New Microsatellite Multiplex PCR for Candida albicans Strain Typing Reveals Microevolutionary Changes

Abstract

Five new microsatellite loci were described and characterized for use as molecular markers for the identification and genetic differentiation of Candida albicans strains. Following the typing of 72 unrelated clinical isolates, the analysis revealed that they were all polymorphic, presenting from 5 to 30 alleles and 8 to 46 different genotypes. The discriminatory power obtained by combining the information generated by three microsatellites used in a multiplex PCR amplification strategy was 0.99, the highest ever reported. The multiplex PCR was later used to test a total of 114 C. albicans strains, including multiple isolates from the same patient collected from different body locations and along episodes of vulvovaginal infections. Three different scenarios for strain relatedness were identified: (i) different isolates that were revealed to be the same strain, (ii) isolates that were the same strain but that apparently underwent a process of microevolution, and (iii) isolates that corresponded to different strains. Analysis of the microevolutionary changes between isolates from recurrent infections indicated that the genotype alterations observed could be the result of events that lead to the loss of heterozygosity (LOH). In one case of recurrent infection, LOH was observed at the CAI locus, and this could have been related to exposure to fluconazole, since such strains were exposed to this antifungal during treatment. The analysis of microsatellites by a multiplex PCR strategy was found to be a highly efficient tool for the rapid and accurate differentiation of C. albicans strains and adequate for the identification of fine microevolutionary events that could be related to strain microevolution in response to environmental stress conditions.

Candida albicans, the most common fungal pathogen, is a commensal yeast that belongs to the normal microbial population of the mouth, vagina, and gastrointestinal tract in humans. However, in people with a variety of transient or permanent immunocompromised conditions, including transplant recipients, chemotherapy patients, underweight neonates, and human immunodeficiency virus-infected individuals, it may become an invasive pathogen (30, 56). Infections by opportunistic fungal agents are a major medical problem due to the growing number of immunocompromised patients with risk factors for such infections. Moreover, this problem is exacerbated by the fact that only a limited array of antifungal drugs is available and by the growing resistance among clinical isolates (39, 46, 55). The development of techniques and strategies that can accurately differentiate clinical isolates is of great relevance. These techniques should provide the ability to differentiate among strains responsible for clinical infections, as well as to trace their epidemiological pathways. Several molecular methods have been used to differentiate C. albicans strains, including electrophoretic karyotyping (2), the use of species-specific probes such as Ca3 or 27A in restriction enzyme analysis (28, 35, 36, 43), PCR-based methods (1, 7, 15, 51), and more recently, multilocus sequence typing, which uses nucleotide sequences from several genes and offers good discrimination (4, 49). Microsatellites or simple tandem repeats (STRs) consist of stretches of tandemly repeated mono- to hexanucleotide motifs dispersed throughout the genome and have a high level of polymorphism compared with those of other molecular markers. In yeast, microsatellite loci have considerable length variations, and this polymorphism quickly made them attractive markers for a variety of types of analyses, including strain typing (9, 16, 45), population structure analysis (13, 18, 25), and epidemiological studies (33, 42). Microsatellite polymorphism is manifested as allelic length differences due to the different numbers of repeated units present in the alleles and is easily assayed by PCR amplification (48).

Several polymorphic microsatellite loci have been identified in the C. albicans genome, most of them located near EF3 (5); CDC3 and HIS3 (3); or inside the coding regions ERK1, 2NF1, CCN2, CPH2, and EFG1 (31), although the discriminatory powers (DPs) calculated for such loci were relatively low (between 0.77 and 0.91). A higher value, 0.97, was estimated by Sampaio et al. (42) for microsatellite CAI, located in a noncoding region; and this discriminatory power was identical to that which had previously been reported for a combination of three microsatellites amplified in a multiplex reaction (3). A higher polymorphism is expected in microsatellite loci from noncoding regions; therefore, in order to obtain a greater resolution, a set of new microsatellite markers was selected from these regions and characterized for use in a multiplex PCR. This multiplex typing system was applied to C. albicans clinical isolates to test its efficiency for strain differentiation. The potential capacity of this system to identify fine microevolutionary events was also evaluated.

MATERIALS AND METHODS

Yeast strains and DNA extraction.

