A Mycobacterium avium subsp. paratuberculosis Predicted Serine Protease Is Associated with Acid Stress and Intraphagosomal Survival

The ability to maintain intra-cellular pH is crucial for bacteria and other microbes to survive in diverse environments, particularly those that undergo fluctuations in pH. Mechanisms of acid resistance remain poorly understood in mycobacteria. Although, studies investigating acid stress in M. tuberculosis are gaining traction, few center on Mycobacterium avium subsp. paratuberculosis (MAP), the etiological agent of chronic enteritis in ruminants. We identified a MAP acid stress response network involved in macrophage infection. The central node of this network was MAP0403, a predicted serine protease that shared an 86% amino acid identity with MarP in M. tuberculosis. Previous studies confirmed MarP as a serine protease integral to maintaining intra-bacterial pH and survival in acid in vitro and in vivo. We show that MAP0403 is upregulated in infected macrophages and MAC-T cells that coincided with phagosome acidification. Treatment of mammalian cells with bafilomcyin A1, a potent inhibitor of phagosomal vATPases, diminished MAP0403 transcription. MAP0403 expression was also noted in acidic medium. A surrogate host, M. smegmatis mc2 155, was designed to express MAP0403 and when exposed to either macrophages or in vitro acid stress had increased bacterial cell viability, which corresponds to maintenance of intra-bacterial pH in acidic (pH = 5) conditions, compared to the parent strain. These data suggest that MAP0403 may be the equivalent of MarP in MAP. Future studies confirming MAP0403 as a serine protease and exploring its structure and possible substrates are warranted.

The ability to maintain intra-cellular pH is crucial for bacteria and other microbes to survive in diverse environments, particularly those that undergo fluctuations in pH. Mechanisms of acid resistance remain poorly understood in mycobacteria. Although, studies investigating acid stress in M. tuberculosis are gaining traction, few center on Mycobacterium avium subsp. paratuberculosis (MAP), the etiological agent of chronic enteritis in ruminants. We identified a MAP acid stress response network involved in macrophage infection. The central node of this network was MAP0403, a predicted serine protease that shared an 86% amino acid identity with MarP in M. tuberculosis. Previous studies confirmed MarP as a serine protease integral to maintaining intra-bacterial pH and survival in acid in vitro and in vivo. We show that MAP0403 is upregulated in infected macrophages and MAC-T cells that coincided with phagosome acidification. Treatment of mammalian cells with bafilomcyin A1, a potent inhibitor of phagosomal vATPases, diminished MAP0403 transcription. MAP0403 expression was also noted in acidic medium. A surrogate host, M. smegmatis mc 2 155, was designed to express MAP0403 and when exposed to either macrophages or in vitro acid stress had increased bacterial cell viability, which corresponds to maintenance of intra-bacterial pH in acidic (pH = 5) conditions, compared to the parent strain. These data suggest that MAP0403 may be the equivalent of MarP in MAP. Future studies confirming MAP0403 as a serine protease and exploring its structure and possible substrates are warranted.

