INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Candida (Torulopsis) glabrata, an imperfect fungus, is a haploid yeast of the genus Candida and has been demonstrated to be a pathogen of opportunistic yeast infections (1). There are increasing concerns over C. glabrata, because it causes not only mucocutaneous but also systemic infections in transplant and immunosuppressed patients (21, 58, 59). Moreover, the extensive use of topical and systemic antifungal drugs has resulted in the appearance of azole-resistant infections with Candida species, including C. glabrata (41, 59). Thus, there is a need to develop new antifungal drugs with novel modes of action and broad spectra.

Fungal cell wall biosynthesis is one possible target for new antifungal drugs, since it is essential for fungal viability and does not occur in mammals (18, 19). Fungal cell wall biosynthesis has been studied quite extensively in Saccharomyces cerevisiae (11, 14, 30) and Candida albicans (5, 11, 19, 36, 37) but not in C. glabrata. However, in addition to the advantage of its haploidy in genetic manipulation, recent progress on the molecular biology of C. glabrata, including development of host-vector systems (28, 29, 60), a controllable gene expression system (40), and the isolation of several structural sequences (17, 28, 35, 44), provides us with an opportunity to study cell wall biosynthesis in this organism.

-1,6-Glucan is a component of fungal cell walls, where it occurs as a polymer covalently attached to glycoproteins (26, 38) and to other cell wall structural polymers such as -1,3-glucan and chitin (14, 30). In S. cerevisiae, many genes involved in -1,6-glucan synthesis were isolated through mutations (kre [killer resistant] mutations) that confer resistance to the K1 killer toxin, which kills sensitive yeast cells following binding to this -1,6-glucan polymer (4, 6, 8, 15, 34, 48, 49). While other genetic studies have identified additional genes affecting cellular levels of -1,6-glucan (24, 25, 46, 55), it still remains unclear how these genes, including the KRE genes, are concerned in -1,6-glucan biosynthesis. Among them, KRE9 and its homolog KNH1, genes encoding cell surface O glycoproteins, are required for -1,6-glucan synthesis in S. cerevisiae (6, 8, 15). The S. cerevisiae kre9 null mutant shows several phenotypes: resistance to K1 killer toxin; slow growth, especially on glucose media; an 80% reduction of alkali-insoluble -1,6-glucan; and defects in cell separation. Overexpression of KNH1 can partially suppress these phenotypes of a kre9 null mutant (15). Although a knh1 null mutant showed no obvious phenotype, disruption of both KRE9 and KNH1 was synthetically lethal (15). Further, the SKN7 gene encoding a yeast homolog of bacterial two-component regulators has also been isolated as a multicopy suppressor of the slow-growth phenotype of the kre9 null mutant (7, 9). Recently, a homolog of the KRE9 gene has been isolated from C. albicans (33).

Here we report isolation of the KRE9 and KNH1 homologs in C. glabrata and several lines of evidence, including the first analysis of cell wall components in C. glabrata, suggesting evolutionary conservation of these molecules as essential components of -1,6-glucan synthesis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains, growth media, and procedures. The S. cerevisiae and C. glabrata strains used in this study are listed in Table 1. YPD and YPGal are complex yeast media with 2% glucose and 2% galactose, respectively, and YNB is a synthetic medium with either 2% glucose or 2% galactose and supplemented for auxotrophic requirements. Yeast transformations were carried out by the modified lithium acetate method (20, 23) and the one-step transformation method (12). Tetracycline assays were carried out as previously described (39). Seeded-plate assays for killer toxin sensitivity were performed as previously described (8). Spotting assays were performed as previously described (31). 5-Fluoro-orotic acid, G418 (Geneticin), and calcofluor white (CFW) were purchased from PCR Inc. (Gainesville, Fla.), GIBCO BRL (Grand Island, N.Y.), and Polysciences Inc. (Warrington, Pa.), respectively. Plasmid DNA was propagated in Escherichia coli XL-1-blue cells (Stratagene, La Jolla, Calif.).

