|Home | About | Journals | Submit | Contact Us | Français|
The environmental niche of each fungus places distinct functional demands on the cell wall. Hence cell wall regulatory pathways may be highly divergent. We have pursued this hypothesis through analysis of Candida albicans transcription factor mutants that are hypersensitive to caspofungin, an inhibitor of beta-1,3-glucan synthase. We report here that mutations in SKO1 cause this phenotype. C. albicans Sko1 undergoes Hog1-dependent phosphorylation after osmotic stress, like its Saccharomyces cerevisiae orthologues, thus arguing that this Hog1-Sko1 relationship is conserved. However, Sko1 has a distinct role in the response to cell wall inhibition because 1) sko1 mutants are much more sensitive to caspofungin than hog1 mutants; 2) Sko1 does not undergo detectable phosphorylation in response to caspofungin; 3) SKO1 transcript levels are induced by caspofungin in both wild-type and hog1 mutant strains; and 4) sko1 mutants are defective in expression of caspofungin-inducible genes that are not induced by osmotic stress. Upstream Sko1 regulators were identified from a panel of caspofungin-hypersensitive protein kinase–defective mutants. Our results show that protein kinase Psk1 is required for expression of SKO1 and of Sko1-dependent genes in response to caspofungin. Thus Psk1 and Sko1 lie in a newly described signal transduction pathway.
The fungal cell wall is critical for interaction with the environment and survival. It is the point of contact between the fungus and target surfaces, and processes such as adhesion, dimorphism, and biofilm formation are dependent on a dynamic cell wall (Nobile and Mitchell, 2005 ; Lesage and Bussey, 2006 ; Ruiz-Herrera et al., 2006 ; Dranginis et al., 2007 ). These processes all contribute to the pathogenicity of Candida albicans, the major fungal pathogen of humans. This organism causes superficial, mucosal, and potentially fatal invasive infections (Rangel-Frausto et al., 1999 ; Rabkin et al., 2000 ). As a fungal-specific structure, the cell wall is also of interest as a mediator of immunological recognition and evasion (Wheeler and Fink, 2006 ) and in addition as a target of antifungal drugs such as caspofungin (Letscher-Bru and Herbrecht, 2003 ). Our interest is in the signaling pathways that govern C. albicans cell wall dynamics.
Caspofungin inhibits beta-glucan synthesis to cause cell lysis (Letscher-Bru and Herbrecht, 2003 ). Caspofungin treatment elicits a broad transcriptional response in the baker's yeast Saccharomyces cerevisiae and in C. albicans (Reinoso-Martin et al., 2003 ; Liu et al., 2005 ; Bruno et al., 2006 ). The S. cerevisiae response is controlled in part by the mitogen-activated protein kinase (MAPK) signaling cascade known as the protein kinase C (PKC) cell wall integrity pathway (Reinoso-Martin et al., 2003 ; Levin, 2005 ; Liu et al., 2005 ; Bruno et al., 2006 ). This MAPK pathway is conserved in C. albicans, where it also governs cell wall integrity (Navarro-Garcia et al., 1998 ; Reinoso-Martin et al., 2003 ). However, there is increasing evidence that the C. albicans response to caspofungin has unique features as well. For example, the C. albicans response includes induction of numerous secretory genes (Bruno et al., 2006 ), a gene class that is largely nonresponsive in S. cerevisiae. Even more striking is the fact that a major mediator of the C. albicans response, transcription factor Cas5, lacks an S. cerevisiae orthologues (Bruno et al., 2006 ). Cas5 is required for induction of genes mainly involved in cell wall biogenesis. Those genes account for a small fraction of caspofungin-responsive genes.
