Phenotypic analysis of a family of transcriptional regulators, the zinc cluster proteins, in the human fungal pathogen Candida glabrata.

Candida glabrata is the second most important human fungal pathogen. Despite its formal name, C. glabrata is in fact more closely related to the nonpathogenic budding yeast Saccharomyces cerevisiae. However, less is known about the biology of this pathogen. Zinc cluster proteins form a large family of transcriptional regulators involved in the regulation of numerous processes such as the control of the metabolism of sugars, amino acids, fatty acids, as well as drug resistance. The C. glabrata genome encodes 41 known or putative zinc cluster proteins, and the majority of them are uncharacterized. We have generated a panel of strains carrying individual deletions of zinc cluster genes. Using a novel approach relying on tetracycline for conditional expression in C. glabrata at the translational level, we show that only two zinc cluster genes are essential. We have performed phenotypic analysis of nonessential zinc cluster genes. Our results show that two deletion strains are thermosensitive whereas two strains are sensitive to caffeine, an inhibitor of the target of rapamycin pathway. Increased salt tolerance has been observed for eight deletion strains, whereas one strain showed reduced tolerance to salt. We have also identified a number of strains with increased susceptibility to the antifungal drugs fluconazole and ketoconazole. Interestingly, one deletion strain showed decreased susceptibility to the antifungal micafungin. In summary, we have assigned phenotypes to more than half of the zinc cluster genes in C. glabrata. Our study provides a resource that will be useful to better understand the biological role of these transcription factors.

The fungal Candida species are the fourth most common cause of hospital-acquired infections and rank just after staphylococci and enterococci (Coleman and Mylonakis 2009; Kim and Sudbery 2011). In the recent years, a new emerging trend has been observed with a shift toward infections with species other than C. albicans [reviewed in (Miceli )]. For example, C. glabrata is now the second most important cause of fungal infections in humans (Roetzer ). Despite its formal name, C. glabrata is more closely related to the nonpathogenic baker’s yeast Saccharomyces cerevisiae. The C. glabrata genome contains 12.3 Mb and approximately 5300 coding genes (Dujon ). C. glabrata has gained genes involved in adhesion to mammalian cells [e.g., EPA genes encoding adhesins (Castano ; Cormack ; Silva )]. Gene loss has occurred in C. glabrata compared with S. cerevisiae. For example, C. glabrata lacks a number of genes for galactose, phosphate, and sulfur metabolism (Dujon ; Roetzer ). In contrast to C. albicans and S. cerevisiae, C. glabrata appears to be asexual and strictly haploid. Pseudohyphal growth has been reported for this organism (Csank and Haynes 2000); however, there is no evidence for hyphal formation or secretion of hydrolases that are associated with C. albicans virulence. C. glabrata can survive in the environment for many months. As a commensal, it is found on mucosal surfaces and, in contrast to C. albicans, tissue penetration is rarely observed (Roetzer ). In addition, this fungus can survive for an extended period of time in phagocytic cells. Little is known about factors involved in C. glabrata virulence.

A very important class of transcriptional regulators is composed of zinc cluster proteins (or binuclear cluster) that form a subfamily of zinc finger proteins. Zinc cluster proteins are exclusively found in fungi and amoeba (Clarke ; MacPherson ). These proteins possess the well-conserved motif CysX2CysX6CysX5-12CysX2CysX6-8Cys. The cysteine residues bind to two zinc atoms, which coordinate folding of the domain involved in DNA recognition (Vallee ). The vast majority of zinc cluster proteins act as transcriptional regulators [reviewed in ref. (MacPherson )]. The family of zinc cluster proteins is best characterized in S. cerevisiae. The genome of this organism encodes more than 50 known (or putative) zinc cluster proteins (MacPherson ). The first and best-studied zinc cluster protein is Gal4, a transcriptional activator of genes involved in the catabolism of galactose (Bhat and Murthy 2001). Many other zinc cluster proteins have been characterized; they control a large number of cellular processes such as the metabolism of amino acids, carbon (sugars and nonfermentable carbon sources), pyrimidine, fatty acid, as well as drug resistance (MacPherson ; Turcotte ). A number of zinc cluster proteins are positive regulators, but some function as both activators and repressors [e.g., Rds2 (Turcotte )], whereas Rdr1 appears to only down-regulate expression of target genes (Hellauer ).

Functional domains of zinc cluster proteins

Quite often, the DNA binding domain (comprising the cysteine-rich region) of zinc cluster proteins is located at the N-terminus whereas an acidic activating domain is located at the C-terminus. A region of low homology of about 80 amino acids, termed the middle homology region, is found among many zinc cluster proteins and is located between the DNA binding and activation domains and may be involved in controlling the transcriptional activity of zinc cluster proteins (Schjerling and Holmberg 1996). In many cases, deletion of the region that bridges the DNA binding domain to the activation domain results in constitutive activity of the transcriptional activator (MacPherson ). Many zinc cluster proteins bind to DNA as homodimers through a coiled-coil dimerization domain located at the C-terminus of the zinc finger but binding as heterodimers or monomers has also been reported (Akache ; Cahuzac ; Mamnun ; Rottensteiner ).