INTRODUCTION
A universal theme among bacteria is the ability to persist, replicate, and expand their territory despite which environment they occupy. Likewise several universal stressors are found within every environment such as changes in nutrients, pressures, temperatures, pHs, etc. Intrinsic and developed acid response mechanisms have been well-studied in enteric pathogens, such as Salmonella spp., H. pylori, Shigella spp., and E. coli (Gorden and Small, 1993;Castanié-Cornet et al., 2010;Valenzuela et al., 2011;Ryan et al., 2016). Bacteria and other microbe resistance to and survival in acid has allowed for the development of diagnostics and drugs as well as the understanding of host responses that regulate acid stress. For example, investigations into acid stress have elucidated host components of the mycobacterial disease process, such as the acidification of M. tuberculosis containing phagosomes (pH = 4.5-4.8) during infection (Sprick, 1956). It is also important to note that the first-line drug in TB, pyrazinamide, becomes active due to acidification of the mycobacteria containing phagosome (Zhang et al., 2014). It is therefore all the more surprising why few studies have explored mechanisms of acid resistance in mycobacteria. Current research in acid resistance mechanisms within mycobacteria have largely been focused in M. tuberculosis and show the involvement of several phosphate-sensing signal transduction systems (Ramakrishnan et al., 2016), cytoplasmic redox sensors (Saini et al., 2012;Mehta et al., 2016), proteases/peptidase (Vandal et al., 2008), lipoglycans (Shui et al., 2011), and transcriptional repressors (Healy et al., 2016) unrelated to phagosome maturation arrest. It is likely that similar acid resistance mechanisms exist in other mycobacterial species.
Mycobacterium avium subsp. paratuberculosis (MAP) is a unique member of the mycobacteria genus due to its ability to colonize and penetrate the intestinal epithelium (Bannantine and Bermudez, 2013), causing a progressive and chronic enteritis in ruminants termed Johne's disease (JD) (Sweeney, 2011). MAP successfully survives several acid exposures found in extracellular (soil and water) (Elliott et al., 2015), intrahost (stomach) and intracellular environments. While acid stress in MAP has been studied in vitro (Sung and Collins, 2003;Wu et al., 2007), no study to date has investigated potential mechanisms of acid resistance employed during the initial stages of cell infection. MAP preferentially infects subepithelial dome (SED) and lamina propria macrophages and localizes in the phagosome or phagolysosome that ranges in pH from 4.5 to 6.2 (Bannantine and Bermudez, 2013). Successful phagosome maturation leads to the increased destruction of engulfed pathogens or particles by a series of increasingly acidified and oxidatively stressed membrane-bound vesicle fusions and fissions (Kinchen and Ravichandran, 2008). Current studies have found that mycobacteria stall or inhibit phagolysosome biogenesis through (1) dysregulation Rab GTPases trafficked to the phagosome (Kelley and Schorey, 2003;Seto et al., 2011), (2) suppression of macrophage activation cytokines, particularly gamma-interferon (IFN-) (Clemens Dl, 1995;Giacomini et al., 2001), (3) neutralization of reactive oxygen intermediates (Miller et al., 2004;Nguyen and Pieters, 2005), and (4) inhibition of vacuolar-ATPases (vATPases) crucial for establishing and maintaining acidification (Wong et al., 2011;Kissing et al., 2015). MAP must tightly regulate the phagosome acidification process to allow for host adaptation and recruitment of other host cells and factors necessary for pathogen survival and replication (Lamont et al., 2012), while maintaining its intrabacterial pH (pH IB ). How MAP achieves this balance in the phagosome is incompletely understood.
In order to identify possible genes and mechanisms involved in acid resistance within the early stages of MAP infection, we conducted a microarray analysis of MAP in bovine MDMs treated with and without bafilomycin A1, a potent inhibitor of vATPases. We discovered an acid stress network with MAP_RS02055 (herein referred to by its original nomenclature, MAP0403) serving as the central node. Computational analysis predicted MAP0403 as a transmembrane protein that shares an 86% amino acid identity to a recently characterized serine protease, Rv3671c, found in M. tuberculosis. A M. tuberculosis transposon library screen showed that a loss-offunction insertion in Rv3671c resulted in bacteria hypersensitivity to acid (pH = 4.5) and failure to maintain pH IB in acid in vitro and IFNactivated macrophages (Vandal et al., 2008). Furthermore, the Rv3671c mutant showed a growth defect using a mouse model (Vandal et al., 2008). Expression, biochemical analyses and crystallization of the Rv3671c periplasmic domain confirmed its function as a transmembrane serine protease; therefore, the protein was renamed MarP for mycobacterial acid resistance protein (Biswas et al., 2010;Small et al., 2013). Rv3671c will be referred to as MarP for the remainder of this manuscript.
This study characterizes MAP0403 in response to intracellular and in vitro acid stress. We show that MAP0403 is upregulated during initial infection in multiple cell types and exposure to extracellular acid stress. Using a MAP surrogate, M. smegmatis mc 2 155, expressing MAP0403, we demonstrate that MAP0403 is associated with increased bacteria survival and maintenance of pH IB . Thus, we propose that MAP0403 is likely a functional equivalent of marP in MAP.

Ethics Statement
All research was conducted in accordance with the University of Minnesota's Intuitional Biosafety Committee (IBC) approval protocol 0806H36901. All animal studies were conducted in compliance with the recommendations of the University of Minnesota's Institutional guidelines and approved animal care and use committee (IACUC) under approval protocol 1207A17288.

Mammalian Cell Culture
Unless otherwise noted, all mammalian cell culture and experiments were conducted at 37 • C in a humidified chamber containing 5% CO 2 . MDMs from JD-free dairy cows were elutriated using a 58% percoll gradient and matured in teflon wells as described in Coussens et al. (2002) and Janagama et al. (2006). MDMs were maintained in RPMI 1640 supplemented with 20% autologous serum. Bovine mammary epithelial cells (MAC-T) were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS).

Construction of M. smegmatis mc 2 155 Expressing MAP0403
The ORF of MAP0403 was PCR amplified from MAP K-10 genomic DNA using a high-fidelity taq polymerase and primers 0403-HF (5 ′ CCC AAG CTT GTG ACG CAC TCG AAT GA 3 ′ ), 0403-SR (5 ′ ACA TGC ATG CTC AAC TGA CGC AGG A 3 ′ ) engineered with Hind III and Sph restriction sites at the 5 ′ end, respectively. The amplified fragment was cloned into a restricted pSM417 vector (herein referred to as pSM417-MAP0403 using T4 DNA ligase and electroporated (single pulse generated at 2.5 kV, 1000 ) into competent M. smegmatis mc 2 155 cells. Competent M. smegmatis mc 2 155 were created using the method described by Goude et al. (2015). Insert orientation and sequence fidelity were confirmed by classic Sanger sequencing at the University of Minnesota's Genomics Center. A vector control strain (pSM417 alone) in M. smegmatis mc 2 155 was also created.

