction mixtures were removed by Ni-NTA. Figure 3 shows the Western blot of the methylation products of rOmpB(AN) and rOmpB(K) catalyzed by rRP789 and rRP027-028. Antibodies against trimethyl lysine reacted with OmpB(AN) only after methylation by either rRP789 or rRP027-028. Recombinant OmpB(AN) or OmpB(K) or rRP789 and rRP027-028 alone did not react with the antibodies against trimethyl lysine. Thus, the Western blot analyses confirmed the protein lysine methyltransferase activity of rRP789 and rRP027-028 and demonstrated their activity of catalyzing trimethylation of lysine residues in rOmpB fragments.
The molar ratios of incorporated methyl group per protein molecule were assessed. The molar ratios at saturating levels of methylation corresponded to 10 methyl groups each per AN or K molecule for rRP789 (Table 2). To ensure that the radioactivity was associated with rOmpB fragments, we extracted the corresponding protein bands from SDS gel after overnight methylation of rOmpB(AN) and rOmpB(K). Similar molar ratios were obtained (Table 2). The high molar ratios are consistent with methylation at multiple sites in rOmpB(AN) and rOmpB(K) by rRP789. For rRP027-028, because no plateau of methylation could be reached and catalytic efficiency was very low, the limiting molar ratio cannot be reliably determined. For the rOmpB(K) fragment, detectable trimethylation by rRP789 or rRP027-028 was not observed by Western blot analysis under the same conditions.
The enzyme kinetic parameters of rRP789 are summarized in Table 3. The Michaelis-Menten constant for S-adenosylmethionine was slightly different dependent on whether the protein substrate rOmpB(AN) or rOmpB(K) was used. The Michaelis-Menten constants for rOmpB(AN) and rOmpB(K) were different. The causes of the difference are not known at present, but the difference could reflect structural differences between rOmpB(AN) and rOmpB(K). The effects of KCl, MgCl2, and pH on rRP789 activity were examined. As shown in Fig. S2 in the supplemental material, KCl inhibited the initial rates of enzymatic methylation of both rOmpB(AN) and rOmpB(K). Similar to KCl, MgCl2 (1 to 10 mM) also inhibited rRP789 (data not shown). The effects of pH on enzymatic methylation are shown in Fig. S3. The optimal pH was close to pH 8, which is lower than that for SET7 methyltransferase.
Unlike that with rRP789, rRP027-028-catalyzed methylation of rOmpB fragments was very slow. The reason for the slowness of the catalysis is not known. Attempts were made to search for possible cofactors that might enhance the catalytic efficiency. Metal ions are cofactors for numerous enzymes. Three divalent cations, Zn(II), Cu(II), and Ni(II), were tested. Inhibition of methylation was found for all three metal ions. None of the metal ions stimulated rRP027-028 methyltransferase activity (see Fig. S4 in the supplemental material). Addition of dithiothreitol (DTT) did not enhance the initial rate of methylation up to 30 min but did enhance the level of methylation after 30 min. In the absence of DTT, the time course gradually plateaued, but it was linear in the presence of DTT. The difference at 120 min is shown elsewhere (see Fig. S4). The rRP027-028 activity with the addition of DTT remained at the same order of magnitude as that in the absence of DTT. No synergistic effect was observed when both rRP789 and rRP027-028 were included in the assay mixtures to methylate OmpB fragments (data not shown). We were unable to carry out Michaelis-Menten kinetic studies of rRP027-028 under conditions where the substrate in large excess of the enzyme is required, because high concentrations of rRP027-028 were required to obtain appreciable methylation of rOmpB.
Since recombinant OmpB AN and K used in the present study were from R. typhi, while rRP027-028 was from R. prowazekii, the slow catalysis could be due to the nonhomologous substrate, which may be not a suitable substrate of the protein methyltransferase from another species. The rRP027-028 ortholog from R. typhi shows 93% identity to rRP027-028 from R. prowazekii. The recombinant ortholog from R. typhi (hypothetical protein RT0101; accession number YP_067069) was prepared by the same method as that used for rRP027-028. R. typhi rRP027-028 showed specific methyltransferase activity toward rOmpB(AN) and rOmpB(K) within 2-fold of that of R. prowazekii rRP027-028 (data not shown). Similar to R. prowazekii rRP027-028, no stimulation by Zn(II), Cu(II), or Ni(II) of the initial rates of methylation of rOmpB(AN) by R. typhi rRP027-028 was found.