Zinc cluster proteins in C. glabrata

In C. glabrata, only a handful of zinc cluster proteins have been characterized (Table 1). CgPdr1, the homolog of S. cerevisiae Pdr1/Pdr3, confers drug resistance by positively controlling the expression of various genes including the ABC transporters CDR1, PDH1, and SNQ2 (Vermitsky ; Vermitsky and Edlind 2004) that act as drug efflux pumps. CgPdr1 is activated by direct binding of various compounds, including azoles that are antifungal drugs (Thakur ). As observed in S. cerevisiae, mutations in the CgPDR1 gene result in hyperactivation of the transcription factor, causing increased resistance to various drugs such as azoles and, unexpectedly, increased virulence (Berila ; Ferrari ; Tsai ; Vermitsky ). There are two functional homologs of S. cerevisiae Upc2/Ecm22 and they were named CgUpc2A and CgUpc2B (Nagi ). CgUpc2A is an activator of ergosterol biosynthetic genes whereas both CgUpc2A and B are positive regulators of the CgAUS1 gene encoding a sterol transporter (Nagi ). Deletion of CgUPC2A (but not B) results in sensitivity to azoles in analogy to S. cerevisiae, where we reported that a ∆upc2 strain is sensitive to ketoconazole whereas no effect was observed with a ∆ecm22 strain (Akache and Turcotte 2002). CgSTB5 encodes a repressor of the transporter genes CDR1, PDH1, and YOR1 (Noble ). Finally, CgCEP3 encodes a centromeric protein and is the functional homolog of S. cerevisiae CEP3 (Stoyan and Carbon 2004). In this study, we were interested in characterizing the whole family of zinc cluster proteins in C. glabrata. Toward this end, we have generated a panel of strains carrying deletions of zinc cluster genes. Results show that two zinc cluster genes are essential. Using our panel of deletion strains of nonessential zinc cluster genes, we performed phenotypic analysis under various conditions. Phenotypes identified in our screen include sensitivity to oxidative stress, increased tolerance to salt stress, and thermosensitivity. In addition, altered susceptibility to antifungal drugs was observed with a number of deletion strains.

