Rediscovery of PF-3845 as a new chemical scaffold inhibiting phenylalanyl-tRNA synthetase in Mycobacterium tuberculosis

Mycobacterium tuberculosis (Mtb) remains the deadliest pathogenic bacteria worldwide. The search for new antibiotics to treat drug-sensitive as well as drug-resistant tuberculosis has become a priority. The essential enzyme phenylalanyl-tRNA synthetase (PheRS) is an antibacterial drug target because of the large differences between bacterial and human PheRS counterparts. In a high-throughput screening of 2148 bioactive compounds, PF-3845, which is a known inhibitor of human fatty acid amide hydrolase, was identified inhibiting Mtb PheRS at Ki ∼ 0.73 ± 0.06 μM. The inhibition mechanism was studied with enzyme kinetics, protein structural modeling, and crystallography, in comparison to a PheRS inhibitor of the noted phenyl–thiazolylurea–sulfonamide class. The 2.3-Å crystal structure of Mtb PheRS in complex with PF-3845 revealed its novel binding mode, in which a trifluoromethyl–pyridinylphenyl group occupies the phenylalanine pocket, whereas a piperidine–piperazine urea group binds into the ATP pocket through an interaction network enforced by a sulfate ion. It represents the first non-nucleoside bisubstrate competitive inhibitor of bacterial PheRS. PF-3845 inhibits the in vitro growth of Mtb H37Rv at ∼24 μM, and the potency of PF-3845 increased against an engineered strain Mtb pheS–FDAS, suggesting on target activity in mycobacterial whole cells. PF-3845 does not inhibit human cytoplasmic or mitochondrial PheRS in biochemical assay, which can be explained from the crystal structures. Further medicinal chemistry efforts focused on the piperidine–piperazine urea moiety may result in the identification of a selective antibacterial lead compound. Mycobacterium tuberculosis (Mtb) remains the deadliest pathogenic bacteria worldwide. The search for new antibiotics to treat drug-sensitive as well as drug-resistant tuberculosis has become a priority. The essential enzyme phenylalanyl-tRNA synthetase (PheRS) is an antibacterial drug target because of the large differences between bacterial and human PheRS counterparts. In a high-throughput screening of 2148 bioactive compounds, PF-3845, which is a known inhibitor of human fatty acid amide hydrolase, was identified inhibiting Mtb PheRS at Ki ∼ 0.73 ± 0.06 μM. The inhibition mechanism was studied with enzyme kinetics, protein structural modeling, and crystallography, in comparison to a PheRS inhibitor of the noted phenyl–thiazolylurea–sulfonamide class. The 2.3-Å crystal structure of Mtb PheRS in complex with PF-3845 revealed its novel binding mode, in which a trifluoromethyl–pyridinylphenyl group occupies the phenylalanine pocket, whereas a piperidine–piperazine urea group binds into the ATP pocket through an interaction network enforced by a sulfate ion. It represents the first non-nucleoside bisubstrate competitive inhibitor of bacterial PheRS. PF-3845 inhibits the in vitro growth of Mtb H37Rv at ∼24 μM, and the potency of PF-3845 increased against an engineered strain Mtb pheS–FDAS, suggesting on target activity in mycobacterial whole cells. PF-3845 does not inhibit human cytoplasmic or mitochondrial PheRS in biochemical assay, which can be explained from the crystal structures. Further medicinal chemistry efforts focused on the piperidine–piperazine urea moiety may result in the identification of a selective antibacterial lead compound. Protein synthesis is the cellular process targeted by many commercial antibiotics. It has always been a focal point of modern antibacterial drug discovery (1Fischbach M.A. Walsh C.T. Antibiotics for emerging pathogens.Science. 