A highly polymorphic degenerate microsatellite for molecular strain typing of Candida krusei

Revital Shemer1, Ziva Weissman1, Nehama Hashman2 and Daniel Kornitzer1

Department of Molecular Microbiology, B. Rappaport Faculty of Medicine1, and Department of Clinical Microbiology2, Rambam Medical Center, Haifa 31096, Israel

Author for correspondence: Daniel Kornitzer. Tel: +972 4 829 5258. Fax: +972 4 829 5254. e-mail: danielk{at}tx.technion.ac.il


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Simple sequence repeats, due to their high variability, are widely used for molecular epidemiology of pathogenic micro-organisms. However, their usefulness is restricted by their high instability and low information content. Here, a locus, CKTNR, in the fungal pathogen Candida krusei is described which displays considerable sequence, as well as length, heterogeneity. Alleles of this locus, which contains a degenerate trinucleotide repeat, appear to be stable. The CKTNR polymorphism could serve as the basis for a molecular typing system of C. krusei. Furthermore, analysis of the CKTNR allele distribution suggested that C. krusei reproduces mainly clonally.

Keywords: molecular typing, fungal VNTR, ploidy level, sexual versus clonal reproduction

Abbreviations: AP-PCR, arbitrarily primed PCR; VNTR, variable numbers of tandem repeats

The GenBank accession numbers for the sequences determined in this work are AF326279AF326292.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Candida krusei, the probable anamorph (asexual form) of Issatchenkia orientalis (Barnett et al., 2000 ), is an emerging opportunistic pathogen, which is characterized by a high level of resistance to fluconazole, and increasingly also to amphotericin B (Samaranayake & Samaranayake, 1994 ; Rex et al., 2000 ). The low but increasing incidence of nosocomial C. krusei bloodstream infections, particularly in intensive care units and among neutropenic patients, together with the poor prognosis of these patients (Abbas et al., 2000 ), underscore the need for molecular epidemiological tools to trace potential sources of C. krusei fungemia.

The most widely used and versatile approach for the molecular epidemiology of microbial infections is DNA fingerprinting, which encompasses a variety of methods that detect variations at the DNA sequence level (Soll, 2000 ). The amount of variation between isolates gives a measure of their relatedness. Importantly, implied in this type of analysis is that the organism reproduces clonally, or at least that sexual reproduction, and its associated recombination of markers, is infrequent within the studied population. The choice of fingerprinting method will depend on the type of polymorphism that needs to be detected, which in turn depends on the purpose of the strain comparison. For the determination of the origin of a nosocomial infection, the main question is whether a set of isolates are epidemiologically related or not. For this purpose, the markers to be used should be stable within the time frame of transmission from one patient to another, and sufficiently polymorphic to allow a reasonable amount of discriminatory power (Hunter, 1991 ). An additional important consideration in the clinical setting is the ease and reproducibility of the method.

One method of choice to detect genetic polymorphisms is arbitrarily primed PCR (AP-PCR), where short oligonucleotides are used to amplify random genomic fragments (Welsh & McClelland, 1990 ). This method is technically simple, and if enough primers are used alone or in combinations, the large number of bands obtained ensures that some of the genomic polymorphisms between unrelated strains will be detected with a reasonable probability. However, AP-PCR suffers from a high sensitivity to reaction conditions (Caetano-Anolles, 1993 ; Meunier & Grimont, 1993 ), resulting in high centre-to-centre, and even experiment-to-experiment, variability (Becker et al., 2000 ; Taylor et al., 1999 ; Tyler et al., 1997 ). An alternative approach is to use locus-specific primers for known highly polymorphic loci. Such loci generally consist of variable numbers of simple tandem repeats, or VNTRs (Taylor et al., 1999 ; van Belkum, 1999 ). Variability of VNTRs is probably due to DNA polymerase ‘slippage’ during replication (Strand et al., 1993 ). VNTRs often consist of microsatellites, or short (1–6 nucleotide) sequence repeats, but can also consist of repeats of longer sequences, or minisatellites. One such variable minisatellite is the C. krusei repeated sequence 1 (CKRS-1), which consists of a 165 bp repeat within the rDNA locus (Carlotti et al., 1997 ). One drawback of VNTRs is that because of their simple structure, alleles can be identical by mutation rather than by descent, which reduces their value for phylogenetic studies (Metzgar et al., 2001 ; Orti et al., 1997 ). Furthermore, VNTRs of Candida albicans have been observed to vary over time even within a single patient, suggesting that the speed with which VNTRs evolve may be too high even for the relatively short time frame of epidemiological research (Metzgar et al., 1998 ). To increase the confidence of VNTR analysis and to prevent misclassification of related strains as unrelated, multiple loci need to be assayed.

