Accepted Article

Structural basis for targeting the ribosomal protein S1 of Mycobacterium tuberculosis by pyrazinamide1

Juanjuan Yang1, Yindi Liu1, Jing Bi2, Qixu Cai3, Xinli Liao1, Wenqian Li1, Chenyun Guo1, Qian Zhang3, Tianwei Lin3, Yufen Zhao1, Honghai Wang2, Jun Liu2,4*, Xuelian Zhang2,*, and Donghai Lin1,*

1

MOE Key Laboratory of Spectrochemical Analysis & Instrumentation, Key Laboratory for Chemical Biology of Fujian Province, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen, 361005, China 2 State Key Laboratory of Genetic Engineering, School of Life Sciences, Fudan University, Shanghai, 200433, China 3 State Key Laboratory of Cellular Stress Biology, School of Life Sciences, Xiamen University, Xiamen, 361102, China 4 Department of Molecular Genetics, University of Toronto, Toronto, M5S1A8, Canada

*

To whom correspondence should be addressed: Donghai Lin, [email protected], Tel.: 86-592-2186078; Fax: 86-592-2186078; Xuelian Zhang, [email protected], Tel.: 86-21-55665073; Fax: 86-21-55665073; or Jun Liu, [email protected], Tel: 1-416-946-5067; Fax: 1-416-978-6885.

Keywords: pyrazinamide, Mycobacterium tuberculosis, RpsA, ribosomal protein S1 Running title: Structural basis for targeting Mtb RpsA by pyrazinamide

Footnote: Atomic coordinates and structure factors have been deposited in the Protein Data Bank (http.www.pdb.org) with accession codes 4NNK (apo-MtRpsACTD), 4NNG (MtRpsA△A438CTD), 4NNH (MsRpsACTD) and 4NNI (MtRpsACTD-POA complex).

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/mmi.12892

1 This article is protected by copyright. All rights reserved.

Accepted Article

SUMMARY Pyrazinamide (PZA) is a first-line drug for tuberculosis (TB) treatment and is responsible for shortening the duration of TB therapy. The mode of action of PZA remains elusive. RpsA, the ribosomal protein S1 of Mycobacterium tuberculosis (Mtb), was recently identified as a target of PZA based on its binding activity to pyrazinoic acid (POA), the active form of PZA. POA binding to RpsA led to the inhibition of trans-translation. However, the nature of the RpsA-POA interaction remains unknown. Key questions include why POA exhibits an exquisite specificity to RpsA of Mtb and how RpsA mutations confer PZA

resistance. Here, we report the crystal structures of the C-terminal domain of RpsA of Mtb and its complex with POA, as well as the corresponding domains of two RpsA variants that are associated with PZA resistance. Structural analysis reveals that POA binds to RpsA through hydrogen-bonds and hydrophobic interactions, mediated mainly by residues (Lys303, Phe307, Phe310 and Arg357) that are essential for tmRNA binding. Conformational changes induced by mutation or sequence variation at the C-terminus of RpsA abolish the POA binding activity. Our findings provide insights into the mode of action of PZA and molecular basis of PZA resistance associated with RpsA mutations.

2 This article is protected by copyright. All rights reserved.

Accepted Article

INTRODUCTION Tuberculosis (TB), caused by Mycobacterium tuberculosis (Mtb), remains one of the leading causes of

death worldwide. The World Health Organization estimated that there were 1.4 million deaths and 8.7 million new cases of TB in 2011. Current TB treatment comprises four first-line drugs, i.e., isoniazid, rifampin, ethambutol, and pyrazinamide (PZA). The inclusion of PZA in the regimen was pivotal in shortening the duration of TB therapy from 9-12 months to 6 months (Mitchison, 1985), although it is still longer than that for other bacterial infections. Shortening TB treatment is a key objective in order to reduce patient noncompliance and to combat drug resistance. The ability of PZA to shorten TB therapy is widely attributed to its ‘sterilizing’ activity primarily by

reducing relapse rate in patients on 6-month multidrug treatment (Mitchison, 1985). PZA is a pro-drug that requires amide hydrolysis to pyrazinoic acid (POA) by mycobacterial pyrazinamidase encoded by pncA (Scorpio & Zhang, 1996), and PZA resistance often correlates with the loss of pyrazinamidase activity through pncA mutations (Scorpio & Zhang, 1996, Sreevatsan et al., 1997). Efforts to develop a functional improvement of PZA, however, have been thwarted by the lack of knowledge on its mode of action. Although many mycobacterial species such as Mycobacterium smegmatis (M. smegmatis) produce functional pyrazinamidase (Boshoff & Mizrahi, 1998, Guo et al., 2000), they are naturally resistant to PZA. The lack of activity of PZA against M. smegmatis has been attributed to the presence of an active efflux of POA in this organism (Zhang et al., 1999). As such, the intracellular accumulation of POA

appears to be an essential feature of PZA activity. Supporting this, in vitro studies found that PZA is only effective against Mtb in media with acidic pH (e.g., pH 5.5) (Mc & Tompsett, 1954), primarily because low extracellular pH enhances the intracellular accumulation of POA (Zhang et al., 1999). 3

This article is protected by copyright. All rights reserved.

Accepted Article

Despite the importance of PZA in TB therapy, the molecular targets of PZA remain elusive until

recently. A previous study found that a PZA analog, 5-Cl-PZA, inhibits fatty acid synthetase I (FAS-I) of Mtb (Zimhony et al., 2000). However, this activity was found to be unique for 5-Cl-PZA and not shared by PZA (Boshoff et al., 2002). POA, the active form of PZA, was found to interfere with the proton

motive force and ATP synthesis (Lu et al., 2011, Zhang et al., 2003), but the underlying mechanisms remain unknown. Recently, Shi et al. identified RpsA of Mtb as a target of POA (Shi et al., 2011). RpsA, also known as ribosomal protein S1, is an essential protein required for translation initiation of mRNAs when their Shine-Dalgarno sequence is degenerated (Sorensen et al., 1998). In addition, RpsA of Mtb is

involved in trans-translation, which is an effective system, mediated by tmRNA-SmpB, to release stalled ribosomes from mRNA in the presence of rare codons (Keiler, 2008). Shi et al. found that POA binds to

RpsA of Mtb and disrupts the formation of RpsA-tmRNA complex (Shi et al., 2011). In contrast, POA did

not bind to an RpsA mutant protein of Mtb associated with PZA resistance, an alanine deletion at the C-terminal (RpsA△A438), nor did it bind to RpsA of M. smegmatis. Based on biochemical evidence, Shi et al. proposed that POA directly binds to the C-terminal region of RpsA and disrupts the site of tmRNA interaction (Shi et al., 2011). However, molecular details of the RpsA-POA interaction remain unknown. Given the indispensable role of PZA in TB therapy and that RpsA represents the first tangible target of

PZA, it is important to elucidate the molecular nature of the RpsA-POA interaction. Key questions include why POA exhibits an exquisite binding specificity to RpsA of Mtb and how such interaction

prevents RpsA from binding to tmRNA. To address these questions, we determined the crystal structures of the C-terminal domain of RpsA of Mtb, and its complex with POA. We also solved the crystal structures of the equivalent domains of two RpsA variants that do not bind POA. Comparison of these 4 This article is protected by copyright. All rights reserved.

