The genomes of 28 bacterial strains, including mycobacterial species Mycobacterium tuberculosis and Mycobacterium bovis, were analyzed for the presence of a special class of microsatellite, that of trinucleotide repeat sequences (TRS). Results of a search of all 10 possible TRS motifs (i.e., CCT, CGG, CTG, GAA, GAT, GTA, GTC, GTG, GTT, and TAT) with five or more repeating units showed that (CGG)(5) was highly represented within the genomic DNA of M. tuberculosis and M. bovis. Most of the (CGG)(5) repeats in the genome were within the open reading frames of two large gene families encoding PE_PGRS and PPE proteins that have the motifs Pro-Glu (PE) and Pro-Pro-Glu (PPE). (CGG)(5)-probed Southern hybridization showed that some mycobacterial species, such as Mycobacterium marinum, Mycobacterium kansasii, and Mycobacterium szulgai, possess many copies of (CGG)(5) in their genomes. Analysis of clinical isolates obtained from Tokyo and Warsaw with both IS6110 and (CGG)(5) probes showed that there is an association between the fingerprinting patterns and the geographic origin of the isolates and that (CGG)(5) fingerprinting patterns were relatively more stable than IS6110 patterns. The (CGG)(5) repeat is a unique sequence for some mycobacterial species, and (CGG)(5) fingerprinting can be used as an epidemiologic method for these species as well as IS6110 fingerprinting can. If these two fingerprinting methods are used together, the precise analysis of M. tuberculosis isolates will be accomplished. (CGG)(5)-based fingerprinting is particularly useful for M. tuberculosis isolates with few or no insertion elements and for the identification of other mycobacterial species when informative probes are lacking.