Potential VNTRs can be identified by computerized searches for simple sequence repeats. Beyond a certain repeat length, such regions have a high probability of being polymorphic (Field & Wills, 1998 ). This approach, which has been successfully used, for example, with C. albicans (Field et al., 1996 ; Lunel et al., 1998 ), requires knowledge of a substantial part of the genome sequence of the organism. However, for pathogens for which little sequence information exists, other methods are required to isolate highly variable regions. Here, we describe the identification by AP-PCR of a highly polymorphic locus in C. krusei. Surprisingly, sequence analysis indicated that the polymorphism occurred within a degenerate trinucleotide repeat. Haplotype analysis suggested that, unlike simple sequence repeats, the degenerate repeats in this locus are relatively stable. Since most haplotypes differ by length, this polymorphism allows the easy and rapid characterization of isolates with high discriminatory power. Furthermore, analysis of the haplotype distribution provides insight into the mode of reproduction of this yeast.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Sample origin.
The origin of the clinical isolates is indicated in Table 1. Six reference strains were obtained from the Centraal Bureau voor Schimmelcultures, Utrecht, The Netherlands.


View this table:
[in this window]
[in a new window]
 
Table 1. Origin and CKTNR diplotypes of clinical C. krusei isolates

 
Molecular biology.
For isolation of genomic DNA, 1 ml of overnight culture in YPD (1% yeast extract, 2% Bacto-peptone, 2% glucose; all from Difco) was resuspended in 0·2 ml 1% SDS, 2% Triton X-100, 100 mM NaCl, 10 mM Tris, 1 mM EDTA, pH 8, and vortexed for 5 min in the presence of 0·2 ml phenol/chloroform (1:1) and 0·3 g glass beads (diameter 0·45 mm). The mixture was then centrifuged for 5 min at room temperature, and the aqueous supernatant was ethanol-precipitated. The DNA pellet was resuspended in 10 mM Tris, 1 mM EDTA, pH 8. PCR reactions were performed with Red-Taq polymerase and associated buffer (Sigma) in 0·2 ml reaction tubes using an Eppendorf ‘Mastercycler personal’ PCR apparatus. For the AP-PCR procedure, 10 ng genomic DNA and 1 µM of each primer were used in 20 µl reactions. For cloning of PCR products, the DNA bands were purified from an agarose gel using a silica matrix (Qiagen) and ligated to the pGEM-T cloning vector (Promega). Sequencing was performed on both strands, either directly on PCR products, or on plasmid clones (in this case, three independent clones were sequenced).

AP-PCR.
Reactions were performed with nine random 10-mers, including UBC143 and UBC734 (Zeng et al., 1996 ) and seven others (sequences available upon request) at annealing temperatures of 40–45 °C, each yielding on average 10 bands of median size 0·5–1 kb, i.e. 5–10 kb were covered per experiment. ‘Coverage’ may have been lower, as short-length polymorphisms may not have been detectable in the larger fragments. The primer that detected the CKTNR polymorphism was RAPD4 (5'-AAGAGCCCGT-3').

Amplification of CKTNR sequence.
The following primers were designed for specific amplification of the CKTNR polymorphic sequence: CKTNR5 (5'-ACAGCAGTCGCAGGCCC-3') and CKTNR3 (5'-GTCGGAGACATAACCGC-3'). All primers were produced by Sigma-Genosys, Cambridgeshire, UK. For each 20 µl reaction, 10 ng genomic DNA was used; alternatively, a small amount of material from a pure fungal colony was directly resuspended in the PCR reaction mix. Amplification conditions were: 94 °C for 4 min, then 30 cycles of 94 °C for 30 s, 58 °C for 30 s and 72 °C for 30 s. Reaction products were separated by electrophoresis for 60 min at 200 V on 6 cm polyacrylamide gels (40:1, 10%) supplemented with Spreadex polymer (Elchrom Scientific). For sequencing of individual alleles, the bands were cut out of the gel, eluted and reamplified, and the PCR products were subjected to direct sequencing using an ABI PRISM 310 automated sequencer. For the GENESCAN procedure, a TET-modified CKTNR3 primer was used, and the reaction products were run on an ABI PRISM 310 machine together with the MapMarker 400 TMRA-marked size standards (BioVentures).

