Characterization of Mycobacterium orygis as M. tuberculosis Complex Subspecies

The oryx bacilli are Mycobacterium tuberculosis complex organisms for which phylogenetic position and host range are unsettled. We characterized 22 isolates by molecular methods and propose elevation to subspecies status as M. orygis. M. orygis is a causative agent of tuberculosis in animals and humans from Africa and South Asia.

. Two other distinct branches of the M. tuberculosis complex phylogenetic tree exist, the dassie and oryx bacilli, causative agents of tuberculosis in the animal species after which they are named. Neither has been validly described as separate taxa, nor have they been associated with disease in humans (1)(2)(3)(4).
Oryx bacilli have been isolated from members of the Bovidae family, i.e., oryxes, gazelles (3), deer, antelope, and waterbucks (5), although their exact host range remains unsettled. No human disease caused by the oryx bacilli has been reported. These bacilli most likely constitute a separate phylogenetic lineage; however, their exact position has not been established with valid phylogenetic markers, such as large genomic deletions or single nucleotide polymorphisms (SNPs). To settle the phylogenetic position and host range of the oryx bacilli, we collected all oryx bacillus isolates from our laboratory database to establish their sources and subjected the isolates to extended phylogenetic analysis.

The Study
We selected 22 isolates on the basis of >90% similarity of the IS6110 restriction fragment-length polymorphism (RFLP) pattern to that of established and previously published oryx bacillus strains; 11 isolates originated from animals, and 11 originated from 10 human patients ( Figure  1) (1-3). All isolates yielded smooth to greasy domed nonchromogenic colonies in culture (online Technical Appendix Figure, wwwnc.cdc.gov/EID/pdfs/11-0888-Techapp.pdf).
For phylogenetic analysis, we performed SNP and region of difference (RD) analysis (2,6). RD and SNP typing showed a consistent pattern among the isolates, with presence of regions RD1, RD2, RD4, RD5a (Rv2348), RD6, and RD13-RD16 and absence of regions RD3, RD5b (plcA), and RD7-RD12 (online Technical Appendix Table  1). The deleted region for RD12 (RD12 oryx ) was larger than that for M. bovis and M. caprae. Analysis of the fl anking regions indicated an IS6110 insertion at the M. tuberculosis  Table 2). Isolates also showed the RDoryx_1, RDoryx_4, and RDoryx_wag22 deletions and the mmpL6 551 AAG mutation (online Technical Appendix Table 1). Results agreed with those from previous studies (1,6).
Using pncA-1F 5′-GGC CGC GAT GAC ACC TCT-3′, pncA1-R 5′-GCC GCA GCC AAT TCA GCA GT-3′, pncA-2F 5′-CGA AGC GGC GGA CTA CCA TCA CG-3′, and pncA-2R 5′-CCC CAC CTG CGG CTG CGA ACC-3′ primers, we partially sequenced Rv2042c, Rv2044c, and the full pncA gene. The pncA sequences of the isolates from animals and humans were identical to those of M. tuberculosis H37Rv; in codon 38 of the Rv2042c gene, directly upstream from pncA, a GTC to GGC (Ser→Ala) mutation was noted in all 22 isolates; the partial Rv2042c sequence is stored in GenBank (accession no. JF417976). To assess the specifi city of the Rv2042 38 GGC mutation, we screened 2 isolates of all M. tuberculosis complex (sub) species and 2 isolates of all M. tuberculosis groupings, on the basis of >60% IS6110 similarity, for this mutation; we did not fi nd it in any of the strains tested (data not shown).
Baseline clinical data of humans were extracted from the anonymized National Tuberculosis Register. Ethical approval was waived for this retrospective laboratorybased study. Nine of the 10 human patients were of South Asian origin; the other was of Southeast Asian origin ( Figure 1); patients' average age was 41 years (range: 0-69 years). Clinically, 6 patients had pulmonary tuberculosis, 3 had lymphadenitis, and 1 child had tuberculosis diagnosed by gastric fl uid culture. All isolates were susceptible to all fi rst-line antituberculosis drugs, including pyrazinamide, and hence the standard treatment regimen was started for all patients. Patients received treatment for an average of 9 months; no details about individual regimens were available. No bacteriologically proven relapses were noted. No information was available about contact-tracing studies.

Conclusions
The oryx bacillus is a phylogenetically distinct lineage of the clonal M. tuberculosis complex and thus deserves a separate subspecies status; we propose the name M. orygis (Latin: oryx, genitive: orygis, of the oryx) to convey that this subspecies was fi rst characterized after its isolation from an oryx (Figure 2).
The animal-adapted M. tuberculosis complex lineage is thought to have evolved in Africa when an M. africanumlike clone diverged from M. tuberculosis, as shown by the loss of the RD9 locus. Consecutive loss of DNA during the adaptation to novel hosts led to the distinct subspecies with its distinct host range that we know today (1,4,5,11). This matches geographically with the habitats of Oryx species, gazelles, and waterbucks.
For M. orygis, the host range remains unknown but may include oryxes, waterbucks, and gazelles in eastern Africa and the Arabian Peninsula; cows and rhesus monkeys in South Asia; and humans. The evolutionary explanation for the diversity in geographic distribution and hosts of M. orygis remains elusive. This diversity contrasts starkly with the conserved VNTR and spoligotype patterns.
The presence of M. orygis in diseased cows and a monkey in Bangladesh, unique RFLP patterns, and lack of onward transmission suggest animal-to-human transmission. As for M. bovis; humans may be accidental, dead-end hosts.
Molecular characteristics defi ne the isolates previously labeled as oryx bacilli as a distinct subspecies in the M. tuberculosis complex for which we propose the name M. orygis. The Rv2042 38 GGC mutation is a novel, useful genetic marker to identify M. orygis, which is otherwise characterized by the presence of genomic regions RD1, Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 18, No. 4, April 2012 Figure 1. Spoligotyping and 24-locus variable number tandem repeat (VNTR) typing results for Mycobacterium orygis. A) Spoligotyping patterns for the oryx bacillus isolates in this study; ST587 is the most common pattern (labeled SB0422 at www.mbovis.org), with minor deviations. B) Minimum spanning tree based on 24-locus VNTR typing results for the oryx bacillus isolates in this study. One type dominates, with few strains representing minor variations. The dominant clone includes isolates from humans and animals. P, pulmonary; LN, lymph node; G, gastric juice. Both panels were created by using BioNumerics version 6.1 software (Applied Maths, Sint-Martens-Latem, Belgium); similarity coeffi cients were calculated by using Dice (spoligotyping) and Pearson (VNTR) methods; cluster analysis was done by the UPGMA (unweighted pair group method with arithmetic mean). Isolate 10-1562 (cow, Bangladesh) could not be included in (B) because of insuffi cient DNA.