Highly Reduced Genome of the New Species Mycobacterium uberis, the Causative Agent of Nodular Thelitis and Tuberculoid Scrotitis in Livestock and a Close Relative of the Leprosy Bacilli

Species name and phylogeny.As shown in Fig. 1, the new species forms a distinct branch lying between M. haemophilum and the most recent common ancestor of M. leprae and Mycobacterium lepromatosis. In view of its initial identification from udder, this species is named Mycobacterium uberis.

• Open in new tab
• Download powerpoint

FIG 1

Phylogenetic tree of Mycobacterium uberis and selected mycobacterial species. The tree was created in MEGA7 from concatenated amino acid sequences (3,696 positions) of 10 proteins (DnaN, RplI, GrpE, MetG, RplY, PheT, FtsQ, HolA, MiaA, and FtsY) (18) and inferred by using the maximum likelihood method based on the JTT matrix-based model. The tree is drawn to scale, with branch lengths measured at the number of substitutions per site. Bootstrap support values, estimated from 500 replicates, are given below each branch. Mycobacteroides abscessus (previously Mycobacterium abscessus) was used as the outgroup.

Mycobacterium uberis genome sequence.DNA was isolated from a skin biopsy of bovine udder with nodular thelitis and subjected to Illumina sequencing. The de novo sequence assembly resulted in 3,571 contigs that were larger than 1 kb and showed average coverage of over 10×. Most contigs matched sequences from a variety of bacterial species. On the basis of sequence similarity to M. haemophilum and M. leprae, we retrieved contigs that unmistakably belonged to M. uberis. Care was taken not to exclude any other potential M. uberis sequences by manually checking all the remaining contigs that displayed GC content similar to that of M. uberis and by repeating the analysis with another assembly program (see Text S1 in the supplemental material for details). The draft M. uberis genome assembly consists of 54 contigs with an average length of 58 kb and totaling 3.12 Mb. All of the contigs harbored genes that closely match those of M. haemophilum (85.7% average nucleotide identity) or M. leprae (82.4% average nucleotide identity), with no obvious outlier that would indicate an erroneous assembly with sequences from another bacterial species.

Genome downsizing and pseudogene formation.At a sequence length of 3.12 Mb and containing 1,081 pseudogenes, the genome of M. uberis is as reduced as that of M. leprae (Table 1). The two species share 1,318 functional protein-coding genes, which corresponds to 75% of the total number of protein-coding genes in M. uberis and 82% in M. leprae. Similarly, among the 1,309 pseudogenes in M. leprae, only 212 orthologs were predicted to be functional in M. uberis; conversely, among the 1,081 pseudogenes in M. uberis, only 126 orthologs are predicted to be functional in M. leprae (see Data Set S1 in the supplemental material).

The pseudogene content of M. uberis differs from that of M. leprae. For example, 465 pseudogenes in M. uberis do not have an ortholog in M. leprae, and 607 pseudogenes in M. leprae do not have an ortholog in M. uberis, suggesting that genome reduction was, at least in part, an independent process in each species. However, this observation does not exclude the possibility that the initial pseudogenization occurred in the ancestor of M. uberis and M. leprae. If this were the case, the pseudogenes that are present in both species would be more likely to share the same deleterious mutations, such as frameshifts and premature stop codons. We manually checked 50 random orthologous pseudogenes and identified only three pairs that shared one or more stop codons and/or frameshifts. Although some signals were lost or blurred by sequence drift, this observation suggests that some of the pseudogenization had already started in the ancestor of M. uberis and M. leprae and that the pseudogenization processes probably continued independently as the two species diverged. Note that roughly 500 genes which are missing in both species (compared to M. haemophilum) were likely deleted in the ancestor of M. uberis and M. leprae, which was probably adapting to a strict intracellular niche (6).

Envelope biogenesis and other specific features.M. leprae contains no methoxy-mycolates, probably because it has lost the MmaA2 and MmaA3 methoxy mycolic acid synthases (7). M. uberis has retained a functional MmaA3 (M. uberis BE_04680 [MUBE_04680]), which might influence the envelope composition. On the other hand, M. uberis has the same reduced set of five mmpL genes as M. leprae. It remains to be determined whether M. uberis produces a glycolipid similar or equivalent to the characteristic and highly antigenic phenolic glycolipid 1 of M. leprae.

