Genome analysis reveals three genomospecies in Mycobacterium abscessus

Background: Mycobacterium abscessus complex, the third most frequent mycobacterial complex responsible for community-and health care-associated infections in developed countries, comprises of M. abscessus subsp. abscessus and M. abscessus subsp. bolletii reviously referred as Mycobacterium bolletii and Mycobacterium massiliense. The diversity of this group of opportunistic pathogens is poorly described. Results: In-depth analysis of 14 published M. abscessus complex genomes found a pan-genome of 6,153 proteins and core-genome of 3,947 (64.1%) proteins, indicating a non-conservative genome. Analysing the average percentage of amino-acid sequence identity (from 94.19% to 98.58%) discriminates three main clusters C1, C2 and C3: C1 comprises strains belonging to M. abscessus, C2 comprises strains belonging to M. massiliense and C3 comprises strains belonging to M. bolletii; and two sub-clusters in clusters C2 and C3. The phylogenomic network confirms these three clusters. The genome length (from 4.8 to 5.51-Mb) varies from 5.07-Mb in C1, 4.89-Mb in C2A, 5.01-Mb in C2B and 5.28-Mb in C3. The mean number of prophage regions (from 0 to 7) is 2 in C1; 1.33 in C2A; 3.5 in C2B and five in C3. A total of 36 genes are uniquely present in C1, 15 in C2 and 15 in C3. These genes could be used for the detection and identification of organisms in each cluster. Further, the mean number of host-interaction factors (including PE, PPE, LpqH, MCE, Yrbe and type VII secretion system ESX3 and ESX4) varies from 70 in cluster C1, 80 in cluster C2A, 74 in cluster C2B and 93 in clusters C3A and C3B. No significant differences in antibiotic resistance genes were observed between clusters, in contrast to previously reported in-vitro patterns of drug resistance. They encode both penicillin-binding proteins targeted by beta-lactam antibiotics and an Ambler class A beta-lactamase for which inhibitors exist. Conclusions: Our comparative analysis indicates that M. abscessus complex comprises three genomospecies, corresponding to M. abscessus, M. bolletii, and M. massiliense. The genomics data here reported indicate differences in virulence of medical interest; and suggest targets for the refined detection and identification of M. abscessus.