Tec1p-Independent Activation of a Hypha-Associated Candida albicans Virulence Gene during Infection

Abstract

The Tec1p transcription factor is involved in the expression of hypha-specific genes in Candida albicans. Although the induction of the hypha-associated SAP5 gene by serum in vitro depends on Tec1p, deletion of all Tec1p binding site consensus sequences from the SAP5 promoter did not affect its activation. In two different animal models of candidiasis, the SAP5 promoter was induced even in a Δtec1 deletion mutant, demonstrating that the requirement for Tec1p in gene expression in C. albicans depends on the environmental conditions within the host.

The ability to switch from a yeast to a filamentous, hyphal form, together with the expression of hypha-associated genes, is important for virulence of the opportunistic fungal pathogen Candida albicans (4). Morphogenesis and expression of hypha-associated genes are regulated in response to environmental cues by several signal transduction pathways, including a mitogen-activated protein kinase pathway that ends in the transcription factor Cph1p and a cyclic AMP-dependent pathway involving the Efg1p transcription factor (3, 5, 12, 15, 24). Depending on the growth conditions used to induce hyphal formation, mutants that are defective in these signaling pathways are unable to form filaments and do not express hypha-associated genes (2, 10, 13, 14, 23).

The TEA/ATTS transcription factor Tec1p also has been shown to be necessary for proper hyphal formation in C. albicans in vitro (19). C. albicans mutants with a deletion of the TEC1 gene displayed attenuated virulence in a mouse model of systemic candidiasis, although they formed hyphae in the infected tissue (19). It has been proposed that Tec1p regulates hypha-associated genes in C. albicans by binding to the sequence 5′-CATTCY-3′, which is the consensus binding sequence for fungal TEA/ATTS transcription factors (1, 6, 19). This sequence is found in the promoter region of hypha-associated genes, and recent evidence suggests that different signal transduction pathways involved in hyphal growth of C. albicans converge on the Tec1p transcription factor, which then mediates the expression of hypha-associated genes (10).

The SAP5 gene encodes one of the secreted aspartic proteinases of C. albicans which have been shown to contribute to virulence of the fungus (9, 17). Its expression is induced upon hyphal formation in vitro and during experimental infection in vivo, and full SAP5 induction by host signals depends on Cph1p and Efg1p (7, 20, 21). The SAP5 promoter contains three direct repeats of a 25-bp sequence in which a Tec1p consensus binding sequence is present, suggesting that Tec1p might induce SAP5 expression by binding to this sequence. Therefore, using a recently developed in vivo expression technology that is based on FLP-mediated deletion of a mycophenolic acid resistance marker (MPA^R) from the genome as a reporter of gene activation (20, 22), we investigated whether SAP5 induction within the host is dependent on Tec1p and whether the putative Tec1p binding sites mediate this activation.

To delete the three 25-mer repeats containing the putative Tec1p binding sites from P_SAP5, a 0.8-kb fragment containing the SAP5 promoter was PCR amplified with the primer pair SAP5P3 (5′-GTATAAATGCTCtaGAATTCTGTTTGGCG-3′) and SAP5P6 (5′-ACATTgtcgacTTGAGCTTAACTTTGGATTAGTTATAAAGGAGTGAAT*GACTATTTGAATG-3′). The lowercase letters represent nucleotide exchanges introduced to create an upstream XbaI site and a SalI site in front of the start codon (restriction sites are underlined, and the reverse sequence of the start codon is shown in bold); the position of the 75-bp deletion is indicated by an asterisk. The PCR product was digested with XbaI and SalI, and the mutated SAP5 promoter (P_SAP5Δ3) was substituted for the wild-type SAP5 promoter of plasmid pSFL53 (20) to generate pSFL56. Strain CFI1, which contains the FLP-deletable MPA^R marker, was transformed by electroporation (8) with the insert from this plasmid. The two SAP5 alleles of the parental strain can be distinguished by a BglII restriction site polymorphism (20), and two independent transformants which contained the P_SAP5Δ3-ecaFLP fusion integrated into the SAP5-1 (strain S5FI5A) or the SAP5-2 allele (strain S5FI5B) were selected. Replacement of the original SAP5 promoter by the mutated version was confirmed by reamplification and sequencing.

To detect activation of the wild-type SAP5 promoter in the absence of the Tec1p transcription factor, the FRT-MPA^R-FRT cassette from plasmid pAFI3 containing the FLP-deletable MPA^R marker between ACT1 flanking sequences (22) was first integrated into one of the ACT1 alleles of the Δtec1 mutant strain CaAS15 (19). The resulting strain, CFI6, was then transformed with the P_SAP5-ecaFLP reporter gene fusion from pSFL53, and two independent transformants in which integration had occurred into the SAP5-1 (strain C6S5F1D) or the SAP5-2 allele (strain C6S5F1E) were selected. All integration events were confirmed by Southern hybridization (data not shown).

