FEMS Microbiology Letters, 364, 2017, fnw294

doi: 10.1093/femsle/fnw294Advance Access Publication Date: 24 December 2016Research Letter

RESEARCH LETTER –Physiology & Biochemistry

Insights into the function of FhaA, a celldivision-associated protein in mycobacteriaGopinath Viswanathan, Sangya Yadav, Shrilaxmi V. Joshi†

and Tirumalai R. Raghunand∗

CSIR—Centre for Cellular and Molecular Biology, Uppal Road, Hyderabad 500007, India∗Corresponding author: CSIR—Centre for Cellular and Molecular Biology, Uppal Road, Hyderabad 500007, India. Tel: +91-40-27192924;Fax: +91-40-27160591; E-mail: raghu@ccmb.res.in†Present address: Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore, India.One sentence summary: Mycobacterial FhaA interacts with PbpA and regulates peptidoglycan biosynthesis.Editor: Jana Jass

ABSTRACT

FhaA is a forkhead-associated domain-containing protein, the depletion of which leads to accumulation of peptidoglycan(PG) precursors at the septum and poles in Mycobacterium smegmatis (M. smegmatis), by a mechanism undefined thus far. Toelucidate its function, we constructed an fhaA (MSMEG 0035) knockout (�fhaA) strain in M. smegmatis and demonstratedthat this gene is dispensable for in vitro growth. The mutant showed a short cell length phenotype due to a probable defectin cell elongation/cell wall synthesis, which was reversed by complementation with both M. smegmatis and Mycobacteriumtuberculosis (M. tb) fhaA (Rv0020c), confirming their association with the observed phenotype. The identification of penicillinbinding protein A (PbpA), a PG biosynthesis enzyme as an interacting partner for mycobacterial FhaA, provided a hint intothe functioning of FhaA. A drastic reduction in the levels of ectopically expressed PbpA in the �fhaA mutant vs wild-type M.smegmatis suggested that FhaA interacts with and stabilises PbpA. In addition, the fhaA deletion mutant was sensitive tomultiple classes of antibiotics pointing to a general permeability defect. Our findings uncover a role for FhaA in PGbiosynthesis and suggest its involvement in the maintenance of mycobacterial cell envelope integrity.

Keywords: FhaA, PbpA; peptidoglycan; Mycobacterium smegmatis; Mycobacterium tuberculosis; cell envelope

INTRODUCTION

As in other eubacteria, mycobacterial cell division is tightlyregulated through membrane receptors as well as cytoplas-mic sensors that transcriptionally and post-translationallyregulate cell division-associated proteins in response to theenvironment (Kieser and Rubin 2014). One of these regu-latory systems involves phosphorylation of a few cell divi-sion proteins through serine/threonine protein kinases (STPKs)(Molle and Kremer 2010). PknA and PknB, 2 of the 11STPKs encoded by Mycobacterium tuberculosis (M. tb), are partof an operon that includes the cell division genes rodA(Rv0017c, predicted to be involved in septum-peptidoglycan (PG)

biosynthesis (http://tuberculist.epfl.ch/) and maintaining cellshape (Henriques et al. 1998) and pbpA (Rv0016c, predicted tobe involved in PG biosynthesis (Fedarovich, Nicholas and Davies2010), (http://tuberculist.epfl.ch/), indicating a probable associ-ation of these kinases with cell division. The synteny of thisoperon is conserved across the mycobacterial genus indicat-ing a conserved functional relationship between these genes(Narayan et al. 2007). Consistent with this observation, PknA andPknB were found to be essential, and involved in regulating cellshape (Sassetti, Boyd and Rubin 2003; Kang et al. 2005). Theirsubstrates include the cell division-associated proteins FtsZ,Wag31, MurD, GlmU and PonA1(Thakur and Chakraborti 2006;

Received: 13 July 2016; Accepted: 23 December 2016C© FEMS 2016. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

1

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018

2 FEMS Microbiology Letters, 2017, Vol. 364, No. 2

Thakur and Chakraborti 2008; Parikh et al. 2009; Jani et al. 2010;Kieser et al. 2015). Phosphorylation of these substrates has beendemonstrated to directly regulate their function or their locali-sation (Thakur and Chakraborti 2006; Parikh et al. 2009; Jani et al.2010; Kieser et al. 2015). Adding a layer of complexity to STPKphosphorylation are the six forkhead-associated (FHA) domain-containing proteins in the M. tb genome which regulate pro-tein function by binding to phosphothreonine residues via theirFHA domain. Among these are FhaA and FhaB which are en-coded in the rodA/pbpA/pknA (Rv0015c) and pknB (Rv0014c) clus-ter, indicating their probable functional association with thesekinases and cell division (Pallen, Chaudhuri and Khan 2002). Ho-mologues of FhaA and FhaB are present across mycobacteria(http://tuberculist.epfl.ch/), implying a conserved role for theseproteins in mycobacterial physiology. We have earlier reportedthat a transposon mutant of Mycobacterium smegmatis (M. smeg-matis) fhaA (MSMEG 0035) exhibited a small colony phenotypeand showed increased susceptibility to oxidative stress, proba-bly due to increased permeability resulting from a compromisedcell envelope (Viswanathan et al. 2015). Additionally, the knock-down of fhaA inM. smegmatis led to the accumulation of PG pre-cursors at the bacillary septum and poles, indicating a probabledefect in PG biosynthesis (Gee et al. 2012). The authors observedthat M. smegmatis FhaA interacts with MviN, a putative lipid IIprecursor flippase, and hypothesised that due to this interac-tion FhaA regulates the activity of MviN, thereby controlling theaccumulation of PG precursors. These observations however arepoorly understood at the mechanistic level.