A total of 112 clinical isolates of C. albicans, obtained from two hospitals and a health care center located in Braga and Porto (northern Portugal), and reference strain WO-1, as well as type strain PYCC 3436 (ATCC 18804), were used in this study. The isolates were obtained from primary cultures, and one colony of each different phenotype was selected. The type strains C. parapsilosis PYCC 2545 (ATCC 22019), C. krusei PYCC 3341 (ATCC 6258), C. tropicalis PYCC 3097 (ATCC 750), C. glabrata PYCC 2418 (ATCC 2001), C. guilliermondii PYCC 2730 (ATCC 6260), C. lusitaniae PYCC 2705 (ATCC 34449), and C. dubliniensis CBS 7987 (ATCC MYA-646) were also included. All reference strains were obtained from the Portuguese Yeast Culture Collection (PYCC), New University of Lisbon, Lisbon, Portugal, except C. dubliniensis, which was obtained from Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands. The isolates were previously identified by PCR fingerprinting with primer T3B (7). Prior to DNA isolation, yeast cells were grown overnight on Sabouraud broth medium at 30°C. A Zymolyase-based method was used to extract DNA by following procedures described previously (22). After DNA extraction, the cultures were frozen in 30% (wt/wt) glycerol.

Microsatellite selection and PCR primers design.

A search of the C. albicans genome sequences, available in databases from Stanford's DNA Sequencing and Technology Center (http://www-sequence.stanford.edu/group/candida), was performed in order to identify sequences containing microsatellite repeats. The search was performed with the aim of identifying microsatellite units of tri-, tetra-, and pentanucleotides according to criteria described previously (42). The sequences that were obtained and selected for locus specific amplification are presented in Table 1. Primers for each locus were designed by using the software Primer3, available from http://www-genome.wi.mit.edu/cgi-bin/primer/primer3_www.cgi.

TABLE 1.

Microsatellite DNA sequences selected from the database search^a

Microsatellite designation	Sequence codeb	Primer sequencec	No. of repeated units	PCR product size (bp)
CAI	396062C04.sl.seq	F-5′-ATGCCATTGAGTGGAATTGG-3′	41	252
		R-5′-AGTGGCTTGTGTTGGGTTTT-3′
CAIII	265080D05.sl.seq	F-5′-TTGGAATCACTTCACCAGGA-3′	11	113
		R-5′-TTTCCGTGGCATCAGTATCA-3′
CAIV	265080D05.sl.seq	F-5′-TGCCAAATCTTGAGATACAAGTG-3′	24	263
		R-5′-CTTGCTTCTCTTGCTTTAAATTG-3′
CAV	265147C06.yl.seq	F-5′-TGCCAAATCTTGAGATACAAGTG-3′	50	237
		R-5′-CTTGCTTCTCTTGCTTTAAATTG-3′
CAVI	385031H11.xl.seq	F-5′-ACAATTAAAGAAATGGATTTTAGTCAG-3′	21	294
		R-5′-TGCTGGTGCTGCTGGTATTA-3′
CAVII	396033C06.sl.seq	F-5′-GGGGATAGAAATGGCATCAA-3′	10	222
		R-5′-TGTGAAACAATTCTCTCCTTGC-3′

^aThe sequence codes and primers used for PCR amplification are presented. The number of repeated units and the predicted size of PCR amplification products were calculated according to the sequence of the strain SC5314 used in the sequencing project.

^bCode attributed to the sequence in the Stanford database.

^cFor each primer pair, the letter F indicates the forward primer and the letter R indicates the reverse primer.

Chromosomal localization of microsatellite markers.

The sequences selected were searched by BLAST against the latest release (assembly 19) of C. albicans genome sequence, to give a location to a sequence contig (http://www-sequence.stanford.edu/group/candida).

PCR amplification conditions. (i) Singleplex amplification.

For all microsatelite loci, singleplex PCRs were performed with several different strains in order to evaluate the locus-specific amplification and to obtain alleles for sequencing. The locus CAI was amplified by following the conditions described by Sampaio et al. (42). The CAIII, CAV, CAVI, and CAVII loci were amplified by using 25 ng of genomic DNA in a 25-μl reaction volume containing 1× PCR buffer, 0.2 mM of each of the four deoxynucleoside triphosphates, 1.5 mM MgCl₂, 0.25 μM of each primer, and 1 U of Taq polymerase (GIBCO). After a 95°C preincubation step of 2 min, PCR amplifications were performed for a total of 30 cycles by using the following conditions: denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min, with a final extension step of 7 min at 72°C.

