Control of Filament Formation in Candida albicans by Polyamine Levels

Abstract

Candida albicans, the most common fungal pathogen, regulates its cellular morphology in response to environmental conditions. The ODC gene, which encodes ornithine decarboxylase, a key enzyme in polyamine biosynthesis, was isolated and disrupted. Homozygous null Candida mutants behaved as polyamine auxotrophs and grew exclusively in the yeast form at low polyamine levels (0.01 mM putrescine) under all conditions tested. An increase in the polyamine concentration (10 mM putrescine) restored the capacity to switch from the yeast to the filamentous form. The strain with a deletion mutation also showed increased sensitivity to salts and calcofluor white. This Candida odc/odc mutant was virulent in a mouse model. The results suggest a model in which polyamine levels exert a pleiotrophic effect on transcriptional activity.

Candida albicans can reversibly alter its mode of growth from a unicellular budding yeast to a filamentous form in the presence of inducing environmental signals. It has been observed that hyphae are able to adhere to and to invade host tissues more efficiently than the yeast form (10, 13, 37). For a mouse infection model it was recently reported that strains unable to form filaments in response to the known inducers of filamentous growth are avirulent (27). This finding supports the hypothesis that the morphological switch contributes to the virulence of this organism.

Identification of the signaling pathways that promote filamentous growth is currently under way. For C. albicans, the gene products of the mitogen-activated protein kinase (MAPK) cascade, CPH1, HST7, and CST20, have been isolated (24, 26, 27, 32). Homozygous null strains carrying mutations of these genes are defective in hyphal formation under certain conditions that induce wild-type Candida strains to form filaments. However, such mutant strains still form hyphae in response to human serum. For Saccharomyces cerevisiae, a branched pathway which is involved in pseudohyphal growth, depends on RAS2, and involves the ASH1 gene product has been described (7, 29). These findings are in agreement with previous observations that cyclic AMP promotes filamentous growth (36, 48). A putative third pathway (non-MAPK), which appears to act in parallel, involves the product of the Candida EFG1 gene (51), which is homologous to the S. cerevisiae PHD1 gene (17). A C. albicans cph1 efg1 double mutant is unable to accomplish germ tube emission in the presence of serum and is avirulent in a mouse model (27).

Polyamines are basic molecules required for cellular growth and differentiation in many organisms (21, 22, 52, 53). They stabilize RNA, stimulate DNA and RNA synthesis in vitro, and improve translation fidelity (35). In mammals and fungi, they are made via a pathway initiated by ornithine decarboxylase (ODC), which forms putrescine from ornithine. The level of ODC activity in quiescent cells is extremely low but strongly induced by a wide variety of stimuli, for example, during fungal spore germination (5, 53). Transient increases in the levels of ODC and polyamines take place during the yeast-hypha transition of the dimorphic fungi Mucor racemosus (23), Mucor rouxii, and Mucor bacilliformis (6), and ODC inhibitors, such as 1,4-diaminobutanone, inhibit the yeast-hypha transition in Aspergillus nidulans (50), M. rouxii and M. bacilliformis (45), Yarrowia lipolytica (18), and C. albicans (33). Recently, a dependence of the mode of growth on polyamine levels has been reported for the basidiomycete fungus Ustilago maydis (19).

In order to investigate whether polyamines play a role in the dimorphism of C. albicans, we isolated the gene encoding C. albicans ODC and observed that during germ tube emission a transient increase in ODC activity occurs. This increase was not accompanied by a rise in transcription levels, suggesting that ODC activity is regulated at the posttranscriptional level (28). Here, we report that an odc/odc mutant at low polyamine levels (putrescine 0.01 mM) fails to form filaments in response to serum or other known inducers of filamentous growth and is virulent in mouse models. Null strains were hypersensitive to calcofluor and salts, suggesting that polyamine levels control a broad range of gene functions.

MATERIALS AND METHODS

Yeasts strains, media, and growth conditions.

