Rapid Hypothesis Testing with Candida albicans through Gene Disruption with Short Homology Regions

Abstract

Disruption of newly identified genes in the pathogen Candida albicans is a vital step in determination of gene function. Several gene disruption methods described previously employ long regions of homology flanking a selectable marker. Here, we describe disruption of C. albicans genes with PCR products that have 50 to 60 bp of homology to a genomic sequence on each end of a selectable marker. We used the method to disrupt two known genes, ARG5 and ADE2, and two sequences newly identified through the Candida genome project, HRM101 and ENX3. HRM101 and ENX3 are homologous to genes in the conserved RIM101 (previously called RIM1) and PacC pathways of Saccharomyces cerevisiae and Aspergillus nidulans. We show that three independent hrm101/hrm101 mutants and two independent enx3/enx3 mutants are defective in filamentation on Spider medium. These observations argue that HRM101 and ENX3 sequences are indeed portions of genes and that the respective gene products have related functions.

Candida albicans is an opportunistic fungal pathogen. It exists as a benign commensal organism in healthy individuals but causes infections in susceptible individuals, such as those with diminished immune function (14). Molecular genetic analysis of C. albicans has permitted evaluation of antifungal drug targets and elucidation of requirements for infection and pathogenesis (16).

New C. albicans genes have been identified frequently through sequence homology to known genes or gene families. Gene discovery has been facilitated greatly by access to much of the C. albicans genomic sequence (11). Now, the rate-limiting step in analysis of gene function in this diploid organism is the creation of a homozygous disruption mutant. Gene disruption has been accomplished through successive transformations with insertion/deletion alleles that are constructed in vitro (2, 7, 12). These methods have thus far required isolation of substantial DNA segments, and yet new genes of interest are often identified through DNA sequences of 400 to 600 bp (3a). We report here a rapid method for disruption of C. albicans genes with PCR products that contain short regions of homology to the genome.

MATERIALS AND METHODS

Strains.

The C. albicans strains used in this study are SC5314 (wild type [2]) and its derivatives CAI4 (ura3Δ::λimm434/ura3Δ::λimm434 [2]) and RM1000 (ura3Δ::λimm434/ura3Δ::λimm434 his1::hisG/his1::hisG [12]). Strain Arg-het1 (ura3Δ::λimm434/ura3Δ::λimm434 arg5::hisG/ARG5) was constructed through transformation of strain CAI4 with the arg5::hisG-URA3-hisG disruption fragment derived from plasmid pUC-ARG-U (12); uridine-prototrophic (Uri⁺) transformants were then purified and subjected to 5FOA selection for excision of URA3. Presence of the hisG cassette was verified by PCR detection with primers hisG-N and hisG-C (Table 1). Strain BWP17 (ura3Δ::λimm434/ura3Δ::λimm434 his1::hisG/his1::hisG arg4::hisG/arg4::hisG) was constructed through two successive transformations of strain RM1000 with an arg4::hisG-URA3-hisG disruption fragment derived from plasmid pRS-ArgBlaster (see below); each transformation was followed by Uri⁻ counterselection on 5FOA medium. Escherichia coli DH5α was used for plasmid propagation.

TABLE 1.

