The genus
Malassezia comprises lipophilic yeasts found in the normal flora of human skin and other mammals. These yeasts were described by Eichstedt in 1848 as being associated with pityriasis versicolor (PV) lesions (
13). The taxonomy and nomenclature of the genus
Malassezia was controversial for many decades. Indeed, until 1990 only three species were recognized:
M. furfur,
M. sympodialis, and
M. pachydermatis, a non-lipid-dependent species (
17,
21,
38). The species
M. globosa,
M. restricta,
M. obtusa, and
M. slooffiae were described in 1995 by morphology, ultrastructure, physiology, and molecular techniques (
18,
20). In the last few years,
M. dermatis,
M. japonica,
M. nana, and
M. yamatoensis (
26,
40-
42) have been reported as
Malassezia species. Recently, Cabañes et al. described two new species,
M. equina and
M. caprae, which were isolated from domestic animals (
4).
Nine of the 13 species within the genus,
M. furfur,
M. sympodialis,
M. globosa,
M. restricta,
M. slooffiae,
M. obtusa,
M. dermatis,
M. japonica, and
M. yamatoensis, are associated with normal human flora and pathologies. Four species,
M. pachydermatis,
M. nana,
M. equina, and
M. caprae, are associated with animals (
4,
6,
11,
17,
18,
20,
21,
26,
34,
38,
40-
42).
Malassezia species have been associated with diverse dermatological pathologies, including PV, seborrheic dermatitis dandruff, atopic dermatitis, folliculitis, psoriasis, onicomycosis, and blepharitis.
M. furfur and
M. pachydermatis have been associated with systemic infections in patients with underlying diseases and those receiving intravenous lipid emulsions (
6,
7,
9-
11,
16,
29,
33,
34).
Although the role of
Malassezia species in the development of these diseases is not clear, some authors suggest that
M. globosa is the causal agent of PV, while others have found a greater percentage of isolates of
M. sympodialis associated with the disease. Differences in diagnosis might be due to sampling methods and differences between the culture media used, leading to controversies in clinical studies of these dermatological pathologies (
1,
7,
9,
10).
New physiological patterns for identification have been described, and recently the availability of molecular biology and sequencing techniques has allowed the species to be distinguished more clearly (
17-
21). Despite the difficulty in isolating, maintaining, and identifying these yeasts, different characteristics of the genus, such as macroscopic and microscopic morphology and some physiological aspects (e.g., the presence/absence of catalase, selective growth on Cremophor EL, β-glucosidase activity, and growth and pigment production on a pigment-producing medium [p-agar]), allow them to be differentiated (
18,
22,
30,
31). The assimilation of Tween 20, 40, 60, and 80 by
M. furfur,
M. sympodialis, and
M. slooffiae yields a specific pattern that easily differentiates them from other species (
18,
22).
In a previous study, we isolated a total of 154 strains of
Malassezia spp., 7 of which were omitted and remained uncharacterized because of an atypical Tween assimilation pattern (
35). The aim of the present study was to characterize these seven isolates by investigating their physiological features and conducting morphological and molecular characterization by PCR-restriction fragment length polymorphism (RFLP) and the sequencing of the 26S and 5.8S ribosomal DNA-internal transcribed spacer 2 (rDNA-ITS2) regions for all of the strains. Establishing a relationship with reported species through phylogenetic analyses will support our findings.
MATERIALS AND METHODS
Strains and growth conditions.
The strains of
Malassezia spp. with a Tween assimilation pattern different from that of the other strains were obtained as uncharacterized isolates from a previous study (
35). All strains were preserved in 10% skim milk at −80°C (
8) and recovered in modified Dixon agar by incubation for 4 to 5 days at 32°C (
18,
22).
Phenotypic characterization.
The morphological characteristics of the colonies were recorded, and Gram staining was performed. The morphology of the stained cells was assessed by light microscopy. Physiological characteristics were assessed as previously described. The following tests were performed five times: assimilation (Tween 20, 40, 60, and 80 and Cremophor EL), enzyme production (β-glucosidase and catalase), pigment production on tryptophan-based medium, and growth (on Sabouraud, Sabouraud plus 10% Tween 20, and Dixon agar at 37 and 42°C) (
18,
22,
30,
31).
