Protein kinase A encoded by TPK2 regulates dimorphism of Candida albicans


Joachim F. Ernst. E-mail; Tel. (+49) 211 811 5176; Fax (+49) 211 811 5176.


External signals induce the switch from a yeast to a hyphal growth form in the fungal pathogen Candida albicans. We demonstrate here that the catalytic subunit of a protein kinase A (PKA) isoform encoded by TPK2 is required for internal signalling leading to hyphal differentiation. TPK2 complements the growth defect of a Saccharomyces cerevisiae tpk1-3 mutant and Tpk2p is able to phosphorylate an established PKA-acceptor peptide (kemptide). Deletion of TPK2 blocks morphogenesis and partially reduces virulence, whereas TPK2 overexpression induces hyphal formation and stimulates agar invasion. The defective tpk2 phenotype is suppressed by overproduction of known signalling components, including Efg1p and Cek1p, whereas TPK2 overexpression reconstitutes the cek1 but not the efg1 phenotype. The results indicate that PKA activity of Tpk2p is an important contributing factor in regulating dimorphism of C. albicans.


The virulence of pathogenic fungi is owing to multiple fungal traits and factors that allow growth and development in the infected host. The human fungal pathogen, Candida albicans, is triggered by host conditions, for example serum components, N-acetylglucosamine (GlcNAc) and elevated temperature, to switch from a unicellular yeast form to a multicellular hyphal form that supports adhesion and penetration of host cells or tissues (for reviews, see Odds, 1988; Calderone and Braun, 1991). Hyphae continuously bud off yeast cells (blastospores), thus allowing systemic distribution and infection of organs. The dimorphic switch of C. albicans is a model of cellular differentiation and its analysis may reveal new targets for antifungal agents.

It has been shown that cAMP and protein kinase A (PKA) regulate several morphogenetic processes in fungi; cAMP is required to activate PKA by binding and inactivating its regulatory subunits. In the yeast Saccharomyces cerevisiae, pseudohyphal morphogenesis depends on elevated cAMP levels directed by the Gpa2 and Ras2 proteins (Mösch et al., 1996, 1999; Kübler et al., 1997; Lorenz and Heitman, 1997). Recently, PKA encoded by TPK2 has been shown to be essential for filamentous growth, whereas morphogenesis is inhibited by the Tpk3 and possibly the Tpk1 PKA isoforms (Robertson and Fink, 1998; Pan and Heitman, 1999). Elevated cAMP levels and PKA activity are necessary for appressorium formation of the rice blast fungus Magnaporthe grisea (Mitchell and Dean, 1995). In contrast, elevated cAMP levels are associated with budding growth and low cAMP levels are correlated with filamentous growth of the corn smut fungus Ustilago maydis (Dürrenberger et al., 1998). In C. albicans, elements of the cAMP pathway, including a Ras protein and PKA, have not been defined, although cAMP addition was shown to augment filamentation (Sabie and Gadd, 1992; Niimi, 1996). Target proteins activated by PKA that regulate morphogenetic responses in fungi are unknown. Recently, the Sfl1 protein was shown to interact with the S. cerevisiae Tpk2 protein and to be situated downstream of Tpk2p in the regulation of FLO11, which encodes a cell-surface component required for pseudohyphal development (Robertson and Fink, 1998); in addition, PKA activation of FLO11 appears to be mediated by the transcription factor Flo8p (Pan and Heitman, 1999; Rupp et al., 1999).

Signalling pathways leading to hyphal development of C. albicans are poorly defined. The Efg1 protein is an essential regulator of hyphal formation in most inducing conditions (Lo et al., 1997; Stoldt et al., 1997) and belongs to a conserved class of bHLH transcription factors regulating morphogenesis in fungi (Dutton et al., 1997; Stoldt et al., 1997), referred to as APSES proteins (Aramayo et al., 1996). Recently, we discovered that Efg1p is also an important element of other morphogenetic events, including spontaneous phenotypic switching, and of chlamydospore formation (Sonneborn et al., 1999a,b). Elements or pathways regulating Efg1p activity and Efg1p target genes are unknown. Compared with Efg1p, components of a conserved MAP kinase pathway, including Cst20p, Hst7p and Cph1p, appear less significant for C. albicans hyphal morphogenesis because mutants lacking these proteins are able to form hyphae in the presence of serum and other induction conditions (Liu et al., 1994; Köhler and Fink, 1996; Leberer et al., 1996; Csank et al., 1998). Similarly, with regard to virulence in the mouse model of systemic infection, the efg1 mutation has a stronger attenuating effect (Lo et al., 1997) compared with mutations of the MAP kinase pathway (Leberer et al., 1996; Csank et al., 1998).

Here, we describe the characterization of Tpk2p, an isoform of a PKA catalytic subunit in C. albicans and show its importance for hyphal development. By complementation of a PKA-deficient S. cerevisiae strain and in vitro phosphorylation of an authentic PKA substrate, we demonstrate that Tpk2p is a typical A-type kinase. We present evidence that elevated expression of TPK2 stimulates morphogenesis and invasive properties, whereas wild-type levels of Tpk2p are required for efficient morphogenesis from the yeast to the hyphal growth form as well as for full virulence of C. albicans.


