Development of a Multiplex PCR Short Tandem Repeat Typing Scheme for Candida krusei

ABSTRACT

Candida krusei is a human-pathogenic yeast that can cause candidemia with the lowest 90-day survival rate in comparison to other Candida species. Infections occur frequently in immunocompromised patients, and several C. krusei outbreaks in health care facilities have been described. Here, we developed a short tandem repeat (STR) typing scheme for C. krusei to allow the fast and cost-effective genotyping of an outbreak and compared the identified relatedness of 10 isolates to single nucleotide polymorphism (SNP) calling from whole-genome sequencing (WGS). From a selection of 14 novel STR markers, 6 were used to develop two multiplex PCRs. Additionally, three previously reported markers were selected for a third multiplex PCR. In total, 119 C. krusei isolates were typed using these nine markers, and 79 different genotypes were found. STR typing correlated well with WGS SNP typing, as isolates with the same STR genotype varied by 8 and 19 SNPs, while isolates that differed in all STR markers varied by at least tens of thousands of SNPs. The STR typing assay was found to be specific for C. krusei, stable in 100 subcloned generations, and comparable to SNP calling by WGS. In summary, this newly developed C. krusei STR typing scheme is a fast, reliable, easy-to-interpret, and cost-effective method compared to other typing methods. Moreover, the two newly developed multiplexes showed the same discriminatory power as all nine markers combined, indicating that multiplexes M3-1 and M9 are sufficient to type C. krusei.

INTRODUCTION

The emergence of multidrug-resistant yeasts is a well-recognized problem in global health care (1). Every year, an estimated 250,000 people are affected by candidemia, a bloodstream infection caused by yeasts of the genus Candida and associated with high morbidity and mortality rates, particularly in immunocompromised patients (2, 3). Candida krusei, also known under synonymous names as Pichia kudriavzevii, Issatchenkia orientalis, and Candida glycerinogenes, causes candidemia with the lowest 90-day survival rate, although it is not as frequently reported as other Candida species (4). Outside the United States and Europe, a common empirical treatment of candidemia is the administration of fluconazole, an antifungal to which C. krusei is intrinsically resistant (5). In addition, the prophylactic use of fluconazole could increase the risk of C. krusei infections, as it can suppress the growth of other microbes and allow fluconazole-resistant species, like C. krusei, to proliferate (6). Reports of decreased susceptibility to other antifungals like flucytosine, amphotericin B, and caspofungin highlight the potential threat of this pathogen (7–9).

Previously reported outbreaks of C. krusei in hospitals include outbreaks in The Netherlands (2001) (10), Finland (2005) (11), South Africa (2014 and 2015) (12), and India (2013, 2014, and 2018) (13, 14). In general, enhanced infection prevention measures curb an outbreak and greatly reduce the risk of further outbreaks from developing (15, 16). For example, during an outbreak in The Netherlands, hand hygiene was enforced, fluconazole prophylaxis was changed to an amphotericin B oral suspension, and patients were cohorted according to positivity. Thereafter, no C. krusei outbreaks have occurred (10). However, in case of an outbreak, it is crucial to identify its source to prevent the further spread of the pathogen. A recent outbreak in a pediatric ward in India demonstrated the need for a fast and reliable way to genotype C. krusei. Amplified fragment length polymorphism (AFLP) typing of C. krusei isolates collected from patients showed clustering of environmental isolates with the majority of the blood isolates, and a washbasin in the neonatal unit of the pediatric emergency ward was identified as the source of the outbreak (14).

The first attempts to fingerprint C. krusei isolates employed restriction endonuclease analysis of genomic DNA with the restriction enzyme HinfI followed by conventional electrophoresis (17) and with PCR-based amplification of species-specific repetitive polymorphic sequences (18). Currently, whole-genome sequencing (WGS), multilocus sequence typing (MLST), and AFLP typing are the methods available to genotype C. krusei (14, 19, 20). However, all of these techniques have several disadvantages. WGS is costly and takes more time than other techniques, which is not desired when dealing with a potential outbreak. MLST has a considerably lower resolution than other genotyping methods, as phylogenetic relationships can be masked due to the analysis of slowly evolving housekeeping genes (21). Finally, AFLP typing requires a large number of isolates, and results are often difficult to interpret, with inconsistencies in genotyping between different laboratories. A more suitable method to type strains of a single organism is short tandem repeat (STR) or microsatellite genotyping (22). This technique relies on the differences between repeat numbers of specific repeating sequences in the genome, which originate from strand slippage of the DNA polymerase during replication (23). Different strains of the same species have different numbers of repeats, which makes it possible to distinguish these strains from each other. A major advantage of STR compared to AFLP genotyping is the easy implementation of the technique across different laboratories, while interpretation of the data is also more straightforward. Afterwards, it is possible to analyze any number of isolates and directly compare the results to those for previously identified isolates. However, this technique must be set up separately for each species, which requires a large variety of isolates with diverse genotypes to identify markers with the largest variation between strains.