(ii) Multiplex amplification.

Multiplex PCR was performed by combining 1× PCR buffer (20 mM Tris HCl, pH 8.4, 50 mM KCl), 0.2 mM of each four of the deoxynucleoside triphosphates, 2 mM MgCl₂, 40 ng of genomic DNA, and 2 U of Taq Gold polymerase (Applied Biosystems) carried in a 25-μl final volume. The PCR program consisted of an initial denaturation step at 95°C for 10 min, followed by 30 cycles of 30 s at 94°C, 30 s at 58°C, and 1 min at 72°C, with a final extension step of 60 min at 68°C. The multiplex reaction was designed by combining microsatellites CAI, CAIII, and CAVI; and the primer concentrations used are depicted in Table 2.

TABLE 2.

Characteristics of the microsatellite loci selected

STR	No. of alleles	No. of genotypes	Size (bp)	Repetitive motif	DP	Chromosome no.	Primer concn (μM)	Dye labela
CAI	27	46	189-303	(CAA)2CTG(CAA)n	0.967	4	0.050	FAM
CAIII	6	12	95-110	(GAA)n	0.853	5	0.050	FAM
CAIV	—b	—	—	(ATT)n	—	—	—	—
CAV	21	21	102-198	(ATT)n	0.574	3
CAVI	30	39	224-388	(TAAA)n	0.933	2	0.225	HEX
CAVII	5	8	200-235	(CAAAT)n	0.670	1

^aFAM, 6-carboxyfluorescein; HEX, hexachlorofluorescein.

^b—, not determined.

DNA sequencing and fragment size determination.

All alleles observed for each locus were sequenced. Before sequencing, the alleles were separated and reamplified as described previously (42). The reamplified DNA fragments were purified with Microspin S-300 HR columns (Amersham Pharmacia Biotech, Quebec, Canada) and subjected to a dideoxy sequencing reaction with the BigDye terminator cycle sequencing ready reaction kit (Applied Biosystems). Following the sequencing reactions, the products were purified with AutoSeqG-50 columns (Amersham Pharmacia Biotech). Finally, the samples were dried, re-suspended in 15 μl of formamide, and run on an ABI PRISM 310 DNA automatic sequencer (Applied Biosystems). The results were analyzed by using Sequencing 3.7 analysis software. Determination of the fragment sizes of the PCR products and the allele sizes were done automatically with GeneScan 3.7 analysis software. The alleles have been designated by the number of repeated units determined after sequencing.

Reproducibility and stability.

The reproducibility of the method and the stability of the microsatellite markers were assessed according to the procedures described previously (42). The reproducibility of the method was also tested by comparing the results obtained by the singleplex analysis with the ones observed by the multiplex analysis.

RESULTS

Screening and selection of repeat regions in Candida albicans sequence database.

The search of the genomic DNA database of C. albicans performed in this study provided a total of 1,086 sequences, 368 obtained with the CAA query, 245 obtained with the GAA query, 220 obtained with the ATT query, 201 obtained with the TAAA query, and 52 obtained with the CAAAT query. Selection of the sequences was primarily based on the number of simple units repeated and their location in the genome. There should be at least 10 repeats, given that these sequences have a higher probability of showing greater genetic variability (11, 37), and they should be located outside known coding regions, since a higher polymorphism is expected in regions less prone to selective forces (32). Subsequently, a third selection was made for the sequences to allow the design of primers with an annealing temperature of about 60°C to ensure specificity and good reproducibility. A total of five sequences were then selected, and specific primers were designed for their amplification. Sequence 265080D05.s1.seq presented two different regions containing microsatellite motifs that were used separately to develop CAIII and CAIV (Table 1). The nomenclature chosen for the new markers was CA, after C. albicans, followed by a roman numeral notation that corresponded to the order of the analysis.

Microsatellite locus analysis.

