A Unique Hydrophobic Cluster Near the Active Site Contributes to Differences in Borrelidin Inhibition among Threonyl-tRNA Synthetases*

Borrelidin, a compound with anti-microbial and antiangiogenic properties, is a known inhibitor of bacterial and eukaryal threonyl-tRNA synthetase (ThrRS). The inhibition mechanism of borrelidin is not well understood. Archaea contain archaeal and bacterial genre ThrRS enzymes that can be distinguished by their sequence. We explored species-specific borrelidin inhibition of ThrRSs. The activity of ThrRS from Sulfolobus solfataricus and Halobacterium sp. NRC-1 was inhibited by borrelidin, whereas ThrRS enzymes from Methanocaldococcus jannaschii and Archaeoglobus fulgidus were not. In Escherichia coli ThrRS, borrelidin binding induced a conformational change, and threonine was not activated as shown by ATP-PPi exchange and a transient kinetic assay measuring intrinsic tryptophan fluorescence changes. These assays further showed that borrelidin is a noncompetitive tight binding inhibitor of E. coli ThrRS with respect to threonine and ATP. Genetic selection of borrelidin-resistant mutants showed that borrelidin binds to a hydrophobic region (Thr-307, His-309, Cys-334, Pro-335, Leu-489, Leu-493) proximal to the zinc ion at the active site of the E. coli ThrRS. Mutating residue Leu-489 3 Trp reduced the space of the hydrophobic cluster and resulted in a 1500-fold increase of the Ki value from 4 nM to 6 M. An alignment of ThrRS sequences showed that this cluster is conserved in most organisms except for some Archaea (e.g. M. jannaschii, A. fulgidus) and some pathogens (e.g. Helicobacter pylori). This study illustrates how one class of natural product inhibitors affects aminoacyl-tRNA synthetase function, providing potentially useful information for structure-based inhibitor design. Aminoacyl-tRNA synthetases (aaRSs) catalyze the acylation of transfer RNAs with their cognate amino acids and are therefore essential enzymes for protein synthesis in all organisms (1). They display exquisite specificity in discriminating between similar amino acid or tRNA substrates. Even though the core architectures of the active sites of individual aaRSs are relatively conserved, protein sequence alignments showed significant divergence among the three domains of life. However, finely tuned structural differences between aaRS orthologs provide ample opportunities for the evolution of species-specific inhibitors of any given tRNA synthetase. For example, mupirocin, an antibiotic produced by Pseudomonas fluorescens, selectively inhibits the prokaryotic isoleucyl-tRNA synthetases but has little or no effect on the eukaryotic enzymes (2, 3). Indolmycin, a secondary metabolite of Streptomyces griseus and analogue of tryptophan, inhibits only one of the two tryptophanyl-tRNA synthetases in Streptomyces coelicolor (4). ThrRS is a class II aaRS containing an N-terminal editing domain, a C-terminal tRNA binding domain, and a zinc-binding catalytic domain with the class II conserved motifs 1, 2, and 3 that provide critical ATP binding determinants (5). ThrRSs from Chinese hamster ovary cells, Saccharomyces cerevisiae and Escherichia coli (reviewed in Ref. 6) are specifically inhibited by borrelidin, an 18-membered macrolide-polyketide (Fig. 1) produced by Streptomyces spp (7–9). Borrelidin was shown to have anti-malarial activity (10). Furthermore, borrelidin was found to interfere with capillary tube formation, possibly through anti-angiogenesis effects that are mediated through the ThrRS inhibition and the caspase activation pathways (11). Borrelidin also inhibited a S. cerevisiae cyclin-dependent kinase (Cdc28/Cln2), an indication of anti-tumor activity (12). Gene expression profiling of S. cerevisiae revealed that borrelidin up-regulated GCN4 leucine zipper mRNA synthesis (13), and this in turn induced the expression of amino acid biosynthetic enzymes. Because of its intriguing biological activities, the chemical synthesis (14, 15) and biosynthesis (16) of borrelidin have been explored. Presumably as a consequence of borrelidin inhibition of human ThrRS, the compound may also have undesired cytotoxic effects. To develop borrelidin into an important anti-microbial, anti-angiogenesis, and anti-tumor drug candidate, a fuller understanding of the mechanism of borrelidin inhibition of ThrRS is required. Previous reports showed that borrelidin is a noncompetitive inhibitor with respect to threonine (17) and inhib* This work was supported by grants from the National Institute of General Medical Sciences (to D. S. and C. S. F.) and the Department of Energy (to D. