Comparative analysis of homologous aminopeptidase PepN from pathogenic and non-pathogenic mycobacteria reveals divergent traits

Mycobacterium tuberculosis (Mtb) secretes proteases and peptidases to subjugate its host. Out of its sixty plus proteases, atleast three are reported to reach host macrophages. In this study, we show that Mtb also delivers a lysyl alanine aminopeptidase, PepN (Rv2467) into host macrophage cytosol. Our comparative in silico analysis shows PepNMtb highly conserved across all pathogenic mycobacteria. Non-pathogenic mycobacteria including M. smegmatis (Msm) also encode pepN. PepN protein levels in both Mtb (pathogenic) and Msm (non-pathogenic) remain uniform across all in vitro growth phases. Despite such tight maintenance of PepNs’ steady state levels, upon supplementation, Mtb alone allows accumulation of any excessive PepN. In contrast, Msm does not. It not only proteolyzes, but also secretes out the excessive PepN, be it native or foreign. Interestingly, while PepNMtb is required for modulating virulence in vivo, PepNMsm is essential for Msm growth in vitro. Despite such essentiality difference, both PepNMtb and PepNMsm harbor almost identical N-terminal M1-type peptidase domains that significantly align in their amino acid sequences and overlap in their secondary structures. Their C-terminal ERAP1_C-like domains however align much more moderately. Our in vitro macrophage-based infection experiments with MtbΔpepN-expressing pepNMsm reveals PepNMsm also retaining the ability to reach host cytosol. Lastly, but notably, we determined the PepNMtb and PepNMsm interactomes and found them to barely coincide. While PepNMtb chiefly interacts with Mtb’s secreted proteins, PepNMsm primarily coimmunoprecipitates with Msm’s housekeeping proteins. Thus, despite high sequence homology and several common properties, our comparative analytical study reveals host-centric traits of pathogenic and bacterial-centric traits of non-pathogenic PepNs.

Introduction Annually, worldwide, atleast a million people die of Tuberculosis (TB) [1]. To establish infection and hijack its host, Mycobacterium tuberculosis (Mtb) injects a battery of arsenal [2][3][4][5]. Mtb's stockpile is predicted to include lipids, proteins, sugars and small molecules. Over the years, though several aspects of the Mtb's biology have been discovered, to this day, only few of Mtb's effectors that manipulate host cellular processes have been identified and their roles determined [2][3][4][5][6]. For example, SapM is a secreted lipid phosphatase that prevents phagosome-lysosome fusion [6]. ESAT-6 is an early secretory antigenic target protein that induces apoptosis [7], inhibits generation of reactive oxygen species [8] and suppresses antigen presentation by MHC1 [9]. ManLAM is a mannose-capped Lipoarabinomannan that also inhibits phagosome-lysosome fusion and T-cell receptor-mediated signaling [10].
Often, bacterial pathogens exploit their proteases and peptidases to target host-specific functions [11,12]. Mtb encodes sixty plus proteases and peptidases [13]. Among them, Zmp1, Msh1 and Rv3668c are known to access host macrophage cytosol. Zmp1 is a Zinc metalloprotease that inactivates inflammasome and arrests phagosome maturation [14]. Msh1 (Rv2672) is a protease that aids the pathogen to utilize host lipids, especially during hypoxic conditions [15]. Rv3668c is a serine protease that modulates inflammatory responses of the host [16].
Thus far, no literature exists on Mtb-encoded peptidases directly accessing host macrophages. However, in in-vitro lab cultures, few Mtb aminopeptidases atleast reach spent media (SM). For example, MapB (Rv2861c), an Iron-binding metallo-L-methionyl aminopeptidase that helps remove L-Methionine from selective nascent Mtb proteins reaches SM [17]. Similarly, PepC (Rv0800), a predicted aminopeptidase reaches SM [4], but its specific functions are yet undetermined.
PepN is also encoded by non-pathogenic mycobacteria including M. smegmatis (Msm). It (MSMEG_4690) also harbors an M1 peptidase domain including the GXMEN and HEXXH motifs (https://pfam.xfam.org/protein/A0R1B3) [19]. However, thus far, no traits of PepN Msm have been reported. Here, using both in silico and in vitro approaches, we compared PepNs from both pathogenic (Mtb, H37Rv) and non-pathogenic (Msm, mc 2 155) mycobacteria to find common and distinct traits. Identification of their distinct traits also helped us predict their possible roles.

Molecular cloning
Employing standard restriction enzymes and/or Gateway cloning system (Thermo Fisher Scientific, USA), we generated the required plasmids (S1 Table). Derivatives of pDONR221 were generated and used as Gateway entry vectors. All expression constructs were derivatives of Tetracycline-(Tet) inducible, Gateway destination plasmid pTetSG [25] (kind gift of Dr. Sarah Fortune). Gateway cloning was performed as per manufacturer's recommendations (Thermo Fisher Scientific, USA). To generate knockout, suicidal plasmid pJM1 (kind gift of Dr. Chris Sassetti) was first modified to insert required multiple cloning sites (pKA1) and then employed as base vector to clone both upstream and downstream flanking regions to pepN. To generate recombinant PepN Mtb for antibody generation, an expression vector pET28a (Merck, USA) [28] was used. To generate pepN Msm ::ssrA, using KAP447, KAP 469 and KAP470, we PCR amplified 1130 bp fragment from 3' end of pepN Msm and cloned the fusion (3'-pepN Msm ::ssrA) into pKA2 to generate pNS38. Plasmid pNS39 is a derivative of pNS38 that lacks the ssrA tag. Plasmids and primers used/generated are listed as S1 and S2 Tables respectively.
Plasmid DNA from DH5α was extracted using pDNA miniprep kit (MDI, India) as per manufacturer recommendations. Msm and Mtb genomic DNA and total RNA were extracted as per recommended protocols [29]. PCRs were performed using high fidelity Phusion or Q5 DNA polymerases (New England Biolabs, USA) in Vapo-protect ProS Mastercycler Systems (Eppendorf, Germany). Amplicons obtained were electrophoresed on 0.8% agarose gel and eluted using Gel extraction kit (MDI). DNA ligations were performed using T4 DNA ligase-Quick Ligation kit (New England Biolabs, USA).

