Quantitative variation of biofilms among strains in natural populations of Candida albicans

Xiaogang Li, Zhun Yan and Jianping Xu

Department of Biology, McMaster University, 1280 Main St West, Hamilton, Ontario, Canada L8S 4K1

Correspondence
Jianping Xu
jpxu{at}mcmaster.ca


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
This study examined the quantitative variation of biofilm formation and its relationship to multilocus genotypes in 115 strains of the human pathogenic fungus Candida albicans. These strains were isolated from three sources: 47 from oral cavities of healthy volunteers, 31 from the environment and 37 from the vaginas of patients with candidiasis. For each strain, biofilm formation was quantified as the ability to adhere to and grow on polystyrene plastic surfaces. Confocal laser scanning microscopy was used to visualize and confirm biofilm formation. Two methods were used to quantify biofilm formation abilities: (i) the XTT reduction assay, and (ii) absorbance following staining by crystal violet dye. Results obtained by the two methods were significantly correlated. Furthermore, biofilm formation ability was positively correlated with cell surface hydrophobicity. The analyses indicated that strains from each of the three sources varied widely in biofilm formation abilities. However, little correlation was observed between biofilm formation and multilocus genotypes as determined by PCR-RFLP at 16 polymorphic loci, regardless of source of strain. Strains with the same or similar multilocus genotypes often showed very different biofilm formation abilities. The results demonstrated that natural clones and clonal lineages of C. albicans exhibited extensive quantitative variation in biofilm formation.


Abbreviations: CLSM, confocal laser scanning microscopy; CSH, cell surface hydrophobicity; CV, crystal violet; XTT, 2, 3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Biofilms are aggregates of unicellular micro-organisms forming multicellular structures that adhere to surfaces. Their formation occurs in response to a variety of cues, including high cell density, nutrient deprivation and physical environmental stresses (O'Toole et al., 2000). Pathogenic bacteria and fungi can form biofilms on inert surfaces of implanted devices such as catheters, prosthetic heart valves and joint replacements (Tunnet, 1996; Hawser & Douglas, 1994; Donlan, 2001). Because biofilms are generally more resistant to both host defence mechanisms and to antibiotics than planktonic unicellular microbes, they represent an ongoing source of infection for many patients (Bagge et al., 2000; Baillie & Douglas, 1998; Donlan, 2001), and therefore are increasingly recognized as important health problems for patients with microbial infections. Although the majority of these infections are caused by bacteria, fungal infections are becoming increasingly common, especially those caused by species in the genus Candida, including Candida albicans.

Candida infection is particularly prevalent among hosts infected with the human immunodeficiency virus (HIV), and candidiasis is now recognized as one of the most important hospital-acquired infections (Cox & Perfect, 1993; Holmstrup & Samaranayake, 1990; Odds, 1988). The evolution of Candida species into important nosocomial pathogens is related to specific risk factors associated with modern medical therapeutics, such as the use of broad-spectrum antibiotics, cancer chemotherapy, immunosuppressive agents following organ transplantation and implanted medical devices (Cox & Perfect, 1993; Odds, 1988; Kuhn et al., 2002). While biofilm formation by human pathogenic yeasts has been recognized as a potentially important medical problem (Baillie & Douglas, 1999; Chandra et al., 2001), relatively little is known about the variation and evolution of biofilm formation within populations of these yeasts.

The increasing understanding of the epidemiology of fungal infectious diseases and the expanding collection of Candida species allow us to examine the patterns of variation and evolution of pathogenicity traits among natural isolates of these fungal pathogens. Here we refer to natural strains as those isolated from healthy hosts, patients or environments but that have not been experimentally manipulated in laboratories. Genetic analyses have identified that natural populations of C. albicans were predominantly clonal with some, perhaps cryptic, recombination (Pujol et al., 1993; Xu et al., 1999a, b; Xu & Mitchell, 2002). Following traditional terminology, strains with the same multilocus genotype are referred to as the same clone and strains with highly similar genotypes as the same clonal lineage. It has been hypothesized that in clonally evolving species, phenotypes should be constrained, with strains from each clonal lineage possessing phenotypic values more similar to each other than to strains from different clones or clonal lineages (Lynch & Walsh, 1998; Xu & Mitchell, 2002). This is because in the absence of sexual reproduction and recombination, genes controlling a phenotype cannot be exchanged among strains to create recombinant genotypes with diverse phenotypic values (Xu & Mitchell, 2002). Interestingly, previous studies have demonstrated that some strains of C. albicans showed tremendous phenotypic plasticity in the laboratory. One of the best-studied examples was colony morphology switching by laboratory strains of C. albicans (Soll, 1992, 2002). However, the extent of phenotypic diversity and its relation to genotypes are largely unexplored among natural strains of C. albicans.

The objective of this study was to test for potential constraints of genotypes on biofilm formation and to examine the extent of phenotypic diversity among natural strains of C. albicans. We examined a total of 115 strains of C. albicans isolated from three different sources: oral cavities of healthy volunteers, the environment and the vaginas of female patients with candidiasis. The phenotypic variation was analysed in the context of sources of isolation and multilocus genotypes as determined by PCR-RFLP at 16 polymorphic loci (Xu et al., 1999b). Polymorphisms at these 16 loci are presumed to reflect the genetic and evolutionary relationships among strains. To test the genetic basis for variation in biofilm formation among strains, we also quantified cell surface hydrophobicity (CSH) for all strains. Our results demonstrated significant correlation between biofilm formation abilities and CSH. However, neither biofilm formation nor CSH was correlated to strain genotype as determined by PCR-RFLP. These results suggested extensive diversity in biofilm formation among strains within the same clone or clonal lineage in natural populations of C. albicans.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Strains.
The strains used in this study were obtained from three sources: 47 from oral cavities of healthy volunteers, 31 from the environment and 37 from the vaginas of female patients with clinical candidiasis. For strains from humans, each was isolated from a different host. Environmental isolates were obtained from the effluents of sewage treatment plants. However, these strains are likely transient in the environment because currently, persistent populations of C. albicans are not known to exist in any ecological niche other than those in mammalian hosts and their immediate surrounding environments (Odds, 1988). In this study, since the original sources of these environmental strains are not known, we will use the term ‘environmental isolates' to reflect their current source of isolation.