Statistical genetics.
For the calculation of deviations from the Hardy–Weinberg equilibrium, Guo and Thompson’s Markov chain approach was used, because of its appropriateness for large numbers of alleles and small sample sizes (Guo & Thompson, 1992 ). It was implemented using the GENEPOP software package (Raymond & Rousset, 1995 ). Where alleles characterized by sequencing were considered together with alleles characterized only by length, the ambiguous alleles were grouped together. This is a conservative assumption with regard to rejection of the null hypothesis of panmixia. When the whole sample was considered, the null hypothesis could be rejected with P<0·0001. When only the sequenced subset of isolates was considered, the same result was obtained, with P=0·0001. Finally, due to the fact that our sample contained isolates from different geographical locations, which may distort the result due to a populations admixture effect, we performed the test with only the Haifa subset of the sample; here, the null hypothesis was still rejected with P=0·0193.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
In the course of investigating a cluster of four cases of C. krusei sepsis at a large regional tertiary care centre, we applied AP-PCR with a panel of 10-mers (see Methods) to estimate strain relatedness. We also assayed for the previously described CKRS-1 polymorphism with primers Arno1 and 2 (Carlotti et al., 1997 ). The four clinical isolates were compared with the reference strain CBS 573. By both methods, the four isolates appeared to be unrelated to each other and to the reference strain; however, we found the AP-PCR patterns were only partially reproducible, and depended on the amount of starting DNA, as well as on minor variations in experimental conditions (results not shown). The CKRS-1 polymorphism consists of a VNTR, and although it was shown to be stable within patients (Carlotti et al., 1997 ), its stability in patient-to-patient transmission has not been characterized. To identify additional, reproducibly detectable polymorphisms, we scanned the AP-PCR data for bands of variable size. Such bands should represent sequence length polymorphisms, which could then reliably be detected with specific PCR primers. Young et al. (2000) recently identified over 1000 simple tandem repeats (mono- to tetranucleotide repeats) within the 12 Mb genome of Saccharomyces cerevisiae, i.e. approximately 1 repeat every 12 kb. A substantial proportion of these repeats may be expected to be polymorphic based on the observations of Field & Wills (1998) , who found that of 20 trinucleotide repeats surveyed, all were length-polymorphic. Assuming that the same ratios hold true for other fungi, we estimated that 50–100 kb worth of sequence should be scanned by AP-PCR to identify sequence length polymorphisms associated with simple tandem repeat variations.

To obtain these data, twelve PCR reactions were performed on four clinical isolates and one reference laboratory stock strain with nine different AP-PCR primers, alone or in combinations (see Methods). A single primer identified a variable band of about 500 bp (Fig. 1). This fragment was cloned from three strains. Sequencing of one clone from each strain revealed a length polymorphism within a degenerate trinucleotide repeat. We designated this locus CKTNR. The sequence information was then used to design locus-specific primers CKTNR5 and CKTNR3. The CKTNR locus was isolated and subjected to sequence analysis in additional clinical isolates. Most of the clinical isolates yielded two different alleles (Fig. 2 and Tables 1 and 3). This observation supports the suggestion, based on segregation of a uracil auxotrophy following UV irradiation, that C. krusei is diploid (Whelan & Kwon-Chung, 1988 ). The CKTNR primers did not react with C. albicans DNA (Fig. 2). One or two slower-migrating bands were detectable in most lanes (indicated with an asterisk in Fig. 2), the exception being the homozygous reference strain. Upon extraction from the gel and reamplification, each of these bands yielded the original gel pattern again, whereas reamplification of the faster-migrating bands yielded the single amplified band only (not shown). We concluded that these slower-migrating species represent heteroduplexes of the two alleles.



