|Home | About | Journals | Submit | Contact Us | Français|
The transcription factor Flo8 is essential for filamentous growth in Saccharomyces cerevisiae and is regulated under the cAMP/protein kinase A (PKA) pathway. To determine whether a similar pathway/regulation exists in Candida albicans, we have cloned C. albicans FLO8 by its ability to complement S. cerevisiae flo8. Deleting FLO8 in C. albicans blocked hyphal development and hypha-specific gene expression. The flo8/flo8 mutant is avirulent in a mouse model of systemic infection. Genome-wide transcription profiling of efg1/efg1 and flo8/flo8 using a C. albicans DNA microarray suggests that Flo8 controls subsets of Efg1-regulated genes. Most of these genes are hypha specific, including HGC1 and IHD1. We also show that Flo8 interacts with Efg1 in yeast and hyphal cells by in vivo immunoprecipitation. Similar to efg1/efg1, flo8/flo8 and cdc35/cdc35 show enhanced hyphal growth under an embedded growth condition. Our results suggest that Flo8 may function downstream of the cAMP/PKA pathway, and together with Efg1, regulates the expression of hypha-specific genes and genes that are important for the virulence of C. albicans.
Candida albicans is the most frequently isolated opportunistic fungal pathogen of humans. Invasive candidiasis is currently the fourth most common cause of bloodstream infections in hospitals and is associated with the highest mortality rate, ~40%, of all the nosocomial bloodstream infections (Kullberg and Filler, 2002 ). The ability to undergo reversible morphogenetic transitions between yeast, pseudohyphae, and hyphae has been shown to be important for its pathogenicity in systemic infections. Mutants defective in the morphogenetic transition show a much reduced virulence in mouse models of systemic infection (Lo et al., 1997 ; Whiteway and Oberholzer, 2004 ).
Flo8 is a transcription factor critical for invasive growth and flocculation in haploids and pseudohyphal growth in diploids of Saccharomyces cerevisiae (Liu et al., 1996 ). It is required for the expression of a family of FLO genes that encode glycerol phosphoinositol (GPI)-anchored cell surface adhesin genes (Kobayashi et al., 1996 ). Among them, FLO11 is essential for invasive/filamentous growth and flocculation (Lo and Dranginis, 1998 ). Flo8 functions downstream of the cyclical AMP (cAMP)-dependent protein kinase A (PKA) pathway because flo8 mutants block the effect of an activated PKA/cAMP pathway on FLO11 expression (Rupp et al., 1999 ). Flo8 has been shown to bind to the promoter of FLO11, and the binding is regulated by Tpk2 (Pan and Heitman, 2002 ), one of three catalytic subunits of PKA in S. cerevisiae. Phosphorylation of Flo8 by Tpk2 is required for Flo8 interaction with the FLO11 promoter both in vivo and in vitro (Pan and Heitman, 2002 ).
In C. albicans, several signaling pathways can regulate the yeast-to-hypha transition (Whiteway and Oberholzer, 2004 ). Among them, the cAMP/PKA pathway plays a major role in hyphal development and virulence, because many mutants in the pathway are defective in hyphal growth and show reduced virulence. The adenylate cyclase Cdc35 and its associated protein Cap1 are required for hyphal development under all hyphal-inducing conditions, including serum (Bahn and Sundstrom, 2001 ; Rocha et al., 2001 ). The cyclase activity is regulated by two G proteins, Ras1 and Gpa2, in C. albicans (Feng et al., 1999 ; Sanchez-Martinez and Perez-Martin, 2002 ; Miwa et al., 2004 ; Maidan et al., 2005 ). ras1 mutants are defective for hyphal formation under induction with serum filtrate in liquid media (Feng et al., 1999 ), whereas gpa2 mutants are defective in hyphal growth on solid hyphal-inducing media (Miwa et al., 2004 ; Maidan et al., 2005 ). The G proteins act through PKA to induce morphogenesis in C. albicans. C. albicans PKA consists of one regulatory subunit, Bcy1, and two catalytic subunits, Tpk1 and Tpk2 (Sonneborn et al., 2000 ; Bockmuhl et al., 2001 ; Cassola et al., 2004 ). Tpk1 and Tpk2 have distinct functions in hyphal development because their mutants have differential effects in different media (Bockmuhl et al., 2001 ; Cassola et al., 2004 ). One potential target of the cAMP/PKA pathway is Efg1, a basic helix-loop-helix protein similar to StuA of Aspergillus nidulans and Sok2 and Phd1 of S. cerevisiae (Stoldt et al., 1997 ). efg1/efg1 mutants are unable to form hyphae in all liquid-inducing media (Stoldt et al., 1997 ), but the mutants show enhanced hyphal formation under embedded growth conditions (Giusani et al., 2002 ). Transcription profiling of cAMP signaling in C. albicans shows that Ras1 regulates a subset of Cdc35-regulated genes, but Efg1-regulated genes are distinct from those modulated by Cdc35 except for the class of genes induced during the yeast-to-hypha transition (Harcus et al., 2004 ). The hypha-specific genes include GPI-anchored cell wall proteins, secreted proteases, and a G1 cyclin-like gene, and they have been shown to be important for virulence in systemic infections (Liu, 2001 ; Zheng and Wang, 2004 ).