All strains of C. albicans were identified based on the germ tube test, the formation of green colonies on CHROMagar Candida (CHROMagar), growth at 45 °C, and the fermentation and assimilation profiles on API 20C (BioMérieux). Isolates were maintained in 18 % (v/v) glycerol at -80 °C. For analysis of genotypic and phenotypic traits, strains were first transferred from stock cultures onto YEPD plates (1 % yeast extract, 2 % Bacto peptone, 2 % D-glucose, 1·5 % agar) and incubated at 37 °C. These fresh cultures were further prepared for quantifications of biofilm formation and CSH and for DNA extraction.

Genotype analysis by PCR-RFLP.
DNA extraction and genotyping by PCR-RFLP used methods as described previously by Xu et al. (1999a, b, 2000b). For each strain, the genotype was obtained at all 16 polymorphic restriction sites. These polymorphic sites were distributed among nine DNA fragments amplified by PCR and digested by 4-base cutter restriction enzymes (DdeI, HinfI, AluI, CfoI and MspI; Gibco BRL) (Xu et al., 1999a, b). Restriction endonuclease digests were performed according to the manufacturer's instructions. PCR products and restriction digests were electrophoresed in 1 % agarose in 1x times; TAE, stained with ethidium bromide, and viewed and photographed by the image analysis system Fluorochem-8800 (Alpha InnovTech). Genotypes at individual restriction sites were scored as previously described (Xu et al., 1999a, b).

Biofilm formation.
Protocols for inoculum preparation and biofilm growth followed those of Baillie & Douglas (1998, 1999) and Chandra et al. (2001). Briefly, to prepare inoculum, all strains were first streaked onto YEPD agar and incubated at 37 °C for 48 h. For each strain, a large loop of actively growing cells was transferred to sterile Yeast Nitrogen Base (YNB) broth (Difco Laboratories) containing 0·9 % D-glucose. After incubation at 37 °C for 24 h, the cells were centrifuged and washed twice with 0·5 ml PBS (0·14 M NaCl, 2·7 mM KCl, 8·5 mM Na2HPO4, 1·8 mM KH2PO4, pH 7·4) by vortexing and centrifuging at 5000 g for 5 min. The washed cells were then resuspended in 1 ml YNB broth. Optical density of cells was determined for each suspension (Ultraspec 2100 Pro spectrophotometer, BioChrom) and adjusted to a final OD600 value of 1·0 with YNB broth. These cell suspensions were then used to grow biofilms.

Many types of materials have been used for growing Candida biofilms in vitro, including polystyrene (e.g. San Millan et al., 1996; Ramage et al., 2001; Shin et al., 2002; Bachmann et al., 2002); polyvinyl chloride (Hawser & Douglas, 1994); silicone elastomer (Chandra et al., 2001); and polymethylmethacrylate (Chandra et al., 2001). Although polystyrene is not a material used in indwelling medical devices, it has been used widely for in vitro diagnostics and shown to be an excellent material for promoting adherence of cells (Merritt et al., 2000). In addition, a standardized method for biofilm formation based on polystyrene 96-well plates has been established (Ramage et al., 2001). The wide commercial availability of polystyrene plates in combination with standardized protocols for this material make it a material of choice for population surveys.

For each strain, 100 µl of the suspension (OD600 1·0) was inoculated into individual wells of polystyrene 96-well plates (flat bottom; Nunc). Three repeats were performed for each strain using a randomized block design (Sokal & Rolhf, 1981). YNB broth containing no inoculum was used as a negative control. The plates were incubated at 37 °C for 90 min (adhesion period). Supernatant including planktonic cells and liquid medium was then discarded and wells were gently washed twice with PBS to get rid of any non-adherent cells. For biofilm growth, 100 µl fresh YNB broth was then added to each well. The plates were covered, wrapped with Parafilm to prevent evaporation and incubated at 37 °C for 48 h. After biofilm formation and growth, planktonic cells were discarded through two rounds of washing with 200 µl sterile PBS buffer. Remaining cells sticking to the plastic surfaces were confirmed as biofilm cells through microscopy and these were quantified using two different methods. Both microscopy and biofilm quantification are described below.

Microscopy.
Following protocols described in the study by Chandra et al. (2001), confocal laser scanning microscopy (CLSM) was used to confirm that mature biofilms were formed after 48 h incubation. Because 96-well plates are not conducive for microscopic observations, biofilms for microscopy were prepared on polystyrene Petri-dishes. Culture preparation, cell adhesion and biofilm growth were performed exactly as those in 96-well plates. After biofilm formation, 4 ml PBS solution containing the fluorescent stains FUN-1 (10 µM, Molecular Probes) and Concanavalin A–Alexafluor 488 conjugate (Con A, 25 µg ml-1; Molecular Probes) were added to stain biofilms for 45 min at 37 °C. FUN-1 (excitation wavelength, 543 nm; emission, 560 nm long-pass filter) was taken up by metabolically active fungal cells and converted from a diffuse yellow colour to red. In contrast, Con A (excitation wavelength, 488 nm; emission, 505 nm long-pass filter) selectively binds to cell wall polysaccharides to produce green fluorescence. Stained biofilms were observed with a Bio-Rad Radiance 2000 CLSM, and images were captured and processed using LaserSharp 2000 software (Bio-Rad).