The strains of C. albicans used were SC5314, CAF4-2 Δura3::imm434/Δura3::imm434 (15), ABH1 (ura3/ura3 ODC/odc::hisG-URA3-hisG), ABH2 (ura3/ura3 ODC/Δodc::hisG), ABH3 (ura3/ura3 odc::hisG-URA3-hisG/odc::hisG), and ABH4 (ura3/ura3 Δodc::hisG/Δodc::hisG) (this work). Strains were maintained by periodic transfer to slants of yeast extract-dextrose (YED) medium (1% yeast extract, 1% glucose, 2% agar). Yeast growth was obtained in Lee medium (25) (28°C, pH 6.8) containing 1.25% glucose as a carbon source and supplemented with 0.2 mM uridine. Solid medium was obtained by adding agar (2%). Our solid medium for inducing the yeast-hypha transition was Lee medium in which glucose was replaced by mannitol (1.25%). The dimorphic transition was induced by changing the temperature to 37°C, by adding 4% bovine calf serum (GIBCO/BRL), or by changing the carbon source (glucose was replaced by 1.25% N-acetylglucosamine [GlcNAc]). When necessary, putrescine was added.

The Escherichia coli strains used for transformation and amplification of recombinant DNA were DH5α supE44 Δ(lacU169(φ80 lacZΔM15)hsdR17 recA1 endA1 gyrA96 thi-1 relA1 (20) and MV1190 Δ(lac-proAB) thi supE Δ(sr1-recA)306::Tn10(Tet^r) (F′ tra D36 proAB lacI^qZ ΔM15) (Bio-Rad). These strains were grown in Luria-Bertani or M9 medium plus the antibiotics necessary for selection (49).

DNA manipulations.

Total DNA from C. albicans was prepared as described previously for filamentous fungi (41) and was purified by centrifugation in CsCl gradients (49). Restriction enzyme digestions and DNA ligations were performed according to the recommendations of the manufacturers. Isolation of plasmid DNA from E. coli was performed by standard procedures (49). DNA fragments used as probes were labeled by random priming with [α-³²P]dCTP (Boehringer Mannheim) and used according to the instructions of the manufacturer. C. albicans cells were transformed by the spheroplast protocol (43).

RNA preparations and Northern analysis.

RNA was prepared by the method of Ausubel et al. (1). Prehybridization and hybridization were performed according to standard procedures (49).

Construction of C. albicans strains and plasmids.

Plasmid pSGS3 contains the Candida ODC gene (see Fig. 1) in a 2,217-bp fragment subcloned into the pGEM-T vector (Promega) (28). To construct an odc null mutant, we inserted the hisG-URA3-hisG cassette derived from plasmid pMB7 (15). The SalI-BglII fragment was inserted into plasmid pSGS3 between the EcoRI and BglII sites (nucleotide positions 1108 to 1863 of the ODC gene), giving rise to plasmid pAHH1 (Fig. 1). The EcoRI and SalI sites were blunted with the Klenow fragment of DNA polymerase I. Plasmid pAHH1 was linearized with the SphI-PvuII fragment and transformed into the Ura⁻ C. albicans strain CAF4-2 (15) to partially replace the coding region of one of the chromosomal ODC alleles with the hisG-URA3-hisG cassette by homologous recombination. Ura⁺ transformants were selected on medium without uracil, and integration of the cassette into the ODC locus was verified by Southern blot analysis. Spontaneous Ura⁻ derivatives of two of the heterozygous disruptants were selected on medium containing 5-fluoro-orotic acid (1 mg/ml; Diagnostic Chemicals Limited, Oxford, Conn.) as described by Boeke et al. (4), except that uracil was replaced by uridine (25 μg/ml). These clones were screened by Southern blot hybridization to identify those which had lost the URA3 gene via intrachromosomal recombination mediated by the hisG repeats. The procedure was then repeated to delete the remaining functional allele of ODC.

FIG. 1.

Restriction maps of the plasmids used in this work. pSGS3 contains a 2,217-bp XbaI fragment carrying the ODC gene of C. albicans (CaODC) in pGEM-T. pAHH1 is a plasmid containing the hisG-URA3-hisG cassette disrupting the ODC open reading frame. pAHH7 and pAHH8 are derivatives of pYPB1-ADHp7 constructed by insertion of the Candida ODC open reading frame under the control of the promoter of ADH1 and by insertion of the Candida ODC gene under the control of its own promoter (thus deleting the ADH promoter), respectively.

