INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

Cryptococcus neoformans is an important fungal pathogen that produces deep-seated infections in immunocompromised patients. After inhalation of spores into the lung, infection spreads via the bloodstream to the brain where it causes life-threatening infection of the central nervous system. With the advent of AIDS, this infection has become more prevalent. Even after successful treatment of these patients, life-long suppressive therapy with fluconazole is required (24). Of concern is the significant number of treatment failures observed after long-term therapy and attributed to fluconazole-resistant strains (28, 53). In addition, two cases of azole-resistant cryptococcal infection that were contracted during long-term fluconazole prophylaxis therapy have recently been reported (4). Multiple modes of azole resistance have been reported, including direct effects on the target enzyme 14-demethylase as well as mechanisms resulting in decreased cellular concentration of drug (28, 58). In some cases, azole resistance is concomitant with amphotericin B resistance (23, 58). Thus, the search for new, safe, and effective therapies continues.

The therapeutic challenge in developing agents against eukaryotic pathogens is identification of enzymes or processes that are essential to the pathogens but are absent or sufficiently different in the human host that mechanism-based toxicity is unlikely. In this context, cell wall synthesis is an attractive fungal specific target because the continuous synthesis of cell wall glucan and chitin is required for viability and these processes lack homologous counterparts in mammalian cells (8, 18, 19, 49). Thus, cell wall inhibitors pose low risk of mechanism-based toxicity. Furthermore, the synthesis and assembly of cell wall components occur vectorially through the plasma membrane and into the periplasmic space (6). Thus, inhibitors of cell wall synthesis may not need to penetrate the cell to exert their effect.

The glucan component of the fungal cell wall has been well characterized in Saccharomyces and Candida. In these organisms, glucan accounts for 30 to 60% of the cell wall and is comprised of predominantly 1,3--D-glucan with a few percent of 1,6- linkages (17). Echinocandin lipopeptides are a class of compounds known to inhibit 1,3--D-glucan synthesis (8). Recent clinical success in treating Candida infections with the lipopeptide MK-0991 (50), a potent inhibitor of 1,3--D-glucan synthase (GS), makes this enzyme an especially attractive target for safe, broad-spectrum antifungal drugs. The spectra of MK-0991 and other semisynthetic GS inhibitors are broad, including a wide range of Candida spp., Aspergillus spp. (2, 43), and other less common pathogenic molds (9). Unfortunately, C. neoformans is notably less susceptible to all known GS inhibitors (1, 22, 27). Understanding the reasons for the poor susceptibility of C. neoformans strains to GS inhibitors is important in order to assess the potential of glucan synthesis as an anticryptococcal target.

We have gained insight into the function and structure of GS in fungi through studies of Saccharomyces cerevisiae, and we now know that many of the structural features of the enzyme from the model organism are also found in the C. albicans and Aspergillus enzymes (11, 25, 34). From studies in Saccharomyces, we know that GS is a heteromeric enzyme complex comprising of at least one large, 215-kDa integral membrane protein (12, 32) and one small subunit, Rho1p, which is more loosely associated with the membrane (15, 31, 45). Fks1p is the proposed catalytic subunit, and Rho1p is a small GTP-binding protein that regulates the activity of GS. Rho1p has several regulatory roles in the S. cerevisiae cell cycle (26, 37, 44, 61). Recent evidence derived from echinocandin cross-linking studies in Candida suggest that additional components might also be involved (46).

In Saccharomyces, two highly homologous genes encode the large integral membrane subunit (FKS1/GSC1/ETG1/CND1/CWH53/PBR1/YLR342W and FKS2/GSC2/G4074/YGRO32W). The enzyme complex containing Fks1p constitutes the majority of the activity found in vegetatively growing cells in S. cerevisiae (12, 32). Lipopeptides affect the function of the Fksp component from either FKS gene (13, 32). The FKS2 gene product is needed for sporulation (32). Single disruptions of either FKS1 or FKS2 are viable; however, the double disruption is a lethal event, suggesting the two genes are at least partially functionally redundant (32). A third FKS homolog in Saccharomyces, discovered as a result of the genome sequencing effort, is dispensable (47). In Candida albicans, three genes (FKS1/GSC1, GSL1, and GSL2) are related by sequence homology to Saccharomyces and Aspergillus FKS genes (34). However, expression of mRNA was demonstrated only for C. albicans FKS1 and GSL1. Evidence that one of the Candida genes, FKS1, is indispensable suggests that this gene may predominate in Candida and implies that GSL1 is either dispensable or possesses a different function (11). In Aspergillus fumigatus, only a single genomic band was observed to cross-hybridize to a S. cerevisiae FKS1 probe (12), and thus far, only a single glucan synthase gene has been cloned from either A. fumigatus (GenBank accession no. U79728) or Aspergillus nidulans (25).

