|Home | About | Journals | Submit | Contact Us | Français|
The success of Candida albicans as a major human fungal pathogen is dependent on its ability to colonize and survive as a commensal on diverse mucosal surfaces. One trait required for survival and virulence in the host is the morphogenetic yeast-to-hypha transition. Mds3 was identified as a regulator of pH-dependent morphogenesis that functions in parallel with the classic Rim101 pH-sensing pathway. Microarray analyses revealed that mds3Δ/Δ cells had an expression profile indicative of a hyperactive TOR pathway, including the preferential expression of genes encoding ribosomal proteins and a decreased expression of genes involved in nitrogen source utilization. The transcriptional and morphological defects of the mds3Δ/Δ mutant were rescued by rapamycin, an inhibitor of TOR, and this rescue was lost in strains carrying the rapamycin-resistant TOR1-1 allele or an rbp1Δ/Δ deletion. Rapamycin also rescued the transcriptional and morphological defects associated with the loss of Sit4, a TOR pathway effector, but not the loss of Rim101 or Ras1. The sit4Δ/Δ and mds3Δ/Δ mutants had additional phenotypic similarities, suggesting that Sit4 and Mds3 function similarly in the TOR pathway. Finally, we found that Mds3 and Sit4 coimmunoprecipitate. Thus, Mds3 is a new member of the TOR pathway that contributes to morphogenesis in C. albicans as a regulator of this key morphogenetic pathway.
Eukaryotic cell growth and morphogenesis is affected by numerous environmental signals. These environmental signals are integrated by highly conserved regulators, including Ras, protein kinase C (PKC), and the target of rapamycin (TOR). These regulators govern the activity of signal transduction pathways, which generally promote changes in gene expression affecting an appropriate cellular response to the environmental signal. TOR is an essential kinase conserved throughout eukaryotic evolution. In mammalian systems, mTOR is required for embryogenesis (38) and for cellular morphogenesis in neurons (33), vacuolar smooth-muscle cells (53), and T cells (19). In the model yeast Saccharomyces cerevisiae, TOR plays a fundamental role in morphogenesis, being required for pseudohyphal growth and sporulation (14, 80). TOR governs morphogenesis by regulating numerous biological processes, including autophagy, translation, and ribosome biogenesis (79). Thus, in eukaryotic cells, TOR responds to environmental signals to promote growth and morphogenetic changes.
In the fungal opportunistic pathogen Candida albicans, the most well-studied morphogenetic transition is the yeast-hypha switch. The switch between the yeast and hyphal growth forms plays a critical role in the ability of C. albicans to colonize as a commensal and cause disease as a pathogen. Yeast cells are small, are nonadherent, and are less immunogenic and divide more rapidly than hyphal cells, which may allow C. albicans to successfully outcompete the faster-growing bacterial flora and the immune responses as well as pass through capillary beds during disseminated disease. Hyphal cells are long and extremely adherent, promote profound immune responses, and divide more slowly. Further, hyphal cells secrete numerous degradative enzymes and are invasive (32, 56). Hyphal cells may allow C. albicans to maintain itself on mucosal surfaces and, in immunocompromised hosts, enter the bloodstream. Both yeast and hyphal cells are observed in commensally colonized and diseased sites, demonstrating the relevance of this morphogenetic switch for C. albicans survival in the host. This idea is further supported by genetic analyses demonstrating that C. albicans mutants locked in a yeast or filamentous form are unable to cause disease (48, 58). Thus, the yeast-hypha morphogenetic switch is critical for pathogenesis.
Because the yeast-hypha transition is critical for pathogenicity, much work has been devoted to identifying the environmental sensors and signal transduction pathways that respond to these signals. The yeast-hypha morphogenetic switch is governed by a plethora of environmental cues, including pH, nutrient availability, temperature, and host factors (3, 10, 17, 56). For example, neutral-alkaline environmental pH is sensed by the Rim101 signal transduction pathway, resulting in the proteolytic activation of the Rim101 transcription factor, which promotes hyphal formation (16, 62, 63). Loss-of-function mutants in the Rim101 pathway do not form hyphae in response to neutral-alkaline environmental pH and show reduced virulence in animal models of infection (15, 61). Similarly, nutrient levels are sensed by the TOR pathway, which also promotes hyphal formation and is required for pathogenesis (4, 14, 46). These studies highlight the profound link between morphogenesis and virulence in the host in response to environmental signals, such as pH and starvation.
To gain further insights into C. albicans adaptation to environmental pH, we identified additional signal transduction pathways that contribute to adaptation to environmental pH. For example, calcineurin and its associated transcription factor, Crz1, were found to act in parallel with Rim101 for adaptation to alkaline pH in C. albicans (43). Further, using a forward genetics approach, MDS3 was identified as a positive regulator of alkaline pH responses in C. albicans (18). Like calcineurin, Mds3 was shown to act in parallel to Rim101 to promote neutral-alkaline pH responses. Mds3 also governs adaptation to neutral-alkaline pH responses in S. cerevisiae; however, this function originally was masked by the presence of the redundant Mds3 paralog, Pmd1. While the well-studied Rim101 pathway is required for adaptation to environmental pH, Mds3 plays a key role in adaptation to environmental pH.
MDS3 originally was identified in S. cerevisiae as a positive regulator of meiosis, a morphogenetic process that requires an alkaline environment (5, 24). In C. albicans, Mds3 is involved in a variety of morphogenetic processes, including the yeast-hyphal transition, chlamydospore formation, and biofilm formation (18, 60, 65). Despite a clear role for Mds3 in diverse morphogenetic processes in fungi, how Mds3 promotes morphogenesis is unknown. Work done in S. cerevisiae suggested that Mds3 transmits starvation signals to Ras (5, 54). However, we demonstrate that Mds3 contributes to morphogenesis in C. albicans through the negative regulation of the TOR pathway. We find that an mds3Δ/Δ mutant has transcriptional defects indicative of a hyperactive TOR pathway and that these transcriptional defects are rescued by rapamycin, a TOR pathway inhibitor. Further, the filamentation defects associated with the mds3Δ/Δ mutant are rescued by rapamycin. We also demonstrate that the loss of Sit4, a downstream effector of TOR, mimics the loss of Mds3. In C. albicans, the loss of Sit4 or Mds3 results in resistance to rapamycin, whereas in S. cerevisiae this loss results in sensitivity to rapamycin, suggesting that TOR functions are distinct in these two species. Finally, we demonstrate that Mds3 and Sit4 interact, suggesting a mechanism of TOR pathway regulation. In total, our results establish that Mds3 is a member of the TOR pathway and suggest that Mds3 regulates C. albicans development and morphogenesis through the TOR pathway.
All yeast strains used in this study are listed in Table Table1.1. DAY938 was constructed by deleting the second copy of MDS3 from VIC1, the MDS3/mds3::ARG4 parent of strain VIC3 (18), using an mds3::URA3-dpl200 disruption cassette amplified with primers MDS3 5DR and MDS3 3DR (Table (Table2).2). Strains DAY1122 and DAY1123 were constructed by transforming strains DAY938 and DAY286, respectively, with the tor1::HIS1 disruption cassette, which was amplified using primers TOR1 5DR and TOR1 3DR (Table (Table2).2). All deletions are from the start to the stop codon and were generated by chemical transformation (78). The correct integration of the disruption cassettes was verified by PCR using the Mds3null 5-detect and Mds3null 3-detect primers or TOR1 5′detect-2 and TOR1 3′detect primers (Table (Table2).2). DAY1118 and DAY1119 were generated by transforming DAY938 with the PmeI-digested plasmids pDDB343 and pDDB353, respectively.