Because Flo8 is a target of the cAMP/PKA pathway essential for invasive/filamentous growth in S. cerevisiae, we were interested to determine whether a similar regulator exists in C. albicans. We cloned a C. albicans FLO8 homolog by functional complementation of a S. cerevisiae flo8 mutant. FLO8 deletions in C. albicans completely blocked hyphal development and the expression of hyphal genes, and led to avirulence in a systemic model of candidiasis. Genome-wide transcription analysis of flo8/flo8 and efg1/efg1 mutants suggests that Flo8 specifically regulates subsets of Efg1-regulated genes, mostly hypha-specific genes. We present evidence to suggest that Flo8 and Efg1 function together in regulating the hyphal transcriptional program in C. albicans.
The C. albicans and S. cerevisiae strains used in this study are listed in Table 1. The pseudohyphal colony formation and invasive growth of S. cerevisiae were examined as described previously (Liu et al., 1993 ; Roberts and Fink, 1994 ). Lee's and YPD + 10% bovine serum media were used for hyphal induction (Lane et al., 2001b ). YPS with 1% agar was used for colony morphology assay under embedded conditions (Brown et al., 1999 ).
The clone pCF56 was isolated from a C. albicans genomic library (Liu et al., 1994 ) based on its ability to suppress the invasive defect of a S. cerevisiae flo8 mutant (HLY850) on SC-Ura medium. pCF56 contains a 3.4-kb insert, with a 2454 base pair open reading frame (ORF), corresponding to a protein (Flo8) of 817 amino acids. A 2.4-kb BglII-BamHI fragment in pCF56 was replaced with a 4-kb BglII-BamHI hisG-URA3-hisG fragment from pCUB6, generating plasmid pCF56-FLO8NΔ. A 1.5-kb SpeI fragment from pCF56 was inserted into the SpeI site of pCF56-FLO8NΔ to generate pCF56-FLO8Δ for disruption of FLO8. pCF56-FLO8Δ was digested with XhoI and SstI and transformed into CAI4 to produce FLO8/flo8 and flo8/flo8 strains. Spontaneous Ura– derivatives were selected on 5-fluoro-orotic acid-containing medium. The disruption was confirmed by Southern blotting (Supplemental Figure 1). The plasmids used in this study are listed in Table 2. All the clones and plasmids were confirmed by DNA sequencing.
pBA1 was constructed by subcloning an ADH1 promoter fragment with NotI and EcoRV at each end into BES116 (Feng et al., 1999 ) at the NotI and EcoR V site.
pBA1-CaFLO8 for the expression of C. albicans FLO8 in C. albicans was constructed by placing a 2.45-kb PCR fragment containing FLO8 coding sequence into the BglII-ClaI site of plasmid pBA1. Primer 1 and primer 2 were used for PCR amplification.
pBA1-CaFLO8ΔN was constructed by placing a 2.1-kb PCR fragment into the BglII-ClaI site of plasmid pBA1. Primer 5′CTGGGATCCGTATGCTTCCTCTTATACAGCAG and primer 2 in Table 3 were used for PCR amplification.
pBES116-CaFLO8, a 3-kb HindIII-EcoRV fragment from pCF56 was inserted into BES116 (Feng et al., 1999 ). Then, a 2.4-kb KpnI-HindIII PCR fragment containing C. albicans FLO8 promoter was amplified (primer 3 and 4) and inserted before the FLO8 coding region.
For pVTU-CaFLO8, a 2.45-kb FLO8 containing PCR product (primers 7 and 8) was subcloned into the SstI-XhoI site of pVT102U to express the C. albicans FLO8 under the control of ADH1p in S. cerevisiae. The full-length coding sequences of FLO8 and EFG1 were PCR amplified (primers 9–14) and inserted into pEG202 and pJG4–5, generating pEG202-CaFLO8, pEG202-EFG1, and pJGCaFLO8 for the two-hybrid assay. SC5314 genomic DNA was used as template for PCR amplification. All constructs were verified by DNA sequencing. The plasmids used in this study are listed in Table 2. The primers used for PCR amplification are listed in Table 3.
For CaFLO8-MYC13, a 580-base pair fragment containing 13xMYC was PCR amplified from the S. cerevisiae vector pFA6a-13MYC-HIS3MX6 (Longtine et al., 1998 ) with oligonucleotides 5 and 6 (Table 3), and cloned into the BamHI and SphI sites of the C. albicans ACT1p-FLAG-HIS1 vector (Umeyama et al., 2002 ) to create pPR671. A Not1 and an Mlu1 site were introduced between the BamHI and myc13 sequence in pPR671. The FLO8 gene was amplified from C. albicans genomic DNA (SC5314) using oligonucleotides 15 and 16 (Table 3). The 2.5-kb PCR product was digested with BamHI and MluI and then inserted into the BamHI-MluI sites of pPR671, to produce pCaFLO8–13MYC-FLAG-HIS1 (pPR672). pPR672 was digested with StuI to target the integration of the plasmid into the genomic RP10 locus under HIS1 selection. Expression of Flo8p under the ACT1 promoter in YPD at 30°C was verified by Western analysis using an anti c-myc-conjugated peroxidase antibody (Roche Diagnostics, Indianapolis, IN) enabling detection of Flo8p at ~150 kDa.
pMSCTAP was constructed by cloning a codon-optimized TAP-tag that contains two copies of protein A sequence followed by a single copy of the calmodulin binding peptide (Rigaut et al., 1999 ) into a BlueScript vector that carries C. albicans URA3 (Shrivastava and Liu, unpublished data). The TAP-CaURA3 can be used as a cassette for PCR amplification and insertion of the TAP in frame to the C terminus of a gene of interest.
