INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Candida lusitaniae is now well documented as an emerging human opportunistic pathogen (8, 22). Albeit much less frequent than other Candida species, C. lusitaniae has been described as a nosocomial agent (4, 26, 30) and can cause severe fungemia in immunocompromised hosts. In some cases, therapeutic failure has been correlated with the propensity of this yeast to acquire antifungal resistance, mainly to amphotericin B (6, 18, 23, 36) but also to azoles (1) and to flucytosine (26). Most of the resistance events can easily be explained by mutations, whose expression is favored by the haploid genetic status of this species. Surprisingly, it was recently reported that a high rate of amphotericin B resistance could also result from an unusual mechanism of switching from susceptible to resistant phenotypes (37).

In such a context, careful attention should be paid to an accurate identification of C. lusitaniae infection to avoid the danger of selecting antifungal resistance during treatment. Unfortunately, routine identification of C. lusitaniae by conventional means (colony morphology and metabolic characteristics) may prove lengthy (up to 72 h) and difficult (21). A series of misidentifications have led in the past to confusion with C. parapsilosis (9, 23, 29), C. tropicalis (6, 19), and even Saccharomyces cerevisiae (7). Identification relying upon carbon assimilation (rhamnose) and fermentation (cellobiose) may sometimes be insufficient for clear distinction from atypical C. tropicalis (5). Advances in molecular biology have provided a set of DNA-based tools for the characterization of Candida species, including C. lusitaniae. However, few of these organisms have been identified to the species level (17, 31), and the great majority of the tools have been developed for epidemiological purposes (11, 20, 26, 33, 35).

On the other hand, sexual reproduction within the genus Candida is a diagnostic element that has been underestimated thus far. Indeed, among the 21 Candida species described as human pathogens, 13 have the potential to reproduce sexually (8). Interestingly, the four species most frequently isolated from cases of fungemia, i.e., C. albicans, C. tropicalis, C. parapsilosis, and C. glabrata, have no known sexual cycle, even though recent experiments with genetically modified C. albicans strains have demonstrated that in vitro mating and in vivo mating are at least possible (10, 16). Taking advantage of this fact, we investigated whether the sexual cycle of C. lusitaniae could be used to provide additional information for its reliable identification. The teleomorph of C. lusitaniae (Clavispora lusitaniae) is a heterothallic ascomycete yeast (28). The sexual cycle is controlled by the biallelic locus MATa/MAT. Mating is possible only between two haploid cells of opposite mating types. Meiosis gives rise to clavate, echinulate ascospores that are liberated by ascus disruption (5, 12, 28). Like that of many other yeasts and fungi, the sexual cycle is triggered and completed during nutrient starvation.

In this study, we first evaluated the culture conditions supporting the in vitro sexual cycle of C. lusitaniae. The meiosis efficiency was then measured at different incubation temperatures using two different combinations of sexually compatible auxotrophic mutants. The major cytological events leading to ascospore formation were characterized by scanning and transmission electron microscopy. Labeling cells with concanavalin A (ConA)-fluorescein isothiocyanate (FITC) before conjugation of different sexually compatible strains enabled us to distinguish the nucleus donor cell from the nucleus acceptor cell, the latter being the cell which undergoes meiosis and turns into the ascus. Finally, we studied the sexual reproduction ability and determined the mating types of a collection of 76 isolates recovered from 60 humans in a hospital context. We demonstrate that sexual reproduction can be used as a criterion for C. lusitaniae identification, and we question the existence of anamorphs in this species.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

C. lusitaniae reference strains and isolates. Mating type determination was performed with a total of 78 strains and isolates. The collection included 2 reference strains, 6936 MATa and 5094 MAT (Centraalbureau voor Schimmelcultures, Baarn, The Netherlands), and 76 isolates recovered from patients hospitalized in intensive care units. Among them, 4 isolates were derived from the University of TexasHouston Medical Center (15) and were recovered from 4 patients; 72 isolates were derived from five French medical centers, some of them already having been described (3, 21, 25), and were recovered from 56 patients (multiple isolates were obtained from some patients: 2 isolates from 8 patients, 3 isolates from 2 patients, and 5 isolates from 1 patient ). Clinical isolates were recovered from the upper respiratory tract (n = 34), mouth (n = 7), sputum (n = 4), stools (n = 9), urine (n = 9), bedsore tissue (n = 1), vagina (n = 2), blood (n = 9), and intravenous catheter (n = 1). Identification was performed by standard methods (34) and with the API 32C system (bioMérieux, Marcy l'Etoile, France).