The microsatellites selected were used to type 72 unrelated C. albicans strains, isolated from different patients, in order to test their specific amplification and polymorphism. This analysis revealed that they were all polymorphic and presented from 5 to 30 alleles and 8 to 46 different genotypes (Table 2). Analysis of the CAIV locus revealed the presence of null alleles; therefore, this marker was discarded.

The different alleles observed at all the loci were sequenced in order to determine the nature of the polymorphisms observed. The sequencing results confirmed that the consensus sequences of all new microsatellite loci analyzed were in accordance with the ones deposited in Stanford's DNA Sequencing and Technology Center database, confirming the amplification of the correct loci. CAIII, CAIV, CAV, CAVI, and CAVII were simple STRs with one variable repetitive motif, and only CAI was a compound microsatellite with two different variable units (Table 2). For all microsatellites the differences in the molecular weights of the distinct alleles reflected the differences in the number of the repeated motifs, which allowed the nomenclature for the alleles to be based upon the number of repeated units. This nomenclature prevents problems if further variation is found in the constant stretches of the alleles when other C. albicans strain typing studies are performed (12, 42).

The distribution of genotypes according to the Hardy-Weinberg equilibrium was investigated for all the microsatellite loci analyzed in the 72 unrelated isolates. A significant departure from Hardy-Weinberg equilibrium expectations was found (P < 0.001), which supported previous findings that the inheritance in infecting C. albicans populations appears to be clonal (15, 25, 50).

The discriminatory power was calculated for each marker according to the Simpson index of diversity:

where N is the number of strains, s is the total number of different genotypes, and n_j is the number of strains of j genotype (20). The results indicated that CAI is the microsatellite with the highest DP value, 0.97, while CAV presented the lowest DP value, 0.57 (Table 2).

Stability and specificity.

The in vitro stability of the microsatellite markers was assessed by growing four independent strains over approximately 300 generations. For all strains tested, the genotypes were always the same after the 300 generations, suggesting an expected mutation rate of less than 3.33 × 10⁻³ for all microsatellite loci. These markers were also species specific, since no amplification products were obtained when the primers and PCR conditions described above were used for the amplification of other pathogenic Candida species, namely, C. glabrata, C. krusei, C. parapsilosis, C. tropicalis, C. guilliermondii, C. lusitaniae, and C. dubliniensis. The specificity regarding C. dubliniensis, which is very closely related to C. albicans and which was just recently recognized as a different species, is noteworthy (47). Similar results were found in previous studies with microsatellites by testing other Candida species (5, 10, 31, 42).

Optimization of multiplex amplification conditions.

To increase the discriminatory capacity, simultaneous analysis of the different loci was carried out in a multiplex PCR. The CAI, CAIII, and CAVI loci were the ones selected for the multiplex PCR, since they presented the highest DP values in the singleplex assay and a DP value of 0.998 when analyzed together, while they are located on different chromosomes (Table 2). The differences in the sizes of the alleles amplified at each of these loci and the possibility of combining different fluorescent dyes made possible their simultaneous amplification (Fig. 1). For this optimization step, 10 C. albicans strains were used, and the results observed in the multiplex PCRs were always compared with the ones observed in the singleplex assay. The primer concentrations used for the multiplex reaction had to be adjusted, and the best results of multiplex coamplification with the three primer pairs were obtained by using the concentrations described in Table 2. Different DNA template concentrations were also tested, from 1 ng to 50 ng, and the optimal DNA concentration needed to coamplify these loci was 40 ng of template DNA in a final PCR volume of 25 μl. If a smaller amount of DNA was used, the preferential amplification of the CAI locus was observed. Comparison of the alleles obtained in the optimized multiplex reaction with the ones observed in the corresponding singleplex showed identical molecular weights, confirming the accuracy and reproducibility of the technique.

FIG. 1.

GenScan profiles depicting the results of automatic fragment sizing for the microsatellite multiplex analysis for (A) one strain from patient E (Table 4) and (B) strain 19C (Table 3). Each marker is represented by a different shade: black, CAI; dark gray, CAVI; light gray, CAIII.

Use of the multiplex assay for strain differentiation.