Bioinformatics tools for comparative analysis
Employing CLUSTALW, a multiprotein sequence alignment tool [30], we aligned PepN Mtb from Mtb-H37Rv (UniProt entry L7N655), and PepN Msm from Msm-mc 2 155 (UniProt entry A0R1B3) and determined their percentage identity, similarity and differences. To identify significant alignment hits, we employed BLAST P [31] and SMARTBLAST local alignment search tools and separately queried with PepN Mtb and PepN Msm . To determine possible differences in the PepNs from pathogenic and non-pathogenic mycobacteria, we employed Expresso [32], a T-COFFEE flavor that aligns multiple protein sequences using structural information. To align all query sequences structurally, Expresso compares each queried sequence to its closest protein crystal structures in PDB archive. Narrowing on identified structure(s) as reference, it aligns multiple sequences structurally [32]

Transformation of Msm and Mtb
Msm and Mtb were transformed as per standard protocol [29]. Briefly, using a GenePulser Xcell (Bio-Rad, USA), plasmid DNA (100/300 ng for Msm and Mtb respectively) was used to transform freshly made Msm/Mtb electrocompetent cells. Plasmid DNA and required volume of electrocompetent cells were mixed, then transferred to a fresh, 2 mm, sterile electrocell/ cuvette and pulsed at 2.5kV, 25uF and 1000 Ohms. Electroporated cell mixture was immediately recovered in three ml of Middlebrook 7H9 broth (supplemented with 1X OADC/ADC with 0.2% Glycerol and 0.05% Tween 80) by incubating for 3/24 h (Msm/Mtb respectively) at 37˚C and 200 rpm. Cells were pellet down at RT, 3000 RPM, resuspended in 100 μl of fresh 7H9 broth and plated on 7H11 agar supplemented with ADC/OADC and appropriate antibiotics (when necessary) and transformants selected.

Generation of MtbΔpepN knockout and its complementation
To generate MtbΔpepN, we employed homologous recombination-based gene knockout strategy using SacB as counter selection marker [33]. To avoid polar effect on Rv2466c (pepN's flanking gene), we retained 324 base pairs (bp) from 5' start of Rv2467 ORF and 24 bp at its 3' end. The retained portions harbor neither the peptidase active site domain nor the C-terminal ERAP1-C_like domain. We PCR-amplified one kb flanking regions to pepN using KAP307 & KAP308 (upstream) and KAP309 and KAP336 (downstream), digested them (with EcoRV & SpeI and XhoI & SphI respectively), purified and cloned them into similarly digested pKA1 to obtain pNS22 such that they flank either sides of the Hygromycin (Hyg)/Chloramphenicol resistance cassette. We electroporated two μg of pNS22 into freshly prepared H37Rv electrocompetent cells [29] and selected transformants on 7H11 agar plates containing 1X OADC and Hyg 50 μg/ml. We screened the obtained transformants for sensitivity to sucrose. Those that grew on 10% sucrose turned out to be false positives for double crossover. Using 10 colonies verified for single crossover at the expected locus (no growth on 10% sucrose), we grew them in fresh 7H9 broth + 1X OADC and Hyg 50 ug/ml to an optical density (O.D. at A 600nm ) of 0.1 and spread plated them on 7H11 agar plates containing 1X OADC, Hyg 50 μg/ml and 10% sucrose. We PCR-confirmed few colonies that emerged for loss of sacB region. Then we PCR-confirmed them for double crossover and loss of pepN. We finally confirmed pepN deletion by Southern [34] and western analysis [35].
genomic DNA was loaded onto 0.8% agarose gel and electrophoretically resolved. The gel was washed in autoclaved Milli-Q water, depurinated (0.2 N HCl, 10 min), washed twice with autoclaved Milli-Q water (5 min each), and denatured (1.5 M NaCl, 0.5 M NaOH for 45 min). Gel was rinsed for 10 min in autoclaved Milli-Q water, neutralized for 45 min with 1 M Ammonium acetate, washed for 10 min with autoclaved Milli-Q water and transferred O/N onto HyBond Nylon + membrane by capillary transfer with 10X SSPE buffer. After transfer, genomic DNA was UV cross-linked with energy of 120 mJ/cm 2 (CL-1000 Ultraviolet Crosslinker, UVP, UK), prehybridized in hybridization bottles for 3 h at 42˚C and probed overnight at 42˚C with Digoxigenin (DIG) [8]-labelled probe in DIG Easy Hyb buffer in hybridization oven. The probe amplicon was generated with KAP8 and KAP474 and labelled as per manufacturer's recommendations (Thermo Fisher Scientific, USA). The probed membrane was washed sequentially at 60˚C twice each with 2X SSPE containing 0.1% SDS and 0.5X SSPE containing 0.1% SDS, then rinsed 15 min in washing buffer (0.1 M Maleic acid, 0.15M NaCl, pH 7.5, 0.3% Tween 20 (v/v)), blocked (1 h with blocking buffer diluted to ratio of 1:10 in Maleic acid buffer (0.1 M Maleic acid, 0.15M NaCl, pH 7.5), and incubated for 30 min with anti-DIG antibody at (1:10,000 in blocking buffer). The blot was then washed again 4 times in washing buffer (15 min each), equilibrated with detection buffer (0.1M Tris-HCl, 0.1 M NaCl, pH 9.5) for 5 min and developed with CSPD substrate and chemiluminescent signal monitored on the gel documentation system (BioRad, USA).

Site-directed mutagenesis
To generate the double mutant of pepN Mtb (mpepN Mtb ; GAMEN to GAAAN and HEXXH to AAXXH), we performed site-directed mutagenesis as per Phusion site-directed mutagenesis kit recommendations (Thermo Fisher Scientific, USA). Briefly, using pDONR221 +pepN Mtb as template, KAP225 and KAP226 as primers and Phusion High-Fidelity polymerase (New England Biolabs, USA), we first mutated GAMEN to GAAAN the active site of M1 peptidase. Then, using the mutant clone (pNS30; confirmed by sequencing) and KAP227 and KAP228 as primers, we mutated HEXXH (zinc-binding motif) to AAXXH. We confirmed the double mutant (pNS32) by sequencing and recombined it to pTetSG [25] to obtain pNS35. As per manufacturer's recommendations, we used DpnI to digest off the template pDNA.