A characteristic feature of M. leprae and M. lepromatosis is the presence of the horizontally acquired gene proS, encoding a eukaryote-like prolyl tRNA synthetase, which is both displaced and inverted with respect to the M. tuberculosis genome (7, 8) and is similar to those present in various members of the Nocardiaceae family. The same proS homolog is also present in M. uberis (MUBE_09850) and M. haemophilum (B586_RS07325) at the same genomic location, indicating that the gene was acquired by their ancestor. In addition, M. uberis has a cytochrome P450 (MUBE_02130) of unknown function that is similar to those present in more distantly related mycobacterial species but that is not present in M. leprae, M. lepromatosis, or M. haemophilum.

Growth.All attempts to grow M. uberis have failed, a result which was expected given its highly reduced genome. As in the cases of M. haemophilum and M. leprae, M. uberis lacks the mycobactin synthesis gene cluster present in Mycobacterium tuberculosis. Moreover, the gene coding for 50S ribosomal protein L25 (MUBE_04325) is truncated in M. uberis and is probably not functional. Disruption of this gene results in growth defects in M. tuberculosis (9) and Escherichia coli (10), due to reduced efficiency of the ribosome.

Virulence.The ESX-1 system is the main determinant of virulence in M. tuberculosis and in a number of other mycobacterial pathogens (11). While M. leprae and M. lepromatosis lost some components of ESX-1 (12), M. uberis retained the protein-coding capacity of the entire system (see Fig. S1 in the supplemental material). However, we identified a break in the genomic synteny downstream of espJ (MUBE_00800) and an insertion of a putative proline-proline-glutamate (PPE) gene (MUBE_01185), flanked by remnants of transposases, between espB (MUBE_01195) and eccE₁ (MUBE_01180). It is not clear how these changes impact the ESX-1 system in M. uberis. Curiously, the structural variations occur around the same genes that lost coding capacity in M. leprae and M. lepromatosis (Fig. S1).

The ESX-5 system is the most recently evolved mycobacterial ESX system, which modulates virulence and host response, and is found only in the slow-growing mycobacterial species (13). The two esx genes and the flanking PE/PPE gene pair of the ESX-5 system underwent a series of duplication events that resulted in multiple copies scattered across the genome (14). It was shown that some of the paralog clusters in M. tuberculosis serve as accessory systems that aid in the secretion of a subset of proteins via the prototype ESX-5 system (14). Interestingly, M. uberis lost the core components of the prototype ESX-5 system, similarly to M. leprae, but has retained at least three paralog esx pairs.

Drug susceptibility.No mutations were found in the drug-resistance-conferring regions of RpoB (MUBE_04585), FolP1 (MUBE_01990), GyrA (MUBE_01070), and GyrB (MUBE_01075) (15), indicating that M. uberis is very likely susceptible to the antileprosy drugs rifampin, dapsone, and ofloxacin. Since there are no known molecular markers for resistance, we can only presume that M. uberis is also susceptible to the drugs clofazimine and clarithromycin, as is M. haemophilum (16).

Mycobacterium leprae cluster organisms.A recently reported causative agent of feline leprosy, “Candidatus Mycobacterium lepraefelis,” was found to be a close relative of M. leprae (17). The partial sequence of the groEL2 (hsp65) gene from this pathogen is 89% to 90% identical to those of M. uberis, M. haemophilum, M. leprae, and M. lepromatosis. Phylogenetic reconstruction of the groEL sequence placed “Candidatus M. lepraefelis” between M. uberis and M. leprae (Fig. S2), so it is likely that the genomic structure of “Candidatus M. lepraefelis” resembles those of M. uberis and M. leprae. Efforts to close the genome sequences of the M. leprae-like pathogens are needed to facilitate more-detailed genomic comparisons, which, coupled with biological data, will provide further insights into the evolution and pathogenicity of this particular group of mycobacteria.

Significance and molecular detection of Mycobacterium uberis.The availability of the genome sequence of M. uberis allowed us to design specific PCR primers for M. uberis. We confirmed the presence of the bacteria in the three cases of bovine nodular thelitis and the two cases of caprine nodular thelitis reported before (4, 5), as well as in a new case of bovine nodular thelitis and two cases of caprine nodular thelitis from different farms in France (Table 2). Moreover, we detected M. uberis in three cases of nodular tuberculoid scrotitis, confirming the implication of the bacterium in the two diseases. While definitive evidence is still lacking, these results strongly suggest that M. uberis is the causative agent of nodular thelitis and tuberculoid scrotitis.

Early detection and diagnosis of infectious diseases are crucial in animal husbandry to prevent disease outbreaks and contamination of animal products. Molecular tools from this study can be used for routine screening of the pathogen and will facilitate epidemiological investigations.