Recently, 33 STR loci in the genome of C. krusei were identified, and a collection of isolates originating from 15 hospitals across China were genotyped (24). All 48 isolates were discriminated from one another using eight markers with the highest discriminatory power. Here, we further improved C. krusei STR typing by selecting the three most discriminating markers reported by Gong et al. (24) and developed two novel multiplex PCR assays containing three trinucleotide repeats and three nonanucleotide repeats, both with more discriminatory power than previously reported markers.

MATERIALS AND METHODS

Isolates.

For the STR analysis, 119 C. krusei isolates of a clinical origin or isolated from the environment were used. The isolates were identified by matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry as C. krusei with log(score) values of ≥2 according to the standard MALDI Biotyper protocol with a Microflex LT mass spectrometer (Bruker, Bremen, Germany) (25). For a complete overview of isolates, see Table S1 in the supplemental material. Additionally, 13 other Candida and Cryptococcus species were included for a specificity test. Isolates were stored at −80°C, according to standard procedures.

Culture and DNA extraction.

Candida krusei isolates were taken from storage at −80°C, grown on Sabouraud agar plates at 30°C, and stored at 4°C. For STR analysis, a single colony was inoculated into 50 μL physiological saline (154 mM NaCl) and incubated for 5 min at 37°C after the addition of 200 U lyticase (Sigma-Aldrich, St. Louis, MO, USA). Subsequently, 450 μL physiological saline (154 mM NaCl) was added, and the sample was incubated for 15 min at 100°C and cooled down to room temperature. For whole-genome sequencing, strains were resuspended in 400 μL MagNA Pure bacterial lysis buffer and MagNA Lyser green beads and mechanically lysed for 30 s at 6,500 rpm using the MagNA Lyser system (all from Roche Diagnostics GmbH, Mannheim, Germany). Subsequently, DNA was extracted and purified with the MagNA Pure LC instrument and MagNA Pure DNA isolation kit III (Roche Diagnostics), according to the recommendations of the manufacturer. After the addition of RNase (5 μg · μL⁻¹), samples were incubated for 1 h and purified with the MagNA Pure LC instrument and MagNA Pure DNA isolation kit III (Roche Diagnostics). The DNA concentration was measured with a Qubit 3.0 fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) using the double-stranded DNA (dsDNA) high-sensitivity settings.

Identification of STR loci.

Three reference genomes of Candida krusei (GenBank assembly accession numbers GCA_003054445.1, GCA_003054405.1, and GCA_003033855.1) were downloaded from the NCBI database. The five chromosomes from each genome were combined into one FASTA file (https://www.bioinformatics.org/sms2/combine_fasta.html), which was uploaded to Tandem Repeats Finder (https://tandem.bu.edu/trf/trf.html) using the advanced search option (alignment parameter, 2,7,7; minimum alignment score to report repeat, 90; maximum period size, 20; maximum tandem repeat array size, 2) (26). The resulting STRs were screened, and repeats that contained insertions or deletions, exhibited a <90% perfect match to the repeat sequence, or had a copy number of <6 were excluded. From the remaining STRs, 6 repeats with a period size of 9 and 8 repeats with a period size of 3 were selected based on a high copy number (>20 and >10 for the trinucleotide and nonanucleotide, respectively) and their presence in at least two reference genomes.

Primer design, PCR, DNA sequencing, and genotyping.