Anti-PepN antibody generation
Using Mtb genomic DNA as template and KAP315 and KAP316 primers, pepN Mtb was PCR amplified and cloned as a NdeI/SacI fragment into similarly digested pET28a to obtain 6X-His::pepN (pNS23; S1 Table). C41 DE3 (pLysS) E. coli strain harboring 6X-His:: pepN was induced with 1 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) for 16 h at 37˚C as per standard protocols [28]. The induced culture was spun down at 4˚C and 10,000 RPM. The washed pellet was boiled in 1X Laemmli buffer for 15 min at 95˚C and loaded on 10% SDS-PAGE to evaluate for over-expression. Since, both HisPur Cobalt and Ni-NTA beads retained several contaminating non-specific proteins, we cut the overexpressed band out, eluted proteins and generated polyclonal antibodies to the mixture in rabbits (Link biotech, India). The specificity of the generated antibody (Ab) was verified by western analysis [35] to whole cell protein lysates of Msm and Mtb. The antisera was further purified as in [36] and eluate immediately neutralized with drops of 1 M Tris, pH 7.5 and stored with 0.1% BSA (Bio Basic, Canada) and 0.02% Sodium azide (Sigma-Aldrich, USA) for future use.

RNA extraction and RT-PCR
Approx. 2 x 10 9 Mtb and Msm cells were used for RNA isolation. RNA was isolated by DNA, RNA and Protein purification Kit as per manufacturer (Machery-Nagel, Germany) protocol. Briefly, cell pellets were resuspended in 200 μl of TE (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) with lysozyme (2 mg/ml) and lysed in the supplied buffer RA1 with β-ME (1:1000). The lysates were then passed through supplied purple colored nucleospin columns. The elutes were mixed with 70% ethanol and the mixtures passed through RNA-binding (blue colored) nucleospin columns. The bound RNA was desalted, treated with DNase, washed and eluted as per manufacturer's recommendations. Three μg of extracted RNA was treated with 1 μl of Turbo DNase (Turbo DNA-free Kit-Thermo Fischer Scientific, USA) and genomic DNA contamination eliminated as per manufacturer's protocol. Two μg of the extracted RNA was used to generate cDNA with PrimeScript 1 st strand cDNA synthesis Kit (Clonetech, Takara, Japan) as per manufacturer's recommendations. Equal volume (1 μl) of generated cDNA was used to set up qRT-PCR with 5X HOT FIREPol Evagreen qPCR Mix Plus (SYBR Green; Solis Biodyne, Estonia) on Stratagene mx3005p system (Agilent technologies, USA). Primer pairs KAP411, KAP412 (for Mtb) and KAP413, KAP414 (for Msm) were used for qRT-PCR (S2 Table). The Ct values were normalized with sigA and mysA [29] as controls for Mtb and Msm respectively. The sigA was amplified with primer pair KAP463 and KAP464. The mysA were amplified with primer pair KAP464 and KAP465. The sequence of primers used are listed in S2 Table

Co-immunoprecipitation assay and interactomes identification
Mid-log cultures of Mtb and Msm were pellet down at 4000 RPM for 10 min, pellets washed and lysed in bead-beating buffer (1 mM Tris (pH 6.8), with 0.5 mM EDTA). Bead beating was performed for 10 cycles, 30 s each, with 1 min incubation on ice between each cycle (Biospec Products, USA). The lysate was filtered through 0.22 μM disc filter (MDI, India) and protein estimated using BCA protein estimation kit (Thermo Fisher Scientific, USA). One mg each of Mtb and Msm lysate were separately incubated on a gentle rocker (30 cycles/min), each with 100 μg of purified anti-PepN antibody for 16 h at 4˚C. The protein complexes were pulled down by further incubating the lysates with 75 μl of SureBeads Protein G Magnetic Beads (BioRad, USA) for 2 h at 4˚C. As negative controls, the bacterial lysates were co-incubated with SureBeads alone or SureBeads together with pre-immune sera. Beads were washed in 1X TBST as per manufacturer's protocol and eluted in 100 μl 7.5 M Urea (2 h at 25˚C with frequent gentle mix) and processed for mass spectrometry (MS).
To subtract out the non-specifically bound lysate proteins from PepN Mtb and PepN Msm interactomes, we simultaneously performed co-immunoprecipitation on same lysates with beads alone and beads with pre-immune sera. To eliminate the non-specifically bound proteins further, in Mtb, we also performed similar pulldowns with lysate of pepN KO. To further shortlist PepN Mtb interactome, we also co-immunoprecipitated proteins with lysates of pepN KO complemented with either wild-type PepN Mtb or mPepN Mtb . All co-immunoprecipitations were performed as per manufacturer (BioRad, USA) recommendations and precipitated pellets were again processed by MS to identify PepN interacting proteins. Technical triplicates of each sample were run on Nano-LC Q-Exactive Plus Orbitrap MS (Thermo Fisher Scientific, USA), MS data collected and analyzed to identify potential interactomes to PepN Mtb and PepN Msm . For every sample, only proteins that were common in at least two replicates with more than one unique peptide were unified to generate a single protein list and the quantitative measurements were averaged using geometric mean. The final protein lists thus obtained were used for comparative analyses to finally generate S3 and S4 Tables.
To shortlist PepN Mtb interactome, we first combined protein lists obtained from (a) wildtype Mtb lysate co-immunoprecipitated with (i) no antibody (NoAb Mtb ) & (ii) with preimmune sera (PI Mtb ); and (b) pepN Mtb KO lysate co-immunoprecipitated with anti-PepN antibodies (KOAb). This combined list includes all non-specifically co-immunoprecipitated Mtb and pepN Mtb KO lysate proteins. As expected, there was no PepN in this list. To generate the interactome of PepN Mtb , all proteins from the above combined list (NoAb Mtb + PI Mtb + KOAb) were subtracted out from individual protein lists of (a) wild-type Mtb lysate co-  Table).
Similarly, we first combined the protein lists of (a) wild-type Msm lysate co-immunoprecipitated with (i) no antibody (NoAb Msm ) & (ii) with pre-immune sera (PI Msm ) to generate the non-specifically co-immunoprecipitated Msm lysate proteins. This combined list was then subtracted out from protein list obtained by co-immunoprecipitating wild-type Msm lysate with anti-PepN antibodies (WT Msm Ab) to finally generate a list of 108 Msm proteins that formed the interactome of PepN Msm (S3 Table).
To dissect these lists further, we first analyzed both protein lists for gene ontology (GO)/ biological processes using Uniprot [37], and Mycobrowser [18]. Then, we broadly classified them into different functional groups manually merging similar functions.

