The genetic diversity of clinical isolates of Streptococcus pyogenes serotype M5 has been characterized. Strain genotypes were defined by macrorestriction profile, 16S ribotype, emm gene subtype, insertion element IS1239 profile, and exotoxin gene determinant. By these criteria, clinical isolates of M5 constituted a multiplicity of strain clusters rather than a homogeneous population as found for certain serotypes. Distance matrices and an unrooted tree were constructed from macrorestriction data with three rarely cutting endonucleases, determined by PFGE. A single IS1239 profile was common to 85% of isolates but there was great diversity of both ribotype and macrorestriction profile, and 18 different emm gene subtypes were detected by PCR-RFLP. DNA sequence analysis of the antigen-coding 5' (hypervariable) region of emm gene amplicons (about 240 bp) showed that 14/18 exhibited up to 6% divergence. Four amplicons had highly divergent sequences--corresponding to those previously determined for emm6, emm11, emm18 and emm77. Further serological and hybridization studies were used to analyse the discrepancy between the Lancefield serotype of these strains (M5) and their emm genotype. Overall, this study shows a high degree of genetic diversity in serotype M5, with implications for the Lancefield scheme itself, for the epidemiology of group A streptococci, and for recombinant DNA strategies for M protein-based vaccine development.