In this study we use a genetic screen to identify new C. albicans transcription factors involved in cell wall damage signaling. We also employ a new resource, a set of caspofungin-sensitive protein kinase mutants (Blankenship, Fanning, Hamaker, and Mitchell, unpublished data), to search for upstream signaling components. We uncover a novel cell wall regulatory pathway that includes the transcription factor Sko1 (ORF 19.1032) and the protein kinase Psk1 (ORF 19.7451). In S. cerevisiae both ScSko1 and the proteins ScPsk1 and ScPSk2 have been characterized. ScSko1 mediates the adaptive response to osmotic stress via the high-osmolarity glycerol (HOG) pathway. ScSko1 is activated through phosphorylation by the MAP kinase ScHog1 and functions as a activator and repressor of osmotic stress–responsive genes (Proft et al., 2001 ; Proft and Struhl, 2002 ). ScSko1 function has not been characterized in the response to cell wall damage. Gene expression studies implicate C. albicans Sko1 in the osmotic stress response (Enjalbert et al., 2006 ), but no sko1 mutant defect has been reported previously (Braun et al., 2001 ). ScPsk1/2 regulates glucose partitioning for either glucan or glycogen synthesis, and Scpsk1 Scpsk2 double mutants are sensitive to cell wall damage (Smith and Rutter, 2007 ). The sole C. albicans orthologue Psk1 has not been characterized previously. Our findings define a new regulatory pathway that governs a critical aspect of C. albicans growth and survival.
C. albicans cultures were prepared in YPD plus uridine (2% dextrose, 2% bacto peptone, 1% yeast extract, and 80 mg/l uridine) at 30°C with shaking at 200 rpm. Synthetic medium (2% dextrose, 6.7% yeast nitrogen base [YNB] plus ammonium sulfate, and the necessary auxotrophic supplements) was used for selection after transformations. In assays monitoring cell wall damage, cells were plated to YPD + uridine supplemented with 125 ng/ml caspofungin (Merck, Rahway, NJ).
All primers used in this study are listed in Table 1. The SKO1 complementing plasmid (pRM03) was constructed as follows: Primers SKO1compfwd and SKO1comprev were used to amplify a 2.4-kb fragment containing 993 bp of promoter, the entire open reading frame (ORF), and 220 bp of the 3′UTR. The recently discovered 109 base pairs of intron sequence in the 5′UTR is included the 2.4-kb fragment. The amplicon was ligated to the pGEMT-Easy vector (Promega, Madison, WI) to create pGEMTE-SKO1 and amplified in Escherichia coli. Purified pGEMTE-SKO1 was digested with NgoMIV and AlwNI and inserted through in vivo recombination in S. cerevisiae into a NotI- and EcoRI-digested pDDB78 (Spreghini et al., 2003 ). The cloned SKO1 insert was verified by DNA sequencing.
The HOG1 complementing plasmid (pRM04) was constructed as follows: Primers HOG1compfwd and HOG1deldet were used to amplify a 2.3-kb fragment containing 1 kb of promoter, the entire ORF, and 189 bp of the 3′UTR. The amplicon was ligated to pGEMT–Easy (pGEMTE-HOG1) and inserted into pDDB78 as described above to generate pRM04.
Construction of a SKO1-V5 epitope-tagged plasmid (pRM05) was performed as follows: Primers SKO1compfwd and SKO1orfrev were used to generate a fragment containing 993 bp of promoter and the entire ORF without the stop codon. The amplicon was inserted into the pYES2.1/V5-His-TOPO vector (Invitrogen, Carlsbad, CA) to create pYES-SKO1-V5. PCR amplification using primers SKO1-V5 fwdpr and CAS5-V5 78 3′ with pYES-SKO1-V5 as a template was done to amplify a fragment containing the V5 epitope tag, His 6x tag, stop codon, and 209 bp of the CYC1 terminator region. This fragment was inserted into linearized pDDB78 as described above.
The PSK1-complementing plasmid (pRM06) was constructed as follows: Primers PSK1compfwd and PSK1comprev were used to generate a 5.3-kb fragment consisting of 965 bp of promoter region, the entire ORF, and 385 bp of the 3′UTR. This fragment was ligated into pGEMT-easy to create pGEMTE-PSK1 and amplified in E. coli. Purified pGEMTE-PSK1 was digested with NgoMIV and SapI and inserted through in vivo recombination in S. cerevisiae into a NotI- and EcoRI-digested pRYS2.
