ABSTRACT
Multilocus sequence typing (MLST) has emerged as a powerful new DNA-typing tool for the evaluation of intraspecies genetic relatedness. This method relies on DNA sequence analysis of nucleotide polymorphisms in housekeeping genes and has shown a high degree of intraspecies discriminatory power for bacterial and fungal pathogens. However, the results of the MLST scheme for Candida albicans have heretofore never been formally compared to those of other established typing techniques. To assess the value of MLST relative to those of other DNA fingerprinting tools for discriminating among strains of C. albicans, we applied it to a previously well-characterized set of 29 C. albicans isolates evaluated by the random amplified polymorphic DNA (RAPD), multilocus enzyme electrophoresis (MLEE), and Ca3 Southern hybridization probe techniques. MLST identified three clusters of genetically related isolates, with 82.3% direct concordance with MLEE, 82.7% with RAPD analysis, and 86.2% with the Ca3 Southern hybridization technique. When MLST was applied to a subset of 22 isolates of unrelated origins, it identified 21 independent diploid sequence types (DSTs), resulting in a discriminatory power of 99.6%. These DSTs were 96.9, 99.6, and 99.6% concordant with the genotypes identified by RAPD analysis, MLEE, and Ca3 Southern hybridization, respectively. These results demonstrate that MLST is a highly effective technique that performs at least comparably to other established DNA fingerprinting techniques.
The introduction of novel antifungal agents has helped stem the steady rise of systemic fungal infections observed over the years (4, 16, 17, 28). Nevertheless, nosocomial Candida albicans infections remain a major cause of morbidity and mortality among immunosuppressed patients (7, 22). In fact, a recent study showed that the mortality rate for patients with nosocomial candidemia is 61%, a 49% increase over that for other matched hospitalized patients (9). Successful treatment and prevention of these infections within the hospital setting depend not only on improved therapy but also on limitation of their spread through rapid and accurate detection of these pathogens. For this purpose, several image-based genotyping techniques have been developed and are widely used to characterize C. albicans strains. Unfortunately, these techniques are not well suited for rapid and high-throughput sample processing. They are also technically demanding and often require assumptions about hybridization and/or gel migration efficiency.
As opposed to image-based techniques, DNA sequence-based genotyping techniques are rapid and often rely on the nucleotide sequences of genes that are under stabilizing selective pressure (e.g., housekeeping genes). Typing schemes that use DNA sequence size and nucleotide polymorphisms have been shown to be effective for the identification of Candida species. For example, the nucleotide polymorphisms of a 396-bp fragment of the mitochondrial cytochrome b gene accurately distinguish between isolates of C. albicans, C. glabrata, C. parapsilosis, C. tropicalis, C. lusitaniae, and C. stellatoidea (1, 30). Similarly, amplicon size variations of the CaACT1 gene intron discriminate among isolates of C. dubliniensis and C. albicans (5). In addition, some degree of intraspecies discrimination was achieved by sequence analysis of these genes, underlining the utility of DNA sequencing for the accurate characterization of yeast pathogens.
In much the same way that it has been used for bacterial pathogens (27, 29), multilocus sequence typing (MLST) has emerged as an alternative typing tool that has a high degree of resolution and that has the capacity to rapidly characterize large numbers of clinical C. albicans isolates. MLST is based on the DNA sequence analysis of nucleotide polymorphisms within housekeeping genes, and it has shown a high degree of intraspecies discriminatory power for bacterial pathogens (13, 14, 23) and, most recently, fungal species, such as C. albicans (2, 26). MLST studies of C. albicans isolates (2, 3, 26) demonstrated that this technique is applicable to a diploid species and can effectively characterize sets of unrelated and related isolates.
However, MLST has not been formally validated by comparison to other conventional fingerprinting methods. Random amplified polymorphic DNA (RAPD) analysis, multilocus enzyme electrophoresis (MLEE), and Ca3 Southern hybridization, among others, have been shown to be effective for the study of both local and global epidemiological populations of Candida spp. (18, 24). MLEE can effectively identify genetic macrovariations that accumulate very slowly and that can thus be used to assess the phylogeny of C. albicans (24). Ca3 Southern hybridization can detect both rapidly and slowly accumulating genetic microvariations within strains, making it a suitable technique for characterization of both local and global populations of C. albicans (19, 25). It is important to determine if MLST is as reliable as other established DNA fingerprinting techniques, and in this study MLST was applied to a panel of C. albicans isolates (n = 29) that had previously been analyzed by RAPD analysis, MLEE, and Ca3 Southern hybridization (18). First, the genetic diversity found among the 29 C. albicans strains by MLST is reported. Second, the discriminatory power of MLST was compared to those of RAPD analysis, MLEE, and Ca3 Southern hybridization for different groups of strains. Finally, the congruence between MLST and the other typing techniques at various genetic depths was determined.
