Strain Typing and Determination of Population Structure of Candida krusei by Multilocus Sequence Typing

Abstract

A multilocus sequence typing (MLST) scheme for Candida krusei was devised, based on sequencing of six gene fragments of the species. The existence of heterozygous results for each of the six fragments sequenced confirms that C. krusei is diploid for at least part of its genome. The C. krusei MLST scheme had a discriminatory index of 0.998, making this system ideal for strain typing of C. krusei clinical isolates. MLST data for 122 independent C. krusei isolates from a range of geographical sources were analyzed by eBURST, structure, and the unweighted-pair group method using average linkages to derive a population structure comprising four subtype strain clusters. There was no evidence of geographical associations with particular subtypes. Data for pairs of isolates from seven patients showed that each patient was colonized and/or infected with strain types that were indistinguishable by MLST. The C. krusei MLST database can be accessed online at http://pubmlst.org/ckrusei/.

Candida krusei is one of the less common Candida species found as a cause of human infections. As an agent of opportunistic bloodstream Candida infections in susceptible hosts, it accounts for up to 6% of Candida sp. isolates in North American surveys (25) and occasionally has higher prevalence rates in surveys from Europe and other parts of the world (25). The ARTEMIS survey indicated a global prevalence for C. krusei of 2 to 3% among all bloodstream isolates (23), and a study of >3,000,000 patients in intensive care units in the United States indicated an overall prevalence of 1.2% among bloodstream isolates, with an annual incidence ranging from 0 to 0.14 cases per 10,000 central venous catheter days (32).

There is considerable institutional variation in C. krusei prevalence rates between different studies (25), with frequencies of 15% or more reported for bloodstream isolates from some Italian hospitals (10, 19). C. krusei fungemia is more likely to be associated with neutropenia in patients with hematological malignancies than among those with solid tumors or located in intensive care units (1, 12, 15, 16). It is relatively seldom associated with indwelling venous catheters but commonly generates tell-tale cutaneous exanthemata (1, 16).

C. krusei first came to attention as an “emerging” cause of bloodstream infection when it was found to be resistant at the species level to fluconazole (34). Isolates of C. krusei are treated as intrinsically resistant to fluconazole regardless of results of susceptibility tests (17), and prior exposure to fluconazole is a statistically significant risk factor in surveys of patients with C. krusei fungemia (1, 16). In fact, C. krusei shows lower in vitro susceptibility than C. albicans to most antifungal classes in vitro, including amphotericin B, caspofungin, and flucytosine (21, 22). The azole-resistant properties of C. krusei and the widespread use of fluconazole prophylaxis might have been expected to result in a rise in the overall incidence of C. krusei infection. Although several surveys of individual institutions document a rise in C. krusei candidemia through the 1990s and early 2000s, the largest surveys reveal no such trend (8, 20, 32).

Multilocus sequence typing (MLST) is a highly accurate and portable system for distinguishing between isolates of a microbial species (14, 33). It depends on sequencing a set of gene fragments that are rich in single-nucleotide polymorphisms (SNPs) between different isolates. Among pathogenic fungi, formal MLST systems have been developed for C. albicans (3), C. glabrata (6), and C. tropicalis (30). MLST facilitates determinations of population structures and epidemiological correlates of properties, such as geographical and anatomical origins of isolates and their transmission within and between patients (2, 6, 29). C. krusei isolates have previously been typed by electrophoretic karyotyping (5), restriction fragment length polymorphism analysis (5), DNA fingerprinting (4), and microsatellite polymorphism approaches (26). In this study, we present a scheme for MLST with C. krusei. We use MLST to show that patients tend to be colonized and/or infected with the same strain type and that the population structure of C. krusei consists of four clusters of closely related strain subtypes, with one comprising just two isolates that were markedly different from the other isolates tested. The results also confirm that C. krusei is diploid (7, 26).

MATERIALS AND METHODS

Isolates.