Accepted Article

structures reveals that RpsA of Mtb interacts with POA through hydrophobic and hydrogen-bond interactions, which are mostly mediated by the conserved residues in the RNA-binding domain. The structural features that distinguish RpsA of Mtb from the RpsA variants incapable of binding POA are also identified. Our findings provide mechanistic insights into the mode of action of POA and a framework for improving the efficacy of PZA.

RESULTS

Crystal structure of the C-terminal domain of RpsA of Mtb In E. coli and other Gram-negative bacteria, RpsA consists of six S1 domains (Salah et al., 2009),

which belong to the oligonucleotide/oligosaccharide binding (OB) fold family and are found in many proteins involved in RNA metabolism (Murzin, 1993). The first two S1 domains at the N-terminal are involved in binding to the ribosome via protein-protein interactions (Subramanian, 1983), whereas the remaining four S1 domains are involved in RNA binding (Aliprandi et al., 2008, Salah et al., 2009). In

Mtb and other actinomycetes, RpsA consists of only four S1 domains (Salah et al., 2009), but has an

additional ~100 amino acids at the C-terminal that are specific for actinomycetes (Fig. 1A) (Salah et al., 2009). Sequence analysis suggests that the first two S1 domains of Mtb RpsA, by analogy to RpsA of E. coli, are likely involved in ribosome binding, while the third and fourth S1 domains might be responsible for RNA binding (Salah et al., 2009). In addition, the RpsA variant of Mtb incapable of binding to POA contains a mutation at the C-terminal (RpsA△A438) (Shi et al., 2011). As such, it was suggested that the C-terminal of RpsA might be responsible for interacting with POA (Shi et al., 2011). To test this

hypothesis, we expressed and purified the full-length RpsA of Mtb, the C-terminal domain (residues 5

This article is protected by copyright. All rights reserved.

Accepted Article

285-481, MtRpsACTD) containing the fourth S1 domain (residues 292-363), and the C-terminus (residues 364-481) (Fig. 1A), and analyzed their POA binding activities by isothermal titration calorimetry (ITC). Previously, Shi et al. reported that the full-length RpsA of Mtb binds POA with a Kd ≈0.13 µM based on ITC analysis while they did not report the binding stoichiometry (Shi et al., 2011). We found that POA

binds the full-length RpsA of Mtb with a Kd  2.810.67 µM and a binding stoichiometry of 2.03 (Table 1, Supporting Figure S1A). The discrepancy in binding affinity may be due to differences in experimental conditions between these two studies (e.g., different instrumentations and assay conditions). Shi et al. did not analyze the POA-binding activity of truncated RpsA proteins (Shi et al., 2011). We

found that MtRpsACTD retains the POA-binding activity of the full-length RpsA protein, with a Kd  3.13 0.06 µM and a binding stoichiometry of 1.94 (roughly two POAs for one MtRpsACTD protein) (Table 1, Supporting Figure S1B). Therefore, we focused our structural studies on the C-terminal domain of RpsA (MtRpsACTD). The crystal structure of MtRpsACTD was determined at 1.6 Å (Supporting Table S1), which includes a

typical S1 domain (Bycroft et al., 1997), a short helix (α2, ≈10 Å) linking β5 and α4 and a long C-terminal helix (α4, ≈40 Å, Glu408-Glu433) (Fig. 1B). Forty-eight residues (residues 434 to 481) at the C-terminus were invisible, presumably due to the structural flexibility. The S1 domain of MtRpsACTD comprises one canonical OB fold characterized by a five-stranded antiparallel β-barrel (β1-β5) and an additional 310-helix between β3 and β4 (Fig. 1B) (Salah et al., 2009). The presence of 310-helix suggests that the S1 domain does not bind oligosaccharides, but rather binds to RNA (Murzin, 1993). Several conserved residues including Phe307, Phe310, His322, Asp352 and Arg357 (corresponding to Phe19, Phe22, His34, Asp64, and Arg68 in a typical S1 domain) spatially juxtaposed at one side of the β-barrel, 6

This article is protected by copyright. All rights reserved.

Accepted Article

forming an RNA-binding site (Bycroft et al., 1997) (Fig. 1C and 1E). The S1 domain of MtRpsACTD demonstrates certain similarities with other nucleic acid binding domains, and can be superimposed with E. coli ribosomal protein S1 domain 4 (PDB accession code 2KHI) or PNPase S1 domain (PDB accession

code 1SRO) with r.m.s. deviations (RMSDs) of 1.3 Å and 1.2Å for Cα atoms, respectively.

Co-crystal structure of MtRpsACTD with POA

The crystal structure of the MtRpsACTD-POA complex was solved in space group P212121 and at a

resolution of 2.6 Å (Supporting Table S1). The structure of POA-bound MtRpsACTD includes 4 additional

residues (residues 434-437) in α4 helix that were invisible in the structure of unbound MtRpsACTD (apoMtRpsACTD), but the rest of the C-terminus (residues 438-481) remained invisible. The crystal asymmetric unit contains two MtRpsACTD-POA complexes with superimposable conformations (RMSD

of 0.1 Å), and interactions between MtRpsACTD and POA in each complex are essentially the same. For

simplicity, we use one MtRpsACTD-POA pair to describe their interactions. The co-crystal structure reveals that two POA molecules bind to one MtRpsACTD (Fig. 1D), which is

consistent with the binding stoichiometry determined in solution by ITC analysis (Kd = 3.13 μM, N = 2) (Table 1). POAs bind to MtRpsACTD at the surface of S1 domain predicted to interact with RNA (compare

Fig. 1C and 1D). Three of the five conserved residues at the RNA binding site (Phe307, Phe310, and Arg357) (Bycroft et al., 1997) are involved in MtRpsACTD-POA interactions. Two POA molecules (POAI and POAII) were modeled into the extra electron density around Phe307, Phe310 and Arg357 unambiguously based on the Fo-Fc map. The refined B factors of POAI and POAII are 81 and 68 Å2, respectively, which are similar to those of neighboring atoms. The POA-binding cavity is complementary 7 This article is protected by copyright. All rights reserved.

Accepted Article

in both shape and charge with POA. Thus the two POA molecules are docked into a groove formed by the strands β2, β3 and β5 of MtRpsACTD (Fig. 1D and 2A). POA interacts with the S1 domain of MtRpsACTD via both hydrophobic and hydrogen-bond interactions, which are mostly mediated by Phe307, Phe310 and Arg357 at the RNA-binding site (Bycroft et al., 1997). As shown in the high-resolution electron density map, the pyrazine ring of POAI forms a π-π interaction with Phe307 located in the loop between β1 and β2, while its O9 atom is hydrogen bonded with the guanidinium group of Arg357 (Fig. 2B and Supporting Figure S2A). Similarly, the pyrazine ring of POAII forms another π-π interaction with Phe310 located within β2, and the O8 atom forms two hydrogen bonds with the guanidinium group of Arg357. In addition, the N5 atom of POAII is hydrogen bonded to the side chain amino group of Lys303, and the C3 and C4 atoms interact with the side chain methylene groups of Glu318 via hydrophobic contact (Fig. 2B and Supporting Figure S2A). The overall structure of the apo-MtRpsACTD is similar to that of the POA-bound MtRpsACTD (RMSD

of 0.6 Å), suggesting that MtRpsACTD may use a preformed surface to interact with POA. However, POA binding induces significant conformational changes at the local structures. There are substantial movements of the β2 (Gly308-Glu314), β3 (Glu317-His322), and β5 (Arg356-Ser361) strands of the S1 domain in the structure of MtRpsACTD-POA compared to that of apo-MtRpsACTD (Fig. 2C). In addition, the side chain of Arg357 in the apo-MtRpsACTD structure is highly flexible and can only be partially

modeled based on the electron density map (Supporting Figure S2B). However, in the MtRpsACTD-POA

structure this side chain becomes rigidly structured and the electron density is unambiguous (Supporting Figure S2C). To confirm the MtRpsACTD-POA interactions revealed by crystal structures, we mutated the 8

This article is protected by copyright. All rights reserved.