Table 1

List of known and putative zinc cluster proteins

Name of Zinc Cluster Gene	Génolevures Code/Name of the Gene	S. cerevisiae Homolog	P Value	Deletion Strain Generated
CgZCF1	CAGL0A00451g CgPDR1 (ref. Vermitsky et al. 2006)	PDR1	2.6e-156	Yes
		PDR3	1.5e-105
CgZCF2	CAGL0A00583g	No homolog	N/A	Yes
CgZCF3	CAGL0A04455g	SEF1	1.9e-257	Yes
		LEU3	2.3e-21
CgZCF4	CAGL0B03421g	HAP1	5.5e-223	Yes
CgZCF5	CAGL0C01199g	UPC2	7e-183	Not generated; this gene is not essential in another strain background (Nagi et al. 2011)
	CgUPC2A (ref. Nagi et al. 2011)	ECM22	2.8e-169
CgZCF6	CAGL0D02904g	PPR1	1.6 e-229	Yes
		STB5	5.78e-16
CgZCF7	CAGL0D03850g	RSC3	1.1e-142	Yes
		RSC30	2.9e-29
CgZCF8	CAGL0E05434g	TEA1	7.7e-186	Yes
		CHA4	9.1e-106
CgZCF9	CAGL0F02519g	YJL206C	3.4e-129	Yes
		ASG1	1.1e-91
CgZCF10	CAGL0F03025g	ARO80	5.8e-151	Yes
CgZCF11	CAGL0F05357g	UME6	6e-51	Essential gene
		LYS14	1.1e-6
CgZCF12	CAGL0F06743g	DAL81	3.8e-184	Yes
		CHA4	7.9e-7
CgZCF13	CAGL0F07755g	CEP3	2.1e-140	Essential gene
	CgCEP3 (ref. Stoyan and Carbon 2004)	YKL122C	7.4e-5
CgZCF14	CAGL0F07865g	UPC2	7e-183	Yes
	CgUPC2B (ref. Nagi et al. 2011)	ECM22	2.8e-169
CgZCF15	CAGL0F07909g	TBS1	4.1e-111	Yes
		HAL9	2.4e-109
CgZCF16	CAGL0F09229g	YER184C	5.4e-80	Yes
		PDR1	2.8e-29
CgZCF17	CAGL0G08844g	ASG1	5.7e-214	Yes
		YJL206C	2.4e-92
CgZCF18	CAGL0G09757g	YLR278C	2.2e-264	Yes
		PPR1	4.4e-10
CgZCF19	CAGL0H00396g	LEU3	2.7e-247	Yes
		SEF1	9.2e-17
CgZCF20	CAGL0H01507g	RSC3	4.2e-99	Yes
		RSC30	3.1e-30
CgZCF21	CAGL0H01683g	URC2	1.8e-186	Yes
CgZCF22	CAGL0H04367g	WAR1	4.2e-133	Yes
CgZCF23	CAGL0H06875g	ARG81	1.1e-106	Yes
CgZCF24	CAGL0I02552g	STB5	7.9e-203	Yes
	CgSTB5 (ref. Noble et al. 2013)	YJL206C	2e-12
CgZCF25	CAGL0I07755g	HAL9	8.8e-196	Yes
		TBS1	6.5e-184
CgZCF26	CAGL0J07150g	OAF1	4.6e-165	Yes
		PIP2	1e-147
CgZCF27	CAGL0K05841g	HAP1	8.7 e-159	Yes
CgZCF28	CAGL0K06985g	ERT1	1e-142	Yes
		GSM1	6.1e-30
CgZCF29	CAGL0K11902g	LYS14	3.2e-240	Yes
CgZCF30	CAGL0L01903g	RGT1	1.9e-197	Yes
		EDS1	4.9e-54
CgZCF31	CAGL0L03377g	SIP4	2.6e-91	Yes
		CAT8	1.9e-13
CgZCF32	CAGL0L03674g	GSM1	1.7e-81	Yes
		RDS2	1.3e-15
CgZCF33	CAGL0L04400g	YRR1	9.9e-115	Yes
		YRM1	2.1e-112
CgZCF34	CAGL0L04576g	YRM1	2.3e-134	Yes
		YRR1	2.2e-122
CgZCF35	CAGL0M11440g	CHA4	7.8e-171	Yes
		TEA1	3e-94
CgZCF36	CAGL0L09383g	SUT1	3.3e-33	Yes
		SUT2	4.5e-28
CgZCF37	CAGL0L09691g	PUT3	2.1e-200	Yes
		ASG1	6.8e-20
CgZCF38	CAGL0M12298g	OAF1	1.2e-265	Yes
		PIP2	2.1e-183
CgZCF39	CAGL0M02651g	RDS2	1e-126	Yes
		ERT1	5.3e-28
CgZCF40	CAGL0M05907g	OAF3	2.6e-116	Yes
CgZCF41	CAGL0M03025g	CAT8	1.9e-148	Not studied
		ASG1	4.2e-13

C. glabrata zinc cluster genes are numbered 1−41. Systematic names (Génolevures code, www.genolevures.org) are also given as well as their gene names (if available). The S. cerevisiae closest homologs are also listed along with P-values. More information about S. cerevisiae zinc cluster genes can be obtained at www.yeastgenome.org. Essential genes are also indicated. Deletion of CgZCF5 was not obtained in the reference strain used in this study.

Material and Methods

Strains and media

The wild-type S. cerevisiae strain used for construction of plasmids by homologous recombination is BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) (Brachmann ). The wild-type C. glabrata strain 66032ura3 (Vermitsky ) used to generate the zinc cluster gene deletions is a tight 5-fluoroorotic acid selected ura3 derivative of strain ATCC 66032. Yeast cells were grown in YPD (2% yeast extract, 1% peptone, 2% glucose) medium or in SD complete medium lacking appropriate auxotrophic components (Adams ). For selection with the dominant SAT1 marker (Reuss ), YPD agar plates containing nourseothricin (cloneNAT, Werner BioAgents) at 200 μg/mL were used.

Plasmids for gene deletion

The overall strategy used to construct plasmids for deletion of zinc genes is schematically shown in Figure 2, and oligonucleotides used to generate plasmids for gene deletion are listed in Supporting Information, Table S1. Plasmid pRS316 (Sikorski and Hieter 1989) was used as a template to amplify the URA3 marker with oligonucleotides URA3REC-1 and URA3REC-2 that contain sequences homologous to DNA flanking the SmaI site in plasmid pRS423 (Sikorski and Hieter 1989). The polymerase chain reaction (PCR) product was transformed into S. cerevisiae along with plasmid pRS423 (Brachmann ) linearized with SmaI, and transformants were selected on minimal plates lacking histidine followed by selection on plates lacking uracil. Yeast DNA was isolated according to Hoffman and Winston (1987), and plasmids were recovered by transformation into Escherichia coli (DH5α-E) using ElectroMAX electrocompetent cells (Invitrogen) to yield plasmid pRS-URA3. To generate a panel of deletion C. glabrata strains (Table 1), a set of plasmids containing disruption cassettes was generated.