We first assessed activation of the SAP5 promoter during serum-induced hyphal growth in vitro. Five microliters of an overnight culture of each reporter strain was inoculated in 250 μl of fresh serum from a healthy human donor and incubated for 4 h at 37°C. Appropriate dilutions were then spread on MPA indicator plates to determine the percentage of MPA^S cells in which induction of the SAP5 promoter had resulted in FLP-mediated deletion of the MPA^R marker. As shown in Fig. 1A, activation of the SAP5 promoter in a wild-type background (strains S5FI2A and S5FI2B [20]) during serum-induced hyphal growth was readily detected with this reporter system, whereas virtually no SAP5 induction was seen in the Δtec1 mutant, confirming previous results from Northern hybridization experiments (19). Surprisingly, the deletion of the putative Tec1p binding sites did not affect the inducibility of the SAP5 promoter by serum. Since there is no other Tec1p binding consensus sequence within 2 kb upstream of the SAP5 coding region, this result suggests that Tec1p binding sites in C. albicans may differ from those in other organisms, as has recently been reported for the Rim101p transcription factor (16), and Tec1p might bind to sites other than the consensus sequence in the SAP5 promoter to mediate activation of this gene. Alternatively, the involvement of Tec1p in the transcriptional induction of SAP5 may be indirect.

FIG. 1.

Activation of the wild-type (SAP5) and mutated (SAP5Δ3) SAP5 promoters in a wild-type background and of the wild-type SAP5 promoter in a Δtec1 mutant (SAP5Δtec1) by serum in vitro (A) and after intraperitoneal (B) or intravenous (C) infection of mice. Fungal cells were recovered from the liver at 24 h after intraperitoneal infection and from the kidneys at 4 days after intravenous infection. Each bar shows the percentage of MPA^S cells in one infected animal. The light grey bars show the results obtained with strains S5FI2A, S5FI5A, and C6S5F1D, and the dark grey bars show the results obtained with strains S5FI2B, S5FI5B, and C6S5F1E.

To analyze the dependence of SAP5 induction on the Tec1p transcription factor during infection of an animal host, we used a mouse model of Candida peritonitis as described previously (20-22). We observed that in this model hyphal formation within the host was impaired in the Δtec1 mutant strains. In contrast to the wild-type strains, which formed hyphae and invaded into the liver, the Δtec1 mutants attached to the liver surface but did not invade and were surrounded by inflammatory cells (Fig. 2, left panels). Despite the failure of the Δtec1 mutants to invade the liver, activation of the SAP5 promoter was clearly detected in the cells remaining at the organ surface, although the average percentage of MPA^S cells was reduced (20% compared with 43% in the wild-type controls) (Fig. 1B). These results demonstrate that SAP5 activation also occurred in the absence of the Tec1p transcription factor within an infected host. No significant differences were observed when the activations of the wild-type SAP5 promoter and the mutated version with a deletion of the putative Tec1p binding sites were compared in a wild-type background (43 and 36% MPA^S cells, respectively) (Fig. 1B), demonstrating that the Tec1p binding consensus sequences are also not essential for activation of the SAP5 promoter within an infected host. The lower level of detectable activation of the SAP5 promoter in the Δtec1 mutants may be due to Tec1p binding to sites other than the consensus sequence in C. albicans or to indirect activation, as discussed above. However, since the Δtec1 mutants resided in host niches other than the wild-type strains, it seems possible that stronger host signals, and not Tec1p-mediated transcriptional activation, account for the increased SAP5 induction seen in the wild-type strains compared with that in the Δtec1 mutants.

FIG. 2.

Hyphal formation and tissue invasion of wild-type strain SC5314 (TEC1) and reporter strain C6S5F1E (Δtec1). Both Δtec1 strains behaved identically, and all reporter strains carrying the P_SAP5-ecaFLP fusion in a wild-type background were indistinguishable from SC5314. (Left panels) Microscopic appearance of the strains at 24 h after intraperitoneal infection. (Right panels) Microscopic appearance of the strains in invaded kidneys at 4 days after intravenous infection.

To investigate whether the Tec1p-independent SAP5 activation within the host was a peculiarity of the intraperitoneal infection model, we also analyzed the behavior of our reporter strains in mice after intravenous infection. Histological examination showed that, as previously observed (19), the Δtec1 mutants formed hyphae in infected kidneys which were indistinguishable from those formed by wild-type strains, demonstrating that the defect in hyphal formation of a Δtec1 mutant depends on the host niche (Fig. 2, right panels). Importantly, deletion of the putative Tec1p binding sites had no effect on the inducibility of the SAP5 promoter and SAP5 activation was also observed in the absence of TEC1 in this infection model (Fig. 1C). These results demonstrate that the roles of the Tec1p transcription factor in the induction of hypha-associated genes can differ in vitro and in vivo and that SAP5 can be activated within the host by Tec1p-independent signaling mechanisms. Staib et al. have shown previously that SAP5 activation during infection depends on the transcriptional regulators Cph1p and Efg1p (21). In addition, no SAP5 mRNA was detected in vitro in Δefg1 or Δtec1 mutants (18, 19). Therefore, the in vitro expression of the SAP5 gene might require Efg1p-dependent activation by the Tec1p transcription factor, whereas SAP5 induction by host signals involves Tec1p-independent mechanisms, possibly including direct activation of the SAP5 promoter by Cph1p and Efg1p (10, 11). The results of our study clearly demonstrate that the importance of specific signaling pathways for the induction of C. albicans virulence-associated genes depends on the environmental conditions encountered within the host during an infection.