Msb2 Signaling Mucin Controls Activation of Cek1 Mitogen-Activated Protein Kinase in Candida albicans

Abstract

We have characterized the role that the Msb2 protein plays in the fungal pathogen Candida albicans by the use of mutants defective in the putative upstream components of the HOG pathway. Msb2, in cooperation with Sho1, controls the activation of the Cek1 mitogen-activated protein kinase under conditions that damage the cell wall, thus defining Msb2 as a signaling element of this pathway in the fungus. msb2 mutants display altered sensitivity to Congo red, caspofungin, zymolyase, or tunicamycin, indicating that this protein is involved in cell wall biogenesis. Msb2 (as well as Sho1 and Hst7) is involved in the transmission of the signal toward Cek1 mediated by the Cdc42 GTPase, as revealed by the use of activated alleles (Cdc42^G12V) of this protein. msb2 mutants have a stronger defective invasion phenotype than sho1 mutants when tested on certain solid media that use mannitol or sucrose as a carbon source or under hypoxia. Interestingly, Msb2 contributes to growth under conditions of high osmolarity when both branches of the HOG pathway are altered, as triple ssk1 msb2 sho1 mutants (but not any single or double mutant) are osmosensitive. However, this phenomenon is independent of the presence of Hog1, as Hog1 phosphorylation, Hog1 translocation to the nucleus, and glycerol accumulation are not affected in this mutant following an osmotic shock. These results reveal essential functions in morphogenesis, invasion, cell wall biogenesis, and growth under conditions of high osmolarity for Msb2 in C. albicans and suggest the divergence and specialization of this signaling pathway in filamentous fungi.

Candida albicans is an important human fungal pathogen, causing infections that may represent a serious health problem. This yeast is found as a commensal in certain body locations (mainly, the vagina and tractointestinal duct) but is able to gain access to different organs under conditions of altered host immune defenses causing severe diseases. Dimorphism, the environmentally regulated differentiation program that allows this fungus to switch between a yeast-like-form (unicellular) and a filamentous form (multicellular) (25, 59, 95), is considered to play an important—albeit not exclusive—role (27, 51, 79, 82, 94) in the virulence of this fungus. Therefore, besides its clinical importance as an opportunistic pathogen, this microbe represents an interesting model of morphogenesis and differentiation in lower eukaryotes.

As occurs with other microbial pathogens, C. albicans must be able to sense environmental signals and develop adaptive responses to survive within the host environment. Mitogen-activated protein (MAP) kinase (MAPK) signal transduction pathways are responsible, in part, for this process (3, 12, 76), and, perhaps not surprisingly, these pathways are important virulence factors (1, 14, 20, 32). The cell integrity pathway is mediated by the Mkc1 MAPK and participates in construction of the cell wall and in response to stress as well as in invasion under defined conditions (45, 60, 62). The Cek1 pathway is also involved in morphogenesis and cell wall formation (14, 15, 77, 83), while the HOG pathway, mediated by the Hog1 MAPK, enables adaptation to both osmotic and oxidative stress (2, 4, 11, 24). Mutants defective in certain elements of the HOG pathway have been shown to be more susceptible to oxidative stress, an essential microbial adaptive mechanism that greatly influences the outcome of infection (78, 91; see also reference 12). They are also more efficiently killed by phagocytic cells of either mouse or human origin (6, 22). This route also influences morphogenesis, and hog1 mutants are hyperfilamentous (1), since Hog1 is a repressor of Cek1 activation (23, 77).

In Saccharomyces cerevisiae, the HOG pathway is mediated by two upstream branches. The first branch relies on a histidine-kinase two-component system that involves Sln1, Ypd1, Ssk1, and the redundant Ssk2 and Ssk22 proteins (MAPK kinase kinases [MAPKKKs]) (35, 48, 68, 69, 86). Eventually, this results in the activation of the Pbs2 MAPK kinase that interacts with and activates Hog1 (57) as well as downstream transcriptions factors that generate the appropriate adaptive transcriptional response. The second branch requires Sho1, the Cdc42 GTPase, Ste20 (PAK), and Ste11 (MAPKKK)/Ste50 (see references 19 and 34 for recent reviews). Sho1 is an adaptor membrane protein that attaches the kinase complex to regions of polarized growth at the plasma membrane and interacts with Pbs2 (52, 53, 71, 97, 98) via a proline-rich domain present in the Pbs2 MAPK kinase. Both transcriptomal and genetic analyses have implicated the Msb2 membrane protein in the HOG pathway (64). Msb2 encodes a mucin-like protein identified as a multicopy suppressor of a cdc24^ts mutant (7) that interacts with Sho1 and Cdc42 (16). Recent work indicates that Msb2 and Hkr1 are the putative osmosensors of the HOG pathway in S. cerevisiae and act coordinately with Sho1 to promote osmotic adaptation (87). Hkr1, however, seems to be specific to the HOG pathway, whereas Msb2, which also participates in filamentous growth (FG), plays dual roles (67). These two branches seem to exist also in C. albicans, although important functional differences between them have been previously demonstrated (3). Both the Ssk1 and Sho1 homologues are involved in morphogenesis and resistance to oxidative stress (9, 11, 77), but Ssk1 is the main component responsible for the transmission of the oxidative activation signal to the Hog1 MAPK (11, 77). Recent investigations have shown that the CaSSK2 homologue to ScSSK2 and ScSSK22 is the only MAPKKK responsible for transmitting the signal to Hog1 (13), while CaSTE11 is mainly involved in cell wall biogenesis.