MS analysis
Trypsinization and peptides elution was as per manufacturer's recommendations (Thermo Fisher Scientific, USA). Briefly, trypsinized samples was spun down at 10,000 RPM, 4˚C for 20 min to remove debris and undissolved pellet. The clear supernatant was passed thrice through C18 columns. The unbound peptides were washed away with 0.1% FA and bound peptides eluted in 0.1% FA with 50% ACN. The eluates were vacuum dried and subjected to MS. The raw data obtained from mass spectrometry proteomics (for secretion and co-immunoprecipitation) studies has been submitted to public data repositories ProteomeXchange via Massive (http://massive.ucsd.edu/ProteoSAFe/static/massive.jsp) and PRIDE (http://www.ebi.ac.uk/ pride/archive/. The mass spectrometry proteomics data for secretion (Table 1) can be downloaded from https://massive.ucsd.edu/ProteoSAFe/dataset.jsp?task=a5fa419b81a9436b940d 736b4db03648. The data further submitted to PRIDE can be downloaded from ProteomeXchange (http://proteomecentral.proteomexchange.org). ProteomeXchange, MassIVE ID: MSV000081967; ProteomeXchange, PRIDE ID PXD008790.

Trypsinization and peptides processing for MS analysis
Trypsinization and peptides elution was as per manufacturer's recommendations (Thermo Fisher Scientific, USA). Briefly, known amount of protein (BCA, Thermo Fisher Scientific, USA) was resuspended in 50 mM Ammonium carbonate pH 8.0 containing 7.5 M Urea and 10 mM Dithiothreitol (DTT). After 1 h incubation at RT, Iodoacetamide was added to a final of 50 mM, gentle vortexed, and mixture incubated in dark for 1 h at room temperature (RT). Then DTT was added to a final of 35 mM and incubated in dark for 1 h at RT. Urea in the sample was then diluted to 0.5 M with 50 mM Ammonium carbonate containing 1 mM CaCl 2 (pH 7.6). One μg (per 50 μg protein) of MS grade trypsin (Promega, USA) was added and samples incubated at 37˚C for 16 h. Trypsinization was stopped by reducing pH to~3 with formic acid (FA). The samples were lyophilized, resuspended in 0.1% FA (pH~3) and purified using C18 spin columns (Thermo fisher Scientific, USA) as per manufacturer's protocol.
Briefly, before adding samples to C18 columns, columns were spun at 3000 RPM for 1 min and sequentially washed thrice with 50% MS-grade Acetonitrile (ACN), 0.1% FA + 70% ACN and 0.1% FA. Trypsinized samples was spun down at 10,000 RPM, 4˚C for 20 min to remove debris and undissolved pellet. The clear supernatant was passed thrice through C18 columns. The unbound peptides were washed away with 0.1% FA and bound peptides eluted in 0.1% FA with 50% ACN. The eluates were vacuum dried and subjected to MS.

MS for interactome experiments
Nano-LC based reverse phase separation of tryptic peptides. Vacuum-dried tryptic peptide pellets were re-dissolved in 10 to 20 μL of 0.1% Formic acid (FA), vortexed gently and prior to injection centrifuged at 12000 RPM for 15 min. Around 3-5 μL of the sample was bound onto a pre-equilibrated Acclaim PepMap 100, 75 um x 2 cm, Nanoviper, C18 pre-column (Thermo Fisher Scientific, USA) at a flow rate of 4 μL/min. Reverse phase separation of the peptide mixture was performed using the Easy Nano-LC 1200 system (Thermo Fisher Scientific, USA) using a PepMap RSLC C18-Easy spray Analytical column of 75um x 50cm length at a flow rate of 250 to 300 nL/min with a solvent system comprising solvent A (0.1% FA), and solvent B (0.1%FA in 80% ACN, v/v). The column temperature was maintained at 40˚C throughout the run. The gradient conditions were set to achieve 5 to 8% B for 2 min; followed by a linear increase of 8% to 20% B for 150 min, 20% to 40% B for 10 min, 40% to 80% B for 5 min, wash with 80% B for 5mins, 80% to 5% B for 2 min and 5% B for 6 min.

Q-Exactive Plus Orbitrap MS based protein identification. Q-Exactive Plus MS instrument parameters.
Spectral measurements were attained using the Q-Exactive Plus Orbitrap MS platform (Thermo Fisher Scientific, USA) in the positive ion mode with an Electrospray voltage of 2.5 kV, Capillary temperature 300˚C. For MS scan, at 70,000 resolution, the AGC target was set to achieve 3e6, IT-50ms, scan range 350 to 2000 m/z, Resolution-70,000 and for MS2 scan AGC was set to achieve 1e5, IT-100ms, Resolution-17,500. Samples were acquired using the TopN = 15-DDA (Data dependent acquisition) method and MS2 fragmentation was achieved through Higher Energy Collisional Dissociation (HCD) using a NCE (Normalized Collision Energy) value of 27. Singly charged, unassigned and >8 charged species were excluded from acquisition. Dynamic exclusion was set to 30 s, intensity threshold of 1.0e4, peptide match was set to preferred and Exclude isotopes-ON. Mass accuracy during the acquisition was ensured through the lock mass option using the Polysiloxane species (m/z 445.12003). The acquisition parameters were fed into the instrument using the Thermo X-Calibur software version 4.0 through the Tune Plus software interface version 2.8.
Protein identification using the Proteome Discoverer software. To generate protein identities, the raw data (.Raw) files were analyzed using Proteome Discoverer (PD; software version 2.1). Briefly, reference proteome database of Mycobacterium tuberculosis (strain ATCC 25618 / H37Rv) comprising 3,993 protein sequences (UP000001584) and Mycobacterium smegmatis (strain ATCC 700084 / mc 2 155) comprising 6,601 protein sequences (UP000000757) were used to achieve protein identities. Employing the processing and consensus workflow options available within the PD software, the Sequest search engine was used. Search parameters included MS tolerance: 10ppm, MS/MS tolerance: 0.02 Da, Enzyme specificity: Trypsin, Static modification: Carbamidomethylation (Cysteine), Dynamic modification: Methionine Oxidation, N-terminal acetylation, Maximum missed cleavage-2, only those protein entries with a FDR threshold of 0.01 were considered as identified in the current study. Identified proteins were exported for further in silico-based analysis.