Thus far, only rOmpB(AN) and rOmpB(K) were found to be methylated by rRP789 and rRP027-028. In order to determine the substrate specificities of rRP789 and rRP027-028, additional substrates of rRP789 and rRP027-028 were searched for using E. coli cell lysate. Neither rRP789 nor rRP027-028 showed significant methyltransferase activity toward E. coli proteins (data not shown). E. coli lysate and supernatant did not show greater methyl incorporation in the presence of rRP789 and rRP027-028 than in the absence of rRP789 and rRP027-028. The lysate showed methyl incorporation above the baseline, likely due to the presence of endogenous methyltransferases and substrates in the E. coli lysate. Calf thymus histones could not be methylated to any detectable level by rRP789 and rRP027-028 as determined by the radioactivity assay under standard assay conditions (data not shown).
A total of 28 lysine residues in rOmpB(AN) are potential methylation sites, and each can be methylated to mono-, di-, and trimethyl lysine. To determine the sites and types of methylation catalyzed by rRP789, we prepared methylated rOmpB(AN). rOmpB(AN) was methylated by rRP789 under standard assay conditions, and the methylated rOmp(AN) was digested by chymotrypsin. The chymotryptic peptides of methylated rOmpB(AN) were analyzed by LC/MS-MS. Table 4 shows the methylated-lysine residue number, the type of methylation, and the neighboring amino acid sequences. As shown in Table 4, the methyltransferase rRP789 catalyzed mono-, di-, and trimethylation at 14 of the 28 lysines in rOmpB(AN). Of the 14 Lys residues, 14 had monomethyl lysine, nine contained dimethyl lysine, and five were trimethyl lysine. No methylated Lys was detected in control samples of rRP789 and rOmpB(AN). No unique consensus amino acid sequence for the methyltransferase methylation can be discerned at present.
The sites and types of methylation catalyzed by rRP027-028 were analyzed similarly to that catalyzed by rRP789. Table 5 shows the rRP027-028 methylated-lysine residue number, the type of methylation, and the neighboring amino acid sequence. As shown in Table 5, the methyltransferase rRP027-028 catalyzed exclusively trimethylation of rOmpB(AN). A total of seven lysine residues were trimethylated by rRP027-028. All seven sites showed only trimethylation. No mono- or dimethyl lysine residues were detected in rOmpB(AN) that was methylated by rRP027-028. Again, no unique consensus amino acid sequence can be discerned at present.
To confirm the methylation patterns exhibited by rRP789 and rRP027-028, we analyzed the patterns of methylation of rOmpB(K) by rRP789 and rRP027-028 by LC/MS-MS, as was used for rOmpB(AN). Table 6 shows the methylation sites, neighboring sequences, and types in rOmpB(K) with rRP789 (A) or rRP027-028 (B). Again, rRP789 catalyzed mono- and dimethylation of rOmpB(K), while rRP027-028 catalyzed only trimethylation. The rRP789 methyltransferase showed monomethylation at six sites and dimethylation at one lysine residue, while rRP027-028 showed trimethylation at both lysine residues.
Two different protein methyltransferases, rRP789 and rRP027-028, that methylated rOmpB(AN) and rOmpB(K) were found. Of the two methyltransferases, rRP789 is from the avirulent R. prowazekii Madrid E strain, while rRP027-028 is from the virulent strain R. prowazekii Rp22. An ortholog of rRP789 can be found in the virulent R. prowazekii Rp22 genome (ADE30353.1), which has a single conservative substitution, from Glu to Asp. rRP789 and rRP027-028 may complement each other in catalyzing the multimethylation of Lys residues in OmpB in the virulent strains. To our knowledge, rRP789 and rRP027-028 are the first biochemically characterized protein methyltransferases of outer membrane proteins from Gram-negative bacteria. As shown in Table S2 in the supplemental material, the two methyltransferases are conserved in various genomes of Rickettsia. We propose naming the RP789 and RP027-028 methyltransferases PKMT1 (protein lysine methyltransferase) and PKMT2, respectively.