Figure 2

Strategy used to generate cassettes for deletion of zinc cluster genes in C. glabrata. Fragments corresponding to sequences flanking an open reading frame (ORF) of interest were amplified by polymerase chain reaction (PCR). Oligos were designed so that they contain 45 bp of homology to the plasmid pRS-URA3 containing the S. cerevisiae URA3 gene. pRS-URA3 was digested with SmaI and transformed along with the two PCR products into S. cerevisiae. A quadruple recombination between the plasmid backbone, the PCR products and the URA3 marker allows the generation of a plasmid which can be recovered and amplified in E. coli. After digestion with SmaI, the DNA is then transformed into C. glabrata.

The 5′ and 3′ regions flanking of the open reading frames (ORFs) of the zinc cluster genes were amplified by PCR using genomic DNA isolated from strain 66032ura3. Oligonucleotides were designed to contain SmaI restriction sites at the 5′ and 3′ ends and sequences complementary to the 5′ and 3′ end of the URA3 marker in pRS-URA3. The 5′ flanking PCR fragment (termed CgZCXXA, where XX refers to a numbered zinc cluster protein) was homologous to the 5′ end of the URA3 marker in pRS-URA3 and was obtained using primer oligonucleotides CgZCXX-a and CgZCXX-b. Similarly, the resulting 3′ flanking PCR fragment (termed CgZCXXB) was homologous to the 3′ end of the URA3 marker in pRS-URA3, using primer oligonucleotides CgZCXX-c and CgZCXX-d resulting in PCR products that were approximately 500-bp long. Plasmid pRS-URA3 linearized with SmaI was transformed with the 5′ and 3′ PCR products in the S. cerevisiae strain BY4741. The flanking PCR fragments were recombined into the SmaI-digested pRS-URA3 to generate plasmids via a quadruple recombination. Selection was performed on SD agar plates lacking histidine followed by selection on SD plates lacking uracil. Plasmids were recovered as described previously. Plasmids were named pCgZCF1 to 40 (Table 1). Independent clones were verified by DNA sequencing.

Deletion of zinc cluster genes

Plasmids for deletion of zinc cluster genes were digested with SmaI, purified using a QIAquick PCR purification kit (QIAGEN), and 1 µg of plasmid DNA was transformed into the strain 66023ura3 using the lithium acetate procedure (Gietz ) except that dimethyl sulfoxide (10% final concentration) was added before the heat shock (42°, 5 min). Cells were plated on SD agar plates lacking uracil, and colonies were restreaked on SD agar plates lacking uracil. Proper integration of the S. cerevisiae URA3 marker was verified using a reverse PCR primer that overlapped the URA3 marker (either URA3-CHECK or URA3-CHECK#2) and forward PCR primer that was complementary to genomic sequences upstream of the 5′ region used to perform homologous recombination (termed CgZCXX-check, see Table S1). PCR primers specific to the DNA binding domain for zinc cluster genes were used to ensure complete removal of the ORF of a zinc cluster gene (data not shown). In addition to URA3, deletion of CgZCF6 was also obtained using the dominant marker SAT1 (Reuss ). The SAT1 marker was amplified using oligonucleotides CgZCF6-SAT1F-I and CgZCF6-SAT1R-I and plasmid pSFS2A (Reuss ). To extend the length of sequences homologous to CgZCF6, the PCR product was used as a template for a second PCR amplification using oligonucleotides CgZCF6-SAT1F-II and CgZCF6-SAT1R-II. A cassette for deletion of CgZCF23 was obtained by amplifying the Myc-URA3-Myc sequences of plasmid pMPY-3xMyc (Schneider ) using oligos PET-CgZC23-1 and KO-CgZC23-2 followed by a second amplification using oligonucleotides PET-CgZC23-4 and KO-CgZC23-4. Similarly, one deletion strain for CgPDR1 was generated using plasmid pMPY-3xMyc and the oligonucleotides PET-CgZC1-1, KO-CgZC1-2, PET-CgZC1-3, and KO-CgZC1-4.

Complementation assays

Zinc cluster genes were amplified using the Expand Long Template PCR System (Roche) with genomic DNA isolated from strain 66032ura3 and oligonucleotides listed in Table S1. Oligonucleotides were designed so that approximately 200−400 bp of sequences flanking an ORF of interest were part of the PCR product. DNA was purified with a QIAquick PCR Purification Kit (QIAGEN) and used to transform deletion strains carrying the URA3 marker. Cells were then directly plated on FOA plates to select for Ura− cells. With the exception of CgZCF9, at least two complementation strains (usually three or more strains) for each deletion strain were tested for reversion of the phenotype. All complementation strains tested showed wild-type phenotypes.