Molecular characterization. (i) DNA extraction.
Colonies grown on Dixon agar at 32°C for 4 to 5 days were transferred to 600 μl sterile distilled water and centrifuged at 10,000 rpm for 5 min. DNA was extracted as previously described (
36).
(ii) PCR-RFLP of 26S and 5.8S rDNA-ITS2 regions.
The 26S rDNA region was PCR amplified using forward and reverse primers as previously described (
32). The 5.8S rDNA-ITS2 regions were PCR amplified using primers ITS3 and ITS4 (
15). All PCRs were performed in a final volume of 25 μl (
46).
The amplified 26S rDNA product was digested with 10 U of the restriction enzymes CfoI (Sigma-Aldrich) and BstF5I (SibEnzyme, Novosibirsk, Russia) by incubation for 4 h at 37 and 65°C, respectively. The amplified 5.8S rDNA-ITS2 product was digested with 10 U of AluI (Promega) by incubation at 37°C for 4 h. The products were visualized by 2% agarose gel electrophoresis using Quantity One software (Bio-Rad, Hercules, CA).
(iii) Sequencing and phylogenetic analysis.
Single electrophoretic bands from the PCR products were sequenced using a 3730xl DNA analyzer (PE Applied Biosystems). Sequence assembly and editing were performed manually on the software CLC DNA Workbench.
For each region (26S and 5.8S rDNA-ITS2), a data set was formed containing sequences from reported
Malassezia species in GenBank. Each data set was aligned using the ClustalW algorithm (
45) under default settings in the BIODEDIT package (
25). To assess phylogenetic relationships, maximum-parsimony analysis was conducted on the aligned data sets using PAUP* 4.0b8 software under default settings (
43), and
Filobasidiella neoformans was used as the outgroup for both analyses. A bootstrap resampling was conducted for 1,000 replicates to assess relative branch support (
14).
Nucleotide sequence accession number.
The nucleotide sequences (26S and 5.8S rDNA-ITS2 regions) obtained in this study have been deposited in GenBank with accession numbers EU815303 to EU815322 (National Center for Biotechnology Information, Bethesda, MD).
RESULTS
To characterize the isolates of
Malassezia spp., we studied their physiological features and conducted morphological and molecular characterization by PCR-RFLP and the sequencing of the 26S and 5.8S rDNA-ITS2 regions in three
M. furfur strains from healthy individuals, four
M. furfur clinical strains, and eight
Malassezia spp. reference strains (Table
1).
Phenotypic characterization.
All
M. furfur colonies were smooth, opaque, and umbonate. Microscopically, all
M. furfur isolates appeared as small ovoid cells (1 to 1.5 by 2 to 2.5 μm), and buds were formed on a broad base. The physiological tests other than Tween assimilation showed the results expected for
M. furfur (Table
2). The Tween assimilation test result was different from the standard patterns so far reported for different
Malassezia species but coincided with that of the
M. furfur CBS 6094 strain. Tween 80 was the only Tween assimilated, as shown in Fig.
1. However,
M. furfur CBS 6094 does not produce pigment at 25, 30, 37, or 42°C.
Molecular characterization.
The 26S rDNA PCR amplified an ∼580-bp product; two bands of ∼180 and ∼400 bp were obtained after BstF51 digestion, and three bands of ∼260, ∼130, and ∼70 bp were obtained after CfoI digestion. The 5.8S-ITS2 PCR amplified an ∼500-bp product. AluI digestion yielded two bands of ∼290 and 250 bp (Table
3).
Phylogenetic analysis.
The isolates with the atypical Tween assimilation pattern formed a single cluster that contained the reference
M. furfur sequences, with high bootstrap support values of 80 and 100% for the 26S and 5.8S rDNA-ITS2 phylogeny, respectively (Fig.
