EFG1 Null Mutants of Candida albicansSwitch but Cannot Express the Complete Phenotype of White-Phase Budding Cells

C. albicans wild-type strain WO-1 (30) was maintained on agar containing modified Lee's medium (6). Strain Red 3/6, an ade2auxotroph (38), and strain TS3.3, a ura3auxotroph (Table 1), were maintained on agar containing modified Lee's medium supplemented with 0.6 mM adenine and 0.01 mM uridine, respectively. EFG1 mutant strains were maintained on agar containing modified Lee's medium.

We originally set out to clone gene homologs in C. albicans of the APSES family of transcription factors (4) that included Phd1p (11), StuAp (21), and Sok2p (48). Two degenerate primers, P1 and P2, spanning common coding regions derived from Phd1p (11), StuAp (21), and Sok2p (48), were used to amplify a DNA fragment of approximately 380 bp encompassing the conserved region of these genes. The PCR-derived fragment was used to screen a λEMBL3A genomic library ofC. albicans WO-1 (40). Of approximately 50,000 plaques screened, 50 putative lambda clones were identified. Southern analysis with the DNA probe was used to select two lambda clones, λ14.1 and λ39.1, which contained approximately 10 and 12 kb of insert DNA, respectively. Partial sequence analysis demonstrated that both contained the EFG1 open reading frame (ORF) and flanking sequences. To isolate the 5′ flanking region ofEFG1, the λ14.1 and λ39.1 clones were screened by a gene-walking strategy by using the primer EC1 in combination with either the lambda left-arm-specific primer ELA or the lambda right-arm-specific primer ERA (Table 2). DNA fragments encompassing the 5′ upstream region of EFG1were obtained from λpH14.1 and λpH39.1 by PCR with the primers EC8, a sequence spanning −1 to −21 bp of EFG1, and ELA (Table2). The fragments were cloned into pGEM-5Z at the EcoRV site, generating the plasmids pTET14.1 and pB69.11, respectively. Both were sequenced in both directions by using an ABI model 373A automatic sequencing system and fluorescent Big Dye terminator chemistry (Perkin-Elmer–Applied Biosystems Inc., Foster City, Calif.).

Total RNA was extracted by methods previously described (41). Poly(A)⁺ mRNA was extracted by using the Oligotex Spin Column Kit according to manufacturer's specifications (Qiagen Inc., Santa Clarita, Calif.). RNA was separated in agarose formaldehyde gels, transferred to Zetabind nylon membranes, and hybridized with the appropriate probes (17). Prehybridization and hybridization procedures were performed by the methods of Church and Gilbert (8). Autoradiography was performed by exposing membranes at −70°C by using intensifying screens and Kodak XAR film (Eastman Kodak Co., Rochester, N.Y.). To measure the fold difference in transcript levels, a Northern blot containing a series of concentrations of total cell RNA ranging from 0.05 to 5.00 μg was hybridized with a radiolabeledEFIα2 gene probe (44), and the autoradiogram was digitized into the DENDRON program database (Solltech Inc., Oakdale, Iowa). Band intensities were measured and then used to generate a plot of measured intensity versus RNA concentration. Fold differences between Northern blot hybridization bands were then computed from the standards plot.

To confirm the configurations of either the EFG1 or URA3 locus in heterozygotes and null mutants by Southern blot analysis, DNA was extracted by methods previously described (24, 38). DNA (3 μg) was digested with the restriction enzymes described in Results, and the resulting fragments were separated in a 0.8% (wt/vol) agarose gel. The fragments were transferred to Hybond-N nylon membrane (Amersham International, Little Chalfont, Buckinghamshire, England) and hybridized with the DNA probes described in Results. Prehybridization and hybridization procedures were performed by the methods of Church and Gilbert (8), and autoradiography was performed as described above for Northern analyses.

Five micrograms of white- and opaque-phase poly(A)⁺ mRNA were converted into cDNA pools by using the Superscript Choice system and were directionally cloned into ZipLox lambda vector according to the manufacturer's specifications (Life Technologies, Gaithersburg, Md.). The titers of the unamplified white- and opaque-phase-specific libraries were 1 × 10⁶ and 2 × 10⁶, respectively. Approximately 10⁵ plaques of each unamplified library were screened, using the PCR-derived 1.7-kb EFG1 ORF as a probe. Thirty independent clones were identified in each case. Each clone was subjected to a second screen. The DNA from each of 10 positive white- and opaque-phase lambda clones obtained in the second screen was digested with EcoRI, and Southern blots were probed with theEFG1 ORF to confirm that they encoded EFG1. The three largest white- and opaque-phase ZipLox clones were chosen and converted into plasmid derivatives according to the manufacturer's protocol (Life Technologies). The three white-phase-specific cDNAs were named EFW1.1, EFW2.1, and EFW4.1, and the three opaque-phase-specific cDNAs were named EFO1.1, EFO3.1, and EFO5.1. The six selected cDNAs, each larger than 2 kb, were sequenced in both directions as previously described.

