Search tips
Search criteria 


Logo of aacPermissionsJournals.ASM.orgJournalAAC ArticleJournal InfoAuthorsReviewers
Antimicrob Agents Chemother. 2009 January; 53(1): 86–94.
Published online 2008 November 17. doi:  10.1128/AAC.00275-08
PMCID: PMC2612134

Inhibition of Methionyl-tRNA Synthetase by REP8839 and Effects of Resistance Mutations on Enzyme Activity[down-pointing small open triangle]


REP8839 is a selective inhibitor of methionyl-tRNA synthetase (MetRS) with antibacterial activity against a variety of gram-positive organisms. We determined REP8839 potency against Staphylococcus aureus MetRS and assessed its selectivity for bacterial versus human orthologs of MetRS. The inhibition constant (Ki) of REP8839 was 10 pM for Staphylococcus aureus MetRS. Inhibition of MetRS by REP8839 was competitive with methionine and uncompetitive with ATP. Thus, high physiological ATP levels would actually facilitate optimal binding of the inhibitor. While many gram-positive bacteria, such as Staphylococcus aureus, express exclusively the MetRS1 subtype, many gram-negative bacteria express an alternative homolog called MetRS2. Some gram-positive bacteria, such as Streptococcus pneumoniae and Bacillus anthracis, express both MetRS1 and MetRS2. MetRS2 orthologs were considerably less susceptible to REP8839 inhibition. REP8839 inhibition of human mitochondrial MetRS was 1,000-fold weaker than inhibition of Staphylococcus aureus MetRS; inhibition of human cytoplasmic MetRS was not detectable, corresponding to >1,000,000-fold selectivity for the bacterial target relative to its cytoplasmic counterpart. Mutations in MetRS that confer reduced susceptibility to REP8839 were examined. The mutant MetRS enzymes generally exhibited substantially impaired catalytic activity, particularly in aminoacylation turnover rates. REP8839 Ki values ranged from 4- to 190,000-fold higher for the mutant enzymes than for wild-type MetRS. These observations provide a potential mechanistic explanation for the reduced growth fitness observed with MetRS mutant strains relative to that with wild-type Staphylococcus aureus.

Staphylococcus aureus and Streptococcus pyogenes are causative agents for a wide variety of cutaneous infections, including impetigo, cellulitis, subcutaneous abscess, furuncles, staphylococcal scalded skin syndrome, and necrotizing fasciitis (16, 38, 39). The incidence of infections caused by antibiotic-resistant bacterial pathogens has increased significantly in recent years. In the United States, for example, over 60% of staphylococcal infections in intensive care units are caused by methicillin-resistant strains (37). Of particular concern has been the emergence of community-associated methicillin-resistant Staphylococcus aureus strains, which are characterized by expression of a wide range of virulence factors and a greater tendency toward progression to invasive disease (23, 40). Therefore, new antibacterial agents that are active against drug-resistant staphylococci represent an important area for drug development.

REP8839 is a novel diaryldiamine-containing compound (Fig. (Fig.1)1) that inhibits methionyl-tRNA synthetase (MetRS). It is currently being developed as a topical antibiotic. REP8839 shows potent antibacterial activity against clinically important skin pathogens, such as S. aureus (including strains that are resistant to vancomycin, linezolid, mupirocin, and methicillin) and multiply resistant strains of Staphylococcus epidermidis (10). It also exhibits strong antibacterial activity against other gram-positive pathogens, such as S. pyogenes, Enterococcus faecium, and Enterococcus faecalis, including vancomycin-resistant enterococci (10).

FIG. 1.
Chemical structure of REP8839.

REP8839 derives from medicinal chemistry optimization (18, 19) of a hit obtained in a high throughput screen for inhibitors of S. aureus MetRS. Ochsner et al. showed that REP8839 specifically inhibited protein synthesis in macromolecular synthesis assays with a Streptococcus pneumoniae relA-negative strain that was unable to support the stringent response (34). Overexpression of MetRS resulted in increased MICs compared to those for strains expressing normal MetRS levels. Mutations in S. aureus that conferred increased resistance to REP8839 mapped to the metS gene encoding MetRS and resulted in amino acid changes in key residues adjacent to the active site for methionine binding. These studies demonstrated that REP8839 exerts its antibacterial activity by specific inhibition of MetRS.

Aminoacyl-tRNA synthetases (aaRSs) are necessary for protein biosynthesis; inhibition of any individual aaRS should effectively shut down the translation process. In the search for new antibacterial agents, the aaRSs thus represent attractive targets for drug discovery (33, 35, 45). The only currently marketed antibiotic that targets an aaRS is mupirocin (pseudomonic acid), a natural product that inhibits isoleucyl-tRNA synthetase. Although considerable effort has focused on developing antibacterial compounds that target other aaRSs, most of these programs have not progressed to clinical development.

The aaRSs fall into two classes based on structural characteristics. Class I enzymes have a Rossman fold in the catalytic center and contain two signature conserved motifs: HIGH and KMSKS. Class II synthetases contain an antiparallel β-sheet with three conserved motifs in the catalytic core (12, 13). MetRS is a class I aaRS that catalyzes the linkage of methionine (Met) to its cognate tRNAMet. This reaction is a two-step process: methionine + ATP ↔ methionyl adenylate + PPi (reaction 1) and methionyl adenylate + tRNAMet → AMP + Met-tRNAMet (reaction 2). First, both methionine and ATP are bound at the active site of the enzyme, which catalyzes the formation of methionyl adenylate, with the release of pyrophosphate (PPi). Next, the activated methionyl adenylate is transferred to the 3′ end of tRNAMet, with the release of AMP. A unique property of MetRS is its ability to recognize and charge two tRNA substrates: tRNAmMet and tRNAfMet. MetRS thus plays a crucial role in translation during both the initiation and elongation phases.

Two major forms of MetRS have been identified based on sequence similarity and sensitivity to inhibitors (15). MetRS1 (encoded by metS1) is the form commonly found in gram-positive bacteria, such as S. aureus and S. pyogenes. MetRS2 (encoded by metS2) is quite distinct from MetRS1 at the amino acid sequence level; MetRS2 homologs are found in archaea, in eukaryotes, and in many gram-negative bacteria. Notable exceptions to this pattern have been observed with certain gram-positive bacteria which contain both metS1 and metS2. These include Bacillus anthracis and a subset of S. pneumoniae clinical isolates (4, 15). It has been proposed that such metS2 genes were acquired through horizontal gene transfer. Eukaryotic organisms contain both MetRS forms, the cytoplasmic enzyme being of the MetRS2 form and the mitochondrial enzyme exhibiting features characteristic of MetRS1. Structural studies have further subdivided MetRS into four subtype families based on the number of Zn-binding knuckle motifs and the actual number of metal atoms bound (8, 26). MetRS2 orthologs contain two knuckles and either two Zn atoms bound (A family) or one Zn bound (B family). MetRS1 orthologs contain one knuckle and either one Zn bound (C family) or no Zn bound (D family).