Given the relevance of the HOG pathway in the virulence of this pathogenic fungus (1, 4, 9, 22), we have undertaken the analysis of the MSB2 gene in this pathogen. We show here that the protein is not involved in the oxidative stress response but plays an important role in FG and cell wall biogenesis by controlling the activation of the Cek1 MAPK in cooperation with Sho1. Most importantly, it is involved in the resistance to osmotic stress by a Hog1-independent mechanism.

MATERIALS AND METHODS

Strains and growth conditions.

Strains used in this study are listed in Table 1. The prefix Ca or Sc is occasionally used to indicate the corresponding C. albicans or S. cerevisiae gene to avoid confusions in the text. Yeast strains were grown at 37°C (unless otherwise stated) in yeast extract-peptone-dextrose (YEPD) medium (1% yeast extract, 2% peptone, 2% glucose) or SD minimal medium (2% glucose, 0.67% yeast nitrogen base [YNB] without amino acids) with the appropriate nutritional requirements at 50 μg/ml (final concentration) for auxotrophs. The morphology of cells under different growth conditions was tested using YEPD medium (yeast extract with 2% dextrose), YPS medium (yeast extract with 2% sucrose) (8), YPM medium (yeast extract with 2% mannitol) (50), and synthetic low ammonium dextrose (SLAD) medium (30). For experiments under conditions of hypoxia, YPS medium was used. Drop tests were performed by spotting 5-μl drops of serial 10-fold dilutions of exponentially growing cells at an optical density at 620 nm (OD₆₂₀) of 1 onto YEPD plates supplemented with sodium chloride, sorbitol, Congo red, tunicamycin, and caspofungin at the indicated concentrations. Plates were incubated for 24 h at 37°C and scanned. To measure the inhibition of growth caused by zymolyase, cells from an exponentially growing culture were inoculated at an OD₆₂₀ of 0.025 in YEPD medium supplemented with different amounts of zymolyase 100T (dissolved in Tris-HCl [pH 7.5]-5% glucose). The assay was performed using duplicate rows of a 96-well plate, and cells were incubated overnight at 37°C. Growth values were calculated as the percentage of growth of each strain in YEPD medium supplemented with the compound compared with the growth in YEPD medium alone. Graphs represent the mean values of the results of at least three independent experiments. To assess the effect of CaCDC42 alleles on MAPK activation, the indicated strains of C. albicans were grown in 2% CAA-YNB medium (for the induction of the PCK1 promoter) as previously described (85). Actin staining with phalloidin was performed as previously described (38). Growth in liquid medium was assessed by measuring the absorbance of the cultures at OD₆₂₀. Glycerol accumulation was measured in whole-cell extracts as previously described (81).

TABLE 1.