MAP Cell Invasion Assay
MDMs and MAC-T cells were used to study global gene expression profiles during initial MAP infection and/or MAP0403 expression. We used established methods described by our laboratory (Zhu et al., 2008;Lamont and Sreevatsan, 2010;Lamont et al., 2012). Briefly, 2 × 10 6 MDMs/flask were seeded into 25 cm 2 flasks, incubated and allowed to adhere for 2 h. Successful cell attachment was confirmed by phase-contrast microscopy. Non-adherent cells were removed by washing with PBS prior to infection. MAP K-10 and M. smegmatis mc 2 155 expressing pSM417 and pSM417-MAP0403 were pelleted at 500 × g for 15 min, separately re-suspended in RPMI 1640 containing 2% autologous serum, and passed multiple times through a syringe-driven 18 G needle. MDMs were infected separately with the above mycobacterial strains at a MOI of 10:1 for 2 h. Upon completion of incubation, cells were washed thrice with pre-warmed PBS and treated with amikacin for 2 h to remove extracellular bacteria. Fresh medium was added to 25 cm 2 flasks and incubated for 10, 30, and 120 min. After post-infection, MDMs were again washed in PBS and were either lysed in PBS containing 0.1% Triton X-100 or collected for RNA extraction. Triton X-100 lysates underwent differential centrifugation to separate host cellular debris from bacteria cells. Bacteria were resuspended in 1.0 mL of PBS, serially diluted on MB7H9 or LB-Lennox agar, and incubated at 37 • C to determine CFUs. Hygromycin (100 µg/mL) was added to LB-Lennox agar when appropriate. MAC-T cells were seeded at a density of 2 × 10 4 cells/well in a 24 well polysterene plate and allowed to reach 80% confluence. The same invasion method and post-infection processing as MDMs were applied to MAC-T cells with the exception that the initial infection time point was 3 h. Each time point and condition were conducted in biological triplicates. Each experiment was replicated three times.

Phagosome Acidification Blocking Assay
MDMs and MAC-T cells were separately seeded at 2 × 10 4 cells/well in a 24 well plate containing No. 1.5 glass cover slips. MAP invasion using MAP K-10(pWes4)-GFP was conducted using the same conditions described above with the exception of an 1 h pre-incubation step with 50 nM of bafilomcyin A1 (A.G. Scientific Inc., San Diego, CA). Phagosome acidification was determined by a LysoTracker staining method (Lamont et al., 2012). Upon the final 30 min of each post-infection time point, 25 nM of LysoTracker Blue DND-22 (Invitrogen, Carlsbad, CA) was added to cell medium. Cells were subsequently washed thrice in Dulbecco's PBS, incubated in pre-warmed Deep Red CellMask plasma membrane stain (2.5 µg/mL) (Invitrogen, Carlsbad, CA) for 5 min, and rewashed. Cells were fixed in absolute methanol for 5 min at −20 • C and washed twice in ice-cold D-PBS. Slides were stored at 4 • C until visualization by confocal microscopy. An Olympus FluoView 1000 upright confocal microscope (Olympus, South-end-on-sea, Essex, United Kingdom) was used to image infected and control cell slides slides with FITC, Cy5, and DAPI lasers. Z-series for each slide was taken in 1 µM steps and stacked to render a complete image. A minimum of three fields per slide were imaged. Blocking assay was repeated in triplicate for each condition.

RNA Extraction
All work surfaces and equipment were treated with RNase Away (Molecular Bioproducts, Inc., San Diego, CA). Total RNA was extracted from MAP infected MDMs treated with/out bafilomycin A1, acid treated and control bacteria using TRIzol reagent (Invitrogen, Carlsbad, CA) per manufacturer's instructions. Bacterial lysates were homogenized using a reported bead-beating method (Janagama et al., 2010). RNA samples were treated with the TURBO DNA-free Kit (Ambion, ThermoFisher Scientific, Rockford, IL) and subjected to PCR to confirm that the samples were devoid of genomic DNA. RNA quality and concentration was determined by measuring the 260/280 ratio on a NanoDrop ND 1000 spectrophotometer (Nanodrop Products, Wilmington, DE).

RNA Processing and Labeling
RNA extracted from MAP K-10 infected MDMs treated with/out bafilomycin A1at 30 min p.i. were processed and hybridized as described (Janagama et al., 2010). Briefly, total RNA treated with DNase was processed to remove host RNA and the 16S ribosomal RNA using MICROBEnrich and MicrobExpress kits (Ambion, Thermo Scientific, Rockford, IL) as specified by manufacturer. Microbial RNA was amplified using the MessageAmpII Bacteria kit for prokaryotic mRNA per manufacturer's instructions (Ambion, ThermoFisher Scientific, Rockford, IL). Labeled (Cy3 or Cy5) DNA was produced from microbial mRNA using the Superscript Plus Direct cDNA labeling system (Invitrogen, Carlsbad, CA) with aminoallyl-dUTP followed by a coupling of the aminoallyl groups to either Cyanine-3 or Cyanine-5 (Cy-3/Cy-5) fluorescent molecules. cDNA reactions were pooled by treatment group to obtain a sufficient concentration of labeled DNA. Effective labeling was achieved by incubating the aminoallyl-dUTP coupled cDNA with the dye for 2 h at room temperature (RT). Labeled cDNA was hybrized to 70mer oligonucleotide microarray slides (National Animal Disease Center, Ames, IA) overnight at RT, washed in microarray buffer and scanned using the HP Scanarray 5000 (Perkin Elmer Inc., Waltham, MA). All images were stored. Microarray experiments were repeated three times. Raw microarray data files have been submitted to NCBI Gene Expression Omnibus (GEO) (Edgar et al., 2002) and are accessible through GEO series accession number GSE84708.

Microarray Data Analysis
All microarray experiments were conducted using the minimal information about a microarray experiment (MIAME) guidelines. Microarray image analysis software, BlueFuse (BlueGnome Ltd., Cambridge), was used to extract numeric data from stored microarray images. Normalization by global LOWESS was performed and expression data was analyzed by GeneSpring GX 10 (Agilent Technologies, Foster City, CA). Normalized ratios were reported as fold change. Differentially expressed genes (DEGs) were cross-referenced to the MAP K-10 genome and the remaining mycobacterial genomes listed in National Center for Biotechnology Institute (NCBI) using Basic Local Alignment Search Tool (BLAST) alogorithm. MAP gene networks were analyzed by the STRING database (http://www. string-db.org).

