Media

All mycobacterial strains were grown in Middlebrook 7H9 broth (Difco Laboratories, Detroit, MI) with 0.2% glycerol, 0.1% Tween 80, and 10% albumin-dextrose-catalase (Becton, Dickinson and Co., Sparks, MD) with rotation at 37 °C. For colony isolation, the bacteria were plated on Middlebrook 7H10 medium supplemented with 10% oleic acid-albumin-dextrose-catalase (Becton, Dickinson and Co.). To grow MAP, 0.1% mycobactin J (Allied Monitor, Fayette MO) was added to liquid and solid media. The antibiotics kanamycin (50 µg/mL), streptomycin (50 µg/mL), or PANTA antibiotic mixture (Becton, Dickinson and Co.) were included in the media when required.

Generation of streptomycin-resistant strains

To generate streptomycin-resistant (strep-R) strains of MAP K10, MAH 104, M. bovis Ravanel and M. orygis 51145, a K43R mutation was introduced into the rpsL gene of each species using oligo-mediated recombineering as previously described with minor modifications [23]. This mutation is known to confer resistance to streptomycin [24]. In brief, the pNit::ET plasmid (from Kenan Murphy – Addgene plasmid #107692) was first introduced into each strain and selected on kanamycin-containing media. The plasmid was confirmed by PCR of the kanamycin cassette (Additional file 1: Table S1). To introduce the K43R mutation in MAP, M. bovis and M. orygis, 30 mL cultures containing pNit::ET were grown to log phase in 7H9 media with kanamycin and then diluted to an OD600 of 0.1. When cultures reached an OD600 of 0.8, 30 µL of a 1000X stock of isovaleronitrile was added to stimulate the pNit::ET plasmid. When cultures reached an OD600 of 1.0, 3 mL of 2 M glycine was added to the culture. The next day, the cultures were washed 3 times in 10% glycerol. The cells were resuspended in 1 mL 10% glycerol and 200 µL of cells were transferred to a 2 mm gap electroporation cuvette (Fisherbrand, Waltham, MA) containing 1 µg of a 70-mer oligo designed to introduce the K43R mutation (Additional file 1: Table S1). The cells were electroporated using the following settings: 2.5 kV, 1,000 Ω, and 25 µF. The cells were then transferred to 3 mL of 7H9 and incubated at 37 °C. After 5 days (MAP) or 2 days (M. bovis, M. orygis), the cultures were recovered on 7H10 agar with streptomycin.

To generate the rpsL mutation in MAH, the same procedure was followed except for the following modifications: the starting culture was 50 mL, 50 µL of isovaleronitrile was used in stimulation, 7.5 mL of 2 M glycine was added prior to electroporation, washes were performed with pre-warmed 10% glycerol with 2 M sucrose, cells were resuspended in 800 µL 7H9 with 2 M sucrose prior to recovery, and cells were recovered after 1 day.

Colonies which grew on streptomycin were grown up and screened for the K43R mutation by PCR of the rpsL gene followed by Sanger sequencing (Additional file 1: Table S1). Sequences were visualized using Geneious Prime (version 2022.1.1).

Whole genome sequencing and analysis

Genomic DNA was extracted from strep-R MAP and the parental strain using the QIAamp UCP pathogen mini kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions. Paired end sequencing libraries were prepared using the S4 reagent kit (Illumina, San Diego, CA) and shotgun sequencing was performed using the NovaSeq 6000 S4 PE150–35 M reads (Illumina). The sequence was aligned to the MAP K10 reference genome (NC_002944.2) using BWA-MEM [25]. The reads were sorted using SAMtools and visualized using Integrative Genomics Viewer (IGV) [26, 27].

