Gpa2, a G-Protein {alpha} Subunit Required for Hyphal Development in Candida albicans

Candida albicans is able to respond to environmental changesby inducing a distinct morphological program, which is relatedto the ability to infect mammalian hosts. Although some of thesignal transduction pathways involved in this response are known,it is not clear how the environmental signals are sensed andtransmitted to these transduction cascades. In this work, wehave studied the function of GPA2, a new gene from C. albicans,which encodes a G-protein ${alpha}$ -subunit homologue. We demonstratethat Gpa2 plays an important role in the yeast-hypha dimorphictransition in the response of C. albicans to some environmentalinducers. Deletion of both alleles of the GPA2 gene causes invitro defects in morphological transitions in Spider mediumand SLAD medium and in embedded conditions but not in mediumcontaining serum. These defects cannot be reversed by exogenousaddition of cyclic AMP. However, overexpression of HST7, whichencodes a component of the filament-inducing mitogen-activatedprotein kinase (MAPK) cascade, bypasses the Gpa2 requirement.We have obtained different gain-of-function and loss-of-functionmutant alleles of the GPA2 gene, which we have introduced inseveral C. albicans genetic backgrounds. Our results indicatethat, in response to environmental cues, Gpa2 is required forthe regulation of a MAPK signaling pathway.

INTRODUCTION

Top

Introduction

Candida albicans is the most common fungal pathogen in humans.It is frequently associated with systemic infections in immunocompromisedpatients (39). The molecular analysis of the regulation of virulencetraits in this organism may reveal new targets for antifungalagents. C. albicans is a dimorphic fungus, in which the yeast-hyphatransition is triggered by different environmental cues suchas serum, neutral pH, high temperature, starvation, etc. Diseaseprogression is promoted by yeast-hypha morphogenesis, whichis linked to the expression of virulence factors (8, 10). Researchefforts have been directed to the characterization of the transductionpathways involved in this signaling process. Most of the dimorphictransition signals converge into two parallel signal transductionpathways, defined by the transcription factors Efg1 and Cph1(17). The Efg1 protein is thought to be the major regulatorin hyphal development because mutants with null mutations inthe EFG1 gene display severe hyphal defects (33, 48). Currentmodels place EFG1 in the cyclic AMP (cAMP) signaling pathway.Epistatic analyses are consistent with efg1 being situated downstreamof the C. albicans isoforms of protein kinase A (PKA) (4, 47),and Efg1 itself could be a target of PKA phosphorylation (3).The Cph1 protein is part of a mitogen-activated protein (MAP)kinase-based signaling pathway. Cph1 is a transcription factorthought to be activated by this MAP kinase cascade (13, 23,28, 29, 32). Additional pathways are the pH-responsive and embeddedpathways, which involve specific regulators but which are alsodependent on the Efg1 and Cph1 pathways (9, 16, 21, 46).

Although it is known that C. albicans responds to environmentalcues, little is known about the way these stimuli are sensedand transmitted to the different signaling pathways. In eukaryoticcells, the small GTP-binding protein Ras and heterotrimericG proteins are involved in the transmission of external signals.In C. albicans a single Ras protein has been described (18,30). Genetic analyses have shown that RAS1 is located upstreamof the genes involved in both the Efg1 and Cph1 pathways (12,30). G-protein ${alpha}$ -subunit (G ${alpha}$ -subunit)-encoding genes in a growingnumber of fungal species have been identified (5). In thesefungi, deletion of G ${alpha}$ subunits leads to morphological and matingdefects that compromise the ability of the fungi to respondto the appropriate stimuli. In C. albicans, only one G ${alpha}$ subunithas been described so far, the one encoded by the CAG1 gene,but no noticeable morphological phenotypes are present in nullmutants (43). A second G ${alpha}$ -homologue-encoding gene, called GPA2,has been identified during the course of the annotation of assembly6 of the C. albicans genome available at the Stanford GenomeTech-nology Center (http://www-sequence.Stanford.edu/group/candida)and conducted by the European Consortium Galar Fungail (http://www.Pasteur.fr/recherché/unites/Galar_Fungail/).In this study, we sought to investigate the role of the C. albicansGPA2 gene in mediating morphological transitions in responseto environmental inducers. Our results suggest that signal transductionthrough G proteins contributes to dimorphic differentiationin response to environmental cues.

