Research Article: Population genomics shows no distinction between pathogenic Candida krusei and environmental Pichia kudriavzevii: One species, four names

Date Published: July 19, 2018

Publisher: Public Library of Science

Author(s): Alexander P. Douglass, Benjamin Offei, Stephanie Braun-Galleani, Aisling Y. Coughlan, Alexandre A. R. Martos, Raúl A. Ortiz-Merino, Kevin P. Byrne, Kenneth H. Wolfe, Aaron P. Mitchell.


We investigated genomic diversity of a yeast species that is both an opportunistic pathogen and an important industrial yeast. Under the name Candida krusei, it is responsible for about 2% of yeast infections caused by Candida species in humans. Bloodstream infections with C. krusei are problematic because most isolates are fluconazole-resistant. Under the names Pichia kudriavzevii, Issatchenkia orientalis and Candida glycerinogenes, the same yeast, including genetically modified strains, is used for industrial-scale production of glycerol and succinate. It is also used to make some fermented foods. Here, we sequenced the type strains of C. krusei (CBS573T) and P. kudriavzevii (CBS5147T), as well as 30 other clinical and environmental isolates. Our results show conclusively that they are the same species, with collinear genomes 99.6% identical in DNA sequence. Phylogenetic analysis of SNPs does not segregate clinical and environmental isolates into separate clades, suggesting that C. krusei infections are frequently acquired from the environment. Reduced resistance of strains to fluconazole correlates with the presence of one gene instead of two at the ABC11-ABC1 tandem locus. Most isolates are diploid, but one-quarter are triploid. Loss of heterozygosity is common, including at the mating-type locus. Our PacBio/Illumina assembly of the 10.8 Mb CBS573T genome is resolved into 5 complete chromosomes, and was annotated using RNAseq support. Each of the 5 centromeres is a 35 kb gene desert containing a large inverted repeat. This species is a member of the genus Pichia and family Pichiaceae (the methylotrophic yeasts clade), and so is only distantly related to other pathogenic Candida species.

Partial Text

Pathogenic Candida species are ascomycete yeasts that cause over 46,000 invasive infections annually in the US alone, with a 30% mortality rate [1]. C. albicans is the most common and the most extensively studied, but non-albicans candidiasis infections are becoming increasingly common. The top five pathogenic Candida species in order of prevalence in invasive candidiasis worldwide are C. albicans (52% of infections), C. glabrata (21%), C. tropicalis (14%), C. parapsilosis (9%) and C. krusei (2%) (calculated from data in [2]). Among these, C. krusei is the least-well studied. Although uncommon in the normal human flora, C. krusei is sometimes carried intestinally by healthy individuals and in one remote Amerindian community it was found to be present in over 30% of the population, much higher than C. albicans, and was probably acquired from food or the environment [3]. As well as being associated with humans, C. krusei has been detected in feral pigeons and other wild animals [3, 4].

Our results confirm that P. kudriavzevii and C. krusei are the same species and demonstrate that their genomes are collinear. The discovery that clinical and environmental isolates are interspersed in a phylogenetic tree of strains and do not form distinct clades indicates that there is no justification for continuing to use both names for this species. A third name, I. orientalis, is obsolete, having been formally replaced by the name P. kudriavzevii [6]. Furthermore, we found that the species has a fourth name, Candida glycerinogenes. Since its discovery by Zhuge in 1973, ‘C. glycerinogenes’ has been used in China for the industrial-scale production of glycerol by fermentation of plant carbohydrates [19]. Extensive research has been carried out into its osmotolerance, and genetic manipulation methods have been developed (e.g., [61, 62]). We find that 37 of the 38 C. glycerinogenes gene sequences available in NCBI are virtually identical to P. kudriavzevii sequences, including the 18S rDNA. The existence of multiple names for this species has almost certainly impeded research into it. In keeping with the One Fungus One Name principle [63], we suggest that P. kudriavzevii should be the only name used in future.