Strains used in this study

Microorganism	Strain	Genotype	Nomenclature in manuscript and figures	Source
E. coli	DH5αF′	F′ K12Δ (lacZYA-argF)u169 supE44 thi1 recA1 endA1 hsdR17 gyrA relA1 (Øo80lacZΔM15)		33
C. albicans	SC5314			29
C. albicans	CAF-2	ura3Δimm434/URA3	wt	26
C. albicans	RM100	ura3Δ::imm434/ura3Δ::imm434his1Δ::hisG/his1ΔhisG-URA3-hisG	wt	63
C. albicans	CNC13	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG hog1::hisG/hog1::hisG-URA3-hisG	hog1	81
C. albicans	REP3	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG sho1::hisG/sho1::hisG-URA3-hisG	sho1	77
C. albicans	REP4	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG sho1::hisG/sho1::hisG		77
C. albicans	REP12	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1::hisG sho1::hisG/sho1::hisG-URA3-hisG	sho1 ssk1	77
C. albicans	CSSK21	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1::hisG-URA3-hisG	ssk1	9
C. albicans	REP14	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG-URA3-hisG MSB2/msb2Δ::FRT-FLIP-SAT1-FRT		This work
C. albicans	REP15	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG-URA3-hisG MSB2/msb2ϖFRT		This work
C. albicans	REP16	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG-URA3-hisG msb2Δ::FRT/msb2Δ::FRT-FLIP-SAT1-FRT		This work
C. albicans	REP17	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ:: hisG-URA3-hisG msb2Δ::FRT/msb2Δ::FRT	msb2	This work
C. albicans	REP18	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG msb2Δ::FRT/msb2Δ::FRT		This work
C. albicans	REP19	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG msb2Δ::FRT/msb2Δ::FRT SHO1/sho1::hisG-URA3-hisG		This work
C. albicans	REP20	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG msb2Δ::FRT/msb2Δ::FRT SHO1/sho1::hisG		This work
C. albicans	REP21	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG msb2Δ::FRT/msb2Δ::FRT sho1::hisG/sho1::hisG-URA3-hisG	msb2 sho1	This work
C. albicans	REP22	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG- msb2Δ::FRT/msb2Δ::FRT sho1::hisG/sho1::hisG		This work
C. albicans	REP23	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG-URA3-hisG MSB2/msb2Δ::FRT-FLIP-SAT1-FRT		This work
C. albicans	REP24	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG-URA3-hisG MSB2/msb2Δ::FRT		This work
C. albicans	REP25	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG-URA3-hisG msb2Δ::FRT/msb2Δ::FRT-FLIP-SAT1-FRT		This work
C. albicans	REP26	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG-URA3-hisG msb2Δ::FRT/msb2Δ::FRT	ssk1 msb2	This work
C. albicans	REP26-u	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG msb2Δ::FRT/msb2Δ::FRT		This work
C. albicans	REP27	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG msb2Δ::FRT/msb2Δ::FRT SHO1/sho1::hisG-URA3-hisG		This work
C. albicans	REP28	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG msb2Δ::FRT/msb2Δ::FRT SHO1/sho1::hisG		This work
C. alαbicans	REP29	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG msb2Δ::FRT/msb2Δ::FRT sho1::hisG/sho1::hisG-URA3-hisG	ssk1 msb2 sho1	This work
C. albicans	REP30	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG msb2Δ::FRT/msb2Δ::FRT sho1::hisG/sho1::hisG		This work
C. albicans	CASU84	ura3/ura3 CaCDC42/cacdc42::hisG PCK1-CaCDC42::hisG-URA3-hisG		89
C. albicans	CASU64	ura3/ura3 CaCDC42/cacdc42::hisG PCK1-CaCDC42G12V::hisG-URA3-hisG		89
C. albicans	CASU69	ura3/ura3 CaCDC42/cacdc42::hisG PCK1-CaCDC42 D118A::hisG-URA3-hisG		89
C. albicans	CDH25	ura3Δ::imm434/ura3Δ::imm434 cst20::hisG/cst20::hisG	cst20	46
C. albicans	CLJ5	ura3Δ::imm434/ura3Δ::imm434 cla4::hisG/cla4::hisG	cla4	47
C. albicans	CDH12	ura3Δ::imm434/ura3Δ::imm434 hst7::hisG/hst7::hisG	hst7	46
C. albicans	REPc-1	ura3Δ::imm434/ura3Δ::imm434 cst20::hisG/cst20::hisG PCK1-CaCDC42G12V::hisG-URA3-hisG		This work
C. albicans	REPc-2	ura3Δ::imm434/ura3Δ::imm434 cla4::hisG/cla4::hisG PCK1-CaCDC42G12V::hisG-URA3-hisG		This work
C. albicans	REPc-3	ura3Δ::imm434/ura3Δ::imm434 hst7::hisG/hst7::hisG PCK1-CaCDC42G12V::hisG-URA3-hisG		This work
C. albicans	REPc-4	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG sho1::hisG/sho1::hisG PCK1-CaCDC42G12V::hisG-URA3-hisG		This work
C. albicans	REPc-5	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG msb2Δ::FRT/msb2Δ::FRT PCK1-CaCDC42G12V::hisG-URA3-hisG		This work
C. albicans	REPc-6	ura3Δ::imm434/ura3Δ::imm434 his1Δ::hisG/his1Δ::hisG msb2Δ::FRT/msb2Δ::FRT sho1::hisG/sho1::hisG PCK1-CaCDC42G12V::hisG-URA3-hisG		This work
C. albicans	REP30-HG1	ura3Δ::imm434/ura3Δ::imm434 ssk1::hisG/ssk1:: hisG msb2Δ::FRT/msb2Δ::FRT sho1::hisG/sho1::hisG LEU2/leu2::ACT1PR-HOG1-GFP-URA3		This work

Molecular biology procedures and plasmid constructions.

