Identification of the FKS1 gene of Candida albicans as the essential target of 1,3-beta-D-glucan synthase inhibitors.

Pneumocandins and echinocandins are fungicidal antibiotics, currently in clinical development, that inhibit 1,3-beta-D-glucan synthase (GS) in several human fungal pathogens. We have identified a gene from the diploid organism Candida albicans that encodes a target of these inhibitors. A 2.1-kb portion of this gene, designated CaFKS1, has significant homology to the Saccharomyces cerevisiae FKS1 and FKS2 genes, which encode partially functionally redundant subunits of GS. To evaluate the role of CaFkslp in susceptibility to echinocandins, we disrupted CaFKS1 on one homolog each of the spontaneous pneumocandin-resistant C. albicans mutants CAI4R1, NR2, NR3, and NR4. These mutants had been selected previously on agar plates containing the pneumocandin L-733,560. The clones derived from this transformation were either resistant (Ech[r]) or fully sensitive (Ech[s]) to inhibition by L-733,560 in both liquid broth microdilution and in vitro GS assays. The site of plasmid insertion in the transformants was mapped by Southern blot analysis, using restriction site polymorphisms in the CaFKS1 gene to distinguish between the two alleles (designated CaFKS1h and CaFKS1b). For strains CAI4R1 and NR2, the CaFKS1b allele was disrupted in each Ech(r) transformant; for strain NR4, CaFKS1h was disrupted in each Ech(r) transformant. We conclude that (i) strains CAI4R1, NR2, and NR4 are heterozygous for a dominant or semidominant pneumocandin resistance mutation at CaFKS1, (ii) drug resistance mutations can occur in either CaFKS1 allele, and (iii) CaFks1p is a target of the echinocandins. For transformants of strain NR3, all the clones we analyzed were uniformly Ech(r), and only the CaFKS1h allele, either in disrupted or wild-type form, was detected on genomic Southern blots. We believe gene conversion at the CaFKS1 locus may have produced two Cafks1h alleles that each contain an Ech(r) mutation. Transformants derived from the mutants were analyzed for susceptibility to pneumocandin treatment in a mouse model of disseminated candidiasis. Strains heterozygous for the resistant allele (i.e., C. albicans CAI4R1, NR2, and NR4) were moderately resistant to treatment, while strains without a functional Ech(s) allele (i.e., strain NR3 and derivatives of strain CAI4R1 with the disruption plasmid integrated in the Ech[s] allele) displayed strong in vivo echinocandin resistance. Finally, we were unable to inactivate both alleles at CaFKS1 by two-step integrative disruption, suggesting that CaFks1p is likely to be an essential protein in C. albicans.

Characterization of echinocandin-resistant mutants of Candida albicans: genetic, biochemical, and virulence studies.

The pneumocandins are potent antifungal agents of the echinocandin class which are under development for use as broad-spectrum antimycotic therapy. One important consideration for any new therapeutic class for treating serious fungal infections is the potential for drug resistance development. In this study we have isolated and characterized four independent spontaneous Candida albicans mutants resistant to the potent semisynthetic pneumocandin L-733,560. These mutants have many of the properties of FKS1/ETG1 echinocandin-resistant mutants of Saccharomyces cerevisiae, including (i) cross-resistance to other 1,3-beta-D-glucan synthase inhibitors, such as papulacandin and echinocandins, but no change in sensitivity to other antifungal agents; (ii) in vitro glucan synthase activity that is more resistant to pneumocandins than the wild-type parent enzyme; and (iii) semidominant drug resistance in spheroplast fusion strains. The mutants were compared with C. albicans echinocandin-resistant mutants isolated by mutagenesis by L. Beckford and D. Kerridge (mutant M-2) (abstr. PS3.11, in Proceedings of the XI Congress of the International Society for Human and Animal Mycology, Montreal, Canada, 1992) and by A. Cassone, R. E. Mason, and D. Kerridge (mutant CA-2) (Sabouraudia 19:97-110, 1981). All of the strains had resistant enzyme activity in vitro. M-2 grew poorly and had low levels of enzyme activity. In contrast, CA-2 and the spontaneous mutants grew as well as the parents and had normal levels of glucan synthase activity. These results suggest that these resistant mutants may have alterations in glucan synthase. CA-2 was unable to form germ tubes, an ability retained by the spontaneous mutants. The virulence of the spontaneous mutants was unimpaired in a mouse model of disseminated candidiasis, while M-2 and CA-2 were 2 orders of magnitude less virulent than their parent strains. Significantly, mice challenged with the spontaneous mutant CAI4R1 responded therapeutically to lower levels of L-733,560 than would he predicted by the increase in in vitro susceptibility.