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ¶ Current address: Dept. of Biochemistry and Biotechnology, University of Thessaly, 41221 Larissa, Greece. ** To whom correspondence may be addressed: Dept. of Biochemistry, University of Vermont Health Sciences Complex, Burlington, VT 05405-0068. Tel.: 802-656-8450; Fax: 802-862-8229; E-mail: franck@ emba.uvm.edu. §§ To whom correspondence may be addressed: Dept. of Molecular Biophysics and Biochemistry, Yale University, New Haven, CT 06520-8114. Tel.: 203-432-6200; Fax: 203-432-6202; E-mail: soll@trna.chem.yale.edu. 1 The abbreviations used are: aaRS, aminoacyl-tRNA synthetase; ThrRS, threonyl-tRNA synthetase. THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 1, Issue of January 7, pp. 571–577, 2005 © 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

Aminoacyl-tRNA synthetases (aaRSs) 1 catalyze the acylation of transfer RNAs with their cognate amino acids and are therefore essential enzymes for protein synthesis in all organisms (1). They display exquisite specificity in discriminating between similar amino acid or tRNA substrates. Even though the core architectures of the active sites of individual aaRSs are relatively conserved, protein sequence alignments showed significant divergence among the three domains of life. However, finely tuned structural differences between aaRS orthologs provide ample opportunities for the evolution of species-specific inhibitors of any given tRNA synthetase. For example, mupirocin, an antibiotic produced by Pseudomonas fluorescens, selectively inhibits the prokaryotic isoleucyl-tRNA synthetases but has little or no effect on the eukaryotic enzymes (2,3). Indolmycin, a secondary metabolite of Streptomyces griseus and analogue of tryptophan, inhibits only one of the two tryptophanyl-tRNA synthetases in Streptomyces coelicolor (4).
ThrRS is a class II aaRS containing an N-terminal editing domain, a C-terminal tRNA binding domain, and a zinc-binding catalytic domain with the class II conserved motifs 1, 2, and 3 that provide critical ATP binding determinants (5). ThrRSs from Chinese hamster ovary cells, Saccharomyces cerevisiae and Escherichia coli (reviewed in Ref. 6) are specifically inhibited by borrelidin, an 18-membered macrolide-polyketide (Fig. 1) produced by Streptomyces spp (7)(8)(9). Borrelidin was shown to have anti-malarial activity (10). Furthermore, borrelidin was found to interfere with capillary tube formation, possibly through anti-angiogenesis effects that are mediated through the ThrRS inhibition and the caspase activation pathways (11). Borrelidin also inhibited a S. cerevisiae cyclin-dependent kinase (Cdc28/Cln2), an indication of anti-tumor activity (12). Gene expression profiling of S. cerevisiae revealed that borrelidin up-regulated GCN4 leucine zipper mRNA synthesis (13), and this in turn induced the expression of amino acid biosynthetic enzymes. Because of its intriguing biological activities, the chemical synthesis (14,15) and biosynthesis (16) of borrelidin have been explored.
Presumably as a consequence of borrelidin inhibition of human ThrRS, the compound may also have undesired cytotoxic effects. To develop borrelidin into an important anti-microbial, anti-angiogenesis, and anti-tumor drug candidate, a fuller understanding of the mechanism of borrelidin inhibition of ThrRS is required. Previous reports showed that borrelidin is a noncompetitive inhibitor with respect to threonine (17) and inhib-its the transfer of activated threonine to tRNA Thr (18). Characterization of borrelidin resistant mutants showed that resistance resulted from ThrRS overexpression in vivo (19) or from a mutation affecting the K m of threonine (17) After verification of the DNA sequences, the genes were subcloned into the expression vectors described below. Specific mutant alleles were generated by site-directed mutagenesis using the QuikChange system. A library of random E. coli thrS mutants was generated by error-prone PCR using the Mutazyme kit. For in vivo tests, the genes were subcloned into the pCBS vector between the KpnI and BglII sites under trpS promoter control. For protein expression, the genes were cloned into pET20b vectors using the XbaI and BamHI sites.