Quantifying biofilms.
Two methods were used to quantify biofilm formation abilities: (i) the XTT [2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide]-reduction assay; and (ii) crystal violet (CV) staining followed by measures of absorbance. The XTT-reduction assay was performed according to protocols used by Baillie & Douglas (1999) and Ramage et al. (2001). Briefly, 100 µl XTT salt solution (1 mg ml-1) (Sigma) and menadione solution (1 µM) (Molecular Probes; prepared in acetone) were added to each well containing mature biofilms. The cultures were then incubated in the dark at 37 °C for 5 h. Biofilm metabolism would reduce XTT tetrazolium salt to XTT formazan by mitochondrial dehydrogenases and result in a colorimetric change (Hawser, 1996; Ramage et al., 2001). The amount of colorimetric change was measured using a microplate reader at A490 (model EL340, Mandel Scientific). Microtitre wells containing only YNB broth but no microbe were used as negative controls.

In the method involving staining with CV, 100 µl 1 % CV was added to each well and incubated for 20 min at 37 °C. Next, 150 µl 95 % ethanol was added to dissolve the dyed biofilm cells and 100 µl each mixture was transferred to a new 96-well microplate. The absorbance for each well was determined using a microplate reader at A570 (model EL340, Mandel Scientific). Similarly, wells containing only YNB broth but no microbe were used as negative controls.

CSH.
Tests for CSH followed the protocol of Klotz et al. (1985). Briefly, cell suspensions of all strains were first prepared as described above for biofilm formation (OD600 1·0 in YNB broth). For each strain, 1·2 ml of the suspension was placed in a clean glass tube and overlaid with 0·3 ml octane. The mixtures were vortexed for 3 min and then the phases were allowed to separate. The OD600 was determined for the aqueous phase soon after the two phases had separated. OD values of strains in YNB broth without octane overlay were used as a negative control. Three repeats were performed for each strain. Following Hazen et al. (1986) and Klotz et al. (1985), the relative hydrophobicity was obtained as: [(OD600 of the control-OD600 after octane overlay)/OD600 of the control] xtimes;100

Data analyses
Biofilm and hydrophobicity data.
To compare variation and correlation within and among C. albicans samples, means and standard deviations (SD) of biofilm formation and CSH were calculated for each sample. The statistical significance in the difference among samples was compared by the t-test (Sokal & Rohlf, 1981). Pearson's correlation index was used to determine the relationship between hydrophobicity and biofilm formation in our samples and between results obtained from the two different biofilm quantifying methods.

Relationships among strains.
Similarity coefficients based on multilocus genotypes between pairs of isolates were calculated as the ratio of matches over the total number of alleles scored. An unweighted pair group method with arithmetic mean (UPGMA) phenogram showing the similarities of all strains was generated based on the pairwise similarity coefficient matrix. The calculations of similarity coefficients and the generation of the UPGMA phenogram were done using the software PAUP 4.0b5 (Swofford, 2002).

Genotypic diversities.
For quantitative comparisons of genotype diversities among samples, we used two methods. The first is Stoddart's genotypic diversity (Stoddart & Taylor, 1988), , where pi is the frequency of the ith multilocus genotype in the sample. This diversity index has a value that ranges from a minimum of 1, when all isolates are of the same genotype, to a maximum of n (the sample size), when every isolate has a different genotype. Differences of genotypic diversity were measured using a t-test between the percentages of maximum possible diversity determined for each collection. The formula of calculating variance followed that described by Chen et al. (1994), as follows:

The second diversity measure was Simpson's unbiased Index of Diversity (Simpson, 1949). Simpson's diversity ({lambda}) is calculated as where pi represents the frequency of a particular multilocus genotype and n is the sample size. This diversity has a value that ranges from a minimum of 0, when all isolates are of the same genotype, to a maximum of 1, when every isolate is a different genotype. The variance of {lambda} was calculated as (Simpson, 1949):

Statistical significances of differences in Stoddart's and Simpson's genotypic diversities between samples were then determined by t-tests.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Genotype distribution and genotypic diversities
A total of 56 unique multilocus genotypes were identified among the 115 strains. Genetic similarities among these strains and genotypes are presented in Fig. 1. Overall, shared genotypes were dispersed throughout the phenogram, regardless of isolation source. However, there were small clusters of source-specific genotypes or genotype groups (Fig. 1).



View larger version (20K):
[in this window]
[in a new window]
 
Fig. 1. (on facing page) Unweighted pair group with arithmetic means (UPGMA) phenogram describing genetic similarities among 115 strains of C. albicans. Genotypes and genotype groups based on the 16 PCR-RFLP loci were marked for comparison as described in the text. Strain designations reflect the source of isolation, followed by a numerical isolation number. O, strains from oral cavities of healthy volunteers; E, strains from the environment; V, strains from vaginas of patients with candidiasis. Clones and clonal lineages are described in the text (see Methods and Results). Strains marked by an asterisk (*) are specifically mentioned in the text.

 
The most common genotype included strains from all three sources and accounted for 29 of the 115 isolates (Fig. 1). Two other genotypes also contained strains from all three sources and included nine and five strains, respectively (Fig. 1). Four additional genotypes were shared by strains from two of the three sources and accounted for a total of nine isolates. Overall, 51 % (40/78) of isolates obtained from oral and environmental sources shared genotypes; 34 % (23/68) of isolates from environmental and vaginal sources shared genotypes; and 38 % (32/84) of isolates from oral and vaginal sources shared genotypes. Among the 56 multilocus genotypes, 15 were shared by more than one isolate; the remaining 41 were represented by only one isolate each (Fig. 1).

