A G₁ Cyclin Is Necessary for Maintenance of Filamentous Growth in Candida albicans

Abstract

Candida albicans undergoes a dramatic morphological transition in response to various growth conditions. This ability to switch from a yeast form to a hyphal form is required for its pathogenicity. The intractability of Candida to traditional genetic approaches has hampered the study of the molecular mechanism governing this developmental switch. Our approach is to use the more genetically tractable yeast Saccharomyces cerevisiae to yield clues about the molecular control of filamentation for further studies in Candida. G₁ cyclins Cln1 and Cln2 have been implicated in the control of morphogenesis in S. cerevisiae. We show that C. albicans CLN1 (CaCLN1) has the same cell cycle-specific expression pattern as CLN1 and CLN2 of S. cerevisiae. To investigate whether G₁ cyclins are similarly involved in the regulation of cell morphogenesis during the yeast-to-hypha transition of C. albicans, we mutated CaCLN1. Cacln1/Cacln1 cells were found to be slower than wild-type cells in cell cycle progression. The Cacln1/Cacln1 mutants were also defective in hyphal colony formation on several solid media. Furthermore, while mutant strains developed germ tubes under several hypha-inducing conditions, they were unable to maintain the hyphal growth mode in a synthetic hypha-inducing liquid medium and were deficient in the expression of hypha-specific genes in this medium. Our results suggest that CaCln1 may coordinately regulate hyphal development with signal transduction pathways in response to various environmental cues.

Candida albicans is the most prevalent fungal pathogen of humans. During infection, Candida undergoes a complex series of morphogenetic transitions, switching among a round, budding yeast form, a budding psuedohyphal form, and true tubular hyphae. The ability of Candida cells to develop filamentous hyphae has been shown to contribute to its virulence: mutants that are defective in hyphal formation have been shown to have a greatly reduced level of killing in a mouse system, presumably because the hyphal form facilitates translocation across tissues and subsequent penetration (10, 17, 29, 37). Therefore, considerable effort has been directed toward understanding how the switch from yeast to hyphal cells in Candida is regulated. A number of environmental conditions are known to affect cell type switching in vitro; these include pH, temperature, and medium composition. Several genes specifically expressed in hyphal cells have been cloned, but relatively little is known about their regulation or the molecular mechanism governing the developmental switch (2, 6, 21, 53).

The recent observation that Saccharomyces cerevisiae also undergoes morphogenetic switching (20) has led to a new approach toward understanding filamentous growth in C. albicans. The experimental tractability of S. cerevisiae provides an important advantage over approaches used to study morphogenesis in Candida, a much less studied diploid asexual yeast. A number of specific genes and regulatory pathways have been shown to have a role in regulating filamentous growth in Saccharomyces. Based on this information, the homologs of some of these genes have been cloned and mutated in C. albicans and thereby shown to be involved in hyphal development (19, 24, 28, 34, 35, 37, 55). For example, elements of a conserved mitogen-activated protein (MAP) kinase pathway are required for pseudohyphal growth in S. cerevisiae (35, 41, 42, 44, 49). The corresponding components of the same MAP kinase pathway have also been shown to be involved in hyphal development in Candida (24, 28, 34, 35). Two transcriptional regulatory genes, C. albicans EFG1 (CaEFG1), whose homolog, the PHD1 gene, promotes pseudohyphal growth in S. cerevisiae, and the global transcriptional repressor gene C. albicans TUP1 (CaTUP1), have also recently been shown to be involved in regulating the morphogenetic switch in Candida (9, 55). Many other genes have been reported to affect pseudohyphal growth in Saccharomyces (3, 4, 7, 8, 18–20, 27, 36, 38, 39, 43, 48). By analogy to the pheromone-responsive MAP kinase pathway and to PHD1 and TUP1, many of these genes are likely to have homologs in Candida that function in hyphal development.

Studies done with S. cerevisiae have suggested that the transition between round yeast-form cells and long filamentous cells may involve regulation of the cyclin-dependent kinase (Cdk) system. In Saccharomyces, Cdc28 is the major Cdk that controls cell cycle progression at the G₁/S-phase and the G₂/M-phase transitions (45). Specific cyclin subunits bound to Cdc28 dictate the proper timing of cell cycle events, presumably by mediating its specificity. Manipulating the relative concentrations of G₁ and G₂ cyclins or the activity of the Cdc28 protein kinase affects the extent of polarized cell growth. Activation of Cdc28 by the G₁ cyclins Cln1 and Cln2 (but not Cln3) promotes apical growth, while activation of Cdc28 by the mitotic cyclins Clb1 and Clb2 leads to isotropic growth (32, 33). Based on these and other results, Lew and Reed (32, 33) proposed a model in which cyclin-Cdk activities trigger different events in the morphogenesis cycle. Consistent with this model, the grr1 mutant, which stabilizes G₁ cyclins, also has enhanced filamentation (3, 4, 46). Furthermore, comparison of the cell cycle between the yeast form and the pseudohyphal form has revealed differences that implicate cyclins-Cdks in the regulation of filamentous growth. Unlike the yeast form, pseudohyphal cells divide symmetrically, have a shorter G₁ phase and a longer G₂ phase, and may possess an additional level of cell cycle control during G₂ (26). Taken together, these results suggest that the timing of the transition from G₁ cyclin-Cdk predominance to G₂ cyclin-Cdk predominance may be a regulated step involved in controlling the developmental switch between filamentous and yeast growth (25, 32).