In Vivo Testing for ThrRS Activity and Borrelidin Inhibition-ThrRS activity was tested by complementation of E. coli thrS ts strain with thrS mutants at 42°C on minimal medium plate containing casamino acids. Borrelidin-resistant genes were selected on minimal medium plate containing borrelidin. The E. coli thrS mutant library was transformed into W3110, and the transformants were grown in glucose M56 minimal medium (20) in the presence of borrelidin (2 mM) overnight. Then the culture was plated on minimal agar plates containing borrelidin to isolate individual borrelidin-resistant colonies for plasmid extraction and further DNA sequencing. The "borrelidin-resistant" thrS mutants were subcloned into pET20b and expressed for biochemical characterization.
Preparation of His 6 -tagged ThrRS-The pET20b-thrS clones were transformed into E. coli BL21 (Codonplus DE3) cells, and the transformants were grown in 500 ml of LB medium with ampicillin (100 g/ml) to a cell density of A 600 ϭ 0.4. At this point, the expression of C-terminal His 6 -tagged thrS was induced with 1 mM isopropyl 1-thio-␤-D-galactopyranoside for 2 h. Cells were harvested and lysed, and the wild-type and mutant His 6 -ThrRS proteins were prepared from the cell extracts by nickel-nitrilotriacetic acid-agarose chromatography to Ͼ95% purity as judged by SDS-PAGE after Coomassie Brilliant Blue staining. Active site titration (21) was performed for ThrRS preparations by incubating each enzyme in 100 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 10 mM KCl, 5 mM dithiothreitol, 4 mM ATP (pH 7), 1 mM [ 14 C]threonine (200 cpm/ pmol), and 1 unit inorganic pyrophosphatase in a total volume of 0.1 ml at 37°C for 15-20 min. The wild-type ThrRS enzymes were 80 -100% active, whereas the activity of the mutant enzymes varied from 0.1-100% active. These values were used to determine enzyme concentration for calculation of k cat . The enzymes were stored in 40% glycerol buffer (60 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 30 mM KCl, 5 mM dithiothreitol) at Ϫ20°C.
Aminoacylation of tRNA-Thr-tRNA formation was measured by acid-precipitable radioactivity (21) in 0.1 ml of reaction mixtures containing 60 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 30 mM KCl, 5 mM dithiothreitol, 0.02-10 mM ATP, 0.01-1 mM [ 3 H]threonine, 4 mg/ml total tRNA (E. coli tRNA for E. coli ThrRSs and M. jannaschii tRNA for other ThrRSs), 0 -6000 nM borrelidin, and 0.1-100 nM enzyme. Reactions were carried out in triplicate at 37°C for E. coli and Halobacterium ThrRS, at 55°C for S. solfataricus ThrRS, and 65°C for M. jannaschii and A. fulgidus ThrRS. For steady state kinetic constants K m and k cat measurement, the enzymes were diluted to obtain linear initial velocities. K m and k cat were calculated in KaleidaGraph 3.0 using a Michaelis-Menten equation fit. For K i measurements, saturating amounts of substrates (0.5 mM threonine, 1 mM ATP, 4 mg/ml tRNA) were mixed with various amounts of borrelidin (0 -6 M). The reaction mixture (ATP, tRNA, ThrRS enzyme) was preincubated in the presence of borrelidin at 37°C for 3 min, and then threonine was added to initiate the reaction. All experiments were done in triplicate and repeated at least twice. The initial velocities obtained were plotted against borrelidin concentration to obtain IC 50 values. The apparent K i (app) values were calculated based on the simplified equation (22) for noncompetitive tight binding inhibition.