MATERIALS AND METHODS
C. albicans isolates.The 29 C. albicans strains used in this study have been described previously (18). All cells were maintained on glycerol stocks at −80°C and were grown on YPD broth (1% [wt/vol] yeast extract, 2% [wt/vol] peptone, 2% [wt/vol] dextrose [pH 5.7]). The strain collection included isolates FC-1 and FC-2 (switch phenotypes of laboratory strain 3153A) and isolates FC-3 and FC-4 (switch phenotypes of laboratory strain WO-1); the remaining 25 isolates (FC-5 to FC-29) were clinical isolates. As assigned previously (18), the strain collection was analyzed as three different groups: the first included all 29 C. albicans isolates, the second was a subset that included 22 isolates from unrelated hosts, and the third was a different subset that included seven isolate pairs that consisted of pairs of strains acquired from the same or related hosts (FC-11 and FC-12 [same patient], FC-13 and FC-14 [same patient], FC-17 and FC-18 [sexual partners], FC-19 and FC-20 [same patient], and FC-23 and FC-24 [sexual partners]) and the pairs of switch phenotype strains described above (these 14 isolates are referred to as “related-origin isolates”) (18).
Choice of loci.MLST based on seven housekeeping genes was performed for the 29 C. albicans strains. Six C. albicans housekeeping genes have been described previously for use in MLST: loci CaVPS13, CaADP1, CaRPN2, and CaSYA1 (2, 26) and loci CaACC1 and CaGLN4 (2). In this study C. albicans CaPMA1 (plasma membrane H+-ATPase) (15) was also included as a seventh gene to enhance the discriminatory power of MLST (Table 1).
Genes and primers used for MLSTa
Fragment amplification.Fragment amplifications were carried out in a 50-μl reaction volume containing 50 ng of genomic DNA, 0.5 μM each primer, 0.5 mM deoxynucleoside triphosphate mixture (Promega Corp.), 2.5 U of Triplemaster polymerase (Brinkmann), and 5.0 μl of 10× High Fidelity Buffer (Brinkmann). The PCR conditions used for all primer sets were 30 cycles of 95°C for 30 s, 60°C for 45 s, and 72°C for 1 min, followed by a final extension step of 72°C for 5 min; the PCRs were performed in a PTC100 96-well thermal cycler (MJ Research). DNA sequencing was performed by using the same primers used for PCR. All sequencing reactions were carried out in 20-μl reaction volume and analyzed with a CEQ 8000 Genomic Analysis System (Beckman). For all strains all seven loci were sequenced in both directions.
MLST data analysis.MLST was performed with each of the isolates as described previously (2, 26). For each gene, distinct alleles were identified and numbered by using the nonredundant databases program (http://calbicans.mlst.net/ ). The alleles at each of the seven loci constituted a strain's allelic profile, i.e., diploid sequence type (DST). Each distinct allelic profile was considered a unique DST, or genotype. A dendrogram based on the pairwise differences in the allelic profiles of the seven genes was constructed by the unweighted pair group method with the arithmetic mean (UPGMA) by using the START program (http://www.medawar.ox.ac.uk/maiden/software.shtml ). BURST, a noncommercial algorithm previously designed for MLST of bacterial pathogens, was used to divide the 29 C. albicans strains into clusters of genetically related strains. The BURST algorithm groups strains according to their allelic profiles by using a user-specified group definition, which is the number of alleles that the isolates need to have in common to belong to the same group (http://www.mlst.net/ ). The relatedness among the 29 C. albicans strains was also assessed by SplitsTree analysis (11), an alternative algorithm for analysis and visualization of evolutionary data. Typing of the 29 C. albicans strains by RAPD analysis, MLEE, and Ca3 Southern hybridization was done previously (18).
Discriminatory power.The discriminatory power of each typing technique for all three isolate groups was measured with Simpson's index of diversity (10), which calculates the probability that any two isolates will have different genotypes. The genotypes assigned by RAPD analysis, MLEE, and Ca3 Southern hybridization were determined at the most discriminatory level of each previously published respective dendrogram (18); any two isolates not labeled as identical were given a different genotype. For MLST the genotypes were based on individual DSTs.