Since the expression of genes encoding subunits of glucan synthase has been shown to be essential in both S. cerevisiae (32) and C. albicans (11), the lack of susceptibility of C. neoformans to inhibitors of glucan synthase may be due to differences in the cell wall structure of this species that render glucan nonessential. In particular, staining with the 1,3--D-glucan-specific fluorochrome aniline blue suggests that the 1,3--D-glucan component in Cryptococcus cell walls is much less prevalent than in Candida (36). On the other hand, the fact that production of C. neoformans protoplasts requires 1,3--D-glucanase argues for the importance of glucan to the stability of the Cryptococcus cell wall (3). Alternatively, the Cryptococcus glucan synthase may be sufficiently divergent from the other fungal enzymes that it lacks sensitivity to agents that affect the enzyme from other fungal species, or the polysaccharide capsule of Cryptococcus could impede the penetration of GS inhibitors.

To begin to explore the potential for C. neoformans Fks protein in susceptibility to GS inhibitors and cell viability, we have cloned and sequenced the FKS1 homolog from strain H99 and attempted to disrupt the gene. Because C. neoformans does not propagate as a diploid, demonstration of essentiality by sporulation of a heterozygous disruption strain, as commonly performed in S. cerevisiae, is not possible in C. neoformans. We therefore devised a strategy using homologous integrative transformation to address the issue of FKS1 essentiality in C. neoformans. Any strategy which attempts to disrupt a potentially essential gene must rely on an alternative mechanism for viability, such as a plasmidborne copy or failure to recover the marked disruption in a sporulated diploid, methods used extensively for S. cerevisiae. For less developed molecular genetic systems, such as C. albicans, statistical arguments based on exclusive recovery of nondisrupting integrations are used to strongly suggest essentiality (11, 40, 42). In C. neoformans, the statistical approach is further complicated by the fact that the majority of stable integrative transformants are ectopic (16, 55, 56). Thus, a practical requirement of our strategy is a facile method to screen a very large number of transformants for integration at the FKS1 locus. Our results suggest that the FKS1 gene is essential for growth of C. neoformans and it is likely that GS is an essential enzyme. The strategy we have developed to assess FKS1 essentiality should be generally applicable to the study of other genes in C. neoformans.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Strains. C. neoformans MY2062 is an acapsular mutant of a clinical isolate and was obtained from T. Kozel. C. neoformans H99 is a clinical serotype A isolate (41), and M001 is an adenine auxotroph derived from H99 by UV mutagenesis (55). Strains were grown in synthetic complete medium minus adenine (SC-Ade) or yeast peptone dextrose adenine medium (YPAD) (51).

Genomic DNA isolations. C. neoformans genomic DNA for hybridization analyses was prepared by the DNA isolation protocol of Varma and Kwong Chung (57). This protocol involves digestion of the cell wall with mureinase followed by detergent lysis, phenol-chloroform extraction, and ethanol precipitation.

DNA gel electrophoresis and hybridization. DNA was restriction digested in 0.5- to 1× KGB buffer (100 mM potassium glutamate, 10 mM Tris acetate [pH 7.6], 5 mM MgSO₄) (33) and separated on 1% agarose gels using Tris acetate-EDTA (TAE) buffer (40 mM Tris acetate [pH 8.3], 1 mM EDTA). Gels were capillary blotted to charged nylon membranes (Amersham Hybond N+) after denaturation for 15 min in 0.5 N NaOH and neutralization for 15 min in 20× SSPE (1× SSPE is 0.18 M NaCl, 10 mM NaPO₄, and 1 mM EDTA [pH 7.7]) (30). Prehybridization was performed in a solution containing 50% deionized formamide, 5× SSPE, 10× Denhardt's solution (30), 100 µg of denatured salmon testes DNA per ml, and 0.4% SDS. Hybridization conditions were identical except the Denhardt's solution was reduced to 1/10 that used during prehybridization. Hybridizations were performed at 42°C overnight, and blots were washed twice with 2× SSPE-0.1% SDS at room temperature for low-stringency conditions with additional washing in 0.2× SSPE-0.1% SDS at 42°C for moderate stringency. In some cases, hybridization was performed without formamide at 65°C. Radiolabeled probes were prepared by a random oligonucleotide-labeling procedure (Random Primer DNA Labeling System [GIBCO BRL] or PrimeIt [Stratagene]).

Cloning of C. neoformans homolog to S. cerevisiae FKS1. Based on Southern blot analysis of MY2062 genomic DNA with a S. cerevisiae FKS1 coding region probe (data not shown), a 4-kb BamHI-PstI fragment of C. neoformans DNA was targeted for cloning. Genomic DNA digested with these enzymes was separated on an agarose gel, the fragments with a size of ca. 4 kb were purified (Gene Clean [Bio 101]), and a size-selected library was constructed in plasmid pGEM3zf (Promega). Colony replicas (20) were hybridization screened (~3,500 colonies) with S. cerevisiae FKS1 DNA consisting of a 4.1-kb HindIII-PvuII fragment from pFF133 (12). Positive clones were colony purified, and the plasmids they contained were subjected to dideoxy DNA sequencing (Sequenase [Amersham]) across the PstI vector-insert junction with the SP6 primer (Promega).