Standard molecular biology procedures were used for all genetic constructs (6). For the disruption of the MSB2 gene, the primers MSB2UP1 (GGTACCTTTCTTTGTTTGTGGAGTGG) and MSB2LP2 (CTCGAGAAAAAGGAATTGTTTAGTTGG) were used to amplify a 1.75-kbp 5′ region flanking the open reading frame (ORF) and subcloned in pGEM-T (underlined letters indicate the restriction sites introduced for cloning purposes). Similarly, oligonucleotides MSB2UP3 (GCGGCCGCTTGACTGGTGATCCTAATGG) and MSB2LP4 (GAGCTCTTAGCAAGGATTGAAAAAGG) were used to amplify a 1.79-kbp 3′ flanking region of the ORF from C. albicans strain SC5314 and subcloned in pGEM-T. The 5′ and 3′ regions were excised from these constructions by using a combination of enzymes KpnI and XhoI and enzymes NotI and SacI, respectively, and were accommodated in the disruption plasmid pSF2A (75), which comprises the SAT1 marker that confers resistance to nourseothricin, by a four-fragment ligation to generate pDMSB2. DNA was digested with KpnI and SacI to generate a region of DNA used to force recombination at the MSB2 locus following the SAT1 flipping scheme (75). Genomics DNAs were digested with ClaI or BglII and probed with the 1.75-kb 5′ region of the gene for the Southern hybridization. This strategy was used for MSB2 deletion in the wild-type (wt) (RM100) or ssk1 (strain CSSK21) background. Next, SHO1 was disrupted in the msb2 and ssk1 msb2 backgrounds following the URA-blaster scheme (26) and using the constructions already described (77). The strains used in this study were all Ura positive (see Table 1). For the analysis of CaCDC42-mutated alleles we used pSU48 and pSU50 plasmids containing CaCDC42^G12V and CaCDC42^D118A under the control of the PCK1 promoter (89). Each plasmid was made linear using a unique HpaI restriction site and transformed into Ura-negative strains; Ura-positive transformants were selected, and proper integration (at the PCK1 locus) was confirmed by Southern blot analysis before using them. For Hog1 translocation experiments, an ACT1^PR-HOG1-green fluorescent protein (ACT1^PR-HOG1-GFP) fusion was integrated in the LEU2 locus after digestion with KpnI (4). MSB2 reintegration was performed as follows. The plasmid pDS1044-1-MSB2-HA (E. Szafranski and J. F. Ernst, unpublished data) was digested with EcoRV, and the MSB2-hemagglutinin (MSB2-HA) fusion was integrated in the LEU2 locus. Ura-positive transformants were selected and checked.

Protein extracts and immunoblot assays.

Cells were grown and samples were processed as previously described (77). The procedures employed for cell collection, lysis, protein extraction, fractionation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and transfer to nitrocellulose membranes have been previously described (54). Anti-phospho-p44/p42 MAPK (Thr²⁰²/Tyr²⁰⁴) antibody (New England Biolabs) was used to detect dually phosphorylated Mkc1 and Cek1 MAPKs; phospho-p38 MAPK (Thr¹⁸⁰/Tyr¹⁸²) 28B10 monoclonal antibody (Cell Signaling Technology, Inc.) and ScHog1 polyclonal antibody (Santa Cruz Biotechnology) were used to detect the phosphorylated Hog1 and Hog1 protein, respectively. Mkc1 and Cek1 proteins were detected by using polyclonal antibodies against them (61, 77). Western blots were developed according to the manufacturer's specifications using a Hybond ECL kit (Amersham Pharmacia Biotech). To make equal the amounts of protein loaded, samples were first normalized by measuring the absorbance at 280 nm and then by pre-sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Coomasie staining.

Animal experiments.

Virulence assays were performed essentially as described previously (20). Briefly, C. albicans cells from fresh YEPD plates of strains to be tested were collected by low-speed centrifugation and washed twice with phosphate-buffered saline (PBS). A total of 10⁶ yeast cells (in 250 μl) were inoculated into the lateral tail vein of BALB/c mice, and the mortality results were followed during 30 days. Postmortem analyses were carried with three to five animals, and clearance of the infection was assessed by counting CFU on Saboureaud-chloramphenicol solid medium as described previously (20).

Quantitative real-time PCR (qRT-PCR).

Total RNA was isolated and purified from cells by mechanical disruption using an RNeasy Mini kit (Qiagen, Hilden, Germany). RNA (2 μg) was reverse transcribed into cDNA (Promega), and 5 μl of the 1:100 dilutions was used for PCR assays, together with 10 μl of SYBR green mix (Applied Biosystems) and 1.2 μl of each of the forward and reverse oligonucleotide primers. The primers selected were CEK1upRT (TTAGAAATTGTTGGAGAAGGAGCAT), CEK1lowRT (GCAACTTTTTGTTGTGATGGTTTATG), ACT1upRT (TGGTGGTTCTATCTTGGCTTCA), and ACT1lowRT (ATCCACATTTGTTGGAAAGTAGA). Transcript levels of a gene in a single strain were calculated relative to ACT1 transcript levels by considering the amplification efficiencies of the primers (e values) and the cycle numbers at a fixed threshold of fluorescence intensity in the logarithmic phase of amplification (C_T values). For each mRNA, mean C_T values for reference and target transcripts were determined from three amplification reactions.

Fluorescence microscopy.

Yeast strains were grown at 37°C in SD medium to an OD of 0.8. In the case of treated cells, NaCl was added to the concentration specified in the figure legend and the mixture was incubated for 5 min or as indicated. Samples were centrifuged and washed twice with PBS. Cells were fixed with ice-cold 70% ethanol for 1 min, centrifuged, and washed twice with PBS. DAPI (4′,6′diamidino-2-phenylindole) was added to achieve a final concentration of 2 mg/ml to stain the nucleus. Viable cells were directly observed under the microscope and challenged with osmotic shock. Fluorescence microscopy was carried out with a Nikon Eclipse TE2000-U microscope at ×100 magnification. Images were captured using a Hamamatsu ORCA-ER charge-coupled-device camera and AquaCosmos 1.3 software. All images were processed identically using Adobe Photoshop 7.0.