The reaction mechanism has been studied extensively for Escherichia coli MetRS (reviewed in reference 28), which is a MetRS2 ortholog, and for Bacillus stearothermophilus MetRS, which is a MetRS1 ortholog (21, 29, 30, 41). From a mechanistic and structural standpoint, the two enzymes exhibit many similarities (21, 27, 41). Both show similar substrate binding affinities. Each can bind methionine and ATP separately, and the reaction kinetics are consistent with a random-order reaction mechanism for the formation of the methionyl adenylate intermediate (1, 2, 29). In both cases, the rate-limiting step is the methionyl transfer to tRNA (30).

REP8839 and related compounds show the most potent antibacterial activity against strains that express only MetRS1, such as S. aureus, and much weaker antibacterial activity against strains that express MetRS2 (10). Ochsner et al. showed that REP8839 inhibited S. aureus MetRS (MetRS1) with greater potency than it inhibited MetRS2 derived from gram-negative pathogens or rat liver lysates (34). The magnitude of the selectivity could not be determined, however, due to assay limitations.

Inhibition of S. aureus MetRS by REP8839 was originally measured with a tRNA aminoacylation assay, which provided only an upper limit on the potency of the compound (34). Here we explore limitations of the aminoacylation assay and characterize more thoroughly the biochemical potency and mechanism of inhibition of MetRS by REP8839. We examine REP8839 target selectivity in more-quantitative terms, including the effects of REP8839 on activity of human mitochondrial MetRS (hmMetRS) and human cytoplasmic MetRS (hcMetRS).

We have recently identified a number of laboratory-generated S. aureus mutants that exhibited reduced susceptibility to REP8839 in whole-cell assays (34). Compared to wild-type S. aureus, in which REP8839 has an MIC of 0.12 μg/ml, the MICs for the mutant strains range from 2 to 32 μg/ml (34). These mutants showed reduced growth rates and impaired fitness relative to levels for wild-type S. aureus. Consistent with the mechanism of action of REP8839, mutations in the metS gene, which encodes MetRS, were identified. A total of 23 different amino acid substitutions were identified in various strains. Since these mutations reduced bacterial susceptibility to the inhibitor, we postulated that the 50% inhibitory concentration (IC50) for enzymatic inhibition by REP8839 would be shifted (increased) from that seen with wild-type MetRS. We were also interested to learn whether these changes would affect the ability of the enzyme to carry out its normal function. Would the kinetic parameters governing substrate interactions or catalytic efficiency be different for the mutant MetRS enzymes? In this study, we have selected 11 representative REP8839-resistant strains for kinetic evaluation.


Expression and purification of Staphylococcus aureus MetRS.

The gene encoding S. aureus MetRS was amplified from genomic DNA (wild type or mutant) and cloned into an inducible E. coli expression vector. A lysate was prepared as described previously (11). Ammonium sulfate was added to a final concentration of 40% saturation. The precipitate was collected by centrifugation (23,000 × g, 45 min, 4°C) and discarded. Ammonium sulfate was added to the supernatant to 50% saturation. The precipitate containing S. aureus MetRS was collected by centrifugation (23,000 × g, 45 min, 4°C) and resuspended in buffer Q (50 mM Tris [pH 7.5], 10% glycerol, 1 mM EDTA, and 5 mM dithiothreitol [DTT]). This solution was desalted in buffer Q plus 50 mM NaCl and loaded onto a Source 15Q anion-exchange column (Amersham). The column was washed with buffer Q plus 150 mM NaCl. S. aureus MetRS was eluted with a linear gradient from 150 to 275 mM NaCl. Purified proteins were greater than 95% homogeneous.

Purification of Streptococcus pneumoniae MetRS1.

Bacterial strains (DH5α Pro) containing pPROLar plasmids carrying the gene encoding S. pneumoniae MetRS1 (metS1 gene) were obtained from GlaxoSmithKline (Collegeville, PA). Cells were grown, induced, and harvested as described for S. aureus MetRS, with the exception that expression of MetRS1 was induced by the addition of IPTG (isopropyl-β-d-thiogalactopyranoside) to 1 mM and also by the addition of l-arabinose to 0.5 mg/ml. Purification was done as described for S. aureus MetRS, with the exception that a 45% to 55% ammonium sulfate fractionation was used.

Expression and purification of hcMetRS and hmMetRS.

Plasmid pcDNA3.1MRS, containing a gene encoding hcMetRS, and plasmid pcDNA3.1MRSce2, containing a gene encoding full-length hmMetRS, were acquired from GlaxoSmithKline (Collegeville, PA). The hcMetRS gene was PCR amplified using a forward primer (5′-CACCATGAGACTGTTCGTGAGTGATG-3′) and a reverse primer (5′-CACTGTGCTGGATATCTGCAG-3′). The region encompassing the mature form of the hmMetRS gene (42) was PCR amplified using a forward primer (5′-CACCATGTCTCTCCTGGAGGACTTCG-3′) and the same reverse primer described above. PCR products were inserted into pET100/D-TOPO vectors (Invitrogen). For preparation of hmMetRS, cells were grown, induced, and harvested as described above. The cells were lysed by sonication on ice (model 300 V/T ultrasonic homogenizer; BioLogics) and cleared by centrifugation (22,000 × g, 60 min, 4°C). hmMetRS was purified by a batch method carried out by adding Ni-nitrilotriacetic acid resin (Qiagen) equilibrated in suspension buffer (50 mM Tris-HCl [pH 7.5], 40 mM KCl, 7 mM MgCl2, 10% glycerol) and rocking for 1.5 h. The resin was washed with 100 column volumes of wash buffer (50 mM Tris-HCl [pH 7.5], 1 M KCl, 7 mM MgCl2, 20 mM imidazole). The protein was eluted using elution buffer (50 mM Tris-HCl [pH 7.5], 40 mM KCl, 7 mM MgCl2, 200 mM imidazole). For preparation of hcMetRS, cells were grown, induced, and harvested as described above. Preparation of crude lysate was done as described previously (11). hcMetRS was purified as described above, using Ni-nitrilotriacetic acid resin (50% slurry).

