Genetic diversity of dairy Geotrichum candidum strains revealed by multilocus sequence typing

The introduction of multilocus sequence typing (MLST) for strain characterization provided the first sequence-based approach for genotyping many fungi, leading to reproducible, reliable, and exchangeable data. A MLST scheme based on the analysis of six housekeeping genes was developed for genotyping Geotrichum candidum. The scheme was first developed using 18 isolates for which the complete sequences of the alanyl-tRNA synthetase (ALA1), pyruvate kinase (CDC19), acetyl-coA acetyltransferase (ERG10), glutaminyl-tRNA synthase (GLN4), phosphoglucoisomerase (PGI1), and phosphoglucomutase (PGM2) housekeeping genes were determined. Multiple sequence alignments of these genes were used to define a set of loci showing, as closely as possible, the same phylogenetic resolution level as complete gene sequences. This scheme was subsequently validated with 22 additional isolates from dairy and non-dairy sources. Overall, 58 polymorphic sites were indexed among 3,009 nucleotides analyzed. Depending on the loci, four to eight alleles were detected, generating 17 different sequence types, of which ten were represented by a single strain. MLST analysis suggested a predominantly clonal population for the 40 G. candidum isolates. Phylogenetic analysis of the concatenated sequences revealed a distantly related group of four isolates. Interestingly, this group diverged with respect to internal transcribed spacers 1 (ITS1), 5.8S, and ITS2 analysis. The reproducibility of the MLST approach was compared to random amplification of microsatellites by PCR (RAM-PCR), a gel profiling method previously proposed for G. candidum strain typing. Our results found MLST differentiation to be more efficient than RAM-PCR, and MLST also offered a non-ambiguous, unique language, permitting data exchange and evolutionary inference.