All 129 C. krusei isolates used for this study (Table 1) were originally cultured from clinical samples. The identities of all isolates were confirmed by standard morphological and physiological criteria. Several isolates came from our collection of pathogenic fungi, and fresh isolates were recently received for routine test purposes or were kindly supplied by the Mycology Reference Laboratory in Bristol, United Kingdom, the regional Mycology Laboratory in Leeds, United Kingdom, and the University of Iowa. Fourteen of the isolates were duplicates from seven patients, either from different anatomical sites or taken at different times (Table 1). A set of 122 single-source isolates was therefore available, including one from each of the pairs of isolates. The isolates represented probable genetic diversity, based on their dates, anatomical sites, and geographical sources of isolation. The yeasts were maintained on Sabouraud agar (Oxoid, Basingstoke, United Kingdom).

TABLE 1.

List of isolates typed, with their year of isolation, geographical and anatomical sources, and DST

Isolate no.	Yr isolated	Source		DST	Note
		Geographical	Anatomical
J931254	1993	United States	Unknown	1
J931263	1993	United States	Unknown	2
J931298	1993	United States	Unknown	3
J932202/1	1993	Germany	Oropharynx	4
J941366		Sri Lanka	Oropharynx	5
J941795	1994	The Netherlands	Blood	6
J950256	1995	United States	Unknown	7
J950817	1995	Belgium	Blood	8
J990508	1999	United States	Vagina	9
AM2000/0021	2000	United Kingdom	Blood	10
AB432780/1	2001	United Kingdom	Unknown	11
AM30411/12/04	2004	United Kingdom	Oropharynx	12
AM30723/01/04	2004	United Kingdom	Oropharynx	13
AM30846/01/04	2004	United Kingdom	Blood	14
AM30850/01/05	2005	United Kingdom	Ascitic fluid	15
AM30277/02/05	2005	United Kingdom	Blood	16
AM30325/02/05	2005	United Kingdom	Blood	17
AM30487/02/05	2005	United Kingdom	Blood	18
AM30219/03/05	2005	United Kingdom	Unknown	19
AM30308/03/05	2005	United Kingdom	Oropharynx	20
AM30274/04/05	2005	United Kingdom	Blood	21
AM30332/04/05	2005	United Kingdom	Oropharynx	22
AM30455/04/05	2005	United Kingdom	Oropharynx	23
AM31194/04/05	2005	United Kingdom	Unknown	19
AM31300/04/05	2005	United Kingdom	Blood	24	Same patient as AM31301/04/05
AM31301/04/05	2005	United Kingdom	Blood	24	Same patient as AM31300/04/05
AM30133/05/05	2005	United Kingdom	Unknown	25
AM30926/05/05	2005	United Kingdom	Ascitic fluid	26
AM31108/05/05	2005	United Kingdom	Urine	27
AM2000/0008	2000	United Kingdom	Unknown	28
B68404	1992	Belgium	Unknown	29
J931255	1993	United States	Unknown	30
J931256	1993	United States	Unknown	30
J931257	1993	United States	Unknown	30
J931258	1993	United States	Unknown	30
J931259	1993	United States	Unknown	31
J931260	1993	United States	Unknown	32
J931261	1993	United States	Unknown	6
J931262	1993	United States	Unknown	33
J932681	1993	Belgium	Oropharynx	34
J950511	1995	United States	Unknown	35
J990529	1999	United States	Vagina	36
73/126	1973	United Kingdom	Oropharynx	37
NCPF3165		United Kingdom	Unknown	38
AMR29-01	1998	United Kingdom	Unknown	39	Same patient as AMR29-02
AMR29-02	1999	United Kingdom	Unknown	39	Same patient as AMR29-01
AMR49-01	2002	United Kingdom	Unknown	33	Same patient as AMR49-02