Accepted Article

POA-interacting residues (Lys303, Phe307, Phe310 and Arg357), expressed these MtRpsACTD mutants, and examined their POA-binding activities by ITC analysis (Table 1, Supporting Figure S1D-I). Individual mutations of these residues significantly impaired the ability of MtRpsACTD to bind POA

(Table 1). Notably, the F310G mutation reduced the POA binding affinity by >3 fold (Kd = 9.341.04 μM, Table 1). This mutation may completely disrupt one of the two binding sites of POA (i.e., the binding site for POAII, Fig. 2B), as such only one site is accessible to POA in the F310G mutant (N=1.07, Table 1). Furthermore, the MtRpsACTD mutants F307A/F310L and F307A/F310L/R357A exhibited unmeasurable POA-binding activities (Supporting Figure S1H and S1I), indicating that double or triple mutations abolished the POA binding ability. These results support the involvement of Lys303, Phe307, Phe310 and Arg357 in MtRpsACTD-POA interactions.

Structural comparison of MtRpsACTD with MtRpsA△A438CTD and MsRpsACTD To gain insights into the specificity of PZA towards Mtb and the molecular basis of PZA resistance

associated with RpsA mutations, we further determined the crystal structures of the corresponding domains of the Mtb RpsA mutant, MtRpsA△A438CTD (residues 285-480) and that of the M. smegmatis, MsRpsACTD (residues 285-479) at 2.0 and 2.3 Å, respectively (Supporting Table S2), and compared them with the structure of MtRpsACTD. Sequence alignment of these RpsACTD proteins is shown in Figure 3A. The MtRpsA△A438 mutation was identified in a PZA resistant Mtb strain (Shi et al., 2011) while M. smegmatis is naturally resistant to PZA. These RpsA variants show little or no POA binding activity (Supporting Figure S1J and S1O), which is consistent with a previous study (Shi et al., 2011). Like MtRpsACTD, the structure of MtRpsA△A438CTD is only visible for residues 285-433. 9

This article is protected by copyright. All rights reserved.

Accepted Article

Superimposition of these two structures reveals that the overall structure of MtRpsA △A438CTD including the S1 domain and α4 helix is similar to that of MtRpsACTD (RMSD of 0.37Å, Figure 3B). However, side chains of the POA-binding residues in MtRpsA△A438CTD adopt different conformations (Fig. 3C). In the crystal structure of MtRpsA△A438CTD, the NH1, NH2 atoms of the guanidinium group of Arg357 shift by 1.6 Å and 3.2 Å, respectively, away from the corresponding atoms of Arg357 in MtRpsA CTD (Fig. 3C), significantly weakening its hydrogen bonding with POA molecules. The amino group (the Nζ atom) of Lys303 shifts by 0.9 Å away from that of Lys303 in MtRpsACTD, potentially weakening its hydrogen bonding with the N5 atom of POAII (Fig. 3C). Moreover, the Cε1, Cζ, Cε2 atoms on the phenyl ring of Phe307 all shift by 0.9 Å, which would not favor the π-π interaction between the phenyl ring of Phe307 and POAI (Fig. 3C). The structure of MsRpsACTD includes additional 11 amino acids (residues 434-444) that are invisible

in MtRpsACTD. As such the α4 helix of MsRpsACTD is longer than that of MtRpsACTD by about 15 Å (Fig. 3B). The overall structure of MsRpsACTD is similar to that of MtRpsACTD (RMSD of 0.67 Å, Fig. 3B). However, conformational differences of residues involved in POA binding are detected (Fig. 3D). In particular, the amino group (the Nζ atom) of Lys303 shifts by 2.1 Å away from that of Lys303 in MtRpsACTD, potentially abolishing its hydrogen bonding with the N5 atom of POAII. The NH1, NH2 atoms of the guanidinium group of Arg357 shift by 1.3 Å and 0.9 Å, respectively, away from the corresponding atoms of Arg357 in MtRpsACTD (Fig. 3D), weakening its hydrogen bonding with POA. In addition, the phenyl ring (Cε1, Cζ, Cε2 atoms) of Phe307 shift by 1.2, 1.1, 0.9 Å and the phenyl ring (Cζ, Cε2 atoms) of Phe310 shift by 0.5, 0.7 Å, respectively, away from the corresponding atoms in MtRpsACTD, thus disrupting the π-π interaction between the phenyl ring of Phe307 or Phe310 and POA (Fig. 3D). 10

This article is protected by copyright. All rights reserved.

Accepted Article

The observed conformational rearrangements in the side chains of the POA-binding residues in

MtRpsA△A438CTD and MsRpsACTD also result in distinct changes in the electrostatic surface potentials (Fig. 3E). Notably, the surface around the POA-binding site of MtRpsACTD is continuously positively charged, contributed by the crucial residues Lys303, Phe310, Leu320, Asp352 and Arg357. However, both the positive potential and the continuity are less striking in MtRpsA△A438CTD and MsRpsACTD (Fig. 3E). Given that POA with a pKa of 2.9 exists as a carboxylate anion at near neutral pH (Zhang et al., 1999), the continuous positively charged surface around the POA-binding site in MtRpsACTD may facilitate POA binding to MtRpsACTD more readily than to MtRpsA△A438CTD or MsRpsACTD. These differences together provide structural basis for the markedly reduced POA-binding affinities of these RpsA variants as opposed to that of MtRpsACTD.

Interaction of the flexible C-terminus with the fourth S1 domain Our analyses above indicate that the S1 domain (residues 292-363) of MtRpsACTD is directly involved

in POA binding. However, the sequence of S1 domain of MtRpsA△A438CTD or MsRpsACTD is either identical or nearly identical to that of MtRpsACTD (Fig. 3A), suggesting that conformational differences of the POA-binding site in these proteins are likely caused by interactions between the S1 domain and the C-terminus that are invisible in the crystal structures. To test this hypothesis, we expressed and purified the S1 domain (residues 285-368) of MtRpsACTD, and performed NMR analysis of the isolated S1 domain and MtRpsACTD. Superimposition of the 2D 1H-15N HSQC spectra of the isolated S1 domain and

MtRpsACTD reveals that a significant fraction of NH signals of the S1 domain does not overlap with that of MtRpsACTD (Fig. 4), suggesting that substantial interactions occur between the flexible C-terminus and 11

This article is protected by copyright. All rights reserved.