To compare the 5′ untranslated sequences of white- and opaque-phaseEFG1 mRNAs, 1 μg of poly(A)⁺ mRNA from each phase was subjected to 5′ rapid amplification of cDNA ends (RACE) analysis using the 5′ RACE kit protocol (Life Technologies). For first-strand cDNA synthesis, two different EFG1-specific primers were used in order to confirm that the derived products were from the same mRNA species. The first-strand-specific primers were EC3, spanning the 3′ end of the EFG1 mRNA, 140 bp downstream from two tandem TAA stop codons, and EC1, spanning the 5′ end of theEFG1 mRNA 210 bp downstream of the ATG start codon (Table2). Following first-strand synthesis and dG tailing, double-stranded 5′ RACE products were generated by a high-fidelity PCR protocol (Roche Biochemicals, Indianapolis, Ind.) by using either the ECI primer for EC3-derived template, or the EC8 primer for EC1-derived template (Table 2). After confirming by Southern analysis that the 5′ RACE products were of predicted molecular sizes, we subcloned these products into pGEM-5Z (Promega Corp., Madison, Wis.) at theEcoRV site in order to determine their nucleotide sequences. Plasmid clones were named pB59W.11 and pB870.1 for white-phase- and opaque-phase-specific 5′ RACE inserts, respectively. Sequences of the two cloned 5′ RACE inserts were determined as described in a previous section.

The 1.2-kb 5′-upstream region of the EFG1 ORF was inserted at the multiple cloning site immediately upstream of the Renillaluciferase (RLUC) ORF in the reporter plasmid pCRW3 (38). Integration was targeted to the ADE2 locus by linearizing the plasmid at an NsiI site in the ADE2 gene or to the EFG1 locus by linearizing at an HpaI site in the EFG1 promoter. Linearized plasmids were used to transform strain Red 3/6 by using the lithium acetate method (27). Five transformants of each construct were analyzed by Southern blot hybridization to confirm both the site of insertion and multiplicity of integration. Transformant clones harboring a single copy of the targeted plasmid in the correct location were chosen for further analysis. Measurements of RLUC activity were made according to methods previously described (38).

The original plasmid p1164 containing the C. albicans URA3 gene was kindly provided by Stewart Scherer of Acacia Biosciences, Richmond, Calif. This plasmid contained a 4.2-kb DNA insert, which included at least 1 kb of DNA flanking theURA3 gene. The DNA fragment was subcloned at theEcoRV site of pGEM-5Z (Promega Corporation), and the resulting plasmid clone was designated p161. In order to construct aURA3 deletion cassette, p161 plasmid DNA was digested withEcoRV and XbaI to delete a central 2.0-kb fragment of the URA3 gene. Following digestion, the plasmid was end-repaired with T4 DNA polymerase and was treated with shrimp alkaline phosphatase. The deleted URA3 gene fragment of the plasmid was then replaced by a 2.4-kb EcoRV fragment of theC. albicans ADE2 gene. The ADE2 DNA was derived from pMC2 (16) as an EcoRV fragment and was subcloned into pGEM-5Z to derive pADE2-5Z. The plasmid containing the URA3 deletion cassetteURA3::ADE2 was designated p161ΔURA3:ADE2. Homozygous deletion of theURA3 gene was performed in strain Red 3/6, anade2 derivative of WO-1 (39). Approximately 50 μg of ApaI/SacI-digested p161ΔURA3:ADE2 DNA was used to transform Red 3/6 by the lithium acetate method (27).ADE2 prototrophic clones were chosen based on their capacity to grow on minimal medium containing no adenine sulfate.URA3 heterozygotes were identified by Southern analysis of genomic DNA probed successively with a 2.4-kb EcoRV fragment of the ADE2 gene (16) and the 0.34-kbXbaI-NsiI fragment of the URA3 gene from the p161 plasmid. Selected heterozygote(s) were subjected to a second round of transformation in order to increase the chances of obtaining homozygotes by integrative gene conversion rather than mitotic recombination. However, because the cassettes used in the first and second transformations were identical, we could not discriminate between the two mechanisms. To generate ura3⁻homozygotes, heterozygous clones were grown to mid-log phase, then were inoculated into fresh yeast extract-peptone-dextrose broth and allowed to grow for one generation. Approximately 4 × 10⁷cells were transformed with 50 μg of the URA3 deletion cassette DNA by the spheroplast method (38). Following transformation, 10⁷ spheroplasts were spread on yeast extract-peptone-dextrose plates containing 1 M sorbitol for 16 h at 30°C. Cells were collected, and approximately 10⁷cells were spread on minimal medium containing 1 mg of 5-fluororotic acid (FOA) per ml and 0.1 mM uridine. These plates were incubated for 4 to 5 days at 30°C for the appearance of FOA-resistant colonies. Thirty independent colonies were tested by Southern analysis for homozygosity of the URA3 locus by using the ADE2and URA3 probes employed in the analysis of heterozygotes. Of 24 clones exhibiting identical patterns and absence of theURA3 region spanning the XbaI-EcoRV restriction sites (9), two clones, TS3.3 and TS3.5, which underwent the white-opaque transition at normal frequencies and were incapable of growing in medium lacking uridine were selected.