TPK2 encodes a PKA homologue

TPK2 was localized on a plasmid and a fosmid clone constructed in the Candida genomic mapping and sequencing project ( html) (see Experimental procedures). The sequence of a 3.3 kb SacI–PstI genomic C. albicans fragment revealed an open reading frame (ORF) with the potential to encode a catalytic subunit of PKA (accession no. AF134300). Because the length of the deduced protein was 442 residues, which is unusual for a PKA catalytic subunit, we determined the transcript 5′ end by rapid amplification of cDNA ends (RACE). Two transcript 5′ ends were found, both of which are situated between the first and second ATGs of the ORF (Fig. 1). This result suggested that the second ATG of the ORF corresponds to the translational initiation codon. Based on this ATG, the size of the theoretical PKA is 411 residues (45 kDa).

Figure 1.

. Localization of TPK2 transcript 5′ ends. The open reading frame conceptually encoding Tpk2p, which is situated on the 3.3 kb genomic fragment (S, SacI; P, PstI), is shown on top. To decide whether only part of the ORF (black region) or the full-length ORF (black and grey regions) are translated, the 5′ ends of the TPK2 transcripts were determined using 5′ RACE. Transcript ends are situated between the two possible ATG translational initiation sequences (arrows at bottom), indicating that the translated sequences start at the second ATG.

The deduced PKA is 65%, 78% and 65% identical to the Tpk1, Tpk2 and Tpk3 proteins respectively, which represent the PKA isoforms (catalytic subunits) of S. cerevisiae (Toda et al., 1987). For this reason, the deduced PKA was designated Tpk2p. Tpk2p has 48% and 49% identity to the mouse and human Cα PKA subunits (Uhler et al., 1986; Maldonado and Hanks, 1988). Tpk2p contains the GTGSFG sequence of the nucleotide binding loop, the RDLKPEN sequence of the catalytic loop and the TDFGFAK sequence involved in Mg2+ binding, which are conserved among PKA isoforms (Knighton et al., 1991). Tpk2p also contains identical or homologous residues required for the recognition of the consensus phosphorylation sequence. Residues L256, P260 and L263 bind the hydrophobic residue in the +1 position relative to the phosphorylation site, residues E228 and E288 bind R in the −2 position and residue E185 binds R in the −3 position (Knighton et al., 1991). A striking feature of the 29 N-terminal residues of Tpk2p is its high content of glutamate residues. Glutamine stretches, which can be important for protein–protein interactions, have also been observed in some other protein kinases of fungi and slime moulds (Haribabu and Dottin, 1991; Buhr et al., 1996).

Tpk2p has PKA activity

To verify that Tpk2p indeed encoded an A-type kinase, we constructed an Escherichia coli expression vector encoding a fusion of six consecutive N-terminal histidine residues to Tpk2p. However, (His)6-Tpk2 protein isolated from cell extracts was inactive in PKA enzymatic assays (not shown). Therefore, we constructed expression vector pBI-HT, designed to express the (His)6-Tpk2 fusion under transcriptional control of the PCK1 promoter (Leuker et al., 1997) in the homologous host, C. albicans. Immunoblottings using an anti-His6 antibody revealed the presence of the fusion protein in the soluble fraction of strain CAI4[pBI-HT] cell extracts, which in SDS–PAGE migrated at about 50 kDa. The protein was enriched by absorption to a Ni-NTA column, as revealed by immunoblotting (Fig. 2A). Using a commercial non-radioactive PKA assay (Promega), in which a fluorescent dye-tagged ‘kemptide’ peptide serves as the acceptor, we demonstrated elevated PKA activities in elution fractions compared with identical elution fractions of control strain CAI4[pBI-1]. In this assay, phosphorylated kemptide peptide migrates to the anode, while the unphosphorylated derivative migrates to the cathode.

To obtain further evidence for the function of Tpk2p as a PKA, we placed the 3.3 SacI–PstI genomic TPK2 fragment into a centromeric and a multicopy yeast transformation vector (pCENTPK2, p2μTPK2) and transformed these plasmids into the mutant S. cerevisiae host strain SGY446, which only contains a single TPK2 allele encoding a temperature-sensitive Tpk2p variant (Smith et al., 1998). As shown in 2Fig. 2B, this mutant, if it was transformed with both TPK2 expression vectors, was able to grow at the elevated temperature, indicating that Tpk2p can functionally replace PKA proteins in S. cerevisiae. Furthermore, expression of TPK2 under control of the GAL1 promoter in S. cerevisiae (pGalTPK2) led to slow growth on galactose medium, suggesting partial toxicity at high-expression levels, a characteristic that is also known for S. cerevisiae Tpk proteins. As expected, the temperature sensitivity of strain SGY446 was also complemented by expression vectors carrying the S. cerevisiae TPK1 gene or a fusion of the GAL1 promoter to the S. cerevisiae TPK2 gene (data not shown). Thus, the sequence identities to known PKA proteins, its function in a PKA enzymatic assay and the complementation of a defined PKA defect indicate that Tpk2p represents an isoform of a PKA catalytic subunit in C. albicans.