Strains DAY1120, DAY1121, DAY1124, DAY1125, and DAY1321, which contain the TOR1-1 allele, were generated as previously described, with minor modifications (13). Briefly, primers TOR1 A-C and TOR1 A-C Rev (Table (Table2)2) were annealed and used for transformation. Transformants were selected on yeast extract-peptone-dextrose (YPD) supplemented with 100 nM rapamycin. Rapamycin-resistant clones were screened by the NheI digestion of a PCR product amplified with primers JOHE6247 and JOHE6248 and were verified by sequencing (13). DAY1255 then was generated by transforming DAY1120 with PmeI-digested pDDB343. The correct genotype of all of the mutants carrying TOR1-1 and/or tor1::HIS1 alleles was verified by the Southern blotting of the NheI/NcoI or NheI/PvuII digestion of genomic DNA (data not shown) and using PCR products amplified with primers TOR1 probe 5′ and TOR1 probe 3′ or TOR1 probe 5′-2 and TOR1 probe 3′ as probes (Table (Table2).2). The probes were radiolabeled with [α-32P]dCTP using the Prime-a-Gene labeling system (Promega).
The mds3Δ/Δ sit4Δ/Δ mutant DAY1233 was constructed by sequentially deleting both MDS3 alleles from the sit4Δ/Δ strain DAY972, using the mds3::URA3-dpl200 disruption cassette amplified with the primers MDS3 5DR and MDS3 3DR described above. The URA3-dpl200 marker was recycled by growing the cells in synthetic complete (SC) medium supplemented with 5-fluoroorotic acid (5-FOA).
The mds3Δ/Δ rbp1Δ/Δ mutant DAY1239 was constructed by sequentially deleting both MDS3 alleles from the rbp1Δ/Δ strain DAY1230. Strain DAY1230 was generated by recycling the MX3::URA3R::MX3 marker in strain DAY1223 (13) in 5-FOA. The mds3::URA3-dpl200 disruption cassette was amplified with primers MDS3 5DR and MDS3 3DR described above, and the URA3-dpl200 marker was recycled by growing the cells in 5-FOA.
DAY1234, which contains a functional SIT4-c-Myc allele, was constructed by introducing a C-terminal MYC-TADH1-URA3 cassette. The MYC-TADH1-URA3 cassette was PCR amplified with primers Sit4 MYC Nter 5′ and Sit4 MYC URA3 Nter 3′ from plasmid pDDB372 (pMG1095 ) (Table (Table2).2). The correct integration of the cassette was verified by PCR using primers 5′ Myc detect and Sit4 3′ detect (Table (Table2)2) and by Western blotting using anti-Myc antibody (R950-25; Invitrogen).
To construct the strains for the coimmunoprecipitation studies, DAY1234 was transformed with NheI-digested pDDB499 (MDS3-HA) to generate strain DAY1236. DAY1234 also was transformed with PmeI-digested pDDB353 (untagged MDS3 vector) to generate the control strain DAY1235.
The MDS3 complementation vector pDDB353 was constructed as follows. Plasmid pDDB343 was generated by replacing the NruI restriction site in pDDB78 with a PmeI site. Primers DDB78 PmeIxNruI 5′ and DDB78 PmeIxNruI 3′ were annealed and used for the gap repair of the NruI-digested pDDB78 through in vivo recombination in trp mutant S. cerevisiae strain L40 to generate plasmid pDDB343. Wild-type MDS3 sequence with flanking promoter and terminator sequences was amplified in two high-fidelity PCRs (Pfu turbo DNA polymerase; Stratagene) using primer pairs MDS3 5comp new and MDS3 5-2comp new as well as MDS3 3comp new-2 and MDS3 3-2comp new-2 from strain BWP17 genomic DNA (Table (Table2).2). Both PCR products were transformed with EcoRI/NotI-double digested pDDB343 into strain L40 to produce plasmid pDDB353 through in vivo recombination. The complete open reading frame (ORF) of MDS3 was verified by sequencing.
The MDS3-HA-tagged vector DDB499 was constructed as follows. To find a location in MDS3's ORF that could be tagged without affecting Mds3 function, we performed the random mutagenesis of MDS3 using the GPS-LS linker scanning system (New England BioLabs). An NheI/AhdI fragment from plasmid DDB353 containing MDS3 and its flanking regions was cloned into an NheI/AhdI-digested pGEM-T Easy vector (Promega), pDDB407, to generate plasmid pDDB408. pDDB408 was used as the substrate for random Tn7 insertion mutagenesis. Plasmids carrying Tn7 integrations within MDS3's ORF were screened by PCR using primer S and primer N (Promega), MDS3null 5-detect, and MDS3null 3-detect (Table (Table2).2). The Tn7 transposon was excised, and plasmids were religated, leaving a 15-bp insertion that contains a PmeI site. Plasmids containing mutated versions of MDS3's ORF were AhdI/NdeI digested, subcloned into NotI/EcoRI-digested pDDB409 by in vivo recombination, and tested for the complementation of an mds3Δ/Δ mutant. Plasmid pDDB500, containing a 5-amino-acid insertion at amino acid 1051, was selected for hemagglutinin (HA) tagging. The HA tag was PCR amplified from plasmid pDDB369 (pMG1921 ) using primers MDS3-HA 56 5′ and MDS3-HA 56 3′ (Table (Table2)2) and in vivo recombined into the PmeI-digested plasmid pDDB500 using S. cerevisiae L40 strain to generate pDDB499, which inserts HA after residue 1056. pDDB499 rescued the mds3Δ/Δ mutant, indicating that Mds3 tolerated the addition of the 29 amino acids of the 3× HA tag.
Plasmid pDDB409 was constructed by replacing the unique NruI site in pDDB78 with a 26-bp linker containing NgoMIV-KpnI-NheI restriction sites. The 26-bp primers NgoMIV KpnI NheI × NruI 5′ and NgoMIV KpnI NheI × NruI 3′ were annealed and ligated in NruI-digested pDDB78 to give pDDB409.
C. albicans was grown routinely at 30°C in YPD (2% Bacto peptone, 2% dextrose, 1% yeast extract). For the selection of Ura+, His+, or Trp+ transformants, synthetic medium without uridine, histidine, or tryptophan was used (0.17% yeast nitrogen base without ammonium sulfate [Q-BioGene], 0.5% ammonium sulfate, 2% dextrose, and supplemented with a dropout mix containing amino and nucleic acids except those necessary for the selection ). Media were buffered at the indicated pH using 150 mM HEPES. The assays for filamentation in the presence of rapamycin were performed in M199 medium (Gibco BRL) buffered at pH 8 and SLAD (0.17% yeast nitrogen base without ammonium sulfate [Q-BioGene], 50 μM ammonium sulfate, 2% dextrose). Filamentation assays were conducted at 37°C except as indicated. Rapamycin (LC Laboratories) was added to the media at the indicated concentrations from a stock solution in 90% ethanol-10% Tween 20. For liquid assays of filamentation in the presence of rapamycin, strains were pregrown in liquid YPD at 30°C, pelleted, resuspended in an equal volume of phosphate-buffered saline (PBS), and diluted 1:100 in M199, pH 8, supplemented with rapamycin or solvent alone. Samples were incubated for 6 h. To determine the percentage of hypha-producing cells, samples were gently sonicated and quantified under the microscope. The values reported represent the averages from at least two independent experiments, in which every strain was tested at least in duplicate. For filamentation in serum, strains were pregrown in liquid YPD at 30°C, diluted 1:100 in liquid YPD supplemented with 10% fetal bovine serum (Gibco), and incubated at 37°C for 5.5 h. For growth assays in the presence of rapamycin, strains were pregrown in liquid YPD at 30°C, diluted in PBS to an optical density at 600 nm (OD600) of 1.6, and then serially diluted 5-fold in PBS, spotted on YPD or Spider medium (1% mannitol, 1% nutrient broth, 0.2% K2HPO4, pH 7.2, before being autoclaved ), supplemented with rapamycin or solvent alone, and incubated at 30°C for 2 or 3 days. All media except that for the selection of Ura+ transformants were supplemented with 80 μg/ml uridine. For solid media, 2% Bacto agar was added, except for Spider medium, which used 1.35% Bacto agar.