Auxotrophic mutants. Auxotrophic strains 14/31 MAT Leu, 5/31 MATa Leu, 69/2 MAT Lys, and 69/3 MATa Lys were derived from ethyl methanesulfonate mutagenesis of reference strain 6936, followed by nystatin enrichment according to standard genetic methods (2). The genome of initial mutants was purified once by a genetic cross with strain 5094 MAT, and MAT auxotrophic progeny were then purified by three successive backcrosses with strain 6936 MATa. The method for isolating ascospores derived from the crosses will be described elsewhere (unpublished data). Genetic analyses showed that mutations in the leucine and lysine pathways affected a single gene, were recessive, were genetically unlinked, and were unlinked to mating type genes (data not shown).

Other yeast species. Various yeast species were assayed in negative control mating tests with C. lusitaniae: C. albicans (ATCC 2091), C. tropicalis (CBS 94), C. parapsilosis (ATCC 22019), C. famata (Debaryomyces hansenii; CBS 1795), C. (Pichia) guillermondii (CBS 6021), C. (Metschnikowia) pulcherrima (IP 82363 [Institute Pasteur]), and S. cerevisiae FL100 MATa (ATCC 28383) and FL200 MAT (ATCC 32119).

Media and culture conditions. C. lusitaniae was routinely cultivated on YPD medium (1% yeast extract, 2% peptone, 2% dextrose) at 35°C. The solid media yeast nitrogen base with amino acids and ammonium sulfate (YNB [concentration, 0.67%]; Difco Laboratories, Detroit, Mich.) and yeast carbon base (YCB [concentration, 1.17%]; Difco), supplemented or not with various nitrogen sources, were tested for their ability to induce and support the sexual cycle. YNB was also used for selecting the recombinant prototrophic meiotic products released from crosses between auxotrophic parents.

Mating type tests and crosses. For the determination of the mating type of an unknown isolate, cells of the isolate and of the reference mating type tester strains 6936 MATa and 5094 MAT (or Cl38 MAT) were separately cultivated to stationary phase in 2 ml of liquid YPD medium under agitation (250 rpm) at 35°C for 16 h (ca. 2 × 10⁸ cells/ml). Equal volumes of the cell suspension (typically 500 µl) from the isolate were distributed in two 1.5-ml microtubes. To one tube was added 500 µl of the MATa tester strain culture, and to the other was added 500 µl of the MAT tester strain culture. The cell mixtures were centrifuged (3 min, 1,000 × g 20°C), and the cell pellet was resuspended with 500 µl of distilled sterile water to obtain a cellular concentration of about 4 × 10⁸ to 5 × 10⁸ cells/ml. Aliquots of 5 µl from each mixture were spotted on solid YCB medium. After 24 to 72 h of incubation at room temperature, cells were removed from the spots with a toothpick and examined in a drop of water with a light microscope (magnification, ×400), ideally in phase contrast. The presence of asci and ascospores indicated that the two strains tested had opposite mating types; the lack of asci and ascospores indicated that the two strains tested had the same mating type. Genetic crosses were performed under the same conditions by mixing cells with opposite mating types. The mating type protocol may be modified in order to save time. Cell suspensions can be made by dispersing fresh colonies isolated from complete solid medium in sterile water.

Scanning and transmission electron microscopy. Yeast cells were mixed in a 1.5-ml microtube with 2% low-melting-point agarose (maintained at 30°C) and centrifuged (3,000 × g, 5 min). After the sample was cooled to 15°C, the solidified pellet was dissected into 1- to 2-mm³ pieces, which were fixed in a mixture of 2% glutaraldehyde and 0.5% paraformaldehyde in pH 7 0.1 M McIlvaine citrate-phosphate buffer and postfixed in 0.5% OsO₄ (24). After successive dehydration in ethanol and propylene oxide, the samples were embedded in Spurr's (32) low-viscosity resin. Ultrathin (70-nm thick) sections were collected on Formvar-coated copper grids, stained with uranyl acetate and lead citrate (27), and observed at 80 kV in a Philips CM 120 electron microscope.