Figure 2.

. Tpk2p has PKA activity. A. A His6–Tpk2 fusion protein in extracts of C. albicans strain CAI4[pBI-HT] was enriched by adsorption to a Ni-NTA column. Strain CAI4[pBI-1] was used as a control. Proteins were eluted sequentially with 200 mM (1), 250 mM (2) and 300 mM imidazole. Fifty microlitres of the elution fractions was analysed by immunoblotting (top) and 10 μl of each fraction was tested for PKA activity (bottom). B. TPK2 complements the PKA deficiency of strain S. cerevisiae SGY446 (Smith et al., 1998). Transformants carrying control vectors YCplac33 (lanes 1 and 2) or YEplac195 (lanes 3–5) are unable to grow at 37°C on SD medium, at which no functional PKA isoform is produced; transformants carrying pCENTPK2 (lanes 6–8) or p2μTPK2 (lanes 9–12) are able to grow at 37°C.

Hyphal morphogenesis is blocked in tpk2 mutants

To explore the function of PKA encoded by TPK2, we deleted both genomic alleles (Fig. 3A). A disruption cassette containing the ‘ura-blaster’ flanked by untranslated regions of TPK2 was used to replace wild-type TPK2 sequences by homologous recombination (Fonzi and Irwin, 1993). Genomic DNA of wild-type cells cut with PstI revealed three fragments with a TPK2 probe. The 1.7 kb fragment is the result of one allele and the 1.2 kb and 0.5 kb fragments arise because of an additional PstI site (asterisk in Fig. 3A) in the coding region of the second allele. After disruption of one TPK2 allele (strains TPK3 and TPK7), either the 1.7 kb or the 1.2/0.5 kb fragment was lost and a larger fragment (12 kb or 9 kb) representing the disrupted allele appeared (Fig. 3A, lanes 2 and 3). Selection on fluoro-orotic acid (FOA)-containing medium of strain TPK7 led to the isolation of Ura cells, in which URA3 of the ‘ura-blaster’ had been deleted by recombination between the hisG repeats. In this strain, TPO7, the disruption fragment, was reduced from 12 kb to 9 kb (Fig. 3A, lane 4). The derivatives of this strain (TPO7.4 and TPO7.5), in which the second allele also was disrupted by the ura-blaster, only contained the 12 kb and 9 kb fragments (Fig. 3A, lanes 5 and 6). Strains AS1 to AS4, the Ura derivatives of TPO7.4, only contained the 9 kb band representing the tpk2 disrupted by one copy of hisG (Fig. 3A, lanes 7–10).

Figure 3.

. Defective hyphal development in strains lacking TPK2. A. Deletion of TPK2 alleles. The scheme (top) demonstrates the replacement of a wild-type allele by a disruption fragment containing the ura-blaster (hisG–URA3–hisG) flanked by homologous sequences. The deletions were verified by Southern blotting (bottom) using the indicated probe. Control strain CAI4 (TPK2/TPK2, lane 1); heterozygous strains TPK3 and TPK7 (TPK2/tpk2::hisG–URA3–hisG, lanes 2 and 3); Ura derivative TPO7 (TPK2/tpk2::hisG, lane 4); homozygous mutants TPO7.4 and TPO7.5 (tpk2::hisG–URA3–hisG/tpk2::hisG, lanes 5 and 6) and Ura derivatives AS1–4 (tpk2::hisG/tpk2::hisG, lanes 7–10). The PstI site marked by the asterisk is present only in one Tpk2 allele. B. Hyphal development in liquid Spider medium. The control strain CAI4[pBI-1] (diamonds) and the tpk2 disruption strain TPO7.4 (triangles) were induced to form hyphae in liquid Spider medium at 37°C. At the indicated times, the percentage of cells carrying hyphae was determined microscopically. Values represent the mean of at least two independent measurements; values of the control strain varied less than 3% of the mean. C. Hyphal formation on solid media. Cells were grown on 5% horse serum (upper row) or on Spider medium (bottom row) at 30°C. Strains tested were the control strain CAI4[pBI-1] (TPK2/TPK2), the half-disruptant strain TPK7 (TPK2/tpk2), the complete disruptant strain TPO7.4 (tpk2/tpk2), the complete disruptant strain AS1 transformed with the control vector pRC2312 [tpk2/tpk2 (vector)] or the expression vector pRC2312CaTPK2 [tpk2/tpk2 (TPK2)].