Candida ODC expression plasmids pAHH7 and pAHH8 (Fig. 1) were constructed by subcloning the Candida ODC open reading frame between the EcoRV and 3′ XhoI sites and Candida ODC under the control of its own promoter between the NotI and 3′ XhoI sites in pYPB-ADHp7 (9), kindly supplied by Al Brown, University of Aberdeen, Aberdeen, United Kingdom.

Phenotypic tests. (i) Zymolyase sensitivity.

Cultures of the wild type (CAF4-2) and of the mutant (ABH4) C. albicans strains were grown in Lee medium at low (0.01 mM) or high (10 mM) putrescine concentrations until the exponential phase. Cells were washed twice in water and resuspended in 10 mM Tris, pH 7.5. Approximately 1.5 × 10⁷ cells were resuspended in the same buffer containing Zymolyase 20T (ICN, Montreal, Quebec, Canada) at a concentration of 25 μg/ml. The optical density at 600 nm was measured at the start of incubation and every 20 min thereafter. The decrease in optical density reflected the proportion of cells that had lysed.

(ii) NaCl, calcofluor white, caffeine, and SDS sensitivities.

Methods for testing the C. albicans strains were similar for NaCl, calcofluor white, caffeine, and sodium dodecyl sulfate (SDS). Cultures were grown as for the Zymolyase assay and diluted to an optical density at 600 nm of 0.5, and 3-μl aliquots of fourfold serial dilutions of each cell culture were spotted onto Lee plates containing 0.01 or 10 mM putrescine (Sigma) and 0.5 M NaCl, 5 μg of calcofluor white (Sigma) per ml, 10 mM caffeine (Sigma), or 0.005% SDS (Sigma). Quantitative differences in the growth of the yeast cells at high or low polyamine levels were recorded after incubation of the plates at 28°C for 72 h.

Animal experiments.

Eight-week-old, male CFW-1 mice (20 to 22 g; Halan-Winkelmann, Paderborn, Germany) were inoculated with 10⁶ cells by intravenous injection (39). Survival curves were calculated according to the Kaplan-Meier method with the PRISM program (GraphPad Software, San Diego, Calif.) and compared by the log rank test. A P value of <0.05 was considered significant.

To quantify C. albicans CFU in kidneys, mice were sacrificed by cervical dislocation 48 h after injection and kidneys were homogenized in 5 ml of phosphate-buffered saline, serially diluted, and plated on YNG medium (0.67% yeast nitrogen base, 1% glucose, 0.2 mM uridine [pH 7.0]). Histological examination of kidney sections stained with periodic acid-Schiff stain was performed as described previously (3).

RESULTS

Chromosomal deletion of the ODC gene in C. albicans CAF4-2.

To investigate the biological function of polyamines in C. albicans, both copies of the ODC gene were deleted in C. albicans CAF4-2 by using the Ura3 blaster strategy (16). A linear 5.3-kb SphI-PvuII Δodc::hisG-URA3-hisG deletion fragment was constructed as described in Materials and Methods (Fig. 1) and used to transform CAF4-2. Deletion of the first ODC allele occurred in strain CAF4-2 at the BclI ODC locus, as seen by the appearance of a new 2.8-kb BclI fragment hybridizing with the 938-bp RasI-EcoRI fragment used as a probe (Fig. 2B, lane 2). After counterselection on 5-fluoro-orotic acid, a 3.2-kb BclI fragment resulting from the recombination between the two hisG direct repeats was detected in strain ABH2. This strain was used for a second round of transformation with the Δodc fragment. Deletion of the remaining BclI ODC allele in strain ABH3 was confirmed by the appearance of the 2.8-kb fragment (Fig. 2B, lane 4). Correct loop-out of the URA3 marker in strain ABH4 was confirmed by the appearance of the 3.2-kb fragment hybridizing with the probe (Fig. 2B, lane 5). Southern blot analysis was also performed with hisG and an ODC internal probe, confirming the correct genotypes of the ABH4 strain (not shown). Northern blots showed that the ODC transcript was absent in the corresponding homozygous deletion strain (not shown).