Strains were pregrown in YPD at 30°C, pelleted, washed with M199 (pH 4 or pH 8) medium, diluted 40-fold in prewarmed M199 (pH 4 or pH 8), and incubated for 4 h at 37°C with agitation. Procedures for RNA extraction, microarray construction, and analysis have been described previously (6). Corrected P values for the gene ontology (GO) categories were obtained using the GO term-finder algorithms available through the Candida Genome Database website (www.candidagenome.org) in February 2009.
Cells were grown overnight in YPD at 30°C. The following day, cells were washed with M199 medium at either pH 4 or 8 and diluted 40-fold into fresh M199 (pH 4 or 8) medium containing solvent or 5 nM rapamycin and incubated for 4 h at 37°C with agitation. For Spider medium experiments, cells were diluted 1:40 in Spider medium containing solvent or 5 nM rapamycin and incubated for 4 h at 37°C with agitation. For YPD medium experiments, cells were grown as described previously (4). Briefly, YPD overnight cultures were inoculated into YPD at an OD of ~0.1. Cells were grown 5 h at 30°C, followed by 1 h at 30°C with solvent or 20 nM rapamycin. Cells then were harvested and frozen in a dry ice-ethanol bath. RNA extraction and Northern blot procedures were previously described (6), except that the PCR products for the probes were purified with a PCR purification kit (Qiagen). RNA concentration was measured using a NanoDrop spectrophotometer ND-1000. Probes for ECE1 and HWP1 were described previously (18).
Statistical analysis for data shown in Table Table55 was performed using the software SAS 9.13 (SAS Institute Inc.). We used repeated analyses of covariance (ANCOVA) (assuming different variances for each strain) and included the day of the experiment as the covariate. We tested for strain-treatment interaction and compared specific strain treatment effects after adjusting for multiple comparisons using Bonferroni's correction (14 comparisons were made; significant P values were <0.05/14, or 0.0036).
Pictures of colonies were taken using a Canon Powershot A560 digital camera on a Zeiss Opton microscope. Images of liquid cultures were captured using a Zeiss Axio camera, Axiovision 4.6.3 software (Zeiss), and a Zeiss AxioImager fluorescence microscope. All images were processed with Adobe Photoshop 7.0 software.
Overnight cultures of C. albicans were diluted 200-fold into M199 (pH 8) or YPD medium and grown for 4 to 5 h at 30°C. Cells were pelleted and resuspended in ice-cold radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris, pH 8, 150 mM NaCl, 1% NP-40, 3 mM EDTA, 0.5% deoxycholate, 0.1% sodium dodecyl sulfate [SDS]) containing 1 μg/ml leupeptin, 2 μg/ml aprotinin, 1 μg/ml pepstatin, 0.1 mM phenylmethylsulfonyl fluoride, and 10 mM dithiothreitol and lysed by vortexing with acid-washed glass beads for 1 h at 4°C. Cell lysates were pelleted and supernatants stored at −80°C. For Western blot assays of the input, ~1.25 mg of total protein was resuspended in 2× SDS gel-loading buffer (100 mM Tris-Cl, pH 6.8; 200 mM dithiothreitol; 4% SDS; 0.1% bromophenol blue; 20% glycerol), boiled at 95 to 100°C for 3 min, and run in an 8% SDS-PAGE gel. Proteins were transferred to nitrocellulose and blocked in 6% nonfat milk in TBS-T (50 mM Tris, pH 7.6, 150 mM NaCl, 0.1% Tween 20). Blots were incubated with anti-HA (F-7 probe; Santa Cruz) or anti-Myc (Invitrogen) at 1:1,000 or 1:5,000 dilution, respectively, in 6% nonfat milk TBS-T, washed in TBS-T, and incubated with anti-mouse antibody-horseradish peroxidase (GE Healthcare) at 1:5,000 in 6% nonfat milk in TBS-T. Blots were washed in TBS-T, incubated with ECL reagent (GE Healthcare), and exposed to film.
Mds3-HA was immunoprecipitated as follows. Total protein (~18 mg) was incubated overnight at 4°C with 20 μl of slurry anti-HA agarose beads (Sigma-Aldrich). Beads were washed four times with Tris-buffered saline (25 mM Tris, 0.15 M NaCl, pH 7.2, 0.05% Tween). Protein was eluted from the beads by resuspension in 50 μl 2× SDS gel loading buffer and boiling for 5 min. The products of two immunoprecipitations were combined, separated by SDS-PAGE, and analyzed by Western blotting.
Mds3 acts in parallel to the Rim101 pathway to promote the yeast-to-hypha morphogenetic switch in response to neutral-alkaline pH (18). To gain insights into how Mds3 contributes to pH responses, we determined the transcriptional profiles of wild-type, rim101Δ/Δ, and mds3Δ/Δ strains grown at pH 8 and 4 (see Table S1 in the supplemental material). The microarray experiments for all three strains were performed concomitantly, but the results for the wild-type and rim101Δ/Δ strains, including the microarray methodology and validation, have been reported previously (6).
Since Mds3 and Rim101 are required for growth and morphogenesis at alkaline pH, we expected that most transcriptional differences observed between the mutant strains and the wild type would occur at pH 8. Indeed, ~80% (287) and ~95% (180) of all differentially expressed ORFs in the mds3Δ/Δ and rim101Δ/Δ mutants, respectively, occurred at pH 8 (Table (Table3)3) (6). Further, because the mds3Δ/Δ and rim101Δ/Δ mutants are defective for filamentation at alkaline pH, we expected to find common changes in the expression of hypha-associated genes. In fact, the expression of genes associated with cell wall/filamentation were decreased in the mds3Δ/Δ and rim101Δ/Δ mutants relative to that of the wild-type strain (Table (Table4)4) (6). Thus, mds3Δ/Δ and rim101Δ/Δ mutants share similar transcriptional defects, as predicted based on their similar phenotypes at alkaline pH.
However, as Mds3 and Rim101 act in parallel, we also expected to find differences between the transcriptional profiles of these mutants relative to that of the wild-type strain. Indeed, several fundamental differences were observed. First, the mds3Δ/Δ mutant affected the expression of twice as many genes as did the rim101Δ/Δ mutant (363 versus 187 total ORFs, respectively) (Table (Table3).3). Approximately 60% (229) of the genes affected by the mds3Δ/Δ mutant were not pH regulated in wild-type cells, compared to ~38% (72) of the genes in the rim101Δ/Δ mutant (Fig. (Fig.1).1). Second, the mds3Δ/Δ mutant affected the expression of more genes at pH 4 than the rim101Δ/Δ mutant (76 versus 7 genes, respectively) (Table (Table3).3). Third, in the rim101Δ/Δ mutant, ~70% of the differentially expressed genes showed reduced expression, suggesting that Rim101 primarily functions as a positive regulator (Table (Table3).3). However, in the mds3Δ/Δ mutant, the differentially expressed genes showed increased and reduced expression at an ~50:50 ratio, demonstrating that Mds3 functions as both a positive and negative regulator (Table (Table3).3). Overall, these results suggest that Mds3 has broader effects on gene expression than Rim101 and that Mds3 has roles in addition to adaptation to environmental pH.