RESULTS

msb2 mutants have cell wall defects.

To analyze the role of the MSB2 gene in C. albicans, we used the SAT1 flipper strategy that makes use of the dominant nourseothricin SAT1 marker (56). 5′ and 3′ fragments adjacent to the MSB2 ORF (from the 19th Candida assembly) were amplified by PCR and introduced in the pSF2A vector, thus obtaining the pDMSB2 deletion plasmid (see Materials and Methods). We also deleted the gene in sho1, ssk1, and sho1 ssk1 genetic backgrounds to analyze the relation of MSB2 to these other elements of the HOG pathway in C. albicans (see list of strains in Table 1).

Given the relationship between Sho1 and Msb2 in S. cerevisiae (16, 90) and the role of Sho1 in the cell wall construction in C. albicans (77), we tested the phenotype of msb2 and double msb2 sho1 mutants treated with compounds that interfere with the cell wall. We used both Congo red (a compound that interacts with chitin and interferes with cell wall construction) and caspofungin [an inhibitor of the β-(1,3)-glucan synthase]. As shown in Fig. 1A, MSB2 deletion resulted in Congo red sensitivity (250 μg/ml), a phenotype which is similar to that observed for sho1 mutants. Deletion of MSB2 in a sho1 background did not result in enhanced sensitivity, suggesting that the two proteins share overlapping roles with respect to this particular trait. Under these conditions, however, hog1 cells were found to be resistant to the drug, in accordance with what has already been described (1) (Fig. 1A). A lower drug concentration (125 μg/ml) gave overall similar results (data not shown). We also observed clear behavior by the use of the β-glucan synthesis inhibitor caspofungin (40, 88): msb2 cells showed a clearly augmented sensitivity to caspofungin, a phenotype that was also observed with sho1 cells but that increased slightly with the double msb2 sho1 mutant (Fig. 1B). This result suggested functional differences in the glucan moiety of the cell wall in msb2 mutants. This led us to test the susceptibility of cells in liquid culture to zymolyase, a glucanase-enriched enzymatic cocktail. Cells were allowed to grow in media supplemented with zymolyase, and we measured the final OD of the cell suspension as indicative of growth. As shown in Fig. 1C, the sho1 and msb2 mutants and the double msb2 sho1 mutant displayed higher susceptibility to zymolyase, as evidenced using a dose of 160 U/ml. Interestingly, the sho1 mutant displayed a more sensitive phenotype than msb2 cells, as evidenced at 80 U/ml, under which conditions the reduction of growth of a sho1 mutant was more than 50% whereas that of an msb2 single mutant was less than 20% (compare 45% to 85% for sho1 and msb2 mutants, respectively). This effect was prevented by using an osmotic stabilizer in the assay (1 M sorbitol) (data not shown), although the final OD reached was lower for all strains. These results indicate that the defect in the MSB2 and/or SHO1 gene led to measurable defects in cell wall formation and suggest that the genes may functionally overlap.

FIG. 1.

Msb2 is involved in cell wall biogenesis. (A and B) Sensitivity of the indicated strains to Congo red (CR) (250 μg/ml) (A) and caspofungin (20 and 30 ng/ml) (B) after 24 h at 37°C. Samples of 10-fold dilutions from exponentially growing cells were spotted on YEPD plates supplemented with the indicated drug concentrations and incubated at 37°C for 24 h before being scanned. (C) Susceptibility to zymolyase 100T. Values represent percentages of growth in YEPD medium supplemented with zymolyase compared to YEPD medium alone. Mean values are given, with bars indicating standard deviations.

Msb2 controls Cek1 phosphorylation.

As Sho1 controls the activation of Cek1 by phosphorylation in C. albicans (77), we tested whether Msb2 is also involved in Cek1 phosphorylation. We tested this assumption by analyzing the pattern of MAPK activation after the addition of either Congo red or caspofungin to exponentially growing cultures of our strains. The concentration of these compounds was chosen to minimize their deleterious effects on growth while it was still maintained at a level over the threshold needed to generate a functional response (data not shown). As shown in Fig. 2A, Congo red (60 μg/ml) induced clear Cek1 phosphorylation which was effectively blocked in sho1, msb2, and msb2 sho1 mutants. Caspofungin (20 ng/ml) also induced Cek1 phosphorylation, and this process was Msb2 and Sho1 dependent, although less intense (3× to 4× less, as determined by densitometry) than that observed with Congo red. Caspofungin also induced Mkc1 phosphorylation, a result consistent with the role of this MAPK in the cell integrity pathway (61, 62). However, phosphorylation of Mkc1 was completely independent of the presence of Msb2 or Sho1 (Fig. 2A), indicating the specificity of the effect of the presence of Msb2 and Sho1 proteins in Cek1 activation. We also tested an additional experimental condition that results in Cek1 phosphorylation at the restart of growth from stationary phase (77); when stationary phase cells are diluted in fresh medium, Cek1 becomes quickly phosphorylated, being detected as early as 5 min. Stationary-phase cells showed no detectable phosphorylated Cek1 despite the presence of Cek1 protein (Fig. 2B) in all strains. After 1 h of growth, Cek1 became quickly phosphorylated in wt cells; Cek1 phosphorylation was, however, severely reduced in msb2 cells and completely blocked in sho1 cells (and also, as expected, in the double msb2 sho1 mutant). These results indicate that Msb2 (as well as Sho1) controls Cek1 activation in response to defined experimental conditions that affect the cell wall.