Deletion of TPK2 did not cause differences in growth rates in any liquid or on any solid medium tested. However, hyphal development in liquid medium (Spider medium) at 37°C was significantly delayed in strains lacking both TPK2 alleles compared with wild-type cells (Fig. 3B). Both strains TPO7.4 and TPO7.5 only formed about 50% hyphae compared with the wild-type control; this difference was apparent even after long induction times (12 h), after which wild-type cells, but not tpk2 mutants, strongly flocculated because of hyphal formation. In liquid medium containing 5% horse serum at 37°C, only slight differences with respect to hyphal development were observed between tpk2 mutants and wild-type strains. In contrast, on solid medium containing 5% serum (Fig. 3C, upper row) or on Spider medium (Fig. 3C, lower row) incubated at 30°C, the effect of deletions in both TPK2 alleles (strains TPO7.4 and AS1 transformed with the control vector pRC2312) was clearly apparent; the deletion of only one TPK2 allele (strain TPK7) had no defective phenotype. Although hyphal development on both types of solid media at 30°C was strongly compromised in tpk2 mutants, this phenotype was reduced or not observed at 37°C.

Therefore, hyphal morphogenesis is not blocked completely by the tpk2 mutation and its phenotypic manifestation depends on other factors, such as temperature and growth on a solid surface.

Overexpression of TPK2 induces filamentation and invasive growth

If PKA was limiting hyphal formation in non-inducing media, it could be expected that overexpression of TPK2 would induce filamentation. To test this possibility, we transformed strain CAI8 with an episomal vector carrying a genomic TPK2 fragment (pATA). Such transformants produced elongated, rod-like cells that often extended into filaments in liquid S4D and SCAA media at 30°C, conditions that did not induce morphogenesis in control cells carrying an empty vector (Fig. 4A). The filaments that were formed were septated tubular structures resembling true hyphae except that they tended to lack side branches and lateral buds. On solid SCAA medium, pATA transformants produced colonies surrounded by a halo of hyphal filaments; again, such hyphae consisted essentially of straight filaments with poor formation of lateral buds or branches. No obvious filamentation was observed in pATA transformants after growth on solid SD medium (Fig. 4B). However, although control transformants were easily washed off the agar surface of SD medium, pATA transformants persisted because they had grown invasively into the agar layer (Fig. 4C). Microscopic inspection of cells invading the agar demonstrated that they grew predominantly as unbranched hyphae; even about 20% of cells in the colony on top of the SD agar formed such filaments, besides growing as regular yeast cells.

Figure 4.

. Overexpression of TPK2 induces filamentation. Strain CAI8 was transformed with the TPK2 expression vector pATA (pCaTPK2) or with control vector pBT-4 (control); the ura3 defect was complemented by plasmid pBI-1 in these transformants. Cells were grown in liquid S4D medium (A) or on solid SCAA and SD media (B). Cells growing on top of SD medium (C, a) were washed off with water (C, b) to demonstrate invasive growth of transformants overexpressing TPK2 (1–4) compared with non-invasive growth of control transformants (control).

These results suggest that PKA activity limits hyphal growth in non-inducing environmental conditions and that Tpk2p is able to act as an inducer of hyphal morphogenesis and of invasive growth.

Epistatic relations of morphogenetic phenotypes

To obtain additional evidence for the function of Tpk2p, we determined the phenotypes generated by the interaction of deleted or overexpressed alleles encoding signalling components (epistasis).

We tested whether defective morphogenesis in strains carrying deletions of tpk2, cek1 or efg1 [strains AS1, CK43B-16L (Csank et al., 1998) and HLC67 (Lo et al., 1997) respectively] could be suppressed by overexpression of genes encoding signalling components (Fig. 5A). In these experiments, overexpression of EFG1 and TPK2 was achieved by placing these genes, under transcriptional control of the PCK1 promoter (Leuker et al., 1997), into a replicating vector, whereas overexpression of CEK1 was achieved by inserting a genomic copy into a replicating vector (Csank et al., 1998). The defective hyphal formation of strain HLC67 on Spider medium was suppressed only by the EFG1 gene, which is mutated in this strain; other overexpressed genes, including TPK2 and CEK1, did not reconstitute hyphal formation. In contrast, the hypha-negative phenotype of tpk2 mutants could not only be suppressed by TPK2, as expected, but also by overexpression of EFG1 and CEK1. On the contrary, the cek1 phenotype (which did not abolish filamentation completely) could not only be suppressed by CEK1 but also by overexpression of TPK2 and EFG1. These results demonstrated that high activities of either of the two kinases PKA (Tpk2p) or MAP kinase (Cek1p) could compensate for loss of the other kinase with respect to hyphal formation. Because an efg1 mutation cannot be suppressed by overexpression of Cek1p or Tpk2p, Efg1p must function either downstream of or in parallel with both kinases.

Figure 5.

. Suppression of morphogenetic defects. C. albicans mutants defective for hyphal formation were transformed with expression vectors for the indicated genes; transformants were grown on solid Spider medium for 5 days. Mutants include HLC67 (efg1), AS1 (tpk2) and CK43B-16L (cek1); expression vectors were pBI-HAHYD (EFG1), pBI-TPK (TPK2) and pCCa4 (CEK1), and the empty vector pRC2312 was used as the control.