Aerosol and parenteral pneumocandins are effective in a rat model of pulmonary aspergillosis.

The pneumocandins are semisynthetic analogs of echinocandin-like compounds that have shown efficacy in animal models of systemic candidiasis, disseminated aspergillosis, and pneumocystis pneumonia. However, the most common form of Aspergillus infection in susceptible patients is pulmonary aspergillosis, which was not directly tested in the mouse models used in the past. We have evaluated three pneumocandins, L-693,989, L-731,373, and L-733,560, in a rat model of pulmonary aspergillosis. Male Sprague-Dawley rats were treated for 2 weeks with cortisone and tetracycline and fed a low-protein diet before being inoculated via the trachea with 10(6) conidia of Aspergillus fumigatus H11-20. In the absence of drug treatment, the animals developed a progressive, rapidly fatal bronchopneumonia. All three pneumocandins at doses of 5 mg/kg (intraperitoneally [i.p.] every 12 h [q12h]) were effective in delaying mortality in this model. Survival at day 7 postinfection was 20% among controls (n = 10 for all groups), while it was 60, 80, and 90% in groups that were treated with L-693,989, L-731,373, and L-733,560, respectively. In another trial, survival at day 7 postinfection was 25% among controls (n = 8 for all groups); it was 87.5% in a group treated with amphotericin B (0.5 mg/kg i.p. q12h) and was 100% in a group treated with L-733,560 (0.625 mg/kg i.p. q12h). In a separate trial, aerosol L-693,989 administered 2 h before infection (5 mg/kg) delayed mortality. Eight of the 10 animals treated with aerosol L-693,989 survived for 7 days, whereas only 2 of 10 control animals survived. We conclude that the pneumocandins we tested were highly effective in an animal model of pulmonary aspergillosis.

The Saccharomyces cerevisiae FKS1 (ETG1) gene encodes an integral membrane protein which is a subunit of 1,3-beta-D-glucan synthase.

In Saccharomyces cerevisiae, mutations in FKS1 confer hypersensitivity to the immunosuppressants FK506 and cyclosporin A, while mutations in ETG1 confer resistance to the cell-wall-active echinocandins (inhibitors of 1,3-beta-D-glucan synthase) and, in some cases, concomitant hypersensitivity to the chitin synthase inhibitor nikkomycin Z. The FKS1 and ETG1 genes were cloned by complementation of these phenotypes and were found to be identical. Disruption of the gene results in (i) a pronounced slow-growth phenotype, (ii) hypersensitivity to FK506 and cyclosporin A, (iii) a slight increase in sensitivity to echinocandin, and (iv) a significant reduction in 1,3-beta-D-glucan synthase activity in vitro. The nucleotide sequence encodes a 215-kDa polypeptide predicted to be an integral membrane protein with 16 transmembrane helices, consistent with previous observations that the etg1-1 mutation results in echinocandin-resistant glucan synthase activity associated with the nonextractable membrane fraction of the enzyme. These results suggest that FKS1 encodes a subunit of 1,3-beta-D-glucan synthase. The residual activity present in the disruption mutant, the nonessential nature of the gene, and results of Southern blot hybridization analysis point to the existence of a glucan synthase isozyme.

A Saccharomyces cerevisiae mutant with echinocandin-resistant 1,3-beta-D-glucan synthase.