AMR49-02	2002	United Kingdom	Unknown	33	Same patient as AMR49-01
AMR49-03	2002	United Kingdom	Unknown	40
AMR49-04	2002	United Kingdom	Unknown	41	Same patient as AMR1465
AMR1465	2002	United Kingdom	Blood	41	Same patient as AMR49-04
AMR18	1994	United Kingdom	Blood	30
AMR62	1994	United Kingdom	Hepatic flexure tubes	42
AMR107	1995	United Kingdom	Blood	43
AMR112	1995	United Kingdom	Blood	43
AMR183	1995	United Kingdom	Blood	44
AMR245		United Kingdom	Unknown	45
AMR275	1996	United Kingdom	Blood	43
AMR455	1997	United Kingdom	Peritoneal fluid	30
AMR458	1997	United Kingdom	Oropharynx	46
AMR695	1998	United Kingdom	Peritoneal fluid	42
AMR743	1999	United Kingdom	Blood	30	Same patient as AMR744
AMR744	1999	United Kingdom	Oropharynx	30	Same patient as AMR743
AMR816	1999	United Kingdom	Blood	47
AMR895	2000	United Kingdom	Blood	5
AMR998		Turkey	Blood	48
AMR1015	2001	United Kingdom	Oropharynx	49	Same patient as AMR1019
AMR1019	2001	United Kingdom	Pus	49	Same patient as AMR1015
AMR1034		United Kingdom	Unknown	50
AMR1041	2001	United Kingdom	Vagina	51
AMR1072	2001	United Kingdom	Bile	51
AMR1423	2002	United Kingdom	Urine	52	Same patient as AMR1424
AMR1424	2002	United Kingdom	Blood	52	Same patient as AMR1423
AMR1569	2003	United Kingdom	Blood	53
AM2005/0525	2004	Venezuela	Blood	54
AM2005/0531	2004	Slovakia	Blood	55
AM2006/0127	2001	Australia	Blood	56
AM2005/0492	2001	Ecuador	Blood	57
AM2005/0494	2002	Chile	Blood	25
AM2005/0496	2002	South Africa	Blood	58
AM2005/0498	2002	Czech Republic	Blood	59
AM2005/0505	2003	Slovakia	Blood	58
AM2005/0507	2003	Australia	Blood	60
AM2005/0508	2003	Australia	Blood	61
AM2005/0510	2003	South Africa	Blood	62
AM2005/0513	2003	Italy	Blood	63
AM2005/0514	2003	Spain	Blood	42
AM2005/0515	2003	Czech Republic	Blood	64
AM2005/0516	2003	Italy	Blood	65
AMR1222	2001	United Kingdom	Blood	66
AM2005/0521	2003	Italy	Blood	67
AM2005/0523	2004	Hungary	Blood	68
AM2005/0528	2004	Czech Republic	Blood	69
AM2005/0529	2004	Poland	Blood	70
AM2005/0532	2004	Italy	Blood	71
AM2005/0533	2004	Slovakia	Blood	55
AM2005/0534	2004	Czech Republic	Blood	67
AM2005/0535	2004	Czech Republic	Blood	17
AM2005/0538	2004	Italy	Blood	72
AM2005/0541	2005	Australia	Blood	73
AMR1491	2002	United Kingdom	Blood	74
AMR1316	2002	United Kingdom	Skin	75
AMR1317	2002	United Kingdom	Blood	76
AM2005/0486	2001	United States	Vagina	77
AM2005/0487	2001	United States	Vagina	78
AM2005/0488	2001	South Africa	Blood	79
AM2005/0490	2001	Spain	Blood	80
AM2005/0491	2001	Ecuador	Blood	81
AM2005/0493	2002	Slovakia	Blood	55
AM2005/0495	2002	Chile	Blood	82
AM2005/0497	2002	Colombia	Blood	30
AM2005/0500	2002	Italy	Blood	83
AM2005/0502	2002	Venezuela	Blood	67
AM2005/0511	2003	Colombia	Blood	84
AM2005/0520	2003	Poland	Blood	42
AM2005/0522	2003	Italy	Blood	85
AMR1411	2002	United Kingdom	Blood	86
AM2005/0530	2004	Slovakia	Blood	87
SCS71987M	2005	United Kingdom	Blood	88
AM2005/0539	2004	Poland	Blood	24
AM2005/0540	2004	Portugal	Blood	10
AM2006/0128	2005	Slovakia	Blood	55
AM2005/0499	2002	Colombia	Blood	89
AM2005/0501	2002	South Africa	Blood	90
AM2005/0503	2002	Spain	Blood	91
AM2005/0512	2003	Mexico	Blood	92
AM2005/0517	2003	Italy	Blood	23
AM2005/0519	2003	Italy	Blood	93
AM2005/0526	2004	Russia	Blood	94