C. albicans strains used in this study are listed in Table 2. All strains were derived from strain BWP17 (genotype: ura3Δ::λimm434/ura3Δ::λimm434 his1::hisG/his1::hisG arg4::hisg/arg4::hisG; Wilson et al., 1999 ). Strain JMR103, the sko1Δ::ARG4/sko1Δ::URA3 mutant was generated by PCR-directed gene deletion using 120mer oligonucleotides SKO1del5′dr and SKO1del3′dr, respectively, to delete the entire ORF (Wilson et al., 1999 ). The SKO1-complemented strain (JMR109) was generated by transforming JMR103 with NruI-digested pRM03 to direct integration to the HIS1 locus. JMR103 was brought to His prototrophy through transformation with NruI-digested pDDB78 to create strain JMR104. Strain JMR114, the hog1Δ::ARG4/hog1Δ::URA3 mutant, was generated using primers HOG1del5′dr and HOG1del3′dr as described above. JMR114 was also brought to His prototrophy through transformation with NruI-digested pDDB78 to create strain JMR121. The HOG1-complemented strain (JMR123) was constructed as described above. Strain JMR167, the psk1Δ::ARG4/psk1Δ::URA3 mutant was generated using primers PSK1del5′dr and PSK1del3′dr as described above. JMR167 was brought to His prototrophy through transformation with SrfI-digested pRYS2 to create strain JMR192. The PSK1-complemented strain (JMR188) was constructed as described above. SKO1-V5 epitope-tagged strains were generated through transformation of NruI-digested pRM05 as described above. Candidate genes related to the transcription process were described previously (Nobile and Mitchell, 2005 ). Construction of the insertion mutant strains followed previously described procedures (Davis et al., 2002 ; Norice et al., 2007 ).
Assays followed previously described procedures (Bruno et al., 2006 ). Briefly, C. albicans overnight cultures were diluted to a starting OD600 nm of 3.0. Samples were serially diluted, spotted onto designated plates, incubated at 30°C, and photographed after 1–3 d of growth.
Overnight cultures of designated C. albicans strains were diluted to a starting OD600nm of 0.200 in 100 ml fresh YPD + uridine media. The cultures were incubated with shaking at 30°C to an OD600 nm of 1.0 and spilt into two 50-ml cultures. A total of 125 ng of caspofungin was added to the experimental culture, and dH2O was added to the control culture. The cultures were incubated for 30–60 min. Cells were harvested by vacuum filtration and stored at −80°C. For kinetic assays a starter culture of 400 ml was prepared as described above, and after caspofungin treatment, 50-ml samples were collected at each designated time point. Total RNA was isolated using the hot acid phenol method (Nobile and Mitchell, 2005 ). RNA yield and purity levels were determined spectrophotometrically, and 5 μg of RNA was DNase digested (RQ1 DNase, Promega; or DNaseI, Ambion, Austin, TX). cDNA was synthesized using the Stratascript first strand synthesis kit (Stratagene, La Jolla, CA). As a control for DNA contamination each sample was treated without reverse transcriptase. Primers are listed in Table 1 and were designed using primer 3 input software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). PCR efficiency (E) was determined for all primers through amplification of C. albicans genomic DNA. Primer pairs yielding E-values between 99 and 103% were used in subsequent real-time (RT) experiments. RT reactions were prepared in triplicate using iQ SYBR supermix (Bio-Rad), and RT-PCR was performed using the Bio-Rad I Cycler thermocycler equipped with an iQ5 multicolor optical unit (Bio-Rad, Richmond, CA), with a program of 95°C for 5 min and then 40 cycles of 95°C for 45 s, followed by 58°C for 1 min. Melt curve analysis confirmed the specificity of the amplification products. Data analysis was conducted using the Bio-Rad iQ5 standard edition optical system software V2.0. Transcript levels were normalized against TDH3 (which encodes glyceraldehyde-3-phosphate dehydrogenase) expression, and gene expression changes were calculated by the ΔΔCT method (Kubista et al., 2006 ). Target gene fold changes for treated or untreated cells were determined by comparison to the wild-type (wt) strain. Significant differences between groups were determined in unpaired t tests (http://graphpad.com/quickcalcs/ttest1.cfm?Format=SD) with a p value of < 0.05 considered to be statistically significant.