Primer sequences

Primer	Sequencea	Purposeb
ARG5-5DR	AAGTGTTTCTTAGAGCAAAACTTGCGTTTGGTGACAGCTTTGGAAAAAATTGGTGTTCACGTTTTCCCAGTCACGACGTT	Disruption of ARG5
ARG5-3DR	ATCTTGGCAAATGTTGTAATAAATCGTGAATTTCCTTGATCTTCAATTTAGTACCATATTTGTGGAATTGTGAGCGGATA	Disruption of ARG5
ADE2-5DR	GTCCATTATATGCTGAAAAATGGTGTCCTTTCACCAAAGAATTGGCTGTGGTTTTCCCAGTCACGACGTT	Disruption of ADE2
ADE2-3DR	GGGTTGCCTTATCACCCAAGACATTCAACATAATAGCATTGGTGATGGAATGTGGAATTGTGAGCGGATA	Disruption of ADE2
ARG4-N2K	ATTCGGATCCGGTACCCCCCTTTAGTAAGATTTTTCAAGAGTAG	ARG4 cloning
ARG4-CS	ATTCTCGAGCCCGGGCAATGCTTGAGGAGAAGAATCAGAACGC	ARG4 cloning
ca-ura-5	TTGGATGGTATAAACGGAAACA	URA3 cloning
ca-ura-3	TCTAGAAGGACCACCTTTGATTG	URA3 cloning
ca-his-5	CCTGGAGGATGAGGAGACAG	HIS1 cloning
ca-his-3	CCAATATATCGGTTGCACCA	HIS1 cloning
5-detect	GTTTTCCCAGTCACGACGTTGTAAAACGAC	Detection of vector sequences flanking disruption marker
3-detect	TGTGGAATTGTGAGCGGATAACAATTTCAC	Detection of vector sequences flanking disruption marker
hisG-N	CGCGATACAGACCGGTTCAGACAGGA	Detection of hisG sequences
hisG-C	TGGTCTTTACTCCATCACAGGGTTCC	Detection of hisG sequences
RIM101-5DR	ACGATCATTGTGTGACGACCATGTTGGTAGAAAGTCTTCGAACAATTTGTCATTGACTTGTGTGGAATTGTGAGCGGATA	Disruption of HRM101
RIM101-3DR	ACATGGACTCTCAAGTGAGAAGTAATGTGATCTCTCTAACTGTAGTTGTGCCACAATTTTTTTCCCAGTCACGACGTT	Disruption of HRM101
RIM101-5a	GGGGAATTCGTGCTAATCAATCTAACACCACAGCTCTGC	Detection of HRM101 alleles
seq7	GGTGAACTCAGCCAGAACCTGCG	Detection of HRM101 alleles
PalA-5DR	GCAGCACAAGAGTTAATTAAGAAAGTAGATAAAATGAAACAATATTTGTTACAGGCTAACAACGGAGATGGGTGGAATTGTGAGCGATA	Disruption of ENX3
PalA-3DR	GGAACGAGTTACTAATAGCTAATTCTAAGTCTCGACTCTCAACTCTTCGTCTATACAAATATTCCTCACTTTCCCAGTCACGACGTT	Disruption of ENX3
PalA5′	ACTGATGATGCAGCACAAGAG	Detection of ENX3 alleles
PalA3′	CCAGGTTTACTAATAGTCGG	Detection of ENX3 alleles

^aBoldface sequences in 5DR and 3DR primers are segments that anneal to plasmids pGEM-HIS1, pGEM-URA3, and pRS-ARG4ΔSpeI for amplification of disruption cassettes. 

^bRelevant use of primer in this study. 

Media.

YPD+Uri medium was used for routine nonselective propagation of C. albicans strains; it contains, per liter, 20 g of dextrose, 20 g of Bacto Peptone, 10 g of Difco yeast extract, and 80 mg of uridine. Synthetic medium (SD) contains, per liter, 20 g of dextrose and 6.7 g of Difco yeast nitrogen base without amino acids (6). SD was supplemented with necessary auxotrophic requirements as described previously (6), except that uridine was added at 80 mg/liter to supplement Uri⁻ strains. Solid Spider medium was prepared as described previously (9), except that it was supplemented with uridine for growth of Uri⁻ strains.

Plasmids. (i) pGEM-URA3.

A 1.2-kbp URA3 fragment was amplified by PCR from strain SC5314 template DNA with primers ca-ura-3 and ca-ura-5 (Table 1). The fragment was gel purified and ligated to vector pGEM-T (Promega) to yield plasmid pGEM-URA3 (Fig. 1A).

FIG. 1.