Conditional expression of zinc cluster proteins

The G418R marker of plasmid pADH1-tc3-3XHA (Kotter ) was replaced by the S. cerevisiae URA3 marker. To this end, oligonucleotides ScURA3-1 and ScURA3-2 were used to amplify the URA3 gene using plasmid pRS316 as a template (Sikorski and Hieter 1989). The PCR product was cut with BamHI and SacI and subcloned into pADH1-tc3-3XHA cut with the same enzymes to yield plasmid pADH1-tc3-3XHA-URA3. This plasmid was used as a template to generate a cassette for integration at a specific promoter using oligonucleotides Tc-CgZCFXX-1 and Tc-CgZCFXX-2. The PCR product was used as a template for a second round of PCR with oligonucleotides Tc-CgZCFXX-3 and Tc-CgZCFXX-4.

Phenotypic analysis and minimal inhibitory concentration (MIC) assays

Fluconazole and ketoconazole were obtained from Medisca (Montréal, Canada). Micafungin and caspofungin were obtained from Astellas (Markham, Ontario, Canada) and Merck Frost (Kirkland, Québec, Canada), respectively. Sensitivity to drugs was assayed in liquid YPD and on YPD agar plates containing various drugs as detailed in the figures. Strains were grown overnight in liquid YPD medium. The cultures then were diluted at 0.2 OD600 and further diluted 5, 25, and 125 times and spotted on appropriate plates. Growth was monitored after 1−2 d. MIC assays were performed as described (Znaidi ).

Results and Discussion

To identify zinc cluster genes in C. glabrata, we used Gal4 and related proteins as queries to perform a BLAST search of the C. glabrata genome. We identified 41 known or putative zinc cluster genes. Alignments of the various zinc cluster motifs are shown in Figure 1. With the exception of CgZCF36, the zinc cluster motifs all match the consensus sequence described previously. CgZCF36 has an extended sequence (77 a.a.) between the third and fourth cysteine. However, a similar spacing is found in some S. cerevisiae zinc cluster proteins (MacPherson ), suggesting that CgZCF36 also encodes a zinc cluster protein. In Gal4, the motif Arg-X2.-Lys-X-Lys (where X is any amino acid) is found between the second and third cysteines. The first arginine and the second lysine form salt bridges with phosphate groups in DNA whereas the first lysine is involved in making base-specific contacts (Marmorstein ). Other S. cerevisiae zinc cluster proteins also harbor this motif, even though some of them have, for example, Arg, His, or Asn residues instead of the first Lys. Strikingly, this motif is also found in the vast majority of zinc cluster proteins in C. glabrata (Figure 1). In summary, the C. glabrata genome contains 41 zinc cluster genes that are highly likely to encode bona fide zinc cluster proteins. A list of the C. glabrata 41 known or putative zinc cluster genes is provided in Table 1 along with their S. cerevisiae homologs.

Figure 1

Alignment of the cysteine-rich motif of S. cerevisiae Gal4 with C. glabrata zinc cluster proteins. C. glabrata zinc cluster proteins were identified by BLAST searches of the C. glabrata genome using S. cerevisiae Gal4 and other zinc cluster proteins as queries and were named CgZCF1 to 41 (C. glabrata Zinc Cluster Factor). The cysteines residues (in yellow) of the 41 putative or known zinc cluster proteins were aligned using Gal4 as a reference. A consensus sequence is shown on top of the figure. Some residues (located between the second and third cysteines) are involved in DNA recognition by Gal4 and are shown in turquoise. Conserved or alternate residues found in other S. cerevisiae zinc cluster proteins are shown in green. Systematic and gene names are listed in Table 1.

Strikingly, 36 of 41 zinc cluster genes in C. glabrata are uncharacterized (Table 1). To obtain insights into the function of these putative zinc cluster proteins, we generated a panel of deletion strains. To this end, we constructed plasmids containing the S. cerevisiae URA3 gene flanked by approximately 500 bp of sequences located upstream and downstream of the ORF of a zinc cluster gene of interest (Figure 2). Linearized plasmids were transformed into a Ura− C. glabrata and transformants were selected on plates lacking uracil. Using this strategy, we successfully deleted 37 of 40 zinc cluster genes (the zinc cluster gene CgZCF41 was not included in the analysis).