The genotype diversities of samples from different sources are presented in Table 1. Analyses with the two diversity measures, Stoddart's genotypic diversity and Simpson's unbiased Index of Diversity, yielded similar results. In both analyses, the population from patients with vaginal candidiasis had the highest diversity (Table 1). For example, the Stoddart's genotypic diversity (±SD) in the vaginal sample was 21·2±11·59. The oral sample from healthy volunteers had the lowest genotypic diversity, at 4·6±1·43. The environmental sample had an intermediate genotypic diversity of 9·2±7·95. Pairwise comparisons revealed that these differences were statistically significant (Table 1; P<0·05).


View this table:
[in this window]
[in a new window]
 
Table 1. Genotype diversities and biofilm formation within samples of C. albicans analysed in this study

 
Significant variation in biofilm formation abilities among strains
After 48 h incubation at 37 °C in YNB broth, mature biofilm formation was confirmed by CLSM (Fig. 2). These biofilms contained both fungal cells and extracellular matrices.



View larger version (53K):
[in this window]
[in a new window]
 
Fig. 2. Confirmation of mature biofilm of Candida albicans by CLSM. Metabolically active cells are shown in red (a) and extra-cellular polysaccharides are shown in green (b). (c) is a composite of (a) and (b). Metabolically active cells with extracellular matrices are shown in yellow. Note extensive green staining in both (b) and (c).

 
The amount of biofilm formation was significantly correlated between the two quantifying methods (correlation coefficient 0·96, P<0·001). Within each strain, there was very little variation among repeats. The standard deviation among repeats was low, typically 5–10 % of the mean by both methods. These results suggest that both quantifying methods were highly reproducible and random errors were small. The summary results obtained by the two methods are described below.

Results from XTT-reduction assay.
Summary results for biofilms from each source are presented in Table 1. In this assay, the mean A490 value for the 115 strains was 0·073 (±0·038, SD) with a range of 0·049–0·329, a 6·7-fold difference between the highest and lowest biofilm-producing strains. The mean biofilm forming abilities were similar for strains from the three different isolation sources. No statistically significant difference was detected among the three samples (P>0·05 in all pairwise comparisons) (Table 1). Variations in biofilm production among clones and clonal lineages are presented in Table 2 and discussed in a later section.


View this table:
[in this window]
[in a new window]
 
Table 2. Biofilm production among clones and clonal lineages in C. albicans

Genotype groupings correspond to those described in Fig. 1 and in the text. Biofilm values are presented as mean±SD (range). Wide variation in biofilm production was observed in most clones and clonal lineages. No statistically significant differences in biofilm formation were observed among these genotype groupings.

 
Results from CV staining.
Summary results for biofilms quantified by this method from each source are presented in Table 1. In general, the CV staining method revealed a greater range of variation of biofilm production among strains. The mean biofilms quantified by this method for the 115 strains was 0·189 (±0·301, SD) with a range of 0·043–2·179, a 50·7-fold difference between the highest and lowest biofilm-producing strains. Both the XTT-reduction assay and CV staining method detected that the same strain (O33) produced the highest level of biofilm. However, they detected different lowest ones. Strain V291 was the lowest biofilm producer (0·049) as determined by the XTT-reduction assay while strain V304 was the lowest (0·043) by the CV staining method. Their respective values by the other methods were 0·089 by CV staining for V291 and 0·050 by XTT reduction assay for V304. Similar to the results obtained by the XTT-reduction assay, the mean biofilm formation abilities were similar for strains from the three different isolation sources. No statistically significant difference was detected among the three samples (P>0·05 in all pairwise comparisons) (Table 1). More detailed analyses of variations in biofilm production among clones and clonal lineages are presented in Table 2 and discussed in a later section.

Positive correlation between biofilm and CSH
The CSH of 115 strains of C. albicans was measured by the biphasic separation method in YNB medium at 37 °C as described in Methods. Positive correlations were observed between biofilm formations as quantified by both methods and CSH. The correlation coefficient was 0·801 (P<0·001) between CSH and biofilms determined by the XTT-reduction assay. Similarly, a correlation coefficient of 0·765 (P<0·001) was obtained between CSH and biofilms determined by CV staining. Two oral strains (O33 and O35) with the highest CSHs (1·001 and 0·745 respectively) were also the biggest biofilm producers. This result suggests that CSH contributes significantly to biofilm formation.

Diversity in biofilm formation among clones and clonal lineages of C. albicans
Our combined genotypic and phenotypic analyses revealed extensive diversity in biofilm formation among clones and clonal lineages in natural populations of C. albicans (Table 2). Here, the diversity for a phenotype is defined as the range of phenotypic values exhibited by a clone or clonal lineage, in relation to phenotypic values from among unrelated clones or clonal lineages. In C. albicans, strains with the same multilocus genotype likely belong to a clone or a clonal lineage (Xu & Mitchell, 2002). The phenotypic diversity within a clone or a clonal lineage is demonstrated by the following two analyses. First, strains in the same clone often exhibit very different biofilm formation abilities (Table 2). For example, clone A in Fig. 1 had a range of biofilm formation ability from 0·051 (strain V303) to 0·105 (strain E104), a twofold difference, as determined by the XTT-reduction assay (Table 2) (a fourfold difference among strains was detected by the CV staining method in clone A). Strains in clones B, C and D also showed a wide variation, with biofilms quantified by CV staining showing greater variation than those by the XTT-reduction assay (Table 2).