MATERIALS AND METHODS

Top

Materials and Methods

Strains and media. C. albicans strains are listed in Table 1. Routine growth wason yeast extract-peptone-dextrose (YPD) plates and liquid mediumat 30°C, and selections for URA3 prototrophy were performedon minimal SD plates (45). Counter-selections against URA3 wereperformed on uridine- and 5-fluoroorotic acid-containing SDplates (19). For specific experiments involving hyphal growth,different media were used. SLAD medium (20) and Spider mediumwere prepared as previously described. For serum-mediated filamentationYPD medium containing 10% fetal calf serum (Gibco BRL) was used.Lee's medium at pH 7 (31) was used to assess filamentation inducedby pH conditions. Embedded conditions were used as previouslydescribed (9).

View this table:
[in this window]
[in a new window]
TABLE 1. C. albicans strains

Knockout of GPA2. Disruption of the two allelic copies of GPA2 was obtained bythe two-step replacement method described by Fonzi and Irwin(19). The GPA2 coding sequence was replaced by the selectionmarker hisG-URA3-hisG by using two different orientations asdescribed previously (7, 40). In first place, we constructeda plasmid containing 5' and 3' GPA2 flanking sequences joinedby a small polylinker formed by recognition sites for the BglII,PstI, and KpnI restriction enzymes. The 600-bp 5' GPA2 regionwas obtained as a HindIII-PstI fragment by PCR amplificationwith oligonucleotides CGPA-A (5'GAAGCTTTTTTGGTTTTGGTTTTGTTCCC3')and CGPA-B (5'CCCTGCAGAAAGATCTGTGTAATGGGGGATGATATAC3'). The600-bp 3' GPA2 region was obtained as a PstI-SacI frament byPCR amplification with the CGPA-C (5'CCCTGCAGAAGGTACCAGTTTTATGCAAATTGTCATA3')and CGPA-D (5'CCGAGCTCATTCAATGGCTTTTTTAAATA3') oligonucleotides.Both PCR fragments were cloned in pUC19 digested with SacI-HindIII.The resulting plasmid was digested with either BglII-PstI orPstI-KpnI and was ligated with the hisG-URA3-hisG cassette obtainedfrom pMB7 as a BglII-PstI or PstI-KpnI fragment, respectively.The final plasmids, pGPA2KO-A and pGPA2KO-B, were digested withSacI and HindIII and used to transform CAI4 cells. Sequentialdisruption steps were analyzed by using oligonucleotides directedto hisG sequences and sequences flanking the integration regionas previously described (40). Positive clones were confirmedby Southern analysis.

Plasmid constructions. To construct plasmids carrying the different GPA2 alleles, thegenomic GPA2 region was isolated by PCR as a 3.1-kbp HindIIIfragment by using DNA from C. albicans SC5314 (19) and clonedin pGEM-T. The GPA2 mutated alleles were obtained by site-directedmutagenesis (27). The oligonucleotides used for the mutagenesiswere GPA2-I (5'TTGTTATTGGGATCAGTTGAAAGTGGTAAATCT3') to generatethe GPA2^G142V allele, GPA2-K (5'TTATTTGATGTTGGTGTTCAAAGGTCAGAAAGA3')to generate the GPA2^G354V allele, and GPA2-N (5'GATGTTGGTTTAAGGTCAGAAAGA3')to generate the GPA2^Q355L allele. The GPA2 alleles were isolatedby PCR as NotI-SacII fragments from the pGEM-derived plasmidswhere mutagenesis was made and cloned into NotI- and SacII-digestedpCNB2 or pCNB3 plasmids. The pCNB2 plasmid carries the ACT1promoter region cloned in front of the multicloning site, aURA3 selection marker, and the FRT site to allow site-specificintegration in the genomes of C. albicans strains carrying FLPcassettes (44). Generation of C. albicans strains able to supportthe FLP-dependent integration system and the integration ofFRT-containing delivery vectors were performed as previouslydescribed (44). pCNB3 is a self-replicating URA3 plasmid whichcarries the ACT1 promoter region cloned in front of the multicloningsite (unpublished data).