Quantitative Real Time PCR
One-step qRT-PCR was performed using QuantiFast SYBR Green One-Step QRT-PCR mix (Qiagen, Valencia, CA) in a LightCycler 480II (Roche, Madison, WI) with corresponding software. The following program was used: 50 • C for 10 min, 95 • C for 5 min (activation), 95 • C for 10 s (denature), and 95 • C for 15 s, 60 • C for 1 min repeated for 40 cycles (PCR amplification). Primers used for qRT-PCR were designed using Primer 3 software (http://frodo.wi.mit.edu/primer3/) and are listed in Table S1. Fold changes were calculated using the 2 − Ct method (Livak and Schmittgen, 2001). The value of the housekeeping gene, secA, was normalized to untreated bacteria cells. All samples were conducted in technical triplicates.

Measurement of In vitro Acid Treatment
MAP K-10 (OD 600 = 0.3) in MB7H9 was vortexed for 5 min and passed 10 times through a sterile 18 G syringe-driven needle to disperse bacterial clumps. The culture was allowed to stand for 5 min to facilitate the sedimentation of bacterial clumps. The upper two-thirds of the culture (containing a single cell suspension) were used in all subsequent experiments. Acidity of test cultures was adjusted to a pH of 5 by 2N HCl, while control cultures were maintained at a pH of 6.6 ± 0.2. All cultures were mixed and allowed to rotate at 37 • C for 10, 30, and 120 min. Each condition was conducted in biological triplicates. All experiments were repeated three times. CFUs were determined by plating serial dilutions of the suspension in duplicate on MB7H9 agar.

Measurement of Intrabacterial pH of Carboxyfluorescin Labeled M. smegmatis mc 2 155-pSM417-MAP0403
Log-phase M. smegmatis mc 2 155 parent and pSM417-MAP0403 transformants were labeled with 5(6)-carboxyfluorescein Nhydroxysuccinimide ester (5(6)-CFDA) (Sigma, St. Louis, MO) as described in Gaggìa et al. (2010). Briefly, 5.0 mL of bacterial cells were harvested by centrifugation at 13,000 rpm for 5 min, resuspended in 980 µl of filter (0.45 µM pore size) sterilized citric acid-phosphate-buffer (pH 7.0) supplemented with 10 µL of 1 M glucose and 10 µL of 5(6)-CFDA, and incubated for 1 h at 37 • C. Upon completion of incubation, cells were sedimented at 13,000 rpm for 5 min, washed thrice in phosphate buffered saline (PBS), and resuspended in LB broth (pH = 7). A standard curve relating fluoresce with decreasing pH was established by irreversible membrane permeablization of M. smegmatis mc 2 155 parent using ethanol (63% v/v) incubation at 37 • C for 30 min. Bacteria were pelleted as described above, resuspended in 5 mL of LB broth adjusted to pH range of 4-8, and incubated at 37 • C for 30 min. to equilibrate pH IB to medium pH. Fluorescence at each pH was measured using a SpectraMax 2000 spectrophotometer (492 nm excitation, 517 nm emission) (Molecular Probes, San Diego, CA) and the mean fluorescence units were plotted against pH. Acid treatment for parent and pSM417-MAP0403 strains was repeated with the exception of irreversible membrane permeabilization. Fluorescence units were compared against the standard curve to determine pH IB . The experiment was thrice repeated.

Statistical Analysis
The results pertaining to relative fold changes and CFUs were analyzed by two-way analysis of variance (ANOVA) with Bonferroni correction in Graphpad Prism software (GraphPad Software, La Jolla, CA). P < 0.05 were considered to be statistically significant. Fitness of M. smegmatis transformants in MDM infection is shown as a box-whisker plot to demonstrate the actual distribution of observed CFUs.