Presteady State Analysis-Intrinsic tryptophan fluorescence changes of the E. coli ThrRS enzyme upon ligand binding were measured by stopped flow and steady-state fluorescence as described previously (23).  (24,25) showed the existence of markedly different bacterial/eukaryal and archaeal versions of ThrRS. There are large deletions or additions in the N-terminal region of these enzymes; archaeal ThrRS has a shorter N-terminal region, an insertion domain at motif 3 (26), and many conserved differences in the catalytic domain. In this work we investigated the S. solfataricus and Halobacterium sp. NRC-1 ThrRS, which are more closely related to the bacterial genre enzyme, and the archaeal genre enzymes from A. fulgidus and M. jannaschii. The influence of borrelidin on the enzyme activity was measured in the aminoacylation assay. As shown in Fig. 2, ThrRS from M. jannaschii and A. fulgidus were not inhibited at 1 M borrelidin, whereas those of S. solfataricus and Halobacterium were inhibited. To compare these enzymes under similar conditions, the ATP-PP i exchange reaction was used for steady-state kinetic measurements. As shown in Table I main was observed between the borrelidin-inhibited ThrRS and the uninhibited type. Also, S. solfataricus contains two thrS related genes (25,27); the archaeal genre ThrRS lacks the putative catalytic domain, and the bacterial genre ThrRS lacks most of the entire N-terminal editing domain. The bacterial genre S. solfataricus ThrRS can be inhibited by borrelidin. To determine whether the N-terminal region contributes to borrelidin binding, we aminoacylated E. coli tRNA with wild-type E. coli ThrRS and M. jannaschii tRNA with the A. fulgidus enzyme in the presence or absence of borrelidin. A similar experiment was carried out with the corresponding N-terminal domain-deleted enzymes. As shown in Fig. 3, the E. coli Nterminal-deleted ThrRS was inhibited by borrelidin to the same extent as the wild-type enzyme; the A. fulgidus N-terminal-deleted ThrRS remained uninhibited. Therefore, the Nterminal domain does not contain a borrelidin binding site.

Borrelidin Inhibition of Archaeal and Bacterial ThrRS Enzymes-Phylogenetic analysis of ThrRS sequences
Borrelidin Is a Slow but Tight Binding Inhibitor of E. coli ThrRS-The K i (app) value (3.7 nM) for borrelidin inhibition of E. coli ThrRS was very close to the enzyme concentration (1-2 nM) used, and a linear dose-response of IC 50 to the enzyme concentration at a fixed substrate concentration was also observed (data not shown). These results suggested that borrelidin binding is tight and appears to be stoichiometric with dimeric ThrRS. Estimates of on and off rates as well as apparent K d values for borrelidin were determined from stopped flow fluorescence experiments. As shown in Fig. 4A, borrelidin binding quenched ThrRS fluorescence; this is clear evidence that the method of inhibition cannot be uncompetitive. Single exponen-tial curve fits were used to measure the apparent first order rate constants at a variety of borrelidin concentrations. The apparent rate constants were plotted with respect to borrelidin concentration as shown in Fig. 4B. Rates varied linearly with increasing borrelidin, suggesting a single borrelidin binding step. Estimates of k on and k off can be read directly from the slope and y-intercept, respectively. From this experiment, k on was estimated to be about 5 ϫ 10 6 M Ϫ1 s Ϫ1 . The y-intercept could not be accurately determined because of a requirement for a relatively high enzyme and borrelidin concentrations, but a conservative upper bound of 0.1 s Ϫ1 fixes the K d (k off /k on ) at no more than 20 nM; reasonably close to the 4 nM inhibitory constant. Therefore, borrelidin is a tight binding inhibitor of ThrRS. Because of the small off rate, overexpression of thrS might be expected to confer apparent borrelidin resistance to E. coli by virtue of titration of the drug with the excess enzyme (19).