C. neoformans FKS1 gene sequencing and sequence analysis. The entire putative C. neoformans FKS1 gene was sequenced on both strands from plasmid pCG2, pCG3, and several subclones. DNA sequencing was performed on an ABI model 370 sequencer, and sequence data were assembled using Sequencer (Gene Codes Corporation). For analysis of the C. neoformans Fks1p open reading frame (ORF), we used the program GAP (Wisconsin Package version 9.1; Genetics Computer Group, Madison, Wis.) to perform pairwise alignments and find the reading frame with strong identity to the large predicted cytoplasmic domain of S. cerevisiae Fks1p. A search for frameshifts or segments of poor homology as we extended this ORF suggested possible introns, and we located potential splice sites using the consensus sequences proposed from other C. neoformans genes (59). To identify exons at either end of the C. neoformans FKS1 ORF, where the predicted protein identity among Fksps is the weakest, we used two approaches. First, PCR primers designed from the genomic sequence were used to amplify fragments (using Taq polymerase from Gibco-BRL and the manufacturer's reaction conditions) from a cDNA library constructed from C. neoformans B-3501 (Clonetech). Second, rapid amplification of cDNA ends (RACE) was performed using mRNA prepared from C. neoformans H99 (mRNA Isolation Kit; Boehringer Mannheim) and the Marathon cDNA amplification kit from Clontech. The PCR products from both approaches were cloned into vector pCR2.1 (Invitrogen) and sequenced. For RNase protection assays of the 5' end of C. neoformans FKS1, a PCR-amplified DNA probe with a T7 promoter element added to the 5' end was transcribed in vitro by using the Maxiscript system from Ambion and [³²P]UTP from Amersham. The resulting radiolabeled RNA probe was hybridized to total RNA isolated from log-phase cells of C. neoformans H99, and the RNA-RNA hybrid was subjected to single-strand RNase digestion, separated on a 6% acrylamide gel containing 8 M urea, and visualized by autoradiography.

Integrative disruption plasmid construction. For the PCR screening strategy (described below), we required PCR priming sites near the genomic FKS1 gene that are not contained in the plasmid sequence. To accomplish this, the ends of the 6-kb PstI fragment of pCG2 were sequenced by using the pUC18 forward and reverse primers, and a fragment slightly smaller than the 6-kb PstI fragment was subcloned by PCR, creating pCG4. The following primers were used to clone the fragment in pCG4: 5'-AAGGAAAAAAGCGGCCGCCAGCCAAATAGTTTCTTCTCTGCC-3' and 5'-AAAGGAAAAAAGCGGCCGCCACACTGGTGATATTCGGAATGC-3'. Each primer consists of a 10- or 11-bp spacer, a NotI site, and 24 or 23 nucleotides (nt) of C. neoformans FKS1 sequence. The primers were used in conjunction with Expand polymerase (Boehringer Mannheim) and buffer conditions recommended by the manufacturer to amplify the fragment from pCG2 for subcloning. The amplification program was 120 s at 92°C followed by 25 cycles, with each cycle consisting of 10 s at 92°C, 30 s at 50°C, and 18 min at 68°C. The PCR product was digested with NotI enzyme (GIBCO-BRL), purified from 1% agarose-1× TAE gels using Gene Clean (Bio 101), and ligated into NotI-digested pGEM5zf (Promega).

Biolistic transformation. Strain M001 was transformed with plasmid pWL42 by the biolistic transformation procedure (55), and adenine prototrophs were selected by plating transformations on SC-Ade medium and incubating at 30°C for several days.