The strep-R MAP sequence was analyzed for deletions and for SNPs outside of the engineered K43R mutation in rpsL. Duplicate reads in sorted BAM files were removed using Picard (http://broadinstitute.github.io/picard). Variant calling was done using Freebayes v1.3.6 with mapping quality 60, minimum read coverage 10, and minimum allele frequency of 0.5 [28]. Variant calls were annotated using SNPEff v.4.3 [29]. Variants identified in strep-R MAP were compared with the parental strain to identify unique variants.

pH experiment

The gastric pH of a mouse is between 3.0 (fed) and 4.0 (after fasting) [30]. To quantify the amount of MAP potentially lost due to the acidic environment of the stomach, strep-R MAP was grown to an OD600 of 1.0 (~ 2 × 108 colony-forming units (CFU)/mL) in 7H9 media. The culture was then split and resuspended in PBS with 0.1% Tween-80 at either a pH of 7.0 or 3.0 and incubated at 37 °C for 1, 2, or 4 h. At the indicated timepoint, the cultures were spun down and resuspended in 7H9 then serially diluted and plated on 7H10 agar.

Animals

Mice were housed in a pathogen-free environment at the Research Institute of the McGill University Health Centre (RI-MUHC). All animal experiments were in accordance with the guidelines of the Canadian Council on Animal Care (CCAC) and all protocols were approved by the animal resource division of the RI-MUHC. C57BL/6 and BALB/c mice were purchased from The Jackson Laboratory (Bar Harbor, ME) at 7-weeks of age and quarantined in-house for 1 week prior to infections. All mouse infections were sex and age matched.

MAP mouse infection model

Infection outcomes were compared using a strep-R MAP inoculum of 108 or 109 CFU. Bacterial inoculum was prepared by growing strep-R MAP to an OD600 of 0.5 (~ 1 × 108 CFU/mL) in 7H9 with streptomycin, washing the cells once in PBS with 0.01% Tween-80, and then resuspending the bacteria in either 1/5 or 1/50 the original volume in PBS with 0.01% Tween-80 to generate a stock of 5 × 108 CFU/mL or 5 × 109 CFU/mL. Prior to all gavage steps, mice were fasted of food and water for 3 h. The first day, mice were given 200 µL of streptomycin (100 mg/mL) by oral gavage. Twenty-four hours later mice were orally inoculated with 200 µL of the prepared strep-R MAP stock (equivalent to a dose of 108 or 109 CFUs). Mice given the higher dose were given a second 109 CFU strep-R MAP infection by oral gavage the following day. When indicated, mice were given 100 µL of 3% sodium bicarbonate 30 minutes [31] prior to oral gavage of strep-R MAP. At the indicated timepoints, mice were sacrificed, and the small intestine, large intestine, MLNs, spleen, liver, and lungs were taken for quantification of organ CFUs. The small intestine, large intestine, and MLNs were also sent for histopathology assessment.

MAH murine infection model

A 5 × 109 CFU/mL bacterial stock of strep-R MAH was prepared by growing the bacteria to an OD600 of 0.5 (~ 1 × 108 CFU/mL). The culture was washed once in PBS with 0.01% Tween-80 and then resuspended in 1/50 of the original volume in PBS with 0.01% Tween-80. Mice were fasted of food and water for 3 h prior to all oral gavage steps. Mice were given 200 µL of streptomycin (100 mg/mL) followed by two consecutive doses of 200 µL of strep-R MAH (5 × 109 CFU/mL) at 24-hour intervals. Mice infected with strep-R MAH were euthanized 48-hours, 4-, 8-, 12-, and 24-weeks post-infection and the small intestine, large intestine, and MLNs were assessed for CFUs.

M. bovis and M. orygis mouse infection model

All M. bovis and M. orygis experiments took place in the containment level 3 facilities at the RI-MUHC as a fully virulent M. bovis strain (M. bovis Ravanel) was used rather than an M. bovis BCG vaccine strain. M. orygis is a cause of zoonotic TB in people in or migrating from South Asia [21], and a clinical isolate collected in Canada (M. orygis 51145) was used [32]. A 5 × 109 CFU/mL bacterial stock of strep-R M. bovis/M. orygis was prepared by first growing strep-R M. bovis/M. orygis in 7H9 to an OD600 of 0.5 (~ 5 × 107 CFU/mL). The culture was washed once in PBS with 0.01% Tween-80 and then resuspended in 1/100 of the original volume in PBS with 0.01% Tween-80. Mice were fasted of food and water for 3 h prior to all oral gavage steps. Mice were given either 200 µL of streptomycin (100 mg/mL) or no pre-treatment followed by two consecutive doses of strep-R M. bovis/M. orygis spaced 24-hours apart. Due to the recognized virulence of M. bovis/M. orygis in mice [33, 34], infected mice were weighed and monitored for survival throughout the 24-week experiment. Mice were euthanized at 4- and 24- weeks post-infection and the small intestine, large intestine, MLNs, and lungs were assessed for CFUs and histopathology.