To construct the plasmid expressing the RAS1^G13V allele, weisolated by PCR a genomic 1.5-kbp DNA fragment and cloned thisfragment in pGEM-T. This plasmid was used as a template forsite-directed mutagenesis with the oligonucleotide QF87 (5'TTGTTGTTGGAGGTGTTGGTGTTGGTAAAT3')(18). The RAS1^G13V allele was isolated by PCR as NotI-SacIIfragments from the pGEM-derived plasmid and cloned into theNotI- and SacII-digested pCNB3 plasmid.

The pYPB1-ADHp-HST7 and pBI-HAHYD plasmids, overexpressing HST7and EFG1 respectively, were gifts from A. J. P. Brown and J.Ernst, respectively, and they have been already described (28,47).

DNA isolation and Southern blotting. Genomic DNAs were isolated from C. albicans strains as describedpreviously (22). Southern blotting, probe labeling, and detectionwere performed according to instructions from the manufacturerof the DIG High Prime DNA labeling and detection starter kitI (Roche).

RESULTS

Top

Results

C. albicans Gpa2 is required for hyphal formation. During the process of the annotation of assembly 6 of the C.albicans genome conducted by the European Consortium Galar Fungail,we found an open reading frame whose conceptual translationproduces a protein with a high sequence similarity to G ${alpha}$ subunitsfrom other fungi. We called the putative corresponding geneGPA2. The open reading frame of the GPA2 gene contains no intronand is predicted to encode a protein of 502 amino acids. Gpa2has sequence identities of 46 and 53% with Saccharomyces cerevisiaeGpa2 (34) and Ustilago maydis Gpa3 (41), respectively (Fig.1A).

View larger version (66K):
[in this window]
[in a new window]
FIG. 1. The C. albicans Gpa2 protein. (A) Sequence alignment of U. maydis Gpa3 (Um), C. albicans Gpa2 (Ca), and S. cerevisiae Gpa2 (Sc) proteins. Identical residues are in boldface. Changed amino acids in the sequences encoded by the different GPA2 alleles are boxed. (B) Deletion of GPA2. Shown is a Southern blot analysis using the 0.6-kbp HindIII-PstI probe located 5' upstream of GPA2 coding region. The genomic samples, digested with HindIII, were prepared from strains CAI4 (GPA2/GPA2; lane 1), CCM8 (gpa2 ${Delta}$ ::hisG-URA3-hisG/GPA2; lane 2), CCS13 (gpa2 ${Delta}$ ::hisG/GPA2; lane 3), CCM12 (gpa2 ${Delta}$ ::hisG/gpa2 ${Delta}$ ::hisG-URA3-hisG; lane 4), and CCS14 (gpa2 ${Delta}$ ::hisG/gpa2 ${Delta}$ ::hisG; lane 5). The wild type was 3.1 kbp, whereas the KO band with the URA3 blaster was 5.7 kbp. After the URA3 fragment was looped out, the band was reduced to 2.8 kbp.

Homologous recombination was used in a multistep procedure todelete both alleles of GPA2 in the SC5314-derived C. albicansstrain CAI4. The deletions were confirmed by Southern blot analyses(Fig. 1B). Strains homozygous for the GPA2 deletion were viable,and mutant cells grew at rates similar to those for wild-typecells in liquid medium. In solid medium, a slight colony sizereduction was apparent (not shown).