The CAI microsatellite was the one presenting the highest discriminatory power. However, several independent or unrelated isolates presented the same allele combination at this locus. These isolates were analyzed with the multiplex system in order to determine if they could indeed be considered the same strain. The genotypes observed with the multiplex assay indicated that, in several cases, the isolates with the same allelic combination at CAI corresponded to different strains or to strains undergoing a microevolutionary process (Table 3). The isolates were defined as identical when the same genotypes were observed at all loci, including the ones not presented in Table 3 (CAV and CAVII), and this was observed in 10 cases. On the contrary, 14 isolates were considered distinct strains since they presented different genotypes at several loci. The remaining three cases could be considered possible microevolutionary events, since minor changes in only one locus were observed. These results clearly showed the need for multilocus analysis, even when a highly discriminatory molecular marker is used.

TABLE 3.

Genotypes observed with microsatellites CAI, CAIII, and CAVI in a selection of strains that showed the same allele combination with CAI

Isolate	Genotype observeda			Institutionb	Body location
	CAI	CAIII	CAVI
IGC3436	17-23	6-6	14-14
37M	17-23	8-8	7-14	HSM	Urine
3	17-23	6-8	7-13	HSJ	Vaginal exudate
61M	17-21	6-8	15-15	HSM	Upper respiratory tract
24C	17-21	6-8	9-21	CSC	Vaginal exudate
1C	17-21	6-8	7-17	CSC	Vaginal exudate
39M	18-18	6-6	7-7	HSM	Upper respiratory tract
77M	18-18	6-6	7-7	HSM	Catheter
5C	18-18	6-6	7-7	CSC	Vaginal exudate
36C	11-18	6-8	6-42	CSC	Vaginal exsudate
45C	11-18	6-8	6-42	CSC	Vaginal exudate
WO1	16-27	6-6	7-7
27C	16-27	6-11	9-21	CSC	Vaginal exudate
26M	20-28	6-9	35-45	HSM	Upper respiratory tract
39C	20-28	6-9	39-44	CSC	Vaginal exudate
88M	20-28	6-9	42-44	HSM	Urine
74M	20-28	6-9	21-47	HSM	Urine
31	21-26	8-9	7-10	HSJ	Vaginal exudate
6C	21-26	8-9	7-10	CSC	Vaginal exudate
49C	21-26	6-7	9-15	CSC	Vaginal exudate
16C	21-26	6-7	9-16	CSC	Vaginal exudate
69M	21-25	6-7	9-15	HSM	Upper respiratory tract
8M	21-25	6-7	9-16	HSM	Urine
1M	21-25	6-7	9-16	HSM	Urine
31C	21-25	6-7	9-17	CSC	Vaginal exudate
11C	21-25	7-7	11-16	CSC	Vaginal exudate
17	21-21	6-7	8-15	HSJ	Vaginal exudate
52M	21-21	6-7	7-7	HSM	Urine
39	21-22	6-7	7-7	HSJ	Vaginal exudate
SER5	21-22	6-7	7-7	HSJ
41M	21-22	7-7	7-7	HSM	Urine
H58	22-22	6-7	7-7	HSJ
64M	22-22	6-7	7-7	HSM	Upper respiratory tract
20	26-26	6-6	7-7	HSJ	Vaginal exudate
4M	26-26	6-6	7-7	HSM	Peritoneal exudate
H37	25-27	6-8	16-22	HSJ	Upper respiratory tract
63M	25-27	6-6	9-17	HSM	Urine
22	23-27	6-8	21-21	HSJ	Vaginal exudate
41	23-27	6-8	18-20	HSJ	Vaginal exudate
35	25-25	7-7	9-15	HSJ	Vaginal exudate
45	25-25	7-7	9-15	HSJ	Vaginal exudate
19C	27-27	6-11	9-21	CSC	Vaginal exudate
46C	27-27	6-8	22-24	CSC	Vaginal exudate

^aCases of microevolution are marked in boldface.

^bHSM, Hospital S. Marcos; HSJ, Hospital S. João; CSC, Centro de Saúde do Carandá.