In Saccharomyces, there are three major G₁ cyclins: Cln1, Cln2, and Cln3. While any of the three is sufficient to promote the onset of the cell cycle, they are not identical in function (13, 31, 56). Cln1 and Cln2 are more similar to each other, and their expression is strongly cell cycle regulated. Cln3 is only distantly related to Cln1 and Cln2, and its transcript level does not vary as dramatically through the cell cycle (45). Furthermore, the primary function of Cln3 is to activate the transcription of CLN1 and CLN2 through Swi4-Swi6 (56, 57), whereas Cln1 and Cln2 are much more potent activators of bud formation, DNA synthesis, and cell polarization (5, 12, 13, 33). To examine if G₁ cyclins are involved in filamentous growth, we mutated the G₁ cyclin genes in Saccharomyces. We found that diploid cln1 cln2 mutants are unable to form filamentous colonies on nitrogen starvation medium, while diploid cln3 mutants develop enhanced filaments (37a). Our genetic results suggest that Cln1 and Cln2 are involved in polarized cell growth during filamentation.

In order to determine whether G₁ cyclins play a role in Candida hyphal development equivalent to that observed for Saccharomyces pseudohyphal growth, we determined the null phenotype of C. albicans CLN1 (CaCLN1). In C. albicans, two putative G₁ cyclins have been isolated by their ability to complement a Saccharomyces strain conditional for G₁ cyclin activity or to confer resistance to pheromone-induced growth arrest upon overexpression (50, 58). However, the sequence similarities between these proteins and the Saccharomyces G₁ cyclins are very low. In this report, we show that CaCLN1 mRNA has a periodic expression pattern similar to those of S. cerevisiae CLN1 and CLN2 mRNAs. We further address the role of Cln1 in filamentous growth by studying the phenotypes of Cacln1/Cacln1 null mutants. As predicted, the Cacln1/Cacln1 mutants are somewhat slower than wild-type cells in cell cycle progression. We demonstrate that CaCln1 is necessary for the maintenance of hyphal growth on solid media and in liquid Lee’s medium (30) but is not required for germ tube formation or hyphal growth in liquid serum-containing medium. Consistent with the morphological phenotypes, Cacln1/Cacln1 strains have a medium-dependent impairment in the expression of hypha-specific genes.

MATERIALS AND METHODS

Strains.

The Candida strains used in this study are listed in Table 1. HLY strains were generated from CAI 4.

TABLE 1.

Candida strains used in this study

Strain	Relevant genotype	Source
SC5314	CLN1/CLN1 URA3/URA3	W. A. Fonzi
CAI4	CLN1/CLN1 ura3/ura3	W. A. Fonzi
HLY1472	cln1::hisG::URA3::hisG/CLN1 ura3/ura3	This study
HLY1476	cln1::hisG/CLN1 ura3/ura3	This study
HLY1488	cln1::hisG/cln1::hisG::URA3::hisG ura3/ura3	This study
HLY1509	cln1::hisG/cln1::hisG ura3/ura3	This study
HLY1586	cln1::hisG/cln1::hisG::CLN1::URA3 ura3/ura3	This study

Plasmid isolation and construction.

CaCLN1 genomic DNA (4.5 kb) (pHL399) and CaCLN2 genomic DNA (4 kb) (pHL402) were isolated by colony hybridization of a Candida genomic library in a Saccharomyces 2μm plasmid with PCR fragments of the respective coding sequences (34, 50, 58). The primers used for CaCLN1 had the sequences 5′CCGGAATTCCCATCCTCATACCATTCC and 5′CCGGAATTCCTGATTTATATTAACGTCAACGTC, and those used for CaCLN2 had the sequences 5′CCGGAATTCCTATCAATCCAAACATAGACAC and 5′CCGGAATTCGAAATGCAGAACATGATATTGTGG.

For CLN1 disruption (pHL419), a 1.3-kb upstream fragment that terminates at the SalI site at R23 and a 1.8-kb downstream fragment starting at the BclI site at L387 were cloned on either side of the hisG::URA3::hisG fusion in plasmid pMB7 (15). The linear fragment for integration was released by digestion with HindIII and KpnI before transformation.

For CLN1 complementation (pHL442), a BamHI/KpnI fragment from pHL399 was subcloned into a pBSII (Stratagene) vector in which the XbaI site had been destroyed. Then, the 5′ end of the insert was truncated to remove an XbaI site by releasing a BamHI/MluI fragment, followed by blunt-end ligation. Finally, a 1-kb PCR fragment containing the Candida URA3 gene was cloned into an XbaI site approximately 300 bp downstream of the termination codon of CLN1. The whole insert was then released by HpaI/KpnI digestion for transformation.