Primers EFG1-F and EFG1-R are used to PCR amplify TAP-CaURA3 from pMSCTAP (EFG-TAP-F 5′-CCTTCACCCCAACAACATCAAGCTAATCAATCAGCTAGCACTGTTGCCAAAGAAGAAAAGAACATGGAAAAGAGAAGATGG-3′; URA3-EFG-R 5′-CGTTCATGTCAATGGATTTGGGAGAAGATTATGATCTATACTATTTCTTTTTTTATTATTCCGCGGTGGCGGCCGCTCTAG-3′). The underlined regions are the sequences identical to the beginning of TAP and URA3, respectively. The 5′ 60 bp are homologous to the C-terminal end of EFG1. The amplified DNA was transformed directly into C. albicans (Wilson et al., 1999 ), and the transformants with the TAP fused to the C terminus of Efg1 by homologous recombination were identified by PCR and verified by Western blot.
A C. albicans microarray containing 6917 elements was printed with a C. albicans 70-mer set (QIAGEN Operon, Alameda, CA) on a code-link activated slide (GE Healthcare, Little Chalfont, Buckinghamshire, UK) by Microarray (Nashville, TN). The 70-mer set includes ~6530 ORFs from the Assembly 6 ORFs released by the C. albicans Genome Sequencing Project at Stanford University (Stanford, CA). Oligonucleotide information is available at www.qiagen.com. The set includes 192 randomly generated 70-mers as negative controls. The mean intensity of the negative controls is used as the basal hybridization intensity in data evaluation.
Total RNA was extracted using the hot acid phenol method with the addition of phase lock gels (Eppendorf, Westbury, NY) and a LiCl precipitation. The quality of RNA was checked by running the samples on a nanochip using the Agilent Bioanalyzer 2100 (University of California Irvine DNA MicroArray Facility, Irvine, CA). Samples of high-quality RNA, as determined by rRNA profiles, were used for cDNA synthesis. For each experiment, RNA from two samples was pooled (12 μg each) and annealed to 10 μg of Oligo(dT) and Random 9mer primers from the Prime-It II kit (Stratagene, La Jolla, CA). cDNA was synthesized using SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA) in a mixture containing 0.5 mM deoxynucleoside triphosphates (aminoallyl-dUTP:dT in a 3:2 ratio), 5× first strand buffer (Invitrogen), and 0.1 M dithiothreitol (DTT) overnight at 42°C. The RNA was hydrolyzed with 0.2 M NaOH and 0.1 M EDTA at 65°C for 15 min and subsequently neutralized with 0.33 M Tris, pH 7.4. The cDNA was washed with double distilled H2O several times and concentrated to a small volume with a Microcon-3 filter (Millipore, Billerica, MA) and stored at –20°C.
The cDNA was thawed at 42°C for 5 min, resuspended in 0.05 M sodium bicarbonate buffer, pH 9.0, and incubated with Cy3 or Cy5 dye (GE Healthcare) for 1 h at room temperature in the dark. Alternatively, the cDNA was coupled to Alexa Fluor 555 or 647 (Invitrogen) according to the manufacturer's instructions. The QIAGEN PCR purification kit was used to remove any unincorporated dye and eluted with 30 μl of 10 mM Tris-Cl, 5 mM EDTA, pH 8.0, twice. The whole sample was loaded into an Ultravette disposable cuvette (Brandtech Scientific, Essex, CT) and scanned from optical density (OD)200 to OD800 using a spectrophotometer to quantify the amount of cDNA generated as well as the amount of dye coupled to the cDNA. Then, volumes containing equal amounts of cDNA for the appropriate experiments were mixed and concentrated using a Microcon-30 filter. Either 5 or 9 μl of the probe (depending on the size of the coverslip) was heated to 100°C for 2 min. Meanwhile, the microarray slide was washed in 0.2% SDS for 10 min, washed in filtered H2O, dipped in ethanol, and spun dry. A hybridization chamber and Millipore buffer #3 was prewarmed (50°C) and two-thirds volume Millipore buffer was added to the probe. The mixture was spun down before adding to the microarray slide. The slides were hybridized in the hybridization chambers for 16–20 h at 50°C.
After hybridization, slides were washed, dried, and then scanned with a GSI Lumonics ScanArray 4000 slide scanner (GMI, Ramsey, MN) using its Scan Array software. The background-subtracted intensity of Cy3 and Cy5 at each spot was determined using the QuantArray software of the scanner. The background-subtracted intensities of Cy3 and Cy5 were used to plot log2(Cy5i/Cy3i) ratio for each element on a microarray against log10(Cy5ixCy3i) (R-I plot) (data not shown). R-I plots were used to predict the quality of hybridization. For example, curved R-I plots (usually because of photobleaching of one dye) or scattered R-I plots (usually indicates degraded RNA) were not used for subsequent data analysis. We routinely carried out four experimental repeats and array hybridizations for each experiment. Data from two high-quality hybridizations for each experiment, as determined by their R-I plots, were used for subsequent data analyses. Because equal amounts of cDNA for each dye were used in hybridization, we used total intensity normalization to normalize the Cy3 and Cy5 for each hybridization. The normalization involved scaling the Cy5 intensities by multiplying them with a normalization factor, which was determined by dividing the sum of Cy3 intensities by the sum of the Cy5 intensities. The ratio of Cy5i to Cy3i was then calculated using the corrected Cy5 values. The ratios that were above 2 times basal cutoff (Cy5i-Cy3i> avgCy5ibasal+avgCy3ibasal) and threefold cutoff were log transformed and clustered with the average linkage cluster in a hierarchical cluster program (Eisen et al., 1998 ). The clustered data were viewed in TreeView. The cluster and TreeView programs are at http://rana.stanford.edu/software/.