Second, the lack of constraints by genetic relationships among strains on biofilms could also be demonstrated based on analyses of clonal lineages (Xu & Mitchell, 2002). In this analysis, 104 of the 115 strains were grouped into four such lineages (I, II, III and IV) based on a genetic similarity of 90 % among strains within a lineage (Fig. 1). The remaining 11 strains were genetically highly diverse and did not constitute a comparable lineage. These 11 strains were therefore classified as the ‘Other’ group (Fig. 1). Our analyses showed no significant differences in biofilm formation ability among these clonal lineages (Table 2; P>0·2 in all pairwise comparisons). All groups had wide variation in biofilm formation abilities as determined by either method (Table 2). In this analysis, genetic distances separating these lineages were arbitrary. However, analyses based on other a priori genetic distances did not change our general conclusion. Overall, there were wide variations in biofilm formation among strains within a lineage but no significant difference among lineages (Table 2). These results further indicated an overall lack of evolutionary constraint on biofilm formation abilities in C. albicans.

While our results suggested extensive diversity in biofilm formation within individual clones and clonal lineages, small clusters of strains with relatively uniform biofilm producing abilities can be found. Two specific groups containing strains with high biofilm formation abilities are worth mentioning. One group contained three strains, E105, E107 and E110 (Fig. 1). All three were from the environment and biofilms produced by these strains were 2–3 times the mean of the whole sample as determined by the XTT-reduction assay (a mean of 0·144 for the three strains vs the mean of 0·073 for the whole sample of 115 strains). When measured by the CV staining method, the differences increased (0·933 vs 0·189). Interestingly, a vaginal strain (V266) with the same multilocus genotype as E107 and E110 (Fig. 1) had a much lower biofilm formation ability of 0·062. The second group consisted of two strains, O33 and O35. Among the 115 strains, these two formed by far the most biofilms: over four times the population mean as determined by the XTT-reduction method (0·312 vs 0·073) and over 10 times by the CV staining method (2·174 vs 0·189).


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
In this study, we applied population genetics to analyse quantitative variation in biofilm formation in the human pathogenic yeast C. albicans. We found significant differences in genotype diversity among samples from the three different sources: oral, vaginal and environmental. Strains from each of the three sources exhibited wide variations in biofilm formation. Results from our two different biofilm quantifying methods were comparable and significantly correlated. The positive correlation observed between biofilm formation and CSH suggested that CSH plays a major role in biofilm formation in C. albicans. This result is consistent and extended previous observations of positive correlation between adhesion to plastic surfaces and CSH among Candida species (el-Azizi & Khardori, 1999; Klotz et al., 1985). Furthermore, the combined analyses of multilocus genotypes and biofilm formation indicated extensive phenotypic diversity in biofilm formation among clones and clonal lineages in natural populations of C. albicans.

Among the three samples, the vaginal sample had the highest genotypic diversity, followed by the environmental sample and the oral sample. While the mechanisms for these differences are not presently known, there are a couple of possibilities. The first is that the observed differences were purely a sampling effect. Samples derived from different groups of hosts, body sites or other physical environments may have patterns of genetic diversities different from the ones observed here. Indeed, subtle differences in genotypic diversity have been observed in oral yeast samples isolated from geographically or ecologically diverse groups of people (Xu et al., 1999a, b, 2000b; J. Xu, unpublished data). The second possibility is that ecological factors that differed among the three sampling sources did play a role in generating and maintaining genotypic diversity of C. albicans. The three isolation sources represented very diverse ecological niches and differ in many biotic and abiotic factors, e.g. other microbial species within the community, host factors (or lack of), temperature, nutrient levels, generation times and accessibility to external sources of yeast populations. Whether and how these and other potential factors contribute to the generation and maintenance of C. albicans genotypic diversity awaits further investigations.

In this study, we used polystyrene plates to grow biofilms. Because polystyrene is not a material used in indwelling medical devices, the results we obtained here may not correspond exactly to clinical biofilm formation abilities. It is possible that using other materials might produce results different from those based on polystyrene plates. Indeed, Baillie & Douglas (2000) found that C. albicans biofilms grown on two different polyvinyl chloride (PVC) catheters (supplied by different manufacturers) showed significant differences in susceptibility to Amphotericin B. However, for several reasons, we believe our results obtained from polystyrene materials are relevant and important for understanding fungal biofilms. First, polystyrene has been used widely for in vitro diagnostics in a variety of microorganisms and shown to be an excellent material for promoting adherence of cells (for a review see Merritt et al., 2000). Second, many studies of C. albicans biofilms used polystyrene (e.g. San Millan et al., 1996; Ramage et al., 2001; Shin et al., 2002; Bachmann et al., 2002), therefore, our results can be compared to these and other future studies. Third, an efficient method for biofilm formation in C. albicans based on polystyrene 96-well plates has been established and standardized (Ramage et al., 2001; Bachmann et al., 2002), and unlike many other materials, polystyrene plates are commercially available and relatively inexpensive. Fourth, the significant correlation between biofilm formation and CSH observed here suggests that adhesion to and growth on polystyrene surfaces reflected intrinsic structural and biochemical differences among strains of C. albicans.