2 and
3). Although the topologies differed somewhat in support values, the 5.8 rDNA-ITS2 topology showed higher bootstrap support values overall than the 26S topology, suggesting that this region provides a better resolution of the
Malassezia spp. Taking both topologies into account, it remains unclear which is the sister group of
M. furfur: the 26S topology places this group next to
M. yamatoensis, while the 5.8S rDNA-ITS2 topology places it next to
M. obtusa.
DISCUSSION
The
Malassezia species are difficult microorganisms to identify and to maintain in culture (
9). To overcome these difficulties, several physiological and molecular techniques have been implemented recently that allow the species to be differentiated and new ones to be described (
2,
15,
17,
18,
20-
24,
27,
38,
44). In this work, we thoroughly characterized seven isolates with an atypical Tween 80 assimilation pattern, and our results support the view that these isolates correspond to
M. furfur.
Because these species are lipophilic, the Tween assimilation pattern allows us to identify different
Malassezia species (
18,
22). This test is based on the ability of
Malassezia spp. to use different fatty acids. Some authors attribute this characteristic to lipase activity and propose that it is related to the adaptation of
Malassezia spp. to host body regions rich in fatty acids under given conditions (
3,
12,
37,
47). Brunke et al. described the lipophilic activity of MfLip in
M. furfur; this extracellular lipase apparently is involved in cellular growth processes and related to pathogenicity mechanisms (
3).
On the other hand,
M. furfur has been described as showing high phenotypic and genotypic variability. Indeed, colonies of this species vary in size and form and show a high degree of cellular pleomorphism, including oval, cylindrical, and spherical cells (
18,
22). Techniques such as randomly amplified polymorphic DNA and amplified fragment length polymorphism have demonstrated high intraspecific variability in this species (
2,
5,
44). The atypical assimilation pattern of our isolates, involving an inability to utilize Tween 20, 40, or 60, reveals a metabolic variation probably resulting from alternate gene expression.
Such variability has been observed in related organisms such as
Candida spp. Lan et al. found a relationship between phenotypic variation and metabolic flexibility in the genus
Candida, which increases its selection according to the availability of nutrients in certain body regions (
28). Phenotypic and genotypic variability such as the atypical Tween assimilation pattern in the case of
M. furfur may, like the variability in
Candida spp., be a determining factor in the infection strategy used by the microorganism, which in turn could be related to the expression of genes involved in pathogenic processes and colonization (
39).
Shifts in metabolic processes could be related to a change of condition from commensal to pathogenic. The resulting high variability also may be influenced by host responses such as the generation of adverse microenvironments, which perturb the host-pathogen balance (
7,
9,
16,
29,
33). Nonetheless, the variation in the Tween assimilation pattern observed in this research (based on the seven atypical strains) was not related to a particular dermatological condition, since it was present both in patients with a dermatological condition and those without apparent injury.
The expected results for 26S (
32) and 5.8S rDNA-ITS2 (
15) were obtained with all 15 isolates, including the reference ones. The PCR products and the products of digestion revealed no differences between the isolates with the atypical Tween assimilation pattern and the reference strains
M. furfur CBS 1878 and CBS 7019. Our sequence analyses based on homology searches of the ribosomal regions identified our isolates as
M. furfur. Furthermore, the phylogenetic analyses for 26S and 5.8S rDNA-ITS2 grouped our isolates and reference strains together in a single cluster with high bootstrap values (Fig.
2 and
3). The phylogenies obtained in this study support our findings and the identity of the isolates as
M. furfur; the atypical assimilation of Tween 80 was found to be a new physiological pattern characteristic of certain strains isolated in Colombia.
In conclusion, we suggest further studies that will explain the genetic background of the atypical Tween assimilation pattern and its relationship to the establishment of the yeast in certain host body regions. The evaluation of the metabolic pathways and the gene products involved in the fatty acid degradation, and even the relationship between gene expression and diverse host conditions, would produce a better understanding of this unique Tween assimilation pattern, which constitutes a new parameter for the identification of M. furfur.
Acknowledgments
We express our acknowledgment to the Science Faculty of the Universidad de los Andes for the financial support of this project.