Choice of loci for MLST.

Initially, amino acid sequences of 11 proteins from C. albicans, C. tropicalis, C. glabrata, and Saccharomyces cerevisiae were aligned to identify regions of homology. Degenerate primers were designed for DNA sequences in areas with high levels of homology to permit PCR amplification of the open reading frames from an isolate of C. krusei. The selected DNA fragments from C. krusei were ligated into the pGEM-T Easy vector (Promega) and sequenced on an ABI 3730 DNA analyzer (Foster City, CA), using the M13fwd and M13rev primers (5′-GTTTTCCCAGTCACGAC and 5′-CAGGAAACAGCTATGAC, respectively). New sets of primers specific for C. krusei were designed, and sequencing was repeated for 20 C. krusei isolates selected to represent wide geographical and anatomical diversity. DNASTAR (Lasergene) software was used to identify polymorphic sites. Final primer sets were designed to amplify fragments of 500 to 750 bp containing the largest possible number of SNPs. For the MLST scheme, we reduced the set of genes to six genes (Table 2), which was the minimum combination that afforded differentiation of all 20 isolates.

TABLE 2.

Genes used for MLST

Gene	Gene product	Primer (direction, sequence [5′-3′])a	Amplicon size (bp)	Sequence start	Sequence end	Gene length (bp)	No. of SNPs	No. of genotypes	No. of genotypes/site
ADE2	Phosphoribosylamino-imidazole carboxylase	Fwd, GTCACTTCTCAGTTTGAAGC	600	AAACAAAT	CTCATTTA	470	9	21	2.33
Rev, ACACCATCTAAAGTAGAGCC
HIS3	Imidazole glycerol phosphate dehydratase	Fwd, GGAGGGGACATATCACTGCC	534	AATCCCAA	GTTGATTG	400	8	14	1.75
Rev, AATCTTTAATTGCCAAAGCC
LEU2	Isopropyl malate dehydrogenase	Fwd, CTGTGAGACCAGAACAGGGG	802	GTAACTTT	AAGCTCTC	619	9	17	1.89
Rev, GCAGAGCCACCCAAGTCTCC
LYS2	l-Aminoadipatesemialdehyde dehydrogenase	Fwd, ATCTGAGAAGCAGTTGGCGC	631	AAAGATTG	TCTGAACT	441	10	19	1.90
Rev, AGACTTGTAAGAATTATCCC
NMT1	N-Myristoyltransferase	Fwd, CTGATGAAGAAATCACCG	738	GCTTCATTT	GAAGTGAA	537	11	22	2.00
Rev, GCTTGATATCATCTTTGTCC
TRP1	Phosphoribosylanthranilate isomerase	Fwd, AGCTATGTCGAGCAAAGAGG	503	ATATGAGG	CAGGTGGG	380	11	22	2.00
		Rev, ACATCAACGCCACAACACCC

^aFwd, forward; Rev, reverse.

DNA extraction, amplification, and sequence determination.

Genomic DNAs were extracted from yeasts grown in yeast extract-peptone-dextrose broth as previously described (30, 31).