Manipulation of DNA. Techniques for manipulation of DNA were performed as previously described (52). Yeast genomic DNA was prepared as previously described (51). Southern blots were performed by using nylon membranes (Hybond N; Amersham Canada Limited, Oakville, Ontario, Canada) and following the instructions of the manufacturer. A PCR fragment harboring the entire coding sequence for S. cerevisiae KRE9 was used as a probe. DNA sequencing was performed by the dideoxy method (53) on an ABI 373A sequencer with Bluescript universal and reverse primers and synthetic oligonucleotides complementary to specific regions of CgKRE9 and CgKNH1.

Plasmids. A 0.7-kbp HindIII fragment harboring the tetO-HOP1 chimeric promoter and a 1.4-kbp NotI fragment harboring the kanamycin resistance gene (Kan^r) were excised from p97t (39) and pKanMX2 (57), respectively, and systematically cloned into Bluescript SKII+ (Stratagene) to generate p97tKan. A 0.4-kbp SpeI-SacII fragment of pMPY-ZAP (54), harboring the hisG sequence, was blunted with T4 DNA polymerase (GIBCO BRL) and cloned into the EcoRV site of Bluescript SKII+ to construct phisG+ and phisG. The latter plasmids have their hisG sequences in opposite orientations. A 0.4-kbp SmaI-EcoRV fragment of phisG+, a 1.1-kbp SmaI-HindIII fragment of pMPY-ZAP (harboring the S. cerevisiae URA3 gene), and a 0.4-kbp HindIII-SmaI fragment of phisG were systematically cloned into Bluescript SKII+ to generate pSNZAP3, harboring a modified hisG-URA3-hisG module.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Disruption of CgKRE9 and CgKNH1 and morphological effects of the deletions. (A) Disruption of CgKRE9. A PCR-amplified fragment (double-headed arrow) from pCGK9T (Materials and Methods) was used for the one-step gene replacement. (B) Disruption of CgKNH1. A KpnI-SacI fragment of pCGK1H (Materials and Methods) was used for the one-step gene replacement. Homologous recombination between the two regions (hatched boxes) resulted in disruption of the chromosomal copy. The wild-type strain, 2001HTU (C), cgkre9 deletion strain SNBG1-7-7 (D), and cgknh1 deletion strain SNBG2-26 (E) as viewed by Normarski optics are shown. Cells precultured on galactose medium were cultured on glucose medium.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Construction of a tetracycline-sensitive mutant of CgKRE9 (Tets CgKRE9). (A) Scheme for replacement of the cognate CgKRE9 promoter region with the tetracycline-responsive promoter. A PCR-amplified fragment (double-headed arrow) from pCGK9tetAB (Materials and Methods) was used for the one-step gene replacement. The solid arrow indicates the ORF of CgKRE9. Open and shaded boxes indicate the S. cerevisiae URA3 gene and the tetracycline-responsive promoter, 97t, respectively. Homologous recombination between the two regions (hatched boxes) resulted in generation of the Tets CgKRE9 mutant. (B) Growth inhibition by tetracycline on the Tets CgKRE9 mutants. A total of 104 cells were inoculated and were cultured on YPD (solid bars) or on YPGal (open bars) for 20 h at 30°C. Growth of cells with tetracycline (50 µg/ml) is expressed as percent of optical density at 600 nm of cells without tetracycline. As the wild type (WT), strain ACG22 (Table 1) was used. Error bars, standard deviations.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Restriction maps and deletional analysis of inserts of C. glabrata genomic DNA on pSB2-1 and pSBG9-1. Open bars indicate the inserts on pSB2-1 (A) and pSBG9-1 (B). Fragments used for deletional analysis are represented by solid bars. The presence and absence of complementation activity in Tets KRE9 knh1 cells are indicated as + and , respectively. Arrows indicate ORFs of CgKRE9 (A) and CgKNH1 (B). Hatched bars indicate regions with homology to the syntenic S. cerevisiae genes.