Congruence between techniques.As defined and performed previously (18), for RAPD analysis, MLEE, and Ca3 Southern hybridization, the average SAB values (i.e., measures of genetic relatedness) for the 29 C. albicans defined the cutoff points used to group the strains into genetically related clusters. For MLST, the number of alleles from the seven loci chosen to group strains into genetically related clusters was determined by exploring the use of different numbers of alleles and determining which resulting cluster was in closest agreement to the clusters identified by the other techniques. To test the agreement between MLST and the other techniques on the placement of isolates into clusters, two-by-two tables were constructed and evaluated by Fisher's exact test (21). Statistical significance was defined as a P value <0.05. The congruence between techniques was further evaluated by cross-classification analysis (12, 21) at the individual genotype and cluster genetic depths of the dendrogram as follows. For each possible pair of isolates (n = 91 for the 14 related-origin isolates, n = 231 for the set of unrelated isolates, and n = 406 for all 29 isolates), it was determined whether the strains comprising the pair were or were not from the same cluster or genotype by each typing technique. A two-by-two table was constructed for each two-technique comparison, and the percent concordance was calculated.
RESULTS
Genetic diversity.MLST was performed by evaluating the DNA sequences of segments from seven different housekeeping genes, which yielded a set of 2,834 nucleotides for each strain. Sixty-one variable nucleotide sites were identified among the 29 C. albicans strains, of which 55 were heterozygous. The number of variable nucleotides at each locus ranged from 5 to 12, indicating a sizable amount of genetic diversity for each of the loci chosen (Table 1). The resulting proportion of variable nucleotide sites for the seven housekeeping genes was 2.2% (61 of 2,834 nucleotides), which is comparable to the results obtained in two recent MLST studies with C. albicans (2, 26) that showed 2.9 and 2.8% nucleotide site variabilities, respectively. The nucleotides present at each variable site and their positions relative to the gene fragment sequenced are shown in Fig. 1. The majority of these variable nucleotide sites had been identified previously (2, 26), and the ones newly identified in this study (n = 15) are highlighted in Fig. 1.
Variable nucleotide sites and alleles identified at each locus. The numbers in the vertical format represent the positions of the variable nucleotides relative to the fragment sequenced. The highlighted nucleotide positions were newly identified as variable in this study; the others were identified previously (2, 26). The nucleotides present at each variable site for the 29 C. albicans isolates analyzed are shown for all alleles; heterozygous variable sites are represented as follows: K, G or T; M, A or C; R, A or G; S, C or G; W, A or T; Y, C or T. The numbers in parentheses represent the number of isolates with that allele.
The number of alleles identified for the 29 C. albicans strains varied from 4 to 15 per gene; CaPMA1 generated the least number of alleles (n = 4), while CaSYA1 generated the most (n = 15). The allelic diversity found at the seven loci resulted in 24 unique DSTs for the 29 isolates (Table 2). Each of the isolates in isolate pairs FC-1 and FC-2, FC-3 and FC-4, FC-5 and FC-6, FC-11 and FC-12, and FC-17 and FC-18 shared the same DST; and the remaining DSTs were each found in a single isolate only. On the basis of the number of alleles found at each of the seven loci and the diploid nature of C. albicans, MLST can theoretically resolve more than 9 million distinct DSTs: (8 × 14 × 8 × 10 × 9 × 15 × 4) × 2.
Isolate descriptions and MLST resultsa
Discriminatory power.Simpson's index of diversity for the genotypes distinguished by each of the four typing techniques was calculated for the three groups of isolates described in Materials and Methods (Table 3). Overall, MLST displayed a very high discriminatory power. As an example, the discriminatory power of MLST was 95.6% for the related-origin isolates and 99.6% for the unrelated isolates. In comparison to other typing techniques, the discriminatory power of MLST was higher than that of RAPD analysis, often equal to that of MLEE, and slightly lower than that of Ca3 Southern hybridization for all groups of strains (Table 3). For example, MLST identified 24 distinct DSTs for the 29 C. albicans isolates, resulting in a discriminatory power of 98.8%, compared to discriminatory powers of 95.0% for RAPD, 98.3% for MLEE, and 99.3% for Ca3 Southern hybridization. When MLST was applied to the 22 unrelated-origin isolates, it identified 21 distinct DSTs, resulting in a discriminatory power (99.6%) that was higher than that of RAPD analysis (96.5%), equal to that of MLEE (99.6%), and slightly lower than that of Ca3 Southern hybridization (100%). The slight differences in discriminatory power between MLST and Ca3 Southern hybridization can be attributed to the fact that MLST identified FC-5 as identical to FC-6 and FC-11 as identical to FC-12, while Ca3 Southern hybridization did not.
