Bacteria strain maintenance

Mycobacterium marinum M bacterial strains were grown on Middlebrook’s 7H11 (Sigma-Aldrich) agar plates supplemented with 10% OADC [oleic acid (Fischer Scientific), dextrose (Sigma and VWR), albumin (Sigma Aldrich), and catalase(Fischer)] and 0.5% glycerol (VWR, Radnor, PA) at 30°C or in Middlebrook’s 7H9 media (Sigma-Aldrich, St. Louis, MO) supplemented with 10% OADC and 0.2% tyloxapol (Chem-Impex Int’l Inc., Wood Dale, IL) at 30°C. Hygromycin (EMP Millipore) and Kanamycin (IBI Scientific, Peosta, IA) were added to final concentrations of 50 μg/mL and 20 μg/mL, respectively, where noted. Escherichia coli was grown on Luria-Bertani (LB) agar or in LB broth at 37°C. Kanamycin, Hygromycin, and Ampicillin (VWR) were added to final concentrations of 50 μg/mL, 200 μg/mL, and 200 μg/mL when stated. Please see S1 Table for a complete list of bacterial strains used in this study.

Mice

Wild type (Jackson Laboratories, Bar Harbor, ME) and MAVS^-/- [Dr. Stanley Perlman, [11, 47]] C57BL/6 mice were bred and housed at the Freimann Life Science Center at the University of Notre Dame [Animal Welfare Assurance (#A3093-01)]. All animals were cared for in accordance with the Association for Assessment and Accreditation of Laboratory Animal Care International under pathogen-free conditions, and all animal experiments were approved by the University of Notre Dame’s Institutional Animal Care and Use Committee.

Isolation of BMDMs

Bone marrow derived macrophages were isolated from the femurs of 6–8 week old C57BL/6 mice (The Jackson Laboratory) following euthanasia by cervical dislocation. Briefly, bone marrow was collected into a 50 mL conical tube by flushing femur with 1X phosphate buffered saline (PBS; Hyclone, Logan, Utah) + 1% penicillin/streptomycin and pelleted by centrifugation at 1000 rpm for 5 minutes at 4°C. Cells were washed once in 1X PBS with gentle agitation before being pelleted again by centrifugation at 1000 rpm for 5 minutes at 4°C. The pellet was resuspended in sterile ACK solution (2.5 mL/mouse; 150 mM NH₄Cl, 1mM KHCO₃, and 0.1 mM EDTA filter sterilized through a 0.2 μM filter) and incubated for 7–10 minutes to lyse red blood cells. The remaining cell suspension was then pelleted, washed twice in 1xPBS, resuspended in BMDM media, and transferred to 100 x 15 mm petri dishes at a density of 1x10⁷ cells/dish in 10 mL media. Cells were grown for seven days, with one media change on day 3, and then stocked down in DMEM + 20% HI-FBS + 10% DMSO at a density of 1x10⁶ cells/mL.

Mammalian cell culture

Bone marrow derived macrophages (BMDMs) were maintained in Dulbecco’s Modified Eagle Medium (DMEM; Sigma-Aldrich and Sigma) supplemented with 20% L929 cell conditioned media, 10% heat-inactivated fetal bovine serum (HI-FBS; Hyclone and Gibco), and 100 U/mL penicillin/streptomycin (Cytiva, South Logan, UT) at 37°C with 5% CO₂. L929 fibroblasts were maintained in DMEM supplemented with 10% HI-FBS and 100 U/mL penicillin/streptomycin (Cytiva; unless otherwise noted) at 37°C with 5% CO₂.

L929 supplement

L929 fibroblasts were rapidly thawed in a 37°C water bath before being washed once in DMEM and 10% HI-FBS and transferred to a T-25 tissue culture treated flask (Sarstedt, Nümbrecht, Germany) in 10 mL DMEM and 10% HI-FBS with 100 U/mL penicillin/streptomycin. At about 70% confluency, the cells were detached in 0.05% trypsin-EDTA (Gibco, Grand Island, NY) and bumped up to T-75 in 20 mL DMEM and 10% HI-FBS with 100 U/mL penicillin/streptomycin. This process was repeated, bumping the cells into a T-125 flask with 40 mL of the same media. At 70% confluency, the cells were again detached and then split equally between five flasks of the same size. This was repeated once more with each of the five flasks seeding an additional five flasks for a total of 25, T-125 flasks with 40 mL DMEM and 10% HI-FBS (no antibiotics). The cells were then allowed to incubate for 10 days at 37°C and 5% CO₂, before the supernatants were harvested, passed through a 0.2 μM filter, and stored in 50mL aliquots at -80°C until further use.

Generation of electrocompetent M. marinum

Electrocompetent M. marinum were generated from single colonies grown in 5mL 7H9 + 0.1% tween 80 for 3–5 days before being expanded to 25 mLs in the same media. Twenty-four to forty-eight hours later, 2 mL 2M Glycine was added and the cells incubated for twenty-four hours. Cells were then pelleted at 3,500 rpm for 5 minutes, washed three times in comp. cell buffer (distilled water containing 10% glycerol and 0.1% tween 80),resuspended in 5mL comp. cell buffer (1/5th the original culture volume), and frozen in 500 μl aliquots at -80C until further use.