The ability of C. albicans cells to switch from a yeast-liketo a hyphal mode of growth was not affected by deletion of onlyone allele of GPA2 (not shown). However, deletion of both allelesblocked the yeast-to-hypha transition in specific media (Fig.2). When morphological switching was induced in Spider liquidmedium or under neutral-pH conditions (Lee's medium, pH 7),we observed that cells were defective in the formation of germtubes or filaments. In contrast, Gpa2-defective cells were ableto respond to 10% serum in liquid cultures. On solid SLAD mediumor as well as in embedded conditions, the normal formation ofhyphae was completely suppressed. These defects were correctedby the ectopic integration of the GPA2 coding sequence underthe control of the ACT1 promoter (Fig. 2).

View larger version (91K):
[in this window]
[in a new window]
FIG. 2. Defects in hyphal formation caused by deletion of GPA2. Wild-type strain CCS2 (GPA2), the isogenic disrupted strain CCS16 ( ${Delta}$ gpa2), and the isogenic disrupted strain carrying the GPA2 coding sequence under the control of the ACT1 promoter, CCS17 (gpa2/ACT-GPA2), were plated in different solid medium. YPD plates were incubated at 37°C for 3 days. SLAD plates were incubated at 37°C for 7 days. Cultures in embedded conditions were incubated at 22°C for 7 days. Log-phase cells were incubated in liquid YPD with fetal calf serum (10% final concentration), Spider medium, and pH 7 Lee's medium at 37°C. Pictures were taken after 8 h of incubation in the test medium.

Different GPA2 mutations affect filamentation. G proteins are highly conserved at both the structural and sequencelevels in all organisms, from yeast to human cells, and severalmutations in other G ${alpha}$ subunits that perturb the proper functionof the G protein have been described. To further characterizethe function of Gpa2, we introduced three independent mutationsin the GPA2 coding sequence. The associated amino acid changeshave shown both positive and negative effects in other G ${alpha}$ subunits,and they are located in specific domains involved in the changefrom the GTP form (active) to the GDP form (inactive). Two ofthem, Gly to Val at residue 354 and Gln to Leu at residue 355,are located in a loop that is hypothesized to participate inthe GTP-triggered conformational switch of GTP-binding proteins.Mutations in this domain are predicted to have either a positiveeffect, mimicking the GTP-bound state (35), or a negative effect,impairing the GTP-induced conformational change (36). The thirdmutation, Gly 142 to Val, was introduced into a loop that contactsthe phosphoryls of GDP. This mutation has different outcomesdepending on the protein: in the mammalian G_s ${alpha}$ , the associatedcoding sequence mutation produces a constitutive allele withlow activity (35), but a similar amino acid change in S. cerevisiaeGpa2 produces a gain-of-function phenotype (34, 49).

The wild-type and mutant alleles were cloned under the controlof the constitutive promoter of the actin gene from C. albicans(ACT1) (14), and they were introduced into a wild-type strainand a strain defective in Gpa2. The modified strains were analyzedfor their abilities to support hyphal development in responseto solid SLAD medium and embedded conditions (Fig. 3), as wellas liquid Spider and Lee pH 7 media (not shown). The presenceof an extra dose of GPA2 does not affect the ability of wild-typecells to respond to stimuli that induce hyphal growth.

View larger version (101K):
[in this window]
[in a new window]
FIG. 3. Effect on the hyphal response to different stimuli in C. albicans strains expressing different GPA2 mutant alleles. Shown are colony phenotypes of strains, either wild type or defective for Gpa2, carrying different GPA2 alleles integrated in the genome: CCS2 and CCS16, empty vector; CCS3 and CCS17, GPA2; CCS4 and CCS18, GPA2^G142V; CCS5 and CCS19, GPA2^G354V; CCS6 and CCS20, GPA2^Q355L. YPD plates were incubated at 37°C for 3 days. SLAD plates were incubated at 37°C for 7 days. Media associated with embedded conditions were incubated at 22°C for 7 days.

