MLST based serotype prediction for the accurate identification of non typhoidal Salmonella serovars.

Abstract

Traditional serotyping based on the phenotypic variation of O- and H-antigen remains as the gold-standard for the identification and classification of Salmonella isolates for last 70years. Although this classification is a globally recognized nomenclature, huge diversity of Salmonella serotypes have made the serovar identification to be very complex. Seven gene multilocus sequence typing (MLST) on the other hand can provide serovar prediction as well as the evolutionary origin between the serovars. In this study non typhoidal Salmonella (NTS) strains (n = 45) isolated from clinical samples (blood, faeces and pus) were identified by traditional phenotypic serotyping and biochemical testing. All the tested Salmonella isolates were designated as serovar Typhimurium based on phenotyping. However, by MLST 60% (27/45) of the isolates were S. Typhimurium, 35.5% (16/45) were S. Agona (ST13), 2.2% (1/45) were S. Kentucky (ST198) and 2.2% (1/45) were S. Saintpaul (ST27). MLST analysis assigned S. Typhimurium isolates as ST36 (18/127), ST19 (7/27) and ST313 (2/27). Mismatches in serovar designation between MLST database and phenotypic serotyping can be due to the misinterpretation of phenotypic serotyping as the antigenic structures of S. Typhimurium, S. Agona differs by a surface antigen. MLST based phylogeny of study isolates showed clustering according to sequence types. Concordance between MLST based sequence type and phenotypic serotype is important to provide insights into genetic population structure of Salmonella.