View larger version (40K):
[in this window]
[in a new window]
 
Fig. 1. AP-PCR pattern obtained with isolates 1–4 and the reference strain CBS 573 (lane R) with the RAPD4 primer. PCR products were separated on a 1% agarose gel. Lane M contains the size marker ({phi}X DNA digested with HaeIII). The arrow indicates the variable band that was analysed further.

 


View larger version (25K):
[in this window]
[in a new window]
 
Fig. 2. Size distribution of the CKTNR alleles. PCR amplification products obtained with C. krusei isolates 1–13 (except 8, which is identical to 9), with the C. krusei reference strain CBS 573 (lane R), or with C. albicans DNA (lane C.a.) reacted with primers CKTNR5 and CKTNR3 were separated on a 10% polyacrylamide gel. Lane M contains the size marker ({phi}X DNA digested with HaeIII). The bands migrating in the>300 bp range (indicated by an asterisk on the right) are heteroduplexes of the two allelic species, which presumably migrate anomalously due to the formation of cruciform structures.

 

View this table:
[in this window]
[in a new window]
 
Table 3. Origin and CKTNR diplotypes of I. orientalis reference strains

 
The sequence data are summarized in Table 2. Rather than consisting of a simple repeat sequence, the variable sequence consists of CAA repeats (indicated by the letter B), interspersed with CAG (‘D’) and CAT (‘E’) trinucleotides. Of the 26 alleles sequenced from the first 13 clinical isolates, 13 distinct haplotypes (a–m) were identified. Sequencing of another 15 isolates yielded only a single new haplotype, o, suggesting that the majority of the haplotypes in the population had been identified. Of these, only two pairs (e and f, and j and o) are of identical length but differ in sequence. Thus, the sequence size is highly informative, which enables the epidemiological monitoring of C. krusei by CKTNR length determination. Towards a partial automation of diplotype determination, sequence length determination of the CKTNR locus was performed using a fluorescently labelled primer (Fig. 3). The next 22 isolates were characterized using this method. To differentiate between the haplotypes assigned by sequencing and those inferred from fragment length, the former are indicated in boldface in Table 1. When only sequence length was considered, 19 distinct profiles (out of 78 possible) were detected in our sample of 50 isolates, with between one and seven isolates displaying any given profile. The numerical index of discriminatory power, D, was calculated to be 0·96 according to Hunter’s equation (Hunter & Gaston, 1988 ), i.e. two unrelated isolates have only a 4% chance of showing the same profile. This calculation assumes that all the isolates in our set are unrelated, but since some of the isolates having the same profile could be epidemiologically related, the actual discriminatory power is probably even higher. For example, we found that isolates 21 and 22, which were isolated from two patients in the same department a day apart, also display the same profile with the highly discriminating CKRS-1 probe (Carlotti et al., 1997 ) (data not shown), and are therefore presumably related.


View this table:
[in this window]
[in a new window]
 
Table 2. Structure of the CKTNR degenerate trinucleotide repeat

 


View larger version (15K):
[in this window]
[in a new window]
 
Fig. 3. Profiles of five different strains upon PCR with a fluorescently labelled primer, followed by chromatography on an automated sequencing machine and analysis with the GENESCAN software (Perkin Elmer). The strain number is indicated on the left within each panel. Sequence length determination was accurate to within 1 bp.

 
To obtain insights into the rate of evolution of the CKTNR polymorphism, we also characterized the CKTNR region of six C. krusei/I. orientalis reference strains, isolated as early as 1926, from different geographical locations (Table 3). Strikingly, of the eight haplotypes observed in the reference strains, four were identical to haplotypes previously identified, two (n and p) were novel, and two differed from previously identified haplotypes by a single point mutation. The high complexity of the CKTNR locus results in a large number of possible haplotypes. Therefore, it is likely that the allele identity between the clinical isolates and the reference strains is due to evolutionary stability, rather than to homoplasy (independent convergence towards an identical allele structure).