Restriction maps. In panels A to E, thin horizontal lines represent DNA segments and thick lines represent the relevant open reading frame with 5′ and 3′ ends on the left and right, respectively. Panel D is drawn at a smaller scale than the other panels. Panels A to C represent the entire cassettes used for primer-directed gene disruption. (A) Insert of plasmid pGEM-URA3. The URA3 PCR product lies between the SphI and SalI sites in vector pGEM-T. (B) Insert of plasmid pGEM-HIS1. The HIS1 genomic fragment lies between the SphI and SalI sites in vector pGEM-T. (C) Insert of plasmid pRS-ARG4ΔSpeI. The ARG4 PCR product lies between the KpnI and SpeI sites in vector pRS314. (D) Genomic ARG5 locus. The open box represents the insertion/deletion allele arg5::hisG (12). Positions of the disruption primers and probe for Southern analysis are shown above the restriction map. (E) Genomic ADE2 locus. Positions of the disruption primers and probe for Southern analysis are shown above the restriction map. (F) Homology relationships between ARG5, the ARG5 disruption primers, and URA3 and HIS1 cassettes. The ARG5-5DR and ARG5-3DR primers are represented by arrows, which point to each primer’s 3′ end. Sets of parallel vertical lines indicate homology between the primers and ARG5 sequences or disruption cassette templates. The 461-bp segment of ARG5 shown above the primers is deleted from the arg5::hisG allele (12). This diagram is not to scale.

(ii) pGEM-HIS1.

A 1.0-kbp HIS1 fragment was amplified by PCR from strain SC5314 template DNA with primers ca-his-3 and ca-his-5 (Table 1). The fragment was gel purified and ligated into vector pGEM-T (Promega) to yield plasmid pGEM-HISX. This insert lacked a few 3′ codons of HIS1 and failed to complement a Saccharomyces cerevisiae his1 mutant, so we used integrative transformation to retrieve additional HIS1 sequences. The integrating plasmid pGEM-URA/HIS was constructed by ligation of the 1.0-kbp NcoI fragment of plasmid pGEM-HISX (which contains HIS1 sequences) into NcoI-cut, phosphatase-treated plasmid pGEM-URA3. One resulting plasmid had the HIS1 3′ end adjacent to URA3 sequences. This plasmid, designated pGEM-URA/HIS, was digested within HIS1 sequences with NruI and transformed into C. albicans CAI4. Genomic DNA purified from Uri⁺ transformants was digested with SalI, ligated, and transformed into E. coli DH5α. The resulting plasmid, pGEM-HIS, had a 2.1-kbp insert with a complete HIS1 gene (Fig. 1B).

(iii) pRS-ARG4 and derivatives.

A 2.3-kbp ARG4 fragment was amplified by PCR from strain CAI4 template DNA with primers Arg4-N2K and Arg4-CS. The gel-purified PCR product was cloned into plasmid pRS314 after digestion of both PCR product and plasmid with XmaI and Asp718. The resulting plasmid was called pRS-ARG4. Sequences outside of ARG4 were deleted by SpeI digestion of pRS-Arg and ligation to yield plasmid pRS-Arg4ΔSpeI (Fig. 1C).

(iv) pRS-ArgBlaster.

Plasmid pRS-Arg4 was digested with XbaI and religated to yield plasmid pRS-Arg4ΔXbaI. Plasmid pRS-Arg4ΔXbaI was digested with BglII and HindIII, thus releasing a fragment with a 3′ portion of ARG4 and some downstream sequences. The digested plasmid was ligated together with a 4.1-kbp BglII-HindIII fragment from plasmid pMB7 (2) containing a hisG-URA3-hisG cassette. The resulting plasmid, pRS-ArgBlaster, was digested with Asp718 and SmaI for transformation into C. albicans.

PCR amplification.

For gene disruption, typical PCR mixtures contained 1 μl of quick-prep template DNA (plasmid pGEM-HIS1, pGEM-URA3, or pRS-Arg4ΔSpeI), 2 μl of a 5 μM stock of each (5DR and 3DR) primer, 10 μl of 10× PCR buffer (Boehringer), 10 μl of a mixture of 2.5 mM deoxynucleoside triphosphates (Boehringer), 2 μl of 0.1 M MgCl₂, 75 μl of water, and 0.5 μl of Taq DNA polymerase (Boehringer or Sigma) (1 U/μl). The mixture was overlaid with mineral oil and incubated at 94°C for 5 min followed by 30 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 4 min. After a final extension at 72°C for 8 min, the reaction mixture was stored at either 4°C or −20°C before further use. Generally, 5 μl of the reaction mixture was examined on an agarose gel to confirm the presence of a product of expected size.