Primers were designed with Primer3web (http://primer3.ut.ee/), using the default settings, except for primer size (minimum [Min], 19 nt; optimum [Opt], 21 nt; maximum [Max], 24 nt), primer melting temperature (T_m) (Min, 57.0°C; Opt, 59.0°C; Max, 62.0°C), Max-poly-X (3 nt [4 nt for 3F_rev and 9C_fwd]), and GC clamp (1) (27). Primers that formed no self- or cross-dimers with 5 or more nucleotides from the last 7 nucleotides of the 3′ end of a primer according to the multiple-primer analyzer from Thermo Fisher Scientific were ordered via Eurogentec (Cologne, Germany). PCR amplification of the STR-flanking sequences was performed on a thermocycler (Biometra; Westburg, Göttingen, Germany) using 1× FastStart Taq polymerase buffer without MgCl₂, deoxynucleotide triphosphates (dNTPs) (0.2 mM), MgCl₂ (3 mM), forward (fwd) and reverse (rev) primers (0.8 μM), 1 U FastStart Taq polymerase (Roche Diagnostics, Germany), and DNA. The same reagents and concentrations were used for multiplex PCRs, except for fwd and rev primer concentrations of 0.1 to 1.0 μM. A thermal protocol of 10 min of denaturation at 95°C followed by 30 cycles consisting of 30 s of denaturation at 95°C, 30 s of annealing at 60°C, and 1 min of extension at 72°C with a final incubation step for 10 min at 72°C was used for PCR amplification. For Sanger sequencing, products were purified according to the AmpliClean method (NimaGen, Nijmegen, The Netherlands), and sequencing PCR was performed using 0.5 μL BrilliantDye premix, 1.75 μL BrilliantDye 5× sequencing buffer (NimaGen), 1 μL fwd or rev primer (5.0 μM), 5.75 μL water, and 1 μL purified DNA. Subsequently, the products were purified using the D-Pure purification protocol (NimaGen), sequenced on a 3500 XL genetic analyzer (Applied Biosystems, Foster City, CA, USA), and analyzed using BioNumerics 7.6.1 (Applied Maths, Kortrijk, Belgium). For STR analysis, PCR products were diluted 1:1,000 in water, and 10 μL together with 0.12 μL of the Orange 600 DNA size standard (NimaGen) were incubated for 1 min at 95°C and analyzed on a 3500 XL genetic analyzer (Applied Biosystems).

Data analysis and discriminatory power.

Copy numbers of all 9 markers were determined using GeneMapper 5 software (Applied Biosystems). For all markers, stutter peaks lower than 50% of the intensity of the highest peak for an allele, minus-A peaks, and bleed-through peaks were discarded. Subsequently, the peak with the highest mass was assigned as the principal peak for the STR allele, and its size was rounded. Copy numbers for repeats were converted to a binary matrix: 1 if an isolate contained the allele and 0 if it did not. Relatedness between isolates was analyzed using BioNumerics software version 7.6.1 (Applied Maths) via the unweighted pair group method with arithmetic means (UPGMA), using the multistate categorical similarity coefficient. The discriminatory power of the STR analysis was determined using the Simpson index of diversity (D) as described previously (28). A D value of 1.0 indicates that the typing method is able to discriminate among all isolates. A D value of 0 indicates that all isolates are identical.

Whole-genome sequencing and variant calling.

Genomic libraries were prepared and sequenced with Illumina technology (Illumina, San Diego, CA, USA) with 2- by 150-bp paired-end-read mode at Eurofins Genomics (Ebersberg, Germany). Read data were uploaded to the Galaxy tool, and FastQC was used to assess the quality of the read data; no trimming was performed (29). Read data were aligned against the genome of C. krusei CBS573 (GenBank assembly accession number GCA_003054445.1) using BWA-MEM, with a mean coverage of 156.34 ± 14.59 (30). PCR duplicates were removed using RmDup, local realignment was performed with BamLeftAlign, and unpaired reads were removed with BAM filter. Finally, reads mapped to mitochondrial DNA were removed. Variants were detected with Freebayes using the default settings except for population model options (ploidy, 3) and allelic scope options (ignore indels, multiple nucleotide polymorphisms [MNPs], and complex events) (31). The resulting VCF files were combined, and variants with a read depth (DP) of <30, a quality (QUAL) of <100, an allele frequency (AO) of <0.25× DP, and an AO of between 0.75× DP and 0.90× DP were removed. Phylogenetic analysis was performed with VCF2PopTree, using the genetic drift algorithm (32).

RESULTS

Preselection of STR markers.

From a list of tandem repeats identified from the genomes of C. krusei strains CBS573, CBS5147, and SJP, eight trinucleotide and six nonanucleotide repeats were selected. To identify conserved regions flanking the tandem repeats, primers were designed approximately 100 to 200 bp upstream and downstream of the STR and amplified in nine different C. krusei isolates and one Candida albicans isolate. PCR products were found for all 14 STRs in the C. krusei isolates and none in C. albicans. For each marker, regions flanking the STR were sequenced, and a consensus sequence was generated. Flanking sequences of one trinucleotide repeat contained indels, and it was discarded. In addition, three other trinucleotide repeats showed little variability in copy number and were also discarded. In total, four trinucleotide repeats and six nonanucleotide repeats were preselected (see Table S2 in the supplemental material).