Construction of tetracycline-sensitive mutants of S. cerevisiae KRE9 (Tet^s KRE9). Replacement of the cognate KRE9 promoter with the tetracycline-responsive promoter, 97t (39), was achieved by the one-step gene replacement method (3, 54) with slight modifications. A DNA fragment was amplified by PCR using p97tKan as a template and a pair of primers (5'-GAATAGAACAGGAGTCTCAAAGCATTCTTGAAGCCAGATTGCAACAGCTATGACCATG-3', 5'-AAAGCACATATGATGGAATTTCTTTGTAAACGCATTATGAATTCT TTTCTGAGATAAAG-3') and subsequently was used for transformation of the S. cerevisiae strain FAHAP4, which harbors the tetR-HAP4AD fusion activator gene (39). After selection on G418-containing plates, the correct integration was confirmed by PCR and the strain was designated SNB50-1. Disruption of the KNH1 gene in SNB50-1 was achieved by using a DNA fragment amplified by PCR using pSNZAP3 as a template and a pair of primers (5'-CTGATAGTATTATTCTTAACATTATTTTGTTCGGTAGTGTTCCGTAAAACGACGGCC AGT-3', 5'-CATTATCTGTGCCTCAAAGCATTAACTTTTCTTGCAGTCAGAGAAACAGCTATGACCATG-3'). The correct integration was confirmed by PCR. The strain was subjected to 5-fluoro-orotic acid selection and finally designated SNB54-5 after the elimination of the URA3 gene was confirmed by PCR.

Cloning of C. glabrata KRE9 and KNH1 genes. SNB54-5 cells were transformed with a pRS424-based C. glabrata subgenomic bank, harboring EcoRI 4- to 7-kbp fragments of C. glabrata genomic DNA, and spread onto both YNB-glucose and YNB-galactose plates containing tetracycline (50 µg/ml). After incubation at 30°C for 3 days, colonies appeared on the plates, cells were collected, and plasmid DNA was recovered from them.

Disruption of CgKRE9 and CgKNH1 and construction of tetracycline-sensitive mutants of CgKRE9 (Tet^s CgKRE9). Disruption of CgKRE9 in strain 2001HTU was achieved by using a DNA fragment amplified by PCR using pCGK9T as a template and a pair of primers (5'-CCATCGATGAATTCATGCTGCTGCTGGCTATACTGCTATC-3', 5'-CAACTGGACAAATATCTAAC-3') (Fig. 1). The correct integration was confirmed by PCR, and the strain was designated SNBG1-7-7. A KpnI-SacI fragment of pCGK1H was used to disrupt CgKNH1 (Fig. 1) in strain 2001HTU. The correct integration was confirmed by PCR, and the strain was designated SNBG2-26.

Cell wall component analysis. The levels of cell wall alkali-insoluble -glucan were determined as previously described (15). The alkali-soluble and alkali-insoluble Zymolyase-resistant cell wall fractions were subjected to a dot blot analysis by using anti--1,6-glucan antibody as previously described (33) with standardization by cell wall dry weight. The content of cellular chitin was determined as previously described (10) with Streptomyces griseus chitinase (Sigma, St. Louis, Mo.) and standardization by cell dry weight.

Sequence analysis and homology search. Sequence analysis was performed by using GeneWorks (Intelligenetics, Mountain View, Calif.) and GeneJockey (Biosoft, Cambridge, United Kingdom) software. A homology search for C. glabrata sequences against S. cerevisiae sequences was performed by using the WU-BLAST2 program in the Saccharomyces Genome Database (Stanford University).

