Typing of Saccharomyces cerevisiae Clinical Strains by Using Microsatellite Sequence Polymorphism

It seems that S. cerevisiae, which was thought for about 30years to be a nonpathogenic yeast, should now be consideredan opportunistic pathogen. In this study, we estimated the discriminationability of the microsatellite sequence amplification techniquewithin a sample of clinical and reference S. cerevisiae strainsand S. boulardii reference strains.

INTRODUCTION

Top

Introduction

Saccharomyces cerevisiae is a saprophyte of the digestive, respiratory,and genitourinary tracts (2-4, 17, 18, 25). However, with theprogress in immunosuppressive therapies, there has been an increasingnumber of reports on yeast systemic infection cases involvingSaccharomyces strains, particularly S. cerevisiae. The epidemiologyof infections caused by Saccharomyces is still unknown. Someauthors suggested a correlation between fungemia caused by Saccharomycesand treatment (24). Moreover, S. cerevisiae is usually consideredto be responsible for infections. Nevertheless, some authorshave suggested the role of other Saccharomyces species in systemicinfections. Some of these infections may be due to treatmentof diarrhea by S. boulardii (9, 20). It would thus be necessaryto develop molecular markers for distinguishing various genotypeswithin S. cerevisiae.

Some epidemiological studies have shown intraspecific diversitywithin S. cerevisiae (11, 12). The present retrospective multicenterstudy exhibited polymorphism in microsatellite sequences of69 clinical strains of S. cerevisiae by comparison to S. cerevisiaeand S. boulardii reference strains. The aim of this work wasto study genotypic polymorphism within clinical strains of S.cerevisiae compared to S. boulardii reference strains.

MATERIALS AND METHODS

Top

Materials and Methods

In this study, 77 strains were tested; 69 came from differentFrench medical centers (Table 1). These strains were isolatedfrom different samples (Table 1). In parallel, five S. cerevisiaeand three S. boulardii reference strains were tested (Table1).

View this table:
[in this window]
[in a new window]
TABLE 1. Characteristics of the Saccharomyces strains studied

Yeast DNA was purified according to the protocol published byQuerol et al. (22). The five microsatellite sequences used inour study have been previously described (10). These sequencesare distributed in the S. cerevisiae genome (Table 2). PCR amplificationwas performed according to the protocol described by Durandet al. (8).

View this table:
[in this window]
[in a new window]
TABLE 2. Characteristics of the five polymorphic microsatellite loci tested in this study

After heat denaturation, PCR fragments were separated by electrophoresison a vertical polyacrylamide gel including formamide in theloading buffer.

The data were analyzed by factorial correspondence analysiswith the PRAXIS-p.c., version 2.0, software package (Praxème,R&D, Biométrie, C.N.R.S., Montpellier, France) (1,6).

Factorial correspondence analysis was presented as a plane projectionof the two most informative axes accounting for the geneticstructure of the electrophoretic types. Allelic frequencieswith each locus were calculated with the Genepop, version 1.2,software program (23).

RESULTS AND DISCUSSION

Top

Results and Discussion

The microsatellite sequence polymorphism analysis gave reproducibleresults. Each locus tested was polymorphic, with 6, 11, 6, 6,and 5 different alleles for the microsatellite sequences ofgenes YKL 172W, YLR 177W, YKL 139W, YKR 072C, and YDR 289C,respectively. The fact that certain strains presented two differentalleles for a given locus increased the number of patterns observed.In the present study, 12 pairs of different alleles were detectedfor the YKL 172W gene, 18 for YLR 177W, 11 for YKL 139W, 8 forYKR 072C, and 6 for YDR 289C. A predominant pair of alleleswas found in microsatellite sequences of genes YKR 172W, YKR072C, and YDR 289C, whereas for genes YLR 177W and YKL 139W,the allele pairs were distributed in a homogeneous way withoutany of them dominating (Table 3).

View this table:
[in this window]
[in a new window]
TABLE 3. Allelic frequencies for the five loci tested^a

In the present work, 34 different alleles were observed andgenerated 64 distinct electrophoretic types. Among these 64electrophoretic types, 58 represented a single strain, threerepresented two strains (electrophoretic types 26, 30, and 53),one represented four strains (electrophoretic type 14), onerepresented three strains (electrophoretic type 64), and onerepresented five strains which were isolated from the same patient(electrophoretic type 54). On the factorial correspondence analysis,electrophoretic types are projected in the most informativeplane defined by axes 1 and 2, which explained 21.29% of theoverall genetic variability.

The electrophoretic type projection exhibited three distinctclusters (Fig. 1). Cluster I included only clinical strainsof S. cerevisiae. Cluster II included reference strains of S.cerevisiae, and cluster III contained reference strains of S.boulardii. These three clusters were separated by axis 1, whichwas the most informative. The loci responsible for the structureof the sample were highlighted by the allelic frequencies indexedin Table 3. Allele pairs responsible for the structure of thesample were the least frequent (i.e., 177-30, 177-31, 13-32,072-33, and 289-34).