The specific activity of rRP789 compares favorably with those of reported lysine methyltransferases such as SET7, which methylates the histone H3 peptide at K9 with a specific activity of 30,000 U/mg. The specific activity of rRP027-028 was 10- to 30-fold lower than that of rRP789. The reason for the low catalytic efficiency of rRP027-028 is not clear at present. The existence of additional modification or cofactors of rRP027-028 cannot be excluded at present. Addition of E. coli proteins did not alter rRP027-028 methyltransferase activity, suggesting that no cofactors were present in E. coli. Addition of divalent cations, such as Zn(II), Ni(II), and Cu(II), did not appreciably enhance methyltransferase activity, suggesting that none of these metal ions could be expected to be a cofactor. The possibility that the use of a substrate from a nonhomologous species caused the low efficiency was eliminated, since comparable low catalytic efficiency was observed even when OmpB and rRP027-028, both from R. typhi, were used.
As analyzed by LC/MS-MS, rRP789 catalyzed mono-, di-, and trimethylation, while rRP-027-028 catalyzed exclusively trimethylation. The qualitative differences in the reaction products for rRP789 and rRP027-028 suggest their different reaction mechanisms. In order to exclusively catalyze trimethylation of lysine, rRP027-028 had to remain bound to the methylation target site of the substrate protein during the entire course of three rounds of methylation of the lysine residue. On the other hand, the methyltransferase rRP789 could dissociate from the target sites of the substrate protein after the addition of a methyl group to a lysine or methylated lysine residue. The trimethylation activity of rRP027-028 is consistent with the notion that RP027-028 accounts at least in part for the hypermethylation of OmpB and consequently the virulence of the virulent strains. LC/MS-MS analysis showed that rOmpB(K) can be trimethylated by rRP027-028, but Western blot analysis did not. The difference could be due to the accessibility of trimethylated lysine in rOmpB(K) and reduced reactivity by the antibodies.
One potential factor that may contribute to the large difference in the catalytic efficiencies between rRP789 and rRP027-028 is different reaction mechanisms. Appreciable differences in the catalytic efficiencies between processive and distributive trimethylation mechanisms have been reported (14). In distributive methylation, methyltransferase dissociates from the protein substrate after the transfer of each methyl group. In processive methylation, methyltransferase transfers methyl groups to the same lysine residue consecutively in three rounds without dissociation from the protein substrate. For each round of methyl transfer, processive enzyme must dissociate S-adenosylhomocysteine and rebind S-adenosylmethionine. The structural similarity between S-adenosylhomocysteine and S-adenosylmethionine would slow down dissociation of S-adenosylhomocysteine and binding of S-adenosylmethionine. The kinetics of rRP027-028 likely follows a processive mechanism.
The molar ratios of methyl groups per OmpB fragment molecules showed that OmpB fragments were methylated at multiple sites. The radioactivity assay provided the average number of methyl groups per OmpB(AN) or OmpB(K). LC/MS-MS analyses of methylated OmpB fragments identified the sites of methylation. The number of sites identified by LC/MS-MS was clearly more than the average number of methyl groups as determined by the radioactivity assays. The cause of the apparent differences is not clear at present. One likely scenario would be that methylated OmpB was microheterogeneous in the number and sites of methylation due to varying degrees and sites of methylation among different substrate protein molecules. The multisite methylation of OmpB catalyzed by the two methyltransferases rRP789 and rRP027-028 is unusual among protein methyltransferases. Most known protein methyltransferases transfer a single methyl group at a specific site of substrate proteins. Multisite methylation opens the possibility of the processive binding by these methyltransferases to consecutive target sites without dissociating from the OmpB molecule. Site hopping has been observed in DNA methyltransferases in epigenetic gene regulation but not in protein methyltransferases.