For analytical PCRs, 1 to 5 μl of C. albicans DNA, prepared by glass bead lysis (5), served as a template in a PCR. Primers 5-detect and 3-detect (Table 1) were used to detect vector sequences flanking a marker. Primers RIM101-5a and seq7 were used to detect integration at HRM101. Primers PalA5′ and PalA3′ were used to detect integration at ENX3.

C. albicans transformations.

For most experiments, we used the method of Braun and Johnson (3). Overnight YPD+Uri cultures were inoculated at a 1/100 dilution into 50 ml of YPD+Uri and incubated at 30°C for two generations (around 4 h). Cells were pelleted, washed with 5 ml of LATE buffer (0.1 M lithium acetate, 10 mM Tris HCl [pH 7.5], 1 mM EDTA), and suspended in 0.5 ml of LATE buffer. Cell suspension (0.1 ml) was mixed with 5 μl of 10-mg/ml calf thymus DNA (Sigma) and transforming DNA and incubated at 30°C for 30 min. Then, 0.7 ml of PLATE buffer (40% polyethylene glycol 3350 in LATE buffer) was added and the tube was vortexed for 2 s. After overnight incubation at 30°C, the mixture was heat shocked for 1 h at 42°C. Cells were pelleted, washed with 1 ml of TE (10 mM Tris HCl [pH 7.5], 1 mM EDTA), suspended in 0.2 ml of TE, and plated on a selective medium. For transformations with PCR products, we used 80 μl of the PCR mixture per transformation. For transformations with linearized plasmids or restriction fragments, we used 2 to 10 μg of digested DNA per transformation.

For disruption of ENX3, we used a Frozen-EZ Yeast Transformation II kit (Zymo Research) and followed the supplier’s instructions.

Identification of HRM101 and ENX3 sequences.

HRM101 sequences were identified through a tblastn search of Candida sequences (3b) with the S. cerevisiae Rim101p protein sequence (20) as a query. Two individual homologous sequences were aligned (Geneworks) to assemble a 627-bp segment. To identify ENX3, we scanned the C. albicans gene list (4a) for genes involved in pH regulation; ENX3 corresponded to a 753-bp segment.

Southern analysis.

Genomic DNA was prepared by glass bead lysis (5). Southern analysis was carried out as described previously (6) with randomly primed probes. The ARG5 probe was a 700-bp ClaI-ScaI fragment of plasmid pUC-ARG-U (12). The ADE2 probe was an 800-bp EcoRI-KpnI fragment from plasmid pRS314-ADE, which contains a 1,587-bp XbaI-EcoRV ADE2 fragment (18, 21), obtained by PCR amplification, in vector pRS314 (19).

RESULTS

PCR product-directed disruption of known genes in C. albicans.

We first used the ARG5 gene in assays for homologous integration in C. albicans of PCR products with short regions of homology. An arg5 homozygous mutant has an Arg⁻ phenotype (12). Our strategy was to transform an arg5::hisG/ARG5 heterozygote (strain Arg-het1) with a PCR product that was capable of homologous integration into the functional ARG5 allele; homology to ARG5 was specified only by synthetic primer sequences (primers ARG5-5DR and ARG5-3DR [Fig. 1F]). We assayed homologous integration through identification of Arg⁻ transformants. The PCR product included URA3 sequences for selection and 60 bp of homology at each end to the ARG5 allele; it had no homology to the arg5::hisG allele (Fig. 1D and F). Transformation of the PCR product into strain Arg-het1 yielded 18 Uri⁺ transformants (Table 2), two of which were Arg⁻. These results indicated that 60 bp of homology on each end of a PCR product was sufficient to direct homologous integration at ARG5.

TABLE 2.