A hisG-URA3-hisG-based cassette (9) was used to create an EFG1 heterozygote in the first round of transformation. To accomplish this, the expression plasmid containing the EFG1 ORF, pEF1α2:EFG1 (Fig.1B), was digested with DraIII to delete 740 bp of the EFG1 ORF (Fig. 1A), was end-repaired with T4 DNA polymerase, and was dephosphorylated by using shrimp alkaline phosphatase. The deleted fragment was substituted with the 4.0-kb hisG-URA3-hisG cassette from pMB7 (9) (Fig. 1B). The hisG-URA3-hisG cassette was prepared by digesting the pMB7 plasmid DNA with SalI/BglII, followed by end-repair using T4 DNA polymerase. Since no suitable restriction enzyme sites flanked the deletion cassette for isolation from the plasmid, a high-fidelity, long PCR protocol (Boehringer Mannheim, Indianapolis, Ind.) with the EFG1-specific primers EC2 and EC3 (Table 2) was used to isolate the cassette for transformation. Approximately 10 μg of PCR-generated cassette DNA was used to transform the ura3 strain TS3.3 (Table 1). All of the recovered clones from selection plates were tested for heterozygosity by digesting the total genomic DNA withBamHI, followed by Southern blot hybridization with theEFG1 ORF and a URA3 gene fragment. After confirming heterozygosity of the targeted EFG1 locus, the clones were subjected to 5-FOA treatment in order to induce “pop outs” of the URA3 gene. For the second round of transformation, a different disruption cassette was constructed. DNA of plasmid pB21.2, containing the EFG1 ORF (Fig. 1C), was digested with NsiI to remove 620 bp of DNA, end-repaired with T4 DNA polymerase, dephosphorylated with shrimp alkaline phosphatase, and substituted with approximately 3.5 kb of a CAT-URA3-CAT cassette from the plasmid pCUC (a gift from Paul T. Magee, University of Minnesota). CAT-URA3-CAT DNA was prepared by digesting pCUC plasmid DNA with BamHI and was end-repaired with T4 DNA polymerase. The derived plasmid was named pB45.1 (Fig. 1C). The disruption cassette used for the second round of transformation was obtained by digesting pB45.1 with the restriction enzyme DraIII, resulting in a digestion fragment with homologous ends that could only target to the functional undeleted chromosomal allele (Fig. 1C). Clones recovered on selection plates were tested for homozygosity by Southern analysis as described earlier.

Cells were washed in double-distilled H₂O, fixed in 2.5% (wt/vol) gluteraldehyde in 0.1 M cacodylate buffer for 1 h, and postfixed in 1% osmium tetroxide in 0.1 M cacodylate buffer for 50 min. Cells were then washed three times in 0.1 M cacodylate buffer and treated with 6% thiocarbohydrazide at room temperature, followed by a second fixation in 1% osmium tetroxide to enhance surface architecture. Cells were then rinsed in double-distilled water, dehydrated through a graded series of ethanol solutions, dried, mounted on aluminum stubs, and sputter-coated with gold palladium. Samples were scanned with a Hitachi S-4000 scanning electron microscope (Hitachi Corp., San Diego, Calif.).

Northern blots of total cellular RNA from white- and opaque-phase cells grown at 25°C were probed with a full-length EFG1 ORF DNA fragment. White-phase-cell RNA contained an EFG1 transcript of approximately 3.2 kb (Fig. 2A). Opaque-phase-cell RNA contained a single, less-abundant EFG1 transcript of approximately 2.2 kb (Fig. 3A). This result was obtained in every case with five additional RNA preparations from independent white- and opaque-phase clones. Because the 3.2-kb hybridization band possessed a hybridization tail in every one of the six northern blot hybridizations performed with white-phase RNA, the possibility existed that a faint 2.2-kb band may have been missed in Northern blots of white-phase RNA. This possibility was resolved when Northern blots of purified poly(A)⁺ mRNA were probed with the full-length EFG1 ORF DNA fragment. White-phase cell poly(A)⁺ RNA contained a 3.2-kb transcript and no 2.2-kb transcript, while opaque-phase cells contained a less-abundant 2.2-kb transcript and no 3.2-kb transcript (Fig. 2B). The ratio of the white-phase EFG1 3.2-kb transcript to the opaque-phase EFG1 2.2-kb transcript was estimated by densitometric analyses to be approximately 20 to 1 for both total RNA and poly(A)⁺ RNA.

