Bacterial strains and growth conditions

Mycobacterium tuberculosis, H37Rv (Mtb) was grown in vitro at 37°C in Middlebrook 7H9 broth (Thermo Fisher Scientific, USA) or 7H11 Middlebrook agar (Thermo Fisher Scientific, USA) supplemented with 10% Oleic acid-Bovine Albumin fraction V-Dextrose-Catalase (1X OADC), 0.2% (for broth)/0.5% (for agar) glycerol and 0.05% Tween 80. Mycobacterium smegmatis, mc²155 (Msm) was grown in vitro at 37°C in Middlebrook 7H9 broth or 7H11 Middlebrook agar supplemented with 10% Bovine Albumin fraction V-Dextrose-Catalase (1X ADC), 0.2% (for broth)/0.5% (for agar) glycerol and 0.05% Tween 80 [25].

Escherichia coli (E. coli) DH5α (Thermo Fisher Scientific, USA) was used for routine cloning and grown in Luria-Bertani (LB) broth/agar-based medium (Thermo Fisher Scientific, USA) at 37°C [26]. For recombinant protein expression, E. coli C41(DE3) was used [27]. For Gateway cloning, donor and destination vectors were maintained in E. coli DB3.1 (Thermo Fisher Scientific, USA).

Where necessary, to maintain plasmids, we used the following final concentrations of antibiotics: For (i) Mtb and Msm–Hygromycin, 50 μg/ml and Kanamycin, 25 μg/ml; and (ii) E. coli–Chloramphenicol, 25 μg/ml, Hygromycin, 150 μg/ml, and Kanamycin, 100 μg/ml.

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—mc²155 (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]. We aligned PepN_Mtb (UniProt entry L7N655), its orthologues viz. M. bovis (UniProt entry Q7TYI7; NCBI Reference Sequence: WP_047709652.1), M. caprae (NCBI Reference Sequence: WP_075744548.1), M. africanum (NCBI Reference Sequence: WP_031669696.1), M. canetti (NCBI Reference Sequence: WP_015303505.1), M. leprae (UniProt entry Q9CBX9), all from Mtb complex and finally PepN_Msm (UniProt entry A0R1B3). ‘Good’, ‘average’ and ‘bad’ are algorithm outputs displayed by Expresso to indicate high, medium and poor levels of structure-based sequence homology.

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].

Using RvΔpepN325-2562 bp (referred in the manuscript as RvΔpepN) cells, we complemented them with either full length PepN_Mtb, or its active site mutant (mPepN_Mtb) or full length PepN_Msm, all expressing under Tet-inducible promoter. We PCR-amplified full length pepN_Mtb with KAP8 and KAP11. Similarly, we PCR-amplified full length pepN_Msm with KAP317 and KAP319. To add attB1 and attB2 regions to their 5’ and 3’ end, we PCR-amplified the first round Gateway amplicons with KAP5 and KAP6. We recombined the amplicons separately with pDONR221 to obtain pDONR221+pepN_Mtb (pNS24) and pDONR221+pepN_Msm (pNS28). We then recombined each with pTetSG [25] to obtain pNS25 (pTetSG+pepN_Mtb) and pNS29 (pTetSG+ pepN_Msm).

Southern analysis

Southern was performed as per standard protocols [34]. Briefly, around 4 μg of genomic DNA of WT Mtb and MtbΔpepN was separately digested O/N with PvuI and NotI. The digested 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/cm² (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 10⁹ 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) wild-type Mtb lysate co-immunoprecipitated with (i) no antibody (NoAb_Mtb) & (ii) with pre-immune 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-immunoprecipitated with anti-PepN antibodies (WT_MtbAb) and (b) lysates of pepN_Mtb KO complimented with either pepN_Mtb or mpepN_Mtb, (KOC_wtAb and KOC_mAb respectively) both co-immunoprecipitated with anti-PepN antibodies. The proteins thus finally shortlisted included those found common to (i) WT_MtbAb & KOC_wtAb & KOC_mAb (36 proteins); (ii) KOC_wtAb and KOC_mAb (31 proteins); (iii) WT_MtbAb & KOC_mAb (20 proteins) and (iv) unique to KOC_mAb. Thus, PepN_Mtb interactome finally constitutes 115 Mtb proteins (S4 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_MsmAb) 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.