Summary of PCR product transformation experiments

Expt	Recipient strain (genotype)	PCR product in transformation	Total no. of transformantsa	No. of integrants at wild-type allele of targeted locus	Screening methodb	Comments
ARG5 disruption	Arg-het1 (arg5::hisG/ARG5)	arg5::URA3	18	2	Arg− phenotype
ARG5 disruption	RM1000 (ARG5/ARG5)	arg5::HIS1	24	4	Southern
	Arg-het2 (arg5::HIS1/ARG5)	arg5::URA3	27	1	Arg− phenotype & Southern	21 transformants were His− and Arg+
ADE2 disruption	RM1000 (ADE2/ADE2)	ade2::HIS1	19	3	Southern & PCR
	Ade-het1 (ade2::HIS1/ADE2)	ade2::URA3	40	3	Ade− phenotype & Southern	12 transformants were His− and Ade+
HRM101 disruption	BWP17 (HRM101/HRM101)	hrm101::ARG4	11	2	PCR
	Hrm-het1 (hrm101::ARG4/HRM101)	hrm101::URA3	14	2	PCR	Arg− transformants were not screened
	Hrm-het2 (hrm101::ARG4/HRM101)	hrm101::URA3	21	1	PCR	Arg− transformants were not screened
ENX3 disruption	BWP17 (ENX3/ENX3)	enx3::URA3	5	2	PCR
	Enx-het1 (enx3::URA3/ENX3)	enx3::ARG4	6	2	PCR	Selection eliminated Uri− transformants

^aResults from a single transformation are reported for each disruption. 

^bDetails of screening methods are described in Materials and Methods and in Results. 

To disrupt ARG5 through the sole use of PCR products, we used a ura3/ura3 his1/his1 double auxotroph (strain RM1000) in successive transformations with PCR products specifying arg5::HIS1 and arg5::URA3 (Fig. 2). PCR products were synthesized with a single pair of primers that anneal to plasmid sequences flanking the HIS1 and URA3 templates (Fig. 1F). Transformation of the arg5::HIS1 PCR product into strain RM1000 yielded 24 His⁺ transformants (Table 2), four of which had incorporated the HIS1 gene into one copy of ARG5, as indicated by two lines of evidence. First, Southern analysis of these four transformants showed that one copy of ARG5 had a novel BamHI site, as expected from the arg5::HIS1 disruption (as shown for one transformant in Fig. 3A by comparison of lanes 1 and 2). Second, PCR analysis of genomic DNA indicated that only these four transformants had retained vector sequences flanking the HIS1 cassette (data not shown). The other His⁺ transformants may have replaced one his1::hisG allele with HIS1 sequences and thus lost vector sequences. Transformation of one arg5::HIS1/ARG5 heterozygote (strain Arg-het2) with the arg5::URA3 PCR product yielded 27 Uri⁺ transformants (Table 2 and Fig. 2), one of which was Arg⁻. Southern analysis confirmed that the Arg⁻ transformant carried one copy of arg5::HIS1, one copy of arg5::URA3, and no copies of ARG5 (Fig. 3A, lane 3). Among the other Uri⁺ transformants, 21 remained Arg⁺ but had become His⁻. Southern analysis indicated that the arg5::URA3 PCR product had replaced the arg5::HIS1 allele in two of the transformants that we examined (Fig. 3A, lanes 4 and 5). Thus the majority of transformants had incorporated the arg4::URA3 PCR product by homologous recombination. These results show that an arg5/arg5 homozygous mutant was created through homologous integration directed by PCR primer sequences.

FIG. 2.

Schematic diagram of de novo ARG5 disruption. Each pair of parallel lines represents the two ARG5 alleles. Successive transformations are indicated by arrows. Strain RM1000 was transformed with an arg5::HIS1 PCR product; 4 of 24 transformants were arg5::HIS1/ARG5 heterozygotes, as indicated by Southern analysis. One heterozygote was transformed with an arg5::URA3 PCR product; 1 of 27 transformants was an arg5::HIS1/arg5::URA3 homozygous mutant, as indicated by Southern analysis and its Arg⁻ phenotype. Twenty-one of the remaining transformants were arg5::URA3/ARG5 heterozygotes, as indicated by their Arg⁺ His⁻ phenotype and, for a representative group, by Southern analysis.

FIG. 3.