FIG. 2.

(A) Deletion of ODC in C. albicans. The open reading frame of the C. albicans ODC locus is shown. An EcoRI-BglII fragment was replaced by the hisG-URA3-hisG cassette. (B) Southern blot analysis with an ODC fragment from the RsaI-EcoRI fragment as the probe. Genomic DNA samples were digested with BclI (giving rise to a BclI-BclI band of 0.9 kb, an internal fragment of the ODC gene, and a second, ever-present BclI-BclI band [arrow] of 1.3 kb, containing part of the ODC gene and part of the promoter 5′ end). Lanes: 1, CAF4-2 (ura3/ura3 ODC/ODC); 2, ABH1 (ura3/ura3 ODC/odc::hisG-URA3-hisG); 3, ABH2 (ura3/ura3 ODC/Δodc::hisG); 4, ABH3 (ura3/ura3 odc::hisG-URA3-hisG/odc::hisG); 5, ABH4 (ura3/ura3 Δodc::hisG/Δodc::hisG).

While disruption of one ODC allele did not produce phenotypic changes (see below) and did affect growth in liquid or solid medium (Fig. 3b), the null mutant obtained displayed no ODC activity and behaved as a polyamine auxotroph (Fig. 3c). This result is in agreement with the finding of an identical single gene in three different C. albicans strains previously reported by members of our group (28). The null C. albicans mutant grown in polyamine-containing medium accumulated a large polyamine pool, which sustained its normal rate of growth in polyamine-free medium for 10 to 12 generations (between 22 and 30 h).

FIG. 3.

Growth of wild-type strain CAF4-2 (ura3/ura3 ODC/ODC) (a and d), ABH2 (ura3/ura3 ODC/Δodc::hisG) (b), ABH4 (ura3/ura3 Δodc::hisG/Δodc::hisG) (c), and ABH4 transformed with plasmid pAHH7 (e) or pAHH8 (f) in the absence of polyamines. Plates contained solid Lee medium (see Materials and Methods).

To check that the observed phenotype of the null mutant was indeed a consequence of the ODC gene disruption, we constructed two plasmids. pAHH7 contained a fragment of 1,508 bp carrying the ODC open reading frame under the control of the ADH1 promoter. pAHH8 contained the ODC gene under the control of its own promoter (Fig. 1; see also Materials and Methods). Both plasmids (pYDB1-ADHp7 served as a control) were used to transform C. albicans ABH4. The desired transformants (eight of each plasmid) were selected on the basis of their expected Ura⁺ phenotype. The transformants obtained with pYDB1-ADHp7 were unable to grow in the absence of putrescine. All the transformants obtained with pAHH7 and pAHH8 showed the phenotype and the growth characteristics of the wild-type strain (Fig. 3e and f). Our results indicate that both our XbaI-XbaI fragment and the ODC open reading frame under the control of the ADH1 promoter are able to complement the ODC null C. albicans mutant. Our results also demonstrate that C. albicans utilizes the ODC pathway as the sole mechanism for polyamine biosynthesis.

Defects in morphology and hyphal formation caused by deletion of both Candida ODC alleles. (i) Role of polyamines.