Accepted Article

the S1 domain in MtRpsACTD. Intriguingly, the isolated S1 domain binds POA with the same affinity and stoichiometry as MtRpsACTD (Kd  3.21 0.08 µM, N=1.96; Table 1, Supporting Figure S1C). Given that

sequences of the S1 domain of MtRpsACTD, MtRpsA△A438CTD and MsRpsACTD are essentially the same (Fig. 3A), this result suggests that the isolated S1 domains of MtRpsA△A438CTD and MsRpsACTD are fully capable of binding POA. It follows then the interaction of the flexible C-terminus with the S1 domain in MtRpsA△A438CTD or MsRpsACTD must have altered the conformation of the POA binding site in the S1 domain and disrupted its binding ability, whereas in MtRpsACTD the conformation of the POA binding site

is not affected by such interaction.

Analysis of tmRNA binding activities of RpsA proteins We found that the full-length RpsA protein of Mtb bound tmRNA at concentration-dependent manner,

and that addition of excessive POA disrupted the RpsA-tmRNA complex (Fig. 5A, lanes 2-4), which is consistent with the previous study (Shi et al., 2011). The truncated protein MtRpsACTD bound tmRNA equally well and this interaction was also inhibited by POA (Fig. 5A, lanes 5-7), suggesting that MtRpsACTD retains the tmRNA binding activity of the full-length protein. The full-length RpsA mutant protein of Mtb, MtRpsA△A438, or the full-length RpsA protein of M. smegmatis (MsRpsA) also bound

tmRNA efficiently. However, their tmRNA activity was not affected by the presence of POA (Fig. 5B, lanes 2-4, lanes 5-7, respectively). Similarly, the C-terminal domain of Mtb RpsA mutant, MtRpsA△A438CTD, or that of the M. smegmatis RpsA, MsRpsACTD, was capable of binding tmRNA, but their interactions were not affected by POA (Fig. 5B, lanes 8-10, lanes 11-13, respectively). These results suggest that residues of the S1 domain that interact with POA (Lys303, Phe307, Phe310 and Arg357) are 12

This article is protected by copyright. All rights reserved.

Accepted Article

likely involved in tmRNA binding. Supporting this, individual mutations of these residues reduced their ability to bind tmRNA (Fig. 5C, lanes 3-6), and the effect is increasingly pronounced in double (F307A/F310L, Fig. 5C, lane 7) and triple mutants (F307A/F310L/R357A, Fig. 5C, lane 8). Not all residues in the S1 domain predicted to be involved in tmRNA binding (e.g. His322, Asp352, Fig. 1E) interact with POA, which may explain the residual tmRNA binding activity of the triple mutants (Fig. 5C, lane 8). Together, our results demonstrate that POA binds to residues of RpsA of Mtb that are also

involved in tmRNA binding, providing an explanation for the POA mediated inhibition of trans-translation (Wower et al., 2000, Felden & Gillet, 2011).

DISCUSSION In this study, we performed a detailed analysis of RpsA-POA interactions. Our findings provide new

insights into the mode of action of POA. Our results suggest that the fourth S1 domain of RpsA of Mtb is responsible for POA binding and that POA inhibits trans-translation of Mtb by directly competing for the tmRNA-binding site in RpsA. This conclusion is supported by several lines of evidence. Firstly, the isolated fourth S1 domain and the C-terminal domain protein MtRpsACTD exhibit nearly identical binding affinity and stoichiometry to POA as the full-length protein (Table 1, Supporting Figure S1). Secondly, the crystal structure of the MtRpsACTD-POA complex reveals that residues Lys303, Phe307, Phe310 and

Arg357 of the S1 domain directly interact with POA (Fig. 2A and 2B), and that mutations of these residues abolish the MtRpsACTD-POA interaction (Table 1, Supporting Figure S1). Thirdly, mutations of residues Lys303, Phe307, Phe310 and Arg357 greatly diminish the ability of MtRpsACTD to bind tmRNA (Fig. 5C), which is consistent with the prediction that they are among the conserved residues involved in 13

This article is protected by copyright. All rights reserved.

Accepted Article

RNA binding. In addition, RNAs mainly interact with the β2, β3 and β5 strands of the S1 domains (Salah

et al., 2009). POA binding to MtRpsACTD induces significant conformational changes in the β2, β3 and β5 stands of the S1 domain (Fig. 2C), which may further disrupt its tmRNA binding activity. Consistently, the RpsA-tmRNA interaction can be efficiently inhibited by POA (Fig. 5A). Together, these data provide strong support for the notion that POA inhibits trans-translation of Mtb by blocking and/or altering the tmRNA binding site in RpsA, thereby preventing the formation of RpsA-tmRNA complex. Since the POA interacting residues Lys303, Phe307, Phe310 and Arg357 are highly conserved in S1

domains found in RpsA proteins of various bacterial species (Fig. 1E), it raises the question as to why those bacterial species are insensitive to POA (or PZA). There are two possible explanations. Firstly, while these residues are highly conserved for their RNA binding activity in S1 domains of various organisms, there are considerable levels of sequence variation between S1 domains of other bacterial species and the fourth S1 domain of Mtb (Fig. 1E), which may explain the lack of POA binding activity of S1 domains of those organisms. Similarly, our results suggest that the first three S1 domains of Mtb RpsA do not bind POA, which may be explained by their overall low sequence homology to the fourth S1 domain (Supporting Figure S3). Secondly, although RpsA of E. coli and Thermus thermophile was shown to bind tmRNA with high affinity (Wower et al., 2000, McGinness & Sauer, 2004, Takada et al., 2007), the physiological relevance of the RpsA-tmRNA interactions remains unknown. There is convincing data which argues against the involvement of RpsA in trans-translation of E. coli and Thermus thermophile (McGinness & Sauer, 2004, Qi et al., 2007, Takada et al., 2007). In contrast, Shi et al. found that POA prevents RpsA of Mtb from binding to tmRNA and causes inhibition of trans-translation, indicating that RpsA is required for trans-translation in Mtb (Shi et al., 2011). The possibility that RpsA of other 14

This article is protected by copyright. All rights reserved.

Accepted Article

organisms such as E. coli and Thermus thermophile are not involved in trans-translation may account for the lack of activity of PZA against these organisms. The apparent difference of RpsA function in Mtb and E. coli or Thermus thermophile may be explained by their sequence variation. As mentioned earlier, RpsA of Mtb and other actinomycetes contain, in addition to the four S1 domains, a C-terminus of ~100 amino acids that does not exist in RpsA from E. coli or other bacteria (Salah et al., 2009). In our crystal structures, the C-terminus consists of a long α-helix (α4), which may be involved in interacting with other essential components of the trans-translation system (e.g., SmpB). The lack of this C-terminus in RpsA of E. coli or other bacteria may prevent its involvement in trans-translation. In our crystal structures, approximately 40-50 residues at the C-terminus are invisible in all RpsACTD

proteins (residues 434-481 in apo-MtRpsACTD, 438-481 in MtRpsACTD-POA, 434-480 in MtRpsA△A438CTD and 445-481 in MsRpsACTD) presumably due to the structural flexibility. Sequences of RpsA proteins

from various mycobacterial species are nearly identical from residues 1-400, including that of the fourth S1 domain (Fig. 3A). However, the sequences are less conserved at residues 440-460 (Fig. 3A), which are part of the C-terminus that are invisible in crystals. While future study is required to resolve the structure of the C-terminus, our initial NMR experiments suggest that there are substantial interactions between the flexible C-terminus and the fourth S1 domain (Fig. 4). This interaction may influence the conformation of the POA-binding site at the fourth S1 domain, resulting in differential binding affinity. Consistent with this notion, we found that the isolated fourth S1 domain (residues 292-363), which is highly conserved in various mycobacterial species, is fully capable of binding POA. Sequence variations at the C-terminus of RpsA of various mycobacterial species may alter its interaction with the fourth S1 domain, leading to conformational differences in the POA-binding site and consequently reduced affinity for POA, as 15

This article is protected by copyright. All rights reserved.