Analysis of transformant strains. (A) Southern blot analysis of ARG5 disruption experiment. Genomic DNA samples, digested with BamHI, were prepared from strain RM1000 (ARG5/ARG5 [lane 1]), Arg-het1 (arg5::HIS1/ARG5 [lane 2]), a Uri⁺ His⁺ Arg⁻ transformant from Arg-het1 (arg5::HIS1/arg5::URA3 [lane 3]), and two Uri⁺ His⁻ Arg⁺ transformants from Arg-het1 (arg5::URA3/ARG5 [lanes 4 and 5]). The blot was probed with a 0.7-kbp ClaI-ScaI fragment of ARG5, as indicated in Fig. 1D. (B) Southern blot analysis of ADE2 disruption experiment. Genomic DNA samples, digested with BglII and HindIII, were prepared from strain RM1000 (ADE2/ADE2 [lane 1]), Ade-het1 (ade2::HIS1/ADE2 [lane 2]), a Uri⁺ His⁺ Ade⁻ (red) transformant from Ade-het1 (ade2::HIS1/ade2::URA3 [lane 3]), and two Uri⁺ His⁻ Ade⁺ transformants from Ade-het1 (ade2::URA3/ADE2 [lanes 4 and 5]). The blot was probed with a 0.8-kbp EcoRI-EcoRV fragment of ADE2, as indicated in Fig. 1E. (C) PCR analysis of HRM101 disruption experiment. PCR amplification with primers RIM101-5a and seq7 was carried out on genomic DNA samples from strain BWP17 (HRM101/HRM101 [lane 1]), Hrm-het1 (hrm1::ARG4/HRM101 [lane 2]), Hrm-het2 (hrm1::ARG4/HRM101 [lane 3]), and Uri⁺ Arg⁺ transformants from Hrm-het1 and Hrm-het2 (hrm1::ARG4/hrm101::URA3 [strains BWP29, BWP30, and BWP31] [lanes 4 to 6]). (D) PCR analysis of ENX3 disruption experiment. PCR amplification with primers PalA5′ and PalA3′ was carried out on genomic DNA samples from strain BWP17 (ENX3/ENX3 [lane 1]), Enx-het1 (enx3::URA3/ENX3 [lane 2]), Enx-het2 (enx3::URA3/ENX3 [lane 3]), and two Uri⁺ Arg⁺ transformants from Enx-het1 (enx3::URA3/enx3::ARG4 [strains DAY23 and DAY24] [lanes 4 and 5]).

To see if this gene disruption strategy was applicable to other loci, we attempted to disrupt the ADE2 locus with ade2::HIS1 and ade2::URA3 PCR products (Fig. 1E). In this case, each PCR product was flanked by 50 bp of homology to ADE2. Transformation of the ade2::HIS1 PCR product into strain RM1000 yielded 19 His⁺ transformants (Table 2), 3 of which retained vector sequences flanking HIS1 (data not shown). A Southern blot verified that one transformant, Ade-het1, was of genotype ade2::HIS1/ADE2 (Fig. 3B; compare lanes 1 and 2). Transformation of the ade2::URA3 PCR product into Ade-het1 yielded 40 Uri⁺ transformants (Table 2), 3 of which were Ade⁻. Colonies of the Ade⁻ transformants were red in color, as expected for ade2/ade2 mutants (4). Southern analysis of one Ade⁻ transformant showed the presence of ade2::HIS1 and ade2::URA3 fragments and the absence of an ADE2 fragment (Fig. 3B, lane 3). Twelve transformants remained Ade⁺ but became His⁻; Southern analysis of two confirmed that ade2::URA3 had replaced the ade2::HIS1 allele (Fig. 3B, lanes 4 and 5). Thus, 15 of 40 transformants had acquired the ade2::URA3 PCR product through homologous recombination. These results show that ade2/ade2 homozygous mutants were created through homologous integration directed by PCR primers.

Primer-directed disruption of newly identified sequences.