Here, using genetic and biochemical approaches, we demon-strate a functional link between mycobacterial fhaA and PGbiosynthesis. Further, the altered susceptibility of a �fhaA strainof M. smegmatis to various classes of antibiotics revealed a cellenvelope-associated function for this gene. Our observationsprovide new insights into the role of fhaA in PG biosynthesis aswell as establish its functional association with the mycobacte-rial cell envelope.

MATERIALS AND METHODS

Bacterial strains, media and growth conditions

Mycobacterium smegmatis mc26, M. smegmatis mc2155 (Snapperet al. 1990), Escherichia coli (E. coli) DH5α and E. coli BL21 (DE3)(Table S1, Supporting Information) were cultured as described(Tiwari et al. 2012). Tween 80 (0.05%) was added to mycobacterialcultures to prevent clumping of strains. The following antibioticswere addedwhen necessary—kanamycin (15 μg/mL forM. smeg-matis and 50 μg/mL for E. coli) and hygromycin (200 μg/mL forE. coli and 50 μg/mL for M. smegmatis).

Construction of fhaA knockout strain in Mycobacteriumsmegmatis (�fhaAMs)

A recombineering strategy was used to construct �fhaAMs

(Table S1) as described previously (van Kessel and Hatfull 2007).

Mycobacterial protein fragment complementationassay

To assess the interactions of FhaAMs-PbpAMs and FhaAMtb-PbpAMtb at the genetic level, mycobacterial protein fragmentcomplementation (M-PFC) assay was performed as described(Singh et al. 2006).

Pull-down assays

For biochemical validation of M-PFC results, pull-down assayswere performed.

Subcellular localisation

The localisation of C-terminal c-myc fusions of PbpAMs andFhaAMs inM. smegmatis strainswas determined as described ear-lier (Cascioferro et al. 2007).

RT-PCR analysis

For co-expression analysis of fhaAMs and pbpAMs in wild-type(WT) M. smegmatis mc26 and to determine the levels of ectopi-cally expressed pbpAMs-cMyc inWT and �fhaAMs strains, RT-PCRwas performed.

Construction of PbpA-GFP and fluorescence microscopy

To determine the cellular localisation of PbpAMs in WT M. smeg-matis and �fhaAMs, a PbpA-GFP fusion was generated and itslocalisation pattern was determined using fluorescence mi-croscopy.

Resazurin based microplate assay

For MIC determination, a resazurin-based microplate assay wasperformed in 96-well plates as described earlier (Taneja andTyagi 2007; Viswanathan, Yadav and Raghunand 2016). Two-folddilutions with following ranges of concentrations were used forthe antibiotics: amoxicillin: 256–2 μg ml−1, ampicillin: 256–2 μgml−1, rifampicin: 32–1 μg ml−1.

Ethidium bromide permeability assay

To check the cell envelope integrity of �fhaAMs in compari-son to the WT M. smegmatis and its complements, an ethidiumbromide (EtBr) permeability assay was performed as described(Danilchanka, Mailaender and Niederweis 2008).

Detailed materials and methods described in Supplemen-tary information include bacterial strains, media and growthconditions, DNA techniques, construction of �fhaAMs, Southernblotting, complementation of �fhaAMs, staining of M. smegmatismc26,�fhaAMs, TR49 [fhaAMs::Tn] and their complements, colonyphenotype documentation, in vitro growth kinetics, M-PFC as-say, in silico analyses, expression and purification of GST-FhaAMs,GST-FhaAMsFHA133 and GST-FhaAMtb, pull-down assays, subcel-lular localisation, RT-PCR analysis, construction of the PbpA-GFPfusion and fluorescencemicroscopy, resazurin basedmicroplateassay and EtBr permeability assay.