A novel, potent, semisynthetic pneumocandin, L-733,560, was used to isolate a resistant mutant in Saccharomyces cerevisiae. This compound, like other pneumocandins and echinocandins, inhibits 1,3-beta-D-glucan synthase from Candida albicans (F.A. Bouffard, R.A. Zambias, J. F. Dropinski, J.M. Balkovec, M.L. Hammond, G.K. Abruzzo, K.F. Bartizal, J.A. Marrinan, M. B. Kurtz, D.C. McFadden, K.H. Nollstadt, M.A. Powles, and D.M. Schmatz, J. Med. Chem. 37:222-225, 1994). Glucan synthesis catalyzed by a crude membrane fraction prepared from the S. cerevisiae mutant R560-1C was resistant to inhibition by L-733,560. The nearly 50-fold increase in the 50% inhibitory concentration against glucan synthase was commensurate with the increase in whole-cell resistance. R560-1C was cross-resistant to other inhibitors of C. albicans 1,3-beta-D-glucan synthase (aculeacin A, dihydropapulacandin, and others) but not to compounds with different modes of action. Genetic analysis revealed that enzyme and whole-cell pneumocandin resistance was due to a single mutant gene, designated etg1-1 (echinocandin target gene 1), which was semidominant in heterozygous diploids. The etg1-1 mutation did not confer enhanced ability to metabolize L-733,560 and had no effect on the membrane-bound enzymes chitin synthase I and squalene synthase. Alkali-soluble beta-glucan synthesized by crude microsomes from R560-1C was indistinguishable from the wild-type product. 1,3-beta-D-Glucan synthase activity from R560-1C was fractionated with NaCl and Tergitol NP-40; reconstitution with fractions from wild-type membranes revealed that drug resistance is associated with the insoluble membrane fraction. We propose that the etg1-1 mutant gene encodes a subunit of the 1,3-beta-D-glucan synthase complex.

Cloning and characterization of the plasma membrane H(+)-ATPase from Candida albicans.

The Candida albicans PMA1 gene was isolated from a genomic library by using a hybridization probe obtained from the PMA1 gene of Saccharomyces cerevisiae. The gene was localized to chromosome III of the Candida genome. An open reading frame of 2,685 nucleotides predicts an amino acid sequence of 895 amino acids that is 83% homologous at both the DNA and protein levels to its S. cerevisiae equivalent. A polyadenylated mRNA transcript of about 4,000 nucleotides contains a highly folded AU-rich leader of 242 nucleotides. The structure of the gene, codon bias, and levels of approximately 100-kDa H(+)-ATPase protein recovered in plasma membranes indicate a highly expressed gene. The plasma membrane ATPase was purified to about 90% homogeneity and appeared to be blocked at the amino terminus. Three hydrophobic membrane sector tryptic fragments from the partially digested ATPase provided internal sequence information for over 50 amino acids, which agrees with the sequence predicted by the cloned gene. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis indicated that the C. albicans enzyme is about 3 kDa smaller than its Saccharomyces counterpart and was consistent with a predicted Mr of 97,398. Antibodies to the S. cerevisiae whole ATPase or its carboxyl terminus bound to the C. albicans enzyme but with lower avidity. Kinetic analysis showed that the Candida and Saccharomyces ATPases respond to glucose activation-starvation in nonidentical fashions. The amino-terminal domain of the C. albicans ATPase is marked by a net deletion of 23 amino acids in comparison with the S. cerevisiae ATPase. These differences maintain net charge, occur in nonconserved regions of fungal ATPases, and are sufficient to account for the observed difference in electrophoretic mobility between the two yeast ATPases.

Antithrombin III Utah: proline-407 to leucine mutation in a highly conserved region near the inhibitor reactive site.

A dysfunctional antithrombin III (ATIII) gene encoding a qualitatively and quantitatively abnormal anticoagulant molecule is responsible for hereditary thrombosis in a Utah kindred [Bock et al. (1985) Am. J. Hum. Genet. 37, 32-41]. Nucleotide sequencing of the entire protein-encoding portion of the cloned ATIII-Utah gene revealed a C to T transitional mutation which converts proline-407 to leucine. Proline-407 is located 14 amino acids C-terminal to the reactive site arginine of ATIII in a core region of the molecule that has been highly conserved during evolution of the serine protease inhibitor (serpin) gene family. The location of this proline in the crystal structure of the homologous serpin alpha 1-antitrypsin suggests that the leucine substitution in ATIII-Utah may interfere with correct folding of the mutant gene product, leading to its rapid turnover and the low antithrombin levels observed in patient plasmas. The Pro-407 to Leu mutation does not interfere with binding of antithrombin III to heparin. Patient antithrombin III, isolated by affinity chromatography on heparin-Sepharose, was reacted with purified thrombin. ATIII encoded by the patient's normal gene formed protease-inhibitor complexes with thrombin, whereas the product of the ATIII-Utah gene did not. The Pro-407 to Leu mutation destroys a restriction site for the enzyme StuI, permitting rapid diagnosis of affected members of the Utah kindred by Southern blotting of genomic DNA.