To test whether the three remaining genes are essential, we adapted a procedure initially developed for S. cerevisiae for use in C. glabrata (Kotter ). The natural promoters of the genes of interest were replaced with the S. cerevisiae promoter ADH1 followed by three aptamers (3XTc) that were inserted just upstream of the initiating codon. The RNA aptamers, located in the 5′ UTR, bind with high affinity to tetracycline, resulting in the formation of a secondary structure that prevents translation, thus verifying if a gene is essential (Figure 3). As expected (Kotter ), the addition of tetracycline did not affect growth of the wild-type strain. As a positive control, we conditionally expressed the topoisomerase CgTop2, a homolog of S. cerevisiae Top2 encoded by an essential gene. Inhibition of translation of the CgTOP2 mRNA by addition of tetracycline completely abolished growth, thus validating this assay in C. glabrata. Similarly, inhibition of CgZcf13 (CgCep3) expression prevented growth, in agreement with a study which showed that the CgCEP3 gene is essential (Stoyan and Carbon 2004). Our results also show that CgZCF11 is an essential gene, whereas it is not clear whether CgZcf5 (a S. cerevisiae ortholog of Upc2/Ecm22) is essential in the strain used for our experiments. The CgZCF5 gene is dispensable in a different strain background (Nagi ). Thus, only two zinc cluster genes are essential in C. glabrata. A similar phenomenon was observed in S. cerevisiae where only two zinc cluster genes are essential, including CEP3 (Akache and Turcotte 2002).

Figure 3

The genes CgZCF11, CgZCF13, and CgTOP2 are essential. Strains (as listed on the left) were grown overnight in rich medium containing 100 µM tetracycline. Cells were then serially diluted and spotted on plates containing tetracycline at concentrations indicated on the top of the Figure and plates were incubated at 30° for 24 h. Spotting experiments were performed with two independent clones for the genes tested. As a control, a C. glabrata ortholog of the essential S. cerevisiae gene TOP2 (encoding topoisomerase II) was used. It is not clear whether CgZCF5 is essential or not because partial growth inhibition could be due, for example, to incomplete translational inhibition.

Phenotypic analysis of strains lacking zinc cluster genes

Using our panel of deletion strains, we performed phenotypic analysis under various conditions (e.g., high temperature, salt stress, exposure to antifungal drugs, etc.) and phenotypes are listed in Table 2. Phenotypes for a number of deletion strains are described herein whereas data for the remaining strains can be found in Figure S1. In addition, complementation assays using at least two revertant strains for all deleted zinc cluster genes (with the exception of CgZCF9 where only one revertant strain was obtained) confirmed that the observed phenotypes were due to deletion of a given zinc cluster gene and not to secondary mutations (see herein and Figure S1). Two deletion strains, Cg∆zcf7 and Cg∆zcf20, are thermosensitive (Figure 4A). Introduction of wild-type alleles in the deletion strains restored growth at high temperature. One deletion strain (Cg∆zcf24) showed high sensitivity to oxidative stress, as assayed with H2O2 (Figure 4B), in agreement with a previous report (Noble ). CgZcf24 is highly homologous to S. cerevisiae Stb5. We previously showed that deletion of STB5 results in sensitivity to oxidative stress and that Stb5 is an activator to genes of the pentose phosphate pathway and other genes involved in the production of NADPH, a cofactor involved in conferring resistance to oxidative stress (Larochelle ). CgStb5 does not appear, however, to regulate genes of the pentose phosphate pathway (data not shown), in agreement Noble . It will be interesting to determine the reason for the sensitivity to oxidative stress of cells lacking CgZCF24.

Table 2

Summary of the phenotypes observed for strains carrying deletions of zinc cluster genes

Zinc Cluster Gene Deleted	Fluconazole	Ketoconazole	Micafungin	H2O2	42°	Caffeine	LiCl	SDS
CgZCF1 (PDR1)	Highly sens.	Highly sens.	Not tested	Not tested	Not tested	Not tested	Not tested	Not tested
CgZCF4	Slightly sens.	Sens.					Res.
CgZCF6			Res.					Sens.
CgZCF7				Slightly sens.	Sens.	Sens.	Sens.
CgZCF9	Slightly sens.	Sens.
CgZCF10	Slightly sens.	Sens.					Res.
CgZCF12	Slightly sens.	Sens.
CgZCF16	Slightly sens.	Sens
CgZCF17							Res.
CgZCF18	Slightly sens.	Sens.
CgZCF20				Slightly sens.	Sens.	Sens.	Slightly sens.
CgZCF23	Slightly sens.	Sens.
CgZCF24 (STB5)	Slightly sens.	Sens.	Slightly sens.	Sens.			Res.
CgZCF26	Slightly sens.	Slightly sens.					Res.
CgZCF27	Slightly sens..	Sens.
CgZCF29	Slightly sens.	Sens.
CgZCF31	Slightly sens.	Sens.
CgZCF33	Slightly sens.	Sens.
CgZCF36	Slightly sens.	Sens.					Res.
CgZCF37	Slightly sens.	Sens.					Res.
CgZCF39	Slightly sens.	Slightly sens.					Res.