Nucleotide sequence accession numbers. The nucleotide sequence data reported in this paper have been submitted to the GenBank database. The accession numbers of the C. glabrata KRE9 (CgKRE9) and KNH1 (CgKNH1) genes are AF064251 and AF064252, respectively.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Construction of tetracycline-sensitive mutants of the S. cerevisiae KRE9 gene. To isolate the S. cerevisiae KRE9 homolog from C. glabrata, we performed complementation screening. As convenient hosts for the screening, tetracycline-sensitive mutants of the S. cerevisiae KRE9 gene (Tet^s KRE9) were constructed. The KRE9 promoter region was replaced with a tetracycline-responsive promoter in a strain, FAHAP4, harboring the tetR-HAP4AD fusion activator gene for tetracycline-controllable gene expression (39). As shown in Fig. 4, addition of tetracycline (50 µg/ml) inhibited growth of cells of Tet^s KRE9 mutant strain SNB50-1 on glucose medium but not on galactose medium. These observations resemble and are consistent with the finding that an S. cerevisiae kre9 mutant grows extremely slowly on glucose medium while growing somewhat better on galactose medium (15) and suggest that the concentration of tetracycline used in the present study is sufficient to repress the expression of KRE9 driven by the tetracycline-responsive promoter. The tetracycline sensitivity of the Tet^s KRE9 mutant was complemented by introduction of an extragenic copy of KRE9 on pRS316 (6) (data not shown). Disruption of KNH1 in a Tet^s KRE9 mutant rendered cell growth tetracycline sensitive on glucose or galactose media (Fig. 4). This result is consistent with the known synthetic lethality between kre9 and knh1 mutations in S. cerevisiae (15). This Tet^s KRE9 knh1 mutant strain, SNB54-5, was used for complementation cloning of a C. glabrata homolog(s).

View larger version (21K): [in this window] [in a new window]   FIG. 4.   Growth of S. cerevisiae Tets KRE9 knh1 cells harboring either pSB2-1 or pSBG9-1. About 104 cells were inoculated and cultured on YNB-glucose for 20 h (solid bars) or on YNB-galactose for 40 h (open bars) at 30°C. Cells were grown with or without tetracycline (50 µg/ml), and growth on tetracycline is expressed as the percentage of optical density at 600 nm of cells grown without tetracycline. The strain FAHAP4 (Table 1) was used as the wild type (WT). Error bars, standard deviations.

Cloning of C. glabrata KRE9 and KNH1 genes. By genomic Southern hybridization using the S. cerevisiae KRE9 sequence as a probe, 5- and 6-kbp EcoRI fragments of C. glabrata genomic DNA were shown to contain sequences hybridizing to S. cerevisiae KRE9 (data not shown). This result allowed us to make a subgenomic C. glabrata bank harboring EcoRI fragments ranging from 4 to 7 kbp to assist in their cloning by functional complementation.

KRE9 knh1

KRE9 knh1

KRE9 knh1

View larger version (48K): [in this window] [in a new window]   FIG. 5.   Sequence comparisons of CgKre9p and CgKnh1p with their S. cerevisiae counterparts. (A) Alignment of the putative Kre9p and Knh1p amino acid sequences deduced from the C. glabrata (CgKRE9 and CgKNH1) and S. cerevisiae (KRE9 and KNH1) nucleotide sequences. The residues with conserved identity in all proteins are underlined in the consensus sequence. The putative N-terminal signals for secretion are underlined in each protein. Gaps (shown as dashes) were introduced to improve alignment. (B) Sequence identities between Kre9p and Knh1p proteins.

KRE9 knh1

Complementation of the killer phenotype of the S. cerevisiae kre9 mutant by CgKRE9 and CgKNH1. Mutations in KRE9 confer resistance to the K1 killer toxin in S. cerevisiae (6, 8). In order to test whether multiple copies of CgKRE9 and CgKNH1 could complement this phenotype, pSB2-1 and pSBG9-1 were transformed into the S. cerevisiae kre9 null mutant strain HAB813 (Table 1) and the killer sensitivities of the transformants were examined by measuring zones of killing in a seeded-plate assay (8). The kre9 mutant cells are known to show no killer zone in the assay, since the mutant has an 80% reduction of -1,6-glucan, which is necessary for the toxin binding. As shown in Table 2, cells harboring pSB2-1 formed killer zones when grown on glucose or galactose plates while cells harboring pSBG9-1 did so only when grown on galactose plates. The killer zone sizes, however, were smaller than those of wild-type strain SEY6210 cells, suggesting that the complementation was partial. We also examined complementation activity of either CgKRE9 or CgKNH1 on a single-copy plasmid as assayed via the killer resistance. Cells harboring CgKRE9-pRS416 formed killer zones in the seeding assay on glucose or galactose plates to the same extent as those harboring multiple copies of CgKRE9 (Table 2), whereas cells harboring CgKNH1-pRS416 failed to form killer zones (data not shown). To show that the partial complementation of the killer phenotype of kre9 mutant was due to restoration of -1,6-glucan levels, alkali-insoluble -1,6-glucan levels in the mutant cells harboring either pSB2-1 or pSBG9-1 were determined. As shown in Table 2, although cells harboring pSBG9-1 showed no restoration, in cells harboring pSB2-1, the alkali-insoluble -1,6-glucan level was partially elevated over that of the mutant when the cells were grown on glucose medium.