C. albicans cells were grown to 3–5 × 107 cells/ml in 50 ml of YPD, SSA, or Lee's media in yeast or hyphal growth conditions. The cells were harvested by centrifugation at 4°C, washed, and resuspended in 0.5 ml of lysis buffer (10 mM Tris-HCl, pH 8, 250 mM NaCl, 0.1% NP-40, 0.5 mM DTT, 0.5 mM phenylmethylsulfonyl fluoride, 2 mM benzamidine, 0.5 μg/ml leupeptin, 1.4 μg/ml pepstatin, 2.4 μg/ml chymostatin, and 17 μg/ml aprotinin) and equal volume of glass beads (Sigma, St. Louis, MO). Cells were lysed at 4°C using a Fast-Prep system (FP120; Thermo Electron, Waltham, MA). Cell lysates were centrifuged for 10 min at 13,000 rpm in a microcentrifuge at 4°C. Protein extract containing 5 mg of protein was subjected to immunoprecipitation using 60 μl of rabbit IgG agarose bead slurry that was preincubated once with 0.2 mg/ml sheared salmon sperm DNA, 0.5 mg/ml bovine serum albumin in phosphate-buffered saline (PBS), and washed once in the lysis buffer. After incubation for 2 h at 4°C, beads were washed six times with 0.5 ml of lysis buffer and once with 1 ml of TE (10 mM Tris-HCl, pH 8, 1 mM EDTA). Bound proteins were eluted from the beads in 60 μl of elution buffer (50 mM Tris-HCl, pH 8, 10 mM EDTA, and 1% SDS) by incubation for 10–15 min at 65°C. Proteins were separated by 8% SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Hybond; GE Healthcare). After blocking in 3% skim-milk powder in 0.05% PBS, Tween 20, a peroxidase-conjugated anti-c-myc antibody (Roche Diagnostics) was used to probe for myc-tagged proteins, which were then detected using the ECL system (GE Healthcare).
Two-hybrid assays were performed as described previously (Gyuris et al., 1993 ). Yeast strain EGY48 containing the LexAop-LacZ reporter plasmid pSH18-34 was cotransformed with pEG202-based plasmids expressing LexA DNA binding domain fusions and pJG4-5-based plasmids containing transcriptional activation domain fusions (Gyuris et al., 1993 ).
RNA extraction and Northern blotting were performed as described by Lane et al. (2001a ). A 3.4-kb FLO8 fragment from pCF56 was used as a probe for Northern analysis. PCR products for C. albicans ECE1, HWP1, ALS1, HGC1, IHD1, and HSP31 were used for probing Northern blots.
The virulence of C. albicans strains was tested as described by Chen et al. (2000 ). ICR male mice (18–21 g) from Shanghai Laboratory Animal Center, Chinese Academy of Sciences (Shanghai, China) were used for the virulence assay.
The GenBank accession number for the C. albicans FLO8 nucleotide sequence is AF414113 and orf19.1093.
Blast analysis of the S. cerevisiae Flo8 protein sequence against the C. albicans genome sequence did not identify a Flo8 orthologue. Therefore, we tried to clone potential C. albicans homologues of ScFlo8 by functional complementation in an S. cerevisiae flo8 mutant. A C. albicans genomic library was transformed into the S. cerevisiae flo8 mutant, and genes that complemented the flo8 defect in invasive growth were isolated. The isolated clones were further examined for their abilities to suppress the invasive growth in a flo11 mutant. Among the several clones isolated, pCF56 showed strongest suppression of invasive growth in flo8 (Figure 1A), and it did not suppress flo11 (our unpublished data). The pCF56 also suppressed the filamentous growth defect of a diploid flo8 mutant (Figure 1B). ScFlo8 has been shown to be essential for the formation of biofilm-like colonies on media with low percentage agar (Reynolds and Fink, 2001 ), and we found that the defect could be partially suppressed by pCF56 (Figure 1C). pCF56 contains a gene that encodes a putative protein of 817 aa, with a small region (amino acids 30–92) highly similar to amino acids 72–154 of ScFlo8 (Figure 2A). We therefore designated this gene C. albicans FLO8. The region conserved between the two Flo8 proteins contains a LUFS domain (LUG/LUH, Flo8, single-stranded DNA binding protein) (Conner and Liu, 2000 ). LUG (Leunig) is a key regulator of flower-specific gene expression during flower development in Arabidopsis (Conner and Liu, 2000 ). The single-stranded DNA binding protein (Ssdp) regulates the activity of LIM-homeodomain protein complexes (van Meyel et al., 2003 ) (Figure 2A). Within the LUFS domain, there is a Lissencephaly type 1-like homology motif (LisH) (Emes and Ponting, 2001 ). The crystal structure of the N-terminal domain of mouse LIS1 shows that the LisH motif is a thermodynamically very stable dimerization domain (Kim et al., 2004a ).