S. pneumoniae MetRS2, E. coli MetRS, and Haemophilus influenzae MetRS were obtained from GlaxoSmithKline (Collegeville, PA).

tRNA aminoacylation assay with scintillation proximity assay detection.

The IC50 represented the concentration of inhibitor required to inhibit enzyme activity by 50%. Aminoacylation reactions for IC50 determinations were carried out at room temperature for 30 min and at 30°C for 1 to 7.5 min for initial rate determinations. Reaction mixtures (50 μl) contained 40 mM Tris-HCl (pH 8.0), 10 mM magnesium acetate, 2.5 mM ATP, 80 mM KCl, 2.5 mM DTT, 40 μg/ml bovine serum albumin, 0.5 to 1 μCi [3H]methionine, 1 μg/ml crude E. coli tRNA, and S. aureus MetRS. Reaction mixtures with other MetRS orthologs were essentially similar, but concentrations of substrates, magnesium, and enzyme were optimized for each system. Reaction mixtures that utilized human MetRS orthologs contained 0.2 to 0.6 mM spermine (pH 7.8). Test compounds dissolved in dimethyl sulfoxide (DMSO) (4 μl) were added to reaction mixtures as serial dilutions from 1 μM to 0.05 nM (final concentration). Reactions were stopped by the addition of 5 μl of 0.5 M EDTA. Charging of tRNAMet was determined using a scintillation proximity assay. Polylysine-coated YSi beads (200 μg; Amersham Biosciences) were added to mixtures for stopped reactions in 150 μl of 300 mM citrate buffer (pH 2.0). Reactions were carried out in white 96-well plates (Costar), and charged tRNA was counted using a TopCount NXT microplate scintillation counter (Packard). IC50s were determined by fitting a four-parameter logistic model (sigmoidal dose response) to the data using XLfit4.1 (IDBS) software.

ATP:PPi exchange assay.

Methods for the ATP:PPi exchange reaction were modified from previously reported examples (5). For determination of MetRS kinetic parameters, reactions were carried out at 30°C and quenched at intervals between 1 and 10 min. Reaction mixtures (50 μl) contained 100 mM Tris-HCl (pH 8.0), 5 to 10 mM magnesium acetate, 80 mM KCl, 2.5 mM DTT, and 30 nM S. aureus MetRS. Assay mixtures contained 2 μCi [32P]NaPPi and cold NaPPi at a concentration equal to that of ATP. ATP and methionine concentrations are indicated in the figure legends. Aliquots (8 μl) were removed at 1-min intervals and quenched with 4 μl of 100 mM EDTA. Aliquots (2.5 μl) of the quenched reaction mixtures were spotted on PEI cellulose flexible thin-layer-chromatography plates (J. T. Baker). ATP and PPi were separated using 4 M urea and 0.75 M KPi (pH 3.5) as a mobile phase. Thin-layer-chromatography plates were quantified with a Storm 840 (Molecular Dynamics) phosphorimager. Initial velocities for the ATP:PPi exchange reaction were plotted against substrate concentration and fit to the Michaelis-Menten steady-state model using XLfit4.1 (IDBS) software.

Inhibition by REP8839.

IC50s for inhibition of tRNA aminoacylation by REP8839 were determined by dosing the inhibitor in serial dilutions from 1 μM to 0.05 nM or from 100 nM to 0.2 nM. The concentrations of ATP and methionine were held constant at 2.5 mM and 7 μM, respectively, in these assays. The inhibitor was dissolved in 100% DMSO, and 4 μl was added to each reaction mixture, giving a final concentration of 8% DMSO. Activity data were expressed as a percentage of the DMSO positive-control value. The data were fit to a sigmoidal dose response based on the formula fit = {A + (BA)/[1 + (C/X)^D]}, where A is the minimum Y value, B is the maximum Y value, C is the IC50, D is the slope, and X is the compound concentration. IC50s that were close to 1.5 nM represented assays in which the measurement of inhibition by REP8839 was limited by the concentration of the enzyme. Therefore, IC50s for wild-type MetRS and MetRS with single mutations L213W, A247E, I57N, V108M, and G223C (with IC50s in the aminoacylation assay ranging from 0.7 to 3.5 nM) did not provide useful information with which to compare the levels of potency of the inhibitors. For these MetRS variants, the IC50 was determined by the ATP:PPi exchange assay under modified substrate conditions of elevated methionine and reduced ATP. The concentration of methionine was raised to 125 mM, and the concentration of ATP was reduced to 100 μM. In this set of assays, REP8839 concentrations ranging from 1 μM to 0.05 nM were used. Under these conditions, the IC50 was shifted up to a level that was not limited by the enzyme concentration.

For REP8839 inhibition of MetRS, there was no difference in IC50 determined with preincubation of enzyme and inhibitor for 5 min versus 2 h before initiating the reaction by addition of [32P]PPi (data not shown). Thus, we found no evidence for a slow conformational change in MetRS with REP8839 binding, as seen with mupirocin inhibition of S. aureus isoleucyl-tRNA synthetase (36).


Limitations of the tRNA aminoacylation assay.

Esterification of radiolabeled methionine to the 3′ end of tRNA was monitored in order to determine tRNA aminoacylation rates catalyzed by MetRS. The REP8839 IC50 for inhibition of tRNA aminoacylation by wild-type S. aureus MetRS decreased in proportion to the decreasing concentration of MetRS enzyme in the reaction mixture (data not shown). The IC50s were close to the enzyme concentration at each condition. This is characteristic of tight-binding inhibitors, for which the measured IC50 reflects the number of inhibitor molecules required to bind half of the enzyme molecules (50% bound). Under these conditions, the IC50 approaches the theoretical lower limit, equal to one-half the enzyme concentration. The enzyme concentration could not be reduced further without an unacceptable reduction in signal-to-background ratio. Therefore, we could not determine the Ki values for potent inhibitors of MetRS by using this assay. An alternative approach was to change the substrate concentrations in order to shift the measured IC50 above the enzyme concentration, as described in the next section.

Adaptation of an ATP:PPi exchange assay.