FIG. 2.

Msb2 controls the activation of the Cek1 MAPK. (A) Effect of Congo red (CR) and caspofungin at the indicated concentrations on the pattern of Cek1 and Mkc1 MAPK activation (phospho-Cek1 [P-Cek1] and phospho-Mkc1 [P-Mkc1] in the figure) in exponentially growing cells (OD₆₂₀ = 1) after 2 h of incubation. Anti-ScHog1 and anti-rabbit-Cek1 antibodies were used for loading protein controls. (B) Cek1 activation (P-Cek1 in the figure) during the resumption of growth from the stationary phase. Cells were diluted in fresh YEPD medium at OD₆₂₀ = 0.2 and grown at 37°C for 1 h. Samples were collected and processed for Western blot analysis.

Inhibition of N-glycosylation activates Cek1 in an Msb2-dependent manner.

Defects in mannose utilization and/or altered protein glycosylation induce the activation of the pheromone response pathway in S. cerevisiae (measured by the increased FUS1-lacZ transcription). This process requires members of the SVG (sterile vegetative growth) pathways and, specifically, the Sho1 protein (17). In C. albicans, both cell wall alterations associated with mannosylation defects (pmt mutants [70]) and tunicamycin-induced altered glycosylation induce the phosphorylation of Cek1 (10). We determined whether Sho1 and Msb2 could play a role in this process by measuring the pattern of MAPK activation after treatment with tunicamycin. We tested the effect of this drug in experiments in which mid exponential phase cells were grown at 37°C; tunicamycin 5 μg/ml was then added and cells were collected after 2 h of additional growth. As shown in Fig. 3A, tunicamycin induced both phosphorylation and synthesis of Cek1 in wt cells (three- to fivefold, as determined by densitometry). Phosphorylation—but not synthesis—was almost completely dependent on Sho1 and was completely dependent on Msb2 (no activation was observed in the double mutant; not shown). Activation of Cek1 was recovered in a reconstituted strain in which an Msb2-HA construct was integrated at the LEU2 locus (not shown). This effect was specific to the Cek1 MAPK, whereas Hog1 was not activated during a similar period (not shown) and Hog1 levels remained constant. The increase in Cek1 protein levels was shown to be specific to the presence of tunicamycin and independent of that of Sho1/Msb2, and the addition of zymolyase (a β-glucan-enriched enzymatic preparation that acts on the cell wall) did not have any effect on Cek1 synthesis but did have an effect on its activation in a Sho1/Msb2-dependent manner.

FIG. 3.

Tunicamycin activates Cek1 in a MSB2-dependent manner. (A) The effect of tunicamycin (5 μg/ml) or zymolyase 100T (2 U/ml) on Cek1 and phospho-Cek1 (P-Cek1) activation was tested on exponentially growing cells (OD₆₂₀ = 1) of the indicated strains after 2 h of treatment at 37°C. (B) CEK1 transcript levels were measured by quantitative PCR. Values represent ratios to untreated CAF-2 levels (−) for each strain after incubation with (+) or without (−) 5 μg of tumicamycin/ml for 2 h in exponentially growing cells (OD₆₂₀ = 1). Mean values are given, with bars indicating standard deviation.

As the final targets of MAPK cascades are transcription factors, we wondered whether this pathway could be subjected to positive transcriptional regulation and whether this would be dependent on P-Cek1. We therefore measured CEK1 mRNA levels by qRT-PCR analysis (see Materials and Methods) under these experimental conditions. As shown in Fig. 3B, CEK1 mRNA levels increased in a 2.5- to 3.5-fold range in response to the drug (in RM100 and CAF2 wt strains). This increase was independent of Cek1 phosphorylation, as it was also found to occur with msb2 and sho1 mutants (2× and 2.5×, respectively), although basal levels were found to be slightly lower in these strains. Collectively, these experiments indicate that the Cek1-mediated pathway responds to tunicamycin-mediated glycosylation inhibition by an increase in Cek1 synthesis and phosphorylation, this latter effect being dependent on the presence of Msb2 and Sho1.

MSB2 deletion prevents Ssk1-mediated Cek1 hyperactivation.