Generation of the ΔsecA2 knockout strain

The ΔsecA2 strain was generated using a previously established allelic exchange approach [48]. Briefly, a 1.5 kb region of DNA upstream and downstream of secA2 M. marinum M genome (NC_010612.1) was amplified by PCR using Phusion polymerase (a complete list of primers and plasmids can be found in S2 and S3 Tables, respectively). Amplicons were introduced into DPNI treated p2NIL vector (Addgene, #20188) by three-part fast cloning [49], transformed into chemically competent DH5ɑ E. coli, and plated on LB agar containing 50 μg/mL kanamycin. Single colonies were picked and grown overnight in LB broth with 50 μg/mL Kanamycin before plasmids were isolated using a QIAprep Spin Mini Kit (Qiagen, Germantown, MD), confirmed by restriction enzyme digest, and quantified by A260 using a Nanodrop 2000 Spectrophotometer (ThermoScientific, Rockford, IL). Verified p2NIL/ΔsecA2 plasmids were digested with PacI (NEB; 37°C for 1 hour, 65°C for 20 minutes), dephosphorylated with antarctic phosphatase (NEB; 37°C for 30 minutes, 80°C for 2 minutes), and ligated with the pGOAL (Addgene, #20190) gene marker cassette by overnight ligation at 4°C with T4 DNA Ligase (NEB). Overnight ligations were transformed into chemically competent DH5ɑ E. coli and selected for on LB agar containing 50 μg/mL Kanamycin and 60 μg/mL X-Gal (5 Prime, Gaithersburg, Maryland). Blue colonies were picked and grown overnight in LB broth containing 50 μg/mL kanamycin before plasmids were isolated, confirmed by restriction enzyme digestion, and quantified as before.

Verified p2NIL/ΔsecA2 GOAL plasmids were transformed into electrocompetent M. marinum M bacteria at room temperature. One 500μl stock of WT electrocompetent cells were thawed at room temperature, pelleted at 13,000rpm for 1 minute, washed three times with glycerol wash buffer (10% glycerol with 0.1% Tween 80), resuspended in 500μl of the same buffer, and placed between the metal plates of a 2 mm electroporation cuvette (VWR). Three micrograms of p2NIL/ΔsecA2 GOAL was UV irradiated (exposed to 1000 μJ/cm2 in a CL-1000 Ultraviolet Crosslinker) was added to the electroporation cuvette. The sample was then electroporated on a BioRad GenePulser Xcell under the following conditions: 2500V voltage, 25μF capacitance, and 1000Ω resistance. Bacteria were allowed to recover overnight in 7H9 + 0.1% Tween80 before being plated on 7H11 agar containing 50 μg/mL Kanamycin, and 60 μg/mL X-Gal and incubated at 30°C for two weeks.

Single blue colonies (merodiploids) were picked and cultured in 7H9 for five days before being counter selected for double crossover events on 7H11 agar containing 50 μg/mL Kanamycin, 60 μg/mL X-Gal, and 2% sucrose and incubated at 30°C for 14 days. White colonies were picked, cultured five days in 7H9 + 0.1% tween 80 and genotyped. Deletion of secA2 was confirmed by Sanger sequencing using primers secA2-X and secA2-Y by the Genomics core at the University of Notre Dame and the resulting sequences analyzed by FinchTV (version 1.5.0). Additionally, strains were checked for the deletion of secA2 using a rapid genotyping process previously described [50].

Generation of complementation strains

M. marinum (MMAR_2698) complementation and M. tuberculosis (Rv1821) cross complementation strains were generated using an integrative vector (pMV306H) under the control of the mycobacterial optimal promoter (MOP; George et al., 1995). To begin, secA2 was amplified from either M. marinum M or M. tuberculosis CDC1551 genomic DNA with primers MMAR_2698 P1 and MMAR_2698 P2 or Rv1821 P1 and Rv1821 P1, respectively, using Phusion High Fidelity polymerase with HF buffer and 3% DMSO (NEB) under the following cycling conditions: 98°C, 30 seconds; 98°C, 10 seconds, 72°C, 30 seconds, 72°C, 30 seconds for 30 cycles; 72°C, 5 minutes. PCR amplicons were checked for size on a 1% agarose gel and imaged using a BioRad GelDoc 1000.

Verified amplicons were combined with the pMV306H integrative vector backbone at a ratio of 1:1 v/v and allowed to ligate overnight at 4°C with T4 DNA ligase (NEB). Ligation products were subsequently transformed into chemically competent DH5ɑ E.coli. Following a 1 hour recovery/out-growth period, transformants were plated onto LB agar containing 200 μg/mL hygromycin. The next day, single colonies were inoculated into LB broth with 200 μg/mL hygromycin and allowed to grow overnight before generated plasmids were purified by mini-prep using a QIAprep Spin Miniprep Kit. Plasmids were confirmed by restriction enzyme digest and concentrations quantified by A260 using a Nanodrop 2000 Spectrophotometer.

Verified plasmids were electroporated into electrocompetent ΔsecA2 M. marinum at room temperature. Electrocompetent bacteria, prepared the same as previously described, were combined with 500 ng of verified plasmid (no UV-irradiation) between the metal plates of a 2mm electroporation cuvette. The samples were electroporated on a BioRad GenePulser Xcell under the following conditions: 2500V voltage, 25μF capacitance, and 1000Ω resistance. Bacteria were allowed to recover overnight in 7H9 + 0.1% Tween80 before being plated on 7H11 agar containing 200 μg/mL Hygromycin and incubated at 30°C for two weeks. Single colonies were picked and cultured in 7H9 for five days before being checked for the complementation using a rapid genotyping process [50].