Our results indicated an overall lack of difference in biofilm production among samples from the three examined sources. However, this result does not exclude the possibility that strains from other samples might show statistically different amounts of biofilms. Indeed, a recent study by Kuhn et al. (2002) showed that invasive isolates of C. albicans produced more biofilms than non-invasive strains when measured by dry weight. However, such a pattern was not observed when biofilms were measured by the XTT-reduction assay (Kuhn et al., 2002). The authors attributed the discordance of their results to several potential factors, including the presence/absence of a blastospore layer, differences in metabolic rates and the amount of extracellular matrices (Kuhn et al., 2002). While biofilm dry weight was not determined for our strains, results from the two methods examined in our study were highly correlated. It should be noted that Kuhn et al. (2002) only analysed a total of ten strains of C. albicans. These ten strains were from seven different body sites, with three strains from blood, two from catheters, and one each from denture, urine, vagina, skin and bronchia. Therefore, the conclusion that invasive isolates produced more biofilms needs to be confirmed using a larger sample from each of the body sites. Such a study could be done in a large medical centre where such collections are available. At present, our samples are limited, with most strains originated from oral cavities of healthy hosts.

To compare biofilm formation at the population level, it is critical to have an efficient and highly reproducible method for quantification. A number of methods and substrate materials have been used to quantify fungal biofilms, and all were adapted from methods reported previously for bacteria (Baillie & Douglas, 1998; Reynolds & Fink, 2001; Merritt et al., 2000). In this study, we used and compared two methods, and both were highly reproducible with little random error. The first method measured metabolic activities of biofilm-forming cells by using an XTT reduction assay. This method has been widely used by the fungal biofilm research community and is regarded as the standard method for testing C. albicans biofilms in vitro (Baillie & Douglas, 1999; Ramage et al., 2001; Chandra et al., 2001). Here we adopted the microtitre format of the XTT-reduction assay as applied by Ramage et al. (2001). In addition, we also used a direct staining method using CV, which is a basic dye. It binds to negatively charged extracellular molecules, including cell surface molecules and polysaccharides in the extracellular matrices in mature biofilms. The CV method is different from the XTT-reduction assay in that CV can stain both active cells and the extracellular matrices in mature biofilms (Hawser, 1996; Ramage et al., 2001). The CV staining method is widely used for measuring biofilms in bacteria (Merritt et al., 1998, 2000; Vidal et al., 1998; Watnick et al., 1999; Djordjevic et al., 2002). Our analyses showed that results from these two methods were significantly correlated, with a correlation coefficient of 0·96 (P<0·001). This correlation indicated that higher biofilm metabolic activities (due to either higher density of cells and/or higher metabolic activity per cell) likely generate higher amounts of extracellular matrices. Compared to the XTT-reduction assay, the CV staining method was cheaper and faster.

A typical laboratory fungal model of biofilm formation involves two operational steps: adhesion, and biofilm growth and maturation (Baillie & Douglas, 1999; Chandra et al., 2001). Using different species of Candida, previous researchers observed that fungal adherence to plastic surfaces was correlated with CSH (Hazen et al., 1986; Klotz et al., 1985; Silva et al., 1995; Samaranayake et al., 1995). In our study, we identified significant positive correlation between CSH and biofilm formation in a large sample of natural strains of C. albicans. While the relative contributions of hydrophobicity to the different steps of biofilm formation in C. albicans are not known, our results were consistent with previous findings that hydrophobicity was a major determinant of biofilm formation in C. albicans. The importance of CSH for adhesion to innate surfaces has also been demonstrated experimentally through gene knockout experiments in the baker's yeast Saccharomyces cerevisiae (Reynolds & Fink, 2001).

Our comparative analyses demonstrated abundant diversity in biofilm formation for clones and clonal lineages of C. albicans. Within most clones and clonal lineages, we found a wide variation in biofilm formation among strains. This pattern of genotype and phenotype relationship has been observed for other traits in C. albicans, including susceptibility to antifungal drugs (Cowen et al., 1999; Xu et al., 2000a), colony morphology (Hellstein et al., 1993) and ecological distributions (Xu et al., 1999a, b, 2000; see also Fig. 1 in this study). We would like to stress that, though the 16 PCR-RFLP markers could help define potentially divergent and robust evolutionary lineages in C. albicans, the objective here was not to use these genetic markers to define such evolutionary lineages. Instead, groupings as defined in our analyses served only to provide a simple way to visualize, analyse and present the quantitative diversities of biofilms in C. albicans. Delineating robust evolutionary lineages in C. albicans could be better achieved using DNA sequence polymorphisms at multiple nuclear and mitochondrial genomic regions (e.g. as done for the pathogenic basidiomycete yeast Cryptococcus neoformans; Xu et al., 2000a).

In the last 20 years, significant progresses have been made to address the long-standing evolutionary question of why a predominantly clonally reproducing species such as C. albicans could survive and persist in human populations and cause diseases (e.g. Soll, 2002). The results from this and other studies suggest that in the absence of sexual reproduction and recombination, populations of C. albicans could adapt to and successfully colonize human populations through its extraordinary phenotypic diversity (for a recent review, see Soll, 2002). Phenotypic diversity in pathogenic microbes could have significant practical implications as well. For example, if genotypes based on neutral markers can't be used as reliable predictors of medically important traits, targeted treatment strategies based on strain genotypes are less likely to succeed for pathogens exhibiting greater phenotypic diversities.

It should be noted that environmental and clinical biofilms rarely, if ever, consist only of cells from a single species. Indeed, earlier observations found that almost all yeast biofilms from medical grade silicone rubber contained bacteria (Neu et al., 1994). Co-aggregation and extensive interactions of Candida spp. with other bacteria and yeast could promote colonization and biofilm formation of yeast cells on host tissues and plastic surfaces (el-Azizi & Khardori, 1999; Adam et al., 2002). Additional studies are needed to determine whether genotypes and evolutionary relationships play any role in biofilm formation involving multiple species or strains.