To identify potential biological functions for Mds3, we analyzed the Mds3-dependent genes by gene ontology (GO). Of the genes upregulated in the mds3Δ/Δ mutant compared to the wild-type strain, the GO categories of glycolysis and translation were significantly overrepresented (P < 1.1E−04 and P < 5.3E−15, respectively) (Table (Table4;4; also see Table S1 in the supplemental material). In wild-type cells, most of the ORFs in translation were expressed preferentially at pH 8 compared to pH 4 (6). This suggests that the loss of Mds3 enhances these gene expression changes observed in wild-type cells at alkaline pH. Of the genes downregulated in the mds3Δ/Δ mutant compared to the wild-type strain, the GO categories of amino acid transport and vacuolar protein catabolic activity (P < 4.7E−04 and P < 1.6E−03, respectively) were significantly overrepresented (Table (Table4).4). Further, the categories of amine transport, lytic vacuole, glycolysis, and translation also were overrepresented when this analysis was done in the mds3Δ/Δ mutant exclusively for genes that are pH independent (data not shown). These results show that Mds3 affects a variety of distinct processes, including growth and starvation responses, and supports the idea that Mds3 has functions in pH-independent processes.
To corroborate the microarray results, we analyzed the expression of several MDS3-dependent ORFs by Northern blotting (Fig. (Fig.22 A). In M199 pH 8 medium, the glycolysis gene CDC19 and the translation genes RPS26A and TEF1 were expressed ~2- to 4-fold more in the mds3Δ/Δ mutant than the wild-type or the complemented mds3Δ/Δ +MDS3 strains (Fig. (Fig.2A,2A, lanes 6 to 8). The increased expression of CDC19, RPS26A, and TEF1 in the mds3Δ/Δ mutant also was observed at pH 4 (Fig. (Fig.2A,2A, lanes 1 to 3), suggesting that Mds3 governs the expression of these genes in a pH-independent manner. GAP2, an amino acid transport gene, was reduced >2-fold in the mds3Δ/Δ mutant compared to the wild-type or the mds3Δ/Δ +MDS3-complemented strain at both pH 4 and 8 (Fig. (Fig.2A,2A, compare lanes 2 and 7 to lanes 1, 3, 6, and 8). These results establish the veracity of the microarray data and support the idea that Mds3 has pH-independent functions.
The GO categories affected by the mds3Δ/Δ mutant reflect processes that are transcriptionally regulated by the TOR pathway. During conditions of active growth, TOR positively regulates the expression of genes involved in ribosomal biogenesis, translation initiation and elongation, rRNA and tRNA synthesis, and glycolysis (79), and it negatively regulates the expression of genes involved in protein degradation and nitrogen catabolite repression (NCR) (4, 8, 35, 68). During starvation conditions, TOR is inactive and genes involved in translation and glycolysis are repressed, while genes involved in amino acid transport and vacuolar degradation are induced (4, 8, 35, 68). Since the mds3Δ/Δ mutant shows gene expression differences that reflect TOR activation, we hypothesized that Mds3 functions as a negative regulator of TOR.
To address the idea that Mds3 is a negative regulator of TOR, we predicted that rapamycin, a TOR inhibitor (36), would rescue the transcriptional defects associated with the mds3Δ/Δ mutant. Indeed, 5 nM rapamycin reduced the expression of CDC19, RPS26A, and TEF1 and restored the expression of GAP2 in mds3Δ/Δ mutant cells compared to that with solvent alone (Fig. (Fig.2A,2A, compare lanes 6 to 8 and 11 to 13). Rapamycin did not appear to affect the expression of RPS26A, TEF1, and GAP2 in the wild-type or mds3Δ/Δ +MDS3 complemented strains; however, the expression of CDC19 was increased ~2-fold compared to that in solvent alone. Regardless, in the presence of rapamycin the expression of CDC19, RPS26A, TEF1, and GAP2 was quantitatively similar in wild-type, mds3Δ/Δ, and mds3Δ/Δ +MDS3 cells. These results demonstrate that rapamycin restores the transcriptional defects due to the loss of Mds3 and corroborates the idea that Mds3 acts as a negative regulator of TOR.
In nutrient-poor Spider medium, we observed similar transcriptional results (Fig. (Fig.2B).2B). The expression of CDC19, RPS26A, and TEF1 was increased in the mds3Δ/Δ mutant compared to that of the wild type (Fig. (Fig.2B,2B, compare lanes 1 and 2). Further, the addition of rapamycin reduced CDC19, RPS26A, and TEF1 expression in the mds3Δ/Δ mutant-, wild-type-, and mds3Δ/Δ +MDS3-complemented strains. We noted that in Spider medium this rapamycin-dependent restoration of gene expression in the mds3Δ/Δ mutant was less than that observed in M199 pH 8 medium. We also found that in Spider medium, GAP2 was expressed poorly in wild-type cells and expressed at greater levels in the absence of MDS3. The addition of rapamycin increased GAP2 expression in all strains tested (Fig. (Fig.2B,2B, compare lanes 1 to 5 with 6 to 10). These results support the idea that Mds3 acts in the TOR pathway and suggests that, not surprisingly, growth medium differences cause distinct effects.
The TOR pathway regulates fungal morphogenesis (14), and rapamycin triggers the expression of hypha-associated genes in C. albicans (4). Since Mds3 also regulates morphogenesis and is required for the expression of hypha-associated genes (18), we asked if rapamycin also rescues the expression of these genes in the mds3Δ/Δ mutant (Fig. (Fig.33 A). Wild-type, mds3Δ/Δ, and mds3Δ/Δ +MDS3 cells were incubated in M199 pH 8 medium at 37°C, which promotes hypha formation, with and without 5 nM rapamycin and analyzed the expression of the hypha-associated genes HWP1 and ECE1 (Fig. (Fig.3A).3A). While the mds3Δ/Δ mutant showed little HWP1 and ECE1 expression, the addition of rapamycin completely restored this expression (Fig. (Fig.3A,3A, compare lanes 2 and 5). We noted that rapamycin also led to the increased expression of HWP1 and ECE1 in both wild-type and mds3Δ/Δ +MDS3 strains compared to that in solvent alone. Thus, rapamycin promotes the expression of hypha-associated genes and rescues the mds3Δ/Δ mutant defects in HWP1 and ECE1 expression.
Since rapamycin rescued the transcriptional defects associated with the loss of Mds3, we asked if rapamycin also could rescue the hyphal formation defect of the mds3Δ/Δ mutant. In liquid M199 pH 8 medium at 37°C, >90% of wild-type cells germinated to form hyphae; <5% of mds3Δ/Δ mutant cells germinated to form hyphae (Table (Table55 and Fig. Fig.3B).3B). However, the addition of 5 nM rapamycin partially restored hyphal formation in the mds3Δ/Δ mutant to 45% (P < 0.0001 compared to mds3Δ/Δ in solvent) (Fig. (Fig.3B3B and Table Table5).5). Rapamycin also promoted hyphal formation in wild-type cells compared to formation in the solvent control (P < 0.001) (Table (Table5),5), and we noted that the hyphae formed in the wild-type, mds3Δ/Δ, and mds3Δ/Δ +MDS3 strains in the presence of rapamycin appeared less branched than in solvent alone. Similar results were observed in liquid SLAD medium, a nutrient-poor medium, indicating that the effect of rapamycin is not specific to M199 medium (Fig. (Fig.4A).4A). These results demonstrate that rapamycin rescues the morphogenetic defect of the mds3Δ/Δ mutant.
We found that the effect of rapamycin on hyphal formation was not due to translational inhibition, because cycloheximide did not rescue the filamentation defect of the mds3Δ/Δ mutant (Fig. (Fig.4B).4B). Further, rapamycin is not a constitutive inducer of hyphal formation, because rapamycin did not rescue the mds3Δ/Δ hyphal formation defect in M199 pH 8 medium at 30°C (Fig. (Fig.3B).3B). These results demonstrate that rapamycin is not a constitutive inducer of hyphal formation and further support a link between Mds3 and TOR.