The GPA2^G142V allele was able to partially suppress the defectsassociated with the gpa2 deletion. No effect was apparent afterthe expression of this mutant allele in a GPA2 wild-type background.These data suggest that the GPA2^G142V allele is a partiallyactive allele.

The GPA2^G354V allele was unable to complement the gpa2 deletionat any condition tested. Furthermore, when expressed in a GPA2wild-type background, it impaired the ability of the cells torespond to solid-medium conditions (Fig. 3) as well as to liquidmedia (not shown). This indicates that the GPA2^G354V alleleis a loss-of-function mutant able to act as a dominant-negativeallele.

The GPA2^Q355L allele improves considerably the response in solidSLAD medium and embedded conditions both in gpa2 mutant andwild-type cells (Fig. 3). However, there was no further improvementin liquid Spider and Lee's pH 7 media with respect to the expressionof a GPA2 wild-type allele, although the mutant allele was ableto fully complement the gpa2 deletion in these media (not shown).These results are compatible with a hyperactive allele, andthey are in agreement with the gain-of-function phenotype producedby similar mutations in other fungal G ${alpha}$ subunits, such as theGpa3 protein from U. maydis, which is also involved in morphogenesis(41).

In summary, we have obtained three different GPA2 alleles, whichconfer on the cells that carry them different abilities to respondto a dimorphic switch. Because these abilities can be strictlycorrelated with the degree of activity reported for mutantswith similar mutations in other G ${alpha}$ proteins, these results reinforcea direct role for Gpa2 in the bud-hypha switch in C. albicans.

GPA2 promotes hyphal growth through the MAP kinase pathway. As mentioned above, the bud-hypha transition in C. albicansis mainly dependent on two parallel signaling pathways, definedby the transcription factors Efg1 and Cph1 (17). To examinethe relationships between Gpa2 and these pathways, we took advantageof the gain-of-function phenotype of the GPA2^Q355L allele. Weexpressed this hyperactive allele in strains defective in eitherEfg1 or Cph1, and then we analyzed the abilities to producehyphal switching in solid SLAD medium and in embedded conditions.In these media, the expression of the GPA2^Q355L allele enhancesthe bud-hypha transition in wild-type cells, as shown above(Fig. 3).

In our hands, the cph1 mutant strain had no defect in filamentousgrowth on SLAD agar but was totally unable to form filamentouscolonies in embedded conditions. Interestingly, the expressionof the GPA2^Q355L allele is unable to suppress the defects occurringin embedded conditions and does not enhance the ability to producehyphal growth in SLAD medium, with respect to the expressionof a wild-type GPA2 allele (Fig. 4A).

View larger version (83K):
[in this window]
[in a new window]
FIG. 4. Relationship between Gpa2 and the MAP kinase and cAMP-dependent pathways (A) Cells of C. albicans with functional or mutated EFG1 and CPH1genes were transformed with integrative plasmids carrying either the GPA2 wild-type allele or the GPA2^Q355L hyperactive allele under the control of the ACT1 promoter. The different strains, CCS3 (GPA2 EFG1 CPH1), CCS6 (GPA2^Q355L EFG1 CPH1), CCS8 (GPA2 EFG1 ${Delta}$ cph1), CCS9 (GPA2^Q355L EFG1 ${Delta}$ cph1), CCS11 (GPA2 ${Delta}$ efg1 CPH1), and CCS12 (GPA2^Q355L ${Delta}$ efg1 CPH1) were assayed for their abilities to support hyphal development in SLAD medium and embedded conditions. (B) C. albicans cells with both alleles of GPA2 deleted and wild-type cells were transformed with an empty control vector (vector) and an HST7-containing vector (ADH1-HST7). The resulting strains, CPJ12 (GPA2 vector), CPJ13 (GPA2 ADH-HST7), CPJ14 ( ${Delta}$ gpa2 vector), and CPJ15 ( ${Delta}$ gpa2 ADH-HST7), were assayed for hyphal growth on SLAD plates, in embedded conditions, and in liquid Spider medium as described for Fig. 2.