A Predicted Serine Protease Network is Expressed during Phagosome Acidification
In order to identify the acid stress transcriptome during initial MAP infection, we first sought to characterize the phagosome maturation process in bovine MDMs. MAP K-10 was allowed to invade MDMS for 10-120 min post-infection (p.i.) and phagosomes were subsequently assessed for acidification using LysoTracker Blue, a fluorescent, acidotropic probe, that emits upon protonation of its basic amine. We observed phagosome acidification as early as 10 min in MDMs ( Figure 1A). Phagosome acidification continued for 1 h ( Figure 1A); however, the acidification process completed by 2 h p.i. (data not shown). Phagosome acidification was validated by pre-treatment of MDMs with bafilomcyin A1, a potent vacuolar ATPase inhibitor. Bafilomcyin A1 treatment of MDMs abrogated phagosome acidification at all p.i. time points (Figure 1A).
We selected a 30 min p.i. time point to perform a microarray analysis of genes expressed during phagosome acidification in MDMs. We compared expression profiles from MDMs infected with MAP K-10 with and without bafilomycin pre-treatment. Microarray analysis identified a MAP acid stress related network composed of genes MAP0403, MAP_RS20120 (MAP3922), MAP_RS12430 (MAP2439c), MAP_RS02045 (MAP0401), MAP_RS02035 (MAP0399c), rlmN, mutY, and nth ( Figure 1B). All genes within the acid stress network were validated by qT-RT-PCR ( Figure 1B).
MAP0403 was the center node of the acid stress network. A Basic Local Alignment Search Tool (BLAST; https://blast.ncbi. nlm.nih.gov/Blast.cgi) protein analysis predicted that MAP0403 encodes a serine protease that is 86% identical and 92% similar to MarP found in M. tuberculosis. MAP0403 is composed of 397 amino acids and is projected by transmembrane helix prediction software (TMHMM) to have 4 transmembrane helices at the Nterminus, similar to MarP (Figure S1A). The C-terminal portion is predicted to contain a trypsin-like protease domain and is likely localized to the periplasm. In MarP, oxidative stress triggers autocleavage of the protease domain and stabilization of the active site through reduction of disulfide bonds and subsequent activation of the domain's protease activity (Cameron et al., 1994;Vandal et al., 2008Vandal et al., , 2009Biswas et al., 2010). Cysteine disulfide bond prediction of MAP0403 using DiANNA ver. 1.1 web software showed a disulfide bond located at amino acid positions 214-397 (SLAPSCQKVLE-VGTGSCVS), which is also present in MarP. Further motif analysis showed that like MarP and other serine proteases, including eukaryotic chymotrypsin proteases, MAP0403 contains the conserved catalytic triad that functions at the active site of transferase enzymes ( Figure S1B). Together these data suggest that MAP0403 is likely the equivalent of MarP in MAP. Given the role of MAP0403 as the central node in an acid stress network and its conservation to MarP, we focused on elucidating the association of MAP0403 with acid stress and pH IB .

MAP0403 Transcription during Initial Cell Infection is Associated with Increased Bacterial Survival
To further understand MAP0403 and its role in bacterial survival during early events in host infection, we analyzed MAP0403 transcription in MAP K-10 exposed to MDMs treated with and without bafilomycin A1. As previously performed, MAP K-10 was allowed to invade MDMs for 10, 30, and 120 min p.i. and MAP0403 transcription was analyzed by qT-RT-PCR. MAP0403 was up-regulated by a 30-fold difference compared to control MAP K-10 within 10 min (Figure 2A). MAP0403 upregulation was sustained throughout all p.i. time points; however, expression was reduced by 14-fold at 120 min (Figure 2A). Pre-treatment of MDMs with bafilomycin A1 decreased MAP0403 expression at all p.i. times (Figure 2A). We next sought to analyze MAP0403 transcription during infection within another cell type to determine if expression occurs only in a specific host cell. We used MAC-T cells, a mammary epithelial cell line, as a surrogate (Patel et al., 2006) for the intestinal epitheliumthe first tissue encountered by MAP during natural infection. In a previous study, we showed that phagosome acidification of MAC-T cells infected with MAP K-10 occurred within 10 min and was sustained until 60 min p.i. (Lamont et al., 2012). Unlike MDMs, MAP0403 differential expression was not detected until 30 min p.i. (Figure 2B). The 30-fold upregulation of MAP0403 compared to the bacteria control corresponded to peak acidification in MAC-T cells ( Figure 2B) (Lamont et al., 2012). MAP0403 expression decreased by 120 min p.i. and showed a 16-fold difference compared to control ( Figure 2B). As expected, pre-treatment of MAC-T cells with bafilomycin A1 abolished MAP0403 expression at all p.i. time points (Figure 2B).
We further interrogated the role of MAP0403 during cell infection by asking whether or not expression was critical to bacteria survival. Attempts by our laboratory to create a MAP0403 deletion in MAP have been unsuccessful. Therefore to answer this question, we utilized the non-pathogenic mycobacteria, M. smegmatis mc 2 155, as a surrogate host to express MAP0403 due to its inability to survive in macrophages past 48 h (Lagier et al., 1998;Kuehnel et al., 2001;Anes et al., 2003Anes et al., , 2006. It is important to note that M. smegmatis mc 2 155 contains a gene, MSMEG_6183, that is predicted to encode a serine protease that shares 66 and 68% amino acid identities with MarP and MAP0403, respectively. However, we found that MSMEG_6183 is not up-regulated during MDM infection or under in vitro acid stress (pH = 5) compared to control bacteria at 10-120 min p.i. (data not shown). M. smegmatis mc 2 155 was electroporated with a plasmid expressing MAP0403 from an Hsp60 promoter (M. smegmatis pSM417-MAP0403) or an empty vector [M. smegmatis pSM417(empty)]. M. smegmatis mc 2 155 parent and transformants were exposed to MDMs for 0, 10, 30, and 120 min p.i. and assessed for bacterial viability. M. smegmatis expressing MAP0403 showed a greater than 1.5 log 10 increase in CFU recovery compared to M. smegmatis controls within 10 min p.i. (Figure 2C). Increased recovery of M. smegmatis pSM417-MAP0403 vs. controls was maintained throughout MDM infection ( Figure 2C). MAP0403 transcription was evaluated in M. smegmatis mc2 155-MAP0403 in response to MDM infection ( Figure S2). However, MAP0403 expression was not differentially expressed to the control time point at 0 min ( Figure S2A). This is likely due to its constitutive expression under the pSM417 Hsp60 promoter. In summary, these data provide initial support that MAP0403 is expressed during host infection in different cell types undergoing phagosome acidification and likely aids in bacterial cell survival.