Western blot

Bacteria were pelleted from mid-log phase cultures at 3,500 rpm for 10 minutes at 4°C. Following centrifugation the supernatants were poured off and the pellets stored at -80°C. These pellets were subsequently thawed on ice, spun down at 3,500 rpm for 10 minutes at 4°C, and any remaining supernatant removed. To generate whole cell lysates, the bacteria were resuspended in 1/10th of the original culture volume of ice cold 1 x Dulbecco’ phosphate buffered saline (DPBS; Hyclone) containing 1 mM phenylmethylsulfonyl fluoride (PMSF) and added to a 2.0 mL O-ring screw cap micro tube (Sarstedt) with 0.1 mm zirconia beads (RPI, Mount Prospect, IL). Bacteria were mechanically lysed by three, thirty-second pulses on a Mini-Beadbeater (BioSpec). The resulting membranes were pelleted by centrifugation at 13,000 rpm for 15 minutes, and the supernatant was transferred to a new tube. The protein concentration for each supernatant was determined using a Pierce BCA Protein Assay Kit (ThermoFischer Scientific).

Mammalian cells were lysed directly in the twenty-four well plates, following the removal of culture supernatants. Monolayers were lysed in 75 μl, ice-cold RIPA lysis buffer (150mM sodium chloride; 50mM Tris-HCl, pH 8.0; 1% Triton X100; 0.5% sodium deoxycholate; 0.1% SDS) containing 1 mM PMSF added directly before use. The resulting whole cell lysates were centrifuged a 12,000 rpm for 5 minutes before quantifying the protein concentration using a Pierce BCA Protein Assay Kit (ThermoFischer Scientific)

Equal amounts of protein [60 μg for bacterial whole cell lysates (WCL); 15 μg for macrophage WCLs) were added to 5X SDS loading dye [250 mM Tris-HCl, pH 6.8 (Fischer Scientific, Fair Lawn, NJ); 10% sodium dodecyl sulfate (SDS; Amresco, Solon, OH); 30% glycerol (VWR); 0.02% Bromophenol blue (Fischer Scientific); and 5% β-Mercaptoethanol (Sigma-Aldrich)] and topped up to a final volume of 40 μl with 1X DPBS before being boiled for 10 minutes and then loaded into lanes of an 8% SDS-PAGE gel. Following electrophoresis, the separated proteins were transferred onto a methanol activated 0.45 μM PVDF membrane (Millipore) at 70V for 120 minutes at 4°C in Transfer Buffer (192 mM Glycine, 25 mM Tris Base, and 10% Methanol). Membranes were blocked in 5% non-fat milk diluted in TBS-T (20 mM Tris Base, 150 mM NaCl, and 0.5% tween 20; pH = 7.6) for one hour at room temperature with gentle agitation. Membranes were then incubated with primary antibody diluted in blocking buffer overnight at 4°C, rocking. The following day, membranes were washed five times in TBS-T for five-minutes before incubation with an HRP-conjugated secondary antibody diluted in blocking buffer at room temperature for 3 hours. Non-specifically and un- bound secondary antibody was removed by washing the membrane five times in TBS-T for 5 minutes. Membranes were coated in SuperSignal West PicoPlus Chemiluminescent or Femto Maximum Sensitivity Substrate (ThermoScientific), exposed to BioBlue-MR Autoradiography film (Alkali Scientific), and imaged on a Konica SRX-101A developer. Please see S4 Table for a list of antibodies utilized in this study.

In vitro growth curve

M. marinum bacterial strains were grown to exponential phase at 30°C in 5 mLs of 7H9 + 10% OADC + 0.2% tyloxapol in 25 mL Erlenmeyer flasks, shaking at 150 rpm. These were then subcultured into triplicate 125 mL Erlenmeyer flasks with 25 mLs of the same media at an optical density of 0.05 using the following equation:

Initial OD readings were measured on an Eppendorf BioPhotometer at the time of subculturing and then once every twenty-four hours for the duration of the experiment. Following data collection, growth curve data were summarized using the R package Growthcurver [51], which fits a logarithmic model [K / (1 + ((K—N0) / N0) * exp(-r * t)] to each curve. Estimated growth rates (r) were compared across bacterial strains to identify changes in growth rates.

SDS sensitivity assay

Previous work in an M. marinum M ΔsecA2 strain showed increased sensitivity to sodium-dodecyl sulfate (SDS), indicating a cell wall defect (Watkins et al., 2012). We examined our ΔsecA2 strain for this same phenotype. WT, ΔsecA2, and ΔsecA2/psecA2_MM were grown in 5 mL 7H9 + 10% OADC + 0.2% tyloxapol for 3–5 days before being bumped up to 25 mLs. Once at mid-log phase, these cultures were subcultured into 5 mL of the same media, 7H9 + 10% OADC + 0.2% tyloxapol + 0.05–0.2% SDS at an OD600 of 0.8, and allowed to grow 24-hours at 30°C, shaking at 150 rpm. Twenty-four hours later, 50μl of each culture was serially diluted in PBS + 0.2% tyloxapol and plated onto 7H11 + 10% OADC agar plates in technical triplicate. Plates were incubated at 32°C with 5% CO₂ for one week before enumeration.