Microscopy.

Candida cell morphology was photographed on a slide by use of a Zeiss Axioskop microscope with a ×100 or ×40 objective and Nomarski imaging. Candida colony morphology was photographed on solid media by use of a Zeiss Stemi 2000 microscope at about ×1.6 with dark-field imaging.

For staining with 4′,6-diamidino-2-phenylindole (DAPI) and Calcofluor, Candida cells were fixed in 70% ethanol and washed with 50 mM Tris–50 mM EDTA before use. For Candida cells in yeast form, DAPI and Calcofluor staining was performed as described previously (47). For Candida cells in hyphal form, prior to DAPI staining, cells were treated with Zymolyase (Sigma) at 25 μg/ml in phosphate-buffered saline for 30 min at 37°C to reduce the level of aggregation. Staining of hyphal cells with Calcofluor results in too much fluorescence all over the cell surface. To reduce the level of background staining, we treated cells with 50 μg of pronase and 20 μg of proteinase A per ml in phosphate-buffered saline for 2 h at 37°C before staining them with Calcofluor. Fluorescence microscopy was performed by use of a Zeiss Axioplan 2 microscope with a digital charge-coupled device camera.

Cell cultures.

C. albicans strains were cultured essentially as described by Sherman et al. (51). Liquid Lee’s medium (30) was modified by the substitution of 1% mannitol (Fisher) for glucose and by the dilution of all the other components by a factor of two (to improve solubility). Solid Lee’s medium was prepared by the addition of Bacto Agar (Difco) to 1.5%. We prepared solid serum-containing medium by spreading 1 ml of newborn calf serum (Sigma) onto a 1.5% agar plate. Transformation of Candida was performed by the method of Ito et al. (23), with the addition of 0.1 M dithiothreitol to the transformation mixture during polyethylene glycol incubation.

Cell synchronization.

To isolate unbudded G₁ cells by centrifugal elutriation (14), Candida cells were grown to the early log phase in yeast extract-peptone–2% raffinose medium at 30°C. Raffinose was chosen as the carbon source because a higher percentage of cells are unbudded in a cycling culture grown in this medium than in YPD (glucose) medium (yeast extract-peptone-dextrose medium). Cells were collected by centrifugation, washed in ice-cold H₂O, resuspended in cold H₂O, sonicated to disperse clumps, and then loaded into the separation chamber of a Beckman JE-5.0 elutriation system maintained at 2,000 rpm. The pump pressure was gradually increased while the outflow was monitored microscopically. Unbudded cells were collected, concentrated by centrifugation, and then released into fresh prewarmed medium. Aliquots were removed periodically and centrifuged, and cell pellets were frozen in liquid N₂.

Northern blotting.

Total RNA was extracted from frozen cell pellets by phenol extraction (22). Formaldehyde gels were prepared and blotted essentially as described previously (22). DNA probes were labeled with a Stratagene Prime-It II random labeling kit and [α-³²P]dCTP (DuPont NEN). An internal ClaI-HindIII fragment of CaCLN1 and a ClaI-SalI fragment of C. albicans ACT1 (CaACT1) were used as probes. Oligonucleotides with the sequences 5′CGGTCTTGACGTTGAAGATTCC and 5′CCTTTGGTACCATGACACCTGA were used to amplify a region outside the conserved cyclin box to probe for CaCLN2. PCRs were used to generate DNA fragments for probing C. albicans ECE1, C. albicans HWP1, and C. albicans HYR1 (CaECE1, CaHWP1, and CaHYR1, respectively). The primers used had the sequences 5′TGACCGTGGAATTCAAGG and 5′GGACCATCTGCACCAGAAAGTG for HYR1, 5′GCCATCCACCATGCTCC and 5′GTGCTACTGAGCCGGCATCTC for ECE1, and 5′TGCTCCAGGTACTGAATCCGC and 5′GGCAGATGGTTGCATGAGTGG for HWP1. The sizes of mRNAs on our Northern blots correlated with the lengths expected based on information from the Candida Genome Database.

RESULTS

CaCLN1 is cell cycle regulated.

To address whether regulation of the cell cycle is important for morphogenesis, we began to characterize G₁ cyclin genes in Candida. We asked whether the previously reported Candida G₁ cyclin-like genes have the characteristic periodic cyclin mRNA expression pattern. In order to observe the periodicity of Candida G₁ cyclin expression, we purified synchronous populations of unbudded wild-type Candida cells by centrifugal elutriation and allowed the culture to reenter the cell cycle. Aliquots were removed from this synchronous culture at 10-min intervals, the percentage of budded cells was counted, and total RNA was prepared for RNA blotting (Fig. 1).

FIG. 1.