Development of multiplex PCR and typing of C. krusei.

Primers were designed in proximity to the STR, screened for potential cross-dimer formation, and coupled to fluorescent probes. After the STR regions were amplified with the new primers, one nonanucleotide repeat did not yield any PCR products and was excluded from further STR selection. Subsequently, the three trinucleotide and three nonanucleotide repeats with the highest copy number variability were combined into two PCR multiplexes, M3-1 and M9, respectively. Additionally, STR markers Cakr005, Cakr019, and Cakr031 previously reported by Gong et al. (24) were combined in a third multiplex, M3-2. The three multiplex PCRs were applied to all 119 C. krusei isolates present in the collection of our hospital (CWZ) that were previously identified by MALDI-TOF analysis, which included isolates from CWZ (n = 78), clinical isolates from India (n = 8), and 14 CBS reference strains (Fig. 1; Table S3). All isolates were successfully genotyped using the multiplex panels, with the exception of one isolate that did not show any PCR products for marker Cakr019 reported by Gong et al. (24). Most isolates had two or three alleles per marker, indicating that these were heterozygous diploid or triploid strains. An overview of repeat characteristics, the number of genotypes found, and Simpson’s index of diversity (D), which ranged from 0.81 to 0.95 with a cumulative D of 0.97, is shown in Table 1. Among 119 C. krusei isolates, 79 different genotypes containing 1 to 15 isolates were identified (Fig. 1). One cluster of 15 isolates was identified, containing 14 isolates from Iran, presumably from a local outbreak, and 1 isolate originating from India. Another cluster composed of four isolates from The Netherlands, the type strain CBS573, and other clinical isolates of unknown origins was found. The majority of the other isolates were all unrelated according to STR typing. Nearly all genotypes varied in at least four STR alleles, with the exception of five genotypes with variation in one allele and two genotypes with variation in two alleles. To investigate whether genetic differences found with STR typing are correlated with single nucleotide polymorphisms (SNPs) between strains, WGS of 10 isolates was performed (Fig. S1 and Tables S4 and S5). Reads from each strain were mapped to the CBS573 reference genome, and variant sites between isolates were determined. Strains that showed the same genotype by STR typing differed in 8 SNPs (isolates 4 and 5) and 19 SNPs (isolates 6 and 7), whereas strains that varied in all markers except Cakr013 were discriminated by more than 45,000 SNPs (isolates 3 and 9). Isolates 1 and 2, which varied by 1 to 4 copy numbers in four markers, differed in 309 SNPs (Fig. 2). These results show that STR typing can identify genetically identical isolates (8 and 19 SNPs), related isolates (309 SNPs), and unrelated isolates (over 45,000 SNPs).

FIG 1.

Minimum-spanning tree of 119 C. krusei isolates originating from different countries. Branch lengths indicate the similarity between isolates with thick solid lines (variation in one allele), thin solid lines (variation in two alleles), thin dashed lines (variation in three alleles), and thin dotted lines (variation in four or more alleles).

TABLE 1.

Overview of PCR primers for selected STR loci, concentrations used in multiplex PCR, details of repeat characteristics, discriminatory indices, and genomic sites