2009; 325: 1089-1093Crossref PubMed Scopus (1199) Google Scholar). Aminoacyl-tRNA synthetases (aaRSs), a family of ∼20 essential enzymes, ligate amino acids to the corresponding tRNAs that decode messenger RNA to produce protein at the translating macromolecular ribosome (2Schimmel P.R. Söll D. Aminoacyl-tRNA synthetases: general features and recognition of transfer RNAs.Annu. Rev. Biochem. 1979; 48: 601-648Crossref PubMed Scopus (455) Google Scholar). Inhibition of bacterial aaRS blocks the translation and ultimately shuts down protein synthesis, which is crucial for pathogens to survive in host or inside host cells (3Francklyn C.S. Mullen P. Progress and challenges in aminoacyl-tRNA synthetase-based therapeutics.J. Biol. Chem. 2019; 294: 5365-5385Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar, 4Kwon N.H. Fox P.L. Kim S. Aminoacyl-tRNA synthetases as therapeutic targets.Nat. Rev. Drug Discov. 2019; 18: 629-650Crossref PubMed Scopus (55) Google Scholar). Between a bacterial aaRS protein and its human counterpart, either the large sequence difference or small variation of key residues in the catalytic core explains the high selectivity of successful and promising aaRS inhibitor drugs, as exemplified by the mupirocin used for the treatment of staphylococcal infection (4Kwon N.H. Fox P.L. Kim S. Aminoacyl-tRNA synthetases as therapeutic targets.Nat. Rev. Drug Discov. 2019; 18: 629-650Crossref PubMed Scopus (55) Google Scholar). Tuberculosis (TB) caused by the single agent Mycobacterium tuberculosis (Mtb) has surpassed AIDS/HIV, becoming a leading infectious disease worldwide (https://www.who.int/news-room/fact-sheets/detail/tuberculosis). Most TB drugs were discovered in the past century and are losing efficacy because of the resistance inevitably arisen in bacteria. New chemical scaffold with novel inhibition mechanism against Mtb is being actively sought. An oxaborole compound GSK3036656 that inhibits Mtb leucyl-tRNA synthetase is currently undergoing clinical trial (5Tenero D. Derimanov G. Carlton A. Tonkyn J. Davies M. Cozens S. Gresham S. Gaudion A. Adeep P. Muliaditan M. Rullas-Trincado J. Mendoza-Losana Alfonso Skingsley A. Barros-Aguirre D. First-time-in-human study and prediction of early bactericidal activity for GSK3036656 , a potent leucyl-tRNA synthetase inhibitor for tuberculosis treatment.Antimicrob. Agents Chemother. 2019; 63: 1-15Crossref Scopus (21) Google Scholar). Bacterial phenylalanyl-tRNA synthetase (PheRS) is a member of class II aaRS based on the active site topology (6Safro M. Moor N. Lavrik O. Phenylalanyl-tRNA synthetases.in: Ibba M. Francklyn C. Cusack S. The Aminoacyl-tRNA Synthetases. Landes Bioscience/Eurekah.com, Georgetown, TX2005: 265-280Google Scholar). A functional PheRS is typically made of two heterodimers (αβ)2, with a whole molecular weight around 250 kD. In Mtb H37Rv genome, two consecutive essential genes, Rv1649 and Rv1650 (pheST), encode the protein subunits (7Dejesus M.A. Gerrick E.R. Xu W. Park S.W. Long J.E. Boutte C.C. Rubin E.J. Schnappinger D. Ehrt S. Fortune S.M. Sassetti C.M. Ioerger T.R. Comprehensive essentiality analysis of the Mycobacterium tuberculosis genome via saturating transposon mutagenesis.MBio. 2017; 8: 1-17Crossref Scopus (203) Google Scholar). Like other aaRSs, PheRS catalyzes the formation of phenylalanyl-tRNAPhe in two steps: (1) activation of phenylalanine (Phe) by hydrolyzing ATP to form Phe-AMP and pyrophosphate (PPi) and (2) subsequent transfer of Phe to the 2′-OH group of adenosine ribose at the 3′-terminal of tRNAPhe and simultaneous release of AMP. In addition to the synthetic site on the α subunit, PheRS has a dedicated domain on the β subunit that can hydrolyze improperly charged tRNA, for example, tyrosinyl-tRNAPhe, to maintain the fidelity of aminoacylation and translation (8Roy H. Ling J. Irnov M. Ibba M. Post-transfer editing in vitro and in vivo by the β subunit of phenylalanyl-tRNA synthetase.EMBO J. 2004; 23: 4639-4648Crossref PubMed Scopus (116) Google Scholar). From a drug discovery standing point, PheRS synthetic activity that involves binding of the three substrates and the editing activity could all be targetable. PheRS should also allow specific binding of compounds that allosterically regulate the enzymatic activities because of its size, molecular mechanism, and other miscellaneous functions (9Lechler A. Kreutzer R. The phenylalanyl-tRNA synthetase specifically binds DNA.J. Mol. Biol. 1998; 278: 897-901Crossref PubMed Scopus (24) Google Scholar). PheRS was a major antibacterial target in a number of screening programs (10Gallant P. Finn J. Keith D. Wendler P. The identification of quality antibacterial drug discovery targets: a case study with aminoacyl-tRNA synthetases.Expert Opin. Ther. Targets. 2000; 4: 1-9Google Scholar, 11Beyer D. Kroll H.P. Endermann R. Schiffer G. Siegel S. Bauser M. Pohlmann J. Brands M. Ziegelbauer K. Haebich D. Eymann C. Brötz-Oesterhelt H. New class of bacterial phenylalanyl-tRNA synthetase inhibitors with high potency and broad-spectrum activity.Antimicrob. Agents Chemother. 2004; 48: 525-532Crossref PubMed Scopus (59) Google Scholar, 12Payne D.J. Gwynn M.N. Holmes D.J. Pompliano D.L. Drugs for bad bugs: confronting the challenges of antibacterial discovery.Nat. Rev. Drug Discov. 2007; 6: 29-40Crossref PubMed Scopus (1805) Google Scholar, 13Montgomery J.I. Toogood P.L. Hutchings K.M. Liu J. Narasimhan L. Braden T. Dermyer M.R. Kulynych A.D. Smith Y.D. Warmus J.S. Taylor C. Discovery and SAR of benzyl phenyl ethers as inhibitors of bacterial phenylalanyl-tRNA synthetase.Bioorg. Med. Chem. Lett. 2009; 19: 665-669Crossref PubMed Scopus (16) Google Scholar, 14Zhang Z.-M. Sun Y. Zhao L.-L. Wang L.-N. Wei Y.-Z. Su J. Zhang Y.-Q. Yu L.-Y. The establishment and application of a high throughput screening assay for inhibitors of Mycobacterium tuberculosis phenylalanyl-tRNA synthetase.Microbiol. China. 2012; 39: 1437-1446Google Scholar, 15Abibi A. Ferguson A.D. Fleming P.R. Gao N. Hajec L.I. Hu J. Laganas V.A. McKinney D.C. McLeod S.M. Prince D.B. Shapiro A.B. Buurman E.T. The role of a novel auxiliary pocket in bacterial phenylalanyl-tRNA synthetase druggability.J. Biol. Chem. 2014; 289: 21651-21662Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar, 16Tommasi R. Brown D.G. Walkup G.K. Manchester J.I. Miller A.A. ESKAPEing the labyrinth of antibacterial discovery.Nat. Rev. Drug Discov. 2015; 14: 529-542Crossref PubMed Scopus (304) Google Scholar, 17Hu Y. Keniry M. Palmer S.O. Bullard J.M. Discovery and analysis of natural-product compounds inhibiting protein synthesis in Pseudomonas aeruginosa.Antimicrob. Agents Chemother. 2016; 60: 4820-4829Crossref PubMed Scopus (13) Google Scholar). Bacterial PheRS-specific hits such as phenyl–thiazolylurea–sulfonamides have emerged from synthetic compound library. To the best of our knowledge, most of them were not for TB drug except one that identified a fungal natural product isopatulin able to inhibit the growth of Mtb strain H37Rv (18Wang L.-N. Di W.-J. Zhang Z.-M. Zhao L. Zhang T. Deng Y.-R. Yu L.-Y. Small-molecule inhibitors of the tuberculosis target, phenylalanyl-tRNA synthetase from Penicillium griseofulvum CPCC-400528.Cogent Chem. 2016; 2: 1-9Crossref Google Scholar). Recently, PheRS from Plasmodium falciparum, which is another pathogen of global health concern, has become genetically and chemically validated target for antiparasitics (19Cowell A.N. Winzeler E.A. Advances in omics-based methods to identify novel targets for malaria and other parasitic protozoan infections.Genome Med. 2019; 11: 1-17Crossref PubMed Scopus (22) Google Scholar). Potent PheRS inhibitors with novel mechanisms at the molecular and cellular levels were discovered (20Kato N. Comer E. Sakata-Kato T. Sharma A. Sharma M. Maetani M. Bastien J. Brancucci N.M. Bittker J.A. Corey V. Clarke D. Derbyshire E.R. Dornan G.L. Duffy S. Eckley S. et al.Diversity-oriented synthesis yields novel multistage antimalarial inhibitors.Nature. 