MAP0403 Transcription and Association with Bacterial Survival and Intrabacterial pH Homeostasis under Extracellular Acid Stress
Although phagosome acidification was observed in both MDMs and MAC-T cells, acid stress represents one of several host stresses against MAP and other factors including Reactive Oxygen Species (ROS) and Reactive Nitrogen Intermediates (RNI), etc. were likely present within the phagolysosome. Therefore, the sole contribution of acid to MAP0403 transcription could not be determined. In order to understand the direct influence of acid stress to MAP0403 expression and bacteria survival, we took a reductionist approach by employing acid stress in vitro to MAP K-10 and M. smegmatis mc 2 155 pSM417-MAP0403. MAP K-10 was separately incubated with acidified (pH = 5) and untreated broth (pH = 6.8) for 10-120 min and interrogated for MAP0403 differential expression. Acid treatment resulted in a fourfold upregulation of MAP0403 within 10 min (Figure 3A) compared to control. MAP0403 transcription increased to ∼8-fold by 30 and 60 min, but decreased by 120 min (Figure 3A).
MAP K-10, M. smegmatis mc 2 155 pSM417-MAP0403, and M. smegmatis mc 2 155 pSM417(empty) were incubated in the same conditions described above and analyzed for recovery using direct plating. Within 10 min of acid treatment, CFUs for all strains were reduced by 0.5-1 log 10 compared to inoculum ( Figure 3B). However, by 120 min of acid exposure MAP K-10 and M. smegmatis mc 2 155 pSM417-MAP0403 recovered and stabilized in contrast to M. smegmatis mc 2 155 pSM417(empty), which showed a 1-fold reduction in CFUs ( Figure 3B). We analyzed MAP0403 transcription in M. smegmatis mc 2 155 pSM417-MAP0403 in response to acid in vitro ( Figure S2B). Similar to pSM417-MAP0403 transcription during MDM infection, data showed that MAP0403 was not differentially expressed ( Figure S2B).
Lastly, we asked if M. smegmatis mc 2 155 pSM417-MAP0403 viability during in vitro acid stress corresponded to maintenance of its pH IB . A 5(6)-CFDA fluorescence ratio microplate assay was used to determine pH IB (Siegumfeldt et al., 1999;Gaggìa et al., 2010). A calibration curve was created to establish the link between fluorescence and pH (range 4-8; Figure S3). Increasing fluorescence was positively correlated to pH (Figure S3). M. smegmatis mc 2 155 pSM417-MAP0403 and M. smegmatis mc 2 155 stained with 5(6)-CFDA were exposed to a pH range of 4-7 for 30 min, measured for fluorescence, and compared against the calibration curve. Both strains maintained a neutral pH IB when exposed to extracellular alkaline pHs as evidenced by an increased fluorescence ( Figure 3C). However, only M. smegmatis mc 2 155 pSM417-MAP0403 was able to sustain a neutral pH IB when incubated with medium acidified to pH = 5 ( Figure 3C). All strains at pH = 4 failed to show a neutral pH IB (Figure 3C).
Overall, data showed that extracellular acid exposure is one signal for MAP0403 expression and similar to intracellular survival, MAP0403 supports bacteria viability when met with an acid stressor.