The measured IC50 of an inhibitor of an enzyme assay can be artificially but predictably increased by raising the concentration of a competitive substrate above the Km for that substrate (36). The observed shift in IC50 is a function of the ratio of the substrate concentration (S) to the Km, described by the Cheng-Prusoff equation for competitive inhibitors (6): IC50 = (1 + S/Km) × Ki. The shift in IC50 thus becomes directly proportional to S/Km when the substrate concentration is well above Km. Conversely, the Cheng-Prusoff equation for uncompetitive inhibitors, IC50 = (1 + Km/S) × Ki, indicates that the IC50 can be shifted upwards by lowering the concentration of an uncompetitive substrate. Appropriate adjustments to substrate concentration can thus achieve a shift in the IC50 to a level well above the assay enzyme concentration, so that the measured IC50 is directly proportional to Ki.

The aminoacyl adenylate condensation half-reaction (reaction 1) catalyzed by MetRS is readily reversible in the absence of tRNA. [32P]PPi incorporation into ATP by the reverse reaction can be detected in an ATP:PPi exchange assay. In this format, both the methionine concentration and the ATP concentration can be varied without a deleterious effect on the signal-to-background ratio. It was thus possible to find substrate concentrations that resulted in IC50s that were well above the enzyme concentration in the assay. Under these conditions, IC50s reflect the binding affinity of the inhibitor rather than the enzyme-limited values seen with the aminoacylation assay.

Synergy between ATP and methionine binding to S. aureus MetRS.

Kinetic parameters for the ATP and methionine substrates of S. aureus MetRS were determined from initial velocities in the ATP:PPi exchange reaction plotted against substrate concentration and fit to the Michaelis-Menten steady-state model, V0 = Vmax/(1 + S/Km). KmMet was 100 μM, and KmATP was 500 μM (data not shown).

In order to calculate a Ki value from an IC50, one must understand the mode of inhibition with respect to each substrate as well as the Km for each substrate. In the simplest case, the Km values for the two substrates are assumed to be independent. It is known for both E. coli and B. stearothermophilus MetRS, however, that there is cooperative binding between the methionine and ATP substrates (2, 21). To determine whether this was true for S. aureus MetRS, we assessed whether KmMet was affected by the ATP concentration. KmMet was redetermined at 100 μM ATP (fivefold below KmATP). KmMet did increase from 100 μM (determined at saturating ATP) to 330 μM (determined at the reduced ATP concentration). This indicated moderately positive cooperative binding of the methionine and ATP substrates in S. aureus MetRS.

REP8839 is uncompetitive with ATP.

The mechanism of inhibition of S. aureus MetRS with respect to ATP was determined by the ATP:PPi exchange assay using various ATP concentrations from 5-fold above Km to 20-fold below Km while keeping methionine at a high (saturating) concentration. The IC50s for REP8839 decreased with increasing ATP concentration (Fig. (Fig.2),2), which is characteristic of an uncompetitive inhibitor. The data were fit (Fig. (Fig.2)2) to the Cheng-Prusoff equation for mixed uncompetitive inhibitors, i.e., an inhibitor that binds more tightly to the enzyme-substrate complex than to the enzyme alone (6). Alternatively, the data were fit to the equation for a pure uncompetitive inhibitor, i.e., an inhibitor that binds only to the enzyme-substrate complex. The mixed uncompetitive model gave a better fit, suggesting that REP8839 can bind weakly to MetRS alone but binds with higher avidity to the MetRS-ATP complex.

FIG. 2.
Inhibition by REP8839 is uncompetitive with ATP. REP8839 IC50s were determined by the ATP:PPi exchange assay. The methionine concentration was fixed at 126 mM, and IC50s were determined at four different ATP concentrations ranging from 25 to 2,500 μM. ...

REP8839 is competitive with methionine.

Compounds related to REP8839 but with lower binding affinity for MetRS have been shown to be competitive with the methionine substrate (20). The mechanism of REP8839 inhibition of S. aureus MetRS with respect to methionine was initially evaluated by the ATP:PPi exchange assay at a saturating ATP concentration. Under these conditions, the binding of REP8839 was so tight that the IC50s were close to the enzyme limit and changed only slightly as a function of methionine concentration up to the highest attainable methionine levels (Fig. (Fig.3A,3A, [ATP] = 2,500 μM). Therefore, in order to assess the mechanism of action with respect to methionine, it was necessary to decrease ATP to levels well below Km (Fig. (Fig.3A,3A, [ATP] = 25 μM). At this suboptimal ATP concentration, the binding affinity of REP8839 was modulated to a point that a response to the methionine concentration was quantifiable. The IC50 increased with an increasing methionine concentration, which is characteristic of a competitive inhibitor (Fig. (Fig.3B3B).

FIG. 3.
Inhibition by REP8839 is competitive with methionine. (A) REP8839 IC50s were determined by the ATP:PPi exchange assay with three different methionine concentrations (12, 36, and 126 mM) and two different ATP concentrations (25 μM and 2.5 mM). ...

Potency of MetRS inhibition by REP8839.

In a two-substrate reaction, the shift in IC50 with respect to various substrate concentrations is the product of the shift due to each individual substrate (6). For the MetRS exchange assay with inhibitors that are methionine competitive and ATP uncompetitive, the relationship is as follows: IC50 = (1 + [Met]/KmMet)(1 + KmATP/[ATP])Ki. This equation was used to calculate Ki for REP8839 with IC50s determined at 126 mM methionine and 100 μM ATP. Due to the synergy between ATP and methionine binding, we utilized the KmMet that had been determined at the exact ATP concentration used for the IC50 determination. In multiple determinations, REP8839 had a mean Ki of 10 (±5) pM for S. aureus MetRS. Thus, REP8839 is a potent inhibitor of the target bacterial enzyme.

REP8839 is a selective inhibitor of bacterial MetRS1.

The relative levels of potency of REP8839-mediated inhibition of different MetRS orthologs were examined in order to assess target selectivity. First, Km values for ATP and methionine were determined for each ortholog (Table (Table1).1). In all cases, the mechanism of REP8839 inhibition was found to be uncompetitive with respect to ATP (data not shown). Ki values were determined (Table (Table2)2) as described above for S. aureus MetRS. Bacterial MetRS1 orthologs were quite sensitive to inhibition by REP8839, with Ki values in the picomolar range. hmMetRS was 1,000-fold less sensitive to inhibition by REP8839 than was S. aureus MetRS. Inhibition by REP8839 was weak against the E. coli and H. influenzae MetRS2 orthologs and nondetectable against the S. pneumoniae MetRS2 and human cytoplasmic orthologs.