   ACKNOWLEDGEMENTS
 
We thank Melissa Quick for strains from the environment and all the volunteers for their contribution to the collection of strains in this study. We thank Dr David Soll and two anonymous reviewers for their comments and suggestions. This research is supported by grants to J. Xu from McMaster University, the Natural Science and Engineering Research Council (NSERC) of Canada, the Premier's Research Excellence Award, the Canadian Foundation for Innovation (CFI) and the Ontario Innovation Trust (OIT).


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Adam, B., Baillie, G. S. & Douglas, L. J. (2002). Mixed species biofilms of Candida albicans and Staphylococcus epidermidis. J Med Microbiol 51, 344–349.[Abstract/Free Full Text]

Bachmann, S. P., VandeWalle, K., Ramage, G., Patterson, T. F., Wickes, B. L., Graybill, J. R. & Lopez-Ribot, J. L. (2002). In vitro activity of caspofungin against Candida albicans biofilms. Antimicrob Agents Chemother 46, 3591–3596.[Abstract/Free Full Text]

Bagge, N., Ciofu, O., Skovgaard, L. T. & Hoiby, N. (2000). Rapid development in vitro and in vivo of resistance to ceftazidime in biofilm-growing Pseudomonas aeruginosa due to chromosomal {beta}-lactamase. APMIS 108, 589–600.[CrossRef][Medline]

Baillie, G. S. & Douglas, L. J. (1998). Effect of growth rate on resistance of Candida albicans biofilms to antifungal agents. Antimicrob Agents Chemother 42, 1900–1905.[Abstract/Free Full Text]

Baillie, G. S. & Douglas, L. J. (1999). Candida biofilms and their susceptibility to antifungal agents. Methods Enzymol 310, 644–656.[Medline]

Baillie, G. S. & Douglas, L. J. (2000). Matrix polymers of Candida biofilms and their possible role in biofilm resistance to antifungal agents. J Antimicrob Chemother 46, 397–403.[Abstract/Free Full Text]

Chandra, J., Kuhn, D. M., Mukherjee, P. K., Hoyer, L. L., McCormick, T. & Ghannoum, M. A. (2001). Biofilm formation by the fungal pathogen Candida albicans: development, architecture, and drug resistance. J Bacteriol 83, 5385–5394.[CrossRef]

Chen, R.-S., Boegen, J. M. & McDonald, B. A. (1994). Genetic stability in a population of a plant pathogenic fungus over time. Mol Ecol 3, 209–218.

Cowen, L. E., Sirjusingh, C., Summerbell, R. C., Walmsley, S., Richardson, S., Kohn, L. M. & Anderson, J. B. (1999). Multilocus genotypes and DNA fingerprints do not predict variation in azole resistance among clinical isolates of Candida albicans. Antimicrob Agents Chemother 43, 2930–2938.[Abstract/Free Full Text]

Cox, G. M. & Perfect, J. R. (1993). Fungal infections. Curr Opin Infect Dis 6, 422–426.

Djordjevic, D., Wiedmann, M. & McLandsborough, L. A. (2002). Microtiter plate assay for assessment of Listeria monocytogenes biofilm formation. Appl Environ Microbiol 68, 2950–2958.[Abstract/Free Full Text]

Donlan, R. M. (2001). Biofilm formation: a clinically relevant microbiological process. Clin Infect Dis 33, 1387–1392.[CrossRef][Medline]

el-Azizi, M. & Khardori, N. (1999). Factors influencing adherence of Candida spp. to host tissues and plastic surfaces. Indian J Exp Biol 37, 941–951.[Medline]

Hawser, S. P. (1996). Adhesion of different Candida spp. to plastic: XTT formazan determinations. J Med Vet Mycol 34, 407–410.[Medline]

Hawser, S. P. & Douglas, L. J. (1994). Biofilm formation by Candida species on the surface of catheter material in vitro. Infect Immun 62, 915–921.[Abstract]

Hazen, K. C., Plotkin, B. J. & Klimas, D. M. (1986). Influence of growth on cell surface hydrophobicity of Candida albicans and Candida glabrata. Infect Immun 54, 269–271.[Medline]

Hellstein, J., Vawter-Hugart, H., Fotos, P., Schmid, J. & Soll, D. R. (1993). Genetic similarity and phenotypic diversity of commensal and pathogenic strains of Candida albicans isolated from the oral cavity. J Clin Microbiol 31, 3190–3199.[Abstract]

Holmstrup, P. & Samaranayake, L. P. (1990). Acute and AIDS-related oral candidoses. In Oral Candidosis, pp. 133–550. Edited by L. P. Samaranayake and T. W. Macfarlane. London: Wright.

Klotz, S. A., Drutz, D. J. & Zajic, J. E. (1985). Factors governing adherence of Candida species to plastic surfaces. Infect Immun 50, 97–101.[Medline]

Kuhn, D. M., Chandra, J., Mukherjee, P. K. & Ghannoum, M. A. (2002). Comparison of biofilms formed by Candida albicans and Candida parapsilosis on bioprosthetic surfaces. Infect Immun 70, 878–888.[Abstract/Free Full Text]

Lynch, M. & Walsh, B. (1998). Genetics and Analysis of Quantitative Traits. Sunderland, MA: Sinauer Associates.

Merritt, K., Gaind, A. & Anderson, J. M. (1998). Detection of bacterial adherence on biomedical polymers. J Biomed Mater Res 39, 415–422.[CrossRef][Medline]

Merritt, K., Hitchins, V. M. & Brown, S. A. (2000). Safety and cleaning of medical materials and devices. J Biomed Mater Res 53, 131–136.[CrossRef][Medline]

Neu, T. R., Verkerke, G. J., Herrmann, I. F., Schutt, H. K., Van der Mei, H. C. & Busscher, H. J. (1994). Microflora on explanted silicone rubber voice prostheses: taxonomy, hydrophobicity and electrophoretic mobility. J Appl Bacteriol 76, 521–528.[Medline]

Odds, F. C. (1988). Candida and Candidosis, 2nd edn. London: Bailliere Tindall.