Cultures of designated C. albicans strains were prepared as described above. Cultures were incubated in the presence of caspofungin for 30 min before harvesting by vacuum filtration. Cells were resuspended in 1.5 ml of ice-cold RNA later (Sigma, St. Louis, MO) to prevent RNA degradation and pelleted. Total RNA was extracted and was DNase treated using the Ribopure yeast kit (Ambion) following manufacturer's instructions. We performed two hybridizations that measured the effects of drug treatment on wt cells, and six hybridizations that compared transcripts from drug-treated mutant cells with drug-treated wt cells. All RNA samples were produced from independent cultures. Transcriptional profiling was performed as previously described (Nantel et al., 2006 ), and the resulting data were normalized and analyzed in GeneSpring GX version 7.3 (Agilent Technologies, Wilmington, DE). The results of this analysis are listed in Supplementary Dataset 1, which includes significantly modulated genes that exhibited a statistically significant (t test; p < 0.05) change in transcript abundance of at least 1.5-fold. Gene annotations were determined using the gene ontology term-finder tool for “process” from the Candida Genome Database Web page (http://www.candidagenome.org/cgi-bin/GO/goTermFinder).
C. albicans overnight cultures were collected and diluted to a starting OD600 nm of 0.200 in 100 ml fresh YPD + uridine media. The cultures were incubated with shaking at 30°C to an OD600 nm of 1.0 and spilt into two 50-ml cultures. The experimental culture was incubated with1.5 M NaCl for 10 min to induce osmotic shock, and the control culture was treated with dH2O. For experiments monitoring cell wall damage, the experimental culture was incubated for 1 h with 125 ng caspofungin, and dH2O was added to the control culture. For kinetic assays a 400-ml starter culture was prepared as described above, and after caspofungin treatment, 50-ml samples were collected at each designated time point. Cells were harvested by vacuum filtration, resuspended in ice cold 20% TCA, and incubated on ice for 30 min. The cells were pelleted at 14,000 rpm for 20 min. A solution of alkaline-buffered acetone was prepared by mixing three parts of 3 M Tris, pH 8.8, to seven parts acetone and was used to wash the pellet twice. The pellet was air-dried and resuspended in 8 M urea. Approximately 100 μl of acid-washed glass beads was added to the cell suspension, and the cells were lysed in a bullet blender (Next Advance, Averill Park, NY). The lysate was pelleted and supernatant was collected. Protein concentration was determined using the Bradford protein assay (Bio-Rad). Cellular lysates were treated with or without calf intestinal phosphatase (New England Biolabs, Beverly, MA) in the presence or absence of phosphatase inhibitors (Sigma). Fifteen micrograms of sample was electrophoresed on 8% SDS polyacrylamide gels, transferred onto PVDF membranes, and stained with Ponceau dye to ensure equal sample loading. Sko1-V5 was probed and detected on immunoblots using anti-V5 monoclonal antibodies conjugated to horseradish peroxidase (Invitrogen) at a 1:2500 dilution and the ECL plus Western blotting chemiluminescent detection system (Amersham, Piscataway, NJ), respectively.