Michaelis constants for MetRS orthologs
Selectivity of REP8839 inhibition of bacterial and human MetRS orthologs

Despite the high selectivity for the bacterial MetRS1 orthologs, the relative susceptibility of mitochondrial MetRS prompted us to assess the effects of REP8839 on mammalian cell growth. REP8839 was tested for cytotoxicity in cultured human A549 and H4IIE cells by use of an MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide] assay at concentrations up to the aqueous solubility limit of ~20 μM. The REP8839 IC50s were >20 μM (data not shown) for both cell lines, indicating that REP8839 was not deleterious to mammalian cell growth or mitochondrial function.

Kinetic analysis of mutant MetRS isoforms.

Eleven MetRS variants were expressed and purified from S. aureus strains that exhibit reduced susceptibility to REP8839 (34) and subjected to detailed kinetic analysis. KmATP, KmMet, and kcat1 values for wild-type and mutant MetRS enzymes are shown in Table Table3.3. The Km values for ATP and methionine were similar (within 4-fold) to the value for wild-type MetRS, with the exception of the G223C and G54A A64P mutants, which exhibited 7-fold and 10-fold increases in KmMet, respectively. The mutant MetRS enzymes ranged from 2- to 25-fold slower in the rate constant (kcat1) for methionyl adenylate formation than wild-type MetRS.

Kinetic parameters of S. aureus mutant MetRS enzymes

The KmtRNA values of the mutant MetRS enzymes remained similar (within 4-fold) to the value for wild-type MetRS, except for the I57N V242F double mutant, for which the KmtRNA was 10-fold higher (Table (Table3).3). However, the turnover rates (kcattRNA) exhibited much greater variability. MetRS with single mutations L213W and V108M had kcattRNA values that were similar to that of the wild-type enzyme, while MetRS with single mutations A247E, I57N, and G223C and double mutations I238F I57N and I57N V242F had kcattRNA values 4- to 20-fold lower than that of wild-type MetRS. MetRS with single mutation G54S and double mutations I57N A247E, I57N G54S, and G54A A64P had kcattRNA values 50- to 500-fold lower than that of wild-type MetRS.

Determination of Ki for REP8839 with mutant MetRS variants.

To gain insight into the effect of mutations on the binding affinity of REP8839, we determined Ki values (Table (Table4).4). The Ki of REP8839 for wild-type S. aureus MetRS was 10 pM. MetRS with certain single amino acid changes (L213W, A247E, I57N, and V108M) had Ki values ranging from 4- to 11-fold higher than that for wild-type MetRS. The Ki for MetRS with single mutation G223C was more than 50-fold greater and the Ki for MetRS with single mutation G54S was more than 8,000-fold greater than that observed with the wild-type MetRS. Ki values for MetRS enzymes with double mutations ranged from 1,500-fold higher (I57N V242F) to almost 200,000-fold higher (G54A A64P) than the wild-type MetRS value.

Inhibitory effects of REP8839 on S. aureus mutant MetRS enzymes


Potency of MetRS inhibitors.

The aaRSs, and MetRS in particular, have long been recognized as potential targets for antibacterial agents (35, 46). Several classes of compounds have previously been reported as bacterial MetRS inhibitors. The most potent examples at the biochemical level, with IC50 values in the low nanomolar range, include oxazolone dipeptides (43) and methionyl adenylate isosteres (45). These compounds showed good selectivity versus human MetRS but lacked antibacterial activity in whole-cell assays. Other methionine analogs have shown much weaker MetRS inhibition, with Ki values in the low micromolar range and a very narrow antibacterial spectrum (24, 25). The 2-pyridyl-pyrazoles (14) demonstrated moderate biochemical potency and selectivity but exhibited off-target activity in whole-cell assays. The biochemical potency of REP8839 reported here (Ki = 10 pM) ranks it among the most potent MetRS inhibitors published to date. REP8839 shows robust antibacterial activity against S. aureus and other skin pathogens (10) and has recently been advanced into clinical development.

Binding cooperativity of MetRS substrates and inhibitors.

Direct binding studies of methionine, ATP, and other substrate analogs have been reported for E. coli MetRS (2, 9) and B. stearothermophilus MetRS (21, 29). Interestingly, cooperative ATP binding was markedly enhanced when a substrate analog such as methioninol, which lacks the carboxylate group, was used. Furthermore, the methionine binding affinity increased dramatically when adenosine and pyrophosphate were substituted for ATP. These observations led to the hypothesis that the enzyme must overcome an electrostatic repulsion between the carboxyl group of the amino acid and the α-phosphoryl group of ATP. Thus, two compensating properties are at work. Favorable coupling between the methionine and ATP sites promotes substrate binding, but an electrostatic repulsion between the substrates must be overcome. E. coli MetRS undergoes distinct conformational changes upon binding of substrates and substrate analogs which may contribute to the cooperative substrate binding (9). We found that MetRS from S. aureus exhibited moderately positive cooperative binding of ATP and methionine, based on the threefold decrease in KmMet observed at high ATP versus low ATP concentrations.

We demonstrated that MetRS inhibition by REP8839 was competitive with methionine binding and uncompetitive with ATP binding. Thus, the naturally high physiological concentrations of ATP (~2.5 mM [3]) would serve to enhance binding of the inhibitor. In fact, the binding of REP8839 was so tight that we were initially unable to demonstrate significant competition by methionine (up to the highest attainable concentrations) in the presence of physiological ATP levels. Methionine could compete for REP8839 binding only under severely limiting ATP concentrations of 25 μM (approximately 100-fold below physiological levels). It is interesting that the positive cooperativity exhibited between REP8839 and ATP parallels the positive cooperativity that has been reported between uncharged methionine analogs and ATP (28). Thus, the inhibitor appears to leverage the marked positive cooperativity derived from occupancy of both the ATP and methionine binding pockets. Unlike methionine, REP8839 carries no negative charge, so it could benefit from the positive cooperativity without having to overcome the charge-charge repulsion that arises when ATP and methionine are juxtaposed in the active site. Thus, the exquisitely tight binding of REP8839 compared to that of methionine may derive from several factors, including its greater size, its lack of negative charge, and its potential ability to stabilize the enzyme in a favorable conformation.

Selectivity of REP8839 for inhibition of MetRS1.