RESULTS

The �fhaA mutant of Mycobacterium smegmatis exhibitsshort cell length and altered colony morphologyphenotypes

To elucidate the function of fhaA, a recombineering strategy wasfollowed to create a knockout strain ofM. smegmatismc26, wherefhaA was replaced by a hygromycin resistance gene (Fig. 1a).Of the six hygromycin-resistant colonies obtained, one colonywas found to carry an fhaA deletion with a hygromycin resis-tancemarker insertion as revealed by PCR screening (Fig. 1b) and

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018

Viswanathan et al. 3

Figure 1. Construction and characterisation of a �fhaA mutant strain of M. smegmatis. (a). Schematic representation of the recombineering strategy used to generatethe fhaA deletion mutant of M. smegmatis. (b) PCR-based detection of the �fhaA mutant. Colonies 1–5 are genetically WT based on sizes of the amplification products.The expected amplification product of the �fhaAmutant is seen in colony 6. (c) Cellular morphologies of carbolfuchsin stained WT and �fhaA strains ofM. smegmatis,along with the fhaAMs and fhaAMtb complemented strains of the deletion mutant (the scale bars are 10μm). The histogram (d) shows the mean cells lengths of the

above strains and the histogram (e) shows their cell length distribution. N = 100; the error bars represent standard error of the mean (SEM), ∗∗∗P < 0.0001. (f) Colonymorphologies of WT, �fhaA strains of M. smegmatis along with the fhaAMs and fhaAMtb complemented strains of the deletion mutant. The scale bars are 0.5 mm.

further verified by sequencing the cognate PCR product. Thegenotype of �fhaAMs and the integrity of its genome were fur-ther confirmed by Southern blotting, which showed the absenceof fhaA and the presence of a single hyg cassette in the dele-tion strain (Fig. S1, Supporting Information), leading us to con-clude that fhaA is a non-essential gene inM. smegmatis. It is likelythat the low recombination frequency we observed is due to the

shorter flanking region (van Kessel andHatfull 2007) used for theconstruction of �fhaAMs(Fig. 1a).

Cells of �fhaAMs were found to be shorter in length than theWT strain (Fig. 1c and d), and the mutant showed a drastic re-duction in the number of cells ranging in length from 3 to 5 μm(Fig. 1e) implying a possible cell elongation defect. A mutantwith a transposon insertion in fhaA (TR49 [fhaAMs::Tn]) that we

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018

4 FEMS Microbiology Letters, 2017, Vol. 364, No. 2

had isolated previously (Viswanathan et al. 2015) also showeda short cell length phenotype, consistent with this observation(Fig. S2, Supporting Information). Importantly, complementa-tion of �fhaAMs and TR49 (Viswanathan et al. 2015) withM. smeg-matis fhaA and itsM. tb homologue reversed their cell length de-fects (Fig. 1c, d, e and S2), confirming their association with theobserved phenotype. Further, colonies of �fhaAMs were smallerand showed a more defined margin in comparison to the WT(Fig. 1f), similar to what we had observed for TR49 (Viswanathanet al. 2015) suggesting a probable alteration in its cell envelope.Complementation of �fhaAMs with M. smegmatis fhaA and its M.tb homologue was observed to restore theWT colony phenotype(Fig. 1f). The small colony morphotype of �fhaAMs prompted usto check its growth kinetics, and we observed amoderate reduc-tion in its growth in comparison to the WT and complementedstrains (Fig. S3, Supporting Information), linking this gene withgrowth-associated processes. These results validated the associ-ation of fhaAMs and its M. tb homologue with mycobacterial cellelongation and provided a pointer to their role in cell envelope-associated functions.

Mycobacterial FhaA interacts with the class B penicillinbinding protein PbpA

To decipher the cell elongation-associated function of FhaA, wetested its interaction with PbpA, a predicted transpeptidase in-volved in PG cross-linking, encoded in the rodA/pbpA/pknA andpknB cluster. Since co-operonic partners often interact with eachother, this seemed a valid approach to identify functional as-sociations among these proteins. Furthermore, PknA and PknBhave already been demonstrated to interact with and phospho-rylate FhaA inM. tb (Grundner, Gay and Alber 2005). Both FhaAMs

and FhaAM.tb were observed to interact with PbpA from theircognate species in the M-PFC assay (Fig. 2a and b), an indica-tion of their linked functions and a reflection of the syntenybetween the two clusters in M. smegmatis and M. tb (Fig. S4aand b, Supporting Information). These observations were cor-roborated biochemically by pull-down assays using lysates ofM.smegmatis expressing 6xHis-tagged PbpAMs and PbpAM.tb and pu-rified preparations (Fig. S5, Supporting Information) of the cog-nate GST-tagged FhaA (Fig. 2c and d). As the interaction of FHAdomain-containing proteins with their partners is primarilyme-diated by this domain (Nott et al. 2009; Gee et al. 2012), we testedits role in mediating FhaAMs-PbpAMs interaction. A 133 aa frag-ment (Fig. S5 and S6, Supporting Information) containing theFHA domain (Finn et al. 2014) and an additional N-terminal re-gion homologous to the MviN interacting sequence of FhaAM.tb