Final interactome lists (S3 and S4 Tables) were analyzed for gene ontology (GO)/biological processes using Uniprot [37], and Mycobrowser [18].

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=a5fa419b81a9436b940d736b4db03648. The data further submitted to PRIDE can be downloaded from ProteomeXchange (http://proteomecentral.proteomexchange.org). ProteomeXchange, MassIVE ID: MSV000081967; ProteomeXchange, PRIDE ID PXD008790.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Upon overexpression of PepN, Msm secretes excess PepN into spent media.

https://doi.org/10.1371/journal.pone.0215123.t001

The data obtained from interactome proteomics (co-immunoprecipitation experiments) can be downloaded from ProteomeXchange (http://proteomecentral.proteomexchange.org). ProteomeXchange, PRIDE ID PXD009164.

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 CaCl₂ (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

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.

Mass spectra and tandem mass spectra were recorded in positive-ion and “high-sensitivity” mode with a resolution of ~35,000 full-width half-maximum. Peptides were injected into the mass spectrometer using 10 μm SilicaTip electrospray PicoTip emitter (New Objective Cat. No. FS360-20-10-N-5-C7-CT), and the ion source was operated with the following parameters: ISVF = 2100; GS1 = 20; CUR = 25.

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 (mc²155); 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 UniProt 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.

Mtb-mediated THP-1 infections and immunofluorescence

THP-1 (from Cell repository, NCCS, Pune, India; authenticated (on 25^th Nov 2016) by STR profiling at Lifecode Technologies Pvt. Ltd., INDIA; also verified by multiplex-PCR for no cross contamination with cell lines derived from Chinese hamster; grivet monkey, rat and mouse) were in vitro-cultured in RPMI-1640 medium containing 2 mM L-Glutamine, 10 mM HEPES (Thermo Fisher Scientific, USA), and supplemented with 10% fetal bovine serum (FBS), 0.05 mM β-Mercaptoethanol and 100 U/mL Penicillin-Streptomycin (all from Thermo Fisher Scientific, USA). THP-1 cells were seeded at 1x10⁵/well in 24-well plates on glass cover slip and differentiated before infection for 3 days at 37°C, 5% CO₂ by adding 10 ng/ml Phorbol-12-myristate-13-acetate (PMA) (Sigma-Aldrich, USA). Differentiated-THP1 were used for infection and for immunofluorescence microscopic analysis.

Mtb-mediated infections of THP-1 were performed as described earlier [38]. Briefly, in vitro grown Mtb (~0.3 OD (A_600nm)) were pellet down, washed in RPM1-1640 (having 2 mM L-Glutamine, 10 mM HEPES and 10% FBS) and filtered through 5μm syringe filters. Known number of Mtb cells were then added to differentiated THP-1 at an MOI of 1:10 and incubated for 4 h at 37°C, and 5% CO₂. The uninfected controls and infected cells were washed thrice with 1X PBS containing Amikacin (200 μg/ml) and fresh RPM1-1640 (having 2 mM L-Glutamine, 10 mM HEPES, 10% FBS and 100U/ml of Penicillin-Streptomycin) added. Five days post-infection, cells were washed thrice with 1X PBS and fixed in 4% Para-formaldehyde. Before fixing, where necessary, Lysotracker green DND-26 (Thermo Fisher Scientific USA; final concentration: 50 nmol) was added to media and incubated for 1 h. The cells were washed thrice with 1X PBS and fixed in 4% Para-formaldehyde. Cells were blocked for 1 h with 1 X PBSAT buffer (1X PBS, pH 7.4 with 1% BSA, 0.5% Tween 20) and incubated for 1 h with primary antibody (anti-GRP94 antibody (Source–Rat; Abcam, USA) and anti-PepN antibody (This study; Source–Rabbit; both at 1:250 dilution in 1X PBSAT). The coverslips were washed thrice with 1X PBSAT and incubated for 1 h with appropriate secondary antibodies (Goat anti Rat for GRP94 and Goat anti-Rabbit for pepN) conjugated to either Alexa Flour 488 (1:500; for detection of GRP94) or Alexa Fluor 647 (1:500; for detection of PepN) (Southern Biotech, USA), washed thrice and mounted on slides with mounting media containing DAPI (Molecular Probes, Thermo Fisher Scientific, USA). Samples were then visualized under a 60X objective on a FV1000 Olympus Confocal microscope (Tokyo, Japan). The accompanying Olympus FV10-ASW version 2.01.03.10 software was used for image analysis.

