Phosphatase-defective DevS sensor kinase mutants permit constitutive expression of 1 DevR-regulated dormancy genes in Mycobacterium tuberculosis

The DevR-DevS/DosR-DosS two-component system of Mycobacterium tuberculosis , that 26 comprises of DevS sensor kinase and DevR response regulator, is essential for bacterial 27 adaptation to hypoxia by inducing dormancy regulon expression. The dominant phosphatase 28 activity of DevS under aerobic conditions enables tight negative control, whereas its kinase 29 function activates DevR under hypoxia to induce the dormancy regulon. A net balance in these 30 opposing kinase and phosphatase activities of DevS calibrates the response output of DevR. To 31 gain mechanistic insights into the kinase-phosphatase balance of DevS, we generated alanine 32 substitution mutants of five residues located in DHp α1 helix of DevS, namely Phe-403, Gly- 33 406, Leu-407, Gly-411 and His-415. For the first time, we have identified kinase positive 34 phosphatase negative (K + P - ) mutants in DevS by single-site mutation in either Gly-406 or Leu- 35 407. M. tuberculosis Gly-406A and Leu-407A mutant strains constitutively expressed the DevR 36 regulon under aerobic conditions despite of the presence of negative signal, oxygen. These 37 mutant proteins exhibited ~2-fold interaction defect with DevR. We conclude that Gly-406 and 38 Leu-407 residues are individually essential for phosphatase function of DevS. Our study 39 provides new insights into negative control mechanism of DevS by demonstrating the 40 importance of an optimal interaction between DevR and DevS, and local changes associated with 41 individual residues, Gly-406 and Leu-407, which mimic ligand-free DevS. These K + P - mutant 42 strains are expected to facilitate the rapid aerobic screening of DevR antagonists in M. 43 tuberculosis, thereby eliminating the requirement for hypoxic culture conditions.

DevR-DevS/DosT is a widely studied TCSs of Mtb wherein, DevR is the RR and DevS/DosT are 80 the two SKs which control its activity [4,5]. DevR and its ~48-gene regulon play an essential role 81 in bacterial adaptation to dormancy in response to various gaseous signals such as hypoxia, CO,82 NO or vitamin C [6]. Therefore it is considered as a novel target for developing strategies against 83 non-replicating or dormant bacteria [6-9]. DevS, like NarX and DesK, belongs to HisKA_3 84 subfamily of sensor kinases and we showed that in the presence of negative signal such as oxygen 85 (O2), under aerobic conditions the phosphatase activity of DevS is dominant which prevents 86 regulon induction, while under inducing conditions (hypoxia, CO, NO or vitamin C), the kinase 87 activity is prominent which favors phosphorylation of DevR [10]. We further showed the 88 involvement of the highly conserved His-395 residue and the DxxxQ motif (residues 396 to 400 89 in DevS) located in the α1 helix of Dimerization and Histidine phosphotransfer (DHp) domain of 90 DevS in both kinase (positive) and phosphatase (negative) functions of DevS; and found that a fine 91 tuning of these opposing activities determines the output response of this TCS [10]. A complete 92 understanding of the negative control mechanism of DevS is limited by the absence of kinase-93 positive, phosphatase-defective (K + P -) missense substitutions in DevS. Furthermore, the 94 identification of residues exclusively involved in negative control will provide a DevS mutant 95 strain (having constitutive DevR regulon expression) that would facilitate the rapid screening of 96 anti-DevR/DevS compounds under aerobic conditions. 97 Analysis of missense substitution mutants in the transmitter module (TM) of NarX SK, also a 98 member of the HisKA_3 family, has shown that the substitutions exhibiting K + Pphenotype are 99 located in the α1 and α2 helices of the DHp domain [11,12]. In the present study, we have analyzed 100 five residues, namely Phe-403, Gly-406, Leu-407, Gly-411 and His-415, located in the vicinity of 101 All the resultant plasmids (Table S2) were sequence verified and for in vitro experiments, were 138 transformed into E. coli BL21 to express the recombinant DevS WT/mutant proteins as described 139 earlier [4]. For in vivo analysis, the selected plasmids were electroporated into Mtb ∆dosS∆dosT 140 (DKO). Mtb transformants denoted by DevS WT, G406A and L406A were selected on 7H11 agar 141 containing hygromycin and kanamycin (Table S1). All the mutant strains of Mtb were confirmed 142 to express DevS protein (data not shown).  DevR antibody (generated in rabbit) was then added to each well at a dilution of 1:5000 and the 195 plate was incubated for 2 h at 25°C. The plate was again washed thrice with 1x PBS and 0.1% 196 Tween-20 and twice with 1x PBS, followed by the addition of goat anti-rabbit IgG-HRP conjugate 197 secondary antibody (1:5000, GeNei™, India) for 1 h at 25°C. The plate was again washed, 198 developed with 3,3', 5,5' tetramethylbenzidine substrate and absorbance was measured at 450 nm. 199