The mode of reproduction of a micro-organism – clonal versus sexual – greatly influences its epidemiological tractability by molecular fingerprinting. When a single locus is considered, as is the case here, the predictions for a primarily clonal mode of reproduction are fixed heterozygosity and absence of segregation genotypes, or, more generally, deviations from the Hardy–Weinberg equilibrium (Tibayrenc et al., 1991 ). Given the nature of our sample, deviations from the Hardy–Weinberg equilibrium could be most rigorously tested. In the case of panmixia (frequent exchange of alleles by random mating between all members of the population), the alleles are predicted to be at equilibrium. We found that the null hypothesis of panmixia could be rejected with a very high probability (see Methods), suggesting that C. krusei/I. orientalis reproduction is primarily clonal.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The usefulness of microsatellites derives from their high level of polymorphism, which is due to variations at the level of the repeat number. However, the instability of this type of allele, which is due to strand slippage during replication, may be so high as to lead to misclassification of related strains. Degenerate microsatellites such as that found in the CKTNR locus were shown to be considerably more stable (Rolfsmeier & Lahue, 2000 ). Independent support for the stability of the CKTNR haplotypes comes from a number of observations. One is the fact that identical haplotypes were found in isolates from widely divergent locations and dates of isolation (Tables 1 and 3). The complex structure of the CKTNR locus makes it unlikely that these identical haplotypes represent homoplasies. A second observation is that a single nucleotide polymorphism differentiates between haplotypes e and f, making it likely that haplotype f originated from a point mutation in e. The fact that e is still extant in the population indicates that the CKTNR repeats expand and contract within time frames similar to the occurrence of point mutations. The third and most compelling observation is that a specific single nucleotide polymorphism was found associated with, and only with, haplotype m, in strains isolated from different geographical locations and at different times (Table 1). Of note, the probable parental allele of m, m*, was recovered from a strain isolated in Finland in 1964 (Table 3). This again suggests that the stability of the CKTNR haplotypes is of the same order as that of single nucleotide polymorphisms. However, the possibility that some of the CKTNR haplotypes are less stable cannot be excluded. This is a distinct possibility for haplotype b, found in a single isolate, which contains 21 uninterrupted CAA repeats.

An epidemiological typing system should ideally be (1) easy to apply, (2) stable enough over time to prevent misclassification of related strains as unrelated, and (3) polymorphic enough to prevent misclassification of unrelated strains as related (Hunter, 1991 ; Soll, 2000 ). The CKTNR locus conforms to these three requirements: (1) locus-specific primers allow rapid and unequivocal genotyping of any given strain, without the need to test the whole set of strains together; (2) the CKTNR haplotypes appear to be extremely stable over time, which should prevent misclassification of related strains as being unrelated; and (3) the considerable length polymorphism of the CKTNR locus and the occurrence of two alleles results in a high discriminatory power, i.e. a low probability that two epidemiologically unrelated strains will exhibit the same profile. For these reasons, the CKTNR locus could be used to test a suspicion that a cluster of isolates are related. If the CKTNR profiles support the suspicion of relatedness, additional assays such as AP-PCR or CKRS-1 profile determination could be used for confirmation.

We found that both original C. krusei strains and I. orientalis strains, including the I. orientalis type strain CBS 5147, react with the CKTNR PCR primers and carry many alleles identical to those found in our clinical C. krusei isolates (Table 3). Thus, our data support the reclassification of both sets of strains as a single species (Barnett et al., 2000 ). Some strains of I. orientalis have been shown to be able to sporulate, mate and exchange markers (Kurtzman & Smiley, 1976 ; Kurtzman et al., 1980 ). The ability to epidemiologically track a micro-organism by molecular fingerprinting presupposes a primarily clonal, rather than sexual, mode of reproduction. The question of whether C. krusei reproduces sexually has additional implications, most notably with regard to resistance to antifungals. Even if an organism exhibits the ability to reproduce sexually in the laboratory, whether it does so in nature can only be addressed by a population genetics approach (Tibayrenc et al., 1991 ). The diplotype distribution is not consistent with panmixia, suggesting that the reproductive mode of C. krusei is primarily clonal. However, a low level of sexual reproduction cannot be ruled out. It actually may explain the observation that identical alleles are found in association with various different partners.