Accepted Article

demonstrated by the structure of MsRpsACTD (Fig. 3D). This may explain why the majority of mycobacterial organisms, other than the Mtb complex, are naturally resistant to PZA. Comparison of the structures of MtRpsACTD and two RpsA variants (MtRpsA△A438CTD and MsRpsACTD)

that do not bind POA reveals that the POA-RpsA interaction is highly specific and any subtle changes in conformation can lead to the loss of POA binding activity, as demonstrated by the deletion of alanine at 438, even though this residue is not directly involved in POA binding. Ala438, although invisible in the crystal structure of MtRpsACTD, is located at the α4 helix of MsRpsACTD and is approximately 20 Å away from the POA binding sites (Supporting Figure S4). Deletion of Ala438 does not cause gross structural changes but induces conformational changes of the POA-binding sites (Fig. 3C), which is likely mediated by altering the interaction of the flexible C-terminus with the fourth S1 domain. Recently, mutations in rpsA of clinical strains of Mtb that are resistant to PZA have been described in the literature (Alexander et

al., 2012, Tan et al., 2014). Notably, in one study, three of the seven PZA-resistant Mtb strains that contain a wild type pncA gene harbored single mutations at the C-terminus (R474L, R474W and E433D) (Fig. 3A) (Tan et al., 2014). Similar to the Ala438 deletion mutant, mutations of these residues may cause conformational changes in the POA-binding site, resulting in reduced affinity for POA. Mutations at the POA-binding sites (Lys303, Phe307, Phe310 and Arg357) of RpsA in Mtb clinical strains have not been reported in the literature. This may be due to the fact that these residues are also required for tmRNA binding (Fig. 5C) and trans-translation, which is essential for Mtb viability. In contrast, mutations at the C-terminus of RpsA are less likely to affect the tmRNA binding activity, as demonstrated by MtRpsA △A438 (Figure 5C). As such, it is not surprising that mutations of rpsA associated with PZA resistance are mostly found in the C-terminus. On the other hand, the frequency of rpsA mutations appears to be quite low in 16

This article is protected by copyright. All rights reserved.

Accepted Article

clinical strains (Alexander et al., 2012, Simons et al., 2013), suggesting that other unknown PZA

resistance mechanisms or additional targets of PZA may exist. Our finding also lays a structural ground for designing compounds with new scaffolds that target

PZA-resistant Mtb strains. As mentioned above, majorities of PZA-resistant Mtb strains contain mutations in pncA, which cannot convert PZA to POA (Scorpio & Zhang, 1996, Sreevatsan et al., 1997). PZA does not bind RpsA (Supporting Figure S5), possibly because the hydrogen bond formed between O9 atom of POA with the guanidinium group of Arg357 of RpsA is critical for their interaction, which cannot be achieved with the NH2 group at the same position in PZA. POA by itself, however, is inactive against Mtb due to its poor ability to penetrate the lipid-rich mycobacterial cell wall and accumulate inside Mtb cells (Zhang et al., 1999). Based on our results, the C3 and C4 position of the POA molecule appear to be less critical in the interaction with RpsA (Fig. 2B, Supporting Figure S2A), which may serve as sites for modification. For example, substitution with alkyl chains at these positions may facilitate POA penetration across the mycobacterial cell wall, which would circumvent the resistance caused by pncA mutation.

EXPERIMENTAL PROCEDURES Protein expression and purification - The gene sequences encoding the full-length RpsA of Mtb

(MtRpsA), the isolated fourth S1 domain (residues 285-368) of Mtb RpsA, the C-terminal domains of Mtb RpsA (MtRpsACTD, residues 285-481) and M. smegmatis RpsA (MsRpsACTD, residues 285-479) were amplified from Mtb and M. smegmatis genomic DNA, respectively. The amplified DNA was inserted into the pET-28a vector (Novagen) between NdeI and Hind III sites, resulting in an N-terminal hexahistidine 17

This article is protected by copyright. All rights reserved.

Accepted Article

tag for purification. The MtRpsACTD mutant (MtRpsA△A438CTD, residues 285-480, missing an alanine at residue

438)

and

MtRpsACTD

mutants

(K303A,

F307A,

F310G,

R357A,

F307A/F310L,

F307A/F310L/R357A) were generated by site-directed mutagenesis using the pET-28a-MtRpsACTD vector as the template and confirmed by sequencing. All the recombinant proteins were overexpressed in Escherichia

coli

strain

BL21

(DE3)

cells

and

induced

with

0.5

mM

isopropyl-β-D-1-thiogalactopyranoside (IPTG, Amresco) at an OD600 of 0.6-0.8 in LB media. The cells were further grown for 8 h at 25 °C, harvested by centrifugation, and the pellet was stored at -80 °C. Bacterial pellets were resuspended in 20 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10 mM imidazole and lysed by sonication on ice. The proteins were purified by Ni 2+ affinity chromatography (Qiagen)

following

the

protocol

recommended

by

the

manufacturer.

Selenomethionine

(SeMet,

Sigma-Aldrich)-labeled MtRpsACTD (SeMet-MtRpsACTD) was produced by inhibiting endogenous methionine biosynthesis in M9 minimal medium supplemented with specific amino acids as well as SeMet and then purified in the same manner as the native protein. All protein purifications were performed at 4 °C. The protein purity was verified by polyacrylamide gel electrophoresis, and the protein concentration was measured by the BCA assay (Sigma). Dynamic light scattering - The states of the RpsACTD proteins were determined by dynamic light

scattering (DLS). The purified proteins were concentrated to approximately 10 mg ml −1 and the concentration of POA was 10 mM. Experiments were carried out in a Dynapro dynamic light scattering instrument (Marvern Zetaszier, Nano-ZS90). The protein and complex solutions were incubated at least 30 min at 25 °C and then centrifuged at 13,000 g for 10 min before measurement. Data was acquired at 25 °C. Regularization histogram analyses of DLS results were carried out using the software zetasizer 18

This article is protected by copyright. All rights reserved.