The efg1 mutant is totally unable to produce hyphal growth inSLAD medium, and this defect is not suppressed by the expressionof the GPA2^Q355L allele. However, filamentation in an efg1 mutantunder embedded conditions is actually improved (in agreementwith previous reports [21, 46]), and, strikingly, it is furtherenhanced by the expression of the GPA2^Q355L allele (Fig. 4A).

Since these experiments suggest that Gpa2 may mediate signalsto the filament-inducing MAP kinase signaling pathway, we triedto reverse the observed defects in a gpa2 mutant strain by introducingan overexpression plasmid carrying HST7 under the control ofthe strong ADH1 promoter. We had observed previously that astrain defective in HST7, encoding a MAP kinase proposed toact upstream of Cph1 (28), was unable to produce hyphal growthin solid SLAD medium and liquid Spider medium and in embeddedconditions and that these defects were not corrected by theexpression of the GPA2^Q355L allele (not shown). We found thatoverexpression of HST7 is able to suppress the filamentationdefects of a gpa2 mutant, supporting a role for Gpa2 upstreamof the filament-inducing MAP kinase signaling pathway (Fig.4B). Of note, overexpression of HST7 enhanced the ability ofwild-type cells to induce hyphal growth in embedded conditions,reinforcing the role of the MAP kinase signaling pathway inthe response to these conditions.

Exogenous addition of cAMP does not rescue the gpa2 defects. Studies of other fungi have demonstrated that there is a clearconnection between G ${alpha}$ subunits and cAMP metabolism (6, 15). Althoughthe results presented above locate Gpa2 upstream of the MAPkinase pathway, they do not rule out some role for Gpa2 in C.albicans in the regulation of intracellular cAMP levels. Biochemicalstudies of C. albicans implicate cAMP level increases in bud-hyphatransitions (11, 37, 38). Previous reports have shown that exogenousaddition to C. albicans cells of cAMP or dibutyryl cAMP (dbcAMP;a membrane-permeable derivative) increases the frequency ofbud-hypha transitions and is able to suppress a defect in adenylatecyclase activity (2, 42).

In our hands, addition of dbcAMP to wild-type cells enhancesthe ability to respond to SLAD solid media and produces onlya slight acceleration in hyphal formation in Lee's pH 7 andSpider liquid media. In contrast, in embedded conditions, thefilamentation is impaired after the addition of dbcAMP (Fig.5). When a similar approach was carried out with gpa2 mutantcells, there were no apparent differences in embedded conditions,liquid Spider medium, and liquid Lee's pH 7 medium between additionof dbcAMP and no addition. A slight filamentation can be seenonly in SLAD medium amended with dbcAMP (Fig. 5). These resultsindicate that, in C. albicans, Gpa2 is not related to the cAMPpathway. This conclusion agrees with our inability to suppressgpa2 defects by the overexpression of EFG1 under the controlof the regulatable PCK1 promoter (not shown).

View larger version (82K):
[in this window]
[in a new window]
FIG. 5. Addition of exogenous cAMP to gpa2 mutants. Cells from the CAI4 (GPA2) and CCM12 ( ${Delta}$ gpa2) strains were incubated in different media with (+cAMP) or without dbcAMP (10 mM final concentration). For solid media, SLAD plates were incubated at 37°C for 5 days, while for embedded conditions plates were incubated for 15 days at 22°C. For liquid assays log-phase cells were incubated in liquid Spider medium and pH 7 Lee's medium at 37°C. Pictures were taken after 8 h of incubation in the test medium.