Concluding remarks
The large-scale AP-PCR methodology we used to identify the CKTNR locus may be applied to isolate additional polymorphic microsatellites, whether the sequence of the organism in question is known or not. Computer-based searches aimed at the detection of simple microsatellites depend on extensive sequence information and may often miss degenerate microsatellites (depending on the allele that is present in the sequenced isolate: note that many of the CKTNR alleles have no more than four contiguous CAA repeats). The combination of length polymorphism, stability and ease of detection potentially provided by degenerate microsatellite loci such as CKTNR makes them useful molecular epidemiology tools.


   ACKNOWLEDGEMENTS
 
The authors thank Itzhack Polacheck (Hadassah Hospital, Jerusalem) for providing isolates and for critical reading of the manuscript, Eli Leffler (Elisha Hospital) and Anna Goldschmiedt-Reuven (Tel Hashomer Hospital) for providing isolates, Sara Selig for critical reading of the manuscript, Nadav Reuven for technical assistance, and Raymonde Szargel and Elie Sprecher for help with GENESCAN. This research was supported by the fund for the promotion of the research at the Technion and the Technion VPR – Hirshenstraus-Gutman Medical Research Fund.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Abbas, J., Bodey, G. P., Hanna, H. A., Mardani, M., Girgawy, E., Abi-Said, D., Whimbey, E., Hachem, R. & Raad, I. (2000). Candida krusei fungemia. An escalating serious infection in immunocompromised patients. Arch Intern Med 160, 2659-2664.[Abstract/Free Full Text]

Barnett, J. A., Payne, R. W. & Yarrow, D. (2000). Yeasts: Characteristics and Identification, 3rd edn. Cambridge: Cambridge University Press.

Becker, K., Badehorn, D., Deiwick, S., Peters, G. & Fegeler, W. (2000). Molecular genotyping of Candida species with special respect to Candida (Torulopsis) glabrata strains by arbitrarily primed PCR. J Med Microbiol 49, 575-581.[Abstract/Free Full Text]

van Belkum, A. (1999). The role of short sequence repeats in epidemiologic typing. Curr Opin Microbiol 2, 306-311.[Medline]

Caetano-Anolles, G. (1993). Amplifying DNA with arbitrary oligonucleotide primers. PCR Methods Appl 3, 85-94.[Medline]

Carlotti, A., Chaib, F., Couble, A., Bourgeois, N., Blanchard, V. & Villard, J. (1997). Rapid identification and fingerprinting of Candida krusei by PCR-based amplification of the species-specific repetitive polymorphic sequence CKRS-1. J Clin Microbiol 35, 1337-1343.[Abstract]

Field, D. & Wills, C. (1998). Abundant microsatellite polymorphism in Saccharomyces cerevisiae, and the different distributions of microsatellites in eight prokaryotes and S. cerevisiae, result from strong mutation pressures and a variety of selective forces. Proc Natl Acad Sci USA 95, 1647-1652.[Abstract/Free Full Text]

Field, D., Eggert, L., Metzgar, D., Rose, R. & Wills, C. (1996). Use of polymorphic short and clustered coding-region microsatellites to distinguish strains of Candida albicans. FEMS Immunol Med Microbiol 15, 73-79.[Medline]

Guo, S. W. & Thompson, E. A. (1992). Performing the exact test of Hardy–Weinberg proportion for multiple alleles. Biometrics 48, 361-372.[Medline]

Hunter, P. R. (1991). A critical review of typing methods for Candida albicans and their applications. Crit Rev Microbiol 17, 417-434.[Medline]

Hunter, P. R. & Gaston, M. A. (1988). Numerical index of the discriminatory ability of typing systems: an application of Simpson’s index of diversity. J Clin Microbiol 26, 2465-2466.[Medline]

Kurtzman, C. P. & Smiley, M. J. (1976). Heterothallism in Pichia kudriavzevii and Pichia terricola. Antonie Leeuwenhoek 42, 355-363.[Medline]

Kurtzman, C. P., Smiley, M. J. & Johnson, C. J. (1980). Emendation of the genus Issatchenkia Kudriavzev and comparison of species by deoxyribonucleic acid reassociation, mating reaction, and ascospore ultrastructure. Int J Syst Bacteriol 30, 503-513.