MS for PepN secretion experiments
All samples were analyzed by reverse-phase high-pressure liquid chromatography electrospray ionization tandem mass spectrometry using an Ekspert-nanoLC 415 system (Eksigent; Dublin, CA) which is directly connected to an ABSCIEX 5600 Triple-TOF (AB SCIEX; Concord, Canada) mass spectrometer, referred as Triple TOF system.
Reverse Phase -HPLC was performed via a trap and elute configuration using Ekspert-nanoLC 415 systemcolumns (Eksigent); the trap column (200 μm × 0.5 mm) and the analytical column (75 μm × 15 cm) were both manufacturer (Eksigent)-packed with 3 μm ChromXP C-18 (120 Å). Reverse-phase mobile phase consisted of mobile phase A: 2% acetonitrile/98% of 0.1% FA (v/v) in water, and mobile phase B: 98% acetonitrile/2% of 0.1% FA (v/v) in water. All samples were eluted from the analytical column at a flow rate of 250 nL/min using a initial gradient elution of 10% B from 0 to 5 min, transitioned to 40% over 120 min, ramping up to 90% B for 5 min, holding 90% B for 10 min, followed by re-equilibration of 2% B at 10 min with a total run time of 150 min. The analytical column temperature was maintained at 35˚C to decrease retention time drift. The collected raw files spectra were stored in .wiff format. Autocalibration of spectra occurred after acquisition of every sample using dynamic LC-MS and MS/MS acquisitions of 100 fmol β-galactosidase.
The data acquisition mode in DDA experiments was set to obtain a high resolution TOF-MS scan over a mass range 350-1250 m/z, followed by MS/MS scans of 20 ion candidates per cycle with activated rolling collision energy, operating the instrument in high sensitivity mode. The selection criteria for the parent ions included the intensity, where ions had to be greater than 150 cps, with a charge state between +2 to +5, mass tolerance of 50 mDa and were present on a dynamic exclusion list. Once an ion had been fragmented by MS/MS, its mass and isotopes were excluded from further MS/MS fragmentation for 12 s. Collision-induced dissociation was triggered by rolling collision energy. The ion accumulation time was set to 250 ms (MS) and to 70 ms (MS/MS).

Database search
All DDA mass spectrometry files were searched in Protein Pilot software v. 5.0.1 (AB SCIEX) with the Paragon algorithm. For Paragon searches, the following settings were used: Sample type: Identification; Cysteine Alkylation: methyl methanethiosulfonate (MMTS), Digestion: Trypsin; Instrument: TripleTOF5600; Species: Mycobacterium tuberculosis (H37Rv) and Mycobacterium smegmatis (mc 2 155); Search effort: Thorough ID; Results Quality: 0.05. Only peptides with a confidence score of > 0.05 were considered for further analysis. The search was conducted using a through identification effort of a Swiss-Prot database from the UniProt website (www.uniprot.org). False discovery rate analysis was also performed. Carbamidomethylation (C) was used as a fixed modification. The peptide and product ion tolerance of 0.05 Da was used for searches. The output of this search is a .group file and this file contains the following information that is required for targeted data extraction: protein name and Uni-Prot accession, cleaved peptide sequence, modified peptide sequence, relative intensity, precursor charge, unused Protscore, confidence, and decoy result. The parameters used for identification of proteins includes: 1) Threshold of 1% accepted Global False discovery rate (G-FDR) proteins; 2) At least one unique peptide with 95% confidence. The false positive rates of the aforementioned filter criteria were all below 1%, estimated by using an individual reversed (decoy) sequence database.

Both PepN Mtb and PepN Msm primarily localize to host macrophage cytosol
Given PepN Mtb 's structurally similarity to host ERAP1, and PepN Mtb presence in spent media (SM) of Mtb lab cultures [4], we wondered if Mtb delivers PepN Mtb into host macrophages and if so specifically to ER. Interestingly, insilico analysis (LocSigDB) [42] of PepN Mtb also indicated ER homing-like sequences spread across both domains (S2 Fig). Hence, we tested if upon infection into THP-1, a human macrophage-like cell line [38], Mtb delivers PepN Mtb into ER of macrophages (Fig 1). We immunoflouresced Mtb-infected THP-1 with anti-PepN and anti-GRP94 (ER-specific marker) [43] antibodies. To our surprise, most of the secreted PepN Mtb localized away from ER (Fig 1). Our nucleus-specific staining with DAPI also revealed that PepN Mtb does not localize to THP-1's nucleus (Fig 1). Using, Lysotracker Green, we determined that PepN Mtb does not localize to lysosomes as well (S3 Fig). This absence of localization to ER, nucleus and lysosomes and the obtained localizing pattern (Fig 1) of PepN Mtb indicate that most of the pathogen secreted PepN Mtb localizes to THP-1 cytosol (Fig 1).

PepN Msm is necessary for in vitro growth of Msm
We speculated earlier that differences in PepN Mtb and PepN Msm amino acids sequence (10. 6084/m9.figshare.7873274) and secondary structures (S1 Fig) might influence their localization in host cells. However, given their similar localization patterns (Fig 1), we wondered if any of the above said differences influence PepN Mtb and PepN Msm roles in their cognate parent mycobacterial environment.
To test this, employing standard homologous recombination-based gene knockout (KO) strategy [33], we set out to generate MtbΔpepN and MsmΔpepN. We effortlessly could knock out pepN from Mtb (H37Rv; S5 Fig) and hence could readily complement it (MtbΔpepN) with pepN Msm and determine PepN Msm localization in THP-1 (Fig 1). Besides, MtbΔpepN grew similar to WT-Mtb confirming its non-essentiality in vitro. In contrast, despite several attempts, we failed to generate MsmΔpepN. Our attempts to conditionally KO the genomic copy (by episomally expressing pepN Msm on an inducible promoter) also failed. To atleast obtain a knockdown phenotype, we adopted CRISPRi [44]. Though  To test this, to 3' end of pepN Msm , just before its stop codon, we genetically fused ssrA, the canonical protein degradation signal [45]. We expected that SsrA-mediated depletion of PepN Msm ::SsrA might generate atleast a knockdown phenotype. Despite repeated attempts, to our surprise, no transformants emerged even after incubation of plates for three weeks. When we used a similar construct lacking just the ssrA tag, transformants emerged within 6-7 d (Fig  2A). During one of several such attempts, in the 4th week, only one colony, a potential revertant (to PepN Msm knock down) emerged that grew very slowly (than WT Msm; Fig 2B). This colony did not harbor any compensatory mutation in its entire pepN Msm ::ssrA length as its amino acid sequence was identical to WT PepN. Western analysis also showed PepN Msm ::ssrA fusion protein moving at the expected molecular weight (Fig 2C). Though we are yet to determine the location of the compensatory mutation, these above observations clearly suggest that PepN Msm is possibly necessary for in vitro growth of Msm. In contrast, Mtb's PepN is unnecessary for Mtb growth in vitro. Typically, PepN Mtb migrates at~100 and~90-95 kDa (Fig 3 and S6B Fig) while PepN Msm migrates at~100, 90-95, 60-63 and 35-40 kDa (Fig 3 and S6A Fig). In both, the 100-kDa form constitutes the major pool.