Organ CFU quantification

At each timepoint, the small and large intestines were excised, separated, and cut open longitudinally. The fecal matter and mucus were removed mechanically using the flat end of curved tweezers, and the organs were washed 3 times in PBS with 0.01% Tween-80 before being placed in 1 mL 7H9. The intestines were processed in this way to identify MAP CFUs that had invaded into the organ tissue and to avoid CFUs passing through the intestines as a result of the inoculation at the earlier timepoints. The spleens, livers, and lungs were placed directly in 1 mL 7H9. The organs were then homogenized with an Omni Tissue Homogenizer TH (Omni International, Kennesaw, GA) for 45 s. Finally, the MLN chain was directly pushed through a 70 μm sterile cell strainer (Fisher Scientific, Waltham, MA) into 1 mL 7H9. The resulting homogenates for all collected organs were serially diluted in 7H9 and plated on 7H10 containing PANTA to quantify CFUs. If no colonies were counted on the lowest dilutions plated, the organ was assigned the limit of detection (LOD).

F57 real-time PCR of fecal pellets

DNA was extracted from fecal pellets of C57BL/6 and BALB/c mice that were uninfected or 12-weeks post-MAP infection using the QIAmp PowerFecal Pro DNA kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The MAP K10 genome is 4,829,781 base pairs which corresponds to ~ 5.3 fg per copy [35]. A standard curve for qPCR was prepared from 1 × 107 genome equivalents (ge) (53 ng) to 1 ge (5.3 fg). A Maxima SYBR Green/ROX real-time PCR assay (Thermo Fisher Scientific, Waltham, MA) was performed using primers for the single copy gene F57 (Additional file 1: Table S1) following the manufacturer’s instructions.

Histopathology

Small intestines and large intestines from all infected and uninfected control mice, mesentery from MAP-infected and uninfected mice, and lungs from M. bovis-infected, M. orygis-infected, and uninfected mice were sent to the histology core at McGill University. The organs were paraffin-embedded and 4 μm sections were cut and stained by hematoxylin and eosin (H&E) staining. Positive control slides of enteritis were generously provided from Dr. Laura Sly’s lab who has a model of spontaneous enteritis in SHIP−/− mice [36]. All slides were reviewed by a pathologist at the MUHC. All slides were photographed with a Nikon Eclipse NI microscope.

Lipocalin-2 ELISA

Fecal pellets were collected from infected and uninfected control mice and stored at −80 °C until processed. Fecal pellets were weighed and placed in a 1.5 mL screwcap tube. PBS with 0.1% Tween-20 was added at a volume of 10 μl per 1 mg of feces. The tubes were vortexed at maximum speed for 20 min and then spun down in a microcentrifuge at 12,000 x g for 10 min at 4 °C. The supernatant was transferred into a new screwcap tube and stored at −20 °C until ready for ELISA of lipocalin-2, a broad marker of intestinal inflammation in mice [37]. The assay was performed using the mouse lipocalin-2/NGAL DuoSet ELISA (Bio-techne, Minneapolis, MO) and DuoSet ELISA ancillary reagent kit (Bio-techne) according to the manufacturer’s instructions.

Statistics

Statistical analyses were conducted with GraphPad Prism (version 9.3.1). Grouped data are graphed as individual datapoints with the sample median. Multiple groups comparisons were performed using a two-way ANOVA with Sidak’s multiple comparisons test. To compare 2 groups, an unpaired t-test was used. Analysis of pooled organ data was performed using a Mann-Whitney test.