Virulence tests

The above experiments had indicated that TPK2 is required for hyphal morphogenesis in certain inducing conditions. To verify whether TPK2 was relevant in vivo, in an experimental model of systemic infection, we injected 105C. albicans cells into the tail vein of 10 mice and monitored their survival. Transformants of C. albicans strain AS1 (tpk2/tpk2) were tested either carrying the control vector pRC2312 or plasmid pRC2312CaTPK2, which reconstitutes its morphogenesis phenotype (Fig. 3).

In both groups of infected mice, the mortality was 60% after 4 days (Fig. 6). In the group infected with the transformant carrying pRC2312CaTPK2, all of the remaining mice died during the following days. In contrast, all mice infected with the tpk2 mutant that were alive at day 4 continued to live to the end of the experiment (18 days). Internal organs of animals dying in both groups were colonized with C. albicans at 105 cfu (spleen) and 107 cfu (kidneys) per gram of tissue. Among the survivors infected by the tpk2 mutant, two mice had only one infected kidney and a sterile spleen, whereas a third mouse had both kidneys and spleen infected.

Figure 6.

. Virulence of C. albicans strains in mice. Ten Swiss mice were injected with 105C. albicans cells in their tail veins and the survival of the animals was monitored. Transformants of strain AS1 (tpk2/tpk2) were used for infection, either carrying a plasmid carrying TPK2 (pRC2312CaTPK2) (top) or an empty control vector (pRC2312) (bottom)

Thus, although deletion of TPK2 does not affect virulence drastically in the mouse model, a partial reduction of virulence was observed. At the dosage chosen (105 cells), the survival of mice infected by the tpk2 mutant is higher than the survival of mice infected with cph1 mutants, which lack an essential element of the MAP kinase pathway (Lo et al., 1997).


C. albicans translates certain extracellular signals into changes in gene expression leading to either a yeast or a hyphal growth form. Results in this paper indicate that a catalytic subunit of a cAMP-dependent kinase (Tpk2p) regulates hyphal morphogenesis in C. albicans. Tpk2p is highly homologous to A-type kinases in other organisms and in particular shares conserved residues required for binding of ATP, Mg2+ and the protein acceptor sequence (as detailed in the Results section). Expression of TPK2 in the heterologous host S. cerevisiae complements the growth defects of a tpk1-3 triple mutant and Tpk2p produced in C. albicans is able to phosphorylate the ‘kemptide’ PKA-acceptor peptide. These results indicate that Tpk2p actually functions as an A-type kinase. Previously, a protein with PKA activity has been isolated from extracts of C. albicans, but this enzyme has an unusual size for a PKA (78 kDa) and, therefore, is not identical to the Tpk2 enzyme we describe here (Zelada et al., 1998). Our results strongly suggest that Tpk2p is a member of a signalling pathway that operates in parallel with a pathway containing a conserved MAP kinase (Cek1p).

Deletion of both TPK2 alleles caused a clear block of morphogenesis in several inducing media, including serum-containing solid medium. However, the effect of the tpk2 mutation was reduced or not apparent in optimized induction conditions, in which additional factors, for example elevated temperature, could provide stimuli for morphogenesis. The morphogenesis of tpk2 mutants was severely compromised on solid serum or solid Spider medium at 30°C but not at 37°C, and mutants were defective in liquid but not solid Spider medium at 37°C. It should be pointed out that the morphogenesis-defective phenotype of the tpk2 mutants is more pronounced than in mutants lacking components of the MAP kinase pathway, including Hst7p, Cek1p or Cph1p (Liu et al., 1994; Köhler and Fink, 1996; Leberer et al., 1996; Csank et al., 1998), because hyphal morphogenesis of tpk2 mutants was blocked in liquid medium and when using serum as the inducer. Nevertheless, Tpk2p may be required only in some suboptimal induction conditions or, alternatively, additional signals in optimal conditions recruit new induction components and/or pathways such that Tpk2p becomes dispensable. The Candida genome project has revealed a gene for another putative PKA isoform, designated TPK1, that may contribute to cellular PKA activity ( The presence of at least two TPK genes may also explain why the tpk2 mutant is fully viable, as in S. cerevisiae the presence of at least one of three PKA isoforms encoded by the TPK1–3 genes is required for growth (Toda et al., 1987). In agreement with a partial block in morphogenesis, tpk2 mutants showed reduced virulence in a mouse model of systemic infection. Similarly, it has been reported that the virulence of mutants lacking components of a conserved MAP kinase pathway is attenuated (Köhler and Fink, 1996; Leberer et al., 1996; Csank et al., 1998).