Msm and not Mtb proteolyzes excess of its own PepN
Uniform steady state levels of both PepNs implies that they either have extended half-lives or possible fine-tuning by the bacteria, for unknown reasons. To test this, we deliberately altered the native steady state PepN Mtb and PepN Msm levels by overexpressing them separately in both WT bacteria. Equal amount of total proteins were loaded (S7A and S7B Fig) and PepN levels evaluated (Fig 3). Mtb well tolerated increased levels of PepN Mtb (Fig 3A-   harboring plasmid and vector controls expressed PepN levels similar to their WT controls ( Fig  3A and 3B).
We then tested, if WT-Msm discriminates between active and mutant PepN pools. Interestingly, Msm proteolyzes both WT PepN Mtb and mutant PepN Mtb (mPepN Mtb ; peptidase active site GAMEN mutated to GAAAN and zinc-binding motif HEXXH mutated to AAXXH) with equal efficiency (Fig 3B, compare lanes 9 and 11). We speculate that such proteolytic activity of Msm towards its own and foreign 100 and 95-kDa PepN forms possibly generates the 60-63 and 35-40 kDa minor forms (S6A Fig).

Msm and not Mtb secretes portion of its excess PepN into spent media
Mass spectrometric (MS) analysis of spent media (SM) of Mtb lab cultures had established earlier that Mtb secretes PepN [4]. Using similar approach, we determined that Msm too secretes PepN ( Contrary to our speculation, Mtb did not secrete any excess PepN Mtb into SM (Table 1). Opposingly, Msm when overexpressing its own PepN, secreted it atleast 3-fold more (see number of peptides; Table 1) suggesting that it indeed maintains steady state levels of PepN Msm by not only proteolyzing but also by secreting excess PepN Msm . As a control, we monitored and found EsxB (a well-established secreted protein in mycobacteria) levels uninfluenced by PepN overexpression across all strains evaluated (Table 1).

PepN Mtb and PepN Msm interactomes are markedly different
As an essential aminopeptidase (Fig 2), PepN Msm might interact and modulate several Msm proteins required for Msm growth and survival. To test this, we co-immunoprecipitated PepN Msm together with its interacting partners and identified them by tandem MS. We employed appropriate controls (S9 Fig) and subtracted out non-specific proteins (see Materials and Methods), we identified 107 PepN Msm -specific interacting partner proteins (S3 Table). Among them, 60% are enzymes involved in diverse intermediary metabolism and/or respiration activities, the so-called workhorses of cellular growth and survival (Fig 4A and 4B; S3 Table). Thirty percent of these enzymes are oxidoreductases. Another 30% of the enzymes are a combination of synthases, synthetases, transferases & hydrolases. Eighteen percent of the interactome constituted ATPases or ATP-binding proteins (Fig 4B, S3 Table). The remaining potential interactors included (i) ribosomal proteins, translation initiation factors, Ribonuclease E all necessary for survival and (ii) ABC transporters, SecA2, SecY, signal peptidase 1, Trigger factor and Tat pathway signal sequence domain protein all involved in proteins transport across inner membrane. Importantly, PepN Msm also interacted with Pup deamidase/depupylase and proteasomal core subunit beta, both necessary components of protein degradation machinery (S3 Table). Thus, >90% of PepN Msm interactome constitute several important housekeeping proteins that may have significant role in Msm survival and growth in vitro.
Since PepN Mtb reaches cytosol (Fig 1) and remains nonessential for in vitro growth (S5 Fig), we speculated that it might interact with and may modulate levels of several host and pathogen proteins required in vivo. We attempted to co-immunoprecipitate interacting pathogen and host proteins from Mtb-infected THP-1 lysates (see materials and methods). While we could easily enrich for PepN and several host proteins (details beyond the scope of this manuscript), MS was not sensitive enough to directly detect pathogen proteins present in THP-1 environment.
Consequently, we co-immunoprecipitated PepN Mtb and its partners from Mtb lysate, applied essential controls and identified 114 PepN Mtb -specific partner proteins (S4 Table). Surprisingly, one third of PepN Mtb interactome constitute secreted proteins of Mtb. Around 55% Comparative analyses of mycobacterial PepNs of its interactome constituted proteins nonessential [22] for in vitro growth of Mtb (S4 Table). Only 40% of the total interactome constitute proteins involved in intermediary metabolism/ respiration functions (Fig 4A). Of those enzymes though 1/3 rd constitute oxidoreductases ( Fig  4B), 50% of them are again non-essential for Mtb growth in vitro. Similarly, around 50% of interacting hydrolases and transferases are nonessential. Around 15% of PepN Mtb interactome uniquely constitute proteins of lipid metabolism (Fig 4A; S4 Table), 85% of which are again nonessential. Additionally, around 10% of PepN Mtb interactome contain proteins involved in detoxification/adaptation processes (Fig 4A), most of them again nonessential for Mtb growth in vitro (S4 Table). Most hypotheticals/conserved proteins with no functions (8 of 10), DNA repair/replication proteins (2 of 3) and cell wall and cell processes-mediating proteins (6 of 9) were again found to be nonessential (Fig 4A; S4 Table). Thus, more than 55% of PepN Mtb interactome constitute proteins that are not essential to Mtb growth in vitro. Notably, only eight proteins among the PepN Mtb and PepN Msm interactomes are common (S3 and S4 Tables). However, both interactomes constituted proteins of various functional groups [18] ( Fig 4A).
To eliminate any concerns that the final interactome lists do not contain proteins in random, we queried both interactomes for interacting networks (using STRING-an interaction network database) [46]. We then imported these networks into Cytoscape [47], assigned colors to each functional group and manually arranged them (10.6084/m9.figshare.7873484). Around 80-83% of PepN Mtb and PepN Msm interactome proteins fell into these interacting networks (10.6084/m9.figshare.7873484; 10.6084/m9.figshare.7873487), validating that our co-immunoprecipitates are highly relevant. Finally, based on protein abundance (eMPAI values of > or < 0.75), we also grouped the interactome proteins and fetched known interactors (not present in interactome lists) from STRING [46]. Our fetched primary protein interactors and their second neighbors differed between PepN Mtb and PepN Msm again suggesting their functional diversity (10.6084/m9.figshare.7873487).