One of the hallmarks of a successful antibiotic is that it exhibits a high degree of selectivity for the prokaryotic target compared to any of its potential eukaryotic orthologs, thereby minimizing the chance of mechanism-based side effects. No inhibition of hcMetRS was observed at concentrations up to 20 μM, which was close to the aqueous solubility limit of the compound. Thus, the selectivity of REP8839 for S. aureus MetRS compared to that for hcMetRS is >1,000,000-fold. Although inhibition of purified hmMetRS was detectable, no significant cytotoxicity was observed in mammalian cell culture at concentrations up to 20 μM. REP8839 is highly serum bound, and the presence of 10% serum in the culture medium may provide a protective effect. Uptake of the compound into mammalian cells may also be limited. It is noteworthy that in bacteria only very potent MetRS inhibitors elicit an antibacterial effect (Ki values are typically much lower than MICs). This may reflect upregulation of MetRS expression in response to depletion of charged tRNAMet pools; a similar mechanism may limit the effect of REP8839 in mammalian cells.

The high selectivity for bacterial versus human MetRS isoforms is clearly a desirable property for a potential therapeutic agent. Interestingly, while previous studies of MetRS inhibitors have reported selectivity values for bacterial versus “human” MetRS, none provided specific data on the mitochondrial MetRS ortholog. The results with REP8839 show that the cytoplasmic and mitochondrial MetRS orthologs can be differentially sensitive to inhibition; it is valuable to monitor both enzymes in order to gain a comprehensive view of the prokaryotic selectivity of candidate compounds.

Effects of mutation on MetRS enzymatic fitness and susceptibility to REP8839.

We have analyzed S. aureus MetRS enzymes from laboratory-generated mutant strains that contain key mutations that result in reduced in vitro susceptibility to REP8839. Mutations that resulted in higher REP8839 MICs in whole-cell assays also caused elevated Ki values in enzymatic assays. The Ki for inhibition by REP8839 increased up to 200,000-fold for MetRS with double mutations. As sensitivity to REP8839 decreased, the enzymatic efficiency of the mutant MetRS enzymes decreased concomitantly.

Under normal physiological conditions (e.g., intracellular methionine at ~100 μM [3]), the rate-limiting step of the aminoacylation reaction catalyzed by wild-type MetRS is the second step, the charging of tRNA. We found that this observation held true for the mutant MetRS enzymes. Despite widely variable kcat1/KmMet values for the mutant MetRS enzymes (Table (Table3),3), in each case the value for kcattRNA in the aminoacylation assay was lower than the rate of methionyl adenylate formation ([Met] × kcat1/KmMet). Methionine levels would have to drop substantially below normal physiological levels (e.g., >50-fold for wild-type MetRS) in order for methionyl adenylate formation to become the rate-limiting step. Thus, the increased KmMet values for some of the mutant MetRS enzymes would probably not be significant for normal enzyme function. They may be significant in the presence of REP8839, however, since the inhibitor is competitive with methionine.

The L213W and V108M single mutants were the least deleterious with respect to normal MetRS function. The I57N mutant was intermediate, with a 21-fold reduction in kcattRNA but a favorable 4-fold decrease in KmMet. I57N occurred in a number of double mutations, each of which resulted in dramatic increases in Ki relative to that with I57N alone. A247E occurred as a single mutation and also in combination with I57N. As a single mutation, it gave a sixfold increase in Ki for REP8839, with an accompanying ninefold reduction in kcattRNA. The G223C and G54S mutants were the most deleterious single mutants, showing markedly impaired enzymatic function and greatly decreased REP8839 sensitivity. In particular, the G54S mutant showed a dramatic 8,500-fold increase in Ki compared to that for wild-type MetRS.

Five MetRS double mutants were examined, four of which contained I57N. The I57N V242F double mutant was unique among the mutants studied in having a substantially elevated KmtRNA (~10-fold higher than that of wild-type MetRS). Interestingly, it exhibited a kcattRNA fivefold higher than that for the I57N single mutant, thereby partially rescuing a defect in the I57N mutant. Given the high KmtRNA, however, the net effect on kcattRNA/KmtRNA was still deleterious relative to the effect with the I57N mutant. The I57N I238F mutant showed a similar property of partially rescuing the low kcattRNA of the I57N mutant. The I57N A247E mutant combined the deleterious properties of the I57N and A247E single mutations, with a net 60-fold reduction in kcattRNA. Of the four double mutants containing I57N, the I57N G54S mutant was the most resistant to inhibition by REP8839 and also showed the most dramatic effects on enzymatic efficiency, with substantial defects in both kcattRNA and kcat1/KmMet. The G54A A64P double mutant showed the greatest effects on both REP8839 sensitivity and enzyme efficiency. The Ki for REP8839 was 190,000-fold higher than that for wild-type MetRS, the kcattRNA was 500-fold lower, and the kcat1/KmMet was 270-fold lower.

Potential impacts of enzymatic fitness on bacterial growth fitness.

It is curious that no high-level-REP8839-resistant S. aureus mutants (MIC of >32 μg/ml) were identified (34), even upon serial passage in the presence of the inhibitor and even though the sensitivity to REP8839 at the enzymatic level was vastly decreased. How can a mutant strain such as the I57N G54S mutant, with a 90,000-fold increase in Ki for REP8839, show only a 250-fold increase in MIC? The answer may lie with the substantially impaired enzymatic function of the mutant MetRS enzyme, particularly with respect to aminoacylation turnover rates. Adequate MetRS activity is required to maintain the intracellular pool of charged tRNAMet (Met-tRNAMet), an essential precursor for protein synthesis. When MetRS activity is inhibited, the level of charged tRNAMet drops below the level required to support efficient translation. This activates the stringent response, and the cells cease to divide. If MetRS enzyme activity is impaired by mutation, the resulting strains are probably unable to maintain normal levels of charged tRNAMet. Under those circumstances, even a weak MetRS inhibitor could tip the balance enough to arrest cell growth.

All of the mutant strains studied showed some degree of bacterial growth impairment, manifested by reduced growth rates and reduced fitness in coculture with wild-type, REP8839-sensitive S. aureus strains (34). The impaired enzymatic function exhibited by the mutant MetRS enzymes provides a plausible explanation for the reduced growth fitness. It is noteworthy that no mutational pathways that resulted in substantially reduced affinity for REP8839 emerged without a substantial fitness burden both at the whole-cell level and at the enzymatic level. Collectively, these observations suggest that the mutational avenues available for the development of REP8839 resistance in S. aureus are limited, constrained at least in part by the inherent catalytic requirements of the target enzyme.


[down-pointing small open triangle]Published ahead of print on 17 November 2008.


