Microsatellite typing identifies the major clades of the human pathogen Candida albicans

Abstract

Candida albicans population studies showed that this species could be divided into sub-groups of closely related strains, designated by clades. Since the emergence of microsatellite analysis as a PCR based method, this technique has been successfully used as a tool to differentiate C. albicans isolates but has never been tested regarding clustering of the five major clades. In this study we tested microsatellite length polymorphism (MLP) ability to group 29 C. albicans strains previously defined as belonging to clades I, II, III, E and SA, nine atypical strains from Angola and Madagascar, and 78 Portuguese clinical isolates. MLP typing of the total 116 strains analyzed yielded 87 different multilocus allelic combinations which resulted in a high discriminatory power index, of 0.987, with only two markers, CA1 and CEF3. Cluster analysis of the 29 previously defined strains grouped them according to their clade designation with a matrix cophenetic correlation of r=0.963 after a normalized Mantel statistic. Clustering analysis of the 116 strains maintained the same groupings, clearly defining the five major C. albicans clades. The cophenetic value obtained was of r=0.839, and the one-tail probability of the normalized Mantel statistic out of 1000 random permutations was P=0.0020. The proportion of Portuguese isolates in the groups I, II, III and SA was of 2.7%, 15.4%, 3.8% and 0%, respectively. None of the isolates co-clustered with the atypical strains. These results confirm MLP typing as a good method both to type and differentiate C. albicans isolates and to group isolates, identifying the major C. albicans clades, similarly to Ca3 fingerprinting and multilocus sequence typing (MLST).