We set out to test the hypothesis that a possible C. albicans homolog of RIM101 may have a role in filamentation. In S. cerevisiae, RIM101 (previously called RIM1) specifies a zinc finger protein (20) that is required for haploid invasive growth (8), a form of filamentation (17). We identified a C. albicans sequence that may specify a Rim101p homolog: the predicted translation product showed over 50% identity to the Rim101p zinc finger region. We refer to this putative C. albicans gene as HRM101 (for “homolog of RIM101”). The hypothesis that HRM101 and RIM101 are functional homologs predicts that hrm101 homozygous mutants may be defective in filamentous growth.

The two genomic copies of HRM101 were inactivated by primer-directed integration of hrm101::ARG4 and hrm101::URA3 PCR products in strain BWP17 (ura3/ura3 his1/his1 arg4/arg4). We first isolated 11 Arg⁺ transformants with an hrm101::ARG4 PCR product. PCR amplification with flanking HRM101 primers verified presence of the hrm101::ARG4 insertion in two of these transformants (Fig. 3C; compare lanes 2 and 3 to lane 1). We used these two hrm101::ARG4/HRM101 heterozygotes for transformation with a PCR product specifying hrm101::URA3 and obtained 35 Uri⁺ transformants. PCR amplification of genomic DNA templates with flanking HRM101 primers indicated that three transformants had only hrm101::ARG4 and hrm101::URA3 alleles and lacked the wild-type HRM101 allele (Fig. 3C, lanes 4 to 6). Therefore, we had obtained three independent hrm101::ARG4/hrm101::URA3 transformants.

To determine whether HRM101 may be required for filamentation, we examined filamentous growth on Spider medium. We observed that HRM101/HRM101 and hrm101::ARG4/HRM101 strains produced abundant filaments at the edges of colonies (Fig. 4A to C) and that hrm101::ARG4/hrm101::URA3 strains did not produce filaments (Fig. 4D to F). The filamentation defect was not simply a consequence of the auxotrophies in the transformants, because Arg⁺ Uri⁺ His⁻ derivatives of strain BWP17 that retained HRM101 alleles were capable of filamentation (data not shown). These observations suggest that HRM101 has a positive role in filamentation and that the hrm101::ARG4 mutation is recessive.

FIG. 4.

Filamentation on Spider plates. Colonies were incubated for 6 days on Spider plates at 37°C. Strains were BWP17 (wild type [A]), Hrm-het1 and Hrm-het2 (both hrm1::ARG4/HRM101 [B and C]), BWP29, BWP30, and BWP31 (all hrm1::ARG4/hrm101::URA3 [D to F]), Enx-het1 and Enx-het2 (both enx3::URA3/ENX3 [G and H]), and DAY23 and DAY24 (both enx3::URA3/enx3::ARG4 [I and J]).

In Aspergillus nidulans, the RIM101 homolog (PacC) lies in a pathway with several pal gene products, including PalA (13, 15). The C. albicans sequence ENX3 may specify a PalA homolog. If ENX3 and HRM101 lie in the same pathway, then enx3 homozygous mutants should be defective in filamentous growth. We created two enx3::URA3/ENX3 heterozygotes and, subsequently, two enx3::URA3/enx3::ARG4 homozygotes by primer-directed integration (Table 2). Genotypes were confirmed by PCRs with flanking primers (Fig. 3D). We observed that the enx3::URA3/enx3::ARG4 homozygotes were defective in filamentation on Spider medium (Fig. 4I and J); the enx3::URA3/ENX3 heterozygotes were not (Fig. 4G and H). These findings indicate that both HRM101 and ENX3 have positive roles in filamentation.

DISCUSSION

We describe here a rapid method for disruption of C. albicans genes. The method is based upon PCR primer-directed gene disruption in S. cerevisiae (1, 10) and upon the idea of using two markers to create homozygous disruptions in C. albicans (7, 12). We were able to disrupt four different genes with PCR products, so it is likely that the method will be applicable to many other loci.