To determine whether the LUFS domain is important for C. albicans Flo8 function, we deleted the N-terminal 122 amino acids of Flo8 and introduced Flo8123–817 and Flo81–122 into haploid and diploid flo8 mutants of S. cerevisiae. Whereas the full-length C. albicans Flo8 could completely complement the invasive/filamentous growth defect of the flo8 mutants, neither Flo8123–817 nor Flo81–122 could (Figure 2B; our unpublished data). Therefore, both the LUFS domain in Flo81–122 and other domains in Flo8123–817 are required for Flo8 function. To determine whether the LUFS domain is important for C. albicans Flo8 transcriptional activity, we fused different Flo8 domains to the lexA DNA binding domain, and measured the transcriptional activities of the fusion proteins in S. cerevisiae. Fusion of Flo8 to the DNA binding domain of lexA gave a high level of lexAop-lacZ expression in a S. cerevisiae flo8-1 strain, which carries a nonsense mutation encoding Flo8 with a truncated LUFS domain (Liu et al., 1996 ). This observation is similar to that reported for S. cerevisiae Flo8 (Rupp et al., 1999 ; Pan and Heitman, 2002 ). Therefore, C. albicans Flo8 is likely a transcriptional activator. Deleting the N-terminal 122 amino acids decreased Flo8 transcriptional activity by only 50%, but deleting the C-terminal 200 amino acids reduced the transcriptional activity by 60-fold (Figure 2C), suggesting that the transcriptional activation domain is probably at the C terminus. Alternatively, the C-terminal domain could interact with a transcriptional activator, leading to transactivation of Flo8. Removing the two regions together completely abolished the transcriptional activity. Our data suggest that both the N-terminal LUFS domain and the C-terminal region are important for Flo8 function.
To study the function of Flo8 in C. albicans, we deleted both FLO8 alleles by using a hisG-URA3-hisG cassette. Successful deletion of C. albicans FLO8 was confirmed by Southern analysis (Supplemental Figure 1). flo8/flo8 mutants had no significant differences in growth rate and cell morphology from wild-type cells under yeast growth conditions at 30°C. However, they were completely unable to respond to hyphal induction in all liquid hyphal-inducing conditions examined, including serum-containing media, Lee's medium (Figure 3A), SSA, CAA, GlcNAc, and RPMI 1640 (our unpublished data). flo8/flo8 mutant cells displayed yeast-like cell morphology under the hyphal inducing conditions. flo8/flo8 mutants also failed to form hyphal colonies on solid media and formed only smooth colonies even after extended growth at 37°C (Figure 3A). The defects of flo8/flo8 mutants could be rescued by integrating a wild-type FLO8 under either the ADH1 promoter or under its own promoter (Figure 3, A and B). The complementation of flo8/flo8 was not because of differences in levels of URA3 expression or insertion positions in the genome because a copy of URA3 inserted at the same ADE2 locus did not suppress the defects (Figure 3A). Partial defects in hyphal morphogenesis and hyphal colony formation were observed in the FLO8/flo8 mutant and the flo8/flo8 mutant transformed with a wild-type copy of FLO8 under its own promoter (Figure 3A). This haploid insufficiency has been observed in many mutants of hyphal regulators. Based on the haploid insufficiency and the complementation of flo8/flo8 defects by wild-type FLO8, we suggest that the defects in flo8/flo8 are from the FLO8 deletion.
Consistent with the defect in hyphal morphogenesis, the flo8/flo8 mutants were also defective in the expression of hypha-specific genes (Figure 3B). ECE1 and HWP1 were highly induced in wild-type cells under hyphal-inducing conditions, but they were not induced at all in flo8/flo8 mutants. The FLO8/flo8 heterozygote showed reduced expression of hypha-specific genes. Because Flo8 is required for the expression of several FLO genes of cell surface adhesins in S. cerevisiae, we also examined the expression of ALS1, which is expressed in both yeast and hyphal cells in an Efg1-dependent manner. Our data showed that Flo8 is essential for hyphal morphogenesis and the induction of hypha-specific genes and ALS1. FLO8 was expressed in both yeast and hyphae at low levels, as determined by the Northern analysis (our unpublished data).
The LUFS domain is important for Flo8 functions in hyphal development. A flo8/flo8 mutant transformed with FLO8ΔN, which lacks the LUFS domain, was unable to develop hyphae or express hyphal genes (Figure 3, C and D).
The dimorphic transition ability of C. albicans has been linked with its pathogenicity in mice (Lo et al., 1997 , Chen et al., 2000 ). Therefore, we examined the virulence of flo8/flo8 mutants in a systemic model of infection. Cells (5 × 106) of wild type (CAI4), flo8/flo8, and FLO8-complemented flo8/flo8 strain were inoculated into each mouse by tail vein injection. All three strains carry one copy of UAR3 integrated at the ADE2 locus. We observed that the mice injected with wild-type and complemented flo8/flo8 C. albicans started to lose weight after 1 d and started to die after 2 d, and all of the mice died by day 15 (Figure 4). The mice injected with the flo8/flo8, in contrast, had no symptoms of illness, and all mice survived for more than 25 d (Figure 4). Therefore, flo8/flo8 was avirulent in a mouse model of systemic infection.
The phenotypes of the flo8/flo8 mutant in various hyphal-inducing conditions are very similar to that of efg1/efg1. Both are unable to induce hyphal morphogenesis and express hypha-specific genes in serum-containing media, and both show severely reduced virulence in the systemic infection model in mice. To further define Flo8 functions and its relationship with Efg1, we compared transcription profiles of flo8/flo8 and efg1/efg1 mutants.