It has been reported that strains of C. albicans homozygous for mutations in the CST20, HST7, CPH1, and EFG1 genes (25, 26, 51) are partially defective in hyphal development on solid media and that only a double mutant lacking CPH1 and EFG1 functions fails to form filaments in response to serum or other known inducers of filamentous growth (27). To determine whether the polyamine concentration exerts an effect on germ tube formation, C. albicans wild-type (CAF4-2) (Fig. 4A) and odc/odc (ABH4) (Fig. 4B and C) strains were grown in yeast form in Lee medium (with 0.01 mM putrescine) until the exponential phase. Cells were plated onto solid Lee medium with mannitol, containing 0.01 mM (Fig. 4B) or 10 mM (Fig. 4C) putrescine. Fifty randomly selected colonies were analyzed for each test condition. On agar-containing plates, germ tube formation and hyphal growth were completely blocked in the odc/odc mutant at low (0.01 mM putrescine) polyamine levels (Fig. 4B, row 1). However, the mutant strain regained the ability to form hyphae at high (10 mM) polyamine levels in the medium (Fig. 4C, row 1). Its growth was indistinguishable from that obtained under normal conditions with the wild-type strain. Furthermore, the same behavior was obtained under conditions that induce ODC/ODC strains to form germ tubes and hyphae in liquid medium, such as high temperature, serum, or GlcNAc as the carbon source (Fig. 4, rows 2, 3, and 4, respectively). In the last case, cells were slightly more elongated (Fig. 4B, row 4). Filamentation was restored in all cases by adding putrescine at 10 mM (Fig. 4C), showing that the phenotype of the ABH4 strain is a consequence of the polyamine levels in the culture medium.

FIG. 4.

Colony growth and morphological characteristics of C. albicans homozygous odc strain at different polyamine concentrations. Column A shows the wild-type strain (CAF 4-2 ura3/ura3), column B shows strain ABH4 (ura3/ura3 Δodc/Δodc) grown at 0.01 mM putrescine, and column C shows strain ABH4 (ura3/ura3 Δodc/Δodc) grown at 10 mM putrescine. Wild-type (CAF4-2) and odc (ABH4) cells were grown in Lee medium, pH 6.7, at 28°C and incubated under conditions that promote germ tube formation and hyphal growth. Row 1 shows colonies of C. albicans cells grown for 7 days at 37°C on solid Lee medium with mannitol. Effects of temperature (37°C), serum (4%), and GlcNAc (1.25%) on germ tube emission are shown in rows 2, 3, and 4, respectively.

(ii) Additional phenotypic changes.

Using a screening method designed to identify genes involved in cell surface assembly, Lussier et al. (30) have found that a mutation in the S. cerevisiae ODC promoter produces sensitivity to calcofluor white and hypersensitivity to Zymolyase. These findings prompted us to carry out five phenotypic tests (for Zymolyase, NaCl, calcofluor white, caffeine, and SDS sensitivity) in order to further characterize and better define the odc/odc C. albicans null mutant. The results are shown in Fig. 5 and 6.

FIG. 5.

Resistance to a cell wall-degrading enzymatic complex of C. albicans. The wild type (CAF4-2 [ura3/ura3]) (▴) ABH3 (ura3/ura3 odc::hisG-URA3-hisG/odc::hisG) (▵), and ABH4 (ura3/ura3 Δodc/Δodc) were grown until the exponential phase in Lee medium, pH 6.7, at 28°C with a low (0.01 mM) or a high (10 mM) putrescine concentration, but respective results for low (■) and high (□) concentrations are shown only for ABH4. Aliquots (1.5 × 10⁷ cells) were resuspended in water and treated with 0.025 mg of Zymolyase per ml. The decrease in optical density (OD) (percentage of resistant cells) is plotted as a function of time.

FIG. 6.

Effect of 0.01 mM (A) and 10 mM (B) putrescine on the growth of C. albicans. Cell suspensions of strain ABH4 (ura3/ura3 Δodc/Δodc) were analyzed by spot assay for the ability to grow on solid Lee medium plates containing 5 μg of calcofluor white per ml (row 1), 0.5 M NaCl (row 2), 10 mM caffeine (row 3), or 0.005% SDS (row 4). Growth differences were monitored after 3 days at 28°C. WT, wild type.

The sensitivity of yeast cells to Zymolyase has been used to uncover changes in cell wall composition and arrangement (42). The data shown in Fig. 5 point to a similar kind of behavior regarding Zymolyase sensitivity for all the C. albicans strains tested, whether grown at low or high (0.01 or 10 mM putrescine, respectively) polyamine concentrations. The same results were obtained at higher Zymolyase concentrations (0.5 or 1 mg/ml [data not shown]). Our results differ from those described for S. cerevisiae, in which insertion of a transposon (Tn3::lacZ::LEU2) in the ODC promoter confers Zymolyase hypersensitivity (30).