PCR screening for targeted integrants. The primers employed to identify targeted integrants and reveal their orientation are listed in Table 1. All PCR primers were purchased from GIBCO-BRL. Primer pair A5'-A3' yields a 3,371-bp product (band A) and primer pair B5'-B3' yields a 2,046-bp product (band B) when wild-type C. neoformans H99 genomic DNA is used as the template. Targeted integration of plasmid pWL42 into the C. neoformans FKS1 locus is predicted to eliminate one of the two PCR products (Fig. 1). For initial screening, all four primers were used in a single PCR and the products were analyzed by agarose gel electrophoresis. Putative targeted integrants, with only a single PCR product, were verified in separate reactions with primer pairs A5'-A3' and B5'-B3'. For PCR analysis, 50-µl reaction mixtures were prepared including 2 to 4 µl of miniprep genomic DNA template, 50 pmol of each primer, 2.5 U of Taq polymerase (GIBCO-BRL), 300 µM each of the four deoxynucleoside triphosphates, 1.5 mM MgCl₂, and the manufacturer's buffer. PCR cycle conditions were 25 cycles, with each cycle consisting of 50 s at 94°C, 50 s at 55°C, and 210 s at 72°C. After the last cycle, reaction mixtures were incubated for an additional 5 min at 72°C and held at 4°C until loaded on a 0.8% agarose-1× TAE gel for analysis.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Integration of pWL42 through homologous recombination at the C. neoformans FKS1 (CnFKS1) locus. Two potential homologous integration events are depicted. In both panels, plasmid pWL42 is shown as a circle with the ADE2 gene (stippled box) inserted within a cloned fragment of the CnFKS1 locus (cross-hatched boxes). The thicker cross-hatched regions on the linear maps correspond to the CnFKS1 DNA segments present on the plasmid. Thick lines with arrowheads or vertical bars at the ends show the direction of transcription and the approximate size of the CnFKS1 reading frame. The vertical bars indicate the position of 5' or 3' truncation, and broken lines indicate promoterless reading frames. The ADE2 insertion is transcribed in the direction opposite that of CnFKS1. In panel A, pWL42 is shown integrating at the CnFKS1 locus through a single crossover event (×) within the cloned 5' region upstream of the CnFKS1 ORF. The integrated plasmid (bracket) creates an interrupted copy (CnFKS1 5', 3' ) of CnFKS1, but an intact copy of the ORF remains downstream of the site of integration. The striped box above the intact ORF illustrates the approximate size and position of a PCR product obtained from A-type integration, wild-type, or ectopic integration but not B-type integration. In panel B, disrupting integration of pWL42 is depicted, in which copies of CnFKS1 truncated at either the 3' end (CnFKS1 3' ) or the 5' end (CnFKS1 5' ) are created. The striped box illustrates a PCR product specific to disrupting and wild-type or ectopic pWL42 integrants but not A-type integration.

Glucan synthase enzyme assays. Glucan synthase enzyme assays were performed with a crude microsomal preparation. Enzyme activity was quantitated by measuring the incorporation of UDP-glucose (UDPG) into a trichloroacetic acid (TCA)-insoluble fraction. Cells grown from 0.01 to 2 optical density units at 600 nm at 30°C in YPAD medium were harvested by centrifugation and washed with 50 mM phosphate buffer (pH 7.0) containing 25 mM KF, 1 mM -mercaptoethanol, and 1 mM EDTA. The pellet was suspended in the same buffer containing 10 µM [-S]GTP and protease inhibitors (1 mM Pefabloc and the complete protease inhibitor mixture from Boehringer); cells were broken, crude membranes were prepared, and protein concentration was determined as described previously (13). GS activity was measured in reaction mixtures containing 100 µg of crude membrane protein, 50 mM Tris (pH 8.5), 10% glycerol, 25 mM KF, 0.25% (wt/vol) bovine serum albumin, 20 µM [-S]GTP, 18 U of -amylase (Sigma), UDP-[³H]glucose (2.2 Ci/mmol; Amersham), and 1.5 mM unlabeled UDPG (60). After 2 h of incubation at 30°C with gentle agitation, reactions were terminated with ice-cold TCA, and the radiolabeled product was collected and quantitated as described previously (13). To measure the glucanase susceptibility of reaction products, aliquots of replicate GS reaction mixtures were removed following synthesis and incubated with or without 6 mg of laminarinase (Sigma) per ml in 50 mM acetate buffer (pH 5.3) for 2 h at 37°C. The TCA-insoluble material was collected, washed, and counted (13).

RNA analysis. For total cellular RNA isolation, an RNeasy kit (Qiagen) was employed. Cells grown in 50 ml of YPAD at 30°C to mid-log phase were harvested by centrifugation (3,000 × g, 5°C). All subsequent operations were performed on ice. Cells were washed with 1 ml of cold TE plus 0.1% diethylpyrocarbonate (DEPC) and transferred to a microcentrifuge tube. Cell pellets were suspended in 350 µl of Qiagen RLT buffer with -mercaptoethanol (Qiagen) and 0.1% DEPC. An equal volume of acid-washed 0.45-µm-diameter glass beads was added. Cells were disrupted for 2 cycles of 30 seconds each in a minibead beater (Biospec Products) with 30 s on ice between disruption cycles. The lysate was transferred to a fresh tube, the glass beads were washed with an additional 350 µl of RLT buffer, and the wash was combined with the lysate. The pooled lysate was subsequently treated according to the manufacturer's protocol except fresh 0.1% DEPC was added to all buffers. Final RNA pellets were dissolved in 50 µl of TE plus 0.1% DEPC.

Nucleotide sequence accession number. The genomic sequence of the C. neoformans FKS1 gene has been submitted to GenBank (accession no. AF102882).

