Update on Candida krusei, a potential multidrug-resistant pathogen
Population structure of clinical and environmental strains. Branch supports represent pseudo-bootstrap values. Strains named in red are clinical isolates, and strains named in blue are environmental. For each strain, four circles indicate relative resistance magenta or relative sensitivity green to four drugs as shown in the key. Two strains from a hospital in Ireland form the closest pair in the tree C-IE2, C-IE4 and group weakly with a third C-IE1but three other strains from the same hospital are scattered across the tree.
Variation in drug resistance We assayed the in vitro sensitivity of all the sequenced strains to four antifungal drugs—fluconazole, flucytosine, amphotericin B, and micafungin—using the EUCAST protocol [ 58 ]. We also included two P. Contaggio observed MICs for each strain in each drug are presented in Table 2. The distribution of MICs for our P.
We wanted to focus on variation within P. The strains designated as RR and RS for each drug are highlighted in magenta and green, respectively, in Table 2. As expected, all strains of P. Within the range of resistance shown by P. The two P. Many of the RR strains of P. Previous studies on azole resistance in P.
The Erg11 enzyme of P. In our data, we noted five nonsynonymous polymorphisms in Erg11 Table 2but none of them correlated with variation in fluconazole phenotypes among strains. While cahdida strains of P. In our Illumina data, we identified seven other strains that had noticeably lower sequence coverage at the ABCABC1 locus than in neighboring regions of chromosome 4 Table 2.
Because these seven strains all retain a copy of the bp intergenic region between ABC11 and ABC1 it is present in de novo assemblies of these genomeswhich is absent in ABC chimeras [ 26 ], our interpretation of these data is that, like CBS, the seven strains are heterozygotes with an ABC chimeric gene on one chromosome, and a normal tandem ABCABC1 gene pair on another. Importantly, the group of eight strains with ABC chimeric genes includes 4 of the 5 strains that are RS for fluconazole, and none of the RR strains Table 2.
All P. Amphotericin B. Surprisingly, all six RR coontagio were environmental isolates. All but one of the clinical isolates in Clade 3 krusel RS for amphotericin B, while those in Clades 1 and cwndida were not.
Micafungin is the recommended candoda effective existing treatment for P. None of the krussi strains contained mutations at amino acid positions — of the glucan synthase gene FKS1, which have previously been implicated in echinocandin resistance in this species [ 2829 ]. Candiea all four drugs, environmental isolates contavio as likely as clinical isolates to show relative resistance RR to a drug.
Discussion Our ccontagio confirm that P. 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. Furthermore, we found that the species has a fourth name, Candida glycerinogenes. Extensive research has been carried out into its osmotolerance, and genetic manipulation methods have been developed e.
We find that 37 of the 38 C. The existence of multiple names for this species has almost certainly impeded research into it. One of the most unexpected features of the genome is the structure of its centromeres, which consist of a simple but large IR. The structure of the centromeres most closely resembles those of Komagataella phaffii, another yeast in the methylotrophs clade. However, the K. The only other yeasts in this clade whose centromeres have been characterized are Ogataea polymorpha, whose centromeres contain clusters of Ty5-like retrotransposons and do not seem to have an IR structure, and Kuraishia capsulata which has been reported to have point centromeres [ 4064 ].
The P. Its centromeres do contain pseudogenes of Ty3-like elements, and these are more abundant at the centromeres than elsewhere in the genome, but the only intact Ty3-like elements are not centromeric. Centromeres with similar IR structures also occur outside the family Pichiaceae, in C. The centromeres of Sch. An important remaining question concerns the sexual cycle. When P. Later studies by Kurtzman and colleagues reported that the type strain of P.
Our discovery that this strain is triploid provides a possible explanation for its failure to sporulate, or at least its failure to produce viable spores. Of the 32 strains we studied, 20 have this status Table 2. It also contains orthologs of many genes involved in meiosis, although IME1, the master inducer of meiosis, has not been found in P. Because clinical isolates were found to be closely related to environmental isolates, either infections are being acquired opportunistically cotnagio the environment, or yeast strains from infected humans are colonizing the environment.
In view of the range of sources, the former possibility is more likely. The use of P. Moreover, high resistance to fluconazole is common in environmental isolates.
The resistance to fluconazole is shared with P. In another case of a pathogen with a major biotechnological role, it was suggested that a harmless closely related species should be used as a replacement [ 68 ]. Similarly, it may be advisable to consider non-pathogenic Pichia species as possible alternatives for some industrial applications.
It would also be advisable to set limits on the levels candidq drug resistance permissible in P. For clinical isolates obtained from hospital laboratories, all isolates came from different patients candida krusei contagio where possible we obtained candida krusei contagio that were taken prior to drug treatment.
DNA for Illumina sequencing was harvested from stationary-phase cultures by homogenization with glass beads followed by phenol-chloroform extraction and ethanol precipitation. Illumina sequencing of all other strains was done by the core facility of the University of Missouri, USA, using TruSeq libraries coverage details are given in Table 2. Overlaps between the ends of two pairs of contigs were merged manually to obtain five near-complete chromosome sequences.
Error correction was done using Pilon [ 69 ] and manual comparison to de novo Illumina contigs assembled by SPAdes version 3. This assembly appeared to have a higher level of indel errors than the CBS assembly, so we used CBS as the reference genome sequence for annotation and downstream analyses. Contahio novo assemblies of the other 30 strains were made using SPAdes version 3. Nucleotide candida krusei contagio identity of Ploidy estimation by propidium iodide staining Ploidy was estimated with a modified version of the method of Popolo et al.
The resulting set of 32 GVCF files defined an initial set ofSNP sites that were variable among the 32 strains, and this set was used for ploidy analyses Fig 5A. To analyze patterns of LOH, we divided the genome into consecutive kb windows and calculated the number of heterozygous SNPs in each window with allele frequencies between 0.
The threshold of 30 heterozygous sites was chosen because it is a local minimum in the distribution of heterozygous SNP numbers among all windows in all strains. This filtered dataset containedvariable sites. This program generated datasets in which, for each heterozygous site in each strain, one allele was chosen randomly.
A single unrooted consensus tree was then constructed from these trees Fig 6. It also provides a value of estimated log likelihood ln Pr for the model used. Consequently, the wells of each dilution series yielded mL of twice the recommended series of drug concentrations required for MIC determinations.
To achieve final cell densities of 0. Cell densities were confirmed by plate counting. The last well was filled with mL of sterile distilled water to serve as contaminant control. Accession numbers The sequence data reported in this manuscript has been submitted to the NCBI nucleotide database with the following accession numbers: P.
Supporting information S1 Fig. Black diagonals indicate matches in the same orientation, and red diagonals indicate matches in opposite orientations. Bars at the top of the plots show the locations of annotated protein-coding genes in the CBS genome, conhagio an absence of genes at the centromeres.
TIF S2 Fig. Complex relationship among P. Dashed lines mark the ends of the kb section from each chromosome. TIF S3 Fig. Population structure analysis of P. Each column represents a strain, and the colors represent the proportion of sites belonging to each of the 4 inferred populations.
Populations 1—3 form monophyletic clades in the tree in Fig 6but population 4 does not.