PCRs were used to amplify the gene fragments listed in Table 2. Reaction volumes of 50 μl contained 100 ng of genomic DNA, 2.5 U Taq DNA polymerase (Promega, Madison, WI), 5 μl of 10× buffer (supplied with the enzyme), 1.5 mM MgCl₂, a 100 μM concentration of each deoxynucleoside triphosphate (Promega), and 0.2 μM (each) forward and reverse primers. A Flexigene thermocycler (Techne, Cambridge, United Kingdom) was used with a first cycle of denaturation for 5 min at 94°C, followed by 30 cycles of denaturation at 94°C for 1 min, annealing at 48°C for 1 min, and elongation at 72°C for 1 min, with a final extension step of 10 min at 72°C. The amplified products were precipitated by adding 20% polyethylene glycol-2.5 M NaCl solution to the PCR products in microdilution plates (30), which were sealed, vortexed, incubated at room temperature for 30 min, and centrifuged for 1 h at 2,250 × g (4°C). The supernatant was discarded, and the plate was inverted onto a piece of 3-mm chromatography paper and centrifuged again at 500 × g for 1 min to remove any residual polyethylene glycol from the wells. Pellets were washed with 150 μl of 70% ethanol, precipitated as described above, and resuspended in 60 μl of sterile water. Both strands of purified gene fragments were sequenced on an ABI 3730 DNA analyzer (Foster City, IA), using the same primers that were used in the PCR step. The sequence data were analyzed with DNASTAR software. Heterozygosities were defined by the presence of two coincident peaks in the forward and reverse sequence chromatograms. The one-letter code for nucleotides from the International Union of Pure and Applied Chemistry (IUPAC) nomenclature was used to define results.

Statistical analysis of MLST data.

Phylogenetic analyses by the unweighted-pair group method using average linkages (UPGMA) were conducted with MEGA, version 2.1 (13). The analysis was applied to concatenated polymorphic base sequence data as previously described (30). Briefly, a single concatenated sequence was generated by representing each SNP as two bases, which were identical when the sequence showed a homozygous result and represented the different bases from each allele for a heterozygous result. The majority of the sequences, which were identical for all isolates, were excluded from the analysis. The significance of the UPGMA cluster nodes was determined by bootstrapping with 1,000 randomizations. The eBURST package (9; http://eburst.mlst.net/) was used to determine putative relationships between isolates. The eBURST algorithm placed all isolates differing at only one or two of six sequenced fragments into clonal complexes and, where possible, predicted the founding or ancestral diploid sequence type (DST) of each complex. The output was a display of the most parsimonious patterns of descent of each DST from the ancestral type. Discriminatory power was calculated according to the method of Hunter (11).

To assign boundaries to clusters of closely related strain subtypes, the information from the UPGMA analysis was combined with results from eBURST and from a haplotype analysis with the structure package (24). structure assigns isolates to one of K population groups according to a probability estimate based on their genotypes, as defined by SNPs. PHASE software (27, 28) was first used to determine putative haplotypes at each gene fragment; these data formed the input for the structure package, set to a K of 7 with no admixture, the value that gave the lowest value of α (a measure of the likelihood of haplotypes originating mostly from a single population [24]) in pilot runs with different values of K. The output assigned individual isolates to one of seven populations according to their dominant genotype, Q (24). For most isolates, the result was unequivocal, reproducible assignment to a single population; for some isolates, assignments differed between runs. The package was therefore run three times with the same data, and the final structure population assignment was taken as the result obtained consistently in 2/3 or 3/3 replicate runs.

RESULTS

C. krusei strain differentiation by MLST.

Heterozygous sequencing results were obtained for all six fragments sequenced, suggesting that C. krusei has a diploid, or at least partially diploid, genome. MLST distinguished 94 DSTs among a total of 129 isolates, which indicates a discriminatory power of 0.998 (Table 1). DST 30 was the most common type encountered, representing 9 of the 129 isolates (Table 1). All six gene fragments were similarly discriminatory, with ratios of genotypes to SNPs ranging from 1.75 for HIS3 to 2.33 for ADE2 (Table 2). The C. krusei MLST database is available for open public access at http://pubmlst.org/ckrusei/.