ConA-FITC labeling of cells. ConA-FITC (Sigma Chemical Co., St. Louis, Mo.) labeling was performed with 10 mM MES (2[N-morpholino]ethanesulfonic acid; Sigma) buffer (pH 5.4) by mixing 50 µl of a 10-mg/ml ConA-FITC solution with 500 µl of a MES-washed overnight yeast culture. After 1 h of incubation at room temperature and three washes with 1 ml of MES buffer, labeled cells were mixed with unlabeled cells of an opposite mating type and the mixture was spotted on solid YCB medium. After 96 h of incubation, the mating reaction was observed with a Zeiss fluorescence photomicroscope.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Triggering and completion of the sexual cycle in vitro. Using a mixture of cells from the 6936 MATa and 5094 MAT reference strains, we first tested the solid media previously described as supporting the sexual cycle of C. lusitaniae, i.e., Gorodkowa, sodium acetate, diluted potato dextrose agar (5), and 1% malt agar (28). None of these media was as effective in our experiments as YCB medium (12) in term of swiftness (conjugation events observed within 16 to 48 h) and quantity of ascospores released (within 24 to 72 h). In order to determine the nutritional deprivation signal that triggers the sexual cycle, we further compared the efficiencies of YCB and YNB media. The media are comparable, since they have the same composition, except for the main carbon and nitrogen sources: YCB contains 1% glucose and no ammonium sulfate, whereas YNB contains 0.5% ammonium sulfate and no glucose. Different compatible combinations of strains 6936 and 5094 and of isolates Cl38 and Cl69, together with pure cultures of each, were assayed on YNB and YCB, the latter being supplemented or not with various nitrogen sources. The production of asci and ascospores was monitored over a period of 10 days. The results obtained with the different strain combinations were comparable; only those for 6936 and 5094 are presented (Table 1). As expected, ascospores could be detected only in the 6936-5094 cellular mixture under given conditions and not from the pure cultures, a result which confirmed the heterothallic behavior of these strains. Ascospores were produced on YCB but not on YNB, a result which suggests that mating is triggered in response to nitrogen but not carbon starvation. Supplementing YCB with potassium nitrate or urea had no inhibitory effect on the mating reaction relative to the results obtained with unsupplemented YCB. On the other hand, increasing the ammonium sulfate concentration from 0.01% to either 0.1 or 0.25% resulted in mating inhibition. This result suggests that depletion of ammonium ions could be a major signal for the triggering of the sexual cycle. Nevertheless, other conditions are required for the induction of the sexual cycle. For instance, mating did not occur in liquid YCB. A solid support seems to be required for mating, either because cells of opposite mating types need a prolonged period of contact for recognition or because the morphogenetic events leading to conjugation can take place only with immobilized cells.

Monitoring the meiotic yield on the basis of the temperature and time of incubation. During the study of nutritional factor requirements, we observed variations in mating efficiencies and ascospore yields depending on the strain combinations used. For example, mating between 6936 and 5094 was slower and yielded consistently fewer ascospores than mating between 6936 and Cl38. In order to better quantify the variations in mating efficiencies, we monitored the ascospore yields from two genetic crosses involving auxotrophic strains on the basis of the temperature and incubation time. Each parent of the cross harbored a single-gene recessive mutation affecting either the leucine or the lysine biosynthetic pathway. A preliminary mating experiment allowed us to select strains with high and low mating abilities.

