Genetic variation and evolutionary origin of the direct repeat locus of Mycobacterium tuberculosis complex bacteria

Abstract

The direct repeat region in Mycobacterium tuberculosis complex strains is composed of multiple direct variant repeats (DVRs), each of which is composed of a 36-bp direct repeat (DR) plus a nonrepetitive spacer sequence of similar size. It has been shown previously that clinical isolates show extensive polymorphism in the DR region by the variable presence of DVRs, and this polymorphism has been used in the epidemiology of tuberculosis. In an attempt to better understand the evolutionary scenario leading to polymorphic DR loci and to improve strain differentiation by spoligotyping, we characterized and compared the DNA sequences of the complete DR region and its flanking DNA of M. tuberculosis complex strains. We identified 94 different spacer sequences among 26 M. tuberculosis complex strains. No sequence homology was found between any of these spacers and M. tuberculosis DNA outside of the DR region or with any other known bacterial sequence. Although strains differed extensively in the presence or absence of DVRs, the order of the spacers in the DR locus was found to be well conserved. The data strongly suggest that the polymorphism in clinical isolates is the result of successive deletions of single discrete DVRs or of multiple contiguous DVRs from a primordial DR region containing many more DVRs than seen in present day isolates and that virtually no scrambling of DVRs took place during evolution. Because the majority of the novel spacer sequences identified in this study were confined to isolates of the rare Mycobacterium canettii taxon, the use of the novel spacers in spoligotyping led only to a slight improvement of strain differentiation by spoligotyping.

FIG. 1

Genetic organization of the DR locus in 26 M. tuberculosis complex strains as follows: 1 to 19, M. tuberculosis; 20 to 23, M. bovis; 24, intermediate phenotype between M. tuberculosis and M. bovis; 25, M. microti; and 26, M. canettii. The rectangles depict individual DVRs, which are composed of a DR and the adjacent spacer. Except for the M. canettii strain, vertically aligned rectangles represent DVRs with identical spacers. The sequence of the DR part of the rectangles is identical, except for those in gray. These differ in one or a few nucleotides from the consensus sequence. The hatched spacer in strain 2 differs in a single nucleotide from that in strain 1. The numbers at the top correspond to the spacer numbers listed in Table 2. The presence and the orientation of IS6110 is depicted by an arrow. DNA flanking the DR region is depicted by the bars at the left and at the right. The black parts of these bars depict the stretches that have been sequenced. The size of the DNA sequence missing compared to strain 8 (M. tuberculosis H37Rv) is given in base pairs, after the triangle. DVRs occurring twice in the DR region are depicted as rectangles marked with a letter: a, a duplication of DVR 35 is present 3′ from spacer 45; b and c, duplicated DVRs 43 and 48, respectively, are present as tandem duplications.

FIG. 2

Rearrangements in the DR locus of presumed isogenic variants. (A) Spoligotype of each of the five isogenic pairs determined by using the standard 43 spacers as a probe; the boxed areas have been sequenced and are depicted in panel B. (B) Arrangement of DVRs as determined by DNA sequencing; the numbers correspond to the spacer numbers as given in Fig. 1. Black rectangles depict DVRs that are present; gray rectangles represent DVRs that are absent. Numbers correspond to DVR numbering as shown in Fig. 1. The arrow represents the insertion element IS6110. Strain numbers are depicted on the left. a, b, and d, tandem duplications of DVR43; DVR48, and DVR21, respectively; c, a duplicated copy of DVR4 is located directly to the right of DVR19.

FIG. 3

Presence of spacers in and traditional spoligotypes and IS6110 RFLP patterns of 170 M. tuberculosis complex strains. (A) Presence or absence of any of the 94 spacers; the spacers are ordered as in Fig. 1, and this order corresponds to the presumed order in the genome. (B) Hybridization signals of the 43 spacers used in traditional spoligotyping. (C) IS6110 RFLP patterns. Strains of group 1 belong to the commonly found spoligotypes ST5, ST14, ST22, and ST38.