When this multiplex assay was applied to the differentiation of multiple isolates from the same patient, it could be observed that all isolates showed exactly the same genotypes, confirming our beliefs that only one strain was present in the infecting population (Table 4). To verify that the infecting population was the same at different body locations, isolates were collected from the same patient with infections at multiple body sites. Isolates were collected from patients F, H, I, and J. The results showed that the isolates from the upper respiratory tract of patient I were identical strains but were different from the urine isolate. The same occurred for patient J, for whom distinct genotypes were observed for the two isolates, one from the vagina and the other from urine. However, isolates from patients F and H showed the same genotypes regardless of their body location. These results show that patients could have different clones at different body sites but that the infecting population at each body site is monoclonal, with some isolates presenting the capacity of invading different microhabitats.

TABLE 4.

Characteristics of multiple strains isolated from the same patient and cultured simultaneously

Patienta	No. of isolates	Body location	Genotype observed
			CAI	CAIII	CAVI
A	3	Urine	21-25	6-7	9-16
B	4	Peritoneal exudate	26-26	6-6	7-7
C	2	Upper respiratory tract	17-17	6-8	7-14
D	2	Urine	21-21	6-7	7-7
E	5	Urine	21-22	7-7	7-7
F	1	Urine	18-47	6-11	9-19
F	1	Upper respiratory tract	18-47	6-11	9-19
G	2	Urine	36-36	6-11	6-41
H	1	Upper respiratory tract	21-25	6-7	9-15
H	2	Urine	21-25	6-7	9-15
I	2	Upper respiratory tract	22-22	6-7	7-7
I	1	Urine	20-28	6-9	42-44
J	1	Vaginal exudate	23-24	6-6	16-27
J	1	Urine	17-23	8-8	7-14
R	2	Vaginal exudate	21-25	6-7	9-17

^aEach patient is referred to as a separate letter (A to J and R).

Cases of recurrent vaginitis in eight patients were also analyzed with this multiplex system. The results showed that the infecting C. albicans strains isolated sequentially in different relapses of the illness displayed the same genotype for all loci, except in three cases, patients L, N, and P (Table 5). In patient N, a case of strain replacement occurred, since one of the three strains isolated presented different genotypes at all loci analyzed. In patient P, a microevolutionary event seems to have occurred, since the strains presented only a minor change at the CAVI locus, from genotype 18-21 to genotype 21-21. Analysis of the isolates from patient L showed a change at the CAI locus, from allele 30 to allele 32. In a previous study we suggested that these changes from allele 30 to allele 32 could have been due to expansion of the microsatellite (42). However, the analysis of this case with the remaining microsatellite loci suggests the occurrence of genetic changes that could lead to the loss of heterozygosity (LOH). Perepnikhatka et al. (34) studied the incubation of C. albicans isolates on fluconazole-containing medium and observed the loss of one homologue of chromosome 4 after 7 days of incubation. Interestingly, the CAI microsatellite is located on the same chromosome, and these patients were treated with fluconazole; therefore, this hypothesis could not be ruled out. The possible loss of chromosome 4, or part of it, which includes the CAI locus, would turn genotype 30-32 into genotype 30-30 in only one mutational step.

TABLE 5.

Genotypes of sequential isolates from vulvovaginal recurrent infections

Patienta	No. of recurrent strains	Genotype observedb
		CAI	CAIII	CAVI
K	4	17-23	6-8	7-13
L	3	30-32	6-6	19-23
		30-32	6-6	19-23
		30-30	6-6	19-23
M	5	18-25	6-6	7-7
N	3	21-21	6-7	8-15
		20-29	6-9	39-46
		20-29	6-9	39-46
O	2	26-26	6-6	7-7
P	2	23-27	6-8	18-21
		23-27	6-8	21-21
Q	2	22-23	6-8	24-24
S	2	20-20	9-9	31-37

^aEach patient is referred to as a separate letter (K to S).

^bCases of microevolution are marked in boldface.

DISCUSSION

Candida albicans is an opportunistic yeast that can cause severe invasive infections, especially in immunocompromised patients, thus making the development of methodologies for the accurate discrimination of strains essential. Knowledge of the relatedness of strains involved in infections would be of extreme relevance to the development and application of the correct therapeutic strategy as well as to obtaining a better understanding of the epidemiology of this yeast. Moreover, some strains are capable of invading different body locations, and microevolution may occur as an adaptive response to new environments.