(Gee et al. 2012) interacted with PbpAMs (Fig. 2e) and was there-fore sufficient tomediate FhaAMs-PbpAMs interaction. To rule outthe possibility of false positives in the pull-down assays, we per-formed western blot analysis with lysates of WT and recombi-nant M. smegmatis expressing C-terminal 6xHis-tagged PbpAMs

and PbpAMtb using anti-His antibody. We found that this an-tibody specifically recognised C-terminal 6xHis-tagged PbpAMs

and PbpAMtb and did not cross-react with any other WT M.smegmatis proteins with sizes similar to our proteins of interest(Fig. S7, Supporting Information) establishing the validity of thisassay. For this interaction to be physiologically relevant, it is nec-essary that the two proteins be present in the same cellular com-partment and should be expressed at the same time. Subcellularlocalisation experiments indicated that both FhaAMs and PbpAMs

localise to the cell wall and cell membrane fractions of M. smeg-matis (Fig. 2f and g). As described by Tiwari et al. (2012), Tiwari,

Soory and Raghunand (2014), Tiwari, Ramakrishnan and Raghu-nand (2015), we ensured equal loading for each fraction in thisexperiment (Fig. S8, Supporting Information). We also observedthat fhaAMs and pbpAMs are both expressed during early and latelogarithmic phases of growth (Fig. 2h and i), but could not bedetected during the stationary phase (Fig. S9, Supporting Infor-mation). These observations are consistent with the relevanceof FhaA-PbpA interaction and the functional importance of theproteins during the active growth phase of M. smegmatis. We at-tempted to test interaction between FhaA and RodA, a septum-PG biosynthetic protein present in the same cluster (Fig. S4aand b), but were unable to obtain transformants of the recom-binant M-PFC plasmids in M. smegmatis possibly due to toxicityresulting from rodA overexpression.

FhaA is required for the stability of PbpA

To understand the functional significance of FhaA-PbpA in-teraction, we examined the levels of ectopically expressedC-terminally c-myc tagged PbpAMs in WT M. smegmatis and�fhaAMs. We observed a drastic reduction in PbpAMs- c-myc lev-els in the cell wall and cell membrane fractions of �fhaAMs incomparison to theWT strain (Fig. 3a). Equal loadingwas ensuredfor each fraction in this experiment (Fig. S10, Supporting Infor-mation) as described by Tiwari et al. (2012), Tiwari, Soory andRaghunand (2014) and Tiwari, Ramakrishnan and Raghunand(2015). The levels of ectopically expressed pbpAMs mRNA weresimilar in both the strains (Fig. 3b), clearly indicating that the re-duction in PbpAMs- c-myc levelswas not a consequence of down-regulation of pbpAMs-c-myc transcription in �fhaAMs. These re-sults reveal a role for FhaA in maintaining PbpA stability, likelyvia their interaction with each other.

PbpA localisation is independent of FhaA inMycobacterium smegmatis and its overexpressionrescues the short cell length phenotype of asubpopulation of �fhaAMs

As the localisation of many cell division proteins is driven byprotein–protein interactions (Sureka et al. 2010; Plocinski et al.2011, 2012), we compared the localisation pattern of a PbpAMs-GFP fusion protein in WT M. smegmatis and �fhaAMs. As re-ported earlier (Plocinski et al. 2011), PbpAMs-GFP was found lo-calised to the mid-cell and/or poles (polar localisation—41%,septal localisation—21%, localisation to both the poles and theseptum—38%, n = 100 cells) in WT cells (Fig. 4a and c). An over-all similar pattern of localisation was observed in �fhaAMs (po-lar localisation—45%, septal localisation—14%, localisation toboth the poles and the septum—41%, n = 100 cells) (Fig. 4band c). This finding implied that PbpAMs localises to its proba-ble sites of action independent of FhaAMs. The observed func-tional link between fhaA and pbpA prompted us to examinethe effect of the varying localisation pattern of PbpAMs-GFP aswell as its overexpression on the cell morphology of �fhaAMs.We found that �fhaAMs: pbpAMs-gfp with only polar or only sep-tal PbpAMs-GFP foci showed a significant reduction in the meancell length in comparison to their WT: pbpAMs-gfp counterparts(Fig. 4d). To our surprise, we found no significant differencein the mean cell length between WT: pbpAMs-gfp and �fhaAMs:pbpAMs-gfp with PbpAMs-GFP foci present in both septum andthe poles (Fig. 4d). These observations suggest a role for PbpAin cell elongation and in rescuing the short length phenotype of�fhaAMs specifically in �fhaAMs cells containing both septal and

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018

Viswanathan et al. 5

Figure 2. Identification and validation of FhaA-PbpA interaction. M-PFC analysis of FhaAMtb - PbpAMtb (a) and FhaAMs - PbpAMs (b) interactions. The pictures showthe growth patterns of M. smegmatis cotransformants on Middlebrook 7H11 agar plates in the presence (+) or absence (–) of 50 μg/mL and 30 μg/mL trimethoprim