No unique consensus amino acid sequences of the methylation sites that were targeted by rRP789 or rRP027-028 were found. OmpB homologous proteins occur in all species of Rickettsia, have been used in their phylogenetic analysis (31), and may play roles in rickettsial evolution (5). The roles of the three-dimensional structure in the substrate recognition strategy of the methyltransferases are yet to be explored. The three-dimensional structure of the passenger domain of OmpB is yet to be determined. Inasmuch as rOmpB(AN) and rOmpB(K) are the only active substrates of rRP789 and rRP027-028 found so far, the three-dimensional structure of OmpB likely plays roles in substrate recognition and antibody epitopes. The rOmpB(AN), rOmpB(K), and rOmpB(A) are parts of the passenger domain of OmpB. The observation that five lysine residues in rOmpB(A) failed to be enzymatically methylated but were methylated in rOmpB(AN) (Table 4) could be due to the absence of properly folded tertiary structure in rOmpB(A), unlike that in rOmpB(AN). Prediction of the secondary structure of OmpB by Psipred (19) suggests predominantly alternating short β-strands and loops in the passenger domain. As shown in Tables 4, ,5,5, and and6,6, the secondary structures at the methylation sites of rRP789 and rRP027-028 were predicted to be either coil or β-strand. The passenger domain has significant identity in sizable segments to known β-helical proteins, such as polysialic acid O-acetyltransferase (PDB 2WLG). The β-helices are characterized by parallel β-strands associating in a helical pattern with three faces (26). Similar parallel β-sheets have been found in prions (17, 22) and amyloid cross-β spine (25, 32). The roles of such β-helical structure in the pathogenesis of Rickettsia are yet to be explored.
OmpBs are highly acidic proteins. Methylation of Lys residues enhances the partial charges of ε-amino groups and hence may affect its electrostatic interactions and hydrogen bonding in protein-protein interactions and consequently the function of OmpB. OmpB has been shown to mediate rickettsial attachment to and invasion of endothelial cells (8). It is conceivable that methylation of multiple Lys residues in OmpB may alter its interaction with the host cell surface and the attachment to and invasion of host cells by the bacteria. After entering host cells, its interactions with host intracellular proteins may also alter. Inasmuch as the virulence of R. prowazekii appears to associate with rRP027-028 (39), methylation of OmpB likely facilitates the host attachment and invasion. Identification of rRP789 and rRP027-028 in the present study provides additional tools toward a better understanding of these phenomena. In the past, we have demonstrated that chemical methylation of rOmpB(A) increased its seroreactivity. Some patient-derived sera could not recognize unmethylated rOmpB(A) but yielded very high binding signals with chemically methylated rOmpB(A). Enzymatic methylation of OmpB may provide a new route for producing methylated OmpB as diagnostic antigens. Antibodies against OmpB are found in patients without detectable rickettsial bacteria (1, 9). Enzymatically methylated OmpB may potentially enhance the potency of OmpB in immunochemical reactions and provide diagnostic reagents with enhanced sensitivity and specificity and vaccine candidates with superior effectiveness. Methylated bacterial surface proteins are not limited to OmpB. Similar approaches to the present studies can be readily applied to many cell surface proteins for the development of enhanced diagnostic reagents and vaccine candidates.
We are grateful to the reviewers for their valuable comments. A.H.A. and D.C.H.Y. thank Yi He (NIH) for help with Western blot analysis and P. Boon Chock for discussion. C.-C.C. and W.-M.C. thank Zhiwen Zhang and Hua-Wei Chen (NMRC) for help with preparation of OmpB AN and OmpB K.
The work was supported by Naval Medical Logistic Command award N62645 (to Georgetown University) and work unit number (WUN) 6000.RAD1.J.A0310 (to NMRC). C.-C.C. and W.-M.C. are employees of the U.S. Government. This work was prepared as part of official duties.
The opinions and assertions contained herein are the private ones of the authors and are not to be construed as official or as reflecting the views of the Department of the Navy, the naval service at large, the Department of Defense, or the U.S. Government.
Published ahead of print 21 September 2012
Supplemental material for this article may be found at http://jb.asm.org/.