2016; 538: 344-349Crossref PubMed Scopus (141) Google Scholar), prompting us to initiate a hit discovery campaign targeting Mtb PheRS with new assays and diversity compound libraries, such as recent collection of pharmacologically relevant small molecules (21Janes J. Young M.E. Chen E. Rogers N.H. Burgstaller-Muehlbacher S. Hughes L.D. Love M.S. Hull M.V. Kuhen K.L. Woods A.K. Joseph S.B. Petrassi H.M. McNamara C.W. Tremblay M.S. Su A.I. et al.The ReFRAME library as a comprehensive drug repurposing library and its application to the treatment of cryptosporidiosis.Proc. Natl. Acad. Sci. U. S. A. 2018; 115: 10750-10755Crossref PubMed Scopus (56) Google Scholar). In this work, Mtb PheRS enzyme was enzymatically characterized and analyzed with a synthesized reference compound GDI05-001. High-throughput screening (HTS) assay was adopted to screen a Selleck-2148 library against the PheRS. One screening hit, PF-3845, was confirmed to inhibit Mtb PheRS with Ki ∼ 0.73 μM. It is a well-developed inhibitor of human fatty acid amide hydrolase (FAAH). Its minimum concentration required to inhibit >90% growth of WT H37Rv (MIC) is ∼24 μM, and the antibiotic activity increased against an engineered strain Mtb pheS–FDAS, which expresses PheS contained a C-terminal degradation tag Flag-DAS. The mechanism of inhibition was studied with enzyme kinetics and crystal structures. PF-3845 was shown to be a non-nucleoside bisubstrate competitive inhibitor of the PheRS. The results will aid future medicinal chemistry effort to improve the potency, reduce the toxicity, and develop combination therapy with TB drugs. In Mtb H37Rv genome, pheST genes encode the α and β subunits of PheRS, respectively. The deduced amino acid sequence of PheS is ∼16 to 30% identical to human cytoplasmic hFARS1α or mitochondrial hFARS2. The mitochondrial PheRS does not have a β subunit. The identity of PheT versus hFARS1β is below 12%. Three conserved signature motifs of class II aaRS are identified in Mtb PheS sequence (Fig. S1). The ∼3.5 kb pheST was synthesized based on the DNA sequences in GenBank. After sequencing confirmation, the DNA was cloned for heterologous expression in Escherichia coli. Several constructs were made for copurification of the heterodimeric protein. The best one, pET30a-Mtb PheRS-CHis, under optimized fermentation conditions, gave a yield of ∼1 to 2 mg/l, with the highest purity possible after optimized purification steps (Fig. S2). We also constructed, overexpressed, and partially purified an Mtb tRNAGAAPhe from E. coli (see Experimental procedures section). It was reported that more than ∼90% of the enriched and partially purified tRNA fraction could be the overexpressed tRNA species (22Martin F. Eriani G. Eiler S. Moras D. Dirheimer G. Gangloff J. Overproduction and purification of native and queuine-lacking Escherichia coli tRNAAsp: role of the wobble base in tRNAAsp acylation.J. Mol. Biol. 1993; 234: 965-974Crossref PubMed Scopus (25) Google Scholar, 23Du X. Wang E.D. Discrimination of tRNALeu isoacceptors by the mutants of Escherichia coli Leucyl-tRNA synthetase in editing.Biochemistry. 