View this table: [in this window] [in a new window]   TABLE 2.   Killer phenotypes of alkali-insoluble -glucan levels of the S. cerevisiae kre9 cells harboring either CgKRE9 or CgKNH1d

Disruption of CgKRE9 and CgKNH1 genes and construction of tetracycline-sensitive mutants of CgKRE9 (Tet^s CgKRE9). To explore the physiological essentialness of CgKRE9 and CgKNH1, each gene was disrupted with the C. glabrata TRP1 (CgTRP1) and HIS3 (CgHIS3) genes, respectively (Fig. 1). Transformation for disruption of CgKRE9 was performed on plates containing either glucose or galactose as a carbon source. cgkre9 mutants were obtained from only galactose plates, whereas cgknh1 mutants were obtained from glucose plates. This carbon source dependency on the growth of cgkre9 mutant was confirmed by spotting cells precultured on galactose medium onto plates containing either 2% galactose, 2% glucose, or 2% glucose and galactose. Although cgknh1 cells on all plates and cgkre9 cells on the galactose containing plate grew as well as wild-type cells, the growth of cgkre9 cells was severely impaired on plates containing glucose as a carbon source (data not shown). These results suggest that the presence of glucose is involved in the slow-growth phenotype of the cgkre9 mutant. As shown in Fig. 1D, microscopic examination of cgkre9 cells transferred from galactose to glucose medium revealed the presence of aggregates of cells with abnormal morphology, which are also observed in the S. cerevisiae kre9 null mutant (6). However, cgknh1 cells showed no morphological change compared to the wild type (Fig. 1C and E).

cgkre9

View larger version (17K): [in this window] [in a new window]   FIG. 6.   Growth of S. cerevisiae Tets KRE9 knh1 cells harboring a single copy of either CgKRE9 or CgKNH1. About 104 cells were inoculated and cultured on YNB-glucose for 20 h (solid bars) or on YNB-galactose for 40 h (open bars) at 30°C. Growth of cells with tetracycline (50 µg/ml) is expressed as percent of optical density at 600 nm of cells without tetracycline. FAHAP4 (Table 1) was used as the wild-type. Error bars, standard deviations.

Killer phenotypes and -1,6-glucan levels of cgkre9 and cgknh1 mutants. Although cgkre9 cells showed severe growth defects on glucose medium, spontaneous second-site suppressor mutations partially restoring growth arose when the cells were cultured by serial passage on glucose medium. Since it is known in S. cerevisiae that those second-site suppressors have no effects on the killer phenotypes except for enhanced growth of the original mutants (4, 8, 34, 48), we used such growth-suppressed cgkre9 mutants for further analysis as described below.

cgkre9

cgknh1

cgkre9

cgknh1

cgkre9

cgkre9

cgkre9

cgkre9

cgkre9

Sensitivity to CFW and cellular chitin levels in cgkre9 and cgknh1 mutants. CFW, a negatively charged fluorescent dye that preferentially binds to nascent chains of chitin and interferes with cell wall assembly (16, 50), is a useful compound for surveying a broad range of cell wall defects in S. cerevisiae (32, 46). To test for cell wall defects in cgkre9 and cgknh1 mutants, CFW sensitivities of both growth-suppressed cgkre9 and cgknh1 cells were determined by a spotting assay (31) on plates containing glucose or galactose as a carbon source. Although cgknh1 cells grew as well as wild-type cells even in the presence of 25-µg/ml CFW, growth-suppressed cgkre9 failed to grow at this concentration of CFW when glucose was used as a carbon source (Table 3).