Since Northern blots of poly(A)⁺ mRNA probed with EFG1revealed white- and opaque-phase EFG1 transcripts of different molecular sizes, the possibility was entertained thatC. albicans contained more than one EFG1 gene. The Southern blot hybridization patterns of DNA digested withBamHI, BglII, HindIII,NsiI, and XhoI (data not shown) were consistent with the physical map of the cloned EFG1 locus (Fig. 1A), suggesting that only one copy of EFGI existed in the WO-1 genome, as previously demonstrated for C. albicans strain CAI4 (19).

Two clones from a white-phase-specific cDNA library and two clones from an opaque-phase-specific cDNA library that were greater than 2 kb were isolated and sequenced. All four contained identical ORFs of 1,662 bp beginning with ATG, and two consecutive TAA stop codons. The ORF contained 554 amino acids, compared to 552 reported earlier (43, 44), and exhibited six additional amino acid differences. The two independent white-phase cDNA clones contained long untranslated 5′ regions of 370 and 380 bp, respectively. The length of the untranslated 3′ regions of both white-phase cDNA clones was 407 bp, and the lengths of the poly(A)⁺ stretches were 60 and 70 bp, respectively. The two independent opaque-phase cDNA clones contained short untranslated 5′ regions of 40 and 50 bp, respectively, which contrasted with the long 5′ regions of the white-phase cDNAs. The length of the untranslated 3′ region of both opaque-phase cDNA clones was 407 bp, identical to that of the white-phase cDNA clones. However, the poly(A)⁺ stretches were 15 and 20 bp, approximately one-third that of white-phase cDNA clones. Although this cDNA analysis demonstrated a moderate length difference at the 5′ untranslated ends of white- and opaque-phase EFG1 transcripts, the difference was not sufficient to account for the size differences of the transcripts demonstrated in Northern blots (Fig. 2).

To compare more precisely the 5′ ends of the white- and opaque-phaseEFG1 transcripts, 5′ RACE analysis (10) was performed with purified poly(A)⁺-containing white-phase and opaque-phase mRNA. Sequence analysis of the 5′ RACE products of two independent white-phase RNA preparations revealed 5′ untranslated regions of 1,126 and 1,173 bp in length. The comparable 1,126-bp sequences were identical. Based upon these lengths, the predicted size of the white-phase EFG1 mRNA was approximately 3.3 kb, close to the size estimated from the Northern blots in Fig. 2. Sequence analysis of the 5′ RACE products of two independent opaque-phase RNA preparations revealed 5′ untranslated regions of 145 and 162 bp in length. The comparable 145-bp sequences were identical. Based upon these lengths, the predicted size of the opaque-phaseEFG1 mRNA was approximately 2.1 kb, close to the size estimated from the Northern blots in Fig. 2. The comparable 5′ transcribed untranslated sequences of the white- and opaque-phaseEFG1 transcripts were identical. The sequence of the longest 5′ RACE product of a white-phase RNA preparation is presented with the predicted transcription start sites for both the white- and opaque-phase transcripts in Fig. 3.

To confirm that the sequences of the 5′ RACE products were included in the 5′ untranslated region of white- and opaque-phase EFG1mRNAs, Northern blots of total RNA of white- and opaque-phase cells were individually probed with a DNA fragment spanning the proximal (3′) portion (−1 to −209 bp relative to ATG) (Fig. 3) and the middle portion (−210 to −877 relative to ATG) (Fig. 3) of the 5′ transcribed untranslated EFG1 sequence. The proximal probe hybridized with both the white- and opaque-phase mRNAs (data not shown). The middle portion probe hybridized with the white-phase mRNA, but not the opaque-phase mRNA (data not shown). These results confirm that the 5′ RACE products of white- and opaque-phase cells accurately represented the 5′ upstream untranslated regions of the respective mRNAs.