Major cytological events of the sexual cycle. The cytological events leading to ascospore formation were characterized by scanning and transmission electron microscopy. A cross was made on YCB between reference strain 6936 MATa and clinical isolate Cl38 MAT, which were selected because they produced high levels of ascospores when mated together. Samples of cell mixtures were removed from solid YCB medium after 24, 48, and 72 h of incubation and resuspended in sterile distilled water before being processed for electron microscopy. A clear discrimination could be made between budding yeast cells (Fig. 1A) and conjugating cells (Fig. 1B), owing to the length of the bulging conjugation canal showing a slight constriction at midlength. The ultrastructure of the conjugation canal was better characterized by transmission electron microscopy (Fig. 1C). The cell walls of both partners were tightly merged and formed a perforated septum at the cell junction, through which nuclear transfer from one cell to the other could occur. Conjugated forms were predominantly observed after 24 h of incubation on YCB. Ascospores released by ascus disruption (deliquescence) were detected after 48 h of incubation (Fig. 1D). After 72 h of incubation, numerous disrupted empty asci (Fig. 1E) and free ascospores (Fig. 1F) were easily distinguishable among all other cellular forms. The photographs show that the ascospores are clavate and echinulate. A material that could be traces of ascus cytoplasm often made them stick together. Note that conjugated forms and disrupted asci are easily observable using a simple routine light microscope with a ×40 objective, preferably with phase contrast.

View larger version (75K): [in this window] [in a new window]   FIG. 1.   Major cytological events of the sexual cycle. (A) Budding yeast cell in a pure culture. (B) Conjugation between MATa and MAT cells showing the conjugation canal with its constriction at midlength (arrow). (C) Detail of the conjugation canal showing the merged cell walls and the perforated septum (arrow). (D) Release of ascospores by ascus disruption. (E) Empty ascus (right part) and its head cell (left part). (F) Tetrad of clavate, echinulate ascospores. Scale bars, 1 µm. (A, B, D, E, and F) Scanning electron microscopy. (C) Transmission electron microscopy. Magnifications were about ×11,640 (A, B, and D), ×14,550 (E), ×19,400 (F), and ×37,830 (C).

Study of nuclear transfer. Meiosis in C. lusitaniae leads to a clear distinction between the nucleus acceptor cell, which turns into the ascus, and the nucleus donor cell (the "head" cell), which remains unchanged. Such cellular dimorphism led us to investigate whether the nuclear transfer occurred randomly between the partners of a conjugated form or whether it was directional, i.e., did cells of one strain always act as nucleus donors and did cells of the opposite mating type always act as nucleus acceptors? In order to answer that question, cells of one strain were labeled with ConA-FITC before being mated with unlabeled cells of the opposite mating type. Three different compatible strain combinations were used, each strain being alternatively labeled, and fluorescence was detected after mating (Table 3).

Determination of the mating types of clinical isolates. In order to determine whether clinical isolates of C. lusitaniae had retained their ability to undergo meiosis, we studied the sexual cycle and attempted to determine the mating types for a collection of 76 strains isolated from 60 patients. This determination was done in two steps. In the first set of experiments, 41 strains, including 2 reference strains and 39 clinical isolates, were paired in all possible combinations {resulting in [n × (n  1]/2 = 820 mating combinations, with n = 41}. This was done in order to verify that there was no abnormality in the mating reaction, i.e., to ascertain that each isolate was able to mate with all other isolates of the opposite mating type and not with all other isolates of the same mating type. Nevertheless, this study could not be realized with the whole sample because it would have resulted in too many mating reactions (3,003 exactly). Accordingly, in the second set of experiments, the mating types of the remaining 37 clinical isolates were determined by pairing with reference strain 6936 MATa and clinical isolate Cl38 MAT.

View this table: [in this window] [in a new window]   TABLE 4.   Assignment of mating types and MATa/MAT ratios in a collection of 76 C. lusitaniae clinical isolates

Specificity of the mating test. Since all 76 tested clinical isolates showed sexual reproduction, our results support the idea that the mating test could be used as a complementary identification tool for C. lusitaniae. With that idea in mind, we ascertained that other yeast species (see Materials and Methods) which are sometimes mistaken for C. lusitaniae did not react with C. lusitaniae during a mating test. Each strain was assayed in pure cultures on YCB and subjected to mating experiments with both 6936 and Cl38. No positive mating reaction comparable to those shown in Fig. 2 could be detected, even after 10 days of incubation.

View larger version (81K): [in this window] [in a new window]   FIG. 2.   Examples of positive and negative mating reactions with C. lusitaniae, as observed by phase-contrast microscopy, and proposal for interpreting all the results that could arise from mating an unknown isolate with mating type tester strains. (A, D, E, and F) Representative set of different cellular structures observable from one positive mating reaction after 24 h of incubation. (B, C, G, and H) Negative mating reactions after 96 h of incubation. a, ascospores; as, ascus; c, conjugated cells; da, disrupted ascus. Magnification, ×300.