To find regulators of the cell wall damage response, we attempted to create homozygous insertion mutants for 67 genes that were related to the transcription process (Table 3). We were unable to create mutants in 34 of these genes, some of which may be essential. We note that the S. cerevisiae orthologues of 13 of these genes are essential, but homozygous C. albicans mutants for another six of these genes have been made previously by other methods. We screened the mutants we recovered in 33 genes for altered growth on caspofungin medium and found a caspofungin-sensitive strain with an insertion in SKO1 (Table 3).
Sko1 is orthologous to the S. cerevisiae transcription factor ScSko1, which functions in the osmotic stress response. To verify that Sko1 governs caspofungin sensitivity in C. albicans, we constructed a sko1Δ/Δ deletion mutant. Growth of the sko1Δ/Δ mutant was drastically reduced on caspofungin plates compared with nutrient YPD plates (Figure 1). Similar results were observed using another independent sko1Δ/Δ mutant (derived from an independent heterozygote; data not shown). The caspofungin-hypersensitive phenotype of both mutants was complemented by introduction of a wt copy of SKO1 (Figure 1 and data not shown), indicating that the sko1Δ mutation is the cause of caspofungin hypersensitivity. These findings show that SKO1 is required for normal caspofungin sensitivity.
Transcription factors are often induced under conditions that require their biological activity. Thus, we hypothesized that caspofungin treatment may induce SKO1 expression. We measured SKO1 transcript levels by RT-PCR after caspofungin treatment. SKO1 was up-regulated sixfold in wt cells treated with caspofungin (Figure 2A). SKO1 expression was not detected in the sko1Δ/Δ deletion mutant, thus confirming primer specificity, and was restored to wt levels in the sko1Δ/Δ/+-complemented strain (Figure 2B). To monitor Sko1 protein levels, we constructed a strain carrying a functional epitope-tagged Sko1-V5 (Figure 1). Consistent with our gene expression results, Western blotting analysis showed that there was an increase in the amount of Sko1-V5 protein levels after caspofungin treatment (Figure 2C). We conclude that caspofungin induces SKO1 gene expression and protein accumulation.
We considered the possibility that Sko1 may be required for expression of caspofungin-responsive genes. Alternatively, Sko1 may be required for expression of osmotic stress response genes that promote survival after cell wall damage. To test these hypotheses, we monitored expression of the caspofungin-responsive gene PGA13 and the osmotic stress response genes RHR2 and GPD2. PGA13 specifies a cell wall protein and is induced in response to cell wall damage (Bruno et al., 2006 ) but not in response to osmotic stress (Enjalbert et al., 2006 ). Rhr2 and Gpd2 catalyze the synthesis of glycerol, which is critical in adaptation to osmotic stress (Fan et al., 2005 ; Enjalbert et al., 2006 ). We observed that PGA13 was induced in the wt and sko1Δ/Δ/+-complemented strains, but not in the sko1Δ/Δ mutant (Figure 3A). On the other hand, GPD2 and RHR2 expression was similar in the wt strain and sko1Δ/Δ mutant (Figure 3, B and C). Therefore, although caspofungin treatment induces two osmotic stress-responsive genes, this response is independent of Sko1 function. In contrast, induction of the cell wall protein gene PGA13 depends on Sko1 function.
To define Sko1-dependent genes in broader terms, we performed microarray comparisons of the wt strain and sko1Δ/Δ mutant treated with caspofungin (Supplementary Dataset 1, Worksheet 1 and 2). We found that Sko1 regulates 79 caspofungin-responsive genes, including several cell wall biogenesis genes (Supplemental Dataset 1, Worksheet 3). RT-PCR analysis confirmed the reduced expression of cell wall biogenesis genes CRH11, MNN2, and SKN1 in the sko1Δ/Δ mutant treated with caspofungin (Figure 4, A–C). Gene expression levels were restored to wt in the sko1Δ/Δ/+-complemented strain (Figure 4, A–C). Therefore, Sko1 is necessary for expression of many caspofungin-responsive genes.