CaCLN1 transcripts are differentially regulated during the cell cycle. Synchronous wild-type Candida cells (SC5314) were purified by centrifugal elutriation. After release into the cell cycle, fractions were collected at 10-min intervals for determination of the budding index and total RNA preparation. (Top) Percentage of cells budded. Time is given in minutes. (Middle) The RNA was analyzed by Northern hybridization with CaCLN1, CaCLN2, and CaACT1 (40) as probes. (Bottom) The transcript levels of each gene were quantitated with a PhosphorImager, and the ratios of CaCLN1 to CaACT1 and CaCLN2 to CaACT1 were plotted for each time interval.

CaCLN1 and CaCLN2 genes had different patterns of expression during the cell cycle. CaCLN1 expression was periodic during the cell cycle. It peaked at 80 and 180 min after release, coinciding with the appearance of first and second buds. Bud emergence has been shown to occur as cells enter the S phase in Candida (52). Therefore, the time of CaCLN1 expression corresponds to the G₁/S-phase transition. At 140 min, expression was diminished, and at this point few cells had made a second bud; therefore, this time probably corresponds to the second G₁ phase after release. Thus, CaCLN1 transcripts have a pattern of periodic expression during the cell cycle similar to that reported for S. cerevisiae CLN1 and CLN2. CaCLN2 showed some periodicity, but less than that of CaCLN1. In addition, CaCLN2 was detectable in very small unbudded cells. CaCLN2 accumulated rapidly after release, reaching a maximum at about 60 min. This time precedes bud formation and corresponds to the G₁ phase. However, the transcript level was lower in the second cycle after release.

Disruption of the CaCLN1 gene.

In order to determine whether G₁ cyclins play a role in Candida hyphal development equivalent to that observed for Saccharomyces pseudohyphal growth, we investigated the null phenotype of G₁ cyclin genes. Because CaCLN1 has a transcriptional pattern most similar to that of S. cerevisiae CLN1 and CLN2 and the S. cerevisiae cln1/cln1 cln2/cln2 mutant is the strain most defective in pseudohyphal growth (37a), we decided to focus on this gene. The genomic CaCLN1 gene was isolated, and a construct was made to interrupt the coding sequences by following the two-step knockout strategy of Fonzi et al. (15, 16) (Fig. 2A). Both copies of CaCLN1 were successfully disrupted in more than 10 isolates, which were derived from two independent heterozygous strains. Deletion of the internal portion of CaCLN1 was confirmed by PCR (Fig. 2B) and Southern blotting (data not shown). The loss of the CaCLN1 message in isolates with double deletions was also confirmed by Northern hybridization (data not shown). The growth rates of the heterozygous and homozygous Cacln1 mutant strains are somewhat lower than that of the wild type. At 37°C, the wild-type strain has a doubling time of 60 min in YPD medium, while the Cacln1/Cacln1 mutant has a doubling time of 80 min. At 30°C, wild-type cells double every 70 min, while Cacln1/Cacln1 cells double every 90 min.

FIG. 2.

Targeted disruption of Candida CLN1. (A) Schematic representation of the disruption strategy. The location of the hisG insertion after selection for the loss of the URA3 gene is shown. (B) PCR analysis of transformants with disruption constructs. Putative transformants with the hisG::URA3::hisG construct were passaged over synthetic complete medium containing 5-fluoro-orotic acid and uridine in order to select for strains that had lost the hisG duplication and the intervening URA3 gene. Genomic DNA was isolated, and PCR was performed with appropriate primers spanning the portion of each cyclin gene that was replaced by hisG::URA3::hisG. The hisG insertion is slightly larger than the replaced sequence. Therefore, upon electrophoresis through an 0.8% agarose gel, the heterozygote (HLY1476) produces a doublet pattern, and the cln1/cln1 homozygote (HLY1509) produces a single, larger band. The Ura⁺ progenitors of these representative strains were used for all subsequent experiments (wild type, SC5314; CaCLN1/Cacln1, HLY1472; Cacln1/Cacln1, HLY1488). Lane m, markers.

Cacln/Cacln1 cells are slightly retarded in cell cycle progression.

To address whether there is any defect in cell cycle progression in the yeast form of the Cacln1/Cacln1 strain, we compared the timing of its cell cycle to that of the wild-type strain. We first determined daughter cell formation and nuclear division for Cacln1/Cacln1 and wild-type strains at 30°C. Synchronous cells were collected at 10-min intervals by releasing small unbudded cells prepared by elutriation into YPD medium at 30°C for yeast growth. Bud emergence and DAPI staining were used to measure daughter cell formation and nuclear division. The time from the maximally budded point of the first cell cycle after synchronization to the time of the maximally budded point of the second cell cycle is used as the time needed to complete one round of the cell cycle. In YPD medium at 30°C, wild-type cells took 80 min to complete one round, while mutant cells took about 100 min (Fig. 3B, panels a and b). The slower cell cycle progression associated with the Cacln1/Cacln1 mutant was reproducibly observed in two independent elutriation experiments. Like Saccharomyces G₁ cyclin cln1 cln2 and cln3 mutant cells (13), unbudded Cacln1/Cacln1 mutant cells purified by elutriation are slightly larger than wild-type cells. This larger cell size may explain why Cacln1/Cacln1 mutant cells entered the cell cycle earlier than wild-type cells, since cell size is known to be a critical determinant in cell cycle initiation in Saccharomyces (32).