High levels of Tpk2p activity in conditions not normally leading to hyphal induction led to strong filamentation, suggesting that low PKA activity levels limit hyphal morphogenesis in these conditions. Although the TPK2-induced filaments resembled hyphae formed by wild-type cells with respect to septae formation, diameter and length, they nevertheless were different in that lateral buds (yeast form) or lateral hyphae arose less often. It appears possible that in wild-type cells, at some distance from the growing tip of the filament, TPK2 expression is downregulated to allow lateral budding; downregulation cannot occur in the TPK2-overexpressing strain. Another phenotype associated with TPK2 overexpression is the formation of irregular-shaped colonies on solid medium (not round colonies as for the wild type). Thus, PKA-induced filamentation may be directly correlated to the invasive properties of C. albicans. The question of which molecular mechanisms lead to an upregulation of TPK2 expression in wild-type cells grown in inducing conditions is open. It is known that PKA activity is due to a catalytic subunit that, in the absence of cAMP, is inactivated by a regulatory subunit. Binding of cAMP to the regulatory subunit triggers its release from the catalytic subunit, which becomes enzymatically active. C. albicans genes encoding the PKA regulatory subunits have been identified (L. Giasson, personal communication). As in the case of S. cerevisiae, signals further upstream of cAMP and Tpk2p in the morphogenetic pathway of C. albicans may include one or more Ras proteins. In fact, a RAS gene required for morphogenesis has been identified recently (E. Leberer, personal communication).

Which components are regulated by Tpk2p in the conceptual Ras–cAMP–Tpk2p pathway? Previous work revealed that the Efg1 protein of C. albicans, which belongs to a conserved group of bHLH proteins regulating morphogenesis in fungi (APSES proteins), is essential for hyphal development (Lo et al., 1997; Stoldt et al., 1997), but a signalling pathway comprising Efg1p is not yet known. Interestingly, Efg1p contains a single RVT sequence, corresponding to the consensus sequence for phosphorylation by PKA (Pearson and Kemp, 1991), between residues 204 and 206, which is conserved among APSES proteins with the exception of the S. cerevisiae Phd1 protein. We recently discovered that mutants with a T206→A exchange in Efg1p are blocked in hyphal formation (D. P. Bockmühl, A. Sonneborn, J. F. Ernst, unpublished), suggesting that Efg1p is a direct downstream target of Tpk2p. The fact that the defective phenotype of a tpk2 mutation is completely suppressed by EFG1 overexpression, whereas an efg1 mutation is not suppressed by TPK2 overexpression, is consistent with this theory. Because a deficiency in Efg1p is not suppressed by CEK1 overexpression, it appears that Efg1p is absolutely required for morphogenesis, even in tpk2 mutants in which a low level of Efg1p phosphorylation may still be present as a result of still undefined A-type kinases. Other data also support the theory that PKA and APSES proteins are in a common signalling pathway. Heterologous production of a C. albicans chaperonin subunit (Cct8p) in S. cerevisiae blocked pseudohyphal growth only if it was induced by an activated Ras2 protein, but not if it was induced by hyperactivation of the MAP kinase pathway. In addition, Cct8p acted as a dominant suppressor of pseudohyphal development induced by both the homologous APSES protein Phd1p and the heterologous Efg1 protein (Rademacher et al., 1998). Because morphogenesis-unrelated phenotypes of an activated Ras2p pathway were inhibited also by Cct8p, it appears that PKA is the target of Cct8p suppression, which may sequester but not fold PKA isoforms. Thus, this pattern of suppression suggests that PKA and APSES proteins are in a common pathway different from the MAP kinase pathway. Current models of morphogenesis in S. cerevisiae and C. albicans suggest a pair of signalling pathways, one of which contains an MAP kinase regulating the Ste12 or, respectively, the Cph1 transcription factor and the other containing PKA isoforms (Robertson and Fink, 1998), which we suggest to contain the APSES proteins Phd1p/Sok2p or Efg1p. In S. cerevisiae, high PKA activity and PHD1 overexpression are able to overcome the ste12 defects (Lo et al., 1997; Pan and Heitman, 1999). Furthermore, high PKA activity is able to activate a promoter responding to Ste12p (Mösch et al., 1999). These findings are in agreement with our finding that the cek1 mutation is suppressed by TPK2 in C. albicans. At the present time, the extent and the mechanisms of cross-talk between both signalling pathways, which could explain, for example, why CEK1 overexpression suppresses the tpk2 defect, are unclear. Possible mechanisms could include known members of signalling pathways as well as modulators of the pathways, such as the Cpp1p protein phosphatase (Csank et al., 1997). In addition, both pathways could also converge downstream of Ste12/Cph1 and APSES proteins to integrating components (Rupp et al., 1999).

Experimental procedures

Strains and media

Strains are listed in Table 1. Transformation of S. cerevisiae was performed using the lithium acetate method and of C. albicans using the spheroplast method (Sherman et al., 1986). Yeast cells were grown in YPD or on supplemented SD minimal medium (Sherman et al., 1986) or S4D medium (as SD, but containing 4% glucose). Spider medium, containing 1% mannitol, 1% nutrient broth, 0.2% K2HPO4, pH 7.2, has been described previously (Liu et al., 1994). The PCK1 promoter was induced in SCAA medium [0.67% yeast nitrogen base without amino acids (Difco); 2% casamino acids; Leuker et al., 1997]. Hyphae were induced in C. albicans yeast cells by pregrowth of strains at 30°C in YPD to the logarithmic growth phase. Cells were harvested by centrifugation, washed in water and starved for 1 h at 30°C before dilution (OD600 = 0.1) into 5% horse serum, Spider medium or salt base (0.45% NaCl, 0.335% yeast nitrogen base without amino acids) containing 2.5 mM N-acetylglucosamine (GlcNAc) (Delbrück and Ernst, 1993).