PCR panel and primer name	Primer sequence (5′–3′)		Concnb (μM)	No. of bases of primer-flanking sequence	Repeat unit	No. of repeatsc					No. of genotypes	D valued	Intragenic/locus of protein-coding genee
	Forwarda	Reverse				Min	Max	Ref 1	Ref 2	Ref 3
M3-1
M3-1a	FAM-GAGGCTGCCACTTCTAGAACAG	GGACTGTTGTGCTGGATTTCC	2	68	ACA	18	86	46	49		59	0.94	C5L36_0B12140
M3-1b	JOE-CAGCAGAATCAGTTACAACTACCAC	AGAATTGTTGTTGCCCATGATG	6	94	ACA	49	120	69	96		47	0.95	C5L36_0A08730
M3-1c	TAMRA-TCAGTTTCAACAACACCTCCAG	GAGCGGAGGCAGCAGAAAG	8	96	CAA	22	96	38	46		59	0.94	C5L36_0B06450
M3-2C
Cakr005	FAM-CAGTCAACTCGCCCTCCCT	CAGTGTTTGTGCCTGTGCC	2	282	AAT	5	39	9	13	26	37	0.89	Intragenic
Cakr019	JOE-CGATTTCTAGTGGTGTTAGT	ATACTCTTAGCCCTGATACA	1	171	TCA	9	65	24	44		34	0.81	C5L36_0D05450
Cakr031	TAMRA-CCTTGTTGGTAATAGTTTTC	CTAACGAGGAAGTTGTATGT	10	320	TCT	4	26	12	15		31	0.87	C5L36_0B01640
M9
M9a	FAM-GGACTCTCTGGTTATCTTGTCCC	ACAGGCAGTTGGTGGATACTG	1	141	TTCTTCTCT	10	36	17	20		58	0.93	C5L36_0B08500
M9b	JOE-CTCACCTCGTTGATGTTGTTC	GCCCCACCAGTCAATTTAATCA	2	220	TCATGTTGT	10	27	15	17	19	45	0.92	C5L36_0B00740
M9c	TAMRA-GGGCGATTGGGTCTTTGAC	ATACACATCAAATGCTCCGGG	8	107	TGTTCATAT	17	23	17			18	0.87	C5L36_0E00420

^aFAM, 6-carboxyfluorescein; JOE, 4′,5′-dichloro-2′,7′-dimethoxy-fluorescein; TAMRA, 6-carboxytetramethylrhodamine.

^bConcentrations of forward and reverse primers are identical.

^cMin, minimum; Max, maximum; Ref, reference strain with alleles 1, 2, and 3. The reference strain is CBS5147.

^dDiscriminatory power of the STR assay as determined via the Simpson index of diversity.

^eLocus in strain CBS573.

FIG 2.

Comparison of genetic relatedness among isolates determined by STR typing (A) and SNP typing (B). UPGMA dendrograms of phylogenetic relationships by STR typing and WGS SNP typing, generated by BioNumerics and VCF2PopTree, respectively. Bars represent the number of different alleles or SNPs between isolates. Original isolate identifiers and numbers of SNPs among isolates can be found in Table S5 in the supplemental material.

Reproducibility, stability, and specificity of C. krusei STR typing.

To test the reproducibility of the STR genotyping, DNAs from isolates CBS573 and CWZ ID 10-04-05-40 were independently amplified five times in five different experiments. STR genotyping demonstrated identical results in each experiment for all STR markers, demonstrating that the method is highly reproducible. The stability of the STR markers was tested by subcloning five colonies of CBS573 and CWZ ID 10-04-05-40 for 10 generations, after which the copy numbers for each STR marker were compared with those from previous generations. The copy number of the markers did not change in any of the colonies, indicating high genetic stability (Table S6). To test the specificity of C. krusei STR genotyping, 13 other yeast species were analyzed with all nine markers (Table S7). PCR products were found for markers Cakr019, M3-1c, and Cakr005 in C. albicans, Candida pseudohaemulonii, and Cryptococcus gattii, respectively. None of the species formed PCR products after amplification with more than one marker, suggesting that the STR typing scheme is specific for C. krusei.

DISCUSSION

This study describes the development of a C. krusei STR genotyping scheme and its application to detect heterogeneity among isolates of different origins. The assay consists of two newly developed multiplex PCRs, which amplify three trinucleotide and three nonanucleotide repeats, M3-1 and M9, respectively. To compare M3-1 and M9 to an existing STR genotyping method for C. krusei, three STR markers previously reported by Gong et al. were combined into a third multiplex PCR (M3-2) (24). Surprisingly, there was no mention of the presence of triploid C. krusei isolates, i.e., containing three STR alleles, in the study by Gong et al., while it has been shown that certain strains, including the type strain of P. kudriavzevii (CBS5147), are triploid (19, 33). Typing of 119 C. krusei isolates resulted in 79 different genotypes. Notably, all observed genotypes were also found when these isolates were typed solely with M3-1 and M9 and no other combination of two multiplexes. This resulted in significantly higher D values for M3-1 and M9, suggesting their use for rapid genotyping during a potential outbreak. All markers were stable and specific for C. krusei, as copy numbers were not altered after subcloning two different isolates for 100 generations, and only three nonspecific PCR products were formed after specificity testing with 13 other pathogenic fungi. The inclusion of 14 CBS reference strains will aid in the implementation of the STR typing scheme in different laboratories, as data can easily be compared, in contrast to other typing methods such as AFLP typing.