DISCUSSION
Resistance and tolerance to acid due to multiple exposures to low pH have been well-studied in enteric bacteria. Unlike enteric pathogens, repeated incubations of mycobacteria with FIGURE 3 | MAP0403 aids in bacterial survivability during in vitro acid exposure. (A) MAP K-10 was exposed to acidified (pH = 5.0) MB7H9 for 10, 30, and 120 min, recovered, and analyzed for MAP0403 expression by qT-RT-PCR. MAP0403 was up-regulated by 4-fold within 10 min of acid exposure and reached peak expression by 30 and 60 min. (B) MAP K-10, M. smegmatis mc 2 155 parent, MAP0403 expression strain (pSM417-MAP0403), and vector control [pSM417 (empty)] were exposed to the same acid conditions as (A) and underwent direct plating to determine viability counts. MAP K-10 and M. smegmatis mc 2 155 expressing MAP0403 cells had higher rates of recovery compared to M. smegmatis mc 2 155 parent and vector control strains. (C) Intra-bacterial pH measurement in M. smegmatis mc 2 155 parent and transformants. Intra-bacterial pH was measured by a fluorescent marker, 5(6)-carboxyfluroescein N-hydroxysuccinimide, in M. smegmatis mc 2 155, MAP0403 expression strain (pSM417-MAP0403), and vector control [pSM417 (empty)] exposed to various pHs ranging from 4 to 7. Compared to controls, M. smegmatis pSM417-MAP0403 maintained its intra-bacterial pH in acidified medium (pH = 5). *P < 0.05, n.s. not significant. acid did not significantly increase resistance compared to controls (O'brien et al., 1996). Current studies investigating acid resistance in mycobacteria indicate that the genus has developed specific mechanisms to either persist in or neutralize acid found in multiple environments. Mycobacterial strategies for combating acid are now being elucidated. Because MAP and other known pathogenic mycobacteria, reside and persist in slightly acidic phagolysosomes, we sought to identify upregulated MAP genes involved in the early phagosome acidification process. Microarray analysis uncovered an acid stress network that contained MAP0403 as its central node. We found that MAP0403 is upregulated during the initial events of cell infection and in vitro acid stress. Furthermore, MAP0403 is associated with increased bacterial survival and pH IB homeostasis.
MAP0403 is a predicted membrane serine protease and computational analysis of its secondary structure suggests it belongs to the chymotrypsin family. MAP0403 contains 4 N-terminal transmembrane bound helices and a Cterminal protease domain likely located in the periplasm. In addition to MAP0403, the MAP K-10 genome contains three other hypothetical trypsin-like serine proteases: PepD (MAP_RS04650), PepA (MAP_RS18120), and HtrA (MAP_RS13035; Li et al., 2005). Information pertaining to PepD, PepA, and HtrA in MAP in the scientific literature concerns the presence of these proteins as serological markers or structure and sequence similarities with serine proteases found in enteric species (Cameron et al., 1994;Shin et al., 2010). For example, N-terminal sequencing of MAP_RS13035 determined that the gene shared 30% homology with htrA genes found in E. coli, Salmonella spp., Brucella abortus, and R. henselae (Cameron et al., 1994). PepA in V. cholera (Behari et al., 2001) and E. coli (Maurer et al., 2005) was shown to activate during low pH and may be utilized for pathogen protection against oxidative stress. White et al. demonstrated interaction of PepD with the 35-kDa cell wall antigen in M. tuberculosis. The authors suggest that PepD may represent a general stress response strategy for cell wall maintenance (White et al., 2011). HtrA functions as both a chaperone and protease in E. coli and is critical for cell survival under elevated temperatures and oxidative stress conditions (Skorko-Glonek et al., 2008). With the exception of MarP in M. tuberculosis, mechanisms of serine proteases in mycobacteria remain poorly understood and understudied.
MAP0403 closely resembles MarP in M. tuberculosis and shares an 86% amino acid conservation. The function of MarP as a serine protease integral for pH IB homeostasis was recently confirmed by Sabine Ehrt and colleagues (Vandal et al., 2008). A M. tuberculosis transposon library screen identified hypersensitivity to acid stress at pH = 4.5 within a marP insertional mutant. The marP insertional mutant was unable to maintain its neutral pH IB in IFN-activated macrophages and after 24 h of infection pH IB had lowered to 5.5. This result was also mirrored in in vitro acid exposure for 8 h and corresponded to 3 log 10 reduction in viable bacteria. Complementation of the marP mutant restored acid resistance in M. tuberculosis. Attenuation of the marP mutant was also observed during in vivo infection of mice, in which 21 days p.i. resulted in a 1 log 10 reduction compared to wild type. We also used a macrophage model to analyze MAP0403. Expression of MAP0403 occurred within 10 min p.i. and corresponded with phagosome acidification (Figures 1A, 2A). Inhibition of vATPases using bafilomycin A1 diminished MAP0403 up-regulation (Figure 2A). We further examined MAP0403 transcription in MAC-T cells. MAP interacts with intestinal epithelial cells before macrophage invasion (Bannantine and Bermudez, 2013). In a previous study we showed that MAP induced peak phagosome acidification concomitant with calcium-dependent IL-1β secretion within 30 min p.i. to recruit macrophage recruitment to the site of infection (Lamont et al., 2012). We found that MAP0403 expression coincided with phagosome acidification in MAC-T cells (Figure 2B). Similar to bafilomycin A1 treatment in macrophages, inhibition of vATPases abolished MAP0403 expression. MAP0403 upregulation was also observed in MAP K-10 exposed to acid in vitro (pH = 5) ( Figure 3A).
Attempts by our group failed to grow a MAP0403 mutant. We are currently investigating other loss-of-function mutation methods such as transposonal insertion that was successfully used by Vandal et al. in M. tuberculosis. It is possible that MAP0403 is an essential gene in MAP and that the Tetracyclineinducible system for conditional expression may be necessary (Carroll et al., 2005). Due to the lack of a MAP0403 mutant, we utilized M. smegmatis mc 2 155 expressing MAP0403 to assess a potential role in bacteria viability. M. smegmatis mc 2 155 has a 68% amino acid identical gene, MSMEG_6183, to MAP0403. However, this gene was not upregulated in either in vitro acid stress (pH = 5) or MDM infection. Two scenarios exist for the absence of MSMEG_6183 expression: (1) MSMEG_6183 serves a different function in M. smegmatis or (2) conditions were not optimal or met for expression. Given that MarP and organization of the operon MSMEG_6183 is found in is conserved within the entire Mycobacteria genus, a different function is improbable. The lack of MSMEG_6183 expression at our designated conditions may be a reflection of its saprophytic and fast-growing nature (M. tuberculosis and MAP are categorized as slow-growers). Further studies should simulate conditions found within soil environments. We showed that M. smegmatis mc 2 155 expressing MAP0403 had increased CFU recovery compared to bacteria controls in both intracellular infection and in vitro acid stress (Figures 2C,  3B). Neutral pH IB was maintained in M. smegmatis mc 2 155 expressing MAP0403 during in vitro acid stress (pH = 5); however, at lower pHs the expression and control strains were unable to regulate pH IB (Figure 3C). Anes et al. have noted that M. smegmatis undergoes three distinct phases of killing interspersed with two phases of bacterial replication during macrophage infection (Anes et al., 2006). The authors observed that the majority of phagosomes did not acidify until 8-24 h p.i. However, it was noted that bafilomcyin A1 treatment of macrophages even during early infection time points prevented killing of M. smegmatis suggestive that at least 20-25% of phagosomes acidify prior to 24 h p.i. The early killing of M. smegmatis within macrophages (phase 1) at 4 h p.i. was associated with oxidative stress. Oxidative stress is triggered by phagosome acidification (Vandal et al., 2009); therefore, the possibility remains that acidification and oxidative stress work in tandem to kill mycobacteria and both were abolished due to bafilomcyin A1 treatment. Interestingly, Biwas et al. demonstrate that MarP is a dual response gene and responds equally to oxidative stress (Biswas et al., 2010). Further studies should determine if (1) MSMEG_6183 is upregulated in M. smegmatis at phase 1 killing (p.i. time point at 4 h) in macrophages, (2) whether both oxidative and acid stress are needed for MSMEG_6183 expression, and (3) the contribution of oxidative stress to MAP0403 expression.
While the data suggests that MAP0403 is associated with acid stress and pH IB homeostasis, the mechanism of how MAP0403 regulates this response is unclear. MarP structural studies by Biwas et al. and Small et al. have shown that the protein (1) contains a chymotrypsin fold and disulfide bond that stabilize the protease active site for substrate binding (Biswas et al., 2010) and (2) tethering of the transmembrane helices to the periplasm is necessary for proteolysis critical for pH IB during acid and oxidative stress (Small et al., 2013). The mycobacterial substrate for MarP is unknown. Biswas proposes two potential mechanisms for MarP's protection of bacterial cells in response to acid and oxidative stress (Biswas et al., 2010). The first mechanism revolves around the ability of MarP to degrade an unrelated protein, β-caesin, and suggests that MarP functions to degrade unfolded proteins found within the periplasm resultant of acid and oxidative stress. The second mechanism disassociates MarP from the stress response and states that it is essential for general cell wall maintenance and once function is lost, cell wall integrity is compromised and bacteria fail to regulate pH IB . Identification of the substrate bound to the active site may shed light on the correct mechanism. Comparative studies using MarP, MAP0403, and MarP equivalents in other mycobacteria may aid in identification or narrow potential substrates by using a "mix-and-match" method between substrates and predicted serine proteases (Madej and Kaback, 2013;Madej, 2015).
Characterization of serine proteases and other proteins necessary for acid and oxidative resistance are likely to provide a new avenue for drug and vaccine development against pathogenic mycobacteria. For example, Zhao et al. developed a whole cell screen and high throughput screen (HTS) of natural products and small, synthetic organic compounds (obtained through the National Institutes of Health Molecular Libraries Screening Center) against MarP, respectively. While the natural product screen did not yield binding to MarP (Darby et al., 2013), 9 compounds from the 324, 751 small, synthetic library reduced M. tuberculosis pH IB , inhibited MarP cleavage of β-casein, and failed to autofluoresce and interact with mammalian serine proteases (Zhao et al., 2015). Four benzoxazinones were found within the nine compounds. A specific B series benzoxazinone, BO43 (3.3 µM), decreased M. tuberculosis' pH IB past the limit of detection by covalently acetylating the active site of MarP, which was confirmed by LC-MS/MS (Zhao et al., 2015). It is exciting to speculate that BO43 may also be active against MAP. Drugs targeted against serine proteases involved in pH IB may be further potentiated by modulating the host response to promote acid and oxidative stress. Studies, such as the one conducted by Anes et al. focusing on how macrophages successfully kill non-pathogenic mycobacteria (Anes et al., 2006) will provide an understanding of critical responses that will inform future host related drug design. Serine proteases may ultimately provide a universal route for drug and vaccine design for pathogenic mycobacteria. Future studies confirming MAP0403 as a serine protease as well as mapping of the active site and identification of the substrate are warranted and findings will support mycobacterial treatment goals.