Our second sensitivity test involved comparing the susceptibilities of the CAF4-2 and ABH4 strains (grown at low and high polyamine concentrations) by using a spot assay and four different compounds. Calcofluor white is a fluorescent dye that prevents microfibril assembly and interferes with the supramolecular organization of the cell wall (14, 42). A disturbed or weakened cell wall is not able to support drug concentrations that do not affect normal wild-type cells. Results shown in Fig. 6B, row 1, reveal no differences with respect to calcofluor sensitivity between our ODC/ODC C. albicans control strain (CAF4-2) and the null odc mutant strain (ABH4) grown at a high polyamine concentration (10 mM). However, a clear phenotype in response to calcofluor white appeared when the odc mutant was grown at a low polyamine concentration (0.01 mM) (Fig. 6A, row 1).

Stress due to increases in external osmolarity reduces the growth and viability of yeast cells owing to an array of effects, including the loss of an osmotic gradient across the plasma membrane (32). Therefore, we tested whether the level of polyamines affects osmolarity. The results are shown in Fig. 6, row 2. Growth in cells treated with 0.5 M NaCl (Fig. 6A, row 2) was severely inhibited in C. albicans ABH4 at 0.01 mM putrescine in comparison with the growth of the same strain at 10 mM putrescine or with the growth of the wild-type strain (CAF4-2) (Fig. 6B, row 2).

Two other phenotypic tests, caffeine (Fig. 6, row 3) and SDS (Fig. 6, row 4) treatment, revealed no differences between the wild-type strain and the null odc mutant grown at low (0.01 mM) or high (10 mM) polyamine levels. Taken together, our results demonstrate that although polyamines are essential, small amounts (0.01 mM) can support growth. However, at this low concentration several unexpected phenotypes appeared, and all of them could be reversed by increasing the amount of polyamines in the medium.

Virulence studies.

To determine the role of the Odc protein in virulence, mice were injected intravenously with wild-type and mutant strains and monitored for survival and for fungal invasion of kidneys. In agreement with the findings of a previous study (8), we observed that the Ura⁻ strain CAF4-2 was not pathogenic (Fig. 7A). However, infection with Ura⁺ wild-type cells (strain SC5314) resulted in rapid mortality (Fig. 7A). No difference in morbidity was found between mice infected with cells of Ura⁺ strains that were either wild type for ODC or had deletions of both alleles in ODC (Δodc/Δodc::URA3) (Fig. 7A and B). Also, no differences were observed with cultures grown at low (0.01 mM) or high (10 mM) putrescine levels, nor were any differences found with Ura⁺ strains with deletions of only one allele of ODC (results not shown).

FIG. 7.

Virulence assays. (A) Survival curves for mice (n = 10 for each C. albicans strain at each inoculation dose) infected with 10⁶ cells of C. albicans SC5314 (wild type) (●), CAF4-2 (ura3/ura3) (○), or ABH3 (ura3/ura3 odc::hisG-URA3-hisG/odc::hisG) grown at low (0.01 mM) (▴) and high (10 mM) (▵) polyamine levels. (B) Staining of mouse kidney sections with periodic acid-Schiff stain 24 h after infection with ura3/ura3 Δodc/Δodc::URA3 mutant strain ABH3 grown at low (0.01 mM [right]) or high (10 mM [left]) polyamine concentrations.

Histological examination revealed that cells of strains carrying deletions of both ODC alleles were able to form hyphae in infected kidneys (Fig. 7B). Our results, in contrast with those obtained for ura3/ura3 (8) or cph1/cph1 efg1/efg1 (27) C. albicans strains, indicate that the odc/odc mutant lacking a gene responsible for an essential function remains virulent.