O'Toole, G. A., Kaplan, H. B. & Kolter, R. (2000). Biofilm formation as microbial development. Annu Rev Microbiol 54, 49–79.[CrossRef][Medline]

Pujol, C., Reynes, J., Renaud, F., Raymond, M., Tibayrenc, M., Ayala, F. J., Janbon, F., Mallie, M. & Bastide, J.-M. (1993). The yeast Candida albicans has a clonal mode of reproduction in a population of infected human immunodeficiency virus-positive patients. Proc Natl Acad Sci U S A 90, 9456–9459.[Abstract]

Ramage, G., VandeWalle, K., Wickes, B. L. & Lopez-Ribot, J. L. (2001). Standardized method for in vitro antifungal susceptibility testing of Candida albicans biofilms. Antimicrob Agents Chemother 45, 2475–2479.[Abstract/Free Full Text]

Reynolds, T. B. & Fink, G. R. (2001). Bakers' yeast, a model for fungal biofilm formation. Science 291, 878–881.[Abstract/Free Full Text]

Samaranayake, Y. H., Wu, P. C., Samaranayake, L. P. & So, M. (1995). Relationship between the cell surface hydrophobicity and adherence of Candida krusei and Candida albicans to epithelia and denture acrylic surfaces. APMIS 103, 707–713.[Medline]

San Millan, R., Ezkurra, P. A., Quindos, G., Robert, R., Senet, J. M. & Ponton, J. (1996). Effect of monoclonal antibodies directed against Candida albicans cell wall antigens on the adhesion of the fungus to polystyrene. Microbiology 142, 2271–2277.[Abstract]

Shin, J. H., Kee, S. J., Shin, M. G., Kim, S. H., Shin, D. H., Lee, S. K., Suh, S. P. & Ryang, D. W. (2002). Biofilm production by isolates of Candida species recovered from nonneutropenic patients: comparison of bloodstream isolates with isolates from other sources. J Clin Microbiol 40, 1244–1248.[Abstract/Free Full Text]

Silva, T. M., Glee, P. M. & Hazen, K. C. (1995). Influence of cell surface hydrophobicity on attachment of Candida albicans to extracellular matrix proteins. J Med Vet Mycol 33, 117–122.[Medline]

Simpson, A. (1949). Measurement of diversity. Nature 163, 688.

Sokal, R. R. & Rohlf, F. J. (1981). Biometry: the Principles and Practice of Statistics in Biological Research, 2nd edn. New York: W. H. Freeman.

Soll, D. R. (1992). High-frequency switching in Candida albicans. Clin Microbiol Rev 5, 183–203.[Abstract]

Soll, D. R. (2002). Candida commensalism and virulence: the evolution of phenotypic plasticity. Acta Trop. 81, 101–110.

Stoddart, J. A. & Taylor, J. F. (1988). Genotypic diversity: estimation and predication in samples. Genetics 118, 705–711.[Abstract/Free Full Text]

Swofford, D. L. (2002). PAUP 4.0: Phylogenetic Analysis Using Parsimony. Sunderland, MA: Sinauer Associates.

Tunnet, M. M., Gorman, S. P. & Patrick, S. (1996). Infection associated with medical devices. Rev Med Microbiol 7, 195–205.

Vidal, O., Longin, R., Prigent-Combaret, C., Dorel, C., Hooreman, M. & Lejune, P. (1998). Isolation of an Escherichia coli K-12 mutant strain able to form biofilms on inert surfaces: involvement of a new ompR allele that increases curli expression. J Bacteriol 180, 2442–2449.[Abstract/Free Full Text]

Watnick, P. I., Fullner, K. J. & Kolter, R. (1999). A role for the mannose-sensitive hemagglutinin in biofilm formation by Vibrio cholerae El Tor. J Bacteriol 181, 3606–3609.[Abstract/Free Full Text]

Xu, J. & Mitchell, T. G. (2002). Strain variation and clonality in species of Candida and Cryptococcus neoformans. In Fungal Pathogenesis: Principles and Clinical Applications, pp. 739–750. Edited by R. Calderone and R. L. Cihlar. New York: Marcel Dekker.

Xu, J., Mitchell, T. G. & Vilgalys, R. (1999a). PCR-restriction fragment length polymorphism (RFLP) analyses reveal both extensive clonality and local genetic differences in Candida albicans. Mol Ecol 8, 59–73.[CrossRef][Medline]

Xu, J., Vilgalys, R. & Mitchell, T. G. (1999b). Lack of genetic differentiation between two geographic samples of Candida albicans isolated from patients infected with the Human Immunodeficiency Virus. J Bacteriol 181, 1369–1373.[Abstract/Free Full Text]

Xu, J., Ramos, A. R., Vilgalys, R. & Mitchell, T. G. (2000a). Clonal and spontaneous origins of fluconazole resistance in Candida albicans. J Clin Microbiol 38, 1214–1220.[Abstract/Free Full Text]

Xu, J., Vilgalys, R. & Mitchell, T. G. (2000b). Multiple gene genealogies reveal recent dispersion and hybridization in the human pathogenic fungus Cryptococcus neoformans. Mol Ecol 9, 1471–1481.[CrossRef][Medline]

Received 6 August 2002; revised 28 October 2002; accepted 7 November 2002.