Macrophage infections

Bone marrow derived macrophages (BMDMs) were cultured in BMDM media for seven days, with media changed at 24 and 96 hours. On day seven, macrophage monolayers were washed once with sterile, 1X phosphate buffered saline (Hyclone, Logan, Utah) and dissociated with 0.05% Trypsin EDTA (Gibco). Trypsin was inactivated with the addition of BMDM media and the cells were pelleted by centrifugation at 250 x g for 5 minutes. Cells were then resuspended in fresh BMDM media, enumerated on a 0.1 mm Brightline hemocytometer (Hausser Scientific, Horsham, PA), seeded into wells of 24 or 48-well tissue culture plates and incubated overnight at 37°C with 5% CO₂. For intracellular growth assays, macrophages were seeded into wells of a 48-well tissue culture plate at a density of 1 x 10⁵ cells in 500 μL BMDM media, and at a density of 5 x 10⁵ cells in 1 mL BMDM media in 24-well plates for ELISA based assays. The next morning, macrophages were infected with a single-cell suspension of log-phase bacteria at a multiplicity of infection of 0.2 (1 bacteria for every 5 macrophages) for intracellular growth assays or 1 for ELISA assays.

Prior to infection, bacteria were collected by centrifugation at 3,500 rpm for 10 minutes, washed once with 1X PBS, and then resuspended in 1 mL 1X PBS. Bacterial suspensions were incubated at room temperature for 30 minutes to allow clumps to settle. The top 700 μl of suspension was transferred to a fresh tube and the optical density (OD600) measured. The volume of bacteria needed to infect at a standard MOI for each strain was calculated based on the assumption that at an OD₆₀₀ of 1 there are 7.7 x 10⁷ CFU/mL using the following equation:

Macrophages and bacteria were incubated together for 2 hours at 32°C with 5% CO2 before the supernatant was discarded and monolayers washed three times with 1X DPBS to remove extracellular bacteria. For intracellular growth assays, the 48 and 96 hour time points were further incubated with BMDM media containing 100 ug/mL Gentamycin (VWR) until 6 hours post infection, at which point the supernatant was removed and media replaced with BMDM lacking antibiotics. For intracellular and IFN-β based assays, following the third PBS wash, the 2 hour time point was lysed in 250 (intracellular) or 500 (ELISA) μl sterile water + 0.2% tyloxapol, diluted in 1X DPBS + 0.2% tyloxapol and plated onto 7H11 + 10% OADC agar plates. Plates were incubated for 1 week at 32°C with 5% CO₂ before being counted. At 48 and 96 hours post infection, the culture supernatant was removed prior to lysing, diluting, and plating the macrophage monolayer as described for the 2 hour timepoint. Additionally, media was replaced on the 96 hour timepoint at 48 hours post infection. Plates were incubated at 32°C for one week before colonies were enumerated. Intracellular growth was assessed by fitting a linear regression model to colony forming units counted over time. Estimated growth rates (r) were compared using a one-way ANOVA followed by Dunnett’s multiple comparison test.

ELISA

Macrophages were infected as described under “Bacterial Growth within BMDMs” with a couple exceptions. Macrophage monolayers were infected at an MOI of 1, unless otherwise stated, and following the 2 hour infection, extracellular bacteria were removed by washing macrophage monolayers three times with 1XDPBS (Cytiva/HyClone). Following the third wash, fresh BMDM media (without antibiotics) was added and the plates were placed at 32°C with 5% CO₂. Secreted levels of IFN-β were measured from BMDM supernatants twenty-four hours following the start of the infection following the Sandwich ELISA protocol published by BioLegend. Anti-IFN-β capture antibody (BioLegend, SanDiego, CA) was diluted 1:200 in Carbonate Coating Buffer [0.84% w/v NaHCO₃ (Sigma) and 0.356% w/v Na₂CO₃ (Sigma) dissolved in ddH₂0, pH = 9.5], while the detection antibody (Biotin anti-mouse IFN-β; BioLegend) and Avidin-HRP (BioLegend) were diluted 1:200 and 1:500, respectively, in blocking buffer. Absorbance was measured at a wavelength of 405 nm and secreted levels of IFN-β were calculated from blank corrected data based on absorbance of the IFN-β standard at known concentrations.

IFN-β normalization

Secreted levels of IFN-β were normalized to 10,000 bacteria using the equation a/∑b, where “a” is the amount of IFN-β measured for each technical replicate and “∑b” is the sum of CFUs counted from technical replicates originating from matching wells of the same biological experiment. This number was then multiplied by 10,000 to provide IFN-β values normalized to a standard number of bacteria present in the host cell.

Reverse transcription quantitative PCR

Bacterial whole cell lysates were generated as described previously (see Western Blot methodology), and used as input for RNA purification using the RNeasy Plus Mini Kit (Qiagen). WCL (50μl) was combined with 550 μl RLT buffer containing 1% β-mercaptoethanol (Sigma-Aldrich) before being passed through the included gDNA eliminator column and processed as described by the manufacturer. Whereas, BMDM monolayers were directly lysed in wells of a 24-well plate in 350 μl RLT buffer containing 1% β-mercaptoethanol (Sigma-Aldrich) before passed through a gDNA eliminator column and processed as per the manufacturer’s recommendation. The resulting bacterial and mammalian RNA was assessed for quality and concentration by absorbance at 260nm using a Nanodrop 2000 Spectrophotometer (ThermoScientific). Five-hundred ng to 1 μg of RNA was used as input for DNaseI treatment following the manufacturer’s recommendation (amplification grade; Invitrogen). After DNase I treatment, equal volumes of RNA were used as input for reverse transcription with Superscript III Reverse Transcriptase (Invitrogen), again following the manufacturer’s recommendations. The resulting cDNA was utilized as input for quantitative PCR analysis on a StepOnePlus Real Time PCR System (Applied Biosystems) using the PowerUp SYBR Green Master Mix (Thermo Fisher Scientific). Raw Ct values were quantified using the ΔΔCt method.