Discriminatory power of typing techniques and congruence between MLST and the other typing techniques for the three groups of C. albicans isolatesa
MLST data analysis.Three similar clusters of genetically related strains were identified among the 29 strains by all four typing techniques. For the RAPD, MLEE, and Ca3 Southern hybridization techniques, these clusters were determined previously (18) by using the average SAB relatedness value for the 29 C. albicans strains as a truncation point on each technique's dendrogram. After analysis of the MLST results with the BURST algorithm with different numbers of shared identical alleles as the inclusion criterion for a cluster, it was determined that any three shared alleles from seven loci formed clusters in optimal agreement with those of the other three techniques. This group definition obtained with the BURST algorithm divided the 29 C. albicans strains into three clusters of genetically related isolates and six outliers (isolates that were not included in any cluster): the first cluster was strains FC-1, FC-2, FC-5, FC-6, FC-11, FC-12, FC-15, FC-19, FC-20, FC-21, FC-25, and FC-27; the second cluster was strains FC-8, FC-13, FC-14, FC-17, FC-18, FC-22, FC-23, and FC-24; the third cluster was strains FC-7, FC-28, FC-and 29; and the outlier strains were FC-3, FC-4, FC-9, FC-10, FC16, and FC-26. These groups of strains corresponded to the clusters I, II, and III and outliers, respectively, identified by Pujol et al. (18). The differences between MLST and all three of the other techniques were as follows: isolates FC-3 and FC-4 were identified as outliers by MLST but were placed into a cluster by the other three techniques; and by MLST, isolates FC-19 and FC-20 were both placed into a cluster different from the cluster into which they were placed by the other techniques. A UPGMA dendrogram (left portion of Fig. 2) based on the pairwise differences in the allelic profiles of the seven genes shows that the strains clustered in a manner similar to that according to the results obtained with the BURST algorithm. Furthermore, the relatedness among the 29 C. albicans strains was also assessed by analysis with SplitsTree (11), an alternative algorithm for the analysis and visualization of evolutionary data not always best represented by a standard tree. The SplitsTree graph (right portion of Fig. 2) showed a clustering of strains highly similar to those obtained with both the BURST algorithm and the UPGMA dendrogram. This agreement indicates that the sequence data were reliable and accurately represented by these relatedness algorithms. It is also worth noting that these strain clusters did not correlate with either the geographic distribution or the body site of isolation of the strains.
Dendrogram and SplitsTree analyses. A dendrogram showing the genetic relatedness among the 29 C. albicans isolates is shown on the left; the corresponding allelic profile (in parentheses), DST (sequence type) number, and isolate identifier (FC numbers) are shown. The SplitsTree analysis (right) also demonstrates the genetic relatedness among the same set of isolates.
All (100%) of the isolates in the three clusters determined by MLST with the BURST algorithm were in the clusters determined by the other techniques; 82, 88, and 92% of the isolates in the clusters determined by RAPD analysis, MLEE, and Ca3 Southern hybridization, respectively, were in the clusters identified by MLST (Table 4). Specifically, the results of MLST in terms of the assignment of the strains into clusters or as outliers were in very strong agreement with those of MLEE (P = 0.005) and were in even stronger agreement with those of Ca3 Southern hybridization (P = 0.0006) (Table 4). Similar to an analysis done previously (18), the typing techniques were further compared by using the seven isolate pairs that comprise the 14 related-origin isolates by determining how often MLST versus how often the other techniques assigned the isolates within each pair identical genotypes or nonidentical genotypes. Table 5 lists how each technique classified each isolate pair. The MLST classifications agreed with those of MLEE, RAPD analysis, and Ca3 Southern hybridization for 57, 71, and 86% of the pairs, respectively. The high discriminatory power of MLST and its potential to discern more accurate genetic differences may be the cause of the lower agreement values for the related-origin isolates. Nevertheless, the results of MLST again had the highest levels of agreement with those of Ca3 Southern hybridization, considered the most accurate of the three techniques (18).