To determine if the effect of rapamycin was actually dependent on TOR, we introduced the rapamycin-resistant TOR1-1 allele into the wild-type and mds3Δ/Δ strains and determined if rapamycin still promoted hyphal formation (7, 13). As expected, strains carrying a TOR1-1 allele were resistant to rapamycin (Fig. (Fig.5A).5A). In the absence of rapamycin, the TOR1-1 allele did not affect hyphal formation in either the MDS3/MDS3 or mds3Δ/Δ background (Fig. (Fig.5B5B and Table Table5).5). However, in the presence of rapamycin, the TOR1-1 allele prevented the rapamycin-induced hyphal formation observed in MDS3/MDS3 cells (Table (Table5),5), and the hypha formed showed a branching pattern similar to that of the wild-type strain without rapamycin (Fig. (Fig.5B).5B). Thus, in the MDS3/MDS3 background rapamycin-mediated phenotypes are TOR dependent. In the presence of rapamycin, the TOR1-1 allele reduced the rapamycin-induced hyphal formation observed in mds3Δ/Δ cells by ~50% (Table (Table5).5). We hypothesized that the semiresponsive effect of rapamycin in the mds3Δ/Δ TOR1-1/TOR1 mutant was an attribute of the remaining wild-type TOR1 allele. To address this possibility, we constructed TOR1-1/Δ strains in the MDS3/MDS3 and mds3Δ/Δ backgrounds. The deletion of one copy of TOR1 in either background had no significant effect on hyphal formation (Fig. (Fig.5B5B and Table Table5),5), demonstrating that TOR1 is not haploinsufficient. While the MDS3/MDS3 TOR1-1/Δ strain filamented similarly to the MDS3/MDS3 TOR1-1/TOR1 strain in the presence or absence of rapamycin, the mds3Δ/Δ TOR1-1/Δ strain failed to filament in the presence or absence of rapamycin (Fig. (Fig.5B5B and Table Table5).5). The mds3Δ/Δ TOR1-1/Δ mutant did form hyphae in serum, an MDS3-independent inducer of hyphal formation (data not shown and reference 18), demonstrating that the mds3Δ/Δ TOR1-1/Δ strain did not have an absolute block in hyphal formation. These results demonstrate that rapamycin rescues mds3Δ/Δ phenotypes through TOR.
We noted that in the absence of a wild-type copy of TOR1, the TOR1-1 allele conferred growth defects in both the wild-type and mds3Δ/Δ backgrounds (Fig. (Fig.5A5A and and6),6), demonstrating that the TOR1-1 allele is not completely functional. Thus, we considered that this growth defect could influence our hyphal formation data. To address this, we constructed an mds3Δ/Δ rbp1Δ/Δ double mutant that did not have a growth defect (Fig. (Fig.5A).5A). Rapamycin inhibits the TOR kinase as a complex with Rbp1 (FKBP12 in mammals), and rbp1Δ/Δ mutants are rapamycin resistant (Fig. (Fig.5A5A and reference 13). While rbp1Δ/Δ mutants did not have a defect in hyphal formation, the loss of Rbp1 in the mds3Δ/Δ background completely blocked the ability of rapamycin to rescue the mds3Δ/Δ hyphal formation defects (Fig. (Fig.5B5B and Table Table6).6). The loss of Rbp1 also prevented the transcriptional effects of rapamycin in the mds3Δ/Δ mutant (Fig. (Fig.2A,2A, lanes 16 to 18, and C, lanes 11 and 12). In total, these results clearly demonstrate that the transcriptional effects of the mds3Δ/Δ mutant have biological consequences that are associated with the inappropriate activation of the TOR pathway.
One downstream effector of the TOR pathway is the type 2A-like phosphatase Sit4 (22, 23, 66). Sit4 promotes starvation responses and is inhibited by TOR-dependent phosphorylation during periods of active growth (21, 40). Thus, in the absence of Sit4, TOR-dependent starvation responses are defective. Since the mds3Δ/Δ mutant also is defective for TOR-dependent starvation responses and promotes TOR activation (Table (Table44 and Fig. Fig.2),2), we predicted that a sit4Δ/Δ mutant has transcriptional defects similar to those of the mds3Δ/Δ mutant. Indeed, in the sit4Δ/Δ mutant, CDC19, RPS26A, and TEF1 were expressed at higher levels than those of the wild-type strain at both pH 4 and 8 (Fig. (Fig.2A,2A, compare lanes 4 and 9 to lanes 1 and 6) and in nutrient-poor Spider medium (Fig. (Fig.2B).2B). We noted that at pH 4, CDC19 expression in the sit4Δ/Δ mutant was ~4-fold higher than that in the mds3Δ/Δ mutant. Further, in the sit4Δ/Δ mutant, GAP2 was not expressed at either pH 4 or 8, although it was expressed more in Spider medium, similarly to the mds3Δ/Δ mutant. Thus, the loss of Sit4 has transcriptional effects similar to those of the loss of Mds3.
The addition of rapamycin to sit4Δ/Δ cells reduced the expression of CDC19, RPS26A, and TEF1 to approximately wild-type levels in M199 pH 8 medium (Fig. (Fig.2A,2A, compare lanes 9 and 14) but had a more modest effect on Spider medium (Fig. (Fig.2B).2B). Rapamycin also promoted the expression of GAP2 in sit4Δ/Δ cells. However, in M199 pH 8 medium, GAP2 expression still was ~2-fold lower than that of wild-type cells, indicating that Sit4 is required for the full induction of GAP2. These results demonstrate that the loss of Sit4 promotes the activation of the TOR pathway similarly to the loss of Mds3 and provides independent support for the idea that Mds3 acts in the TOR pathway.
Sit4 also is required for hyphal formation in C. albicans (46) (Fig. (Fig.3B).3B). We found that sit4Δ/Δ mutant cells expressed little HWP1 and ECE1 in M199 pH 8 medium and that this defect was partially rescued by the addition of rapamycin (Fig. (Fig.3A,3A, lanes 7 and 8). Similarly to the mds3Δ/Δ mutant, rapamycin rescued the sit4Δ/Δ hyphal formation defect in M199 pH 8 (12.9% ± 0.1% in solvent versus 60.5% ± 2.6% in rapamycin) and SLAD media (9.2% ± 1.0% in solvent versus 19.2% ± 0.5% in rapamycin) (Fig. (Fig.3B3B and and4A),4A), and this rescue was prevented by the TOR1-1 rapamycin-resistant allele (Fig. 5A and C). Importantly, rapamycin did not rescue the filamentation defects associated with the loss of Ras1 or Rim101, which are not associated with the TOR pathway (Fig. (Fig.3B).3B). In total, these results suggest a functional association between Mds3 and Sit4.