Borrelidin Inhibits E. coli ThrRS at the Step of Threonine Activation-ThrRS catalyzes Thr-tRNA Thr formation in two steps (6); first it activates threonine to form the adenylate (Thr-AMP), and then it transfers the activated amino acid onto tRNA Thr to generate Thr-tRNA Thr . Borrelidin was reported to    Error bars indicate the 95% confidence limit. The slope and y-intercept represent estimates of k on and k off , respectively. inhibit the second step, the transfer of activated threonine to tRNA Thr (18). However, because borrelidin is a tight binding inhibitor, the IC 50 values could vary remarkably depending on the enzyme concentration. Therefore, we measured the IC 50 value of the same amount of E. coli ThrRS enzyme determined by both ATP-PP i exchange and aminoacylation. ATP-PP i exchange reaction measures the kinetic parameter of the first step, threonine activation; aminoacylation determines the overall formation of Thr-tRNA Thr formation. Fig. 5, A and B show that the IC 50 values were similar in both reactions; the calculated K i (app) for E. coli ThrRS in the ATP-PP i exchange reaction (4.2 nM) is very similar to the K i (app) of 3.7 nM in aminoacylation, indicating that borrelidin inhibits E. coli ThrRS at the first step of threonine activation and thus the overall aminoacylation. To confirm this, presteady-state kinetic analysis of the ThrRS-catalyzed adenylation reaction was carried out in the presence of borrelidin. Fig. 5C shows transient changes in the intrinsic tryptophan fluorescence of ThrRS in the presence or absence of substrates and inhibitor. All data traces were well described by fits to a single exponential function. In the presence of threonine and ATP, the enzyme fluorescence is quenched by 10%, with half the quench amplitude occurring within the dead time of the instrument (bottom fitted trace). This instantaneous quench reflects the enzyme response to the rapid binding of the substrates, whereas the fitted transient corresponds to an isomerization that precedes the chemical step of adenylation (23). In the absence of substrates the fluorescence change is negligible (top trace). The addition of various concentrations of borrelidin to ThrRS prior to rapid mixing with threonine and ATP abolishes the adenylationassociated quench but not that associated with substrate binding, as expected for a noncompetitive inhibitor. At the highest borrelidin concentration, the fluorescence quench includes a small contribution from the prebound borrelidin in addition to the rapid binding of threonine and ATP. Taken together, both steady-state and presteady-state kinetic data showed that borrelidin inhibits ThrRS at the activation step rather than at the transfer step (18).
Borrelidin Is a Noncompetitive Inhibitor of E. coli ThrRS with Respect to Threonine and ATP-All the data above indicated that borrelidin may bind near or at the active site. To examine whether borrelidin has overlapping binding sites with the other enzyme substrates, we evaluated the inhibition mechanism with respect to threonine and ATP. The Michaelis-Menten derived double-reciprocal plot is commonly used to distinguish the inhibition pattern of reversible inhibitors with the assumption that the free inhibitor concentration could approximate the total concentration of the added inhibitor. However, borrelidin is a tight binding inhibitor that inhibits ThrRS at a very low concentration that is close to the enzyme concentration. Under these conditions the concentration of the enzyme-inhibitor complex is not negligible compared with the inhibitor concentration, so we can not use the double-reciprocal plot method to determine the inhibition patterns of borrelidin.
The inhibition pattern was then determined according to the equations derived for tight binding inhibitors (22,28,29). Several graphic approaches are available, and the most direct one is to determine the IC 50 values for the inhibitor at a fixed enzyme concentration but at different substrate concentrations. For the competitive type, IC 50 values increase linearly with increasing substrate concentration; that of the uncompetitive type is linear with the reciprocal value of the substrate concentration, i.e. IC 50 values decrease sharply with the increasing substrate concentration. For the noncompetitive type, the relationship between IC 50 values and substrate concentration varies depending on the dissociation constant of the inhibitor-enzyme-substrate complex (␣K i ) and the inhibitor-enzyme complex (K i ). As shown in Fig. 6A, when K i equals ␣K i , the equation for a tight binding competitive inhibitor reduces to IC 50 ϭ K i ϩ [E]/2, indicating that substrate binding does not influence the binding of inhibitor to ThrRS enzyme. If ␣K i Ͻ K i , the inhibitor-enzyme-substrate complex formation is favored over the inhibitor-enzyme complex, indicating that the substrate has a positive influence on inhibitor binding to the enzyme. If ␣K i Ͼ K i , the substrate must have a negative influence on inhibitor binding to the enzyme.