For azoles compounds, a deletion strain was scored as sensitive if the MIC difference with the wild-type stain was 2 or more (see Table 3). CgZCF5, CgZCF41, and the essential genes CgZCF11, CgZCF13 were not included in the phenotypic analysis. See the Results section as well as Figure S1 for spotting assays. SDS, sodium dodecyl sulfate; Sens., sensitive; Res., resistant.

Figure 4

Strains Cg∆zcf7 and Cg∆zcf20 are thermosensitive whereas strain Cg∆zcf24 is sensitive to oxidative stress. Strains were grown overnight in rich medium, serially diluted, and spotted on plates as described in the section Material and Methods. (A) Two independent clones of deletion strains Cg∆zcf7 and Cg∆zcf20 were tested and are Ura+. Cg∆zcf7-1A + ZCF7 and Cg∆zcf20-1A + ZCF20 are deletion strains were a wild-type allele was introduced and the strains are Ura-. B) Two independent clones of deletion strain Cg∆zcf24 were tested and are Ura+. Cg∆zcf24-1A + ZCF24 is a deletion strain were a wild-type allele was introduced and the strain is Ura−.

Table 3

MIC values for fluconazole and ketoconazole as measured in various deletion strains

Strain	MIC Fluconazole, µg/mL	Fold Difference	MIC Ketoconazole, µg/mL	Fold Difference
WT URA3	32	N/A	0.5	N/A
WT ura3	32	N/A	0.5	N/A
Cg∆zcf1 (Cg∆pdr1)	4	8	No growth with 0.03 µg/mL	>8
Cg∆zcf4	32	N/A	0.25	2
Cg∆zcf9	Not tested	N/A	0.25	2
Cg∆zcf10	Not tested	N/A	0.25	2
Cg∆zcf12	Not tested	N/A	0.25	2
Cg∆zcf16	Not tested	N/A	0.25	2
Cg∆zcf18	Not tested	N/A	0.25	2
Cg∆zcf23	32	N/A	0.25	2
Cg∆zcf24	32	N/A	0.25	2
Cg∆zcf26	32	N/A	0.5	N/A
Cg∆zcf27	16-26	≈ 2	0.25	2
Cg∆zcf29	32	N/A	0.25	2
Cg∆zcf31	32	N/A	0.25	2
Cg∆zcf33	32	N/A	0.25	2
Cg∆zcf36	Not tested	N/A	0.25	2
Cg∆zcf37	Not tested	N/A	0.25	2
Cg∆zcf39	32	N/A	0.5	N/A

Deletion strains that showed sensitivity to azoles with spotting assays were used to perform MIC assays. MIC, minimal inhibitory concentration; WT, wild type; N/A, not applicable.

We also tested deletion strains for sensitivity to caffeine, an inhibitor of the target of rapamycin pathway (Reinke ). Cells lacking CgZCF7 or CgZCF20 were sensitive to caffeine (Figure 5) whereas reintroduction of the wild-type alleles in the deletion strains resulted in a wild-type phenotype. High concentrations of sorbitol cause osmotic stress and activation of the high-osmolarity glycerol pathway (Saito and Posas 2012). However, sensitivity to sorbitol was not observed in our screen. Regarding tolerance to salt (150 mM LiCl), only one deletion strain (Cg∆zcf7) showed sensitivity under this condition (Figure S1). Unexpectedly, our results show that deletion of eight zinc cluster genes (CgZCF4, CgZCF10, CgZCF17, CgZCF24, CgZCF26, CgZCF36, CgZCF37, and CgZCF39) rather results in increased tolerance to salt stress (Figure 6 and Figure S1).

Figure 5

Sensitivity of deletion strains to caffeine. Strains were grown overnight in rich medium, serially diluted and spotted on plates as described in the section Material and Methods. Two independent clones of deletion strains Cg∆zcf7 and Cg∆zcf20 were tested and are Ura+. Cg∆zcf7-1A + ZCF7, Cg∆zcf7-1B + ZCF7, Cg∆zcf20-1A + ZCF20, and Cg∆zcf20-1B + ZCF20 are deletion strains were a wild-type allele was introduced and the strains are Ura−.

Figure 6

Strains Cg∆zcf26 and Cg∆zcf37 show increased tolerance to salt stress. Strains were grown overnight in rich medium, serially diluted and spotted on plates as described in the section Material and Methods. Two independent clones of deletion strains Cg∆zcf26 and Cg∆zcf37 were tested and are Ura+. Cg∆zcf26-1A + ZCF26, Cg∆zcf26-1B + ZCF26, and Cg∆zcf37-1A + ZCF37 are deletion strains were a wild-type allele was introduced and the strains are Ura−.