kre9

cgkre9

cgkre9

cgknh1

cgkre9

cgkre9

cgkre9

cgkre9

View larger version (15K): [in this window] [in a new window]   FIG. 7.   Cellular chitin levels in cgkre9 mutants of C. glabrata. (A) Effects of addition of tetracycline on cellular chitin levels in Tets CgKRE9 mutants. About 106 cells were cultured on YPD with (solid bars) or without (open bars) tetracycline (50 µg/ml) at 30°C for 20 h, and the cellular chitin levels were measured. As the wild type (WT), strain ACG22 (Table 1) was used. (B) Effect of switching the carbon source on cellular chitin levels in cgkre9 mutant. Cells precultured on YPGal were inoculated onto either YPD (solid bars) or YPGal (hatched bars) and cultured at 30°C for 20 h, and the cellular chitin levels were measured. As the wild type (WT), strain 2001HTU (Table 1) was used. Error bars, standard deviations.

Overexpression of CgKNH1 and S. cerevisiae KRE9 in cgkre9 cells. We asked if multiple copies of either CgKNH1 or S. cerevisiae KRE9 could complement the phenotypes of a cgkre9 mutant. CgKNH1 was cloned into pRS316 (56), which is known to be a multicopy plasmid for C. glabrata (60). CgKNH1-pRS316 and KRE9-pRS316 (6) were transformed into growth-suppressed cgkre9 cells. As summarized in Table 4, the killer sensitivities and -1,6-glucan levels of the mutant cells were partially restored by multiple copies of S. cerevisiae KRE9 whereas multiple copies of CgKNH1 showed no effect. Further, multiple copies of either CgKNH1 or S. cerevisiae KRE9 allowed growth-suppressed cgkre9 cells to grow as well as wild-type cells on plates containing glucose and CFW (25 µg/ml). In the cells harboring CgKNH1-pRS316, the chitin increase was slightly suppressed (Table 4).

View this table: [in this window] [in a new window]   TABLE 4.   Effects of multiple copies of either CgKNH1 or S. cerevisiae KRE9 on the phenotypes of growth-suppressed cgkre9 cellsf

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The CgKRE9 and CgKNH1 genes have been identified by functional screening using an S. cerevisiae Tet^s KRE9 knh1 mutant. Both C. glabrata gene products have significant overall identity with their S. cerevisiae counterparts (Fig. 5B). Partial restoration of the killer sensitivity and -1,6-glucan levels of kre9 mutant cells harboring multiple copies of CgKRE9 (Table 2) clearly indicates that CgKRE9 is an ortholog of S. cerevisiae KRE9. Furthermore, a single copy of CgKRE9 was sufficient to partially complement the killer phenotype of the kre9 mutant (Table 2). This result also supports the argument for the functional similarity between Kre9p and CgKre9p and implies that the promoter activity of CgKRE9 and the N-terminal signal for secretion of CgKre9p are active in S. cerevisiae.

Disruption of CgKRE9 resulted in cells with phenotypes similar to that of the S. cerevisiae kre9 null mutant (6): a severe growth defect on glucose medium, resistance to the K1 killer toxin, a reduction of -1,6-glucan, and the presence of aggregates of cells with abnormal morphology on glucose medium (Table 3; Fig. 1D). Some of these phenotypes were partially complemented by multiple copies of S. cerevisiae KRE9 (Table 4). Recent cloning of the C. albicans KRE9 (CaKRE9) gene has demonstrated that CaKre9p is also required for -1,6-glucan synthesis in C. albicans (33). These lines of evidence indicate that the function of Kre9p as an essential component for -1,6-glucan biosynthesis is conserved at least among S. cerevisiae, C. albicans, and C. glabrata.

cgknh1 mutants, however, had no phenotype beyond a slightly increased sensitivity to the K1 killer toxin. Further, multiple copies of CgKNH1 failed to restore the killer sensitivity and alkali-insoluble -1,6-glucan levels in cgkre9 cells grown on glucose medium (Table 4). However, in addition to the synthetic lethality suggested by the tetracycline sensitivity of Tet^s CgKRE9 cgknh1 mutant (Fig. 6B), its ability to complement a range of kre9 defects in S. cerevisiae and C. glabrata implies that CgKnh1p is related to Kre9p/CgKre9p and is an ortholog of S. cerevisiae Knh1p. These complementation abilities include S. cerevisiae kre9 mutant phenotypes (Fig. 4 and Table 2), CFW sensitivity, and chitin increase of growth-suppressed cgkre9 cells (Table 4).