AUTHOR CONTRIBUTIONS
EL, AK, FS, and EB wrote the manuscript. AK, EL, and SS designed experiments. AK and EL performed experiments. JB provided microarray slides. AK, JB, HJ, and SS analyzed microarray data.

ACKNOWLEDGMENTS
We thank Dr. Raul Barletta (University of Nebraska-Lincoln) and Dr. Luiz E. Bermudez (Oregon State University) for their generous gifts of MAP K-10 (pWes4) and MAC-T cells, respectively. We thank Drs. Marcela Rodriguez and Issar Smith (Public Health Research Institute, Rutgers New Jersey Medical Institute) for providing pSM417 used in M. smegmatis studies. We appreciate Dr. Wayne Xu (University of Manitoba) for his guidance and advice in microarray analysis. We acknowledge services rendered from the University of Minnesota Genomics Center, the Minnesota Supercomputing Institute, and the Bovine Blood Collection at the University of Minnesota.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fcimb. 2016.00085  M. smegmatis mc 2 155 was electroporated with pSM417 containing the ORF of MAP0403. M. smegmatis mc 2 155 pSM417-MAP0403 was (A) infected in MDMs or (B) exposed to acid (pH = 5) in vitro for 0, 10, 30, and 120 min. MAP0403 transcription was evaluated using qT-RT-PCR by comparing test time points (10-120 min) to control time point (0 min). MAP0403 was not differentially expressed, which is likely due to its constitutive expression by the pSM417 Hsp60 promoter. Figure S3 | M. smegmatis mc 2 155 intra-bacterial pH calibration curve. Intra-bacterial pH was measured by 5(6)-carboxyfluroescein N-hydroxysuccinimide fluorescence when exposed to a pH range of 4-8.
Table S1 | Qt-RT-PCR primers used in this study.