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Cryptococcus glucan synthase cloning. A Saccharomyces FKS1 probe encompassing a portion of the coding region had previously been shown to hybridize to genomic DNA from C. neoformans MY2062 (12) and thus provided a strategy to clone the putative C. neoformans FKS1 homolog. Based on additional hybridization analysis with the Saccharomyces FKS1 gene probe (data not shown), a 4-kb BamHI-PstI C. neoformans genomic fragment from strain MY2062 was targeted for cloning (see Materials and Methods). Six putative clones were isolated, the ends of each clone were sequenced, and the translated protein sequences for five of the six clones (pCG1-2 through pCG1-6) aligned well with the Saccharomyces FKS1 protein sequence. The short stretch of amino acid homology provided the first verification that we had cloned a C. neoformans FKS1 homolog. The location of the homologous region in the Saccharomyces FKS1 protein sequence and the size of the cloned insert suggested that the BamHI-PstI fragment contained DNA representing the amino-terminal half of the C. neoformans FKS1 gene.

C. neoformans FKS1 gene sequence analysis. We determined the sequence of 8,023 bp of genomic DNA from C. neoformans H99 that encompasses the FKS1 promoter, an ORF interrupted by seven introns, and the likely transcription terminator. At the amino and carboxy termini, the C. neoformans Fks1p is shorter than other Fksp family members by roughly 70 and 40 residues, respectively (data not shown). In analyzing the 5' end of C. neoformans FKS1 by RNase protection, 45 nt were digested from a 258-nt probe spanning bases 1805 to 2063, revealing the presence of the first intron (bases 2013 to 2058). Therefore, the predicted protein start site at position 1940 of FKS1 is the furthest upstream AUG codon that is in frame within the exon preceding this first intron. As for the 3' end, a UAA stop codon at position 7468 is within the last exon established by RACE analysis of the cDNA sequence. A hydropathy analysis (52) of the C. neoformans Fks1 protein reveals that, like other members of the Fksp family, this protein has at least 12 transmembrane helices, with an amino terminus and large central hydrophilic region that are predicted to be on the cytoplasmic face of the plasma membrane (data not shown). The highest degree of identity to other Fks proteins is within the central region, and the genomic sequence encoding this portion of C. neoformans Fks1p is not interrupted by introns. Like the other Fks proteins, Fks1p from C. neoformans does not contain the proposed UDP-glucose binding motif QXXRW, but it does have a region with limited homology to BcsAp, the catalytic subunit of cellulose synthase from Acetobacter xylinium (25).

View larger version (19K): [in this window] [in a new window]   FIG. 2.   Glucan synthase parsimony tree. Abbreviations: SpFKSx1, Schizosaccharomyces pombe FKS chromosome 1 (accession no. Z98601); SpFKSx2, S. pombe FKS chromosome 2 (accession no. AL021839); AfFKS, Aspergillus fumigatus FKS (accession no. U79728); AnFKS, Aspergillus nidulans FKS (accession no. U51272); CaFKS1, Candida albicans FKS1 (accession no. D88815); CaGSL1, C. albicans GSL1 (accession no. D88816); CaGSL2, C. albicans GSL2 (accession no. AB001077); ScFKS1, Saccharomyces cerevisiae FKS1 (accession no. S50235); ScFKS2, S. cerevisiae FKS2 (accession no. S50240); ScFKS3, S. cerevisiae FKS3 (accession no. S53976). (All accession numbers are GenBank except for the last three which are PIR.) The numbers in parentheses are the percent identity with C. neoformans Fks1p based on GCG GAP alignments.

Is the putative Cryptococcus FKS1 gene single copy? A genomic Southern blot analysis was performed to determine if the C. neoformans FKS1 gene is single copy. The 4-kb BamHI-PstI fragment from pCG2 was used to probe a Southern blot of H99 genomic DNA individually digested with four restriction enzymes (Fig. 3). Employing low-stringency wash conditions, each of restriction digestions produced only a single hybridizing band. Thus, only a single genomic C. neoformans FKS locus is detectable by low-stringency hybridization.

View larger version (53K): [in this window] [in a new window]   FIG. 3.   Cryptococcus FKS Southern blot. Each lane contains 10 µg of restriction enzyme-digested H99 genomic DNA. The enzymes were BamHI (lane 1), PstI (lane 2), NcoI (lane 3), and SmaI (lane 4). Hybridization probe was 32P-labeled random-primed 4-kb BamHI-PstI H99 genomic fragment derived from pCG2 and encompasses the first 1,221 codons (of 1,643) of the C. neoformans FKS1 sequence. Arrows indicate positions of HindIII-digested lambda standards.

Is the Cryptococcus FKS1 gene essential? An integration plasmid, pWL42, was constructed such that integration by homologous recombination at the FKS1 locus can occur in two possible orientations. This plasmid contains two regions of genomic FKS1 sequence separated by an ADE2 selectable marker. Integration by homologous recombination on one side of the ADE2 marker leaves an intact FKS1 gene, while integration on the other side results in loss of function for the FKS1 gene (Fig. 1). If the gene were nonessential, recovery of integrative transformants in both orientations should be possible. Conversely, if the FKS1 gene encodes an essential function, then only nondisrupting integration events will be recovered. Using this strategy, isolation of a sufficient number of nondisrupting transformants without isolation of disrupting transformants provides a statistical argument that the gene in question is essential for mitotic growth.