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).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Both PepN_Mtb and PepN_Msm efficiently localize to host macrophage cytosol.

Immunofluorescence-based localization analysis of PepN_Mtb in THP-1 infected with either WT Mtb (top two panel rows); MtbΔpepN (middle (third) panel row) or MtbΔpepN expressing pepN_Msm (bottom two panel rows). Panel columns: First: DAPI to track nuclei (blue); Second: virulent Mtb expressing mCherry (pink)—to locate infected THP-1; Third: ER-specific marker (green); Fourth: PepN (red); Final: merger of first four panels (of each row); Scale: 10 μm. Atleast 300 infected macrophages were visualized for confirming consistency of observed results. The immunofluorescence data is a representation of three independent experiments. Each independent experiment had biological duplicates.

https://doi.org/10.1371/journal.pone.0215123.g001

As expected, we did not see any PepN-specific signal in THP-1 infected with MtbΔpepN (Fig 1). Since, PepN_Mtb and PepN_Msm exhibit high sequence homology (10.6084/m9.figshare.7873274) and structural identity (S1 Fig), we also tested if PepN_Msm harbors ER-homing signals and if so can reach ER. Again, insilico analysis (LocSigDB) [42] of PepN_Msm showed ER homing-like sequences spread across both domains (S4 Fig). MtbΔpepN complemented with pepN_Msm was used for infecting THP-1. Similar to PepN_Mtb, most of PepN_Msm also did not reach nucleus, ER (Fig 1) and lysosomes (S3 Fig). Thus, despite (i) harboring ERAP1_C-like domains; (ii) containing ER homing-like signals, and (iii) structurally resembling host ERAP1 (10.6084/m9.figshare.7873451), neither PepN_Mtb nor PepN_Msm localized to ER. Their absence to localize to ER, nucleus (Fig 1) and lysosomes (S3 Fig), the staining pattern observed, and by process of elimination, we infer that both PepN_Mtb and PepN_Msm primarily localize to host macrophage 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 this knockdown strategy reduced pepN_Msm transcripts by ~80% (10.6084/m9.figshare.7873469), surprisingly, the steady state levels of PepN_Msm remained unaltered (10.6084/m9.figshare.7873469). This suggested that either PepN_Msm has an extended half-life or its levels are strictly maintained in Msm.

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.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. PepN_Msm is necessary for Msm’s in vitro growth.

(A) Freshly prepared electrocompetent Msm were transformed with suicidal vector containing 3’ end (~ 1 kb) of pepN alone (pNS39; top single plate; 500 ng pDNA) or pepN fused at its 3’ end with degradation tag ssrA (pNS38; i—500 ng; and ii—2500 ng of pDNA) and plated on Kanamycin (25 μg/ml) containing 7H11 + ADC plates. Similar results were observed with three independent rounds of electroporation. Transformants with pNS39 emerged in 5–6 days post transformation as it is a suicidal plasmid. Top plate was incubated for additional 36–48 h to see robust growth of colonies that emerged. The bottom plates (i) and (ii) were photographed after 3 weeks of incubation. (B) To compare growth rate, equal number (~10,000 bacteria) of WT Msm and possible revertant Msm expressing pepN::ssrA were spotted and then streaked on 7H11 + ADC plates. Both plates did not contain Hygromycin so that the speed of growth could be compared. The image recorded is after 4 days of incubation at 37°C. Broth cultures could not be used for comparison as the Msm expressing pepN::ssrA clumps in broth. (C) Equal amount of total proteins from WT Msm and Msm expression pepN::ssrA (revertant) was resolved on 10% SDS-PAGE gels, proteins transferred onto nitrocellulose membrane and western analysis performed with anti-PepN antibody (1:2500). M–protein marker. The major form of PepN_Msm (arrow) in the potential Msm revertant exhibit increase in size because of degradation tag (SsrA) fusion to PepN.