Previous results from our group have shown that ssk1 (and hog1) mutants display enhanced basal activation of Cek1 and that this correlates with their resistance to certain cell wall inhibitors (1, 77). We tried to determine the role of Msb2 in this cross-talk mechanism. For this reason, we deleted the MSB2 and SHO1 genes in an ssk1 mutant and performed experiments similar to those described above. While ssk1 cells were resistant to Congo red, deletion of SHO1 in ssk1 cells resulted in sensitivity to this compound, indicating that SHO1 deletion is dominant over SSK1 in this assay (Fig. 4A). MSB2 deletion had also an effect, although less drastic, on Congo red sensitivity in an ssk1 mutant, since its deletion in this background partially abolished the ssk1 resistance to this compound but was found to have blocked the enhanced ability of the ssk1 cells to grow on caspofungin plates (a result that was similar to what was observed with wt cells; Fig. 1B). The most drastic sensitivity to caspofungin was observed with a cek1 mutant (Fig. 4B). These results reveal important functional differences between Sho1 and Msb2 in terms of signaling that could be explained by their differential effects on Cek1 activation. This hypothesis was tested by analyzing the MAPK pattern in the resumption of growth from the stationary phase. As shown in Fig. 4C, Cek1 was clearly activated in wt cells at 1 h after dilution of the culture. However, sho1 mutants (single and msb2 background) completely blocked the activation of Cek1, which was undetectable at that time (Fig. 4C) and for an additional 2 to 3 h (not shown). The behavior of an ssk1 mutant was consistent with the presence of a derepressed Cek1 signaling phenotype, as (i) phosphorylated Cek1 was still detectable in overnight cultures and (ii) both the maximum levels and the lengths of the activation period were increased compared to wt cell results (Fig. 4D). Deletion of MSB2 in a ssk1 background only partially blocked Cek1 activation, which was still detectable at 1 and 2 h after growth. Collectively, these results demonstrate that the Ssk1-mediated Cek1 hyperactivation was dependent mainly on Sho1 and only partially on Msb2.

FIG. 4.

MSB2 deletion suppresses Cek1 hyperactivation in ssk1 mutants. (A and B) Sensitivity of the indicated strains on plates supplemented with Congo red (CR) (250 μg/ml) (A) and caspofungin (20 or 50 ng/ml) (B). Plates were incubated at 37°C for 24 h before being scanned. (C and D) Western blots showing Cek1 phosphorylation (phospho-Cek1 [P-Cek1]) and Cek1 protein during the resumption of growth from the stationary phase (st) and 1 or 2 h after dilution in fresh YEPD medium at OD₆₂₀ = 0.2. Samples were processed as described in Materials and Methods. The film was intentionally overexposed to produce the image presented in panel D.

Overexpression of Cdc42p^G12V hyperactivates Cek1 in a Sho1/Msb2/Hst7-dependent manner.

Cdc42 is an essential GTPase of the RAS superfamily (55). Overexpression of mutant alleles of this enzyme in C. albicans results in multinucleate cells that are either unbudded or multibudded, depending on the enzymatic activity of the protein (89). However, no direct evidence has been obtained about the role of this protein in the activation of the SVG pathway in C. albicans despite several reports that indicate its involvement in polarized and pseudohyphal growth in S. cerevisiae. We analyzed the activation of MAPKs during the resumption of growth from the stationary phase in point mutants in which CaCDC42 expression was regulated by the PCK1 promoter (89). These included the G12V mutant, which has an intrinsic decreased GTPase activity that results in a protein locked in a GTP-bound state (hyperactive), and the D118A mutant, which cannot exchange GDP with GTP, resulting in a protein locked in the GDP-bound (inactive) state (dominant negative). To assess the effect of the ectopic expression of the PCK1^PR-CaCDC42 wt and mutants, cells carrying these constructions (see Table 1) were grown for 24 h either in liquid YEPD (repressing conditions) or in YNB-2% CAA (inducing conditions) at 37°C, diluted to an OD₆₂₀ of 0.1 in the same media, and allowed to grow for 3 h. Cells were then collected and analyzed for MAPK activation (see Materials and Methods). Under these conditions, the wt and D118A mutant showed a mainly filamentous morphology, while the G12V mutant displayed several aberrant structures as previously described (89) (data not shown).

As shown in Fig. 5A, Cek1 was hyperphosphorylated compared to wt cells when the GTP-bound locked Cdc42 protein was overexpressed (the G12V mutant in YNB-2% CAA liquid medium). In contrast, a signal that was clearly reduced compared to that seen with the wt was observed when the inactive form of Cdc42 (GDP) was overexpressed (D118 mutant allele). No significant differences with respect to the amount of Cek1 were found in these extracts. Such effects were mainly detected under overexpressing conditions (YNB-2% CAA medium), although in repressing media (YEPD), a faint signal could still be detected for Cek1 in the G12V allele (data not shown). Resting stationary-phase cells switched off Cek1 phosphorylation (as previously shown for wt cells [77]); interestingly, however, this happened even with the G12V mutant, suggesting that dephosphorylation mechanisms are dominant over Cdc42 activation. These results indicate that Cdc42 is a mediator of Cek1 phosphorylation.

FIG. 5.

