Differentiation of Candida dubliniensis from Candida albicans on Pal's Agar

Candida dubliniensis is a novel Candida species first described by our laboratory in 1995 (20). Studies to date indicate that this species is primarily associated with oral carriage and infection in human immunodeficiency virus-infected and AIDS patients and has rarely been identified in blood cultures from patients with candidemia (3, 4, 6, 7, 10, 14, 19, 20). Routine discrimination between C. dubliniensis and the closely related species Candida albicans has been problematic (4, 13, 18). The most accurate means of identifying C. dubliniensis and discriminating it from C. albicans requires PCR-based tests; however, these are not readily applicable to the high-volume throughput of isolates in many routine diagnostic laboratories (5, 8). In contrast, while a number of phenotype-based tests (e.g., determination of colony color on CHROMagar Candida plates and lack of growth at 45°C and carbohydrate assimilation profiling) have proved to be useful for identifying C. dubliniensis isolates, they do not give completely reliable results (4, 12, 13, 15, 18). In a recent study, Staib and Morschhäuser suggested that Staib agar (which contains Guizotia abyssinica seed extract) was a good discriminatory medium (17). This finding was confirmed in a study from our laboratory in which 127 of 130 isolates of C. dubliniensis (97.7%) grew as rough colonies and all 166 C. albicans isolates tested grew as smooth colonies (1). Staib agar was originally developed for the identification of Cryptococcus neoformans, which produces melanin-like pigment on this agar (16). In order to develop a medium with even greater discriminatory ability, we investigated the effect of replacing G. abyssinica seed extract with extracts from the seeds of other plants. One such medium is Pal's agar, which contains sunflower (Helianthus annus) seed extract and is another medium originally developed for the identification of C. neoformans (16). Since both H. annus and G. abyssinica belong to the same botanical family (Asteraceae), and since Staib agar is a good but not an absolutely reliable medium for differentiating between C. dubliniensis and C. albicans, we investigated the use of Pal's medium for this purpose.

The yeast isolates used in this study are shown in Table 1. The identity of all isolates was reconfirmed using the ID 32C yeast identification system (bioMérieux, Marcy l'Étoile, France). In addition, the identities of C. dubliniensis and C. albicans isolates were confirmed by growth at 45°C and by PCR (5, 13). All isolates were also tested for chlamydospore formation on rice agar-Tween agar, and all the C. dubliniensis and C. albicans isolates produced chlamydospores on this medium (20). Pal's agar was prepared freshly with unsalted sunflower seeds (including kernels and shells) and used within 5 days to ensure consistent results. First, an aqueous extract of sunflower seeds was prepared by pulverizing 50 g of seeds in a domestic Moulinex (Dublin, Ireland) model B57 blender for 5 min and then adding the ground seeds to 1 liter of distilled water, followed by boiling for 30 min. Next, the seed extract was cooled and filtered, and the following ingredients were added: glucose (1 g), KH₂PO₄ (1 g), and creatinine (1 g). The pH was adjusted to 5.5, the volume was readjusted to 1 liter, and 15 g of agar (Difco) was added before the mixture was autoclaved at 110°C for 20 min. For each isolate included in the study, part of a single colony grown on potato dextrose agar (Oxoid) at 37°C for 48 h was streaked onto Pal's agar contained in 90-mm-diameter single-vent petri dishes (25 ml of agar per plate) and incubated at 30°C. Colony morphology was examined visually every 24 h for up to 10 days, and the data were recorded. At 48 to 72 h, colonies were examined microscopically for the presence or absence of chlamydospores: 10 to 20 well-separated single colonies were stained with 1 drop of 1% (wt/vol) lactophenol cotton blue stain (which stains chlamydospores preferentially), allowed to stain for 5 min, and then covered with sterile glass coverslips (22 by 22 mm) and examined microscopically under bright-field illumination with a 40× objective (9, 20). Chlamydospore-negative isolates continued to be examined at 24-h intervals for up to 10 days.

All of the yeast isolates tested grew well on Pal's agar. Following 48 to 72 h of incubation, all 124 C. albicans and all 128 C. dubliniensis isolates tested grew as smooth creamy-gray colonies (Fig. 1). However, all the C. dubliniensis isolates exhibited a hyphal fringe (Fig. 1B), whereas none of the C. albicans did (Fig. 1A). It should be noted that 36 of 124 C. albicans isolates (29%) were observed to produce a fringe following 10 days of incubation. Microbiological analysis revealed that the fringe surrounding C. dubliniensis colonies was found to be comprised of hyphae, pseudohyphae, and blastospores. These findings are in contrast to previous data obtained by using Staib agar, on which 127 of 130 C. dubliniensis isolates tested (97.7%) formed rough colonies, many (65%) with a hyphal fringe, and all 166 C. albicans tested grew as smooth colonies (1). In the present study, none of the C. albicans isolates produced chlamydospores on Pal's agar, even after 10 days of incubation, whereas 120 of 128 of the C. dubliniensis isolates tested (93.75%) produced chlamydospores within 48 to 72 h. The remaining 8 of 128 C. dubliniensis isolates (6.25%) were chlamydospore negative even after 10 days of incubation. Interestingly, in the previous study with Staib agar, none of the 166 C. albicans isolates tested produced chlamydospores whereas only 19 of 130 of the C. dubliniensis isolates examined (14.6%) were chlamydospore negative (1). Twelve of these 19 chlamydospore-negative C. dubliniensis isolates were included in the present study; 7 were chlamydospore positive on Pal's agar, and the remaining 5 were chlamydospore negative.

The colony morphologies of selected oral isolates of four other Candida species (Table 1), including five C. glabrata, four C. tropicalis, five C. parapsilosis, and a single C. krusei isolate, were also examined on Pal's medium. One C. tropicalis isolate, two C. glabrata isolates, and three C. parapsilosis isolates yielded smooth colonies similar to those of C. albicans. Three C. tropicalis isolates, two C. parapsilosis isolates, and the single C. krusei isolate formed rough colonies with a fringe similar to that formed by C. dubliniensis.

Similar results to those described above were obtained with all isolates tested on separate batches of Pal's agar prepared from the same batch of sunflower seeds, in each case with seeds purchased from two separate suppliers. However, we found that incubation of the plates at 37°C rather than 30°C resulted in poorer discrimination between the two species. We therefore propose that the formation of a hyphal fringe surrounding C. dubliniensis colonies on Pal's medium following incubation for 48 to 72 h at 30°C provides a definitive means of discrimination between this species and C. albicans. However, since colonies of some isolates belonging to other Candida species also produce a fringe on Pal's agar, we suggest that this medium should be used only to screen germ tube- and/or chlamydospore-positive isolates. Pal's medium has an important advantage over Staib agar in that it permits an absolute discrimination between C. dubliniensis and C. albicans.

• Open in new tab
• Download powerpoint

FIG. 1.

Macroscopic appearance of C. dubliniensis and C. albicans colonies on Pal's medium following 72 h of incubation at 30°C. (A) Smooth colonies exhibited by C. albicans composed exclusively of blastospores; (B) C. dubliniensis colonies displaying a hyphal fringe, containing abundant hyphae, pseudohyphae, and chlamydospores.

• American Society for Microbiology