View larger version (36K): [in this window] [in a new window]   FIG. 4.   PCR screening profiles. The left panel depicts the PCR pattern expected from wild-type (WT), A-type (A), and B-type (B) integration of pWL42. The right gels show actual screening examples from a wild-type control (WT) and an A-type integration (A). A B-type integration was not obtained. The sizes of the A- and B-specific PCR products are 3,371 and 2,046 bp, respectively.

View larger version (33K): [in this window] [in a new window]   FIG. 5.   Southern blot analysis of nondisrupting integrants. Genomic DNA isolated from two pWL42 transformants (lane 1, strain 42-1-3; lane 2, strain 42-1-28) and from strain H99 (lane 3) was digested with PstI (panels A, B, and C) or XmaI and ClaI (panel D), and analyzed on four separate Southern blots. The probes for the blots were as follows: for panel A, the 1.5-kb NotI-XmaI CnFKS1 fragment from pCG4; for panel B, the 3.0-kb pGEM5zf vector made linear with NotI digestion; for panel C, the 0.9-kb XhoI-XhoI CnFKS1 fragment from pCG2 which represents the region of FKS1 that was replaced by ADE2 in pWL42; and for panel D, the 3.0-kb KpnI-XmaI CnADE2 fragment from pCnADE2Apa. Arrows on the side of panels A to D denote the migration positions of HindIII fragments of lambda (from top to bottom, 23.1, 9.4, 6.6, 4.3, 2.3, and 2.0 kb). Maps predicted for the CnFKS1 locus following ectopic, nondisrupting, or disrupting integration are presented below panels A to D. Black bars at the bottom of the map show the positions of the probe fragments on the disrupted pWL42 integrant map. Features of the maps (ORFs, plasmid DNA, and regions of CnFKS1 contained in pWL42) are as described in the legend to Fig. 1. The brackets above each map indicate the sizes and positions of fragments generated by PstI digestion; fragments produced by digestion with XmaI and ClaI are shown below. A bar indicating scale is shown at the bottom right.

Expression and biochemical analysis of indeterminate targeted integrants. To evaluate the expression status of the three targeted integrants with indeterminate genomic structure, GS enzyme activity in membranes prepared from these strains was measured and the strains were also analyzed for expression of FKS1 mRNA. Membranes from all of the transformants contained active GS enzyme as measured by incorporation of glucose from UDP-glucose into a TCA-insoluble product (Table 2). We also showed that the percentage of glucanase-sensitive product was unchanged, confirming that the TCA-insoluble material was unchanged and consists of predominantly 1,3--glucan (Table 2). Similarly, the Northern blot analysis detected a mRNA band identical in size to the wild-type FKS1 message (Fig. 6).

View larger version (92K): [in this window] [in a new window]   FIG. 6.   CnFKS1 Northern analysis. Lane 1, H99 parent strain; lane 2, 42-1-28; lane 3, 42-2-13; lane 4, 42-1-43; lane 5, 42-1-48; lane 6, 42-6-60. Lane 2 and 3 are A-type, nondisrupting integrations. Lanes 4 to 6 contain strains with putative genomic rearrangements. The hybridization probe was 32P-labeled random-primed 4-kb BamHI-PstI H99 genomic fragment derived from pCG2. The arrow indicates the position of the ~9-kb FKS mRNA band.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

An understanding of why C. neoformans is much less susceptible to the known GS inhibitors should help to evaluate the importance of GS as a broad-spectrum antifungal target and thus help focus future drug discovery efforts. We decided to clone the C. neoformans FKS1 homolog as a means to evaluate whether glucan synthase is required for viability in C. neoformans and to compare the sequence of C. neoformans Fks protein to those of other fungal enzymes. At issue is whether the reduced susceptibility shown by Cryptococcus to GS inhibitors can be attributed to divergence of the Cryptococcus glucan synthase with respect to other pathogenic fungi or whether the phenotype is due to other factors, such as unique features in the cell wall structure of Cryptococcus that might make cells less dependent on the cell wall for viability.

We have isolated a C. neoformans gene by cross-hybridization to S. cerevisiae FKS1. Several lines of evidence argue that this gene is the functional FKS homolog in C. neoformans. The deduced amino acid sequence is 58% identical to both S. cerevisiae FKS1 and C. albicans FKS1, and the overall length, 1,643 amino acids, is similar to those of other fungal glucan synthases. Low-stringency genomic Southern analysis with C. neoformans FKS1 hybridization probe failed to show evidence of additional copies. The hybridization conditions employed in that analysis were identical to the conditions used to detect the cross-hybridization between S. cerevisiae FKS1 and C. neoformans FKS1. Thus, we can conservatively conclude that if additional related sequences are present in C. neoformans, they must bear less than the 58% identity shared by S. cerevisiae FKS1 and C. neoformans FKS1. Finally, the results of Northern hybridization analysis confirm that the cloned C. neoformans FKS1 gene is expressed.