Statistical analysis

For each experiment, samples were examined using a Shapiro-Wilk test to identify if data were normally distributed. Normally distributed data were then further analyzed using parametric statistical methods such as a one-way analysis of variance (ANOVA). Data that were non normally distributed were analyzed using non-parametric statistical methods such as a Kruskal Wallis test. Statistical significance was defined as p-values ≤ 0.05 (*** p<0.001, ** p<0.01, * p<0.05, ‘ns’ not significant, p>0.1). All statistical analyses were performed using R version 4.2.2.

Genetic conservation of SecA2 between M. marinum and M. tuberculosis

We sought to explore the conservation of SecA2 and SecA2-dependent pathogenesis (via activation of host cytoplasmic RNA sensors) between M. tuberculosis and M. marinum. To start, we compared the genomic organization and percent protein identity of genes present in the secA2 locus for each species using Mycobrowser and NCBI. We observed that the organization of genes present in the locus of each of these two species is very similar, with only two genes present in M. tuberculosis that are absent in M. marinum. Additionally, the amino acid sequences these genes encode share a percent identity ranging from 78–96%. Notably, SecA2 shares an 88% amino acid identity between the two species (Fig 1 and S1 Fig). This high degree of similarity at the genetic level combined with previous reports of shared function [29, 30, 43–46], provided strong evidence to further explore the conservation of SecA2-dependent Rig-I/MAVS activation in host cells.

Download:

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

Fig 1. Organization and conservation of secA2 genomic locus in M. marinum and M. tuberculosis.

M. tuberculosis amino acid sequences (accession numbers from NC_000962.3) encoding proteins within the same locus as SecA2 were blasted against the M. marinum M genome (taxid: 216594) using NCBI Protein Blast. The percent protein identity for homologous proteins shared between the two mycobacterial species is depicted above each protein pair. Genomic locations, annotations, and gene function descriptions taken from NCBI and Mycobrowser. Fig design adapted from Chirakos et al., 2020 [37].

https://doi.org/10.1371/journal.pone.0281564.g001

Generation and validation of the ΔsecA2, complementation and cross-complementation strains

To test our hypothesis that SecA2 dependent, Rig-I/MAVS activation and subsequent induction of IFN-β secretion by host macrophages is conserved in M. marinum, we generated an in-frame, unmarked deletion of secA2, removing amino acids 7 through 801. We also generated M. marinum complementation (ΔsecA2/psecA2_MM) and M. tuberculosis cross-complementation strains (ΔsecA2/psecA2_MT) to further explore this conservation and ensure observed phenotypes are the result of the deletion of secA2 and not caused by polarity of the deletion on genes downstream. Generated strains were validated by PCR (Fig 2A–2D), qPCR (Fig 2E) and western blot analysis (Fig 2F).

Download:

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

Fig 2. Confirmation of M. marinum ΔsecA2 and complementation strains by PCR, qPCR, and western blot analysis.

M. marinum WCL was used as input for PCR confirmation of the presence of secA2 within the genome (A; secA2_qPCR_F/R; 158 bp), the successful removal of secA2 from it’s endogenous location(B; secA2-X and secA2-Y primers; WT = 2758bp, ΔsecA2 = 370bp), M. marinum secA2 (C; MMAR_2698 P1/P2; 2456 bp) and M. tuberculosis secA2 (D; Rv1821_P1/P2; 2462bp). PCR products were separated on a 1.5% agarose gel and visualized with ethidium bromide on a BioRad GelDoc 1000. M. marinum and M. tuberculosis gDNA were used as positive controls, while nuclease free water served as a negative control where noted. Strains were further characterized by qPCR (E) and Western Blot (F). RNA was isolated from WCLs, reverse transcribed into cDNA, and used to assess the abundance of secA2 relative to housekeeping gene sigA using the ΔΔct method. Statistical significance was determined using a one-way ANOVA followed by a Dunnett’s Multiple Comparison Test relative to WT, with p-values ≤ 0.05 considered significant (*** p<0.001 and ‘nd’ not detected)(E). For western blot analysis, equal amounts of protein (60 μg) were added to lanes of an 8% SDS-PAGE-gel and separated by electrophoresis. Proteins were transferred onto a 0.45 μm, methanol-activated PVDF membrane and blotted for SecA2 and GroEL2 (loading control). The brightness and exposure for each whole blot was adjusted to 1.0 and 0.25 and is representative of three biological replicates. Red arrow indicates SecA2 to distinguish it from lower and higher molecular weight cross-reactive species (F).