To determine if Mds3 and Sit4 act in the same pathway or parallel pathways, we constructed an mds3Δ/Δ sit4Δ/Δ double mutant. If Mds3 and Sit4 function in the same pathway, then the mds3Δ/Δ sit4Δ/Δ double mutant should have phenotypes similar to those of either single mutant; if Mds3 and Sit4 function in parallel pathways, then the mds3Δ/Δ sit4Δ/Δ double mutant should have more severe phenotypes than either single mutant. The mds3Δ/Δ sit4Δ/Δ double mutant expressed CDC19, RPS26A, TEF1, and GAP2 similarly to the mds3Δ/Δ and sit4Δ/Δ single mutants in M199 pH 8 and Spider media (Fig. (Fig.2A,2A, compare lanes 7, 9, and 10). In M199 pH 4 medium, the mds3Δ/Δ sit4Δ/Δ double mutant behaved similarly to the sit4Δ/Δ single mutant, which had a more severe phenotype than the mds3Δ/Δ mutant (Fig. (Fig.2A,2A, compare lanes 2, 4, and 5). In Spider medium, the mds3Δ/Δ sit4Δ/Δ double mutant also behaved similarly to the sit4Δ/Δ single mutant. However, the mds3Δ/Δ single mutant had a more severe phenotype, indicating a dominant effect of the sit4Δ/Δ mutation (Fig. (Fig.2B,2B, compare lanes 2, 4, and 5). These results demonstrate that the mds3Δ/Δ sit4Δ/Δ mutant does not have transcriptional defects beyond that of the single mutants, suggesting Mds3 and Sit4 act in the same pathway. However, the addition of rapamycin did not restore CDC19, RPS26A, and TEF1 expression in the mds3Δ/Δ sit4Δ/Δ double mutant (Fig. (Fig.2A,2A, lanes 12, 14, and 15, and B, lanes 7, 9, and 10), although it did restore GAP2 expression (Fig. (Fig.2A,2A, lanes 10 and 15). Rapamycin also did not rescue the hyphal formation defect associated with the mds3Δ/Δ sit4Δ/Δ double mutant (Fig. (Fig.33 and and4A),4A), which suggests that Mds3 and/or Sit4 has additional parallel functions.
We noted that regardless of pH, wild-type cells grown in M199 medium express relatively little RPS26A and robust levels of GAP2, suggesting that this medium is a nutrient-poor medium (Fig. (Fig.2A,2A, lane 1). Since TOR responds to nutrient availability, we asked if Mds3 and Sit4 affect gene expression in a nutrient-rich environment. Thus, we determined RPS26A and GAP2 expression in rich YPD medium with or without rapamycin. Unlike the results obtained for M199 pH 8 and Spider media (Fig. 2A and B), we observed the robust expression of RPS26A and no expression of GAP2 in wild-type cells, and cells lacking Mds3 and/or Sit4 behaved similarly (Fig. (Fig.2C,2C, compare lane 1 to lanes 2, 4, and 5). The addition of rapamycin reduced the expression of RPS26A and promoted the expression of GAP2 in wild-type cells, as expected if TOR is inhibited (Fig. (Fig.2C,2C, lane 6). The addition of rapamycin had a similar effect on mds3Δ/Δ, sit4Δ/Δ, and mds3Δ/Δ sit4Δ/Δ mutants (Fig. (Fig.2C,2C, lanes 7, 9, and 10). Further, this effect is clearly due to TOR inhibition, as RPS26A and GAP2 expression was similar in the mds3Δ/Δ rbp1Δ/Δ double mutant with and without rapamycin (Fig. (Fig.2C,2C, lanes 11 and 12). We noted that GAP2 expression in the presence of rapamycin was consistently higher in the mds3Δ/Δ mutant, and that this increase was SIT4 dependent. This is in contrast to the results observed in M199 pH 8 medium, where GAP2 expression in the mds3Δ/Δ mutant in the presence of rapamycin appears to be Sit4 independent. These results suggest that M199 medium is a more nutrient-poor medium than YPD and suggest that Mds3 and Sit4 contribute to the regulation of TOR-dependent targets in nutrient-poor conditions.
Since rapamycin rescued the filamentation defects of the mds3Δ/Δ and the sit4Δ/Δ mutants, we wanted to determine if these mutations affected rapamycin sensitivity in C. albicans. On YPD medium, the wild-type, mds3Δ/Δ, and mds3Δ/Δ +MDS3 strains showed similar rapamycin sensitivities, but the sit4Δ/Δ mutant was resistant to rapamycin (Fig. (Fig.77 A). The complementation of the sit4Δ/Δ mutant restored rapamycin sensitivity to wild-type levels. This result for the sit4Δ/Δ mutant is in sharp contrast to the situation in S. cerevisiae, where the sit4Δ mutant is sensitive to rapamycin and this sensitivity can be rescued by C. albicans SIT4, suggesting that Sit4 is functioning similarly in the two organisms (14, 46). These results indicate that MDS3 does not affect rapamycin sensitivity in rich medium, and that unlike the case for S. cerevisiae, Sit4 function promotes rapamycin sensitivity in C. albicans.
Since Mds3 appears to promote TOR-dependent responses in nutrient-poor conditions, we considered that Mds3 affects rapamycin sensitivity on nutrient-poor medium. Indeed, while wild-type cells were sensitive to 5 nM rapamycin on nutrient-poor Spider medium, the mds3Δ/Δ mutant was resistant (Fig. (Fig.7A).7A). Similar results were observed for the sit4Δ/Δ mutant. The mds3Δ/Δ sit4Δ/Δ double mutant behaved like the sit4Δ/Δ single mutant on both media, suggesting that Sit4 and Mds3 do not make independent contributions to rapamycin sensitivity. These results demonstrate that Mds3 and Sit4 promote rapamycin sensitivity in C. albicans and support the idea that different nutrient conditions have distinct constraints on the TOR pathway and the function of Mds3 and Sit4 in the TOR pathway.
Since the function of Sit4 in S. cerevisiae promotes rapamycin resistance (14), we predicted that the function of Mds3 in S. cerevisiae would be similar. To test this hypothesis, we determined the rapamycin sensitivity of congenic wild-type, mds3Δ, pmd1Δ (an MDS3 paralog), and mds3Δ pmd1Δ S. cerevisiae strains (Fig. (Fig.7B).7B). On rich medium, the wild type and pmd1Δ mutant grew similarly in the presence of rapamycin, and the mds3Δ and mds3Δ pmd1Δ mutants were more sensitive to rapamycin. Similar results were observed in nutrient-poor Spider medium; however, we noted that the mds3Δ pmd1Δ double mutant grew extremely poorly on this medium (Fig. (Fig.7B).7B). Since the rapamycin sensitivities of the mds3Δ and mds3Δ pmd1Δ mutants were similar and since the pmd1Δ mutant did not confer rapamycin sensitivity, we conclude that Mds3, but not Pmd1, functions in the TOR pathway. Cells lacking Sit4 or Mds3 have similar phenotypes in relation to rapamycin sensitivity; however, these phenotypes are disparate between S. cerevisiae and C. albicans. Thus, we conclude that Mds3 and Sit4 function similarly in C. albicans and S. cerevisiae but that the effect of these functions has changed since C. albicans and S. cerevisiae split evolutionarily.
Because the mds3Δ/Δ and sit4Δ/Δ mutants share a number of phenotypic traits, we wanted to determine if the sit4Δ/Δ mutant had other mds3Δ/Δ-dependent phenotypes (18). While the sit4Δ/Δ mutant has a slight growth defect on YPD, YPD plus LiCl, or YPD pH 6 medium, the sit4Δ/Δ mutant had a severe growth defect on YPD pH 9 medium that was rescued by the reintroduction of a wild-type copy of SIT4 (Fig. (Fig.7C).7C). The mds3Δ/Δ sit4Δ/Δ double mutant grew more poorly on both YPD and YPD pH 9 medium than either single mutant (Fig. (Fig.7C),7C), supporting the idea that Mds3 and Sit4 have some independent functions. The mds3Δ/Δ sit4Δ/Δ double mutant grew similarly to the mds3Δ/Δ mutant on LiCl medium, suggesting that Sit4 does not play a role in response to this stress medium. These results support a model in which Mds3 and Sit4 function similarly in the TOR pathway but are not completely dependent.