(TRIM), respectively. FhaA-PbpA interaction results in the growth of cotransformants-containing pUAB400:fhaA and pUAB300:pbpA but not the control plasmid setsin TRIM-containing media. Negative control (NC)—pUAB300+pUAB400:fhaA, positive control (PC)—pUAB100+pUAB200. Western blots depicting interactions betweenFhaAMtb - PbpAMtb (c) and FhaAMs - PbpAMs (d) using cell envelope fractions of M. smegmatis expressing PbpA-His6 and purified preparations of GST-FhaA. GSH beadswere used to pull down the complex, and the blot was probed with an antibody against His6. (e) Western blot of pull-down reactions using cell envelope fractions ofM.

smegmatis expressing PbpAMsHis6, with a 133 aa fragment of FhaAMs containing the FHA domain. Purified GSTwas used as a negative control for all pull-down reactions.Immunodetection of PbpAMs (f) and FhaAMs (g) in subcellular fractions of recombinantM. smegmatis strains expressing PbpAMs-myc and FhaAMs-myc. All proteins weredetected using an anti c-myc mAb. CW, cell wall fraction; CM, cell membrane fraction; CY, cytoplasmic fraction; CL, cell lysate. (h) RT-PCR-based detection of fhaAMs

and pbpAMs transcripts using RNA isolated from exponentially growing WT M. smegmatis (OD600 nm of 1.0). (i) RT-PCR-based detection of fhaAMs and pbpAMs transcripts

using RNA isolated from WT M. smegmatis at late log phase (OD600 nm of 1.5). RT, reverse transcriptase; ± indicates the presence or absence of reverse transcriptase inthe RT-PCR reaction mixture. The results represent at least two biological replicates.

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018

6 FEMS Microbiology Letters, 2017, Vol. 364, No. 2

Figure 3. Physiological effect of FhaAMs - PbpAMs interaction. (a) Immunodetection of PbpAMs in subcellular fractions of recombinantWTand�fhaAmutantM. smegmatis

strains expressing PbpAMs-myc. All proteins were detected using an anti c-myc mAb. CW, cell wall fraction; CM, cell membrane fraction; CY, cytoplasmic fraction; CL,

cell lysate. (b) RT-PCR-based detection of ectopically transcribed pbpAMs using RNA isolated from exponentially growingWT and �fhaAmutant strains ofM. smegmatis.sigAMs was used as an internal control for normalisation. Results shown are representative of two biological replicates.

polar PbpAMs-GFP foci (Fig. 4d). Additionally, the PbpAMs-GFP fociwere found to be of reduced intensity in �fhaAMs as deducedthrough image analysis (Fig. 4e). This observation is consistentwith our inference that FhaAMs is required for the maintenanceof PbpAMs stability.

�fhaAMs is hypersusceptible to different classes ofantibiotics and displays increased permeability to EtBr

Changes in colony morphology are often associated with alter-ations in mycobacterial envelope composition (Recht et al. 2000;Recht and Kolter 2001; Nguyen, Chinnapapagari and Thomp-son 2005). This led us to hypothesise that �fhaAMs which dis-plays altered colony morphology could show enhanced per-meability and antibiotic susceptibility, by virtue of a compro-mised cell envelope. To test this hypothesis, we determinedthe MIC values for WT M. smegmatis, �fhaAMs and its cog-nate M. smegmatis and M. tb complements against amoxicillin,ampicillin and rifampicin, respectively. As surmised, �fhaAMs

showed enhanced sensitivity, as indicated by reduced MIC val-ues, to all the above antibiotics in comparison to the wildstrain. Complementation of �fhaAMs with fhaAMs as well asfhaAMtb led to a complete or significant restoration of MICs tothe WT levels (Fig. 5a). This result along with our previousobservation of increased susceptibility of TR49 [fhaAMs::Tn] tooxidative stress causing agents (Viswanathan et al. 2015) in-dicated a generalised permeability defect in both these mu-tants. Confirming this, �fhaAMs exhibited increased permeabil-ity and accumulation of EtBr in comparison to theWT and com-plemented strains (Fig. 5b). These observations demonstratea role for fhaA in maintaining the cell envelope integrity ofmycobacteria.

DISCUSSION

FHA domain-containing proteins mediate STPK-associated sig-nalling processes in prokaryotes and higher eukaryotes, signi-fying their role across different domains of life (Durocher andJackson 2002). The M. tb genome encodes six annotated FHAdomain-containing proteins, but their role in regulating STPK-associated functions (Prisic et al. 2010) awaits detailed charac-terisation. FhaA, belonging to this class of proteins, has beenshown to be associated with PG biosynthesis in mycobacteria(Gee et al. 2012). Its interaction with MviN and accumulation ofPG precursors at the poles and septum in an fhaAMs knockdownstrain has led to the speculation that FhaA controls the accu-mulation of PG precursors by regulating MviN activity (Gee et al.2012). However, the molecular details of how FhaA regulates PGbiosynthesis remain to be elucidated.