In terms of genomic organization of FKS genes, C. neoformans appears most similar to Aspergillus species, which also appear to encode FKS with a single gene (25). The C. neoformans Fks1 protein sequence also shares the highest sequence identity with the Aspergillus Fks proteins (Fig. 2).

Unequivocal demonstration of gene essentially is routinely performed in S. cerevisiae by introducing a genetically marked disruption into a diploid and showing that haploid segregants carrying the disruption are inviable. While C. neoformans possesses a sexual cycle, the diploid phase cannot be cultured. As a result, other strategies for evaluating gene essentiality must be developed for this organism.

We have devised a strategy to assess essentiality that requires only sufficient sequence information to support construction of an appropriate integration plasmid and define suitable PCR primers to rapidly identify homologous integrants. Our strategy involves construction of a plasmid that can integrate by homologous recombination in two orientations, only one of which will disrupt gene function. This approach involves the introduction of two defects into the gene, a truncation of the coding region and an insertion of a selectable marker in the remaining coding region. Additionally, the construct must contain both 5'- and 3'-flanking sequence such that the resulting plasmid can integrate into the genome by a single crossover event in either the 5'- or 3'-flanking region. Integration in either orientation results in gene duplication. In one orientation, both defects remain on the same copy, thus preserving one intact copy of the gene. However, integration in the other orientation distributes one defect to each of the resulting copies and therefore disrupts the function of both gene copies. Thus demonstration of essentiality derives from the exclusive recovery of integrations in the nondisrupting orientation.

To date, only a few attempts to evaluate gene essentiality in Cryptococcus have been reported. Lodge et al. exploited a temperature-sensitive mutant to assess essentiality of the gene for N-myristoyl transferase (29). They used a plasmid construction that, via homologous integration, resulted in replacement of the wild-type allele with a well-characterized temperature-sensitive (ts) mutation. Subsequent demonstration of loss of viability at the restrictive temperature provided convincing evidence for the essentiality of N-myristoyl transferase. A limitation to this approach, however, is the need to obtain a suitable ts allele before the demonstration of essentiality can be performed. Construction and characterization of such ts alleles generally follows the initial sequencing of a new gene. Therefore, our strategy, which relies solely on sequence information, should be generally useful and applicable at an earlier stage in the analysis of a new gene than the ts strategy employed by Lodge et al.

Recently, we evaluated the essentiality of the topoisomerase I gene by showing that homologous integrants of a disruption plasmid could be recovered only if a second ectopically integrated functional gene copy was present (10). However, this strategy was cumbersome and required a more complicated series of molecular manipulations for proof of the essential feature.

Using our strategy, the argument for essentiality is statistical in nature and depends on isolation of a sufficient number of homologous integrations in the nondisrupting orientation in order to be able to state with confidence that the disrupting orientation should have been recovered if the gene is not essential. We identified 23 homologous recombination events in the nondisrupting orientation, while no integrations in the disrupting orientation were recovered. Assuming an equal chance of recombination in either orientation, the probability that only one orientation would be isolated in 23 attempts is (0.5)²³ or 1.19 × 10⁷. Thus, the statistical argument for essentiality of the FKS1 gene is quite strong. This argument relies on an assumption of roughly equal probability of insertion in either of the two possible orientations. In our integration plasmid construct, we introduced a larger segment of flanking sequence (2,871 bp versus 1,204 bp) on the side in which recombination would result in gene disruption. Thus, if length considerations introduce a bias, it should favor recovery of disrupting integrations.

Three additional homologous integrants at the C. neoformans FKS1 locus that contained a complicated genomic rearrangement at or near the FKS1 locus were identified. Subsequent analysis of mRNA expression in these transformants confirmed that, despite the genomic rearrangements, an apparent full-length mRNA from the FKS1 gene is still expressed. Glucan synthase activity was detected in all three integrants for which FKS1 expression status was in question, and the GS specific activity was comparable to those of the A-type, nondisrupting integrants that were also analyzed. Therefore, the three transformants resulting in genomic rearrangements at the FKS1 locus are not functionally disrupted and do not impact the conclusion that the FKS1 gene is essential. The minor (twofold or less) quantitative variations in the observed GS specific activity are not considered significant.

We have not produced direct evidence that the GS activity measured in vitro is derived from expression of the FKS1 gene. However, the alternative possibility that another gene, unrelated to known glucan synthases, encodes GS in C. neoformans appears unlikely. This would require that the FKS1 gene reported here provide an essential function distinct from GS activity. No such bifunctional character has been attributed to any other fungal Fks proteins, and the deduced protein sequence of the C. neoformans Fks1p aligns reasonably well with the other Fks proteins over nearly its entire length. Given the high level of homology of the C. neoformans FKS1 gene with those of other fungi and its single-copy status in C. neoformans, it seems logical to assume that this gene is providing a subunit required for GS activity. Formally, however, we cannot exclude the possibility that the C. neoformans FKS1 gene might be essential for reasons other than providing a subunit of the glucan synthase complex.