Association between the clusters identified by MLST and the other typing techniques
Agreement between techniques for related-origin isolatesa
Cross-classification analysis.To further test the agreement between MLST and the other techniques, cross-classification analyses were conducted at the deep genetic cluster level. The cluster concordance values resulting from these analyses between MLST and the other techniques for each group of isolates are shown in Table 3. Overall, the isolate clusters identified by MLST and the other techniques were substantially concordant, ranging from 72.5% for MLST and all of the other techniques for the related-origin isolates to 90.5% for MLST and Ca3 Southern hybridization for the unrelated isolates. As an example, the latter indicates that 90.5% of the time any two strains that were assigned to matching (i.e., the same) or mismatching (i.e., different) clusters by MLST, the exact matching and mismatching determinations were made by Ca3 Southern hybridization. For all 29 isolates, MLST and Ca3 Southern hybridization were in 86.2% direct concordance for cluster classification.
In addition to the cluster analysis, it was important to test for agreement between the typing techniques at a highly discriminating genotypic level in light of MLST's high degree of resolving power. The results of cross-classification analyses performed by using MLST DSTs and the genotypes distinguished by each of the other techniques are shown in Table 3. Overall, for all three groups of isolates analyzed, there was a very high degree of concordance between the DSTs and the genotypes obtained by the other techniques, ranging from 91.2% for the related-origin isolates by RAPD analysis to 99.6% for the unrelated isolates by Ca3 Southern hybridization and MLEE. Compared with MLST, the results of Ca3 Southern hybridization were less concordant with those of RAPD analysis (96.5%) and equally concordant with those of MLEE (99.6%) for the unrelated C. albicans strains. The very high level of agreement of MLST with Ca3 Southern hybridization at both the cluster and the individual genotype levels is important, as Ca3 Southern hybridization has been determined to be the most discriminating and effective of the fingerprinting techniques (18).
The concordance values between MLST and the other techniques were slightly higher for the unrelated isolates than for the related-origin isolates at the cluster and genotype levels (Table 3). For example, cluster concordance between MLST and all of the other techniques for the 14 related-origin isolates was 72.5%, but it increased to 83.1% with RAPD analysis, to 85.3% with MLEE, and to 90.5% with Ca3 Southern hybridization for the 22 unrelated C. albicans isolates. As indicated above, dendrogram analysis revealed that four related-origin isolates (isolates FC-3, FC-4, FC-19, and FC-20) were classified into clusters differently by MLST than by the other techniques. This resulted in a lowering of the cluster concordance values for the related-origin isolates more so than for the unrelated isolates, which had only two of these four isolates with discordant classifications. This was also the primary reason why genotype concordance values were generally higher than cluster concordance values. Also, at the genotype level it was observed that there was an important difference in the concordance values between related-origin isolates (91.2%) and unrelated isolates (96.9%) only between MLST and the least discriminatory typing method, RAPD analysis. This difference resulted simply because RAPD analysis distinguished fewer genotypes than MLST.
Evaluation of locus utility.In order to explore whether MLST could be based on fewer than seven loci and still cluster strains accurately, BURST analysis was performed with different numbers of loci. The sharing of any two alleles among six loci (i.e., excluding CaPMA1) or any two alleles among only five loci (i.e., excluding CaACC1 and CaPMA1) divided the 29 C. albicans strains into the exact same three clusters and outliers obtained by using the BURST group definition of three shared alleles from seven loci detailed in “MLST data analysis” above. Also, strain and genetic relationships determined with the UPGMA dendrogram based on the pairwise differences in the allelic profiles with the five loci were virtually the same as those determined with all seven loci (data not shown). Furthermore, the discriminatory power of MLST was in no way affected by the exclusion of CaPMA1 and CaACC1 (data not shown). Although nucleotide polymorphisms were detected in CaPMA1, they did not enhance the discriminatory power of MLST because they were in linkage disequilibrium with the genetic variation of the other loci. Such linkage is further evidence for the previously documented notion that C. albicans has a clonal population structure (18).
DISCUSSION
Although MLST has been developed as a tool for the typing of C. albicans, it had never been formally compared to the techniques already established for the study of C. albicans. Our results presented here show that MLST was more discriminating than RAPD analysis and was essentially as discriminating as MLEE and Ca3 Southern hybridization, with Ca3 Southern hybridization usually being accepted as the most discriminating of the image-based techniques and as being effective in epidemiological investigations (18, 24). The overall high discriminatory power of MLST detected here is corroborated by the results of Bougnoux et al. (2) and Tavanti et al. (26). Also, the discriminatory ability of MLST stems from nucleotide polymorphisms, changes that are presumably stable and accurate, whereas elsewhere (6) some image-based techniques have been shown to be hypervariable in certain cases and to inappropriately subdivide known related isolates. The allelic polymorphisms of the seven genes used for MLST in this study can potentially resolve more than 9 million distinct DSTs, and the probability that two isolates would incorrectly be considered identical is very minimal. Therefore, MLST can be used in lieu of other techniques as a highly informative and reliable tool for outbreak investigations and other local epidemiological study settings.