Each IC 50 value shown in Fig. 6, B and C was the average of IC 50 values determined by dose-response curves of ThrRS activity to borrelidin concentration from repeated triplicate experiments with a less than 10% error. The IC 50 values we obtained at different ATP or threonine concentrations showed slightly higher values at low ATP or threonine concentrations but essentially the same at higher substrate concentration, indicating a noncompetitive pattern. Further, we curve fitted data points of IC 50 values against substrate concentrations using the equation for a tight binding noncompetitive inhibitor. threonine indicated that the conformational changes in E. coli ThrRS upon ATP or threonine binding (23) have some positive influence on borrelidin binding to ThrRS enzymes, although the positive effect is too small to be an uncompetitive pattern. Because borrelidin is a noncompetitive inhibitor with respect to threonine and ATP, borrelidin is not likely to have an overlapping binding site with that of ATP and threonine.
Screening of ThrRS Mutants Indicates Possible Binding Sites for Borrelidin-Possible binding sites at the catalytic domain were located by screening a library of randomly mutagenized E. coli thrS for borrelidin-resistant alleles. Eight of ten resistant colonies contained the same Leu489Met thrS allele; another resistant allele is thrS Pro296Ser. Both mutants were subcloned for overexpression as His-tagged proteins, and the purified recombinant enzymes were characterized in vitro. The kinetic data in Table II show that the selected mutant enzymes had a similar K m value for threonine, an up to 2-fold higher K m value for ATP, and an up to 2-fold higher K i value for borrelidin compared with those of the wild type. Therefore, residues at position 489 and 296 are relevant to borrelidin inhibition.
Further, thrS alleles, containing mutations of the predicted contact residues and insertion or deletion constructs of the unique motif three sequence (EKG) of the archeal genre ThrRS, were made by PCR mutagenesis (the mutant changes are described in Fig. 7). After overexpression and purification, the K i (app) for these enzymes was determined by aminoacylation under the same experimental conditions. Fig. 7 shows the K i (app) of 15 ThrRS mutants at fixed enzyme concentration (2 nM). Deleting the N-terminal 167 amino acids as well as a deletion of Motif 3 sequence (⌬485-492) in A. fulgidus ThrRS did not affect enzyme activity, and these enzymes were not inhibited by borrelidin. However, insertion of the three amino acid archaeal motif three sequence (26) into E. coli ThrRS greatly reduced enzyme activity, and the ThrRSHis309Ala enzyme showed very weak aminoacylation activity. The other mutant enzymes had good in vitro activity. E. coli mutant ThrRS enzymes with single amino acid changes (residues 335, 337, 489, 307, and 309) showed 2-1500-fold higher K i (app) val-ues; mutations in other residues had no effect on borrelidin inhibition. The most sensitive residue is leucine 489, located in the middle of the hydrophobic cluster A. Replacement of this Leu with Trp increased the K i value 1500-fold. These in vitro results were confirmed by in vivo tests of ThrRS activity and borrelidin inhibition. ThrRS activity was tested by complementation of a temperature-sensitive E. coli thrS strain, whereas borrelidin resistance was checked by growth on minimal medium containing 2 mM borrelidin (Fig. 8).
With the exception of the E. coli thrS 531GEKG strain all mutant thrS strains grew at the non-permissive temperature (Fig. 8, left panel). N-terminal-deleted thrS and wild-type thrS were borrelidin sensitive; however, thrSLeu489Met, thrSLeu489Trp, thrSHis337Ala, and thrSHis309Ala gave rise to borrelidin-resistant transformants (Fig. 8, right panel). These data support our conclusion that cluster A is the borrelidin binding site.