200
In vivo kinase and phosphatase function assays. To assess in vivo kinase activity, Mtb was 201 cultured in 100 ml of DTA medium to an OD595 of ~ 0.2-0.3. A 50 ml aliquot (10 ml in 50-ml 202 tubes in triplicate for RNA isolation and in duplicate for urea lysate preparation) was centrifuged 203 immediately for 'aerobic' condition sample. The remaining 50 ml culture was distributed in 10 ml 204 aliquots as described above and kept standing for 5 days to generate 'hypoxic' cultures for RNA 205 isolation and urea lysate preparation as described [10]. 206 For in vivo phosphatase assay, DevS WT/mutant Mtb strains were aerobically cultured in 7H9 207 acetate or 7H9 glucose medium (10 ml culture in 50-ml tubes) and the cells were harvested at an 208 Life Technologies TM USA). The cDNA was analysed by qPCR using gene-specific primers (Table  222 S4) and iQ TM SYBR® Green Supermix (in CFX96 Real-Time PCR-detection system, Bio-Rad).

230
DHp α1 helix of HisKA_3 family harbours residues crucial for phosphatase function. 231 Several residues of the α1 helix in the DHp domain have been implicated in kinase positive 232 phosphatase negative (K + P -) functions in NarX SK of the HisKA_3 family [11,12]. DevS belongs 233 to the same family of proteins and to identify residues that confer a K + Pphenotype, we analysed 234 selected residues located distal to the DxxxQ motif in the α1 helix of the DHp domain in DevS. 235 The selected amino acid residues, Phe-403, Gly-406, Leu-407, Gly-411, His-415, are marked on 236 DevS structure modelled on that of DesK structure using PyMOL software ( DevS WT and its alanine substitution mutant proteins were tested for autokinase activity using γ 32 P 248 ATP which leads to phosphorylation at H395 residue [4] ( Fig. 2A). Relative to DevS~P level of 249 WT protein at 60 min (which was taken as 100%), G411A mutant exhibited hyperkinase activity 250 (~188%), F403A showed a major decrease in autokinase activity (~86% decrease), whereas 251 G406A, L407A and H415A exhibited a near-normal or weak kinase activity (~10%, 43% and 252 46% decrease, respectively, Fig. 2B  to that of DevS (t1/2 of DevR~P ~46 min vs. ~3 min in the presence of DevS WT). The phosphatase 257 activity of DevS WT and its mutant proteins was assayed next by measuring the dephosphorylation 258 of DevR~3 2 P (Fig. 2D). DevS WT mediated a rapid and efficient dephosphorylation of DevR~3 2 P 259 (t1/2 of ~3 min, Fig. 2E and F), F403A, G406A and L407A exhibited a significant decrease in 260 phosphatase activity (t1/2 of ~16 to 36 min, Fig. 2E and F), whereas G411A and H415A mutants 261 retained their phosphatase activity (t1/2 of ~0.8 to 3.2 min, Fig. 2E and F). From autokinase and 262 phosphatase analysis of these DevS mutant proteins, we conclude that F403A exhibits a substantial 263 decrease in both kinase (partial, abbreviated as 'par') and phosphatase function, whereas G406A, 264 L407A G411A and H415A preserved their kinase activity to varying extents, and of these, two 265 mutant proteins namely G406A and L407A, exhibited a substantial decrease in phosphatase 266 activity (K + P -, Table 1). 267 268

Phosphotransfer activity. 269
The phosphotransfer assay is a coupled assay wherein we monitored the ability of DevS mutant 270 protein to transfer the phosphosignal to DevR and also dephosphorylate DevR~P (Fig. 3A). The 271 G406A and L407A K + Pmutants exhibited a delay in phosphotransfer activity as compared to 272 DevS WT (Fig. 3B, C). Importantly, the radioactive signal in DevR~P was retained for a longer 273 time duration in the presence of G406A and L407A proteins which was attributed to a substantial 274 decrease in the phosphatase activity of these mutants. In contrast, the G411A and H415A mutants 275 were active in phosphotransfer and phosphatase function (Fig. 3B, C) and the F403A mutant 276 poorly transferred phosphosignal to DevR (Fig. 3B, C). 277 278 Interaction between DevR and DevS. 279 The interaction between purified DevR and DevS WT/mutant derivatives was monitored next 280 using ELISA. A 96-well microtitre plate was coated with DevS WT/mutant proteins and interacted 281 with GST-tagged DevR protein. Protein-protein interaction was monitored using anti-GST 282 monoclonal antibody and HRP-conjugated Goat anti-Mouse IGg antibody. A reduction in 283 interaction was observed for the K + Pmutants, G406A and L407A and for K par P -F403A mutant 284 which might underlie the defect in their phosphatase activity (Fig. 3D). H415A also showed an 285 interaction defect although it did not hamper its phosphatase activity. The G411A mutant protein 286 showed a 2-fold enhanced interaction with DevR as compared to DevS WT which may contribute 287 to its hyperkinase activity (Fig. 3D). The observed defects in the phosphotransfer and phosphatase 288 activities of F403A, G406A and L407A mutant proteins could be attributed to a defect in 289 interaction between DevR and DevS proteins as a consequence of mutation. 290 291