To address this issue, we began with a forward genetic ap-proach and constructed an fhaA deletion strain of M. smegma-tis. The essentiality of fhaA for in vitromycobacterial growth hasbeen debated thus far (Lamichhane et al. 2003; Sassetti, Boydand Rubin 2003; Zhang et al. 2012), but the successful creationof �fhaAMs in this study implied that this gene is dispensablefor in vitro growth at least inM. smegmatis. A high-density trans-poson mutagenesis study in M. tb revealed that the region infhaA encoding the FHA domain could not sustain transposoninsertions. Only 8 out of the 52 potential transposon insertionsites (TA dinucleotides) of fhaA harboured the transposon andall these were found in or near its domain of unknown function(DUF 3669) (Zhang et al. 2012). This highlights the requirement ofthe FHA domain of the protein for the in vitro growth ofM. tb. Thereasons for the differential essentiality of this gene inM. tb vsM.smegmatis is currently unclear to us and needs further investiga-tion. The cell elongation defect of �fhaAMs was complemented

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018

Viswanathan et al. 7

Figure 4. Cellular localisation of PbpAMs and phenotypic consequences of its overexpression in M. smegmatis. Fluorescence micrographs showing the localisation of a

PbpAMs-GFP fusion protein in WT (a) and �fhaAmutant (b) strains ofM. smegmatis. The arrows indicate foci of the fusion protein. The scale bars are 10 μm. Percentage(c) and mean cell lengths (d) of WT and �fhaA transformants of M. smegmatis showing pole, septal and pole+septal localisation of PbpAMs-GFP foci; N = 100. (e) Meanfluorescence intensities of ectopically expressed PbpAMs-GFP foci in WT and �fhaA strains of M. smegmatis. The histogram represents 100 foci each in 44 WT and 43�fhaA transformants of M. smegmatis, respectively. AU, arbitrary units. The error bars represent SEM, ∗∗∗P < 0.0001; ns, not significant.

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018

8 FEMS Microbiology Letters, 2017, Vol. 364, No. 2

Figure 5. Antibiotic sensitivity and EtBr permeability of the �fhaAmutant strainof M. smegmatis. (a) MIC values of WT and �fhaA strains of M. smegmatis, alongwith the fhaAMs and fhaAMtb complemented strains of the deletion mutant to

different classes of antibiotics. Results shown are representative of two biolog-ical replicates and four technical replicates. (b) Permeability of WT, �fhaA mu-tant and the fhaAMs and fhaAMtb complemented �fhaA M. smegmatis strains as

assessed by an EtBr fluorescence assay. Results shown are representative of atleast two biological replicates. Error bars represent SEM; AU, arbitrary units. ∗P <

0.05 for �fhaAMs vs WT and the complemented strains at the 10, 20 and 30 mintime points. The corresponding values of the WT vs the complemented strains

were not significant.

by both fhaAMs and fhaAMtb confirming their orthology and pro-viding evidence for the novel association of fhaAMtb with a shortcell length phenotype. Our result also established �fhaAMs asa suitable model to study cell division-associated functions offhaAM.tb.

Our finding that FhaA interacts with PbpA and is necessaryformaintaining its stability inM. smegmatis provided a new func-tionality to FhaA as well as an alternate explanation for the ac-cumulation of PG precursors observed earlier in anM. smegmatisfhaA knockdown strain (Gee et al. 2012). We propose that the de-crease in PbpA levels in the absence of FhaA leads to poor cross-linking of PG precursors to the existing PG backbone, resulting inthe accumulation of PG precursors at the poles and septum andcompromised cell elongation. Our observation that a PbpA-GFPfusion protein localises to the poles and septum inM. smegmatisindicating its potential sites of action validated our proposition.The rescue of the shorter cell length defect of a subpopulationof �fhaAMs by pbpAMs-gfp upon its overexpression reinforced ourhypothesis. As the deletion of fhaAMs did not affect M. smegma-tis survival, we speculate that other PG-synthesising enzymesmay compensate for the drastically reduced levels of PbpA inthis strain. Additionally, our study indicated that a 133 aa se-quence containing the FHA domain was sufficient for FhaAMs-PbpAMs interaction, revealing the possible involvement of phos-phorylated threonine residues in this interaction.

Although FhaA appears to stabilise PbpA, the mechanism bywhich it does so requires definition. In Bacillus subtilis, PrsA a pro-tein possessing peptidyl prolyl cis-trans isomerise activity hasbeen shown to be involved in maintaining the stability of PG-synthesising proteins Pbp2A, Pb2B, Pbp3 and Pbp4 by assistingin their folding as part of a quality control process (Hyyrylainenet al. 2010). Similarly, FhaA may help in folding of PbpA therebystabilising this protein, a proposition which needs further inves-tigation. In this context, it is also pertinent that the identity ofthe proteases involved in PbpAMs degradation in the absence ofFhaAMs be established.