FIG. 3.

Cell cycle studies of synchronous wild-type and Cacln1/Cacln1 cells in yeast and hyphal growth forms. Synchronous G₁ cells collected by centrifugal elutriation were released into either YPD medium at 30°C or YPD medium with 10% serum at 37°C. Cells were collected at 10-min intervals for fluorescence microscopy. (A) DAPI- and Calcofluor-stained Candida wild-type cells grown in YPD medium with 10% serum at 37°C at various times during the cell cycle. (B) (a) Percentage of budded cells after release into YPD medium at 30°C. (b) Percentage of cells with two nuclei after release into YPD medium at 30°C. (c) Percentage of cells with one chitin ring (circles), two chitin rings (squares), or three chitin rings (triangles) after release into YPD medium with 10% serum at 37°C. (d) Percentage of hyphal cells with two nuclei (circles) or three nuclei (squares) after release into YPD medium with 10% serum at 37°C. Closed symbols, wild type (SC5314); open symbols, Cacln1/Cacln1 (HLY1488). (C) Northern hybridization of synchronous G₁ cells released into YPD medium with 10% serum at 37°C. (Top) Synchronous G₁ wild-type Candida cells (SC5314) were purified by centrifugal elutriation and released into YPD medium with 10% serum at 37°C. Fractions were collected at 10-min intervals for total RNA preparation. Northern blots were probed with CaACT1 and CaCLN1. (Bottom) Transcription levels were quantitated with a PhosphorImager, and the ratio of CaCLN1 to CaACT1 was plotted for each time interval.

The timing of the cell cycle for wild-type and mutant cells during the induction of the hyphal growth mode was examined. Small unbudded cells from the same elutriation experiment as that described above were released into YPD medium with 10% serum at 37°C. Both wild-type and Cacln1/Cacln1 mutant cells were able to develop germ tubes under these conditions. Calcofluor was used to stain chitin ring structures (Fig. 3A). Since hyphal cells grow as a linear tube without any constrictions at each septum, bud emergence is difficult to define. Therefore, we used chitin ring formation as an indicator of hyphal septum formation. Chitin rings in yeast cells appear at the time of bud formation or immediately after bud emergence, indicating the G₁/S-phase transition (32). DAPI staining was used to visualize nuclear division (Fig. 3A). Figure 3B, panels c and d, shows the percentage of hyphal septum formation and the percentage of cells that have completed nuclear division for synchronous cultures at 37°C in YPD medium with 10% serum. The time required for the completion of a cell cycle is calculated from the time at which 50% of cells have one chitin ring to the time at which 50% of cells have two chitin rings. The lengths of the cell cycle were about 80 min for wild-type cells and 110 min for mutant cells (Fig. 3B, panels c and d). Therefore, as at 30°C, mutant cells are slower in cell cycle progression at 37°C than wild-type cells. In this experiment, we observed that cells released into hyphal growth medium became asynchronous more rapidly than elutriated cells released into YPD medium. This finding may reflect an intrinsic feature of cell cycle regulation during hyphal growth.

The elutriation experiment with wild-type cells released into yeast and hyphal growth conditions showed that the length of the cell cycle for serum-induced hyphae was about the same as that for the yeast form. Chitin ring formation and nuclear division occurred at approximately the same time for yeast cells and hyphal cells. These data indicate that the cell cycle during germ tube formation is similar to that during yeast growth. However, this conclusion is necessarily limited to the specific hyphal induction condition used in this experiment. Furthermore, we can only examine the first two divisions after hyphal induction in this manner. Therefore, timing of the later cell cycles for the hyphal form may differ from that for the yeast form, as suggested by Kron and Gow (25).

Polarized apical growth precedes the periodic expression of CaCLN1. Our Calcofluor labeling experiment shows that initial apical growth or germ tube formation preceded chitin ring formation (Fig. 3A). Small unbudded cells released into YPD medium at 30°C did not start to bud until 80 min, while cells released into YPD medium with 10% serum at 37°C initiated apical growth after 40 min. Chitin rings in hyphal cells formed at 80 min, about the same time as bud formation in yeast cells. Therefore, the formation of chitin rings lagged behind initial apical growth by 40 min during hyphal induction. The time of budding and hyphal septum formation coincided with the peak of CaCLN1 expression in several Northern hybridization experiments that we performed with synchronized cells released into serum-containing YPD medium at 37°C (Fig. 3C).

Hyphal growth is defective in Cacln1/Cacln1 strains.