Lunel, F. V., Licciardello, L., Stefani, S., Verbrugh, H. A., Melchers, W. J., Meis, J. F., Scherer, S. & van Belkum, A. (1998). Lack of consistent short sequence repeat polymorphisms in genetically homologous colonizing and invasive Candida albicans strains. J Bacteriol 180, 3771-3778.[Abstract/Free Full Text]

Metzgar, D., van Belkum, A., Field, D., Haubrich, R. & Wills, C. (1998). Random amplification of polymorphic DNA and microsatellite genotyping of pre- and posttreatment isolates of Candida spp. from human immunodeficiency virus-infected patients on different fluconazole regimens. J Clin Microbiol 36, 2308-2313.[Abstract/Free Full Text]

Metzgar, D., Thomas, E., Davis, C., Field, D. & Wills, C. (2001). The microsatellites of Escherichia coli: rapidly evolving repetitive DNAs in a non-pathogenic prokaryote. Mol Microbiol 39, 183-190.[Medline]

Meunier, J. R. & Grimont, P. A. (1993). Factors affecting reproducibility of random amplified polymorphic DNA fingerprinting. Res Microbiol 144, 373-379.[Medline]

Orti, G., Pearse, D. E. & Avise, J. C. (1997). Phylogenetic assessment of length variation at a microsatellite locus. Proc Natl Acad Sci USA 94, 10745-10749.[Abstract/Free Full Text]

Raymond, M. & Rousset, F. (1995). GENEPOP: population genetics software for exact tests and ecumencicsm. J Heredity 86, 248-249.

Rex, J. H., Walsh, T. J., Sobel, J. D., Filler, S. G., Pappas, P. G., Dismukes, W. E. & Edwards, J. E. (2000). Practice guidelines for the treatment of candidiasis. Infectious Diseases Society of America. Clin Infect Dis 30, 662-678.[Medline]

Rolfsmeier, M. L. & Lahue, R. S. (2000). Stabilizing effects of interruptions on trinucleotide repeat expansions in Saccharomyces cerevisiae. Mol Cell Biol 20, 173-180.[Abstract/Free Full Text]

Samaranayake, Y. H. & Samaranayake, L. P. (1994). Candida krusei: biology, epidemiology, pathogenicity and clinical manifestations of an emerging pathogen. J Med Microbiol 41, 295-310.[Abstract]

Soll, D. R. (2000). The ins and outs of DNA fingerprinting the infectious fungi. Clin Microbiol Rev 13, 332-370.[Abstract/Free Full Text]

Strand, M., Prolla, T. A., Liskay, R. M. & Petes, T. D. (1993). Destabilization of tracts of simple repetitive DNA in yeast by mutations affecting DNA mismatch repair. Nature 365, 274-276.[Medline]

Taylor, J. W., Geiser, D. M., Burt, A. & Koufopanou, V. (1999). The evolutionary biology and population genetics underlying fungal strain typing. Clin Microbiol Rev 12, 126-146.[Abstract/Free Full Text]

Tibayrenc, M., Kjellberg, F., Arnaud, J., Oury, B., Breniere, S. F., Darde, M. L. & Ayala, F. J. (1991). Are eukaryotic microorganisms clonal or sexual? A population genetics vantage. Proc Natl Acad Sci USA 88, 5129-5133.[Abstract]

Tyler, K. D., Wang, G., Tyler, S. D. & Johnson, W. M. (1997). Factors affecting reliability and reproducibility of amplification-based DNA fingerprinting of representative bacterial pathogens. J Clin Microbiol 35, 339-346.[Free Full Text]

Welsh, J. & McClelland, M. (1990). Fingerprinting genomes using PCR with arbitrary primers. Nucleic Acids Res 18, 7213-7218.[Abstract]

Whelan, W. L. & Kwon-Chung, K. J. (1988). Auxotrophic heterozygosities and the ploidy of Candida parapsilosis and Candida krusei. J Med Vet Mycol 26, 163-171.[Medline]

Young, E. T., Sloan, J. S. & Van Riper, K. (2000). Trinucleotide repeats are clustered in regulatory genes in Saccharomyces cerevisiae. Genetics 154, 1053-1068.[Abstract/Free Full Text]

Zeng, S., Wu, L. C. & Lehmann, P. F. (1996). Random amplified polymorphic DNA analysis of culture collection strains of Candida species. J Med Vet Mycol 34, 293-297.[Medline]

Received 27 March 2001; accepted 8 April 2001.