Our finding that �fhaAMs showed altered colony morphol-ogy and increased susceptibility to different classes of antibi-otics can be linked to altered cell envelope integrity. Since my-colic acids, the primary components of the mycobacterial outermembrane, are linked to the underlying PG through arabino-galactans (Brennan and Nikaido 1995), a defect in PG synthesiscould conceivably lead to a compromised cell envelope, result-ing in increased permeability of�fhaAMs to the drugs tested. Theobserved increase in the permeability and accumulation of EtBrin �fhaAMs supports this argument.

In conclusion, our study uncovers a novel PbpA-mediatedlink between FhaA and PG biosynthesis in mycobacteria. Thisadds a new facet to the PG-associated function of FhaA andcould lead to an enhanced understanding of regulation of PGsynthesis during mycobacterial cell division.

SUPPLEMENTARY DATA

Supplementary data are available at FEMSLE online.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the assistance of Ms ZebaRizvi in cloning, expression and purification of FhaAMsFHA133.The assistance of Dr. Rajesh Prasad, Mr. Ajay Deepak Verma andDr. Kunal Dayma, CCMB, Hyderabad, India in the EtBr perme-ability assay, Southern blotting and image analysis respectively,is gratefully acknowledged.

FUNDING

This work was supported by a Starting Research Grant from In-stitut Merieux (to TRR) and the Council of Scientific and Indus-trial Research (CSIR), Government of India. GV was supported bya Senior Research Fellowship from the CSIR. SY and SVJ weresupported by fellowships from Institut Merieux.

Conflict of interest. None declared.

REFERENCES

Brennan PJ, Nikaido H. The envelope of mycobacteria. Annu RevBiochem 1995;64:29–63.

Cascioferro A, Delogu G, Colone M et al. PE is a func-tional domain responsible for protein translocation and lo-calization on mycobacterial cell wall. Mol Microbiol 2007;66:1536–47.

Danilchanka O, Mailaender C, Niederweis M. Identification of anovel multidrug efflux pump of Mycobacterium tuberculosis.Antimicrob Agents Ch 2008;52:2503–11.

Durocher D, Jackson SP. The FHA domain. FEBS Lett 2002;513:58–66.

Fedarovich A, Nicholas RA, Davies C. Unusual conformation ofthe SxN motif in the crystal structure of penicillin-binding

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018

Viswanathan et al. 9

protein A from Mycobacterium tuberculosis. J Mol Biol2010;398:54–65.

Finn RD, Bateman A, Clements J et al. Pfam: the protein familiesdatabase. Nucleic Acids Res 2014;42:D222–30.

Gee CL, Papavinasasundaram KG, Blair SR et al. A phosphory-lated pseudokinase complex controls cell wall synthesis inmycobacteria. Sci Signal 2012;5:ra7.

Grundner C, Gay LM, Alber T. Mycobacterium tuberculosis ser-ine/threonine kinases PknB, PknD, PknE, and PknF phospho-rylate multiple FHA domains. Protein Sci 2005;14:1918–21.

Henriques AO, Glaser P, Piggot PJ et al. Control of cell shape andelongation by the rodA gene in Bacillus subtilis. Mol Microbiol1998;28:235–47.

Hyyrylainen HL, Marciniak BC, Dahncke K et al. Penicillin-binding protein folding is dependent on the PrsA peptidyl-prolyl cis-trans isomerase in Bacillus subtilis. Mol Microbiol2010;77:108–27.

Jani C, Eoh H, Lee JJ et al. Regulation of polar peptidoglycanbiosynthesis by Wag31 phosphorylation in mycobacteria.BMC Microbiol 2010;10:327.

KangCM,Abbott DW, Park ST et al.TheMycobacterium tuberculosisserine/threonine kinases PknA and PknB: substrate identifi-cation and regulation of cell shape. Gene Dev 2005;19:1692–704.

Kieser KJ, Rubin EJ. How sisters grow apart: mycobacterialgrowth and division. Nat Rev Microbiol 2014;12:550–62.

Kieser KJ, Boutte CC, Kester JC et al. Phosphorylation of the pep-tidoglycan synthase PonA1 governs the rate of polar elonga-tion in mycobacteria. PLoS Pathog 2015;11:e1005010.

Lamichhane G, Zignol M, Blades NJ et al. A postgenomic methodfor predicting essential genes at subsaturation levels of mu-tagenesis: Application to Mycobacterium tuberculosis. P NatlAcad Sci USA 2003;100:7213–8.

Molle V, Kremer L. Division and cell envelope regulation bySer/Thr phosphorylation: Mycobacterium shows the way.Mol Microbiol 2010;75:1064–77.

Narayan A, Sachdeva P, Sharma K et al. Serine threonine proteinkinases of mycobacterial genus: phylogeny to function. Phys-iol Genomics 2007;29:66–75.

Nguyen L, Chinnapapagari S, Thompson CJ. FbpA-Dependentbiosynthesis of trehalose dimycolate is required for the in-trinsic multidrug resistance, cell wall structure, and colo-nial morphology of Mycobacterium smegmatis. J Bacteriol2005;187:6603–11.