Carrying out and interpreting a mating test in a clinical laboratory. Carrying out a mating test means determining the mating ability and the sexual type of an unknown isolate. For a routine mating test with C. lusitaniae, it is of primary interest to select MATa and MAT tester strains having a high mating potential. This step is particularly important for rapid and easy interpretation. The test consists of mixing cells of the isolate and of both tester strains separately (two mating reactions) and spotting cell mixtures on solid YCB. Negative control reactions can be carried out by spotting a pure suspension of each strain, and positive control reactions can be carried out by mating the tester strains together. Incubation at room temperature is convenient, and it is generally not necessary to prolong incubation beyond 72 h. In our experiments, the majority of the tests could be read with a phase-contrast microscope at 24 h. Some examples of positive and negative mating reactions, with a guideline for interpretation, are presented in Fig. 2. An unknown isolate can be unambiguously identified as a C. lusitaniae strain when mating occurs with only one of the mating type tester strains. In any other situation, the test cannot be interpreted. If no mating occurs with both tester strains, one cannot conclude that the isolate belongs to another yeast species, since we cannot completely preclude the existence of mating-defective isolates of C. lusitaniae. Nevertheless, if it does exist, this phenomenon should be rare, because we never observed it within our collection of 76 clinical isolates.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Mating in C. lusitaniae is readily obtained in vitro on solid medium in response to nitrogen starvation. We showed that ammonium ions are the major nitrogen source whose depletion triggers mating, a finding which is consistent with those of previous studies (12, 38). Other nitrogenous compounds (such as nitrates and urea) had no effect on mating, consistent with the fact that C. lusitaniae cannot assimilate these sources. A key role of ammonium in some important morphogenetic pathways is not surprising. In S. cerevisiae, under nitrogen starvation conditions, an ammonium permease is involved in signaling for triggering the pseudofilamentation pathway (14), which shares a common mitogen-activated protein kinase cascade with the mating pheromone response pathway (13). It is possible that C. lusitaniae uses analogous signaling cascades for both mating and pseudofilamentation. First, we observed that the nitrogen-lacking YCB medium used for mating allowed abundant differentiation of pseudohyphae, suggesting that both pathways respond to nitrogen starvation, although pseudofilamentation also responds to carbon starvation, while mating does not. Second, following ethyl methanesulfonate mutagenesis, we selected several pseudofilamentation-defective mutants, about half of which were concomitantly defective in mating.

Depletion of ammonium is necessary but not sufficient for completion of the sexual cycle, because we failed to observe mating in liquid YCB medium. A previous work reported the apparent lack of diffusible mating factors and of sexual agglutination in both C. opuntiae and C. lusitaniae and concluded that cell-to-cell contact was required for a mating response (12). The fact that in our experiments, mating and meiosis could occur only on a solid support is consistent with that idea. Furthermore, the electron microscopy analysis revealed that complete cell fusion during mating did not occur and that cellular exchange, notably, nuclear migration, was mediated through a conjugation canal about 1 µm long. The ultrastructure of the canal is rather complex: each of both partner cells differentiates a part of the canal, and at the junction, the cell walls appear tightly merged and form a septum which is perforated for allowing communication between cells. In a turbulent liquid environment, it is possible that weak adhesion strength between cells, combined with the relatively long period of time necessary to complete canal differentiation, constitutes an obstacle to conjugation attempts. Alternatively, it is reasonable to consider that cell-to-cell contact may require additional adhesion factors whose expression would be induced in response to a solid surface.

After conjugation, one cell turns into the ascus, where karyogamy and meiosis take place, and the other cell remains unchanged. Such cellular dimorphism during sexual reproduction led us to wonder whether nuclear transfer between two strains was polarized. Our results show that nuclear transfer is predominantly polarized. Cells of one strain act as nucleus donors, and cells of the sexually compatible strain act as nucleus acceptors. Using different isolates, we demonstrated that polarity depends on the strain combination and does not seem to be directly controlled by mating type genes. The cellular message involved in the determination of the nuclear transfer direction is not known, but we suspect that the lectin ConA may disturb this message. Indeed, the contradictory results obtained by labeling alternatively each strain in one cross (6936 × Cl38) could be explained by an interfering cellular activity of ConA.