https://doi.org/10.1371/journal.pone.0281564.g002

PCR validation was conducted to check for (A) secA2 at its endogenous location, (B) presence/absences of secA2 in the genome, (C) M. marinum (MMAR_2698) and (D) M. tuberculosis (Rv1821) versions of the gene. Our results show the presence of secA2 within the genome of every strain except ΔsecA2 (Fig 2A), as well as the presence of full length secA2 in only the WT and ΔesxBA strains as well as the M. marinum gDNA positive control (top bands of 2758bp). Smaller products (370bp) present in the ΔsecA2, ΔsecA2/psecA2_MM, and ΔsecA2/psecA2_MT indicate the deletion of the gene from it’s endogenous location (Fig 2B). Deletion of secA2 was further confirmed by Sanger sequencing (S2 Fig). PCRs with species specific secA2 primers pairs confirmed the complementation and cross-complementation strain genotypes (Fig 2C–2D and S2 Table for sequences of all primers). Using the same primer pairs as Fig 2A, we additionally validated each strain by RT-qPCR. We observed no detectable secA2 in the ΔsecA2 strain, while the ΔsecA2/psecA2_MM and ΔsecA2/psecA2_MT strains showed transcript levels of 5.18 (+/- 0.52 s.d.) and 7.16 (+/- 0.22 s.d.) relative to the housekeeping gene sigA, respectively (Fig 2E). The higher levels of secA2 transcript abundance observed in the complement and cross-compliment strains are likely due to constitutive expression by the mycobacterial optimal promoter (MOP), rather than its native promoter. Finally, we also validated the strains by western blot analysis. Equal amounts of bacterial whole cell lysate protein was probed for SecA2 using a polyclonal ɑ-SecA2 antibody [52] and GroEL2 (loading control). Our results indicate a roughly 90 kDa protein present in all strains except the ΔsecA2 (Fig 2F), indicated by the red arrow to differentiate it from lower and higher molecular weight species in the M. marinum and M. tuberculosis complementation strains. While we are unsure what these additional protein bands represent, they may be the result of post-translational modifications or degradation products of the SecA2 protein, as they are missing from the ΔsecA2 strain. Taken together, our results confirm the generation of an unmarked ΔsecA2 deletion as well as M. marinum and M. tuberculosis complementation and cross-complementation constructs.

Impact of secA2 on in vitro growth under specific media conditions

With our ΔsecA2, ΔsecA2/psecA2_MM, and ΔsecA2/psecA2_MT strains built and validated, we sought to characterize the impact secA2 has on growth in vitro and compare this to previously published data. We characterized growth in vitro under nutrient rich conditions (7H9 supplemented with 10% OADC and 0.2% tyloxapol) and fit logistic growth curves to the resulting data using the R package Growthcurver (S3 and S4 Figs). Our results illustrate a significant growth defect in the ΔsecA2 strain (p<0.001) that was either completely or partially restored by complementation with either the M. marinum or M. tuberculosis version of the gene (Fig 3, S3, S4A and S4B Figs). While the data represented in Fig 3 are based on growth curves completed with the same batch of OADC, we observed that the degree of complementation was variable across batches of OADC (S5 Fig). Additionally, when grown in minimal media (7H9 + 0.5% glycerol + 0.5% glucose + 0.1% tween 80) a severe aggregation phenotype was observed in the ΔsecA2 strain (S6 Fig). These data are contrary to previous observations in ΔsecA2 M. tuberculosis [28], M. marinum [29] and M. smegmatis [53] where in vitro growth curves in nutrient rich liquid media showed no growth defect. However, slower growth and changes in morphology have been described in these studies under specific media conditions [28, 29, 53].

Download:

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

Fig 3. M. marinum ΔsecA2 strain has attenuated growth in vitro and sensitivity to SDS.

M. marinum strains were grown to exponential phase in 7H9 supplemented with 10% OADC and 0.2% tyloxapol before being subcultured into fresh media at an optical density (OD₆₀₀) of 0.05. Bacterial growth was monitored by measuring the OD₆₀₀ every 24-hours for 7 days and plotted on a log transformed scale (A). The same strains were examined for growth in sodium dodecyl-sulfate (SDS) by subculturing exponential phase bacteria into 7H9 supplemented with 10% OADC and 0.2% tyloxapol with and without 0.2% SDS at an OD600 of 0.8. Twenty-four hours later, these cultures were serially diluted and plated onto 7H11 agar plates supplemented with 10% OADC in technical triplicate. Images are representative of three biological replicates each plated in technical triplicate (B). Colony forming units were quantified from agars plates following incubation at 32°C with 5% CO₂ for one week (C). Statistical significance was determined using a one-way ANOVA followed by a Dunnett’s test for multiple comparison relative to WT, with p-values ≤ 0.05 considered significant (C only).

https://doi.org/10.1371/journal.pone.0281564.g003

Two such observations of M. marinum ΔsecA2-related growth deficiencies involved increased susceptibility to antibiotics specifically targeting the cell wall [43] and sensitivity to SDS [29], both of which suggest a cell wall synthesis defect. To examine if this phenotype was present in our ΔsecA2 strain, we tested it for sensitivity to 0.05–0.2% SDS. In support of previous work by Watkins et al., our ΔsecA2 strain also showed an observable growth defect in SDS (0.052 relative growth +/- 0.037 s.d.; Fig 3B) as compared to growth in the same media lacking SDS. This phenotype was partially or fully rescued to WT levels (1.61 relative growth +/- 1.17 s.d.) by complementation with the M. marinum or M. tuberculosis version of secA2 (1.88 relative growth +/- 0.98 s.d.; Fig 3B and 3C and S7A and S7B Fig). Similar to growth in nutrient rich media, the degree of complementation varied across different media preparations.

SecA2 is not required for the intracellular growth of M. marinum in BMDMs

Upon phagocytosis by alveolar macrophages, pathogenic mycobacteria are placed into a phagosome. To evade degradation by the maturation and acidification of these membranous compartments, mycobacteria have evolved mechanisms to subvert this process. Protein secretion systems including ESX-1 and SecA2 play vital roles in the bacterium’s ability to survive intracellularly. For example, the ESX-1 protein secretion system is responsible for the translocation of substrate(s) that act to permeabilize the phagosomal membrane, providing the bacterium access to the host cytosol, and with it, new opportunities for nutrient acquisition and host manipulation [42, 54–56]. The SecA2-dependent Rig-I/MAVS activation of INF-β signaling and secretion in M. tuberculosis has been shown to be dependent on a functional ESX-1 protein secretion system [11]. Without ESX-1, the bacterially derived RNAs which activate Rig-I/MAVS, are trapped within the phagosomal compartment. For this reason, an ESX-1 deletion strain (ΔesxBA) is included as a control for attenuated intracellular growth and low IFN-β inducing bacteria in all infection based studies.