The number of new genotypes found for each fragment sequenced fell with increasing numbers of isolates sequenced, but the number of DSTs continued to increase at an average rate of 7.3 new DSTs per 10 isolates typed.

Nucleotide polymorphisms and amino acid changes.

The changes resulting from the 60 SNPs in the six gene fragments were synonymous in 30 instances. Among the 30 nonsynonymous changes, 16 were nontrivial, e.g., changes of acidic side chains to basic side chains or of aliphatic to aromatic side chains.

Paired C. krusei isolates from single patients.

For the seven patients from whom two separate C. krusei isolates were obtained either at different times or from different anatomical sites, DSTs were identical for both isolates but different for each of the patients (Table 1).

Population structure of 122 isolates from single sources.

eBURST analysis of the C. krusei MLST genotypes revealed seven clonal clusters, but 50 isolates were singletons that did not belong to any clonal cluster. To reduce the proportion of singletons, the eBURST analysis was repeated to allow two genotype differences between isolates, which generated eight clonal clusters with 36 singleton isolates. The largest clonal cluster resulting from tests with four of six genotypes being identical comprised 19 DSTs and was based on DST 55 as the putative ancestral member. Cluster 2 was founded on DST 63 as the most probable ancestor and contained 18 DSTs. Clusters 3, 4, and 5 represented six, five, and five DSTs, respectively, while clusters 6 through 8 comprised just two or three DSTs.

The UPGMA dendrogram for 122 C. krusei isolates, based on P distances for the concatenated SNPs determined by MLST, is shown in Fig. 1. The bootstrap values for the nodes in this dendrogram were extremely low and did not help to robustly distinguish internal structure. At an arbitrarily selected cutoff P distance of 0.07, the dendrogram could be subgrouped into four clusters of isolates, which stood out clearly when the dendrogram was represented as an unrooted radial display (Fig. 1, inset).

FIG. 1.

UPGMA dendrogram based on P distances for 122 C. krusei isolates, each from a separate source. Bootstrap values are indicated for the major cluster nodes. For each isolate, the subtype assignment, DST, eBURST clonal cluster assignment, and structure group assignment are shown. Subtype boundaries are indicated by solid lines, and possible clusters within subtype 1 are indicated with dotted lines. Inset, radial display of the dendrogram to demonstrate the four subtypes with clarity.

The large group of isolates designated subtype 1 (Fig. 1) contained isolates from four eBURST clonal clusters and four structure groups. Within subtype 1, there was good correlation between isolates in clonal clusters 2, 3, and 5 and those in structure groups 4, 6, and 5, respectively (Fig. 1). The isolates in UPGMA subtype 2 fell mainly into structure group 1, although 7 of the 30 isolates in this subgroup were from structure groups 4 and 7; all of these were singletons by eBURST analysis. UPGMA subtype 3 correlated perfectly with structure group 3 and with clonal cluster 1 and therefore represents the most robust subtype of the four. Subtype 4 consisted of just two isolates, AMR1015 and AM2005/0530, which were also the only isolates in structure group 2.

Geographical origins of subtypes.

There was insufficient information on the anatomical sources of the isolates typed to permit statistical analysis for possible associations between subtypes and sites of C. krusei commensalism or infection. The geographical origins of all of the isolates were known (Table 1). The geographical distributions of isolates in the four subtypes were as follows: subtype 1 (56 isolates), United Kingdom (21 [37.5%]), other European countries (24 [42.9%]), North America (6 [10.7%]), and elsewhere (5 [8.9%]); subtype 2 (29 isolates), United Kingdom (13 [44.8%]), other European countries, (2 [6.9%]), North America (4 [13.8%]), and elsewhere (10 [34.5%]); and subtype 3 (35 isolates), United Kingdom (17 [48.6%]), other European countries and North America (7 each [20.0%]), and elsewhere (4 [11.4%]). The two isolates in subtype 4 came from the United Kingdom and Europe. These differences in geographical distributions within subtypes were not statistically significant (Fisher-Freeman-Halton exact test; P = 0.3).