To sequence the EFG1promoter, a lambda clone, λEF39.1, was identified that contained a 6-kb sequence upstream of the EFG1 ORF. The 1.2-kb nucleotide sequence was compared both to the S. cerevisiaedatabase and the global eukaryotic promoter database (SCPD; Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.). Two TATA box protein binding motifs (7) were identified, between nucleotides −1227 and −1220 (TBP-box2) and between nucleotides −1737 and −1731 (TBP-box1). A binding site for the repressor-activator protein RAPI (28) was identified between −1491 and −1480 bp. A J-chain variable-repeat sequence was identified between −1257 and −1249 bp. Seven TGANTN binding sites for the transcription factor GCN5 were identified in the 456-bp region immediately upstream of the untranslated region. Finally, a consensus binding site for a heat shock transcription factor (45) was identified between −2201 and −2191 bp. In the region upstream of the 5′ untranslated region of the opaque-phase EFG1 transcript, two additional TATA box protein binding motifs were identified between −245 and −253 bp (Fig.3).

To demonstrate that the differences in the level of white- and opaque-phase mRNA expression observed in Northern analyses (Fig. 2) were in fact regulated by 5′ upstream promoter sequences, the 1,369-bp region upstream of the ATG start site for EFG1 translation was inserted upstream of the RLUC gene in the reporter plasmid pCRW3 (38) to generate the plasmid pEPB26. pEPB26 integration was first targeted to the ADE2 locus (39), and integration at that locus was demonstrated by Southern analysis (data not shown). In this case, RLUC activity was approximately 33-fold higher in the opaque phase than in the white phase (Table3), the reverse of the result one might predict from Northern analysis (Fig. 2). Next, pEPB26 integration was targeted to the resident EFG1 locus, and integration at that locus was demonstrated by Southern analysis (data not shown). Integration at the EFG1 locus restored the complete EFG1 promoter upstream of the RLUC gene. In this case, RLUC activity was approximately 38-fold higher in the white phase than in the opaque phase (Table 3), which was the expected result. The level of luciferase activity for cells in the white phase was approximately 737 times higher in the strain in which pEPB26 was integrated at theEFG1 locus than in the strain in which it was integrated at the ADE2 locus (Table 3). However, the luciferase activity for cells in the opaque phase were similar in the two strains (Table3). These results demonstrate that in order to express the more abundant white-phase-specific transcript, sequences upstream of the 5′-untranslated region of the white-phase EFG1 transcript are essential. However, this region is sufficient for opaque-phase-specific expression.

To directly assess the role of EFG1 in the white-to-opaque transition, we applied a urablast gene knockout strategy (9) for generating homozygous mutants in C. albicans WO-1. A ura3 strain, TS3.3, was first created as described in the Materials and Methods section. TS3.3 was transformed with the hisG-URA3-hisG-based EFG1Δ cassette, pA42.23 (Fig. 1B), and 24 transformant clones were obtained. One of the transformants, Ef3.1, was chosen and analyzed by Southern blot hybridization with the full-length EFG1 ORF probe. While Southern blots of TS3.3 digested with BamHI contained a single hybridization band at 4.5 kb, Southern blots of Ef3.1 digested with BamHI contained a 4.5-kb band and two additional bands of approximately 2.9 kb and 2.4 kb, resulting from the twoBamHI sites in the hisG cassette (Fig.4A). These two bands represent 5′ and 3′ flanking regions of the disrupted copy of the EFG1 allele. To induce URA3 auxotrophy, Ef3.1 was subjected to a 5-FOA pop out protocol (9) resulting in aura3⁻ auxotroph, Ef3.1.1. Ef3.1.1 was transformed with a CAT-URA3-CAT-based EFG1disruption cassette from pB45.1 (Fig. 2C), and 52 transformant clones were obtained. Three of the transformants, Efc20, Efc25, and Efc30, were analyzed by Southern blot hybridization with the full-lengthEFG1 ORF probe. The 4.5-kb band was absent in all three transformants (Fig. 4A). Instead, Efc20 contained a 7.5-kb band, the expected size for the disruption cassette, while Efc25 and Efc30 contained 9.5- and 10.5-kb bands, respectively (Fig. 4A). To test whether the larger bands in the latter two were due to gross rearrangements of the flanking 4.5-kb BamHI fragment or to internal duplication of the CAT-URA3-CAT cassette, Southern analyses were performed of genomic DNA digested with BamHI,BglII, ApaLI, and HphI, and were probed with sequences that flank the 5′ or 3′ ends of theEFG1 transcript. The patterns of the three EFG1null mutants were identical to those of the wild-type WO-1 and the parental strain TS3.3 (data not shown). Since the flanking regions hybridizing to the probes spanned at least 10 kb, we conclude that the increased sizes of the inserts in Efc25 and Efc30 were due to internal duplication of the CAT cassettes rather than to reorganization of flanking regions.