Removal of secA2 has been shown to attenuate bacterial survival in mouse [46] and zebrafish [29, 43] models infected with M. tuberculosis or M. marinum M or E11, respectively. Intracellular models using primary, mouse bone marrow derived macrophages, reciprocate this data in M. tuberculosis [46] but not for M. marinum [29]. We tested our ΔsecA2 M. marinum strain for intracellular growth in non-activated BMDMs over 96 hours (Fig 4). By fitting linear regression models to colony forming unit counts for each time point, we were able to compare intracellular growth rates across bacteria strains (S8, S9A and S9B Fig). In accordance with data presented by Watkins et al., we observed no attenuation of growth in the ΔsecA2 as compared to the wild type and complemented strains. However, we did observe a 10 fold reduction (0.0042 +/- 0.006 s.d.) in growth rate for the ΔesxBA strain as compared to the WT (0.049 +/- 0.004 s.d.).

Download:

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

Fig 4. ΔsecA2 uptake relative to bacterial input and growth in BMDMs is comparable to WT.

M. marinum strains were grown to exponential phase in 7H9 supplemented with 10% OADC and 0.2% tyloxapol and used to infect triplicate wells of BMDMs, at an MOI of 0.2, for 2 hours. At 2, 48, and 96 hours post infection (hpi), macrophage monolayers were lysed and plated onto 7H11 agar supplemented with 10% OADC in technical triplicate. One week after plating, colony forming units were enumerated for each timepoint (A). Bacterial numbers present at 2hpi were plotted separately (B) to better visualize differences in bacterial counts. Uninfected wells were included as a control for across well contamination (not shown, no contamination observed) and the ΔesxBA strain as a control for attenuated growth. Statistical significance was determined using a Kruskal-Wallis test followed by a Wilcoxon Rank Sum test for pairwise comparison relative to WT, with p-values ≤ 0.05 considered significant (*** p<0.001).

https://doi.org/10.1371/journal.pone.0281564.g004

From this data, we also observed strain-specific differences in the number of bacteria present in host macrophages 2hpi (Fig 4A). When infecting macrophage monolayers, we utilized an established, OD-based calculation that assumes at an OD₆₀₀ = 1, there are approximately 7.7 x 10⁷ bacteria present in one mL of bacterial suspension [57]. Applying this calculation across our five strains, it was notable that the number of colony forming units present in host macrophages infected with ΔsecA2 M. marinum at 2hpi was on average 3 fold lower than those infected with WT bacteria (Fig 4B). When compared to the number of bacteria added to macrophages (bacterial input), our data indicate that the observed reduction in number of bacteria taken up is not likely due to a defect in the macrophage’s ability to phagocytose the bacteria (S10 Fig) but due to an underestimation of the number of bacteria present in suspension at a given optical density.

IFN-β secretion from WT BMDMs is not SecA2-dependent in M. marinum

The results observed in Fig 4 suggested that it is important to take into account the number of intracellular bacteria present in macrophages following infection when comparing levels of secreted IFN-β.Therefore, when examining BMDMs for secreted levels of IFN-β, we infected these cells using the same OD-based calculation as before, and additionally plated macrophage whole cell lysates at the end of the infection period (2hpi). This enabled us to quantify the number of intracellular bacteria following a 2 hour infection (Fig 5A) and secreted levels of IFN-β (Fig 5B) separately. Crucially, it also allowed us to correct for differences in numbers of intracellular bacteria to examine secreted levels of IFN-β based on equal numbers of bacterial colony forming units (Fig 5C).

Download:

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

Fig 5. ΔsecA2 M. marinum induces high levels of IFN-β from WT BMDMs.

To evaluate the IFN-β response from wild type BMDMs when levels of intracellular bacteria are equal, log-phase M. marinum cultures were used to infect wild type BMDMs at an MOI of 1. Following a 2 hour infection, macrophage monolayers were washed three times with PBS before being lysed and plated onto 7H11 + 10% agar plates in technical triplicate (A) or returned to the incubator in fresh media until twenty-four hours post infection. At twenty-four hours, macrophage culture supernatants were removed and examined for secreted levels of IFN-β by ELISA (B). To account for variability in the number of bacteria present at 2hpi, IFN-β levels were normalized to a standard number of bacteria (C). IFN-β Heat killed WT M. marinum was added to macrophages at an MOI of 10 as a control for bacterial cell lysis. Data is representative of 10 biological replicates. Statistical significance was calculated using the non-parametric Kruskal-Wallis test followed by pairwise comparison with a Wilcoxon Rank Sum test relative to WT. Statistical significance was defined as p-values ≤ 0.05; ***p<0.001, **p<0.01, *p<0.05, ns = not significant (p>0.05).