2002; 41: 10623-10628Crossref PubMed Scopus (9) Google Scholar). We used standard aminoacylation assay that detects the charging of 14C-labeled Phe onto tRNAPhe to confirm the purified protein and tRNA were indeed functional together. Michaelis–Menten kinetic parameters of Mtb PheRS were measured with the tRNA aminoacylation assay and shown in Figure S3. The Km's and kcat's generally agree with previous reports on Mtbs and other bacterial PheRSs although some discrepancies are noted. For example, we and others found the Km with regard to Phe at ∼68.0 ± 6.6 μM, but this is much higher than the ∼1 to 7 μM of other bacterial PheRSs (14Zhang Z.-M. Sun Y. Zhao L.-L. Wang L.-N. Wei Y.-Z. Su J. Zhang Y.-Q. Yu L.-Y. The establishment and application of a high throughput screening assay for inhibitors of Mycobacterium tuberculosis phenylalanyl-tRNA synthetase.Microbiol. China. 2012; 39: 1437-1446Google Scholar, 15Abibi A. Ferguson A.D. Fleming P.R. Gao N. Hajec L.I. Hu J. Laganas V.A. McKinney D.C. McLeod S.M. Prince D.B. Shapiro A.B. Buurman E.T. The role of a novel auxiliary pocket in bacterial phenylalanyl-tRNA synthetase druggability.J. Biol. Chem. 2014; 289: 21651-21662Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar, 24Ibba M. Kast P. Hennecke H. Substrate specificity is determined by amino acid binding pocket size in Escherichia coli phenylalanyl-tRNA synthetase.Biochemistry. 1994; 33: 7107-7112Crossref PubMed Scopus (97) Google Scholar, 25Moor N. Klipcan L. Safro M.G. Bacterial and eukaryotic phenylalanyl-tRNA synthetases catalyze misaminoacylation of tRNA Phe with 3,4-dihydroxy-L-phenylalanine.Chem. Biol. 2011; 18: 1221-1229Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar). The parameters are subject to revision in further research of PheRS biology and aromatic amino acid metabolism. Nonradioactive biochemical assays were adopted for screening for PheRS inhibitor and the subsequent analysis of mode of inhibition. In a continuous spectrophotometric assay (26Lloyd A.J. Thomann H.U. Ibba M. Söll D. A broadly applicable continuous spectrophotometric assay for measuring aminoacyl-tRNA synthetase activity.Nucleic Acids Res. 1995; 23: 2886-2892Crossref PubMed Scopus (44) Google Scholar), the PPi production is coupled to generation of phosphate by inorganic pyrophosphatase (PPase); the phosphate then serves as a substrate in a purine nucleoside phosphorylase (PNPase)–catalyzed reaction to cleave nucleoside compound 2-amino-6-mercapto-7-methylpurine ribonucleoside (MESG) into ribose-1-phosphate and base 2-amino-6-mercapto-7-methylpurine that has absorbance at 360 nm. When optimizing the so-called PPi production assay, we noticed that the A360 readout is tRNA dependent as expected, meaning no PPi was released in the absence of tRNA (Fig. 1A). PheRS concentration at 100 nM per reaction gave a linear response in the first 6 min of the progression. By measuring the initial velocity of PPi production, the apparent Km, ATP with regard to ATP was found to be 162.50 ± 11.62 μM, Km, Phe 18.10 ± 2.44 μM, and Km, tRNA 0.13 ± 0.05 μg/ml (Fig. 1, B–D). These Km's were used as references in the following screening and hit confirmation. To screen compound library, we used a commercial kit Kinase-Glo, which measures the remaining amount of ATP in a PheRS reaction. Readout is luminescent signal generated by luciferase-catalyzed oxidation of luciferin in the presence of the ATP. The readout is thus inversely correlated with the amount of PheRS activity. Because ATP at 1 μM to initiate the PheRS reaction is far below its Km, the assay is presumably biased toward ATP-competitive inhibitors (27Baragaña B. Forte B. Choi R. Hewitt S.N. Bueren-Calabuig J.A. Pisco J.P. Peet C. Dranow D.M. Robinson D.A. Jansen C. Norcross N.R. Vinayak S. Anderson M. Brooks C.F. Cooper C.A. et al.Lysyl-tRNA synthetase as a drug target in malaria and cryptosporidiosis.Proc. Natl. Acad. Sci. U. S. A. 2019; 116: 7015-7020Crossref PubMed Scopus (35) Google Scholar). We optimized the assay in 384-well HTS format and determined the amount of PheRS per reaction (Fig. S4). Reaction endpoint was chosen to be at the first 2 h. A known bacterial PheRS inhibitor (11Beyer D. Kroll H.P. Endermann R. Schiffer G. Siegel S. Bauser M. Pohlmann J. Brands M. Ziegelbauer K. Haebich D. Eymann C. Brötz-Oesterhelt H. New class of bacterial phenylalanyl-tRNA synthetase inhibitors with high potency and broad-spectrum activity.Antimicrob. Agents Chemother. 