If GS is essential, why are C. neoformans cells so much less susceptible than other fungi to known GS inhibitors? The availability of the C. neoformans FKS1 sequence gives us the opportunity to compare this sequence to the FKS sequences of other fungi for insight into possible differences that might account for altered susceptibility to the known GS inhibitors. The primary amino acid sequence of C. neoformans Fks1p shares 58 and 62% identity, respectively, with the homologous proteins from the other fungal pathogens, C. albicans FKS1 and A. fumigatus FKS. The amino acid identity between C. albicans FKS1 and A. fumigatus FKS is 64%; thus, the degree of Fks protein sequence conservation is similar between any pairwise combination of these three pathogens.

In S. cerevisiae, three echinocandin-resistant mutations have been mapped to single amino acid changes in FKS genes in a region spanning eight amino acids (14). Using multiple sequence alignment, we closely examined the C. neoformans FKS1 sequence in this important region. Table 3 shows the sequence of this region for the known essential Fks proteins from S. cerevisiae and C. albicans, as well as the sole sequences derived from Aspergillus species and C. neoformans. Six of the eight residues in this region are absolutely conserved in all six FKS sequences. The three residues where specific substitutions can result in echinocandin resistance are all conserved in the C. neoformans sequence. The C. neoformans sequence differs from the consensus in this eight-residue region only at position 6 where Phe replaces Leu. A. fumigatus also has Phe at this position. To test the possible significance of this deviation from the consensus, we constructed a Leu-to-Phe substitution at the homologous position in the S. cerevisiae FKS1 sequence. The substitution had no effect on the echinocandin sensitivity of either whole cells or enzyme activity in vitro (data not shown).

We also used the multiple protein sequence alignment of all known Fks proteins to construct the unrooted parsimony tree in Fig. 2. The C. neoformans and Aspergillus proteins occupy a central position in the tree, bracketed by the Schizosaccharomyces pombe proteins on one side and the Saccharomyces and Candida proteins on the other. The susceptibility of Candida and Aspergillus species to a variety of echinocandins is well established (2, 43). S. pombe has been studied less in this regard but is known to be susceptible to the lipopeptide aculeacin (35, 38), as well as papulacandin B (7, 21, 48). Thus, the C. neoformans sequence is located centrally in the parsimony tree, bracketed by organisms that show susceptibility to known GS inhibitors. It is interesting to note that the known functional FKS genes from S. cerevisiae (FKS1 and FKS2) and C. albicans (FKS1) clustered together. FKS1 is indispensable in Candida despite the presence of two other homologs (GSL1 and GSL2). In S. cerevisiae, both FKS1 and FKS2 are functional and the double disruption is lethal. Another cluster consisting of C. albicans GSL1 and S. cerevisiae FKS3 contains proteins that are either nonfunctional with respect to their contribution to GS enzyme activity or at least are not capable of providing functional subunits of the GS complex that can support growth. FKS3 in S. cerevisiae is dispensable as a single disruption and is not synthetically lethal in combination with either FKS1 or FKS2 deletions (47). In C. albicans, since FKS1 alone is indispensable, it seems that GSL1 is not capable of providing GS activity sufficient to support growth and may be nonfunctional or have a different function altogether. This distinct and separate clustering of functional and unknown function Fks homologs lends support to the validity of the parsimony tree.

We have compared the C. neoformans FKS1 sequence with homologs from other fungal species in terms of overall sequence identities, evolutionary relationships as established by parsimony analysis, and detailed examination of the sequence in a region in which known mutations confer echinocandin resistance. None of these analyses support the notion that the C. neoformans GS enzyme might be less susceptible to GS inhibitors because it is different from other fungi. However, differences in the intrinsic sensitivity of C. neoformans GS could be due to changes that we cannot discern by sequence analysis. Supporting this notion, Williamson et al. (60) have recently shown that echinocandins are three orders of magnitude less potent against the Cryptococcus glucan synthase in vitro than the enzyme derived from Candida (5). Conversely, the notion that the capsule might contribute to the relative lack of susceptibility in Cryptococcus appears unlikely. Our acapsular C. neoformans strain, MY2062, shows a susceptibility to the echinocandin L-733,560 (MIC = 16 µg/ml) that is consistent with values determined for numerous capsular clinical isolates (range, 16 to 32 µg/ml) (2). The genetic data described in this work suggest that, despite a lower content of glucan in the Cryptococcus cell wall, glucan synthesis is required for viability in Cryptococcus. Thus, our data reinforce the potential utility of glucan synthesis inhibition as a broad-spectrum antifungal target. Overall, further study of the biochemical differences between fungal glucan synthases appears most likely to provide information that will be useful in developing broad-spectrum antifungal agents targeting glucan synthesis.