Nott TJ, Kelly G, Stach L et al. An intramolecular switch regulatesphosphoindependent FHA domain interactions in Mycobac-terium tuberculosis. Sci Signal 2009;2:ra12.

Pallen M, Chaudhuri R, Khan A. Bacterial FHA domains: ne-glected players in the phospho-threonine signalling game?Trends Microbiol 2002;10:556–63.

Parikh A, Verma SK, Khan S et al. PknB-mediated phosphoryla-tion of a novel substrate, N-acetylglucosamine-1-phosphateuridyltransferase, modulates its acetyltransferase activity. JMol Biol 2009;386:451–64.

Plocinski P, Arora N, Sarva K et al. Mycobacterium tuberculosisCwsA interacts with CrgA and Wag31, and the CrgA-CwsAcomplex is involved in peptidoglycan synthesis and cellshape determination. J Bacteriol 2012;194:6398–409.

Plocinski P, Ziolkiewicz M, Kiran M et al. Characterizationof CrgA, a new partner of the Mycobacterium tubercu-losis peptidoglycan polymerization complexes. J Bacteriol2011;193:3246–56.

Prisic S, Dankwa S, Schwartz D et al. Extensive phosphoryla-tion with overlapping specificity by Mycobacterium tubercu-losis serine/threonine protein kinases. P Natl Acad Sci USA2010;107:7521–6.

Recht J, Kolter R. Glycopeptidolipid acetylation affects slidingmotility and biofilm formation in Mycobacterium smegmatis.J Bacteriol 2001;183:5718–24.

Recht J, Martinez A, Torello S et al. Genetic analysis ofsliding motility in Mycobacterium smegmatis. J Bacteriol2000;182:4348–51.

Sassetti CM, Boyd DH, Rubin EJ. Genes required for mycobacte-rial growth defined by high density mutagenesis. Mol Micro-biol 2003;48:77–84.

Singh A, Mai D, Kumar A et al. Dissecting virulence path-ways of Mycobacterium tuberculosis through protein-protein association. P Natl Acad Sci USA 2006;103:11346–51.

Snapper SB, Melton RE, Mustafa S et al. Isolation and char-acterization of efficient plasmid transformation mu-tants of Mycobacterium smegmatis. Mol Microbiol 1990;4:1911–9.

Sureka K, Hossain T, Mukherjee P et al. Novel role ofphosphorylation-dependent interaction between FtsZ andFipA in mycobacterial cell division. PLoS One 2010;5:e8590.

Taneja NK, Tyagi JS. Resazurin reduction assays for screen-ing of anti-tubercular compounds against dormant and ac-tively growing Mycobacterium tuberculosis, Mycobacterium bo-vis BCG and Mycobacterium smegmatis. J Antimicrob Chemoth2007;60:288–93.

Thakur M, Chakraborti PK. GTPase activity of mycobacte-rial FtsZ is impaired due to its transphosphorylation bythe eukaryotic-type Ser/Thr kinase, PknA. J Biol Chem2006;281:40107–13.

Thakur M, Chakraborti PK. Ability of PknA, a mycobacterialeukaryotic-type serine/threonine kinase, to transphosphory-late MurD, a ligase involved in the process of peptidoglycanbiosynthesis. Biochem J 2008;415:27–33.

Tiwari B, Kannan N, Vemu L et al. The Mycobacterium tuberculo-sis PE Proteins Rv0285 and Rv1386 modulate innate immu-nity andmediate bacillary survival inmacrophages. PLoS One2012;7:e51686.

Tiwari B, Ramakrishnan UM, Raghunand TR. The Mycobac-terium tuberculosis protein pair PE9 (Rv1088)-PE10 (Rv1089)forms heterodimers and induces macrophage apop-tosis through Toll-like receptor 4. Cell Microbiol 2015;17:1653–69.

Tiwari B, Soory A, Raghunand TR. An immunomodulatory rolefor the Mycobacterium tuberculosis region of difference 1locus proteins PE35 (Rv3872) and PPE68 (Rv3873). FEBS J2014;281:1556–70.

van Kessel JC, Hatfull GF. Recombineering in Mycobacterium tu-berculosis. Nat Methods 2007;4:147–52.

Viswanathan G, Joshi SV, Sridhar A et al. Identifying novel my-cobacterial stress associated genes using a random muta-genesis screen in Mycobacterium smegmatis. Gene 2015;574:20–7.

Viswanathan G, Yadav S, Raghunand TR. Identification of novelloci associated with mycobacterial isoniazid resistance. Tu-berculosis (Edinb) 2016;96:21–6.

Zhang YJ, Ioerger TR, Huttenhower C et al. Global assessment ofgenomic regions required for growth in Mycobacterium tuber-culosis. PLoS Pathog 2012;8:e1002946.

Downloaded from https://academic.oup.com/femsle/article-abstract/364/2/fnw294/2737522by gueston 09 February 2018