DISCUSSION

Polyamines are essential for normal growth, as has been shown in many studies with mutants and pathway inhibitors (12, 19, 40, 44, 53, 54). However, the effect of polyamine starvation has not been clearly established. Some authors have reported that protein and nucleic acid elongation diminishes and that translation fidelity is impaired (34). Also, polyamines have a distributed charge, whose spacing may allow them to interact more flexibly with phosphates, DNA, and RNA (34). To unravel the biological function of polyamines in C. albicans, we deleted the ODC gene (Fig. 2) and analyzed the consequences of this deletion (i) on growth, (ii) on the effect of low (0.01 mM) and high (10 mM) polyamine concentrations on the yeast-hypha transition, (iii) in several phenotypes associated with cell wall defects, and (iv) on virulence.

The null odc mutant of C. albicans behaves as a putrescine auxotroph (Fig. 3), in agreement with the results described for other fungi, such as S. cerevisiae (16), Neurospora crassa (11), and U. maydis (19). Our results support the conclusion that in fungi the only functional pathway for putrescine biosynthesis involves ODC (53).

The fact that reintroduction of the gene in the null mutant restored the capacity to grow confirms this finding (Fig. 1 and 3). The C. albicans odc mutant was able to grow in the absence of polyamines (for about 24 to 30 h, 12 generations), indicating, in agreement with the results described for S. cerevisiae (2) and U. maydis (19), that C. albicans is able to accumulate a very large polyamine pool. Its cellular location is not completely established, but several different pools in the vacuole, in the cytosol, and in the nucleus have been described (12), and in light of our results (see below) this distribution may be relevant. Since we were able to completely deplete our odc mutant of polyamines, we were able to analyze its behavior in the presence of different amounts of polyamines in the growth medium.

Until now, strains of C. albicans homozygous for mutations in genes of the MAPK pathway (24, 26) or in a putative second pathway involving EGF1 (27, 31) still form hyphae in liquid cultures and in response to serum. Only the cph1/cph1 egf1/egf1 double mutants lacking the gene functions involved in both pathways are unable to form hyphae in response to all known inducers of filamentous growth (27). Here we show that in C. albicans, polyamine levels control the switch from the yeast form to a filamentous pattern (Fig. 4) (a similar result has been described for U. maydis [19]). This phenomenon is independent of the growth medium (either liquid or solid) and of the type of inducers of filamentous growth (serum, temperature, or GlcNAc).

Our results indicate that in some hitherto-unknown way, polyamine concentrations control the expression of the genes involved in both developmental pathways. Furthermore, growth at low polyamine levels also affords at least two new phenotypes: hypersensitivities to calcofluor white and salts. Zymolyase, caffeine, and SDS sensitivities remained unaffected (Fig. 5). The altered sensitivity to calcofluor white correlates well with the phenotype described previously for an S. cerevisiae strain with a mutation in the ODC promoter (30). However, our results for Zymolyase do not point to the hypersensitivity phenotype described for the S. cerevisiae mutant in the above-mentioned work (30). SDS disturbs the plasma membrane, and caffeine is an inhibitor of cyclic AMP phosphodiesterases (38). Thus, our results indicate that the PKC1-MPK1 signal transduction pathway is not affected by polyamine levels.

Taken together, all our results demonstrate that polyamines exert a pleiotrophic effect. As a working hypothesis, we propose that polyamine concentrations contribute to the transcriptional regulation of several genes, mainly those involved in cell differentiation. Two basic models could explain how polyamine levels might affect transcriptional activity in a gene-specific manner. One of them involves CpG methylation, which in mammalian DNA is involved in gene silencing. Thus, increased levels of polyamines can inhibit DNA methylation, permitting the expression of specific genes. Supporting this hypothesis, a low methylation level has been described for C. albicans (47), and in fact members of our group have previously shown that polyamines inhibit cytosine-DNA methylases (46). Our second working hypothesis is that polyamine levels alter nucleosomal conformation, which can increase the accessibility of transcriptional regulatory proteins to chromatin templates. Experiments to test both hypotheses are currently under way.

Finally, the Odc protein is not essential for virulence in a mouse model of systemic infections, in contrast to the results described by other authors for essential genes (8). The most likely explanation for this is that C. albicans is able to take up polyamines from both animal tissues and serum. Alternatively, it is possible that C. albicans could survive due to its high capacity to accumulate polyamines during its growth prior to infection in the mouse model.