In vivo activity of Mtb DevS mutants. 292
In view of the limitation of studying truncated proteins lacking the regulatory sensory module 293 (SM) in the in vitro assays described above, we proceeded to expand our understanding of DevS 294 negative regulatory function by performing in vivo studies. For this we analyzed the two K + P -295 mutants, G406A and L407A, that were identified through in vitro analysis, by generating Mtb 296 strains that constitutively express full-length versions of DevS WT/mutant protein in a double 297 knockout strain of devS and dosT (Table S1). The complemented strains were confirmed to express 298 equivalent levels of DevS protein to ensure an unequivocal interpretation of the in vivo assay 299 readouts (data not shown). 300 301 Residues G406 and L407 do not contribute to DevS positive function in vivo. 302 The standing 5-day hypoxia model [10] was used to assess the in vivo kinase activity of the K + P -303 mutants (Fig. 4A). Herein, the expression of DevR regulon genes under hypoxia was used as a To examine the in vivo phosphatase function of G406 and L407 residues, DevR regulon gene 320 expression was monitored in aerobic cultures of Mtb G406A and L407A mutants that were grown 321 in 7H9 acetate medium (Fig. 4D). Briefly in this aerobic model, when Mtb is grown on acetate, 322 DevR is phosphorylated at D54 by acetyl phosphate leading to regulon induction [20]; and the 323 level of induction was compared to that in Mtb cultures grown in 7H9 glucose (non-inducing 324 condition) as described earlier [10]. 325

326
The expression of four representative regulon genes, namely hspX, fdxA, devR, and Rv3130, was 327 examined in this aerobic model. None of these genes were induced in the DevS WT strain grown 328 in either 7H9 acetate or 7H9 glucose media, which was ascribed to its potent phosphatase activity. 329 The regulon expression in 7H9 glucose-grown DevS WT culture was used as a control to determine 330 relative gene expression levels in the mutants. An induction of all the four regulon genes was 331 observed (≥2 to 95-fold) in G406A and L407A Mtb mutant strains irrespective of the culture media 332 that was used to grow the bacteria (Fig. 4E). The constitutive expression of HspX protein was 333 consistent with the RNA expression pattern of their corresponding genes in these mutant strains 334 (Fig. 4F).From this analysis, we conclude that G406 and L407 residues of DevS are individually 335 sufficient for conferring negative (phosphatase) function without majorly influencing its kinase 336 activity. This study has yielded two Mtb strains with K + Pphenotype that display a constitutive 337 expression of the DevR regulon gene in a condition-(aerobic/hypoxic) and culture medium-338 (glucose/acetate) independent manner. The constitutive expression of the DevR regulon in these 339 mutants is attributed to the absence of strong negative regulatory function that resides in G406 and 340 L407 residues of DevS. class of mutants. They were assigned as K + P - (Fig. 5A); mutation in either G406 or L407 residue 363 in the α1 helix led to a near-complete abrogation of phosphatase function and enabled 364 constitutive expression of the DevR regulon in Mtb (ON phenotype, Fig. 5B). This helix in DevS 365 contains the highly conserved phosphorylatable H395 residue and the conserved DxxxQ (396 to 366 400 residues) motif that are central to the kinase and phosphatase activities of SKs [10, 12] and 367

forms a major interaction interface between SK and RR [25-27]. Phosphatase activity in 368 particular is governed by strong interaction between the TM of SK and phosphorylated RR [28]. 369
In this context, the ~50% decrease in interaction between DevS G406A/ L407A and DevR may 370 compromise the interaction interface between DevS and DevR. A similar K + Pphenotype was 371 reported for mutations in the corresponding residues of NarX, namely K410 and M411, where 372 these residues are aligned on the same helical phase as Q404 of the DxxxQ motif in NarX [11]; 373 whereas in DevS these residues are in close proximity but offset to Q400 in the DHp α1 helix.