The capacity of Cacln1 mutant strains to produce filaments was examined under many hypha-inducing conditions. Cacln1/Cacln1 strains have a profound defect in filamentous growth. The most dramatic effect is found on modified Lee’s medium, a defined hypha-inducing medium containing moderate levels of nitrogen as ammonium, salts, amino acids, and mannitol as a carbon source (30). As shown in Fig. 4A, after 3 days on Lee’s medium plates, the wild-type strain produces abundant filaments at the periphery of the colony, while Cacln1/Cacln1 and CaCLN1/Cacln1 mutant strains display none. After 5 days, some short filaments are observed around heterozygous colonies, but filaments are never observed around homozygous colonies. Furthermore, in contrast to the findings for wild-type cells, no filaments are observed under homozygous colonies after scraping or washing the mutant cells on the surface away (data not shown). We believe that the defect in hyphal development is probably not due to the lower growth rate of Cacln1/Cacln1 strains because a longer incubation time did not allow the mutant strains to develop filaments. The phenomenon of a heterozygous mutant exhibiting a phenotype similar to that of the corresponding homozygote has been observed previously for Candida (24).

FIG. 4.

Phenotypes of Cacln1 mutant strains. (A) Colony phenotypes of Cacln1 mutant strains. Ura⁺ strains that were CaCLN1/CaCLN1 (SC5314), CaCLN1/Cacln1 (HLY1472), and Cacln1/Cacln1 (HLY1488) were plated at a density of approximately 20 colonies/plate on solid Lee’s and serum-containing media. The plates were incubated at 37°C for 3 or 5 days and then photographed. (B) Phenotypes of Cacln1 mutant strains in liquid media. Overnight cultures of the Ura⁺ strains were diluted into YPD medium with 5% serum or modified Lee’s medium and incubated at 37°C for 4 and 15 h, respectively.

Cacln1 mutants also have a defect in the production of hyphal colonies on solid serum-containing medium (2% agar plated with 1 ml of serum). Both wild-type and Cacln1 mutant cells are capable of the initial response to serum, the production of long germ tubes (data not shown). However, after 1 day of incubation, the mutant strain begins to make predominantly round cells while the wild-type strain continues to grow exclusively as hyphae (data not shown). Figure 4A shows growth on serum plates after 3 days of incubation. As it does on Lee’s medium, the CaCLN1/Cacln1 strain exhibits an intermediate level of filament formation.

The ability of Cacln1 mutants to produce filaments was also tested with liquid media. We found little difference between mutant and wild-type strains in their ability to initiate germ tubes and produce elongated filaments in 5% newborn calf serum in YPD medium. Figure 4B shows hyphal cells after 4 h of serum induction at 37°C from overnight stationary-phase cultures. However, both wild-type and Cacln1/Cacln1 mutant cells started to convert from hyphal growth form to yeast growth form after 6 to 8 h of incubation, and little difference in sustained hyphal formation was found between wild-type and mutant cells. However, we did observe that the mutant strain was slightly less clumpy during serum-induced hyphal formation than the wild-type strain (data not shown). In contrast to their behavior in liquid serum-containing medium, in liquid modified Lee’s medium, Cacln1/Cacln1 strains were significantly impaired in the maintenance of the hyphal growth mode. Figure 4B shows the defect in filamentous growth after 15 h in Lee’s medium. The Cacln1/Cacln1 mutant cells were able to initiate germ tubes and long cells at the beginning of the induction (data not shown), although the percentage of long cells was slightly lower in the mutant cells than in the wild-type cells. However, after prolonged incubation, the mutant cells mostly switched to a more pseudohyphal and then a yeast-form growth mode, whereas the wild-type cells continued to grow as more than 90% hyphal cells. Consistent with our other results, the CaCLN1/Cacln1 strain exhibited an intermediate phenotype.

To confirm that the defect in hyphal development observed in the Cacln1/Cacln1 mutant was caused by CaCLN1 disruption and not by other mutations introduced during two rounds of transformation, we replaced one of the Cacln1::hisG alleles with a wild-type CaCLN1 allele. Figure 5A is a schematic diagram of the Cacln1::CaCLN1 complementation construct. Reintroduction of a wild-type copy of CaCLN1 into the Cacln1 locus was confirmed by PCR and Southern blotting (data not shown). Two independent Cacln1/Cacln1::CaCLN1 transformants were obtained, and both strains regained competence to produce hyphal colonies. The quality of the hyphae produced by the transformants was intermediate between that of wild-type and CaCLN1/Cacln1 heterozygote strains on solid Lee’s medium (Fig. 5B). The transformants also generated robust hyphal filaments in liquid Lee’s medium, more similar to those of the wild-type strain than to those of the CaCLN1/Cacln1 heterozygote strain (data not shown). Therefore, reintroduction of a wild-type copy of CaCLN1 into the Cacln1/Cacln1 strain can complement the mutant defect in hyphal development.

FIG. 5.

Complementation of Cacln1/Cacln1 strains by wild-type CaCLN1. (A) Schematic representation of the complementation strategy. (B) Colony phenotypes with CaCLN1 complementation. Ura⁺ strains that were wild type (+/+) (SC5314), CaCLN1/Cacln1 (+/−) (HLY1472), Cacln1/Cacln1 (−/−) (HLY1488), and CaCLN1-transformed Cacln1/Cacln1 [−/− (+)] (HLY1586) were plated onto modified Lee’s medium and incubated at 37°C for 4 days.