Even in light of its high discriminatory power, the results of MLST still had a high level of agreement with those of Ca3 Southern hybridization at various genetic depths, as demonstrated by the cross-classification analyses. Furthermore, it was revealed that in comparison with the results of Ca3 Southern hybridization, the results of MLST were, in certain cases, less congruent with those of MLEE and always less congruent with those of RAPD analysis. These results are largely because the RAPD and MLEE schemes produce fewer genotypes than MLST. Differences in the levels of agreement between techniques may also be due in part to the differences in some of the technique-specific methods of analysis used (e.g., some clustering and phylogenetic algorithms) or to MLST’s establishing more accurate genetic relationships because of its use of DNA sequence data. Our results also suggest that the present MLST scheme is as good at clustering related strains as MLEE, the well-established tool of evolutionary and population biologists. Consistent with the latter is the fact that SplitsTree analysis revealed nearly the same results as those revealed by MLST with the BURST algorithm, thus indicating that MLST produced sound phylogenetic information. Therefore, MLST can effectively be used for molecular typing in both global and long-term investigations and local investigations.
Among the techniques of RAPD analysis, MLEE, and Ca3 Southern hybridization, Ca3 Southern hybridization has been singled out as the most effective (18); and the results of MLST were almost always in greater agreement with those of Ca3 Southern hybridization than with those of the other techniques and performed more comparably to Ca3 Southern hybridization than to the other fingerprinting techniques. For the 29 isolates, the results of MLST and Ca3 Southern hybridization were in 86.2% concordance at the cluster level and 99.5% concordance at the genotype level. Such high levels of concordance and the similar discriminatory powers (for all 29 isolates, discriminatory powers were 98.8% for MLST and 99.3% for Ca3 Southern hybridization), along with MLST's DNA sequence-based accuracy and ease of performance, storage, and data comparison and sharing, make MLST more advantageous than Ca3 Southern hybridization. With the trend toward low-cost high-throughput DNA sequencing, MLST will become even more desirable.
A reliable, rapid, objective, and high-throughput molecular typing approach has certainly become more necessary in the health care environment for addressing reported nosocomial C. albicans infections. Gil-Lamaignere et al. (8) pointed out that large amounts of MLST data from clinical isolates accumulated from spatiotemporally diverse locations can easily be analyzed (a website for C. albicans MLST [http://calbicans.mlst.net/ ] developed by d'Enfert and colleagues contains such a database), whereas image-based techniques are usually more difficult because they require processing of isolates together for accurate visual comparisons. Also, as the use of MLST increases, more genotypes will be found and the technique will be further optimized. In this study 15 polymorphic nucleotide sites were newly identified (Fig. 1). We also found that the addition of CaPMA1 to the six other loci used for MLST did not contribute major discriminatory power. However, it added four more individual alleles, and if more isolates are characterized, it may indeed contribute substantially to the discriminatory ability of MLST (potentially up to 7,000 more DSTs). CaPMA1 polymorphisms also verified the strain relationships determined by use of the other loci. Nevertheless, if simplification or cost savings is a priority, five of the seven loci (i.e., all loci except CaPMA1 and CaACC1) perform nearly as accurately as all seven loci together. Four of the seven loci used in this study are the same loci found among a set of seven housekeeping genes recently proposed for use in C. albicans MLST (3). Therefore, if assays with this other proposed locus set were compared to the fingerprinting techniques analyzed here, the results should be similar to ours. Furthermore, by the use of comprehensive MLST data banks, highly informative nucleotide polymorphisms may be identified and the many benefits of MLST could be obtained by detection of just these sites by real-time PCR (20) and allele-specific probing techniques, such as the molecular beacon or Taqman technology. This would allow accurate and even quicker analysis of large numbers of isolates. We are in the process of evaluating whether such an approach is effective for C. albicans.
ACKNOWLEDGMENTS
We are grateful to David Soll for making available the 29 reference isolates used in this study.
FOOTNOTES
- Received 21 February 2004.
- Returned for modification 8 March 2004.
- Accepted 18 March 2004.
- Copyright © 2004 American Society for Microbiology