A hyperactive RAS1 allele can bypass some of the gpa2 defects. A C. albicans homologue of Ras1 has been characterized recently(18, 30). This homologue has been shown to be required for theyeast-to-hypha switch through regulation of the MAP kinase andcAMP signaling pathways. To further examine the relationshipbetween Ras1 and Gpa2, we made use of the dominant-active RAS1^G13Vallele, which carries a mutation that locks the G protein inthe active state (18). The RAS1^G13V and GPA2^Q355L alleles wereexpressed in C. albicans gpa2 and ras1 mutants, respectively,and we studied their response to SLAD medium and embedded conditions.In our hands, the expression of the RAS1^G13V allele in wild-typecells enhanced the formation of hyphae under filament-inducingconditions such as SLAD medium or even in noninducing conditions(not shown), as has been described elsewhere (12, 18, 42). Ofnote, the expression of the RAS1^G13V hyperactive allele doesnot improve the ability of the cells to induce the hyphal switchin embedded conditions. We found that expression of the hyperactiveRAS1^G13V allele in gpa2 mutant cells suppressed the hyphal-formationdefect in SLAD medium but not in embedded conditions (Fig. 6).

View larger version (45K):
[in this window]
[in a new window]
FIG. 6. Epistasis analysis of RAS1 and GPA2. Wild-type and ${Delta}$ gpa2 cells were transformed with pCNB3-RAS1^G13V, a plasmid expressing the hyperactive RAS1^G13V allele under the control of the ACT1 promoter. Similarly, wild-type and ${Delta}$ ras1 cells were transformed with pCNB3-GPA2^Q355L, a plasmid expressing the hyperactive GPA2^Q355L allele under the control of the ACT1 promoter. The resulting strains were grown in SLAD medium and embedded conditions as described for Fig. 2.

On the other hand, the expression of the GPA2^Q355L allele inras1 mutant cells does not suppress the ras1-associated inabilityto produce hyphal growth in SLAD medium. However, the expressionof this gain-of-function allele enhances the ability of ras1mutant cells to produce hyphal growth in embedded conditions(Fig. 6).

DISCUSSION

Top

Discussion

Studies of G ${alpha}$ subunits in several saprophytic and pathogenicfungi have revealed roles for these proteins in the regulationof virulence, morphogenesis, and sexual development (5, 24).A role for GPA2-like genes in morphogenesis in S. cerevisiae,where they are required for pseudohyphal differentiation (26,34), and roles in the pathogenic fungi Cryptococcus neoformans(1) and U. maydis (41) have been well described. In this work,we have characterized the role of GPA2, a gene encoding a G ${alpha}$ subunit, in C. albicans morphogenesis. We found that cells lackingGpa2 have a defect in hyphal development in response to specificenvironmental cues. Furthermore, constitutively active mutantallele GPA2^Q355L, which is predicted to be locked in the active-conformationstate, enhances hyphal growth in response to two of the stimulitested, SLAD medium and embedded conditions. A second mutant,GPA2^G354V, which is predicted to prevent the release from ß ${gamma}$ and the interaction with signaling effectors, is a dominant-negativeallele that impairs filamentation when expressed in a wild-typestrain. Our data also indicate that Gpa2 plays a positive rolein signaling. This conclusion agrees with the fact that in aphylogenetic tree produced by multiple alignment of the fungalG ${alpha}$ subunits, C. albicans Gpa2 is located in the same subgroupas S. cerevisiae Gpa2, U. maydis Gpa3, and C. neoformans Gpa1,among others, all of them having an activating role in signaling(data not shown).