Cacln1/Cacln1 mutants are defective in the hypha-specific transcriptional program.

Because we observed a morphological defect for the Cacln1/Cacln1 strain in liquid Lee’s medium, we began to wonder whether the expression of molecular markers for hyphal growth depends on CaCLN1. We examined the induction of three genes, HYR1, ECE1, and HWP1, known to be hypha specific (2, 6, 54). Overnight cultures were diluted 1:100 either into YPD medium for the yeast form of growth or into YPD medium plus 10% serum or Lee’s medium for hyphal induction. Transcripts of these genes were undetectable in wild-type cells released into YPD medium at 30 or 37°C (Fig. 6). Furthermore, as expected, their mRNA levels increased dramatically within 60 min of the switch to the serum-containing medium. Transcriptional induction of hypha-specific genes in Lee’s medium was slower. The level of transcripts at 3 h was generally lower than the 1-h induction seen in serum (Fig. 6). The levels of hypha-specific transcripts are consistent with the observed morphological effects. Germ tubes form immediately in response to serum, while hyphal development in Lee’s medium starts after 3 h. The expression of hypha-specific genes in the Cacln1/Cacln1 mutant is consistent with the observed morphological phenotypes. The induction of HYR1, ECE1, and HWP1 in Cacln1/Cacln1 cells in response to serum is slightly reduced (about 50%) in comparison to that in wild-type cells (Fig. 6). In liquid Lee’s medium, the Cacln1/Cacln1 strain shows a profound defect in the transcriptional activation of all three hypha-specific genes. The difference in transcriptional induction between wild-type cells and mutant cells is even more dramatic after prolonged incubation. By 8 h of incubation in Lee’s medium, wild-type cells have over several hundredfold the induction seen for CaECE1 and CaHWP1; the level of hypha-specific transcripts in the mutant cells remains at initial induction levels. The transcriptional defect in this strain precedes the detection of the morphological defect and is considerably more pronounced than the morphological impairment. This result suggests that G₁ cyclins may play a relatively direct role in regulating the transcriptional program of hypha-specific genes.

FIG. 6.

Hyphal gene expression is reduced in a Cacln1 mutant. (Top) Wild-type and Cacln1/Cacln1 (HLY1488) cells were diluted from overnight cultures into the indicated liquid media for the indicated times. RNA was analyzed by Northern hybridization with CaHYR1, CaECE1, CaHWP1, and CaACT1 as probes. (Bottom) Transcriptional levels of hypha-specific genes were quantitated against CaACT1 with a PhosphorImager to obtain ratios of CaHYR1 to CaACT1, CaHWP1 to CaACT1, and CaECE1 to CaACT1 for each RNA sample. The ratios of each hypha-specific gene to ACT1 are shown graphically. Bars (black, +/+ genotype; grey, −/− genotype) correspond to lane labels above top panels.

DISCUSSION

G₁ cyclins in Candida.

S. cerevisiae has three major G₁ cyclins. Cln1 and Cln2 are more homologous to each other in protein sequence, and their expression is cell cycle regulated, while Cln3 is only distantly related to Cln1 or Cln2, and its expression does not vary as dramatically through the cell cycle. Two putative G₁ cyclins, CaCln1 and CaCln2, have been cloned from C. albicans by functional complementation in Saccharomyces (50, 58). However, the protein sequence homologies between the Candida proteins and the Saccharomyces G₁ cyclins are very low, with the highest similarity to Cln3 (50, 58). Our finding that CaCLN1 transcripts are periodically expressed during the cell cycle suggests that CaCLN1 is a bona fide cyclin gene. Because its transcript levels peak at the time of bud formation, it is likely to be similar in function to the S. cerevisiae CLN1 and CLN2 genes. This notion agrees with the finding that the overexpression of CaCLN1 can cause pheromone resistance in Saccharomyces, as had been reported for S. cerevisiae CLN2 (11, 58). CaCLN2 shows a subtle periodicity in its transcription. CaCLN2 transcripts are expressed in unbudded G₁ cells and peak earlier than CaCLN1 transcripts. In addition, CaCLN2 expression is lower in the second cell cycle. The pattern of CaCLN2 expression is somewhat similar to that of S. cerevisiae CLN3 (57), suggesting that CaCLN2 might be the counterpart of S. cerevisiae CLN3 in Candida.

The role of CaCln1 in filamentous growth.

Cacln1/Cacln1 mutants are unable to maintain the hyphal growth mode. This phenotype was observed on solid media and in liquid Lee’s medium. The defect in hyphal colony formation agrees with our findings for Saccharomyces cln1 and cln2 mutants (37a). It also supports the previously reported finding that high levels of G₁ cyclins promote polarized cell growth (3, 33). We believe that the defect in hyphal maintenance in Cacln1/Cacln1 mutants is probably not due to its lower growth rate, because the mutant strains converted to yeast growth earlier than the wild-type strains. In addition, a longer incubation time did not allow the mutant strains to develop filaments.