We performed microarray hybridizations with cells grown in two sets of hyphal-inducing media, YPD + serum and Lee's. Wild-type, efg1/efg1, and flo8/flo8 cells were grown in YPD at 25°C for yeast growth and in YPD + serum at 37°C for hyphal growth. Similarly, the three strains were grown in Lee's medium at 25°C for yeast growth and in Lee's at 37°C for hyphal growth. After 3 h growth in YPD media or 6 h in Lee's medium, cells were harvested for RNA extraction. The RNA was reverse transcribed, and equal amounts of cDNA were labeled with Cy3 and Cy5 and hybridized to each slide. Wild type versus efg1/efg1 and wild type versus flo8/flo8 in each yeast growth or hyphal growth condition as well as wild-type hyphae versus wild-type yeast were compared on each slide. Changes in gene expression were measured as a ratio of normalized Cy3 versus Cy5 intensity at each position on each microarray slide. Most experiments had four experimental repeats. The ratio of normalized Cy3 versus Cy5 intensities from two repeats with high quality of hybridization, as determined by R-I plots (see Materials and Methods) were used in clustering analysis (Eisen et al., 1998 ). The clustering result is visualized with TreeView (Figure 5A). The ratios used for the clustering analysis are available in Supplemental Figure 2.
Clustering analysis shows that Flo8 was essential for the expression of subsets of Efg1-regulated genes (Figure 5A). Overall, flo8/flo8 and efg1/efg1 were grouped together under all experimental conditions by the cluster program, indicating that the transcription profiles were more similar between the two mutants than between two growth conditions. We did not identify genes regulated only by Flo8, but not by Efg1, under our experimental conditions. It seemed that genes whose expression was affected in the flo8/flo8 mutant were similarly affected in the efg1/efg1 mutant. Most of the Flo8-regulated genes were hypha-specific, induced only in hyphal cells in both YPD + serum and Lee's and the expression was blocked in both mutants. These include all known hypha-specific genes on the microarray (Figure 5A). IHD1, identified in a genome-wide transcriptional profiling of genes induced or repressed during hyphal development by Nantel et al. (2002 ), was also found in this cluster, and further confirmed by Northern analysis (Figure 5B). The gene was independently found by Murad et al. (2001 ) as one of the genes regulated by the Nrg1 and Tup1 repressors. IHD1 is predicted to encode a GPI-anchored cell wall protein (De Groot et al., 2003 ), and the protein sequence also has similarity to glucan 1,4-α-glucosidase Sta1 in Saccharomyces diastaticus (Yamashita et al., 1985 ). Some hypha-specific genes were expressed at higher levels in Lee's medium than in YPD + serum. These included SAP4, SAP5, and CLN21/HGC1, a G1-type cyclin gene that was recently found to be hypha-specific and essential for hyphal morphogenesis (Zheng and Wang, 2004 ).
Not all Efg1-regulated genes were regulated by Flo8. Several Efg1-repressed genes were not repressed by Flo8 (Figure 5A). As confirmed by Northern blotting, HSP31 was only detected in the efg1/efg1 mutant in Lee's medium and in YPD + serum at 37°C, but not in wild type or flo8/flo8 (Figure 5). Transcriptional profiling of efg1/efg1 and cdc35/cdc35 by Harcus et al. (2004 ) has demonstrated that hypha-specific genes are the only genes regulated by both Efg1 and the cAMP/PKA pathway. Many Efg1-regulated genes are not affected by the cdc35/cdc35 mutant. Similarly, we find that, Flo8 only regulates subsets of Efg1-regulated genes and those genes are mostly hypha-specific.
The fact that all the Flo8-regulated genes we identified so far were similarly regulated by Efg1 suggests that Flo8 and Efg1 might function together to control the expression of these genes. To determine whether Flo8 and Efg1 act together in transcriptional activation, we first determined whether the two transcription factors interact by yeast two-hybrid system. The Efg1 fusion to the DNA binding domain of lexA did not activate the transcription from lexAop-lacZ, in agreement with a recent report (Doedt et al., 2004 ). When using the lexA-Efg1 as bait in the yeast two-hybrid system, we detected a weak two-hybrid interaction of Efg1 with Flo8 (Figure 6A). A reciprocal two-hybrid assay could not be performed because the lexADB-Flo8 fusion gave a high level of basal activity.
To further investigate whether Flo8 interacts with Efg1 in C. albicans, we performed immunoprecipitation experiments with tagged proteins. Flo8 was tagged at its C terminus with Myc13, and expression was under the control of the ACT1 promoter. Efg1 was fused at its C terminus with two copies of protein A sequences followed by calmodulin binding protein. Both fusions were functional (our unpublished data). Immunoprecipitation of Efg1 with IgG beads was able to pull down Flo8 from both yeast and hyphal cells grown in three different growth media (Figure 6B). The interaction was specific because the interaction was barely detectable in the control strain that carried only Flo8-Myc13. Furthermore, the observed interaction was not due to the protein A (fused to Efg1), because protein A-beads could not bring down Flo8myc unless anti-myc antibodies were included in the precipitation (our unpublished data). The immunoprecipitation (IP) data suggest that Flo8 and Efg1 can interact in vivo. It also suggests that the interaction is not regulated in connection to growth forms. This is consistent with the finding that Flo8 and Efg1 are required for ALS1 expression in yeast and hyphae.