1. Blanquet, S., G. Fayat, and J. P. Waller. 1974. The mechanism of action of methionyl-tRNA synthetase from Escherichia coli. Mechanism of the amino-acid activation reaction catalyzed by the native and the trypsin-modified enzymes. Eur. J. Biochem. 44:343-351. [PubMed]
2. Blanquet, S., G. Fayat, and J. P. Waller. 1975. The amino acid activation reaction catalyzed by methionyl-transfer RNA synthetase: evidence for synergistic coupling between the sites for methionine adenosine and pyrophosphate. J. Mol. Biol. 94:1-15. [PubMed]
3. Blanquet, S., M. Iwatsubo, and J. P. Waller. 1973. The mechanism of action of methionyl-tRNA synthetase from Escherichia coli. 1. Fluorescence studies on tRNAMet binding as a function of ligands, ions and pH. Eur. J. Biochem. 36:213-226. [PubMed]
4. Brown, J. R., D. Gentry, J. A. Becker, K. Ingraham, D. J. Holmes, and M. J. Stanhope. 2003. Horizontal transfer of drug-resistant aminoacyl-transfer-RNA synthetases of anthrax and Gram-positive pathogens. EMBO Rep. 4:692-698. [PubMed]
5. Bullard, J. M., Y. C. Cai, B. Demeler, and L. L. Spremulli. 1999. Expression and characterization of a human mitochondrial phenylalanyl-tRNA synthetase. J. Mol. Biol. 288:567-577. [PubMed]
6. Cheng, Y., and W. H. Prusoff. 1973. Relationship between the inhibition constant (KI) and the concentration of inhibitor which causes 50 per cent inhibition (IC50) of an enzymatic reaction. Biochem. Pharmacol. 22:3099-3108. [PubMed]
7. Crepin, T., E. Schmitt, S. Blanquet, and Y. Mechulam. 2002. Structure and function of the C-terminal domain of methionyl-tRNA synthetase. Biochemistry 41:13003-13011. [PubMed]
8. Crepin, T., E. Schmitt, S. Blanquet, and Y. Mechulam. 2004. Three-dimensional structure of methionyl-tRNA synthetase from Pyrococcus abyssi. Biochemistry 43:2635-2644. [PubMed]
9. Crepin, T., E. Schmitt, Y. Mechulam, P. B. Sampson, M. D. Vaughan, J. F. Honek, and S. Blanquet. 2003. Use of analogues of methionine and methionyl adenylate to sample conformational changes during catalysis in Escherichia coli methionyl-tRNA synthetase. J. Mol. Biol. 332:59-72. [PubMed]
10. Critchley, I. A., C. L. Young, K. C. Stone, U. A. Ochsner, J. Guiles, T. Tarasow, and N. Janjic. 2005. Antibacterial activity of REP8839, a new antibiotic for topical use. Antimicrob. Agents Chemother. 49:4247-4252. [PMC free article] [PubMed]
11. Cull, M., and C. S. McHenry. 1990. Preparation of extracts from prokaryotes. Methods Enzymol. 182:147-153. [PubMed]
12. Cusack, S., M. Hartlein, and R. Leberman. 1991. Sequence, structural and evolutionary relationships between class 2 aminoacyl-tRNA synthetases. Nucleic Acids Res. 19:3489-3498. [PMC free article] [PubMed]
13. Eriani, G., M. Delarue, O. Poch, J. Gangloff, and D. Moras. 1990. Partition of tRNA synthetases into two classes based on mutually exclusive sets of sequence motifs. Nature 347:203-206. [PubMed]
14. Finn, J., K. Mattia, M. Morytko, S. Ram, Y. Yang, X. Wu, E. Mak, P. Gallant, and D. Keith. 2003. Discovery of a potent and selective series of pyrazole bacterial methionyl-tRNA synthetase inhibitors. Bioorg. Med. Chem. Lett. 13:2231-2234. [PubMed]
15. Gentry, D. R., K. A. Ingraham, M. J. Stanhope, S. Rittenhouse, R. L. Jarvest, P. J. O'Hanlon, J. R. Brown, and D. J. Holmes. 2003. Variable sensitivity to bacterial methionyl-tRNA synthetase inhibitors reveals subpopulations of Streptococcus pneumoniae with two distinct methionyl-tRNA synthetase genes. Antimicrob. Agents Chemother. 47:1784-1789. [PMC free article] [PubMed]
16. Iwatsuki, K., O. Yamasaki, S. Morizane, and T. Oono. 2006. Staphylococcal cutaneous infections: invasion, evasion and aggression. J. Dermatol. Sci. 42:203-214. [PubMed]
17. Jakubowski, H. 2000. Translational incorporation of S-nitrosohomocysteine into protein. J. Biol. Chem. 275:21813-21816. [PubMed]
18. Jarvest, R. L., J. M. Berge, V. Berry, H. F. Boyd, M. J. Brown, J. S. Elder, A. K. Forrest, A. P. Fosberry, D. R. Gentry, M. J. Hibbs, D. D. Jaworski, P. J. O'Hanlon, A. J. Pope, S. Rittenhouse, R. J. Sheppard, C. Slater-Radosti, and A. Worby. 2002. Nanomolar inhibitors of Staphylococcus aureus methionyl tRNA synthetase with potent antibacterial activity against gram-positive pathogens. J. Med. Chem. 45:1959-1962. [PubMed]
19. Jarvest, R. L., J. M. Berge, M. J. Brown, P. Brown, J. S. Elder, A. K. Forrest, C. S. Houge-Frydrych, P. J. O'Hanlon, D. J. McNair, S. Rittenhouse, and R. J. Sheppard. 2003. Optimisation of aryl substitution leading to potent methionyl tRNA synthetase inhibitors with excellent gram-positive antibacterial activity. Bioorg. Med. Chem. Lett. 13:665-668. [PubMed]
20. Jarvis, T. C., L. S. Green, J. M. Bullard, W. Ribble, and N. Janjic. 2006. Selectivity and potency of methionyl tRNA synthetase inhibition by REP8839. Abstr. 46th Intersci. Conf. Antimicrob. Agents Chemother., abstr. F1-1969.
21. Kalogerakos, T., P. Dessen, G. Fayat, and S. Blanquet. 1980. Proteolytic cleavage of methionyl transfer ribonucleic acid synthetase from Bacillus stearothermophilus: effects on activity and structure. Biochemistry 19:3712-3723. [PubMed]