All three mutants also contained the 2.4- and 2.7-kb bands representing the first disrupted allele (Fig. 4A). When probed with theURA3 ORF, TS3.3 showed no hybridization, as expected, while Ef3.1 showed hybridization of a 2.9-kb fragment, demonstrating that one of the two alleles contained the hisG-URA3-hisG cassette (Fig. 4A). The three mutants, Efc20, Efc25, and Efc30, showed no hybridization at 2.9 kb and hybridization at 7.5 to 10.5 kb, demonstrating that the first allele had lost the URA3 gene through a pop-out and the second allele contained the CAT-URA3-CAT cassette (Fig. 4A). Efc25 also contained a 3.4-kb band of unknown origin (Fig. 4A).

When cells of strains Efc20, Efc25, and Efc30 were plated at 25°C on agar containing modified Lee's medium, 100% of the colonies exhibited an opaque-phase colony morphology. When plated on agar containing phloxin B, all colonies (100%) were bright red (data not shown), an opaque-phase-specific characteristic (3). To test whether mutant cells could be induced to convert en masse to the white-phase phenotype by a temperature shift (23, 26, 30, 40), parental, heterozygous, and mutant cells in the opaque phase were first grown at 25°C to mid-log phase in liquid cultures of modified Lee's medium, then shifted to 42°C and incubated at the latter temperature in the same medium for an additional 16 h. Cells were then plated at low density on agar plates containing modified Lee's medium plus phloxin B, and white- and opaque-phase colony phenotypes were counted. The parental strain TS3.3 formed 94 and 92% white-phase colonies in two independent experiments (Table 4), demonstrating temperature-induced mass conversion from opaque to white phase. The heterozygous strain Ef3.1 formed 77 and 79% white-phase colonies in two independent experiments (Table 4), demonstrating mass conversion again, but at a slightly reduced level. In contrast, the three mutant strains formed no white-phase colonies in two independent experiments (Table 4), demonstrating that in the absence of EFG1expression, cells did not express the white-phase colony phenotype.

The white-to-opaque transition in strain WO-1 involves a dramatic change in cellular phenotype (3, 30, 31). White-phase cells incubated at 25°C are round, produce round daughter cells, and exhibit a relatively homogeneous smooth surface. In contrast, opaque-phase cells incubated at 25°C are approximately twice the size of white-phase cells, are elongate or bean-shaped, and contain pimples on the cell membrane with central pores (3). When opaque-phase cells are transferred from 25 to 42°C and are incubated at the latter temperature for 16 h, they convert en masse to the white phase (23, 26, 30, 31, 40). The parent strain TS3.3 and the heterozygote Ef3.1 both formed smooth, round, budding cells in the white phase at 25°C, and pimpled, elongate, or bean-shaped cells in the opaque phase at 25°C (Fig. 5). When opaque-phase cells of strains TS3.3 and Ef3.1 were shifted from 25 to 42°C and were incubated at the latter temperature for 16 h, they converted en masse to the white phase, forming round, smooth cells characteristic of the white phase (Fig. 5). The three EFG1 null mutants Efc20, Efc25, and Efc30 all formed pimpled, elongate cells exclusively at 25°C, characteristic of the opaque phase (Fig. 5). These cells were indistinguishable from opaque-phase cells formed by strains WO-1, TS3.3, and Ef3.1 (Fig. 5). When opaque-phase cells of the three mutant strains Efc20, Efc25, and Efc30 were shifted from 25 to 42°C and were incubated at the latter temperature for 16 h, they formed cells with the smooth (i.e., devoid of pimples) surface of white-phase cells, but with the elongate shape of opaque-phase cells (Fig. 5).

The expression of several genes has been demonstrated to be phase specific in the white-to-opaque transition. While the gene WH11 is expressed exclusively by white-phase cells (40), the genes PEPI(SAPI) (24), OP4 (23), andCDR3 (5) are expressed exclusively in the opaque phase. Expression of WH11, PEPI(SAPI), and OP4 was examined in strain TS3.3, the heterozygote Ef3.1, and the three mutants Efc20, Efc25, and Efc30. In the white phase at 25°C, TS3.3 and Ef3.1 cells both expressedWH11, but neither expressed OP4 orSAPI. In the opaque phase at 25°C, cells of strains TS3.3, Ef3.1, Efc20, Efc25, and Efc30 expressed Op4 andPEP1, but did not express WH11.

Sixteen hours after opaque-phase cells of strains TS3.3 and Ef3.1 were transferred from 25 to 42°C to induce mass conversion, they no longer expressed OP4 and PEP1 (SAP1) but did express WH11 (Fig. 6). Sixteen hours after opaque-phase cells of mutant strain Efc20, Efc25, and Efc30 were transferred, they also no longer expressed OP4 andPEP1 (SAP1), but did express WH11(Fig. 6). However, the levels of WH11 transcript were severely reduced in all three mutants (Fig. 6).