2004; 48: 525-532Crossref PubMed Scopus (59) Google Scholar) was synthesized as reference compound, named GDI05-001 (NMR and MS data in Supporting Information). Its IC50 against Mtb PheRS was determined to be 1.72 ± 0.13 μM. A Selleck Bioactive Collection purchased before 2018 contains 2148 small molecules with validated biological and pharmacological activities. The library was screened in seven plates, which showed mean robust Z′ at 0.756 to 0.862. When the cutoff was set to 70% inhibition compared with no-enzyme control reaction, four initial hits were identified (Fig. S5), giving a hit rate ∼0.19% (Fig. 2A). A top hit PF-3845 is reportedly a potent covalent inhibitor of human enzyme FAAH with Ki calculated to be 0.23 ± 0.03 μM (28Ahn K. Johnson D.S. Mileni M. Beidler D. Long J.Z. McKinney M.K. Weerapana E. Sadagopan N. Liimatta M. Smith S.E. Lazerwith S. Stiff C. Kamtekar S. Bhattacharya K. Zhang Y. et al.Discovery and characterization of a highly selective FAAH inhibitor that reduces inflammatory pain.Chem. Biol. 2009; 16: 411-420Abstract Full Text Full Text PDF PubMed Scopus (348) Google Scholar), and many analogous compounds are in public database. PF-3845 and an analog PF-04457845 were repurchased (Fig. 2B). The IC50s of the two compounds were determined with the two assays and other similar assays (29Cestari I. Stuart K. A spectrophotometric assay for quantitative measurement of aminoacyl-tRNA synthetase activity.J. Biomol. Screen. 2013; 18: 490-497Crossref PubMed Scopus (50) Google Scholar) (Fig. S6). By the Kinase-Glo assay, the IC50 of PF-3845 was 2.65 ± 0.19 μM (hillslope −0.86 ± 0.05) and the IC50 of PF-04457845 was 9.76 ± 0.73 μM (hillslope −0.81 ± 0.05) (Fig. 2C). PF-3845, at concentrations up to 20 μM, did not inhibit the activities of PPase, PNPase, and luciferase that were used in the coupled reactions (Fig. S7). The inhibition was observed only when Mtb PheRS was present in the coupled reactions. Thus, we identified investigational drug PF-3845 as a new PheRS inhibitor. PF-3845 apparently has a scaffold different from previously known PheRS inhibitors. The mode of inhibition of PF-3845 was analyzed with enzyme kinetics in comparison with GDI05-001. PPi production assay was used to record the progress of PheRS reaction setups, which had excessive amounts of two substrates but varying concentration of a third substrate at various inhibitor concentrations. Each dataset was fit into Michaelis–Menten equation; the resulting Lineweaver–Burk plot of a set of experiments with regard to the third substrate was examined for a pattern of competitive, noncompetitive, uncompetitive, or mixed inhibition mode. Both GDI05-001 and PF-3845 are clearly Phe competitive inhibitors of PheRS as the presence of inhibitor only affected Km but not Vmax (Fig. 3A). In the Phe-competitive mode, Ki of GDI05-001 is 0.20 ± 0.01 μM and Ki of PF-3845 is 1.70 ± 0.11 μM, determined with the PPi production assay. It was reported that GDI05-001 was noncompetitive regarding ATP (11Beyer D. Kroll H.P. Endermann R. Schiffer G. Siegel S. Bauser M. Pohlmann J. Brands M. Ziegelbauer K. Haebich D. Eymann C. Brötz-Oesterhelt H. New class of bacterial phenylalanyl-tRNA synthetase inhibitors with high potency and broad-spectrum activity.Antimicrob. Agents Chemother. 