A representation (Fig. 9) of cluster A was generated in the PyMOL program (32) using the crystallographic coordinates of E. coli ThrRS (Protein Data Bank accession code 1QF6). This region is located at the "back" of the active site. Cys-334 of cluster A directly contacts the zinc ion, which is essential for ThrRS activity. Borrelidin binding to cluster A may cause relocation of Cys-334, resulting in further distortion of the zinc ion coordination and disruption of ThrRS activity (31). As cluster A is very close to the binding sites for threonine, ATP, and the terminal A residue of tRNA Thr , substrate binding may also influence cluster A conformation. The fact that other aaRSs do not have such a hydrophobic core in this part of their active site may explain why borrelidin only inhibits ThrRS but not any other aaRSs.
Comparison of the Putative Borrelidin Binding Region Among ThrRSs-Alignment of a selection of archaeal, bacterial, and eukaryotic ThrRS protein sequences was performed to reveal the extent of sequence conservation of the E. coli ThrRS cluster A residues (Thr-307, His-309, Cys-334, Pro-335, His-337, Leu-489, and Leu-493) position. The Cys-334 residue is conserved in all ThrRS enzymes. The above mentioned cluster A residues (except for Thr-307) are found in the corresponding positions of the human and S. cerevisiae enzyme, both of which are inhibited by borrelidin. Furthermore, Plasmodium ThrRS has the same residues, which may explain its observed antimalarial activity (10). Some bacteria, including some pathogens (e.g. Helicobacter pylori and Mycobacteria), have amino acid replacements in some of the conserved residues in cluster A. The influence of these variations on borrelidin resistance by the corresponding ThrRS enzymes remains to be investigated.
However, the archaeal genre ThrRS, found in M. jannaschii, A. fulgidus, Methanococcus maripaludis, Methanopyrus kandleri, Pyrococci, and Methanosarcineae, contains a different set of amino acids in the corresponding cluster A positions with the exception of the conserved Cys334 (Fig. 10). The absence of cluster A and the resulting borrelidin binding site may explain the observed borrelidin resistance of the M. jannaschii and A. fulgidus enzymes.
General Discussion-The work presented here indicates that borrelidin is a novel type of aminoacyl-tRNA synthetase inhib-itor. In contrast, most tRNA synthetase inhibitors bind at the active site or bind to the tRNA substrate. For example, indolmycin, a tryptophan analog, inhibits tryptophanyl-tRNA synthetase (4), aminoacyl adenylate sulfamates and sulfonamides inhibit aaRS as mimics of the activated substrate aminoacyl-AMP (33), and oligonucleotide inhibitors mimic tRNA features (34). These inhibitors compete with the substrates for binding to the active site and generally are reversible inhibitors. Mupirocin, a topical antimicrobial agent against methicillin-resistant Staphylococcus aureus, binds to the active site of isoleucyl-tRNA synthetase and inactivates the enzyme by incorrect product shuttling between the synthetic and the editing active sites (2,3). Tobramycin, an aminoglycoside, inhibits acylation of tRNA Asp by binding to tRNA Asp (35). Distinct from the active site inhibitors, borrelidin inhibits bacterial genre ThrRS by binding to the unique cluster A proximal to the Zn 2ϩ binding site that includes the Cys-334 residue (31). The absence of the cluster A in other tRNA synthetases and archaeal ThrRS is consistent with the lack of borrelidin inhibition in these cases. The inhibition mechanism of borrelidin explored in this work could facilitate structure-based inhibitor design for effective drugs with anti-angiogenesis and anti-tumor activity.   The ␣ carbon backbone of the ThrRS active site is rendered as a red ribbon. The bound AMP and tRNA are sticks rendered in green and blue, respectively. The coordinated zinc atom is denoted as a yellow sphere. The putative borrelidin binding cluster A (Leu-493, His-307, His-309, and Pro-335) is rendered in space filling mode as orange spheres; Leu-489 is highlighted in green, and Cys-334 is highlighted in purple. ThrRS is colored as red ribbon. The putative hydrophobic cluster B is (Met-509 and Val-338) is rendered as blue spheres.