Mds3 was identified as a protein that interacts with Sit4 in an S. cerevisiae proteomic screen (27, 28). Since we have demonstrated a strong genetic interaction between Mds3 and Sit4, we tested whether Mds3 and Sit4 can indeed physically interact. Thus, we generated a strain carrying a functional C-terminal 13× Myc-tagged SIT4 allele and then introduced a functional MDS3 allele containing the 3× HA epitope. Whole-cell extracts of cells grown in M199 pH 8 medium were immunoprecipitated with α-HA and separated by SDS-PAGE. Sit4-Myc was detected in similar amounts in whole-cell extracts from strains with and without Mds3-HA (Fig. (Fig.8).8). In α-HA immunoprecipitates from cells containing Mds3-HA, Sit4-Myc was pulled down. Sit4-Myc was not detected in anti-HA immunoprecipitates in the absence of Mds3-HA, demonstrating that Sit4-Myc is associating specifically with Mds3-HA. Similar results were observed in whole-cell extracts from cells grown in YPD, indicating that this interaction is not dependent on nutrient availability (Fig. (Fig.8).8). Thus, Mds3 and Sit4 physically interact, demonstrating that Mds3 is a member of the TOR pathway.
Growth and morphogenesis is controlled by numerous signal transduction pathways, which respond to diverse environmental cues. The TOR pathway is a key regulator of growth and morphogenesis in eukaryotes, and in C. albicans, the TOR pathway governs the yeast-to-hyphal morphogenetic transition, which is critical for pathogenesis (14, 77). Microarray analyses suggested that MDS3 is a member of the TOR pathway in C. albicans, and this idea was supported by genetic, pharmacological, and biochemical approaches. Thus, Mds3 represents a new member of the TOR morphogenetic pathway.
Mds3 is a 1,383-amino-acid protein with predicted Kelch repeats in the N-terminal region but no other obvious motifs and functional domains (18). Kelch repeats form a β-propeller structure that functions as a protein-protein interaction domain and is found in an array of proteins, including cytoskeletal proteins and signal transduction proteins (2, 25, 34, 50). To the best of our knowledge, Mds3 represents the first Kelch repeat protein associated with the TOR pathway. However, two components of the TOR complex, Lst8 and Kog1 (GβL and Raptor in mammalian systems), contain WD-40 domains, which also fold into β-propeller structures (9, 31, 41, 42, 49). Thus, we propose that Mds3 acts as a scaffold to facilitate interactions between TOR pathway members to control morphogenetic processes in C. albicans.
Based on our results, which include the transcriptional and morphological similarities between the mds3Δ/Δ and sit4Δ/Δ mutants that are rescued by rapamycin and the interaction between Mds3 and Sit4, we propose two models to explain how Mds3 acts in the TOR pathway (Fig. (Fig.9).9). First, Mds3 may act upstream of TOR as a member of a nutrient-sensitive complex, such as the one composed of Raptor, GβL, and mTOR (41, 42). Support for this model comes from the fact that rapamycin, which inhibits TOR kinase activity, rescues the transcriptional defects and morphogenetic defects associated with the mds3Δ/Δ mutant (Fig. (Fig.22 to to4).4). Further, the rapamycin-mediated rescue of the mds3Δ/Δ mutant is lost with the introduction of a TOR1-1 allele or the loss of Rbp1 (Fig. (Fig.22 and and5).5). Based on these data, Mds3 appears to act upstream of TOR. However, for this model to be accurate, Mds3 interacts indirectly with Sit4.
How could Mds3 function upstream of TOR? One possibility is that Mds3 modulates the interaction of Kog1 and Lst8, Raptor and GβL in mammalian systems, with TOR. GβL promotes TOR kinase activity, and Raptor-GβL constitutively interacts with mTOR. However, Raptor-GβL-mTOR interaction is affected by nutrient signals (41). Under nutrient-poor conditions Raptor-mTOR interaction is stabilized, which interferes with the GβL stimulation of mTOR (42). In yeast, Kog1-Lst8 interaction with Tor1 does not appear to be affected by nutrient status (49). This led to the hypothesis that an additional factor, such as Mds3, is involved that affects the activity of Kog1 and/or Lst8 on Tor1.
An alternative model is that Mds3 acts downstream of Tor1 to promote starvation responses when Tor1 is inactive (Fig. (Fig.9).9). Support for this model comes from the fact that the sit4Δ/Δ mutant has transcriptional and morphological defects similar to those of the mds3Δ/Δ mutant, and that these defects are rescued by rapamycin (Fig. (Fig.22 to to5).5). Additionally, both mds3Δ/Δ and sit4Δ/Δ mutants are sensitive to alkaline pH (Fig. (Fig.7C).7C). Since rapamycin inhibits TOR kinase activity, these results suggest that the transcriptional and morphogenetic defects associated with the mds3Δ/Δ and sit4Δ/Δ mutants are due to an increase in TOR kinase activity or TOR kinase-dependent activities, and that reduced kinase activity by rapamycin is sufficient to rescue the sit4Δ/Δ and mds3Δ/Δ mutants. Finally, the fact that Mds3 and Sit4 coimmunoprecipitate suggests that these proteins interact within the same complex either directly or indirectly. One potential problem for this model is the restoration of GAP2 expression by rapamycin in the mds3Δ/Δ mutant (Fig. (Fig.2).2). However, GAP2 expression also is rescued by rapamycin in the sit4Δ/Δ mutant. Thus, additional PP2As, which are activated by rapamycin treatment, may promote the expression of GAP2 independently of Mds3 or Sit4. A similar phenomenon has been described for S. cerevisiae (11, 29, 69). Finally, we noted that the TOR1-1 allele and mds3Δ/Δ mutation had a synergistic effect, which generally suggests parallel pathways. However, the TOR1-1 allele and similar alleles have altered functions in both S. pombe and mammalian systems (55, 73, 75), and we find that TOR1-1 confers growth defects in the absence of a wild-type TOR1 allele (Fig. (Fig.5A5A and and6B).6B). Thus, we propose that the TOR1-1 mds3Δ/Δ synergism reflects defective kinase function in Tor1-1, which exacerbates the TOR pathway defects associated with the mds3Δ/Δ mutant.
How could Mds3 function downstream of TOR? One possibility is that Mds3 acts as a scaffold bridging TOR activity to its effectors. For example, Sit4 activity is dependent upon its association with Tap42 and several Sit4-associated proteins (SAPs), and Mds3 may govern these associations (20, 22, 51). Mds3 may facilitate or stabilize these interactions to promote Sit4 activity. Support for this idea comes from large-scale protein-protein interaction analyses that identified Sap185, a Sit4 regulatory protein, as an Mds3 binding partner (20, 27, 28, 51).
TOR kinase activity is controlled by nutrient availability (12, 20, 41, 66). In nutrient-rich conditions, TOR kinase activity is stimulated, leading to the activation of growth processes and the inactivation of starvation responses. In starvation conditions, TOR kinase activity is inhibited, preventing the activation of growth responses and leading to the activation of starvation responses. Here, we found that wild-type cells promote the expression of RPS26A and repression of GAP2 when grown in YPD but promote the expression of RPS26A, although to reduced levels compared to those for YPD, and GAP2 when grown in M199 (Fig. (Fig.2).2). This suggests that M199 is nutrient poor compared to YPD. Further, wild-type cells divide slower in M199 pH 8 medium than in M199 pH 4 medium, suggesting that M199 pH 8 is more poor than M199 pH 4 medium. The latter growth difference can be explained by the inhibition of plasma membrane transporters and the dependence on the endocytic uptake of nutrients in alkali environments (57, 71, 76). Thus, YPD > M199 pH 4 > M199 pH 8 in nutrient availability, and cells grown in these media may require different levels of TOR kinase activity.