2004; 48: 525-532Crossref PubMed Scopus (59) Google Scholar), and we had similar observation. In contrast, PF-3845 is more likely in a mixed inhibition mode with regard to ATP over a range of PF-3845 concentrations (Fig. 3B). PF-3845 is an uncompetitive inhibitor with regard to tRNAPhe (Fig. 3C). Crystallographic effort was made to first solve the structures of Mtb PheRS apo protein and its complex with GDI05-001 at 2.83 and 2.71 Å, respectively. Like other bacterial PheRS structures, Mtb PheRS is an (αβ)2 heterotetramer. The α subunit contains catalytic site, and β subunit contains editing site (Fig. 4A). Except for the N-terminal 74 residues of α subunit, all other residues were constructed. The N-terminal region of Thermophilus thermus PheRS was reported to form a coiled-coil domain, playing a critical role in interaction with tRNA (30Moor N. Kotik-Kogan O. Tworowski D. Sukhanova M. Safro M. The crystal structure of the ternary complex of phenylalanyl-tRNA synthetase with tRNAPhe and a phenylalanyl-adenylate analogue reveals a conformational switch of the CCA end.Biochemistry. 2006; 45: 10572-10583Crossref PubMed Scopus (36) Google Scholar). It was flexible and invisible in the solved Mtb PheRS when no tRNA was included. The binding mode of GDI05-001 was illustrated by cocrystallization (Fig. 4B). When looking into the catalytic site of the α subunit, we found that besides the amino acid (Phe) pocket and ATP pocket, there is an additional pocket between the two pockets (Fig. 4C). The key residues forming the additional pocket are mainly hydrophobic, including F143, F148, A154, H175, F254, F255, P256, and F257. GDI05-001 occupies the Phe pocket and the additional pocket, whereas the ATP pocket is empty (Fig. 4, C and D). The phenylsulfonamide group of GDI05-001 extends much deeper into the amino acid pocket than a Phe could. In this pocket, strong π–π stackings can be seen between the phenyl rings of GDI05-001 and F255/F257. Three hydrogen bonds are formed between the urea group of GDI05-001 and the side chains of E217 and S177. In addition, the sulfonamide oxygen forms interactions with the side chain of Q215 and the main chain of G282 and G307. The indole moiety of GDI05-001 is bended almost vertically to the phenylsulfonamide and extends into the additional pocket, which has strong hydrophobic interaction with the protein. Compared with the apo form, two major differences were observed. First, in the apo form structure, the side chain of F257 has two alternative conformations with one conformation that occupies the deeper amino acid pocket, and the side chain of Q180 also stretches toward the pocket. While in the liganded structure, F257 adapts only one rigid conformation, which occupies out of the pocket. Together with the rotation of Q180, they constitute a much deeper amino acid pocket of the complex than the apo form (Fig. S8). Second, the loop 148 to 160 undergoes a conformational change upon GDI05-001 binding and has much higher B factors than the apo structure, indicating it is more flexible when GDI05-001 binds (Fig. S9). On the other hand, preliminary analysis using a deep learning–based docking platform Orbital showed that PF-3845 might have two major putative binding conformations among the top-ranking poses (Fig. S10). In one conformation, PF-3845 is bended to interact with the additional pocket outside Phe pocket by π–π interactions. Another possibility is that PF-3845 could extend deep into the ATP pocket. Further work is needed to validate which binding mode is dominated. To illustrate the binding mode of PF-3845, we obtained the cocrystal structure of Mtb PheRS in complex with PF-3845 at a resolution of 2.3 