When opaque-phase cells of strain WO-1 are shifted from 25 to 42°C, they semisynchronously undergo three cell doublings, at approximately 2, 4, and 6 h (23, 40) (Fig.7A). When cells are returned to 25°C during the first 3 to 4 h at 42°C, they continue to multiply in the opaque phase, forming daughter cells with pimples. However, when cells are returned to 25°C after 4 h at 42°C, they multiply in the white phase, forming smooth, round, white-phase daughter cells (23, 40). Shift experiments have, therefore, identified a phenotypic commitment event to the white-phase between 3 and 4 h (23, 40) (Fig. 7A). When opaque-phase cells of strain WO-1 are incubated at 42°C and then transferred back to 25°C at intervals prior to the commitment event, they do not express the white-phase-specific gene WH11 (44) (Fig. 7A). However, when shifted back to 25°C after the commitment event,WH11 is reexpressed (40) (Fig. 7A). When opaque-phase cells are incubated at 42°C, they stop expressing bothOP4 and PEP1 (SAP1) (23) (Fig. 7A). When opaque-phase cells are then shifted from 42 to 25°C prior to the commitment event, they immediately reacquire transcripts of both genes within 1 h (Fig. 7A). These results demonstrate that although expression of the two phase-specific genes are temperature-sensitive, they can still be rapidly reactivated prior to the commitment event. However, when opaque-phase cells are shifted from 42 to 25°C after the commitment event, neitherOP4 nor PEP1 (SAP1) are reexpressed (23) (Fig. 7A). These results demonstrate that at the time of temperature-induced commitment to the white phase, a switch-associated event blocks subsequent temperature-induced (42-to-25°C) reactivation of opaque-phase-specific gene expression (23) (Fig. 7A).

The phenotype observed after opaque-phase cells were shifted from 25 to 42°C (Fig. 5) suggested that the three EFG1 null mutants underwent at least a portion of the changes in cellular phenotype associated with the opaque-to-white-phase transition. If gene expression flipped in the normal fashion (Fig. 7A) in mutant cells after a shift to 42°C, this would add support to the conclusion that mutant cells undergo the switch event, but they cannot fully express the complete white-phase phenotype without EFG1 expression. When TS3.3 or Efc20 cells in the opaque phase were incubated at 42°C for 1 h, they stopped expressing OP4, and when subsequently returned to 25°C for 1 h, they reexpressedOP4 (Fig. 7B). In neither case did they expressWH11 (Fig. 7B). When opaque-phase cells of strains TS3.3 and Efc20 were incubated at 42°C for 7 h, they stopped expressingOP4, and when subsequently returned to 25°C for 1 h, they still did not reexpress OP4 (Fig. 7B). After 7 h of incubation at 42°C, WH11 was expressed, and when cells were transferred back to 25°C, expression was even greater. Therefore, both the parent TS3.3 and the EFG1 null mutant Efc20 conform to the regulation of WH11 and OP4gene expression previously described for the parental strain WO-1 (23, 40). These results demonstrate that cells lackingEFG1 undergo the changes in phase-specific gene expression that are associated with phenotypic commitment in wild-type cells.

Northern analysis of total cell RNA revealed a 3.2-kbEFG1 transcript in white-phase cells that was missing in opaque-phase cells. It also revealed a less-abundant 2.2-kbEFG1 transcript in opaque-phase cells. Densitometric measurements of the levels of the two transcripts revealed an approximately 20-fold difference in message levels, which may be the reason why Sonneborn et al. (36) did not detect the opaque-phase-specific transcript. Sequence analysis of 5′ RACE products demonstrated that the difference in the molecular sizes of the white- and opaque-phase EFG1 transcripts was due to differences in the transcription initiation sites. The transcription initiation site of the white-phase EFG1 transcript was approximately 1,173 bp upstream of the ATG translation start site, while that of the opaque-phase EFG1 transcript was approximately 162 bp upstream of the ATG translation start site. These results parallel those for the homologue StuA in A. nidulans (49). Two different promoters result in overlapping transcripts, designated StuAα and StuAβ, which encode the same protein (49).

The 1.2-kb 5′ sequence upstream of the ATG translation start site of EFG1 was tested for its ability to function as a promoter by placing it upstream of the Renilla reniformisluciferase gene in the plasmid pCRW3 (38). When integration of the resultant plasmid pEPB26 was targeted to the ADE2locus (39), luciferase expression was 33-fold higher in the opaque phase than in the white phase, which was opposite to expectations derived from Northern analysis. We previously reported that promoters of a variety of genes (OP4, WH11,GAL1, and EF1α2) functioned normally at this ectopic locus with luciferase activities correlating with transcript levels obtained by Northern analysis (38). However, when we targeted the reporter construct to the EFG1 locus, the promoter was reconstituted and luciferase expression was 38-fold higher in the white phase than in the opaque phase, which correlated with transcript levels revealed by Northern analysis. These results suggest that cis-acting regulatory sequences necessary for the synthesis of the higher-molecular-weight EFG1 transcript in the white phase are distal to the 1.2-kb sequence immediately upstream of the translation start site. However, the level of luciferase activity in the opaque phase was similar when integration was targeted to the ADE2 locus or the EFG1 locus, suggesting that the cis-acting regulatory sequences for opaque-phase expression of the low-molecular-weight EFG1 transcript reside in the 1.2-kb sequence immediately upstream of the translation start site. A detailed functional analysis of the EFG1promoter is now in progress.