Discussion
Our search for potential Mtb effectors among its aminopeptidases pool led us to PepN. Since it is (i) reported to be secreted into SM [4]; (ii) not essential for Mtb's in vitro growth [22], and (iii) essential for Mtb's growth in vivo [24], we speculated that this might be a potential effector in hosts. Our in vitro macrophage infection studies with Mtb showed that PepN not only secretes into SM [4] but also reaches macrophages (Fig 1). Presence of (i) a C-terminal ERAP1_C-like domain [19] that resembles ERAP1, an essential host aminopeptidase of ER [39] and (ii) identification of ER-homing like signals along its length (S2 Fig) furthered us to hypothesize that PepN not only enters macrophages but also reaches ER. However, surprisingly, our THP-1 infection experiments showed that PepN barely co-localizes with GRP94, an ER-specific marker. (Fig 1). Further evaluation also indicated that PepN fails to localize with THP-1's nuclei (Fig 1) and lysosomes (S3 Fig). Thus, our in vitro infection experiments with THP-1 indicate that PepN is largely localized to host cell cytosol. Himar 1 insertion into pepN Mtb showed that mutated PepN led to attenuation of Mtb-mediated virulence in mice [24]. In contrast, a Tn5370 insertion into pepN facilitated Mtb to become hypervirulent in BALB-C mice [23]. It is unclear at this point as to what probably is the function of PepN Mtb post its delivery into host macrophages.
Thus, if PepN Mtb has evolved to play a major role in vivo [23,24], we wondered what role PepN Msm has in store. Msm not only expresses pepN, but also maintain uniform steady state levels across all growth stages in vitro (S6 Fig). Our failure to generate an MsmΔpepN indicates PepN's essential role during Msm growth in vitro. While we also confirmed its necessity for in vitro growth by fusing a degradation tag, SsrA (Fig 2A and 2B), emergence of one slow-growing revertant after THREE weeks on selection further suggested PepN Msm role towards Msm survival. Uniform levels of PepN Msm indicates possible regulation. Interestingly, failure to even knockdown PepN Msm protein levels by CRISPRi-Cas9 [44] (0.6084/m9.figshare.7873469) reinforces such regulation. Our overexpression studies also indicate that Msm senses excess levels of PepN and proteolyzes it (Fig 3B). Interestingly, Msm proteolyzes both native and foreign PepN with equal efficiency (Fig 3B). It seems to sense quantity over quality as it proteolyzes the active site and zinc-binding motif double mutant, mPepN Mtb with similar efficiency ( Fig  3B).
One wonders why proteolysis instead of transcriptional or post-translational modifications is the way to fine-tune protein level of an aminopeptidase. Most M1 members play pivotal roles in survival, cell maintenance, growth and development, virulence and pathogenesis [11]. They normally cleave proteins at their lysines, alanines, arginines and leucines [18]. Our coimmunoprecipitation studies showed more than 90% of PepN Msm partners to be housekeeping proteins (S3 Table). Perhaps managing their cellular abundance requires active role of several proteases/peptidases including PepN. Thus, expressing altered PepN levels might have global consequences including death and hence proteolysis of excess PepN becomes a necessary step. We carved out gel pieces at the 60-70 and 30-50 kDa positions, performed MS/MS and detected peptides unique to PepN Msm minor forms (60-63 and 35-40 kDa) confirming that those signals in western are indeed specific to PepN. To also ensure that no partially proteolysed (the 60-63 and 35-40 kDa forms) PepN remain active in the cell, perhaps Msm secretes the excess PepN out too (Table 1). Interestingly, MS of both 60-63 and 35-40 kDa forms detected peptides that span the entire length of PepN Msm indicating that they are indeed proteolytic products of full length PepN Msm and not its truncated forms.
In contrast to Msm, Mtb tolerates its own PepN and selectively proteolysis the non-native kind (Fig 3A). Additionally, it does not secrete out any excess PepN it continues to tolerate (Table 1). If PepN Mtb were to have a specific in vivo role in host cytosol, this control of secretion makes sense, as it needs to modulate/proteolyse host and pathogen protein levels only to the extent necessary. Interestingly, when we monitored PepN Mtb localization especially of THP-1 infected with ΔpepN expressing PepN Mtb (in trans), despite higher levels of induction (~96 h; S7C Fig), we did not find excess PepN Mtb signals (fluorescence intensity). Its structure resemblance to ERAP1 (10.6084/m9.figshare.7873451) and localization to cytosol (Fig 1) lets us speculate that it modulates members of its pathogen interactome that reach the host. Almost one third of PepN Mtb interacting partners are Mtb secreted proteins. Given the aminopeptidase role, and it ability to modulate virulence [23,24], it is tempting to hypothesize that PepN Mtb may cleave pathogen and host proteins in host macrophages to regulate virulence levels. We are currently performing experiments to evaluate these possibilities.
Thus, our comparative analyses provides ample insights into common and possible opposing roles for PepN Mtb and PepN Msm . Their divergent interactomes (S3 and S4 Tables) to 78% identity (10.6084/m9.figshare.7873274) sheds light on their potential specific roles. Only eight proteins were common to both interactomes (S3 and S4 Tables). Though the remaining 106 Mtb and 99 Msm proteins fell into common functional groups (Fig 4A and 4B), proteins constituting each category were strikingly different (Fig 4; S3 and S4 Tables). For eg. one third of PepN Mtb interactome contained secreted proteins potentially involved in lipid metabolism, adaptation to stress, detoxification, and intermediary metabolism, all essential for pathogen's survival in the host. Unlike PepN Mtb interactome, PepN Msm interactome lacked any proteins involved in lipid metabolism (S3 and S4 Tables).
In contrast, one fifth of the PepN Msm interactome consist of oxidoreductases that help in redox maintenance (S3 Table). Around 10% of the PepN Msm interactome consists of ABC transporters or proteins including SecA2, SecY, signal peptidase and Tat pathway signal sequence domain protein all necessary for protein translocation across inner membrane (S3 Table). While SecA2 might play an accessory role in protein translocation, cell cannot survive without SecY [48]. Perhaps PepN Msm plays a role in controlling SecY levels. Additionally, around 15% of the interactome constitutes proteins involved in DNA repair/transcription/ translation/protein degradation including DNA polymerase III γ-subunit, DNA partitioning proteins, ribosomal proteins, transcription factors and pup deamidase/depupylase (S3 Table) all again perhaps performing essential functions in Msm. The PepN Msm co-precipitant also included a transglycosylase that has important functions to play in cell wall integrity [49]. Interestingly, PepN Msm interacts with trigger factor (tig) that interacts with several important players including SecA1, a major player of protein translocation [50] (S3 Table and 10.6084/ m9.figshare.7873484; 10.6084/m9.figshare.7873487). Tig is a chaperone that binds to nascent proteins to keep them in a translocation compatible confirmation [51]. Such interactions again reinforce the necessity of PepN Msm for in vitro growth.