We propose that Gpa2 acts upstream of the filamentous MAP kinasecascade. This model is consistent with the observation thatthe hyphal growth defect caused by homozygous deletion of theGPA2 gene is reversed by overexpression of HST7 (the MEK homologueof the MAP kinase cascade). We also found that the hyphal enhancementcaused by the hyperactive GPA2^Q355L allele in SLAD medium andin embedded conditions is prevented by homozygous deletion ofCPH1 and HST7. Surprisingly, we found no evidence of Gpa2-mediatedsignaling through the cAMP-dependent cascade. Although the GPA2^Q355L-dependenthyphal enhancement observed in SLAD medium is totally preventedby deletion of EFG1, a synergy in the promotion of hyphal developmentwas observed when the GPA2^Q355L allele was expressed in Efg1-defectivecells in embedded conditions. In addition, in contrast to whathas been observed with other G ${alpha}$ proteins (1, 25, 26, 34), wewere unable to suppress the defect in hyphal promotion of gpa2mutant cells by the addition of exogenous cAMP. Furthermore,we were unable to compensate the deletion of GPA2 by overexpressionof EFG1 (not shown).

We have also analyzed the genetic interactions between the GTP-bindingproteins Ras1 and Gpa2. In C. albicans, Ras1 has been provedto act through the cAMP and MAP kinase pathways for the promotionof filamentous growth (30). Our epistasis analysis using RAS1and GPA2 hyperactive alleles further supports a role for Gpa2upstream of the MAP kinase cascade. The expression in Gpa2-defectivecells of the RAS1^G13V allele suppresses the hyphal growth defectin SLAD medium, while the expression of the GPA2^Q355L allelein Ras1-defective cells suppresses only slightly the defectin hyphal growth in the same growth conditions. It is well knownthat hyphal growth in SLAD medium is dependent on both cAMP-dependentand MAP kinase pathways (17).

Promotion of hyphal growth in embedded conditions deserves aseparate comment. Previous data indicated that Cph1 was requiredfor hyphal morphogenesis in response to matrix embedding (9).Our results agree with these observations. Furthermore, we foundthat the overexpression of HST7 enhances the response to embeddedconditions, while the deletion of HST7 prevented hyphal formationin these conditions. All these data support a positive rolefor the MAP kinase in the promotion of hyphal growth in embeddedconditions. Our data also suggest that a cAMP-dependent pathwayhas a negative role in hyphal growth in embedded conditions.We found that addition of cAMP impairs hyphal growth in embeddedconditions and that cells defective in Efg1, which has beenproposed to lie downstream of PKA, have an enhanced responseto embedded conditions (in agreement with previous reports [21,46]). The different effects of hyperactive Ras1 and Gpa2 genealleles in embedded conditions could be also explained by assumingopposite roles for MAP kinase and cAMP-dependent pathways inembedded conditions. In a wild-type cell, Ras1- and Gpa2-mediatedactivation would result in a balanced activation of both MAPkinase and cAMP-dependent pathways that results in the developmentof hyphal growth in embedded conditions. Defective signalingthrough the cAMP-dependent pathway (as is produced by an efg1deletion) would result in a release of the negative effectsand then an enhancement of hyphal growth. Displacement of thesignals through MAP kinase, either by specific overactivationof the MAP kinase pathway (due to, e.g., overexpression of HST7or GPA2^Q355L) or by inability to activate the cAMP pathway (whichoccurs in a ras1 mutant) would also result in an enhanced response.

In summary, we propose that in C. albicans Gpa2 is requiredfor hyphal development in response to specific environmentalcues and that this response involves the filament-inducing MAPkinase cascade. We also propose that a strict coordination withother effectors, such as Ras1, is required for a fine adjustmentof the response to environmental changes.

ACKNOWLEDGMENTS

We thank J. F. Ernst and A. J. P. Brown for sending us plasmidsand strains, and Galar Fungail Consortium by helpful discussions.

This work was supported by EU contract QLK2CT 2000-00795 andCAM 08.2/0023/1998.

FOOTNOTES

* Corresponding author. Mailing address: Department of Microbial Biotechnology, Centro Nacional de Biotecnología-CSIC, Campus de Cantoblanco-UAM, 28049 Madrid, Spain. Phone: 34 91-585 4704. Fax: 34 91-585 4506. E-mail: jperez{at}cnb.uam.es.

REFERENCES

Top