We have demonstrated that cellular chitin levels were significantly increased in cgkre9 mutants on glucose medium (Table 3 and Fig. 7). It is known that chitin levels are also increased in several cell wall mutants of S. cerevisiae such as gas1, fks1, and knr4 mutants (22, 27, 45, 47). Based on genetic interaction between gas1 and chs3 mutations and the sensitivity to nikkomycin Z (a competitive inhibitor of chitin synthases) of a gas1 mutant, it has been hypothesized that such a chitin increase is essential for growth as a compensation mechanism to support the impaired cell wall integrity of these mutants (27, 45, 47). However, the increase of chitin in cgkre9 cannot simply be concluded to be the result of such a compensation mechanism, since it is correlated with a severe growth defect on glucose medium and is independent of the reduction of -1,6-glucan. This idea that increased chitin levels slow the growth of cgkre9 mutants is supported by several observations in the present study. First, considerable amounts of cellular chitin were detected in both tetracycline-treated Tet^s CgKRE9 cells grown on glucose medium (Fig. 7A) and cgkre9 cells transferred from galactose to glucose medium (Fig. 7B). Second, there was no obvious increase in chitin levels in cgkre9 cells grown on galactose medium (Table 3 and Fig. 7B), on which they grew as well as the wild type did, in spite of a 50% reduction of alkali-insoluble -1,6-glucan (Tables 3 and 4).

The mechanism and physiological relevance of the chitin increase in cgkre9 mutants and its apparent glucose dependence remain to be elucidated. In S. cerevisiae, at least five genes have been known to be involved in the chitin synthase activity (11, 14). Cloning of these homologs and an enzymatic analysis of chitin synthesis in C. glabrata will be helpful in addressing this question. It will be useful to see if a chitin increase is common to S. cerevisiae kre9 and other kre mutants, since second-site mutations suppressing growth defects have been isolated in many kre mutants and act without restoration of killer sensitivity or -1,6-glucan levels (4, 8, 34, 48). Glucose-specific cross-linking changes in the cell wall of cgkre9 cells may result in elevated chitin levels and a severe growth defect on glucose medium.

Extensive sequencing of regions around both the CgKRE9 and CgKNH1 loci show that genomic organization in the 3' regions of both homologs is conserved between C. glabrata and S. cerevisiae (Fig. 3). This synteny in regions of two chromosomes further indicates a close evolutionary relationship between C. glabrata and S. cerevisiae, consistent with the phylogenetic trees deduced from comparison of 5S (2) and 18S (43) rRNA genes. Further, CgKre9p and CgKnh1p have lower overall identity between themselves than to their orthologous S. cerevisiae counterparts (Fig. 5B). This observation implies that the duplication of the KRE9 and KNH1 genes took place before the divergence of these two fungi from a common ancestor. In contrast, no chromosomal conservation between S. cerevisiae and C. albicans was found in the 8-kbp fragment containing the CaKRE9 locus (data not shown). This result supports the idea of a more distant relationship of C. albicans and S. cerevisiae based on phylogenetic trees deduced from the distribution of the serine-tRNA gene (42, 43) and comparison of rRNA genes (2, 43). Although the presence of a KNH1 homolog in C. albicans still remains a possibility, this result suggests that extensive genomic reorganization around the CaKRE9 locus has occurred since its divergence from a common ancestor with S. cerevisiae. For example, it is possible that the duplication event leading to the KRE9 and KNH1 pair in S. cerevisiae and C. glabrata occurred after the divergence of these yeast lineages from that of C. albicans.

In summary, although the molecular functions of the Kre9p/Knh1p proteins still remain to be characterized, the evolutionary conservation of the essentiality of these proteins supports the idea that compounds that interfere with their functions would be new antifungal drugs affecting a broad spectrum of pathogenic fungi. Our data also indicate that C. glabrata is a useful model pathogenic fungus for understanding biological processes, including cell wall biosynthesis.