Multilocus sequence typing scheme for Staphylococcus aureus: revision of the gmk locus.

Multilocus sequence typing (MLST) is a sequence-based genotyping method based on polymorphisms (each variant is termed an allele) in seven housekeeping genes (loci) (arcC, aroE, glpF, gmk, pta, tpi, and yqiL) in Staphylococcus aureus providing unique allelic profiles known as sequence types (STs) (1). The data obtained by MLST are highly reproducible and are available in the S. aureus MLST database (http://saureus.mlst.net/). Since the publication of Enright and coworkers’ article (1), MLST has been extensively used to study evolution and population dynamics among S. aureus lineages (our search identified 1,000 citations of this paper indexed at ISI Web of Knowledge as of 16 April 2012). We occasionally have had difficulties assigning allele numbers for the gmk locus due to suboptimal DNA sequence quality in the 3= end of the gmk amplicons, which prompted us to investigate the problem. DNA sequence analysis revealed a 12-bp overlap between the 3= end of the 429-bp region that is used to define the gmk alleles and the 3= end of the primer gmk-Dn (5=-TCATTAACTAC AACGTAATCGTA-3= [the overlapping 3= end is underlined]). This means that the 12 bp (TACGATTACGTT) in the 3= end of the 429-bp gmk region should always be conserved given that the gmk-Dn primer is integrated into the amplicons. Nevertheless, among the 183 gmk alleles available in the S. aureus MLST database as of 16 April 2012, we identified 12 gmk alleles with a polymorphism(s) in the 12-bp region corresponding to 48 STs (Table 1). The most common polymorphism (8/12) was a G instead of a T at position 429 (Table 1). Based on our experience, many of the polymorphisms in the 12-bp region are likely to be “false” due to suboptimal DNA sequence quality toward the end of the gmk amplicons. On the other hand, we acknowledge that some of the polymorphisms may be “true” if they have been identified by PCR amplification using alternative downstream primers or by analysis of whole-genome sequence data. We argue that exclusion of the 12 bp from the 429-bp gmk region is the most straightforward procedure to ensure the robustness of S. aureus MLST in the future. The required changes in the S. aureus MLST database and the software used to assign gmk alleles based on the new 417-bp region have recently been made, and the issue identified here has been highlighted on the homepage (http://saureus.mlst.net/). Where changes of the ST have oc-