Accepted Article

version 6.20. These measured average size distributions are consistent with the monodisperse particles of average molecular weight of ~21 kDa, which is within the expected size of monomeric RpsACTD

(Supporting Fig. S6). Circular dichroism spectroscopy - The secondary structures of the RpsACTD proteins were determined

by circular dichroism (CD). The purified proteins were concentrated to approximately 0.5 mg ml−1 and the concentration of POA was 0.5 mM. CD spectra were recorded on a JASCO-810 spectropolarimeter (Jasco, Tokyo, Japan) with a quartz cell of 0.1 cm path length. Each protein sample was incubated at 25 °C for 30 min. Spectra were recorded from 190 to 260 nm by accumulating three consecutive scans with a bandwidth of 1.0 nm and a response time constant of 1.0 s. The spectra were processed by first subtracting a blank spectrum followed by baseline correction, and then normalized for the mean residue weight. The CD spectra demonstrate that main secondary structures of RpsACTD proteins are random coils, β-sheets and α-helices (Supporting Fig. S7). Crystallization and data collection - For crystallization, the purified protein was concentrated to

approximately 10 mg ml−1 in 10 mM Tris-HCl, pH 7.5, 150 mM NaCl. The crystallization of the complex of POA (Sigma-Aldrich) and MtRpsACTD was performed by mixing 10 mg ml−1 protein with 10 mM POA

at 22 °C for about 12 h. The crystals of MtRpsACTD, MtRpsA△A438CTD, MsRpsACTD, SeMet-labeled MtRpsACTD and the MtRpsACTD-POA complex were grown by the hanging-drop vapor diffusion method at 22 °C, using equal volumes of proteins and well solution. The crystals of MtRpsA CTD, SeMet-labeled MtRpsACTD and the MtRpsACTD-POA complex were grown from the same well solution (0.1 M Tris-HCl, pH 8.5, 0.3 M magnesium formate). The crystals of MtRpsA△A438CTD and MsRpsACTD were obtained in the different crystallization solutions (0.1 M sodium chloride, 0.1 M HEPES pH 7.5, 1.6 M ammonium 19

This article is protected by copyright. All rights reserved.

Accepted Article

sulfate and 0.2 M magnesium acetate tetrahydrate, 20% w/v PEG 3350, respectively). These crystals grew to their maximal sizes within 7 days at 22 °C. All the X-ray diffraction data sets were collected on beamline 17U at the Shanghai Synchrotron Radiation Facility using a charge-coupled device (CCD) detector.

Structure determination and refinement - The diffraction data sets were processed and scaled by

HKL2000 (Otwinowski & Minor, 1997). Initial phases were solved with the PHENIX package (Adams et al., 2002) by the SAD phasing method using SeMet-derived data. The model was manually built with COOT (Emsley & Cowtan, 2004), and structure refinement was accomplished with the CCP4 suite (Collaborative Computational Project, 1994). The structures of two RpsA variants associated with PZA-resistance and the MtRpsACTD-POA complex were solved by molecular replacement with the CCP4 suite, using the MtRpsACTD structure as the searching model. Ramachandran plots of the final four models display that all residues are within allowed regions, with above 96% in favored regions. Crystal diffraction data and refinement statistics are shown in Supporting Table S1 and Supporting Table S2. All the structure figures were prepared with PyMOL (http://www.pymol.org/). Sequence alignment was carried out with ClustalX (Chenna et al., 2003), and the alignment figure was generated with ESPript

(Gouet et al., 1999). The POA-binding site of MtRpsACTD was plotted by LIGPLOT (Wallace et al., 1995).

Isothermal titration calorimetry (ITC) assays - The dissociation constant (Kd) and stoichiometry of the

interaction between POA and MtRpsACTD or other protein mutants were measured by ITC using an ITC200 calorimeter (GE Healthcare). Calorimetric titration of POA (350 μM in the syringe; 2 μl 20

This article is protected by copyright. All rights reserved.

Accepted Article

injections) to MtRpsACTD or other mutant proteins (15 μM in the cell, 200 μl) was performed at 25 °C in assay buffer containing 10 mM Na2HPO4-NaH2PO4, pH 6.0, 150 mM NaCl. Time between injections was 150 s. ITC data were analyzed by integrating the heat effects after the data were normalized to the amount of injected protein. Data fitting was conducted to determine the dissociation constant and stoichiometry based on a single-site binding model using the Origin software package (MicroCal). NMR Experiments –

15

N-labeled S1 domain (residues 285-368) and MtRpsACTD (residues 285-481)

were produced in M9 minimal medium containing [15N]-ammonium chloride and then purified in the same manner as the unlabeled protein. The protein samples were concentrated to 0.3 mM in 10 mM Na2HPO4-NaH2PO4, pH 6.0, 150 mM NaCl. 2D 1H-15N HSQC spectra were acquired at 25°C on a Bruker Avance III 850 MHz spectrometer. In the HSQC experiment, 128 transients per increment and 128 increments were collected into 2048 data points, with spectral widths of 12 kHz in 1H dimension and 2.24 kHz in 15N-dimension, respectively. NMR spectra were processed using NMRPipe/NMRDraw (Delaglio et al., 1995) and further analyzed with the program SPARKY(Goddard). Gel mobility shift assay - [32P]tmRNA of Mtb was heated at 70 °C for 10 min and then slowly cooled

to room temperature before added to the binding mixtures. Binding reaction mixtures (40 μl) containing 20 μM [32P]tmRNA, 20 units of RNasin and 10μM or 200 μM of recombinant RpsA proteins in binding buffer (10 mM Tris-HCl pH 7.5, 100 mM NH4Cl, 10 mM MgAc, 1 mM DTT, 100 μg ml-1 bovine serum albumin and 4% glycerol) were incubated at 4 °C for 1 hr. For POA inhibition experiments, POA (300 μg ml-1) was added to the mixtures. Samples were analyzed by electrophoresis on a 5% native polyacrylamide gel (40:1).

21 This article is protected by copyright. All rights reserved.

Accepted Article

Acknowledgments: We thank Professors Lingling Chen, Hao Huang, Riqiang Fu, Zhiliang Ji and Zhichao Zhang for helpful discussions. We thank Yuchuan Zhou and Yanfei Ru for help in mobility shift assay. We also thank the staff at Shanghai Synchrotron Radiation Facility beamline 17U for assistance with data collection. This work was supported by funds from the National Natural Science Foundation of China (Nos. 31270777, 31170717, 91129713, 81261120558, 30901828, and 21232005), the National Basic Research Program of China 973 (No. 2013CB910700) and NFFTBS (No. J1310024), the Canadian Institutes of Health Research (MOP 15107).

22 This article is protected by copyright. All rights reserved.

Accepted Article

References

Adams, P.D., R.W. Grosse-Kunstleve, L.W. Hung, T.R. Ioerger, A.J. McCoy, N.W. Moriarty, R.J. Read, J.C. Sacchettini, N.K. Sauter & T.C. Terwilliger, (2002) PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr D Biol Crystallogr 58: 1948-1954.

Alexander, D.C., J.H. Ma, J.L. Guthrie, J. Blair, P. Chedore & F.B. Jamieson, (2012) Gene sequencing for routine verification of pyrazinamide resistance in Mycobacterium tuberculosis: a role for pncA but not rpsA. J Clin Microbiol 50: 3726-3728.

Aliprandi, P., C. Sizun, J. Perez, F. Mareuil, S. Caputo, J.L. Leroy, B. Odaert, S. Laalami, M. Uzan & F. Bontems, (2008) S1 ribosomal protein functions in translation initiation and ribonuclease RegB activation are mediated by similar RNA-protein interactions: an NMR and SAXS analysis. J Biol Chem 283: 13289-13301.

Boshoff, H.I. & V. Mizrahi, (1998) Purification, gene cloning, targeted knockout, overexpression, and biochemical characterization of the major pyrazinamidase from Mycobacterium smegmatis. J Bacteriol 180: 5809-5814.

Boshoff, H.I., V. Mizrahi & C.E. Barry, 3rd, (2002) Effects of pyrazinamide on fatty acid synthesis by whole mycobacterial cells and purified fatty acid synthase I. J Bacteriol 184: 2167-2172.

Bycroft, M., T.J. Hubbard, M. Proctor, S.M. Freund & A.G. Murzin, (1997) The solution structure of the S1 RNA binding domain: a member of an ancient nucleic acid-binding fold. Cell 88: 235-242.