Å. Initial Fo–Fc map of PF-3845 (along with GDI05-001) is presented in Figure S11. PF-3845 binds clearly in the amino acid pocket as well as the ATP pocket, whereas the additional pocket is empty (Fig. 4C). An extensive hydrophilic and hydrophobic interaction network in two pockets can be found between PF-3845 and Mtb PheRS. Like the phenylsulfonamide group of GDI05-001, the 4-trifluoromethyl-2-pyridinyl group and phenyl ring of PF-3845 inserts deeply into the amino acid pocket. Two strong π–π stackings also can be seen between the phenyl rings of PF-3845 and F255/F257 residues of the PheRS (Fig. 4E). A hydrogen bond is formed between the amide of the trifluoromethylpyridinyl group and the main chain oxygen of G218. Two more hydrogen bonds are formed between the oxygen linker and the side chain of E217 and the main chain of F306. Unlike GDI05-001, the other end of PF-3845, including the 3-aminopyridine and piperidine groups, completely inserts into the ATP pocket (Fig. 4, C and E). The carbonyl group of PF-3845 forms a hydrogen bond with the main chain amide of G309. The amide of 3-aminopyridine forms a water-mediated interaction with the side chain of E311 and the main chain of E311/R312. The 3-aminopyridine of PF-3845 also forms a strong π–π stacking with F213 and a π–cation interaction with R312. Surprisingly, a sulfate ion (SO42−) is found near PF-3845 to bridge the interaction between the amide linker of PF-3845 and R201/R312 of the PheRS. Compared with the apo form, the side chain of R312 in the complex rotates nearly 90° to get closer to the compound. The SO42− most likely stabilizes R312, which in turn forms the π–cation interaction with the 3-aminopyridine. When this is compared with two other structures, hFARS2 complexed with Phe–AMP and E. coli PheRS (EcPheRS) complexed with Phe and AMP, the SO42− occupies an approximate space of AMP phosphate group in either case (Fig. 5). In the hFARS2–PheAMP cocrystal structure (31Klipcan L. Levin I. Kessler N. Moor N. Finarov I. Safro M. The tRNA-induced conformational activation of human mitochondrial phenylalanyl-tRNA synthetase.Structure. 2008; 16: 1095-1104Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar), the PheAMP forms hydrogen bond with R143 that is corresponding to R201 in Mtb PheRS. In the EcPheRS–AMP cocrystal structure, the AMP forms hydrogen bond with R301 corresponding to R312 in Mtb PheRS. Overall, the SO42− here engages R201 and R312 in the ATP pocket and enables multiple interactions to facilitate PF-3845 binding into the ATP pocket of Mtb PheRS. Although SO42− has been found at the active site of several other aaRS structures, none of them contributes to the binding of a ligand as in the case of PF-3845 (32Itoh Y. Sekine S. Kuroishi C. Terada T. Shirouzu M. Kuramitsu S. Yokoyama S. Crystallographic and mutational studies of seryl-tRNA synthetase from the archaeon Pyrococcus horikoshii.RNA Biol. 2008; 5: 169-177Crossref PubMed Scopus (22) Google Scholar, 33Zhou M. Dong X. Shen N. Zhong C. Ding J. Crystal structures of Saccharomyces cerevisiae tryptophanyl-tRNA synthetase: new insights into the mechanism of tryptophan activation and implications for anti-fungal drug design.Nucleic Acids Res. 2010; 38: 3399-3413Crossref PubMed Scopus (12) Google Scholar). To confirm the role of the SO42−, we reperformed and analyzed the PF-3845 inhibition kinetics with 5 mM (NH4)2SO4 in the PheRS enzyme reaction. This time, PF-3845 clearly acted as an ATP competitive inhibitor (Fig. 3D), with the Ki calculated to be 0.73 ± 0.055 μM, which is lower than Ki 1.08 ± 0.245 μM with regard to ATP when SO42− was absent (Fig. 3B). The experiment demonstrated that sulfate ion increased the binding of PF-3845, making it a true bisubstrate compet