Though, our co-immunoprecipitation studies shed light on possible role of these PepNs, identifying their cognate substrate proteins, evaluating when and how their levels get modulated might be necessary to determine the exact functions of PepN-like aminopeptidases. While this requires the comparative quantitative proteomic analysis between WT Mtb and Msm with their cognate ΔpepNs, proteolysis by ATP-independent aminopeptidases are generally considered distal to ATP-dependent steps of proteolysis [52]. This could mean that the potential PepN substrates in both Mtb and Msm backgrounds could be first proteolyzed by other ATP utilizing proteases into peptides before they are handed over to PepNs for further proteolysis. Since PepN Msm is necessary for in vitro growth, MsmΔpepN could not be generated and hence the comparative quantitative proteomics could not be performed in Msm. Given the PepN Mtb redundancy (in Mtb growth in vitro), we are currently exploring use of quantitative proteomics to determine what host proteins does WT Mtb and MtbΔpepN influence. Perhaps this would shed light on understanding how the pathogen modulates host cellular environment. In summary, we show that despite high homology, the PepN aminopeptidases of the slow-growing pathogenic and fast-growing non-pathogenic mycobacteria have evolved to carry divergent traits that define their host and/or pathogen-specific functions. To the best of our knowledge, there is no report thus far reporting on such divergent traits in any bacterial aminopeptidase.
Supporting information S1  [37] IDs obtained from MS analysis of PepN Msm interactome. GO (Gene Ontology)-Molecular function: Potential biological process/ function (as listed in UniProt, [37] to genes listed in column one. Functional Group: Genes in column one were searched in Mycobrowser [18] to identify the major functional group to which they belong. All genes belonging to intermediary metabolism and/or respiration are further sub-grouped in different categories of enzymes.  [37] IDs obtained from MS analysis of PepN Mtb interactome. GO (Gene Ontology)-Molecular function: Potential biological process/function as listed in UniProt [37] to genes listed in column one. Secretion: Yes-Proteins corresponding to genes in column one that are secreted into spent media; No-Proteins corresponding to genes in column one that are NOT secreted into spent media; Essentiality: Primarily corresponds to whether the genes in column 1 are essential (Yes) or not essential (No) for Mtb growth in vitro; when indicated (B) (next to Yes)-essential both in vitro and in vivo. Functional Group: Genes in column one were searched in Mycobrowser [18] to identify the major functional group to which they belong. All genes belonging to intermediary metabolism and/ or respiration are further sub-grouped in different categories of enzymes. Step 1: Rectangular dotted (dull grey) boxes indicate PCR-amplified upstream and downstream regions of pepN that were cloned into the suicidal vector pKA1 to obtain pNS22. 'X' drawing indicates potential recombination occurring regions between pepN Mtb (in the genome) and regions of pNS22. P1 to P9 -primers used for confirmation/analyses. Single line arrows below/above primers indicate forward and reverse directions. Broad arrows in pNS22 and genome indicate open reading frames.
Step 2: schematics of single crossover that occurred between the upstream region to pepN Mtb and pNS22 upstream fragment. Upon single crossover, the entire pNS22 is recombined into the pepN Mtb locus.
Step 3: schematics of pepN Mtb deletion that occurred as a result of the second recombination between downstream region to pepN Mtb and pNS22 downstream fragment. This led to pepN Mtb being replaced by the Hygromycin resistance cassette flanked by unidirectional loxP sites.
Step 4: schematics of the pepN Mtb locus lacking both pepN Mtb and Hygromycin resistance cassette. The resistance cassette was removed by use of pCre-Zeo encoded Cre-recombinase. colonies (from (e)) were pelleted down, lysed by bead beating, total proteins boiled in 1X Laemmli's buffer and loaded on a 10% SDS-PAGE. 10 fold more total protein was loaded in the two ΔpepN Mtb colonies lanes. Total protein of E. coli C41(DE3) harboring pNS23 (S1 Table; expressing pepN Mtb ) was used as positive control. Midlog culture of Msm (mc 2 155) were pelleted down, lysed by bead beating, total proteins boiled in 1X Laemmli's buffer and loaded on a 10% SDS-PAGE to detect for the major and minor forms of PepN Msm that migrate in the gel. The nitrocellulose membrane with transferred proteins was developed with Anti-PepN (1:2500) as primary and anti-rabbit Goat IgG as secondary (1:10000). M-protein marker. After resolving, the gel was processed as per protocol (see materials and methods) and probed with DIG-labelled amplicon of 421 bp (amplified with KAP8 and KAP474 as indicated in (A)) to detect specific size fragments (as mentioned to the right of the Hybond Nylon + membrane image). The signal in lanes 8 and 9 indicate the correct size of the probe amplicon. A standard measuring ruler (right of blot) was used to measure the distance the marker fragments (unlabeled) had resolved for accurate estimation of signal mobility. The signal was obtained by using CSPD substrate and chemiluminescent signal generated was monitored and recorded on the gel documentation system (BioRad, USA). The probe location is as depicted.  (Table 1), exponential phase lab cultures of WT Msm (mc 2 155) and Mtb (H37Rv) were grown, cell pellets collected, lysed by beat beating, lysate filtered twice and total proteins estimated by BCA kit (Thermo Fisher Scientific, USA). The spent media were also filtered twice, TCA precipitated and total protein estimated. Equal amount of lysates (left blots) and 10-fold higher amount of culture filtrate (CF) proteins (right blots-CF) were loaded onto 10% SDS-PAGE, proteins resolved and westerns performed. Blots were developed with specific anti-PepN antibody (α PepN blots) and Hsp65-specific antibody (Abcam, UK). Hsp65 (Rv0440 in Mtb and MSMEG_0880 in Msm) is used as lysis control. Anti-PepN Mtb antibody (1:2500) and anti-Rabbit IgG Goat secondary antibody (1:10000) were used for detection of PepN. Anti-Hsp65 antibody (1:2500) and anti-mouse IgG Goat secondary antibody (1:10000) were used for detection of Hsp65. White tiny bands in each blot indicate protein markers (kDa). These are representative blots of three independent experiments and their biological duplicates.