Microsatellites are among the most frequent markers used for differentiation purposes due to their hypervariability, the ease of PCR amplification and interpretation, their codominant profiles, and their potential for use in automated assays. These loci are relatively abundant in C. albicans, and several markers have been developed (5, 12, 26). However, the majority of microsatellites appear to present low levels of heterozygosity, which may be explained by the fact that they are situated in coding regions (12). In these situations, the discriminatory power is rather low, but it can be compensated for by surveying a larger number of loci, facilitated by multiplexing the PCR. The best approach was obtained by combining three loci, CDC3, EF3, and HIS3, thus yielding a discriminatory power of 0.97 (3). In the present work, a multiplex strategy was developed with three new microsatellite markers, located in noncoding regions, which presented the highest DP observed for C. albicans differentiation, 0.99.

The analysis of 72 independent isolates with CAI identified 16 different genotypes shared by two or more isolates. Application of this multiplex assay to assessment of the genetic relatedness of those isolates allowed the discrimination of 14 of 27 (51.8%) strains that otherwise would have been considered identical. It also allowed the confirmation of the identities of 10 isolates, which were, in fact, the same strain, and three cases of microevolution. Microevolution resulted from minor changes of the strain genotypes, i.e., a change at only one locus that could be explained by a single mutational step. Overall, the improved discriminatory potential of the methodology described here revealed the presence of three basic scenarios: (i) isolates that were the same strain, (ii) isolates that were the same strain(s) but that were apparently undergoing a process of microevolution, and (iii) isolates that corresponded to distinct strains. These results clearly indicate the need for multilocus analysis even when isolates are studied with a highly discriminatory molecular marker. Application of this methodology to the analysis of isolates from recurrent infections also revealed the same set of basic scenarios described previously (23, 24, 44). Of the 15 recurrent cases analyzed, 12 were due to the same strain, 2 were due to the same strain which was undergoing microevolution, and only 1 was due to strain replacement. This multiplex system demonstrated that C. albicans strains can undergo microevolutionary events at body sites of carriage during colonization between recurrent episodes of infection and that these events seem to be relatively more frequent than strain replacement (6, 23, 24).

Analysis of the microevolutionary events observed in this study suggested that two different mutational events could be occurring: strand slippage of the DNA polymerase in the microsatellite region and LOH as the result of chromosomal rearrangements; both of these mutational events take place at similar rates. The reported rates for chromosomal rearrangements were from 1.2 × 10⁻³ to 3.0 × 10⁻³ (8, 41), while for microsatellite loci the values varied between 10⁻² and 10⁻⁶ (14, 19, 53, 54). Chromosomal alterations have been widely documented in C. albicans, with some being associated with the loss of entire or parts of chromosomes (27, 40, 41, 50, 52). Several conditions are known to induce LOH by chromosomal alterations, i.e., heat shock (17); exposure to different carbon sources, such as l-sorbose and d-arabinose (21, 41); and exposure to fluconazole (34). Metzgar et al. (31) also reported changes in microsatellites in vivo between the pretreatment isolate and the isolate obtained after treatment with different doses of fluconazole. The association of highly polymorphic STRs to different chromosomes and subchromosomal locations could be of great use for the evaluation of the loss of the entire chromosome or parts of chromosomes in C. albicans. Although LOH was observed at CAI, additional experiments are needed to confirm if it can be associated with rearrangements at chromosome 4 in response to fluconazole pressure.

Several studies have supported the concept that C. albicans contains a source of potentially beneficial genes that are activated by changes in chromosome number and that this elaborate mechanism regulates the utilization of food supplies and possibly other important functions in response to environmental stress. As C. albicans clinical populations are usually clonal, we suggest that in the absence of genetic exchange through sexual reproduction, microevolution could conceivably confer a selective advantage under adverse environmental conditions.

It is clear that the microsatellite multiplex PCR-based system described here enables high-speed typing, which makes it useful in large epidemiological studies. Furthermore, its capacity to detect microevolutionary events can make it useful for the detection of strain microevolution in response to environmental stress conditions, providing support for therapeutic adjustments, especially in patients with recurrent infections.

The standardization of microsatellite typing systems, including the primers and the separation techniques used and the allele nomenclature, is an issue that should be accomplished to allow interlaboratory comparisons. The creation of public databases that would make microsatellite allele data available worldwide, similar to those already in use for human microsatellites (29, 38), is another essential topic that deserves attention.