The functional relationship between Efg1 and Flo8 was further studied by epistasis experiments in S. cerevisiae. Overexpression of EFG1 or C. albicans FLO8 in S. cerevisiae enhanced invasive growth in wild-type haploids and bypassed the requirement of the Kss1 mitogen-activated protein kinase pathway. But the effect of EFG1 overexpression was completely blocked in a flo8 mutant (Figure 7). Therefore, Flo8 is required for Efg1-mediated transcriptional activation of invasive/filamentous growth in S. cerevisiae. Overexpression of EFG1 from the PCK1 promoter in a C. albicans flo8/flo8 mutant did not induce filamentous growth, and the flo8/flo8 mutant carrying the EFG1 overexpression construct showed similar yeast growth morphology as the control flo8/flo8 strain carrying a vector (our unpublished data). Although the result was in agreement with the epistasis study in S. cerevisiae, the extent of data interpretation was limited by the relatively weak phenotype observed in wild-type C. albicans expressing EFG1 under the PCK1 promoter, because it only generated pseudohyphal filaments. Reciprocal epistasis experiments with FLO8 overexpression in efg1/efg1 was hindered by the lack of any detectable phenotypes from FLO8 overexpression (our unpublished data).
The ability of C. albicans to sense the presence of surrounding matrix may play a role during infection. Although efg1/efg1 is defective in hyphal development under many laboratory conditions at 37°C (Lo et al., 1997 ), it forms filaments on the tongue of immunosuppressed piglets (Riggle et al., 1999 ) and undergoes filamentous growth when embedded in YPS agar (Giusani et al., 2002 ). Because flo8/flo8 mutants had a similar phenotype to efg1/efg1 in aerobic conditions at 37°C and Flo8 could interact with Efg1 in vivo, we examined flo8/flo8 mutants under embedded conditions. Like efg1/efg1, colonies of flo8/flo8 produced filaments in 1 d when grown embedded in YPS agar at 25°C. In contrast, wild-type strain SC5314 just started to produce filaments after 2 d at 25°C, and even after 3 d, the wild type formed only limited amounts of filaments and remained predominantly smooth (Figure 8). One observable difference between efg1/efg1 and flo8/flo8 was that filamentation seemed heterogeneous in the efg1/efg1 mutant with long hyphal filaments surrounded by branches covered with yeast cells, whereas the flo8/flo8 mutant formed homogeneous hyphal filaments in each hyphal colony. In parallel to efg1/efg1 and flo8/flo8, we also examined whether the cAMP/PKA pathway represses filamentous growth in microaerophilic conditions. Similar to the flo8/flo8 mutant, a cdc35/cdc35 mutant also formed homogenous hyphal filaments in each colony, although the cdc35/cdc35 mutant grew much slower than flo8/flo8. Therefore, like Efg1,the cAMP/PKA pathway and Flo8 activity are inhibitory to filamentation in response to growth within a matrix at low temperature.
We identified a C. albicans gene, FLO8, by functional complementation of a flo8 mutant of S. cerevisiae. Like S. cerevisiae Flo8, Flo8 does not possess an obvious DNA binding motif, but it has a conserved LUFS domain that is found in several other regulatory proteins such as Leunig, Flo8, and Ssdp (Conner and Liu, 2000 ). The N-terminal half of the LUFS domain has a conserved LisH motif, which exists in >100 eukaryotic proteins of various functions (Emes and Ponting, 2001 ). The crystal structure of a mouse LIS1 fragment has shown that the LisH motif is a thermodynamically very stable dimerization domain, located adjacent to a coiled-coil fragment in LIS1 (Kim et al., 2004a ). Interestingly, the LUFS domain also has a conserved region with two predicted α-helices adjacent to the LisH motif (Figure 2A). It is likely that the LisH motif in Flo8 is also involved in dimerization, and together with the α-helices, forms a structure similar to the structure found in LIS1 (Kim et al., 2004a ). The dimerization could be homodimerization of Flo8 itself or heterodimerization of Flo8 with another protein with a LUFS domain. In fact, ScFlo8 has been shown to physically and functionally interact with Mss11, another LUFS-containing transcription factor, in the transcription of their target genes (Kim et al., 2004b ), although it remains to be seen whether the interaction is mediated through the LUFS domains in these proteins. Dimerization in some transcription factors is required for DNA binding, and despite the absence of an obvious DNA binding domain, ScFlo8 is a sequence-specific DNA binding protein, and its binding to target promoters is regulated by PKA activity (Pan and Heitman, 2002 ; Kim et al., 2004b ). Based on our study, the LUFS domain is required for Flo8 function in vivo. A Flo8 truncation lacking the LUFS domain could not restore invasive/filamentous growth in Scflo8 mutants or in C. albicans flo8/flo8. Because the LUFS domain is not required for Flo8 transcriptional activities, it may be important for DNA binding either directly or via other regulators. In addition to the LUFS domain, we also show by fusions of Flo8 fragments to a lexA DB domain that a transcriptional activation domain is located at the C terminus of Flo8. Alternatively, the C terminus of Flo8 is transcriptionally active through association with other transcription activators.
Flo8 is essential for hyphal development in C. albicans. Deleting FLO8 completely blocks hypha formation under all aerobic hyphal-inducing conditions investigated. In terms of the extent of cell elongation in yeast and hyphal growth conditions, the flo8/flo8 mutant shows a more specific and tighter phenotype than the efg1/efg1 mutant, because efg1/efg1 cells are slightly elongated in response to serum at 37°C, whereas flo8/flo8 cells remain yeast-like. Whole-genome transcription analysis of flo8/flo8 and efg1/efg1 shows that Flo8 regulates subsets of Efg1-regulated genes, and most of these are hypha-specific genes, including a recently reported hypha-specific G1-cyclin (Zheng and Wang, 2004 ) and a potential cell wall protein (Ihd1) with similarity to glucan 1,4-α-glucosidase Sta1 in S. diastaticus. We did not identify a class of genes that are regulated only by Flo8 under the growth conditions we used. In contrast, some Efg1-regulated genes are not affected by flo8/flo8. Therefore, Flo8 is essential and specific for hyphal development.