We noted that carbohydrate metabolic genes, such as the glucose transporter gene HGT6, were significantly overexpressed in the sko1Δ/Δ mutant (Supplementary Table S1, Worksheets 1 and 2). These genes are not induced by caspofungin. RT-PCR assays showed that HGT6 is overexpressed in the sko1Δ/Δ mutant with or without caspofungin treatment (Figure 4D). These findings indicate that Sko1 is a negative regulator of carbon metabolic genes.
To identify upstream regulators of Sko1 activity, we first considered the S. cerevisiae paradigm. The protein kinase ScHog1 activates ScSko1 by phosphorylation in response to osmotic shock, thereby causing a change in ScSko1 electrophoretic mobility (Proft et al., 2001 ). Thus, we considered that C. albicans Hog1 may be a regulator of Sko1 in response to caspofungin treatment. Prior studies have shown that the C. albicans the HOG pathway is important for cell wall biosynthesis and stability (Eisman et al., 2006 ; Enjalbert et al., 2006 ; Munro et al., 2007 ). However, we observed that a hog1Δ/Δ mutant was only slightly hypersensitive to caspofungin compared with the sko1Δ/Δ mutant (Figure 1), and it expressed SKO1 normally (Figure 2A). Protein analysis from wt cells treated with caspofungin showed that Sko1 does not undergo an electrophoretic shift (Figure 2C). On the other hand, we observed a Sko1 electrophoretic shift after osmotic shock in wt cells but not in the hog1Δ/Δ mutant strain (Figure 5A). The Sko1 electrophoretic shift was sensitive to phosphatase treatment (Figure 5B). These results suggest that Hog1 phosphorylates Sko1 after osmotic stress, but argue that the HOG pathway does not regulate Sko1 after caspofungin-induced cell wall damage.
We have recently identified insertion mutants in several protein kinase–related genes that are hypersensitive to caspofungin (Blankenship, Fanning, Hamaker, and Mitchell, unpublished data). Those protein kinases are additional candidate SKO1 regulators. We found that SKO1 expression was similar to wt in eight mutants, reduced about twofold in four mutants, and increased about twofold in four mutants. We note that SKO1 expression was increased in all mutants of the PKC-signaling pathway (Figure 6). SKO1 expression was most severely reduced in the psk1−/− mutant (Figure 6). Indeed, several independent psk1Δ/Δ deletion strains were hypersensitive to caspofungin (Figure 1 and data not shown), a phenotype that was complemented by a wt PSK1 allele (Figure 1). SKO1 was expressed at its uninduced level in three independent psk1Δ/Δ mutant deletion mutants, regardless of caspofungin treatment (Figure 7A and data not shown). Therefore, Psk1 is a positive regulator of SKO1 expression in caspofungin-treated cells.
Our observations predict that a psk1Δ mutation will affect expression of Sko1 target genes. RT-PCR assays showed reduced expression of PGA13 and MNN2 and the increased expression of HGT6 in psk1Δ/Δ cells, compared with wt or complemented strains (Figure 8, A and B). Interestingly, HGT6 was overexpressed in the psk1Δ/Δ mutant only after caspofungin treatment (Figure 8C), the circumstance in which the mutant has reduced expression of SKO1 (Figure 7). These results support the model that Psk1 is required for functional expression of SKO1 in response to caspofungin.
The fungal cell wall has vital roles in growth, survival, morphogenesis, and pathogenicity. Critical for the coordination of these activities is the dynamic nature of the cell wall, its ability to respond to external and internal stimuli. We propose that the distinct evolutionary paths of each fungal species may be reflected in unique cell wall regulatory pathways. Our identification of a C. albicans Psk1-Sko1 pathway (Figure 9) lends support to this idea. Inhibition of cell wall biogenesis by caspofungin causes an increase in SKO1 expression. This increase is dependent on the protein kinase Psk1 and culminates in the expression of diverse genes that are necessary for cell wall stability. Although aspects of Sko1 and Psk1 function are conserved in S. cerevisiae, the connections among Sko1, Psk1, and cell wall perturbation may be unique to C. albicans.