A direct assessment of EFG1function in the white-opaque transition required the genesis of a homozygous disruptant. Three EFG1 null mutants uniformly formed phloxine-B-stained colonies at 25°C, which we interpreted to represent the opaque-phase phenotype (3). When cells from select colonies of the three mutants grown at 25°C were examined by scanning electron microscopy, in all cases they exhibited the elongate morphology of opaque-phase cells, and the majority of cells possessed pimples, which are a signature feature of the opaque-phase phenotype (3). Cells of all three mutants also expressed the opaque-phase-specific genes OP4 andPEP1(SAP1) and did not express the white-phase-specific gene WH11 at 25°C (32, 33). We initially interpreted these results to mean thatEFG1 null mutants were jammed in the opaque-phase phenotype. However, an analysis of phenotypic commitment during temperature-induced mass conversion from the opaque- to the white-phase phenotype (23, 26, 30, 40) proved otherwise.

When opaque-phase cells of strain WO-1, the ura3 parent strain TS3.3, and the heterozygote Ef3.1 were transferred from 25 to 42°C, they retained the capacity to multiply in the opaque phase for approximately 3 h then semisynchronously converted to white-phase growth (23, 40). The time at which 50% of cells underwent this conversion was 3.5 h, the average time of the commitment event for the differentiation from the opaque- to the white-phase phenotype. When cells of the three EFG1 null mutants were transferred from 25 to 42°C and examined after the expected time of phenotypic commitment, they formed daughter cells with the smooth wall of white-phase cells (i.e., no opaque-phase-cell pimples), but the elongate, bean-shaped morphology of opaque-phase cells. This result suggested that the temperature shift induced a switch, but mutant cells were unable to express the complete phenotype of white-phase budding cells. The cellular phenotype of temperature-shifted opaque phase cells of the three EFG1 null mutants was similar to that of a transformant of C. albicans strain CAI8 engineered to express very low levels of EFG1(36).

At the point of phenotypic commitment, wild-type cells also undergo dramatic changes in phase-specific gene expression (23, 40). When shifted to 42°C, transcription of the two opaque-phase-specific genes OP4 andPEPI (SAPI) immediately turns off. However, when cells shifted to 42°C are shifted back to 25°C prior to the commitment event, transcription of these genes is immediately resumed. After the point of commitment, however, a shift back to 25°C will not turn on transcription of these genes, demonstrating that a fundamental change occurs in the regulation of phase-specific genes at the commitment point. In support of this conclusion, transcription of the white-phase-specific gene WH11 turns on at the commitment event. All three null mutants underwent these “yin-yang” changes in gene expression at the commitment point when shifted from 25 to 42°C. Therefore, EFG1 null mutant cells in the opaque phase undergo temperature-induced phenotypic commitment, switching from an opaque-phase to a white-phase pattern of gene expression, and they undergo changes in wall morphology that include the loss of opaque-phase pimples. Since EFG1 is homologous to known transcription factors in both S. cerevisiae (11) and A. nidulans (21), it seems reasonable to suggest that it plays a similar role in C. albicans and directs the expression of a subset of white-phase-specific genes necessary for the genesis of a round, white-phase budding cell phenotype. EFG1 expression, therefore, is not integral to the basic switch event but, rather, plays a role downstream in the genesis of part of the white-phase phenotype.

We have identified a phenotypic defect in white-phase cells of EFG1null mutants that suggests that the white-phase EFG1transcript plays a role in the genesis of the normal, round, shape of white-phase cells. We have found no similar phenotypic defect in opaque-phase cells of the three EFG1 null mutants that would suggest a role for the less-abundant opaque phase transcript. However, before concluding that this latter transcript plays no role in the opaque phase, two points must be considered. First, observing no morphological difference between wild-type and mutant opaque-phase cells does not prove phenotypic equality. Opaque-phase cells differ from white-phase cells in a number of physiological and virulence characteristics (1, 15, 17, 24, 31) that may be expressed independently of the unique opaque-phase cell morphology. Second, we have not demonstrated that the EFG1 protein is in fact expressed in opaque-phase cells. Experiments to resolve these two latter issues are in progress.