We had expected that the ARG4 and URA3 cassettes would integrate more efficiently than HIS1 at targeted loci. The URA3 cassette in plasmid pGEM-URA3 lacks sequences to direct integration at the URA3 locus, because the entire URA3 locus has been deleted in strain CAI4 (2). The ARG4 cassette in plasmid pRS-ARG4ΔSpeI also lacks sequences to direct integration at the ARG4 locus, because it does not extend beyond one end of the arg4::hisG insertion/deletion allele in strain BWP17. On the other hand, the HIS1 cassette in plasmid pGEM-HIS1 extends far beyond the HIS1 coding region, so it should be capable of integration into the his1::hisG allele. However, targeted integration of hrm101::ARG4 and enx3::URA3 in strain BWP17 occurred with efficiency comparable to that of arg5::HIS1 and ade2::HIS1 in strain RM1000. These results are not directly comparable, because they involve different targeted loci and different transformation recipients. However, they suggest that all three markers may be equally useful for future gene disruption experiments.

The ARG4 and URA3 cassettes lack known sequences for homologous integration at each respective locus in strain BWP17, yet we recovered transformants that did not carry the markers at targeted loci. We are uncertain whether the PCR products integrate into the genome or are maintained in an extrachromosomal state. However, the markers are much more stable than the ARS plasmid pRC2312 (4) in our hands (unpublished results), so we believe that they are integrated. These observations underscore the importance of verifying targeted integration through either Southern analysis or PCRs with outside primers.

We observed a significant allelic integration bias in the second of successive transformations. For example, transformation of the arg5::HIS1/ARG5 strain with the arg5::URA3 PCR product yielded more frequent integration into the arg5::HIS1 allele than into the ARG5 allele. A similar bias, though less severe, was observed during disruption of ADE2. The bias may reflect a greater recombination efficiency between molecules that have more extensive homology: the ends of the arg5::URA3 PCR product have 160 and 210 bp of homology to the arg5::HIS1 allele but 60 bp (at each end) to the ARG5 allele. Also, our use of unpurified PCR primers for creation of disruption constructs may contribute to allelic integration bias: contaminating primers with 5′ truncations will yield PCR products with little or no homology to a wild-type allele. In practice, this problem is not a significant impediment because undesired integrants may be eliminated by selection. Where cost is not a factor, the problem might be eliminated entirely through use of nonhomologous primers for amplification of each disruption cassette.

The main value of a rapid gene disruption method in C. albicans is to provide functional information about a sequence before investing significant effort in its characterization. In this study, we examined the function of two C. albicans genomic segments that were identified through Blast searches of the genomic sequence database. They were relatively short for traditional gene disruption strategies (HRM101, 627 bp; ENX3, 753 bp); in addition, both sequence records include uncertain nucleotides, so their isolation by PCR amplification might be difficult. Both sequences are also too short to specify an entire protein. By primer-directed disruption, we created homozygous mutant strains within 4 weeks of sequence identification. Our hypothesis predicted that hrm101 and enx3 homozygotes would have similar phenotypes, and this prediction was verified. However, we might have found that enx3 and hrm101 homozygotes have different phenotypes, thus suggesting that the respective gene products have distinct physiological functions. We might have found that one or both homozygotes have no detectable phenotype, thus suggesting that the sequences may be pseudogenes, that they may have functional homologs elsewhere in the C. albicans genome, or that our phenotypic assays are too crude to detect their function. We might have been unable to isolate homozygous mutants, thus suggesting that the respective gene products may be essential for growth, essential for recovery from transformation, or perhaps essential for completion of recombination. Each of these possible outcomes would affect our priorities for further characterization of a possible RIM101 pathway in C. albicans. Also, the outcomes might change our view of the RIM101 pathway in S. cerevisiae (8). Thus it is extremely valuable to be able to characterize a homozygous mutant at the start of a research effort.

Our information about HRM101 and ENX3 function is preliminary, because we have not complemented or reverted the defects. It is formally possible that the phenotypes of mutant strains arise from coincident mutations that were inadvertently isolated during transformation. However, the finding that three independent hrm101/hrm101 homozygotes have a filamentation defect supports the idea that the phenotype arises from the hrm101 mutation, not from a coincidental secondary mutation. A similar argument applies to the two enx3/enx3 homozygotes. These observations provide preliminary support for the idea that HRM101 and ENX3 sequence fragments are parts of genes, that they specify products, and that Hrm101p and Enx3p may have related functions.