https://doi.org/10.1371/journal.pone.0281564.g005

Using this adapted methodology we observed an average 2.2 fold reduction in bacterial CFUs for all other bacterial strains as compared to WT (p<0.001; Fig 5A). Despite having fewer bacteria present at 2hpi, the ΔsecA2, ΔsecA2/psecA2_MM, ΔsecA2/psecA2_MT induced IFN-β at levels comparable to WT (Fig 5B), while the ΔesxBA strain induced significantly less IFN-β (14.13 pg/mL +/- 12.08 s.d. versus 49.07 pg/mL +/- 20.29 s.d.) (Fig 5B). When normalized to an equal number of bacterial CFUs (10,000) at 2hpi, our results showed that the ΔsecA2 and both complementation strains induced higher levels of IFN-β from host macrophages as compared to WT M. marinum (Fig 5C). It additionally confirmed previous findings showing the ΔesxBA as a low IFN-β inducing control following normalization. As the removal of secA2 impacts cell wall integrity, we included heat killed wild type M. marinum to control for IFN-β secretion induced by lysed mycobacterial components and saw levels of secreted IFN-β (1.71 pg/mL +/- 2.14 s.d.) comparable to uninfected controls (8.93 +/- 5.57 s.d. pg/mL). We additionally assessed the impact increasing numbers of ΔsecA2 M. marinum (with MOIs ranging from 1 to 10) had on secreted levels of IFN-β (S11A–S11C Fig). Higher CFU counts at 2hpi induced a strong linear increase in the IFN-β in response for each biological replicate, although with variation across replicates in terms of the magnitude of the response elicited (slope of the line; S12 Fig). Notably, when normalized, ΔsecA2 M. marinum induced an average amount of IFN-β around 7 times higher than WT for all MOIs tested (S11C Fig).

As the kinetics of an M. marinum infection are faster than that of M. tuberculosis, with a faster doubling rate and quicker access to the host cytosol, we also examined the host response to M. marinum infection earlier in the infection by quantitative PCR (S13 Fig). We assessed BMDMs infected with out panel of M. marinum strains for relative levels of transcript abundance for IFN-β, the transcription factor Irf7, and upstream signalling molecule Rig-I relative to the housekeeping gene GAPDH at 8 hours post infection. Results from this assay showed no difference in transcript abundance for any of the genes tested across all strains except ΔesxBA, where there was a significant reduction in transcript abundance for all three genes as compared to WT (S13 Fig).

IFN-β secretion from MAVS^-/- BMDMs is not SecA2-dependent in M. marinum

Following activation of Rig-I by immunostimulatory RNAs (such as those derived from bacteria and viruses), Rig-I undergoes a conformational change that enables its interaction with MAVS. MAVS is an essential adapter protein that relays this activation signal onwards, ultimately inducing the transcription of high levels of type I interferons [27, 58]. While this response was originally attributed to defense against viruses, it is increasingly recognized for its complex role in bacterial infections [1, 6, 59–61]. In M. tuberculosis, for example, it has been shown that MAVS^-/- deficient mice are better able to control bacterial proliferation in the lungs six weeks after aerosol challenge with wild type M. tuberculosis [11].

We therefore sought to examine the contribution of this signaling pathway to the INF-β response in M. marinum and explore how this IFN-β response is impacted by the presence or absence of a functional SecA2-protein secretion system using the same experimental model utilized in Fig 5 but with MAVS^-/- BMDMs. Similar to our observations with WT BMDM, we observed an approximately 2.5 fold reduction in bacterial CFUs 2 hpi across all bacterial strains compared to WT (Fig 6A). Despite fewer bacteria measured at 2hpi in the ΔsecA2, ΔsecA2/psecA2_MM, and ΔsecA2/psecA2_MT strains, we observed levels of secreted IFN-β comparable to WT (Fig 6B). While the ΔesxBA strain had CFU numbers comparable to ΔsecA2 and its complementation strains, we measured significantly less secreted IFN-β by ELISA (15.1 pg/mL +/- 6.7 versus 56.2 pg/mL +/- 14.4 for WT and 51.2 pg/mL +/- 28.1 for ΔsecA2).

Download:

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

Fig 6. ΔsecA2 M. marinum induces high levels of IFN-β from MAVS^-/- BMDMs.

To evaluate the contribution of the MAVS intracellular RNA signaling pathway in the IFN-β response and examine whether this response is dependent on a functional SecA2 protein secretion system, log-phase M. marinum cultures were used to infect MAVS^-/- BMDMs at an MOI of 1. Following a 2 hour infection, macrophage monolayers were washed three times with PBS before being lysed and plated onto 7H11 + 10% agar plates in technical triplicate (A) or returned to the incubator in fresh media until twenty-four hours post infection. At twenty-four hours, macrophage culture supernatants were removed and examined for secreted levels of IFN-β by ELISA (B). To account for variability in the number of bacteria present at 2hpi, IFN-β levels were normalized to a standard number of bacteria (C). Heat killed WT M. marinum was added to macrophages at an MOI of 10 as a control for bacterial cell lysis. Data is representative of 5 biological replicates. Statistical significance was calculated using the non-parametric Kruskal-Wallis test followed by pairwise comparison with a Wilcoxon Rank Sum test relative to WT. Statistical significance was defined as p-values ≤ 0.05; ***p<0.001, **p<0.01, *p<0.05, ns = not significant (p>0.05).

https://doi.org/10.1371/journal.pone.0281564.g006

When normalized to 10,000 bacterial CFUs at 2hpi, the ΔsecA2, complementation and cross-complementation strains were shown to induce INF-β secretion from host MAVS^-/- macrophages to concentrations equal to or greater than levels than WT (Fig 6C). Consistent with previous observations, the ΔesxBA strain induced significantly lower levels of secreted IFN-β from macrophages infected with WT M. marinum (p<0.01)(10,20). To ensure the results reported above were representative of a MAVS^-/- BMDM model, we additionally probed macrophage WCLs, derived from the same infection experiments, for MAVS by western blot (S14 Fig). MAVS was present in WT but not in the MAVS^-/- BMDMs, confirming the genotype of each cell type used.