First Isolation of Reddish-Pigmented Candida (Torulopsis) glabrata from a Clinical Specimen

Reddish-pigmented yeast isolates from human specimens generally belong to the genus Rhodotorula and are usually regarded as contaminants. Recently, we isolated a reddish-pigmented yeast from a surveillance culture of stool of an immunocompromised patient (culture collection no. GHP 1911, Institute of Medical Microbiology, Aachen, Germany). The isolate was the predominant stool species and was cultured as pure yeast culture from two independent stool specimens of this patient. Surprisingly, the pigmented isolate could be unequivocally assigned to the taxon Candida (Torulopsis) glabrata.

Identification of this yeast isolate was confirmed by several means, including (i) biochemical profiling (ID 32 C); (ii) gas-liquid chromatographic (GLC) analysis of whole-cell fatty acid methyl ester (FAME) (5); and (iii) sequence analysis of a 500-bp fragment of the 5′ end of the nuclear large ribosomal subunit rRNA gene (4).

In addition, micromorphology and susceptibility testing were also performed with GHP 1911. Small spherical yeast cells were observed in the germ tube test, which is in keeping with identification of C. glabrata for which there is no tendency towards either pseudohyphal or hyphal growth. Rhodoturula species isolates generally present more ovoidal or elongated cells with a tendency to form pseudohypha. However, overall, the micromorphologic traits of Rhodoturula spp. are not characteristically different from those of C. glabrata. Antimicrobial susceptibility of GHP 1911, as determined by the Etest (AB-Biodisc, Solna, Sweden), revealed a fluconazole MIC of 0.064 mg/liter, an itraconazole MIC of 0.003 mg/liter, and an amphotericin B MIC of 0.25 mg/liter, signifying susceptibility.

When using the ID 32 C, GHP 1911 only assimilated trehalose and glucose, giving it a biochemical profile number 0001000001. This biochemical profile corresponds to a 99.1% probability of correct identification as C. glabrata in comparison with all the other taxa available in the database, an extremely high concordancy. Thus, our isolate was allocated to the taxon C. glabrata. In stark contrast, the second choice, Kloeckera japonica, was selected with only 0.4% probability of correct identification. Furthermore, when the database was used, the biochemical profile of our isolate was estimated as close as possible to the typical set of reactions for this taxon (T value = 1.0). In contrast, some pigmented Rhodotorula species would typically assimilate saccharides besides trehalose and glucose with varying degrees of probability; this would hence translate into another biochemical profile, e.g., Rhodotorula mucilaginosa (synonym Rhodotorula rubra; biochemical profile number 5421650101) or Rhodotorula glutinis (biochemical profile number 4461650101). FAME analysis of GHP 1911 with the Microbial Identification System (MIS; Sherlock, MIDI, Inc., Newark, Del.) revealed a similarity index (SI) of 0.397 when using the environmental database (YST28). This numerical value indicates a hierarchy of possible strain fits which express how closely the fatty acid composition of the measured sample compares with the mean fatty acid composition of the strains used to create the library entry listed as its match. An exact match of the fatty acid composition with the mean of the library entry would result in an SI of 1.0. As a second choice, and with a slightly lower SI of 0.393, Saccharomyces cerevisiae was listed. The problematic issue of separating these two species by FAME analysis has been discussed previously (5). The whole-cell main fatty acid composition of GHP 1911 did not differ from that of typical C. glabrata strains, comprising around 50% of palmitoleic acid and around 30% of oleic acid. FAME analysis of typical strains of R. mucilaginosa would reveal a higher percentage of around 50% oleic acid but no palmitoleic acid. Furthermore, in contrast to what is observed for C. glabrata, around 30% linoleic acid would be detected. Thus, it can be expected that all the other pigmented yeasts included in the respective database would, by whole-cell FAME analysis, clearly differ from GHP 1911. The above-mentioned sequence analysis revealed 100% identity with the respective sequence of a C. glabrata isolate CBS 138^T (U44808) (Fig. 1). Our isolate GHP 1911 was submitted to the Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands, and is accessible under the strain number CBS 8947.

FIG. 1.