22. Kohda, D., S. Yokoyama, and T. Miyazawa. 1987. Functions of isolated domains of methionyl-tRNA synthetase from an extreme thermophile, Thermus thermophilus HB8. J. Biol. Chem. 262:558-563. [PubMed]
23. Le, J., and J. M. Lieberman. 2006. Management of community-associated methicillin-resistant Staphylococcus aureus infections in children. Pharmacotherapy 26:1758-1770. [PubMed]
24. Lee, J., M. K. Kang, M. W. Chun, Y. J. Jo, J. H. Kwak, and S. Kim. 1998. Methionine analogues as inhibitors of methionyl-tRNA synthetase. Bioorg. Med. Chem. Lett. 8:3511-3514. [PubMed]
25. Lee, J., S. U. Kang, M. K. Kang, M. W. Chun, Y. J. Jo, J. H. Kwak, and S. Kim. 1999. Methionyl adenylate analogues as inhibitors of methionyl-tRNA synthetase. Bioorg. Med. Chem. Lett. 9:1365-1370. [PubMed]
26. Mechulam, Y., E. Schmitt, L. Maveyraud, C. Zelwer, O. Nureki, S. Yokoyama, M. Konno, and S. Blanquet. 1999. Crystal structure of Escherichia coli methionyl-tRNA synthetase highlights species-specific features. J. Mol. Biol. 294:1287-1297. [PubMed]
27. Mechulam, Y., E. Schmitt, M. Panvert, J. M. Schmitter, M. Lapadat-Tapolsky, T. Meinnel, P. Dessen, S. Blanquet, and G. Fayat. 1991. Methionyl-tRNA synthetase from Bacillus stearothermophilus: structural and functional identities with the Escherichia coli enzyme. Nucleic Acids Res. 19:3673-3681. [PMC free article] [PubMed]
28. Meinnel, T., Y. Mechulam, F. Dardel, J. M. Schmitter, C. Hountondji, S. Brunie, P. Dessen, G. Fayat, and S. Blanquet. 1990. Methionyl-tRNA synthetase from E. coli—a review. Biochimie 72:625-632. [PubMed]
29. Mulvey, R. S., and A. R. Fersht. 1976. Subunit interactions in the methionyl-tRNA synthetase of Bacillus stearothermophilus. Biochemistry 15:243-249. [PubMed]
30. Mulvey, R. S., and A. R. Fersht. 1978. Mechanism of aminoacylation of transfer RNA. A pre-steady-state analysis of the reaction pathway catalyzed by the methionyl-tRNA synthetase of Bacillus stearothermophilus. Biochemistry 17:5591-5597. [PubMed]
31. Myers, E. W., and W. Miller. 1988. Optimal alignments in linear space. Comput. Appl. Biosci. 4:11-17. [PubMed]
32. Nureki, O., T. Kohno, K. Sakamoto, T. Miyazawa, and S. Yokoyama. 1993. Chemical modification and mutagenesis studies on zinc binding of aminoacyl-tRNA synthetases. J. Biol. Chem. 268:15368-15373. [PubMed]
33. Ochsner, U. A., X. Sun, T. Jarvis, I. Critchley, and N. Janjic. 2007. Aminoacyl-tRNA synthetases: essential and still promising targets for new anti-infective agents. Expert Opin. Investig. Drugs 16:573-593. [PubMed]
34. Ochsner, U. A., C. L. Young, K. C. Stone, F. B. Dean, N. Janjic, and I. A. Critchley. 2005. Mode of action and biochemical characterization of REP8839, a novel inhibitor of methionyl-tRNA synthetase. Antimicrob. Agents Chemother. 49:4253-4262. [PMC free article] [PubMed]
35. Pohlmann, J., and H. Brotz-Oesterhelt. 2004. New aminoacyl-tRNA synthetase inhibitors as antibacterial agents. Curr. Drug Targets Infect. Disord. 4:261-272. [PubMed]
36. Pope, A. J., K. J. Moore, M. McVey, L. Mensah, N. Benson, N. Osbourne, N. Broom, M. J. Brown, and P. O'Hanlon. 1998. Characterization of isoleucyl-tRNA synthetase from Staphylococcus aureus. II. Mechanism of inhibition by reaction intermediate and pseudomonic acid analogues studied using transient and steady-state kinetics. J. Biol. Chem. 273:31691-31701. [PubMed]
37. Rice, L. B. 2006. Antimicrobial resistance in gram-positive bacteria. Am. J. Infect. Control 34:S11-S19. [PubMed]
38. Rogers, R. L., and J. Perkins. 2006. Skin and soft tissue infections. Prim. Care 33:697-710. [PubMed]
39. Rosen, T. 2005. Update on treating uncomplicated skin and skin structure infections. J. Drugs Dermatol. 4:s9-s14. [PubMed]
40. Sabol, K. E., K. L. Echevarria, and J. S. Lewis. 2006. Community-associated methicillin-resistant Staphylococcus aureus: new bug, old drugs. Ann. Pharmacother. 40:1125-1133. [PubMed]
41. Schmitt, E., M. Panvert, Y. Mechulam, and S. Blanquet. 1997. General structure/function properties of microbial methionyl-tRNA synthetases. Eur. J. Biochem. 246:539-547. [PubMed]
42. Spencer, A. C., A. Heck, N. Takeuchi, K. Watanabe, and L. L. Spremulli. 2004. Characterization of the human mitochondrial methionyl-tRNA synthetase. Biochemistry 43:9743-9754. [PubMed]
43. Tandon, M., D. L. Coffen, P. Gallant, D. Keith, and M. A. Ashwell. 2004. Potent and selective inhibitors of bacterial methionyl tRNA synthetase derived from an oxazolone-dipeptide scaffold. Bioorg. Med. Chem. Lett. 14:1909-1911. [PubMed]
44. Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680. [PMC free article] [PubMed]
45. Vaughan, M. D., P. B. Sampson, E. Daub, and J. F. Honek. 2005. Investigation of bioisosteric effects on the interaction of substrates/ inhibitors with the methionyl-tRNA synthetase from Escherichia coli. Med. Chem. 1:227-237. [PubMed]
46. Vaughan, M. D., P. B. Sampson, and J. F. Honek. 2002. Methionine in and out of proteins: targets for drug design. Curr. Med. Chem. 9:385-409. [PubMed]

Articles from Antimicrobial Agents and Chemotherapy are provided here courtesy of American Society for Microbiology (ASM)