Our findings argue that Sko1 functions in the Hog1-dependent osmotic stress response, a relationship well established in S. cerevisiae (Proft et al., 2001 ; Rep et al., 2001 ). This role was foreshadowed by microarray analysis (Enjalbert et al., 2006 ), which revealed that SKO1 expression is induced 1.5-fold by osmotic stress, dependent on HOG1. Our results point to a second aspect of this relationship: Sko1 undergoes Hog1-dependent phosphorylation after osmotic stress. Hog1 may phosphorylate Sko1 directly, as known for the S. cerevisiae orthologues, because the ScSko1 phosphoacceptor sequence is well conserved in C. albicans Sko1 (Krantz et al., 2006 ). Indeed, sko1Δ/Δ mutants are slightly sensitive to osmotic stress (our unpublished results), so these modes of Sko1 regulation may be functionally significant. Therefore, aspects of the Hog1–Sko1 relationship are conserved in the C. albicans osmotic stress response.
Our findings establish that Sko1 is necessary for the cell wall damage response. In principle, the caspofungin hypersensitivity of the sko1Δ/Δ mutant might have reflected an aberrant osmotic stress response. This response is induced by cell wall perturbation in both S. cerevisiae (Boorsma et al., 2004 ) and, as we show here, in C. albicans. However, two C. albicans osmotic stress genes are induced by caspofungin independently of Sko1. Furthermore, the major Sko1-dependent genes that are induced by caspofungin, such as CRH11, PGA13, and MNN2, are not induced by osmotic stress (Enjalbert et al., 2006 ). The fact that SKO1 is induced by caspofungin in both wt and hog1Δ/Δ strains, along with our failure to detect caspofungin-induced Sko1 phosphorylation, further underscore the independence of Sko1 and Hog1 activities after cell wall perturbation. Therefore, the Hog1–Sko1 paradigm does not account for the role of Sko1 in the cell wall damage response.
Our hypothesis is that the caspofungin-inducible genes that depend upon Sko1 for full expression contribute to the sko1Δ/Δ mutant's caspofungin hypersensitivity. We have identified 26 genes of this class in our experiments. This number includes 25 genes that were induced by caspofungin in the wt strain, as detected (≥1.5-fold) with our current array platform, as well as PGA13, for which induction was detected only by RT-PCR (Supplemental Dataset, Worksheet 4). (Based on the caspofungin-inducible gene set defined by Bruno et al. (2006) with a different array platform, there are 14 genes of this class, as summarized in the Supplemental Dataset Worksheet 5). For example, KRE1, SKN1, PHR1, CRH11, PGA13, PGA31, and MNN2 have all been implicated in cell wall biogenesis (Boone et al., 1991 ; Mio et al., 1997 ; Popolo and Vai, 1998 ; De Groot et al., 2003 ; Pardini et al., 2006 ). In addition, we have observed that mnn2 and pga13 homozygous insertion mutants are caspofungin hypersensitive (our unpublished data). These observations suggest that Sko1-dependent induction of these genes may be critical for an effective response to cell wall damage.
It seems likely that additional Sko1-regulated genes may also influence the sko1Δ/Δ mutant's caspofungin hypersensitivity. Most Sko1-regulated genes are not induced by caspofungin under our treatment conditions. A major subset of these genes is involved in carbohydrate metabolism (p = 6.51 × 10−5 for 46/447 genes; http://www.candidagenome.org/cgi-bin/GO/goTermFinder), such as PFK2 (glycolysis), PCK1 (gluconeogenesis), and REG1 (carbon regulation). Several hexose transporter genes, such as HGT6, are also regulated by Sko1. The cell wall is composed mainly of glucose polymers, so altered flux through carbon metabolic pathways may have significant consequences for cell wall biogenesis. Thus we suggest that multiple classes of Sko1-regulated genes impact the integrity of the cell wall.
