Mycobacterium bovis Strain Ravenel Is Attenuated in Cattle

Abstract

Mycobacterium tuberculosis variant bovis (MBO) has one of the widest known mammalian host ranges, including humans. Despite the characterization of this pathogen in the 1800s and whole genome sequencing of a UK strain (AF2122) nearly two decades ago, the basis of its host specificity and pathogenicity remains poorly understood. Recent experimental calf infection studies show that MBO strain Ravenel (MBO Ravenel) is attenuated in the cattle host compared to other pathogenic strains of MBO. In the present study, experimental infections were performed to define attenuation. Whole genome sequencing was completed to identify regions of differences (RD) and single nucleotide polymorphisms (SNPs) to explain the observed attenuation. Comparative genomic analysis of MBO Ravenel against three pathogenic strains of MBO (strains AF2122-97, 10-7428, and 95-1315) was performed. Experimental infection studies on five calves each, with either MBO Ravenel or 95-1315, revealed no visible lesions in all five animals in the Ravenel group despite robust IFN-γ responses. Out of 486 polymorphisms in the present analysis, 173 were unique to MBO Ravenel among the strains compared. A high-confidence subset of nine unique SNPs were missense mutations in genes with annotated functions impacting two major MBO survival and virulence pathways: (1) Cell wall synthesis & transport [espH (A103T), mmpL8 (V888I), aftB (H484Y), eccC₅ (T507M), rpfB (E263G)], and (2) Lipid metabolism & respiration [mycP₁(T125I), pks5 (G455S), fadD29 (N231S), fadE29 (V360G)]. These substitutions likely contribute to the observed attenuation. Results from experimental calf infections and the functional attributions of polymorphic loci on the genome of MBO Ravenel provide new insights into the strain's genotype-disease phenotype associations.

Keywords: Mycobacterium bovis; SNPs; attenuation; bovine TB; pathogenomics; strain Ravenel; tuberculosis.

Figure 1

Interferon-γ (panel (A)) and delayed type hypersensitivity (DTH, panel (B)) responses upon aerosol M. bovis Ravenel (n = 3) infection. Panel A. Whole blood cultures were stimulated with 20 mg/mL M. bovis PPD (Prionics AG) or medium alone (no stimulation) for 48 h. Stimulated plasma were harvested, and IFN-γ concentrations were determined by ELISA (Bovigam, Prionics, Ag). Values represent mean (±standard error) responses to antigen minus the response to media alone (D OD) for M. bovis Ravenel-infected cattle. Response kinetics to rESAT-6/CFP10 were similar to M. bovis PPD responses (data not shown). * Differs (p < 0.05) from pre-challenge (Day 0) responses. Panel B. Approximately 2 months after challenge, 0.1 mL (100 mg) of M. bovis PPD and 0.1 mL (40 mg) of M. avium PPD were injected intradermally at separate clipped sites in the mid-cervical region of each calf. Values represent mean (±standard error) change in skin thickness (i.e., 72 h post-injection minus pre-injection).

Figure 2

Interferon-γ responses upon M. bovis infection of cattle. Whole blood cultures were stimulated with 1 mg/mL rESAT-6:CFP-10 (panel (A)), 20 mg/mL M. bovis PPD (panel (B)), or medium alone (no stimulation) at 39 °C/5% CO₂ for 20 h. Stimulated plasma were harvested, and IFN-γ concentrations were determined by ELISA (Bovigam, Prionics, Ag). Values represent mean (±standard error) responses to antigen minus the response to media alone (D OD) for non-infected cattle (controls, closed squares) or cattle infected with M. bovis strain Ravenel (open triangles) or strain 95-1315 (closed inverted triangles). All responses elicited after challenge with M. bovis (10⁵ CFU MBO Ravenel by aerosol either strain) exceeded (p < 0.05) respective responses in non-infected (control) cattle. Responses elicited by M. bovis Ravenel infection did not differ (p > 0.05) from responses elicited by M. bovis 95-1315.

Figure 3

Delayed type hypersensitivity responses upon aerosol M. bovis infection. Approximately 3.25 months after challenge, 0.1 mL (100 mg) of M. bovis PPD and 0.1 mL (40 mg) of M. avium PPD were injected intradermally at separate clipped sites in the mid-cervical region of each calf according to USDA, APHIS uniform methods (APHIS 91-45-011). Values represent mean (±standard error) change in skin thickness (i.e., 72 h post injection minus pre-injection). Differences between treatment groups are indicated on each graph.

Figure 4

Genome-wide comparisons of Ravenel and 95-1315. The outermost circle is the scale in kilobase pairs. The 1st two rings depict coding regions in the positive and negative strands, respectively. The 3rd ring depicts the predicted subcellular localization of each protein, with blue = cell wall, green = cytoplasmic, black = cytoplasmic membrane, yellow = extracellular, and grey = unknown. Rings 4–6 depict SNPs that are shared by both strains Ravenel and 95-1315 (Ring 4), unique to strain 95-1315 (Ring 5), and unique to strain Ravenel (Ring 6). Ring 7 in yellow depicts SNP density across the M. bovis genome in 10-kb increments. Ring 8 in red depicts G+C content in a 3-kb sliding window.

Figure 5

Position of 32 high-confidence missense polymorphisms in the Ravenel assembly. The Ravenel genome assembly (GCA_018305025.1) was downloaded, and contigs longer than 1000 bp were visualized using Circleator. Working inwards, the rings represent: red, genome %GC; black, genes-fwd; black, genes-rev; blue, total unique polymorphisms (n = 486); central blue ring with labeled loci, 32 high-confidence missense polymorphisms in genes with annotated function. Of these polymorphisms, 31 are SNPs, and 1 is a multi-nucleotide polymorphism (lipN). The manuscript discusses nine SNPs in this innermost subset in greater detail.