Rooted (outgroup Absidia, belonging to the division Zygomycota) phylogenetic tree inferred by analysis of the D1/2 domain of the LSU ribosomal DNA using a neighbour-joining algorithm. Bootstrap values of >800 are displayed at respective nodes.

Phylogenetic analysis with bootstrap analysis (1,000 trees) using a neighbor-joining algorithm was performed with the support of the programs DCSE (vers. 2.60) and TREECON (vers. 1.3b). By these means, information can be provided about the relationship of GHP 1911 with other similar pigmented yeasts, i.e., Sporobolomyces roseus, R. glutinis, and R. mucilaginosa (Fig. 1). Ascomycetal yeasts, including GHP 1911, cluster with substantial phylogenetic distance as a monophyletic group to the pigmented yeasts with basidiomycetal affilation (Fig. 1).

To determine the color more precisely, spectrograms of the water-soluble pigment (after 48 h of cultivation at 37°C) of GHP 1911 and a respective solution from a nonpigmented C. glabrata isolate (CBS 138^T) were compared as shown in Fig. 2. Furthermore, according to the software program used (WhatColor v4.02e) (http://www.hikarun.com/e/) for classifying the color of the colonies of GHP 1911, the name sandy orange with the respective hexadecimal RGB triplet value of 249,183,85 was attributed (Fig. 3).

FIG. 2.

Spectrophotometric analysis of Candida glabrata isolates. Red line, water soluble eluate of non-pigmented isolate; blue line, water soluble eluate of sandy orange-pigmented C. glabrata isolate (GHP 1911).

FIG. 3.

Growth of the reddish-pigmented isolate GHP 1911 on Sabouraud agar after 48 h of incubation at 37°C. The color name sandy orange, corresponding to the hexadecimal RGB triplet value 249,183,85, was attributed. Around an adenine-soaked filter disk (arrow), no feedback inhibition of the synthesis of the precursor substance of the adenine pathway that turns red during oxidation occurs. This indicates that GHP 1911 has a potentially different pathway for pigment synthesis from adenine auxotrophy.

Besides yeasts of the genus Rhodotorula, red mutants, which were caused by UV irradiation, have been described for several (white) yeast species, e.g., S. cerevisiae, Candida krusei, and Candida albicans. Their reddish color is due to a blockage in the adenine biosynthetic pathway, resulting in the accumulation of a precursor compound that turns red during oxidation (2). Experimentally, this accumulation can be prevented by growing the isolate on a medium with a high concentration of adenine; this leads to a feedback inhibition of the synthesis of the precursor substance. Clinical isolates have been reported extremely rarely; e.g., a red-pigmented adenine auxotrophic isolate of C. albicans was reported in the case of a patient with cystic fibrosis (3). In contrast to these strains, pigment formation of the isolated sandy orange GHP 1911 could not be prevented by adenine supplementation, which indicates a potentially different pathway for pigment synthesis (Fig. 3).

C. glabrata occurs as a saprophyte on the human body. It is now emerging as an opportunistic pathogen and has been reported to be the yeast species isolated second most frequently from clinical infections (1). Due to the commonly occurring innate or acquired resistance of C. glabrata to fluconazole, the widespread use of this azole might result in selection of this yeast species (5). In addition, C. glabrata has even been reported to be responsible for 75% of fungemias in patients receiving fluconazole (6). Since fungemia is associated with high mortality rates, as the majority of the patients have a rapidly progressive or ultimately fatal disease, immediate implementation of an appropriate antimycotic therapy is mandatory, and this requires reliable identification of C. glabrata. However, antimicrobial susceptibility testing of GHP 1911 did not reveal resistance to the commonly used antimycotics (the azoles and amphotericin B).

In conclusion, one should be aware the reddish-pigmented yeasts do not exclusively belong to Rhodotorula spp. and efforts should be made to achieve detailed and accurate characterization.

ADDENDUM IN PROOF

Since we became aware of the recently described phenomenon of phenotypic switching in Candida glabrata (S. A. Lachke, T. Srikantha, L. K. Tsai, K. Daniels, and D. R. Soll, Infect. Immun. 68:884-895, 2000), we have performed additional experiments using an indicator agar containing 1 mM CuSO₄. Culturing of our strain (CBS 8947) on this indicator agar as recommended did not result in switching of the color phenotype.