https://doi.org/10.1371/journal.pone.0215123.g002

Msm and not Mtb proteolyzes excess of its own PepN

Such opposing dependencies of Mtb and Msm to PepN in vitro led us to test if pepN_Mtb and pepN_Msm exhibit qualitative/quantitative differences in their transcript and/or steady state levels of PepN. Though pepN_Mtb and pepN_Msm exhibited some differences in their transcript levels in rich and minimal media (10.6084/m9.figshare.7873478), overall, by mid-log phase, significant dampening in transcript levels of pepN occurred in both Mtb and Msm (10.6084/m9.figshare.7873478). In contrast, western analysis showed both PepN_Mtb and PepN_Msm steady state levels being uniform across all stages of in vitro growth including stationary phase (S6 Fig). 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.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Msm and not Mtb proteolyzes excess of its own PepN.

Equal protein lysates from mid-log grown Mtb (A) or Msm (B) and their PepN overexpressing strains were resolved on 10% SDS-PAGE gels, western analysis performed, and steady state levels of PepN monitored. Anti-PepN_Mtb antibody was used for detection of PepN. Rabbit polyclonal anti-PepN antibody—(1:2500); anti-Rabbit IgG Goat secondary antibody (1:10000). (A): lane 2—WT Mtb; lane 3 & 4—Mtb with vector alone; lane 6 & 7 –Mtb overexpressing PepN_Mtb; lane 8 & 9 –Mtb overexpressing mutant PepN_Mtb; lane 10 & 11 –Mtb overexpressing PepN_Msm (B): lane 2—WT Msm; lane 3 & 4—Msm with vector alone; lane 6 & 7 –Msm overexpressing PepN_Msm; lane 8 & 9 –Mtb overexpressing PepN_Mtb; lane 10 & 11 –Mtb overexpressing mutant PepN_Mtb. M- Protein marker.–and + indicate absence or presence of AnTc (100 ng/ml) respectively. Bl–Blank lane. The experiments were independently repeated thrice. ‘PepN’: indicates whether the overexpressing version is wild type or a mutant form (both expressed as episomal). ‘Source’: Indicates the bacterium from where the ‘PepNs’ are derived from. ‘Bacterium’ indicates whether it is Mtb or Msm that expresses those ‘PepNs’. These also express their native PepNs from their genome.

https://doi.org/10.1371/journal.pone.0215123.g003

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—compare lanes 2, 4 and 7; S7C Fig). However, when we expressed pepN_Msm, it proteolyzed over accumulating PepN_Msm (Fig 3A—compare lanes 7 and 11). It similarly proteolysed accumulating high levels of 3XFLAG::PepN_Msm (S7D Fig). Unlike Mtb, Msm proteolysed excess of its own (PepN_Msm) (Fig 3B, compare lanes 2, 4 and 7) and foreign PepN (here PepN_Mtb; Fig 3B, compares lanes 2, 7 and 9). Thus, while Mtb selectively proteolyses over accumulating PepN_Msm, Msm proteolyses both PepN_Msm and PepN_Mtb. As expected, Mtb and Msm 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 (Table 1; see Materials & Methods for details). Though western analyses detected secreted PepNs, they required loading of atleast 15 fold excess culture filtrate proteins (S8 Fig). Therefore, we employed MS approach to detect secreted PepNs.

We had observed that, in Mtb, though pepN transcripts significantly scale down by mid-log (10.6084/m9.figshare.7873478), PepN steady state levels remain uniform across all phases of in vitro growth (including stationary phase; S6B Fig). As a result, we hypothesized that unlike Msm, Mtb perhaps maintains uniform PepN_Mtb levels by secreting out excess PepN. To evaluate this, through MS, we monitored PepN levels in SM of Mtb culture overexpressing its 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.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. PepN_Mtb and PepN_Msm interactomes are markedly different.

(A) Comparative representation of biological processes/functional groups into which the PepN_Mtb and PepN_Msm interactomes fall. (B) Comparative breakdown of intermediary metabolism/respiration group into different subgroups of enzymes to which PepN_Msm and PepN_Mtb interactomes belong. Detailed grouping shown in S3 and S4 Tables respectively. Biological triplicates were processed for MS/MS.

https://doi.org/10.1371/journal.pone.0215123.g004

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% 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).