Table 1. . Strains.Thumbnail image of

Cloning, sequencing and expression of TPK2

Plasmid p2191, which contains a 1.7 kb genomic PstI fragment inserted in pUC18, and fosmid 1D7 were identified during the Candida genomic sequencing project (; both vectors were kindly supplied by B. B. Magee. Because p2191 did not contain the entire TPK2 coding region, the insert of fosmid 1D7 was subcloned by inserting a 8.5 kb EcoRV fragment into the EcoRV site of pUC21. The insert in the resulting plasmid pUC21/RV-FOS-A was subcloned again by inserting its 1.6 kb SacI fragment (which contained the 5′ portion of TPK2 as shown by Southern analysis) into the SacI site of pUC18 (pTPK2-promB). The sequences of both CaTPK2 fragments were determined using various subclones.

TPK2 was assembled by inserting the 1.7 kb SacI–PstI fragment into pARS/AatII, which contains the CARS replicator in pUC18. The 1.6 kb SacI fragment was cloned into the resulting vector. The vector, which contains the 3.3 kb TPK2 fragment, was completed by inserting the 2.25 kb EcoRV fragment carrying ADE2 into the single EheI site, resulting in expression vector pATA. The assembled TPK2, as a filled-in 3.3 kb EcoRI–HindIII fragment, was also inserted into the SmaI site of pRC2312 to generate pRC2312CaTPK2. To place TPK2 under control of the regulatable PCK1 promoter, the coding region of TPK2 was amplified using PCR with pATA as template and the primers TPK-1 (5′-TTAGGA TCCATCAATGGACAATCATC) and TPK-2 (5′-TTAAGATCT CCTCTCAATTC) (BamHI and BglII sites underlined; ATG start codon, bold). The 1.25 kb BamHI–BglII fragment was cloned first into SmaI site of pUC18 using the SureClone kit (Pharmacia), thus generating plasmid pUC/TPK2. This fragment was also cloned into the single BglII site of pBI-1 downstream of the PCK1 promoter to generate pBI/TPK. An expression vector containing the ADE2 selection marker pARS/PTA was constructed by inserting the BamHI–BglII PCK1p–TPK2 fusion fragment from pBI/TPK into the BamHI site of pARS/AatII, followed by insertion of the ADE2 fragment as described above. Two expression vectors for S. cerevisiae, pCENTPK2 and p2μTPK2, were constructed by inserting the 3.3 kb SacI–PstI fragment into plasmids YCplac33 and YEplac195 respectively (Gietz and Sugino, 1988). TPK2 was also placed under transcriptional control of the S. cerevisiae GAL1 promoter by inserting the 1.27 kb BamHI–BglII fragment into the single BamHI site of YEplacGal1/10 (Rademacher et al., 1998).

An E. coli expression vector, pQE-TPK2, was constructed by inserting the 1.25 kb BamHI TPK2 fragment of PUC/TPK2 (downstream BamHI site is derived from pUC18) into the single BamHI site of pQE-10 (Qiagen). This vector directs the expression of an enzymatically inactive (His)6 fusion to Tpk2p in E. coli (not shown). For production of (His)6–Tpk2p in C. albicans, a PCR was performed using primers C.TPK2(pQE10) (5′-TTAAGATCTATGAGAGGATCTC) and TPK-2, which generates a fusion gene flanked downstream and upstream of the coding region (bold, translational start) by BglII sites (underlined). The resulting fragment was inserted into the single BamHI site of pBI-1, downstream of the PCK1 promoter (Stoldt et al., 1997), to generate expression vector pBI-HT.

Other plasmids

An expression vector for a tagged version of EFG1, pBI-HAHYD, was constructed in several steps. First, the 2.3 kb BamHI–SphI fragment of pBIST (Stoldt et al., 1997), which carries the EFG1 coding region, was ligated with the large BamHI–SpeI fragment of plasmid YCpIF17 (Foreman and Davis, 1994); the ligation of SphI and SpeI sites was aided by an adapter (5′-CTAGCATG). The resulting plasmid, YCpIF17-BIST, encodes Efg1p fused at its N-terminus to 25 amino acids containing the HA epitope; its expression in S. cerevisiae is directed by the GAL1 promoter. The HA–EFG1 fusion gene was amplified using PCR with primers HA-HYD1-P1 (5′-TTTAAGATCTAATGAGTCGATACCCAT AC) and HA-HYD1-P2 (5′-TTTAAGATCTGCGAATTGGAG CTCCAC), which introduce BglII sites (underlined) upstream of the HA–EFG1 coding region (bold) at a position corresponding to the SphI site in the 3′ untranslated region of EFG1. The 2.3 kb BglII PCR fragment was inserted into the single BglII site of pBI-1, downstream of the PCK1 promoter, to create plasmid pBI-HAHYD.