Chenna, R., H. Sugawara, T. Koike, R. Lopez, T.J. Gibson, D.G. Higgins & J.D. Thompson, (2003) Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res 31: 3497-3500.

Collaborative Computational Project, N., (1994) The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol Crystallogr 50: 760-763.

Delaglio, F., S. Grzesiek, G.W. Vuister, G. Zhu, J. Pfeifer & A. Bax, (1995) Nmrpipe - a Multidimensional Spectral Processing System Based on Unix Pipes. J Biomol Nmr 6: 277-293.

Deryusheva, E.I., A.V. Machulin, O.M. Selivanova & I.N. Serdyuk, (2010) The S1 ribosomal protein family contains a unique conservative domain. Mol Biol+ 44: 642-647.

Emsley, P. & K. Cowtan, (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 60: 2126-2132.

Felden, B. & R. Gillet, (2011) SmpB as the handyman of tmRNA during trans-translation. RNA Biol 8: 440-449.

Goddard, T.D., Kneller, D.G., SPARKY 3. San Francisco: University of California. Gouet, P., E. Courcelle, D.I. Stuart & F. Metoz, (1999) ESPript: analysis of multiple sequence alignments in PostScript. Bioinformatics 15: 305-308.

Guo, M., Z. Sun & Y. Zhang, (2000) Mycobacterium smegmatis has two pyrazinamidase enzymes, PncA and pzaA. J Bacteriol 182: 3881-3884.

Keiler, K.C., (2008) Biology of trans-translation. Annu Rev Microbiol 62: 133-151. Lu, P., A.C. Haagsma, H. Pham, J.J. Maaskant, S. Mol, H. Lill & D. Bald, (2011) Pyrazinoic acid decreases the proton motive force, respiratory ATP synthesis activity, and cellular ATP levels. Antimicrob Agents Chemother 55: 5354-5357.

Mc, D.W. & R. Tompsett, (1954) Activation of pyrazinamide and nicotinamide in acidic environments in

23 This article is protected by copyright. All rights reserved.

Accepted Article

vitro. Am Rev Tuberc 70: 748-754.

McGinness, K.E. & R.T. Sauer, (2004) Ribosomal protein S1 binds mRNA and tmRNA similarly but plays distinct roles in translation of these molecules. Proc Natl Acad Sci U S A 101: 13454-13459.

Mitchison, D.A., (1985) The action of antituberculosis drugs in short-course chemotherapy. Tubercle 66: 219-225.

Murzin, A.G., (1993) OB(oligonucleotide/oligosaccharide binding)-fold: common structural and functional solution for non-homologous sequences. EMBO J 12: 861-867.

Otwinowski, Z. & W. Minor, (1997) Processing of X-ray diffraction data collected in oscillation mode. Method Enzymol 276: 307-326.

Qi, H., Y. Shimizu & T. Ueda, (2007) Ribosomal protein S1 is not essential for the trans-translation machinery. J Mol Biol 368: 845-852.

Salah, P., M. Bisaglia, P. Aliprandi, M. Uzan, C. Sizun & F. Bontems, (2009) Probing the relationship between Gram-negative and Gram-positive S1 proteins by sequence analysis. Nucleic Acids Res 37: 5578-5588.

Scorpio, A. & Y. Zhang, (1996) Mutations in pncA, a gene encoding pyrazinamidase/nicotinamidase, cause resistance to the antituberculous drug pyrazinamide in tubercle bacillus. Nat Med 2: 662-667.

Shi, W., X. Zhang, X. Jiang, H. Yuan, J.S. Lee, C.E. Barry, 3rd, H. Wang, W. Zhang & Y. Zhang, (2011) Pyrazinamide inhibits trans-translation in Mycobacterium tuberculosis. Science 333: 1630-1632.

Simons, S.O., A. Mulder, J. van Ingen, M.J. Boeree & D. van Soolingen, (2013) Role of rpsA gene sequencing in diagnosis of pyrazinamide resistance. J Clin Microbiol 51: 382.

Sorensen, M.A., J. Fricke & S. Pedersen, (1998) Ribosomal protein S1 is required for translation of most, if not all, natural mRNAs in Escherichia coli in vivo. J Mol Biol 280: 561-569.

Sreevatsan, S., X. Pan, Y. Zhang, B.N. Kreiswirth & J.M. Musser, (1997) Mutations associated with pyrazinamide resistance in pncA of Mycobacterium tuberculosis complex organisms. Antimicrob Agents Chemother 41: 636-640.

Subramanian, A.R., (1983) Structure and functions of ribosomal protein S1. Prog Nucleic Acid Res Mol Biol 28: 101-142.

Takada, K., C. Takemoto, M. Kawazoe, T. Konno, K. Hanawa-Suetsugu, S. Lee, M. Shirouzu, S. Yokoyama, A. Muto & H. Himeno, (2007) In vitro trans-translation of Thermus thermophilus: ribosomal protein S1 is not required for the early stage of trans-translation. RNA 13: 503-510.

Tan, Y., Z. Hu, T. Zhang, X. Cai, H. Kuang, Y. Liu, J. Chen, F. Yang, K. Zhang, S. Tan & Y. Zhao, (2014) Role of pncA and rpsA Gene Sequencing in Detection of Pyrazinamide Resistance in Mycobacterium tuberculosis Isolates from Southern China. J Clin Microbiol 52: 291-297.

Wallace, A.C., R.A. Laskowski & J.M. Thornton, (1995) LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions. Protein engineering 8: 127-134.

Wower, I.K., C.W. Zwieb, S.A. Guven & J. Wower, (2000) Binding and cross-linking of tmRNA to ribosomal protein S1, on and off the Escherichia coli ribosome. EMBO J 19: 6612-6621.

Zhang, Y., A. Scorpio, H. Nikaido & Z. Sun, (1999) Role of acid pH and deficient efflux of pyrazinoic acid in unique susceptibility of Mycobacterium tuberculosis to pyrazinamide. J Bacteriol 181: 2044-2049.

Zhang, Y., M.M. Wade, A. Scorpio, H. Zhang & Z. Sun, (2003) Mode of action of pyrazinamide: disruption of Mycobacterium tuberculosis membrane transport and energetics by pyrazinoic acid. J Antimicrob

24 This article is protected by copyright. All rights reserved.

Accepted Article

Chemother 52: 790-795.

Zimhony, O., J.S. Cox, J.T. Welch, C. Vilcheze & W.R. Jacobs, Jr., (2000) Pyrazinamide inhibits the eukaryotic-like fatty acid synthetase I (FASI) of Mycobacterium tuberculosis. Nat Med 6: 1043-1047.