Cacln1/Cacln1 mutants have immediately detectable defects in both cell cycle progression and transcription of hypha-specific genes. We found that Cacln1/Cacln1 mutant cells are slower in cell cycle progression than wild-type cells. Although we did not have a marker for the S phase in our experiments, it is possible that CaCln1 is involved in the timely progression of the cell cycle into the S phase in a manner analogous to that of S. cerevisiae Cln1 and Cln2. Our data also suggest that the G₁ cyclin kinase may play a role in regulating the transcription of hypha-specific genes. We observed that Cacln1/Cacln1 mutants have rapidly arising defects in the transcription of several hypha-specific genes in both serum-containing medium and liquid Lee’s medium and that this defect precedes the development of the morphological phenotype. The transcriptional defect is much more pronounced after longer incubation in Lee’s medium. The defect in the transcriptional program of hypha-specific genes may account for part of the requirement for CaCln1 for the maintenance of hyphal growth. Several transcription factors, such as Tup1, Efg1, and Cph1, have been shown to regulate hyphal growth (9, 37, 55), suggesting that some of the hypha-specific genes regulated by these transcription factors are responsible for hyphal development. The three known hypha-specific genes, ECE1, HWP1, and HYR1, are not required for morphogenesis, as mutations in them do not affect filamentation (2, 6, 54). However, this fact may be due to the existence of other, functionally redundant genes, since genes with high sequence similarities to ECE1 and HYR1 have been discovered in the Candida Genome Sequencing Project. Another possibility is that other hypha-specific genes required for filamentation have not been discovered.

CaCln1 may have additional roles in filamentous growth, perhaps maintaining the polarization of the actin cytoskeleton, which is thought to contribute to the highly active apical growth at the hyphal tip (1). Like Saccharomyces cells, Candida cells in the yeast form of growth display a temporal change in the organization of the actin cytoskeleton during cell cycle progression (1). Studies with Saccharomyces have suggested that this temporal regulation of the actin cytoskeleton may be regulated by the Cdk Cdc28 (32). High G₁ cyclin-Cdk levels will result in a polarized actin cytoskeleton, and high B-type cyclin-Cdk levels will cause a depolarized actin cytoskeleton. In contrast to the situation during yeast growth, the majority of the actin cortical patches are concentrated at the tip of hyphal filaments during hyphal growth (1). Therefore, one role of CaCln1 in hyphal growth may be to maintain the polarization of the actin cytoskeleton. This activity of CaCln1 may be necessary for repressing the effect of the B-type cyclin-Cdk activity in the depolarization of the actin cytoskeleton in cycling cells (32) and is not required for polarized growth in cells with low B-type cyclin-Cdk activity. This notion is in agreement with our observation that stationary-phase Cacln1/Cacln1 cells are able to initiate apical growth when diluted into liquid Lee’s medium. Similarly, initial apical growth from early G₁ cells precedes hyphal septum formation (Fig. 3A and C) and the expression of cyclin genes. On the other hand, during prolonged hyphal growth post-Cdk activation, CaCln1 may be necessary to balance the effect of Cdk activities on the actin cyctoskeleton to prevent isotropic growth throughout the cell cycle.

Relationship between CaCln1 and signal transduction pathways in regulating hyphal development.

We observed that Cacln1/Cacln1 cells are defective in hyphal development in liquid Lee’s medium but that their hyphal growth is virtually unaffected in serum-containing liquid medium. This result suggests that two different signal transduction pathways may be responsible for hyphal development in serum-containing and Lee’s media. Apparently, CaCln1 is more important for hyphal growth in Lee’s medium than in serum-containing medium. This idea of two different signaling pathways is also in agreement with the fact that hyphal induction in serum-containing medium is much faster than that in Lee’s medium. Several independent signal transduction pathways or regulatory proteins have been shown to be involved in hyphal development. These include a conserved MAP kinase pathway, an integrin-mediated pathway, and transcriptional factors, such as Tup1 and Efg1 (9, 17, 24, 28, 29, 34, 37, 55). Interestingly, the MAP kinase pathway and Efg1 have been mapped to two parallel pathways, because a double mutant blocks hyphal development under all growth conditions tested, while a single mutant does not (37). The CaCln1-Cdk kinase may function in parallel with these signal transduction pathways or as a potential target of one of the signal transduction pathways.

A crucial question concerning the role of cyclin-dependent kinase in filamentous growth is whether it is regulated in response to any extracellular signals. A careful comparison of budding and nuclear division between yeast growth and serum-induced hyphal growth did not reveal any changes in the timing of the cell cycle (Fig. 3B). However, our data cannot exclude the possibility that certain extracellular signals for hyphal development can induce the dimorphic switch by regulating the cell cycle machinery. Considering that Candida cells can integrate a large number of extracellular signals and growth conditions as part of their dimorphic regulation, there are undoubtedly many mediators of these signals. Given the central role of the Cdk system in directing the localization of cell growth, it presents a likely target for some of these regulatory pathways.