We found that the mds3Δ/Δ and sit4Δ/Δ mutants did not affect the expression of TOR-dependent genes when grown in YPD, as expected if Mds3 and Sit4 promote TOR-dependent starvation responses (Fig. (Fig.2).2). In M199 pH 8 medium and Spider medium, the mds3Δ/Δ and sit4Δ/Δ mutants had similar defects, suggesting that both media are nutrient poor and require TOR-dependent starvation responses. In M199 pH 4 medium, the mds3Δ/Δ and sit4Δ/Δ mutants did not repress CDC19, RPS26A, or TEF1, nor did they induce GAP2 to wild-type levels, supporting the idea that M199 pH 4 medium is a semistarvation medium that requires some level of the TOR-dependent starvation response. In M199 pH 4 medium, the sit4Δ/Δ mutant clearly had a more severe defect in the repression of CDC19 and RPS26A and induction of GAP2 than the mds3Δ/Δ mutant. This represents the first demonstration that Sit4 negatively regulates RP synthesis and is in contrast to work with S. cerevisiae, which suggests that Sit4 does not govern RP synthesis (23, 67). However, these results suggest that Sit4 plays a more general role in promoting TOR-dependent starvation responses than Mds3. For example, Sit4 may be required under both semistarvation and starvation conditions, whereas Mds3 may function only under starvation conditions. This idea is supported by the fact that the mds3Δ/Δ mutant was resistant to rapamycin in nutrient-poor medium but not in rich medium, whereas the sit4Δ/Δ mutant was resistant to rapamycin in both media (Fig. (Fig.7A7A).
MDS3 was first identified in S. cerevisiae as a multicopy suppressor of the Δmck1 sporulation defect and subsequently was identified as a regulator of the C. albicans yeast-to-hypha transition (5, 18). These morphogenetic processes are promoted by neutral-alkaline environmental pH, which led to the idea that Mds3 is a pH response regulator that is similar to Rim101. This idea was supported by the fact that Mds3 is required for wild-type growth at alkaline pH (18). However, Mds3 also is required for chlamydospore and biofilm formation in C. albicans, neither of which require neutral-alkaline pH (60, 65). We also found that the mds3Δ/Δ mutant has a slight growth defect on rich medium, which is acidic, and our microarray studies demonstrated that the mds3Δ/Δ mutant showed transcriptional changes at both acidic and alkaline pH (Fig. (Fig.11 to to3A3A and and66 and Tables Tables33 and and4;4; also see Table S1 in the supplemental material). Thus, Mds3 clearly plays a broader role than neutral-alkaline pH regulation in C. albicans, and we propose that the morphogenetic defects observed in the mds3Δ/Δ mutant reflect defects in general cellular processes that also are required for growth and filamentation at alkaline pH.
In S. cerevisiae, Mds3 and Pmd1 are proposed to act upstream of or in parallel to the Ras/PKA pathway (5, 54), because the sporulation defect of the mds3Δ pmd1Δ double mutant was rescued by the hyperactive RAS2V19 allele (5). However, several lines of evidence suggest that Mds3 is unlikely to function upstream of Ras independently of TOR. First, the hyperactive RAS2V19 allele rescues TOR pathway mutants in S. cerevisiae, suggesting that Ras2 acts in parallel or downstream of the TOR pathway (14, 67, 81). Second, the filamentation defect of the ras1Δ/Δ mutant is not rescued by rapamycin, unlike the mds3Δ/Δ mutant (Fig. (Fig.77 and reference 26, 45). Third, we found that both the Δsit4 and Δmds3 mutants are sensitive to rapamycin in S. cerevisiae. Thus, we propose that Mds3 acts in the TOR pathway in both S. cerevisiae and C. albicans, but that the hyperactive RAS2V19 allele is able to bypass this by acting either in parallel or downstream of TOR.
Why does S. cerevisiae encode two MDS3 paralogs? In S. cerevisiae, phenotypes associated with the Δmds3 mutant are masked by the functionally redundant PMD1 (5, 18). However, we found that the S. cerevisiae Δmds3 mutant, like the Δsit4 mutant but not the Δpmd1 mutant, was sensitive to rapamycin, which is the first demonstration that Mds3 and Pmd1 are not completely redundant in S. cerevisiae. In C. albicans, the mds3Δ/Δ and sit4Δ/Δ mutants behave similarly; however, our results suggest that Mds3 has Sit4-independent functions. First, the mds3Δ/Δ mutant is sensitive to high concentrations of LiCl, whereas the sit4Δ/Δ mutant is not (Fig. (Fig.7C7C and reference 18). Second, the mds3Δ/Δ sit4Δ/Δ double mutant has more severe growth defects on YPD and at alkaline pH than either single mutant (Fig. (Fig.7C).7C). Third, the filamentation defect of the mds3Δ/Δ sit4Δ/Δ double mutant is not rescued by rapamycin (Fig. (Fig.33 and and4).4). Thus, in C. albicans, Mds3 has Sit4-dependent and -independent function. This raises the interesting possibility that in S. cerevisiae, Mds3 governs the Sit4-dependent functions, rapamycin sensitivity, and Pmd1 governs the Sit4-independent functions. It is not yet known if Sit4-independent functions of Mds3 in C. albicans occur through the TOR pathway or completely independent of the TOR pathway.
While Mds3 appears to act in the TOR pathway in both C. albicans and S. cerevisiae, our results have demonstrated that the roles of the TOR pathway in these organisms have diverged. In S. cerevisiae, sporulation requires both nitrogen and carbon starvation and is negatively regulated by Mds3. This suggests that Mds3 functions as a positive regulator of the TOR pathway (5, 54). However, we have clearly shown that Mds3 functions as a negative regulator of the TOR pathway in C. albicans during starvation conditions. Thus, while the TOR pathway is mechanistically similar in C. albicans and S. cerevisiae, the signals coming in to and/or responses coming out of the TOR pathway appear to be distinct. This idea is supported by the rapamycin resistance observed for the mds3Δ/Δ and sit4Δ/Δ mutants in C. albicans compared to the sensitivity observed for these mutants in S. cerevisiae (Fig. (Fig.7B7B and reference 14, 46). Regardless, several signal transduction pathways have been identified that, while mechanistically similar, lead to distinct responses in C. albicans and S. cerevisiae, which is not surprising given the ~300 million years of evolutionary distance between these organisms (37, 39, 52, 70).
Despite the fact that rapamycin initially triggered interest because of its antifungal properties against C. albicans (72), the TOR pathway has received relatively little attention in this organism. In fact, most work has focused on the role of TOR in multicellular organisms in processes as diverse as angiogenesis and embryogenesis. However, TOR appears to have conserved functions in growth and stress responses in both multicellular and unicellular eukaryotes (64, 79). Our studies demonstrate that the kelch repeat protein Mds3, which promotes morphogenetic processes critical for pathogenesis, such as the yeast-to-hypha transition and biofilm formation, is a new member of the TOR pathway (15, 44, 59). Our study represents the first example in which a kelch repeat protein has been implicated in the TOR pathway and raises the possibility that an Mds3-like protein exists in mammalian systems to promote TOR-dependent responses.
This work was supported by the Investigators in Pathogenesis of Infectious Disease Award from the Burroughs Wellcome Fund to D.A.D. and by the NIH National Institute of Allergy and Infectious Diseases award R01-AI064054 to D.A.D.
We thank Eric Bensen and Zhen Jin Tu for assistance with microarray analysis and Gabriela Vazquez for statistical analysis. We also are indebted to Ed Winter, Doreen Harcus, Joseph Heitman, and Malcolm Whiteway for strains. Finally, we are grateful to Judith Berman, Kirsten Nielsen, Timothy Brickman, Do-Hyung Kim, Thomas Neufeld, and the members of the Davis laboratory for numerous helpful discussions and the critical reading of the manuscript.
Published ahead of print on 10 May 2010.
†Supplemental material for this article may be found at http://mcb.asm.org/.