Table 1. Thermodynamic parameters of RpsA-POA interactions determined by ITC

Protein Kd (µM) Na H (kcal/mol) MtRpsA 2.03 2.810.67 -27.960.20 CTD MtRpsA 1.94 3.130.06 -62.100.12 4th S1 domainb 1.96 3.210.08 -68.320.21 K303Ac 1.44 7.810.86 -47.870.17 F307A 1.61 8.030.35 -52.210.65 F310G 1.07 9.341.04 -77.270.52 R357A 1.85 8.470.53 -49.960.10 a N: Stoichiometry of the binding (POA/RpsA) b th 4 S1 domain: residues 285-368 of MtRpsACTD c K303A, F307A, F310G, R357A: mutants of MtRpsACTD

25 This article is protected by copyright. All rights reserved.

TΔS(kcal/mol) -20.38 -54.53 -60.77 -40.83 -45.29 -70.33 -42.91

ΔG(kcal/mol) -7.580.05 -7.570.03 -7.540.06 -7.040.05 -6.920.21 -6.940.18 -7.050.02

Accepted Article

FIGURE LEGENDS

FIGURE 1. Crystal structures of the C-terminal domain of Mtb RpsA (MtRpsACTD) and its complex with POA. (A) Domain organization of RpsA of Mtb. The boxes denote the predicted S1 domains based on

sequence analysis. The residue numbers corresponding to each S1 domain and the predicted α-helix are labeled. The C-terminal domain (MtRpsACTD, residues 285-481), which includes the fourth S1 domain and the C-terminus, is indicated. (B) Crystal structure of MtRpsACTD. The S1 domain is boxed and the invisible C-terminus is indicated. Secondary structure elements, N- and C-termini are indicated. (C) Conserved residues of MtRpsACTD involved in RNA binding. Side chains of residues (Phe307, Phe310, His322, Asp352 and Arg357) predicted to be involved in RNA binding are highlighted in red in the structure of MtRpsACTD. The views of (B) and (C) are related by ∼ 90° rotation. (D) Crystal structure of the MtRpsACTD-POA complex. Two POA molecules (POAI, POAII) are shown in stick representation and colored by elements, with carbon in yellow, oxygen in red, and nitrogen in blue. (E) Sequence alignment of specific S1 domains from different bacterial species. Mt, Mycobacterium tuberculosis (4th S1 domain); Te, Thermosynechococcus elongates (3rd S1 domain); Se, Staphylococcus epidermidis (4th S1 domain); Ec, Escherichia coli (3rd S1 domain); Lr, Leptospirillum rubarum (2nd S1 domain); Sk, Spiroplasma kunkelii (1st S1 domain). Conserved

residues are in red background. Similar residues are shown in red color and blue-squared boxes denote homologous residues. Black asterisks indicate conserved residues involved in RNA 26

This article is protected by copyright. All rights reserved.

Accepted Article

binding (Bycroft et al., 1997, Salah et al., 2009, Deryusheva et al., 2010). The numbering is based on a typical S1 domain sequence where MtRpsACTD 289-363 corresponds to the S1 domain residues 1-74.

FIGURE 2. Structural and molecular nature of MtRpsACTD-POA interactions. (A) Electrostatic potential surface representation of the structure of the MtRpsACTD-POA complex

with a close-up view of the ligand-binding sites. A color-code bar shows an electrostatic scale from -77 to +77 eV. (B) Details of the interactions between the substrate-binding site and POA. The ‘omit’ electron density for POA is contoured at 2.5σ, shown as a blue mesh; POA contacting residues are labeled and shown as stick representations. Hydrogen bonds are indicated as black dashed lines. (C) Structural superimposition of apo-MtRpsACTD (in light blue) and MtRpsACTD-POA (in pale cyan). Different conformations of the β2, β3, and β5 strands are detected.

FIGURE 3. Structural comparisons of MtRpsACTD with MtRpsA△A438CTD and MsRpsACTD. (A) Sequence alignment of MtRpsACTD, MtRpsA△A438CTD and MsRpsACTD. Strictly conserved residues are in red background. Similar residues are shown in red color and blue-squared boxes denote homologous residues. The mutations conferring PZA-resistance are shown in green box. (B) Structural superimposition of MtRpsACTD, MtRpsA△A438CTD and MsRpsACTD, which are colored with light blue, wheat and pink, respectively. No gross structural changes are detected. (C) Superimposition of the POA-binding sites of MtRpsACTD, MtRpsA△A438CTD colored in light 27

This article is protected by copyright. All rights reserved.

Accepted Article

blue and wheat, respectively. POA-contacting residues are labeled and shown as stick representations, and the linear distances between the corresponding atoms of residues Lys303, Phe307, Phe310 and Arg357 were measured with PyMol. (D) Superimposition of the POA-binding sites of MtRpsACTD and MsRpsACTD colored in light blue and pink, respectively. POA-contacting residues are labeled and shown as stick representations, and the linear distances between the corresponding atoms of residues Lys303, Phe307, Phe310 and Arg357 were measured with PyMol. (E) Electrostatic surface potentials of MtRpsACTD (left), MtRpsA△A438CTD (middle) and MsRpsACTD (right) around the POA-binding site (shown in the same orientation and using the

same color code as in Fig. 2A). The area of prominent change is boxed.

FIGURE 4. Superimposition of the 1H-15N HSQC spectra of the isolated S1 domain (residues 285-368) and MtRpsACTD (residues 285-481) colored with red and blue, respectively. Not all NH signals of the S1 domain are overlapped with that of MtRpsACTD.

FIGURE 5. Gel mobility shift assays of RpsA-tmRNA interactions. (A) tmRNA binding activities of the full-length RpsA of Mtb (MtRpsA) and its C-terminal

domain (MtRpsACTD). Lane 1: 20 μM [32P]-labeled tmRNA of Mtb; lanes 2 and 3: MtRpsA was added to the reaction mixtures at 10 and 200 μM, respectively; lane 4: MtRpsA at 200 μM and POA at 300 μg ml-1 were added; lanes 5 and 6: MtRpsACTD was added at 10 and 200 μM, respectively. lane 7: MtRpsACTD at 200 μM and POA at 300 μg ml-1 were added. 28

This article is protected by copyright. All rights reserved.

Accepted Article

(B) tmRNA binding activities of the two RpsA variants (MtRpsA△A438 and MsRpsA) and their C-terminal domains (MtRpsA△A438CTD and MsRpsACTD). Lane 1: tmRNA; lanes 2 and 3: MtRpsA△A438 was added at 10 and 200 μM, respectively; lane 4: MtRpsA△A438 at 200 μM and POA at 300 μg ml-1 were added; lanes 5 and 6: MsRpsA was added at 10 and 200 μM,

respectively; lane 7: MsRpsA at 200 μM and POA at 300 μg ml -1 were added; lanes 8 and 9:

MtRpsA△A438CTD was added at 10 and 200 μM, respectively; lane 10: MtRpsA△A438CTD at 200 μM and POA at 300 μg ml-1 were added; lanes 11 and 12: MsRpsACTD was added at 10 and 200 μM, respectively; lane 13: MsRpsACTD at 200 μM and POA at 300 μg ml-1 were added.

(C) tmRNA binding activities of RpsA mutants. MtRpsACTD and indicated MtRpsACTD mutants were added at 200 μM in lanes 2-8, respectively, with 20 μM [32P]-labeled tmRNA in each lane. Lane 1: tmRNA control.

29 This article is protected by copyright. All rights reserved.

Accepted Article mmi_12892_f1.tif

Accepted Article mmi_12892_f2.tif

Accepted Article mmi_12892_f3.tif

Accepted Article mmi_12892_f4.tif

Accepted Article mmi_12892_f5.tif

Structural basis for targeting the ribosomal protein S1 of Mycobacterium tuberculosis by pyrazinamide.

Pyrazinamide (PZA) is a first-line drug for tuberculosis (TB) treatment and is responsible for shortening the duration of TB therapy. The mode of acti...
2MB Sizes 0 Downloads 6 Views