The expression vectors for CEK1, pCCa4 and, for CPH1, pLJ19 (Csank et al., 1998), as well as the empty control vector pRC2312 (Cannon et al., 1992), have been described previously.

Rapid amplification of cDNA ends (RACE)

Poly(A) RNA of strain SC5314 was isolated and cDNA was prepared according to the manufacturer's instructions (Boehringer Mannheim; 5′/3′RACE kit) using the TPK2-specific primer SP1 (5′-GACTGTGCAAATACTCCAAAGC). The cDNA was tailed with terminal transferase and the TPK2 cDNA was amplified by two rounds of PCR using two nested primers (SP2: 5′-GCTACTGGATTGGGAAATCTCTGAGAC; SP3: 5′-CAGTAGAACGTTCAGGCAACAACGATC) in combination with the PCR anchor primer. cDNAs were inserted into the SmaI site of pUC18 using the SureClone system (Pharmacia) and sequenced.

Disruption of CaTPK2

To construct a gene disruption cassette, the 5′ and 3′ regions flanking the TPK2 coding region were amplified using PCR with pATA as template. The 5′ region was amplified using primers TPK-dis1 (5′-ATTAGATCTACCACACTATTTTGA) and TPK-dis2 (5′-ATTGGATCCATTGATGTGTTTGAG) (BamHI and BglII sites, underlined) and the BamHI–BglII fragment was inserted into the BglII site next to the ‘ura-blaster’ in p5921 (resulting vector B14/CaTPKdisI). The 3′ region, which was amplified using PCR with primers TPK-dis3 (5′-ATTAGATCTGTATTTCCTTGACTTTTG) and TPK-dis4 (5′-ATTGGATCCTACTTTACCTCG) (BglII and BamHI, underlined), was inserted as a BglII–BamHI fragment into the single BamHI site of the resulting vector. In the resulting plasmid, pCaTPK2/blast, the ‘ura-blaster’ is flanked by 5′ and 3′ untranslated regions of TPK2. The BglII–BamHI disruption fragment was isolated and used to transform strain CAI4 to uridine prototrophy. Subsequent steps, including deletion of the URA3 selection marker and disruption of the second TPK2 allele, were as described previously (Fonzi and Irwin, 1993). The correct integration of the disruption cassette was verified by means of a Southern blot using DNA of transformants cut with PstI using the 1.7 kb PstI fragment of p2191 (Fig. 2).

Production and enzymatic assay of (His)6–Tpk2p

Strain CAI4 transformed with pBI-HT was grown in 100 ml of SCAA medium to an OD600 = 1. Cells were pelleted by centrifugation and washed in 20 mM MOPS, pH 7.0, 150 mM NaCl. Cells were resuspended in 300 μl of ice-cold breaking buffer (20 mM MOPS, pH 7.0, 300 mM NaCl, 10 mM imidazole) containing protease inhibitors [one tablet ‘complete mini’ protease mix (Boehringer Mannheim) per 10 ml of lysis buffer] and disrupted by shaking with glass beads for 10 min. The cell extract was centrifuged for 10 min at 13 000 r.p.m. in a microfuge and the supernatant was collected. To enrich the His6–Tpk2 fusion proteins of the crude extract were adsorbed to Ni-NTA spin-columns according to a commercial protocol (Qiagen). The columns were washed three times with 20 mM MOPS, 300 mM NaCl and 20 mM imidazole, and proteins were eluted stepwise with increasing imidazole concentrations (50, 100, 150, 200, 250, 300 mM). Fractions of 200 μl were collected, of which 50 μl was separated by SDS–PAGE (12.5% acrylamide). The His6–Tpk2 fusion was visualized by immunoblotting [using the anti-RGS-His-antibody (Qiagen) at 1:2000 dilution]. In parallel, equal amounts of proteins in the crude extracts of control strain CAI4[pBI-1] were separated. Ten microlitres of the elution fractions was used in a commercial assay for PKA activity using fluorescent dye-coupled kemptide peptide as the phosphoacceptor (Promega). Protein concentrations were determined using the Bradford assay (BioRad).

Virulence tests

Animal experiments were performed using 3-week-old Swiss mice. Transformants of C. albicans strain AS1 (tpk2/tpk2), either carrying the control vector pRC2312 or plasmid pRC2312CaTPK2, were cultivated in Sabouraud medium and inoculated intravenously in the tail vein. Each strain was used to infect 10 mice. Survival of mice was recorded over a period of 18 days. Animal experiments were performed according to the Swiss Federal regulations on animal protection.


We thank P. T. Magee and B. B. Magee for distribution of fosmids and the TPK2 clone. We are grateful to F. Herberg and B. Zimmermann for advice on the PKA assay and thank B. Strickling for contributing to plasmid constructions. We acknowledge C. Csank, G. Fink, J. Koehler, H. Liu and E. Leberer for generously supplying plasmids and strains. Supported by the Deutsche Forschungsgemeinschaft and grant BMH4-CT96-0310 of the European community.