DISCUSSION

We have demonstrated that MLST can be used to distinguish strain types within the species C. krusei, thus adding C. krusei to the existing list of Candida species (C. albicans, C. glabrata, and C. tropicalis) that can be subtyped by the MLST approach (3, 6, 30). We consider the four subtypes we defined by an arbitrary similarity cutoff point for the UPGMA dendrogram in this study to be reasonably robust, despite very low bootstrap values, since they correlate reasonably with clusters defined by eBURST and with groups determined by triplicate runs of haplotypes in the structure package (Fig. 1). A larger database of isolates from more geographically disparate sources may ultimately lead to better subtype definition and reveal enrichment of MLST-based subtypes by geographical source, as shown for the other three Candida species (6, 29, 30). However, it was not easy to assemble even the panel of 129 C. krusei isolates used in this study, since the species is the least frequently isolated among species regarded as “medically associated” yeasts.

C. krusei is the anamorph of the ascomycetous yeast Issatchenkia orientalis. However, no reports exist to suggest that clinical isolates of C. krusei can be mated to generate the teleomorph. The low bootstrap values in the UPGMA dendrogram and the high proportion of singletons found in eBURST analyses may be evidence of departure from clonal reproduction in C. krusei, and deeper study of mating types in the species may be warranted.

Carlotti et al. used a DNA fingerprinting probe and differentiated 12 clusters of related types among 58 C. krusei isolates (4). Shemer et al. found 17 haplotypes of C. krusei, based on microsatellite typing of a highly polymorphic locus, among 50 isolates (26). Both studies therefore indicated a high inherent level of diversity among isolates of C. krusei. By using MLST, we found 94 distinct DSTs in a panel of 122 separate-source isolates, also suggesting a high level of interisolate diversity. The level of strain diversity indicated by the UPGMA dendrogram for C. krusei isolates (Fig. 1) is at least as big as that for the equivalent C. albicans dendrogram (29), which means that MLST, with its high discriminatory index, its portability, its ability to detect new SNPs (unlike microsatellite typing), and the availability of a public database of results (http://pubmlst.org/ckrusei/), is the current tool of choice for typing new isolates of the species.

Because the reported incidences of C. krusei bloodstream infections vary considerably between institutions (25), MLST could be used to investigate possible single-institution outbreaks of C. krusei infection. Although we found no evidence for geographical enrichment of isolates within C. krusei subtypes, it is perhaps worth noting that a set of five isolates from different patients in a single hospital in Leeds, United Kingdom, in the year 2002 (AMR49-04, AMR49-03, AMR1316, AMR1491, and AMR1317) coclustered with a high similarity within subtype 3 (Fig. 1) and that four isolates (AM2005/0531, AM2005/0493, AM2005/0533, and AM2006/0128) originating from Slovakia, possibly from a single institution, were indistinguishable by MLST, suggesting that the patients from whom the isolates were obtained were infected by a single C. krusei strain type located in the institutions concerned.

Results for the seven pairs of isolates from different patients show that, as with C. albicans (2, 18), patients tend most often to be colonized or infected at different sites with the same strain, thus indicating endogenous infection with a clonal strain type. The appearance of heterozygosities for some isolates in each of the six gene fragments sequenced confirms previously published suggestions that the genome of C. krusei is diploid (7, 26).

Studies with C. krusei have been relatively limited compared with those for more commonly encountered Candida species, and they typically focus on the innate resistance of C. krusei to fluconazole and other antifungal agents. The introduction of a portable and discriminatory MLST system for C. krusei will greatly facilitate future studies with the species by allowing different laboratories to compare results and will contribute to the growth of the central web database for strain types.