Phylogenetic analysis of C. krusei by STR typing.

The application of the C. krusei STR typing scheme revealed the genetic relationship among strains isolated from patients in different countries. We identified 79 different genotypes in 119 C. krusei isolates and two clusters with more than 10 isolates. The largest cluster consisted of 15 isolates, with 14 isolates from Iran and 1 from India. These 14 detected isolates are possibly a clonal strain distributed throughout the country and demonstrate how the STR scheme is able to identify outbreaks. The other cluster contained 12 isolates and included 7 isolates of unknown origin, 4 from The Netherlands, and strain CBS573, which was isolated in 1935 in Sri Lanka. To differentiate strains by STR typing, the absolute number of differing STR alleles was used, without taking the difference in copy numbers per marker into account (Fig. 1). The latter was not feasible as some isolates were homozygous for a particular STR allele and presented 1 copy number, whereas others were heterozygous triploid for the same STR and presented 3 copy numbers (i.e., one STR marker of one strain will have three PCR products, since the strain is heterozygous triploid for the particular STR allele). As a consequence, we opted for a method that directly compares the presence of alleles between isolates and marks isolates as related only if they share identical STR alleles. A potential caveat to this method is the perception of heterozygous alleles as homozygous when the presence of SNPs in the primer binding region flanking the STR prevents the amplification of one or two alleles. To determine whether the difference of one STR allele was caused by an SNP, the primer binding regions of these particular alleles were sequenced for all isolates that varied in one or two STR alleles from each other (data not shown). From these 19 isolates, CBS2055 contained an SNP in the forward primer binding sequence of M3-1a on one of its three alleles. Additionally, an SNP in the reverse primer binding sequence of M9c on an allele of isolate CWZ ID 10-06-14-67 resulted in a difference of one allele from isolate CWZ ID 10-06-14-60. Thus, a difference in the absence or presence of a single STR allele is certainly not sufficient to identify isolates as genetically distinct, as this difference can be caused by a single SNP.

Comparison STR typing and SNP typing.

To investigate whether the relatedness of isolates identified by the newly developed STR typing scheme for C. krusei is comparable to that determined by WGS, the relatedness of 10 C. krusei isolates was determined with SNP calling of WGS data. These isolates included two pairs of isolates with identical STR genotypes, two isolates that varied by a few copy numbers in four markers, and four isolates that varied in almost all markers (see Table S3 in the supplemental material). The SNP calls were filtered and included in the matrix if they had a minimum read depth of 30 and a quality of 100 and were found in more than 90% (homozygous SNPs) or between 25% and 75% (heterozygous SNPs) of the base calls at a position. The pairs of isolates with identical STR profiles varied in 8 and 19 SNPs from one another. These isolates originated from the same hospitals in India and were likely part of a clonal outbreak, which explains their genetic relatedness. The number of SNPs between C. krusei isolates with the same STR profiles is comparable to the number of Candida auris isolates with identical STR profiles, as all of these isolates were differentiated by fewer than 20 SNPs (34).

The two isolates that differed by 1 to 4 copy numbers in 4 markers varied in 309 SNPs, which is more than an order of magnitude higher than the isolates with identical copy numbers for each marker. These two isolates are likely closely related but not identical, and both originated from the same hospital in India. The fact that such closely related strains are differentiated by STR analysis is due to the high mutation rates of repeat regions in comparison to nonrepeating sequences (35). This instability causes a change in the STR profiles in relatively small time frames and, thus, closely related strains. The four isolates that were identified as completely unrelated by STR typing varied by 20,000 to 47,000 SNPs from all other isolates, over 3 orders of magnitude higher than for genetically identical isolates with the same STR profile (Fig. 2). These results show that C. krusei STR typing can identify related and unrelated isolates comparably to SNP typing while being significantly faster and less expensive. In the future, additional studies will be required to determine the minimal number of differing STR alleles or SNPs needed to differentiate between related and unrelated isolates.

In summary, we developed an STR typing method for C. krusei, which is reproducible, reliable, specific, and much faster than current typing methods such as AFLP typing and WGS. Using STR analysis to type 119 C. krusei isolates, we discriminated 79 different genotypes and found that its results were highly comparable to WGS variant calling. Ultimately, this assay makes genotyping of C. krusei more accessible for many laboratories during a potential outbreak.