Like Efg1, Flo8 seems to have a dual function in filamentous growth. Although the flo8/flo8 mutant blocked hyphal development in aerobic conditions at 37°C, it formed long hyphal filaments in a microaerophilic condition at room temperature. This suggests that Flo8 could act as an activator of the hyphal response program to certain stimuli but could also function as a repressor of hyphal development in a matrix. Unlike the efg1/efg1 mutant, which formed heterogenous filaments surrounded with yeast cells in the matrix, the flo8/flo8 mutant formed uniform hyphal filaments when embedded in agar. Therefore, flo8/flo8 seems to have a stronger and more complete phenotype than efg1/efg1 under both aerobic and microaerophilic conditions.
The flo8/flo8 mutant is avirulent in a mouse model of systemic infection. This is probably directly linked to the critical role of Flo8 in hyphal development. The flo8/flo8 mutant completely blocked hyphal development and specifically prevented the expression of hypha-specific genes, many of which are known to contribute to virulence. In addition, the flo8/flo8 mutant is completely filamentous under embedded conditions. It seems that the mutant is unable to switch between yeast and filamentous forms under a given condition. This is consistent with its avirulent property in the mouse model, because the virulence of C. albicans is considered to be linked to the ability of cells to undergo dimorphic transitions. The subtle morphological difference between the efg1/efg1 and the flo8/flo8 mutants is also consistent with the observation that flo8/flo8 is avirulent, whereas the efg1/efg1 mutant shows reduced virulence.
Several lines of evidence suggest that Flo8 may function downstream of the cAMP/PKA pathway, and together with Efg1, may regulate the expression of hyphal genes. First, C. albicans FLO8 could complement the defects of Scflo8 in invasive/filamentous growth, and therefore it is likely to function in the same position as ScFlo8 in S. cerevisiae, which is downstream of Tpk2. Second, the flo8/flo8 mutant has similar phenotypes as cdc35/cdc35 and efg1/efg1; they are defective in hyphal development and the induction of hypha-specific genes including HGC1 under many liquid hyphal-inducing media, including serum-containing media, but they show elevated filamentation under embedded conditions. Third, transcription profiling of flo8/flo8 and efg1/efg1 shows that Flo8 regulates the hypha-specific set of the Efg1-regulated genes. We did not find genes regulated by Flo8, and not by Efg1. Interestingly, hypha-specific genes are the only overlapping genes affected by both efg1/efg1 and cdc35/cdc35 in a genome-wide profiling study (Harcus et al., 2004 ). Fourth, an in vivo interaction between Flo8 and Efg1 was detected by yeast two-hybrid and immunoprecipitation. This places both Efg1 and Flo8 downstream of the cAMP/PKA pathway. Whether one or both of them are regulated by PKA remains to be determined. Considering that some Efg1-repressed genes are not regulated by Flo8, we predict that, when in complex with Flo8, the Efg1/Flo8 complex can activate transcription. This is consistent with the notion that Efg1 is an inhibitor of transcription under both yeast and hyphal growth conditions (Doedt et al., 2004 ), whereas Flo8 seems to act as a transcriptional activator as the lexADB-Flo8 fusion has high transcriptional activity. It is possible, that by interacting with different regulators, Efg1 can exert different effects on transcription.
Flo8 may interact with additional regulators to integrate responses from different signaling pathways to regulate the hyphal transcriptional program. LUFS-containing regulators have been found to interact with other transcriptional regulators. In Arabidopsis, the LUFS domain of Leunig is both necessary and sufficient for its interaction with Seuss, and the Leunig/Seuss complex regulates gene expression during flower development (Sridhar et al., 2004 ). In Drosophila, the Ssdp protein interacts through its LUFS domain with Chip, a cofactor of the LIM homeodomain transcription factors, and regulates the activity of the LIM–homeodomain protein complexes in various developmental processes (van Meyel et al., 2003 ). In S. cerevisiae, the Flo8/Mss11 complex interacts with Ste12 and Tec1 to activate STA1 expression (Kim et al., 2004b ). In C. albicans, we show that Flo8 interacts with Efg1. Because the flo8/flo8 mutant has a tighter phenotype than that of efg1/efg1, we predict that Flo8 may interact with other coactivators or coinhibitors besides Efg1 to regulate hyphal development. So, we propose that Flo8 acts downstream of several pathways to integrate various signals in hyphal development. In conclusion, we have identified a conserved regulator Flo8 that plays an essential role in the dimorphic switch and virulence of C. albicans.
We thank the Stanford DNA sequencing facility for the complete Candida albicans genome sequence and the Galar Fungal Consortium for CandidaDB. Sequencing of C. albicans was accomplished with the support of the National Institute of Dental Research and the Burroughs Wellcome Fund. The DNA microarray was established in collaboration with Scott Filler. This work was supported by National Institutes of Health Grant GM-55155 and UC University-wide AIDS Research Program (F03-1-208) to H. L., by National Institutes of Health (DE013974) to Scott Filler, and by Chinese National Natural Science Foundation Grants 30330010 and 30028010 and Chinese National 863 Grants 2004AA223120 and CAS2004-2-8 to J.Y.C.
This article was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E05–06–0502) on November 2, 2005.
The online version of this article contains supplemental material at MBC Online (http://www.molbiolcell.org).