View larger version (21K):
[in this window]
[in a new window]
FIG. 1. First factorial correspondence analysis plane projection in two informative axes of all electrophoretic types observed. Axes 1 and 2 represent 11.73% and 9.56% of the overall variability, respectively. Cluster I includes all S. cerevisiae clinical strains, cluster II includes S. cerevisiae reference strains, and cluster III includes S. boulardii reference strains. Clusters are separated on axis 1, which is the most informative.

Genotype differences between the electrophoretic types of clusterI were investigated with a second factorial correspondence analysis(data not shown). Genotypes were not linked to a geographicalorigin or anatomic site. No conclusions can thus be drawn atpresent, but it would be interesting to conduct a further studywith a larger sample of isolates.

In this study the sample assessed was interesting, related exclusivelyto clinical S. cerevisiae strains. This sample should be increased,but it is very difficult to obtain strictly clinical strainsof S. cerevisiae, which explains why the present sample waslimited in size and spatiotemporally heterogeneous.

Sixty-four electrophoretic types for the 77 strains were foundby the microsatellite sequence polymorphism technique. Theseresults are in agreement with those of Zerva et al. by usingrestriction fragment length polymorphism (26). Joly et al. (11)with randomly amplified polymorphic DNA and Duarte et al. (7)with multilocus enzyme electrophoresis on S. cerevisiae found48 types for 54 isolates and 32 types for 35 studied strains,respectively. Hennequin et al. (10) differentiated 52 typesfor 91 strains tested by the microsatellite technique. Thiscould be explained by the high homogeneity of the clinical strainsamples studied. In the work of Lewicka et al. (12), only 22types for 52 strains by multilocus enzyme electrophoresis wereobserved. However, it appears that in all of these studies,there was high diversity in the clinical and nonclinical S.cerevisiae strains.

In the present study, the clinical strains of S. cerevisiae(cluster I) seemed to have genotypes distinct from those ofthe S. cerevisiae reference strains studied (cluster II). Thisis in agreement with the work of Clemons et al. (5) in whichthe clinical and nonclinical strains showed different degreesof virulence. These results are of epidemiological importance,because in the present work, environmental strains had electrophoreticprofiles which differed from those of strains isolated frompatients. The genotypes of the S. cerevisiae strains responsiblefor infection and those isolated from the environment were verydifferent. Microsatellite typing also revealed a genotypic differencebetween the S. cerevisiae and S. boulardii strains tested. Thisis in agreement with Mallié et al. (13) and McFarlandet al. (15), who differentiated S. cerevisiae from S. boulardiiby multilocus enzyme electrophoresis.

However, some authors such as Molnar et al. (16), Perapoch etal., and McCullough et al. demonstrated, by randomly amplifiedpolymorphic DNA and restriction fragment length polymorphism(14, 19), respectively, that S. cerevisiae and S. boulardiihave identical genetic profiles. Moreover, Hennequin et al.(10) found the same profile for clinical Saccharomyces strainsand S. boulardii strains. Nevertheless, no variations in thegenotype of S. boulardii have ever been observed (10, 21), whichis in agreement with the results of the present study. Indeed,the same electrophoretic type was found for the three S. boulardiireference strains tested. This work exhibited a difference betweenclinical strains of S. cerevisiae (cluster I) and referencestrains of S. boulardii (cluster III). These results shouldbe confirmed by a comparative study on a larger sample of S.cerevisiae and S. boulardii strains.

By the description of 64 different electrophoretic types forthe 77 strains tested, the microsatellite sequence polymorphismtechnique highlighted significant intraspecific diversity withinS. cerevisiae for both the clinical and reference strains. Microsatellitesequence polymorphism is a powerful molecular typing tool. Thepresent work demonstrated a difference between clinical S. cerevisiaeand S. boulardii strains, suggesting that S. boulardii was notresponsible for the systemic fungal infections observed withinthe patients studied. It would be useful to carry out a comparativestudy with a broader sample of S. cerevisiae and S. boulardiiin order to determine their roles in human pathological infections.This could help to clarify the transmission mechanisms of thesefungi and identify the contamination risk factors so that moreeffective prevention measures can be implemented.

ACKNOWLEDGMENTS

We thank Biocodex laboratoires, A. Blancard, B. Couprie, F.Dromer, G. Galeazzi, C. Guiguen, M. D. Linas, and J. Reynesfor supplying strains.

FOOTNOTES

* Corresponding author. Mailing address: Laboratoire de Parasitologie et Mycologie Médicale, MNERT EA 2413, Faculté de Pharmacie, 15 Av. Charles Flahault, B.P. 14 491, 34093 Montpellier cedex 5, France. Phone: 04.67.63.52.02. Fax: 04.67.41.16.17. E-mail: michele.mallie{at}univ-montp1.fr.

REFERENCES

Top