The main goal of this study was to determine whether sexual reproduction could be used as a complementary identification system for C. lusitaniae. Of the 76 isolates tested, 100% were able to mate and to undergo meiosis when mixed with a sexually compatible strain of the opposite mating type. We verified that the clinical isolates were able to mate not only with reference strains but also with each other when associated in compatible combinations. We failed to detect any other genetic system, such as cytoplasmic incompatibility, in addition to the biallelic locus MATa/MAT, which controls the mating ability of two strains. Finally, no deviation from an equal distribution of the mating types MATa and MAT were observed in the sample. These results contrast with those of previous studies (5, 28, 38) which reported either the inability of some strains to reproduce sexually ("sterile" strains) or a marked imbalance in the MATa/MAT ratio or both. It should be emphasized that these previous studies were performed with much smaller samples (13, 9, and 5 strains, respectively) and with different culture conditions. From our experience, the emergence of sterile strains may be explained in several ways. First, we observed that the mating reactivity of C. lusitaniae was variable, some isolates exhibiting a high potential to mate and others having a lower reactivity. Monitoring mating and meiosis efficiencies in two different crosses of auxotrophic mutants provided evidence that high-mating-potential strains could yield up to 70% of meiotic products (relative to the total number of CFU) within 72 to 96 h of incubation in the optimal 18 to 28°C temperature range. In contrast, the ascospore yield from low-mating-potential strains was about 1,000 times lower, even though they were derived from the same lineage as high-mating-potential strains. Obviously, the use of a nonoptimal medium for mating poorly reactive strains, resulting in the release of few ascospores, may lead to a false interpretation. Second, some mutant strains can exhibit a sterile phenotype, such as pseudofilamentation-deficient mutants and some auxotrophic mutants, which can be prevented from mating until the medium is adequately supplemented. Third, a sterile phenotype may be assigned to a strain that has been misidentified as C. lusitaniae. In light of our results, there is no argument suggesting that an anamorphic state exists in clinical specimens of C. lusitaniae; all isolates analyzed are able to reproduce sexually and belong de facto to the species Clavispora lusitaniae (28). Accordingly, there is no reason to maintain two names for a species that is constituted by a homogeneous population with regard to sex. For historical and convenience reasons, it would be more advisable to keep the name Candida lusitaniae.

Imbalance in the mating type ratio is generally interpreted as a consequence of the loss of the ability to reproduce sexually, to the benefit of asexual dissemination. The fact that a 1:1 ratio for mating types a and  was observed in our sample leads us to speculate that sexual reproduction, i.e., meiosis, could still be used by C. lusitaniae as a source of genetic variability.

In our experiments, none of the media generally described as supporting the sexual reproduction of C. lusitaniae (5) was as effective as YCB medium, whose use was reported for mating the cactophilic yeast C. opuntiae (12). Conjugating cells, asci, and free ascospores were generally observed within 24 to 48 h after mixing and incubating in the 18 to 28°C temperature range two sexually compatible strains. This medium offers other advantages: it is delivered as a preformulated dehydrated powder and is completely synthetic, a feature which minimizes lot-to-lot variations. We verified that YCB did not support cross-mating reactions with reference strains from species that are sometimes mistaken for C. lusitaniae by standard methods (21) and thus confirmed a previous comparable work done with other Candida species (28). These data make the mating test suitable for C. lusitaniae identification.

Given that a clinical laboratory has mating type tester strains with a high potential to mate, developing a mating test is easy and inexpensive (a petri dish containing solid YCB medium costs less than U.S. $0.20). With a standard phase-contrast microscope, the results are simple to interpret by observing at least disrupted asci, which are easier to detect than ascospores. During a mating test, any C. lusitaniae isolate is supposed to mate with only one tester strain and not the other, depending on its sexual type. This is the sole situation in which a yeast isolate can be unambiguously identified as C. lusitaniae from a mating test. Thus, the mating test allows easy verification that takes no longer than any other identification system, and we believe that it is a reliable complementary tool in case of doubtful identification of C. lusitaniae.