Although several studies have revealed that transcription factors have been rewired in C. albicans compared with S. cerevisiae (Kadosh and Johnson, 2001 ; Khalaf and Zitomer, 2001 ; Ihmels et al., 2005 ; Martchenko et al., 2007 ; Banerjee et al., 2008 ), seldom have the relevant upstream regulators been identified. Here we have identified protein kinase Psk1 as a regulator of SKO1 expression. Psk1 is a PAS-domain protein, and PAS-domain proteins of prokaryotes and eukaryotes regulate diverse physiological processes (Rutter et al., 2001 ; Gilles-Gonzalez and Gonzalez, 2004 ). The S. cerevisiae PAS protein kinases ScPsk1 and ScPsk2 control glucose partitioning. S. cerevisiae ScPsk1/2 phosphorylates the enzyme UDP-glucose pyrophosphorylase to stimulate the formation of UDP-glucose, the precursor for glycogen and glucan synthesis (Smith and Rutter, 2007 ). Thus, Scpsk1/2Δ double mutants are sensitive to cell wall–perturbing agents (Smith and Rutter, 2007 ). In this context, it is not surprising that the C. albicans psk1Δ/Δ mutant is caspofungin-hypersensitive. However, its connection to Sko1 is unexpected.
Our conclusion that Psk1 acts upstream of Sko1 is based on two lines of evidence. First, we found that that psk1 insertion and deletion homozygotes express SKO1 RNA at its basal level, even after caspofungin treatment. Thus Psk1 is required specifically for the induction of SKO1 by caspofungin. Second, we observed that two Sko1-dependent cell wall genes, PGA13 and MNN2, are expressed at reduced levels in psk1Δ/Δ mutants. In addition, the Sko1-repressed gene HGT6 is expressed at elevated levels in psk1Δ/Δ mutants. Interestingly, the altered regulation of HGT6 in psk1Δ/Δ mutants argues that the induction of SKO1 by the cell wall damage has functional consequences: HGT6 is expressed at normal levels in psk1Δ/Δ mutants in the absence of caspofungin, when SKO1 is expressed at its basal level. However, HGT6 is overexpressed in psk1Δ/Δ mutants in the presence of caspofungin, when SKO1 induction is defective. We suggest that caspofungin treatment increases the demand for Sko1 activity, which is limiting in psk1Δ/Δ mutants. Limitation of Sko1 activity may partially recapitulate a sko1Δ/Δ mutant phenotype, resulting in elevated HGT6 expression. This observation, along with the caspofungin hypersensitivity of the psk1Δ/Δ mutant, argues that Psk1-dependent induction of SKO1 is critical for an effective response to cell wall perturbation.
The Psk1–Sko1 relationship represents a new cell wall damage signaling pathway. Our gene expression data provide some insight into the outputs of the pathway, though we have not yet distinguished direct Sko1 target genes. Two key aspects of the pathway remain to be discovered. One is the mechanism by which Psk1 regulates Sko1 RNA accumulation. A simple possibility is that Psk1 phosphorylates and activates another transcription factor, which in turn activates SKO1 expression. Transcription factor mutant screens, as reported here and in Bruno et al. (2006) , may identify this component. A second area for future analysis is the mechanism by which Psk1 may sense cell wall perturbation. The similarity of overall Psk1 biological function in S. cerevisiae and C. albicans may indicate that upstream signaling components are conserved, so that gene discovery strategies carried out in both organisms may converge upon these genes. Finally, our results argue that Sko1 lies at the intersection of two C. albicans stress response pathways, defined by Hog1 and Psk1. An interesting possibility is that Sko1 may coordinate these responses.
We thank members of our lab for their advice and discussions, and Carmelle T. Norice for providing useful preliminary observations. We are grateful to Merck Research Labs for providing caspofungin. This is NRC publication number 49558. This work was supported by NIH grant 5R01AI057804, its supplement S1, and fellowship F32AI71439 to JRB.
This article was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E08-02-0191) on April 23, 2008.