Phenotypic analysis also was performed with antifungal drugs. Azoles, such as fluconazole or ketoconazole, are fungistatic antifungal drugs. These compounds target lanosterol 14α-demethylase involved in the synthesis of ergosterol and this enzyme is encoded by the ERG11 gene. Antifungal activity is caused by decreased ergosterol levels and increased production of toxic ergosterol derivatives (Lupetti ). As expected (Vermitsky ; Vermitsky and Edlind 2004), deletion of CgPDR1 (CgZCF1) greatly increased susceptibility to fluconazole and ketoconazole (≥eightfold difference in MIC, Table 3). Unexpectedly, increased susceptibility to azoles (in particular ketoconazole) was observed in many deletion strains (total of 15, Table 3). These strains showed a twofold reduced MIC compared with the wild-type strain. Figure 7 shows spotting assays for deletion strains Cg∆zcf4 and Cg∆zcf37. In agreement with MIC values, both strains showed increased susceptibility to ketoconazole, whereas slightly increased susceptibility was observed for fluconazole. We also note the presence of some small colonies in the presence of fluconazole or ketoconazole. These colonies are probably resistant to the azoles due, for instance, to mutations in CgPDR1 (Tsai ).

Figure 7

Increased susceptibility of deletion strains Cg∆zcf4 and Cg∆zcf37 to azoles. Strains were grown overnight in rich medium, serially diluted, and spotted on plates with or without drugs as indicated in the figure. Two independent clones of deletion strains Cg∆zcf4 and Cg∆zcf37 were tested and are Ura+. Cg∆zcf4-2A + ZCF4, Cg∆zcf4-2B + ZCF4, Cg∆zcf37-2A + ZCF37 and Cg∆zcf37-2B + ZCF37 are deletion strains were a wild-type allele was introduced and the strains are Ura−.

Echinocandins (e.g., caspofungin, micafungin) are the latest class of antifungal drugs used in the clinic, and they have fungicidal activity (reviewed in Chen and Mayr ). Echinocandins inhibit the activity of a two-subunit enzyme involved in the synthesis of the polysaccharide 1,3-β-glucan, which is a major and essential component of the cell wall. In S. cerevisiae, one subunit is encoded by the genes FKS1, FKS2, and FKS3 whereas the second one is encoded by RHO1. Resistance to echinocandins has been attributed to mutations in the FKS1 and FKS2 genes (Johnson ; Kahn ; Katiyar ; Thompson ). A strain carrying a deletion of CgZCF24 showed slightly increased susceptibility to micafungin (data not shown). Interestingly, deletion of CgZCF6 resulted in reduced susceptibility to micafungin, as determined by spotting assays (Figure 8, top panel). Moreover, with a Cg∆zcf6 strain, we observed a twofold increase in MIC for micafungin whereas only a slight difference was observed with caspofungin (Figure 8, bottom panel). The Cg∆zcf6 strain is also sensitive to 0.04% sodium dodecyl sulfate (data not shown), a phenotype that is indicative of cell wall defects. These phenotypes were also observed using a Cg∆zcf6 deletion strain generated with the dominant marker SAT1 instead of URA3 (Figure S1).

Figure 8

Strain Cg∆zcf6 shows reduced susceptibility to micafungin. Strains were grown overnight in rich medium, serially diluted, and spotted on plates as described in the section Material and Methods. Two independent clones of deletion strains Cg∆zcf6 were tested and are Ura+. Cg∆zcf6-1A + ZCF6 is a deletion strain were a wild-type allele was introduced and the strain is Ura-. MIC values are given at the bottom of the figure.

The S. cerevisiae zinc cluster protein Ppr1 is highly homologous to CgZcf6 (P-value 1.6 X10−229). Ppr1 is an activator of the URA genes involved in pyrimidine synthesis (Losson and Lacroute 1981; MacPherson ). However, it is not clear whether the two factors perform the same function. For example, a ∆ppr1 strain does not show altered susceptibility to micafungin (data not shown). Transcription factor rewiring may explain the apparent functional difference between CgZcf6 and ScPpr1. It will be interesting to determine the molecular basis for the decreased susceptibility to micafungin of a Cg∆zcf6 strain.

In this study, we have performed phenotypic analysis of a C. glabrata family of transcriptional regulators, the zinc cluster proteins. Results show that only two zinc cluster genes are essential (Figure 3). Their gene products may be potential targets for antifungal drugs because zinc cluster proteins are fungal (and amoebae) specific. Phenotypes have been identified for more than half of the zinc cluster genes, strongly suggesting that these genes do encode functional proteins. However, we were unable to assign phenotypes for a number of zinc cluster proteins. Some of them may perform functions related to a specific environment (e.g., survival in macrophages) or may show redundancy. In summary, our panel of deletion strains along with our phenotypic analysis will provide useful tools to the researcher community for the study of this family of regulators in an important fungal pathogen.