Effect of ectopic expression of hyperactive (G12V) or dominant-negative (D118A) CaCDC42 alleles on the Cek1-mediated pathway. (A) Strains CASU64 (ectopic expression of CaCdc42p^G12V in the heterozygous disruptant strain CaDH85), CASU69 (ectopic expression of CaCdc42p^D118A in the heterozygous disruptant strain CaDH85), and CASU84 (CDC42^wt) were grown overnight either in YEPD medium or in YNB-2% CAA (a medium that induces the PCK1 promoter). Cells were diluted at OD₆₂₀ = 0.1 in the same media and allowed to grow 3 h before collection and processing for phospho-Cek1 (P-Cek1) and Cek1 detection. (B) Ectopic expression of CaCdc42p^G12V in wt and sho1, msb2, msb2 sho1, cla4, hst7, and cst20 mutant cells was induced by growing the cells in YNB-2% CAA overnight at 37°C. Cells were diluted in fresh medium at an OD₆₂₀ of 0.2 and allowed to grow 3 h at 37°C before being processed for Western blotting. (C) Cells (stationary phase) of the msb2 sho1 mutant with the PCK1-CaCDC42^G12V::hisG-URA3-hisG construction integrated were washed with PBS and diluted into 2% glucose (for PCK1 promoter repression) or YNB-2% CAA medium (for PCK1 promoter expression). Cells were photographed after 6 h of incubation at 37°C.

In order to determine which elements are involved in signaling to Cek1, we integrated the PCK1^PR-CDC42^G12V construction in different mutant strains (sho1, msb2, cla4, hst7, cst20, and msb2 sho1) and checked the activation of the pathway under similar sets of inducing conditions (Fig. 5B). When cells were transfer to YNB-2% CAA medium, no phosphorylation of Cek1 could be detected when Sho1, Msb2, or Hst7 was absent. However, even 15 min after the dilution into fresh medium, Cek1 was detected in a phosphorylated state in cla4 and cst20 mutants, suggesting either that these proteins do not participate in signaling or that they are redundant in this process. Interestingly, under these conditions, activation of the cell wall integrity pathway MAPK Mkc1 occurred when Sho1 and/or Msb2—but not the Hst7 MAPK kinase—was absent, suggesting that the lack of these upstream mediators of Cek1 results in cross-talk toward the cell integrity pathway, maybe as a rescue mechanism. Morphological alterations associated with CDC42^G12V expression were still observed for msb2 sho1 mutants under these conditions (Fig. 5C), indicating that Cek1 activation is not the main kinase responsible of Cdc42-mediated morphological effects.

Collectively, these results indicate that Cdc42 may be an upstream mediator of Cek1 activation in C. albicans and implicate Sho1, Msb2, and Hst7 proteins in this process.

Msb2 contributes to the osmosensitivity of ssk1 sho1 mutants by an Hog1-independent mechanism.

Previous work in our laboratory suggested the existence of a third input component of the HOG pathway in C. albicans, as double sho1 ssk1 mutants, which are impaired in the two upstream branches of the pathway in S. cerevisiae, are not as osmosensitive as the hog1 mutant (77). This behavior clearly contrasts with what occurs in S. cerevisiae, for which a double deletion mutant is osmosensitive (64, 66). We addressed the role of MSB2 in resistance to osmotic stress by spotting exponentially growing cells onto YEPD plates supplemented with different osmolytes such as sodium chloride and sorbitol at different concentrations. As shown in Fig. 6A, deletion of MSB2 did not render cells osmosensitive and growth of msb2 mutants was quite similar to that of wt cells (RM100 strain) on 1 M NaCl solid YEPD plates. In contrast, the hog1 mutant failed to grow under similar conditions, which is consistent with previously described results (81). The situation was essentially similar for a different osmolyte (sorbitol), and no effects were observed for msb2 mutant cells when grown under conditions of 1.5 or 2 M (not shown) sorbitol plates (Fig. 6A).

FIG. 6.

Osmotic response in msb2 mutants. (A) Exponentially growing cells of the strains indicated were spotted onto YEPD plates supplemented with 1.5 M sorbitol or 1 M sodium chloride (NaCl). (B) Growth under osmotic stress conditions (1.5 and 2 M NaCl) in liquid media after 24 h at 37°C. Mean values of final ODs (y axis) are given, with bars indicating the standard deviations. (C) Phalloidin staining of an ssk1 msb2 sho1 mutant after 24 h of growth in YEPD medium (−) or 2 M NaCl-YEPD medium (+). (D) Hog1 MAPK phosphorylation in the presence of increasing amounts of NaCl in exponentially growing cells (OD₆₂₀ = 1) of the indicated mutants after 10 min of incubation. (E) Internal glycerol (Glyc) quantification of the results of three different experiments after treatment with 1 M NaCl of exponentially growing cells at the times indicated. (F) An ssk1 msb2 sho1 strain with ACT1^PR-HOG1-GFP integrated in the LEU2 locus was grown in SD medium and exposed to 1 M NaCl for 5 min. Cells were fixed, stained with DAPI, and visualized with a fluorescence microscope.

Deletion of MSB2 had no effect on the osmosensitivity of ssk1 or sho1 cells as well, and double mutant ssk1 msb2 or msb2 sho1 cells showed behavior similar to that seen with single sho1 or ssk1 mutants on solid media. However, the triple ssk1 msb2 sho1 mutant displayed significant osmosensitivity that was close to that shown by hog1 cells. These effects were also reproduced in liquid medium. Cells were grown in 1.5 or 2 M NaCl YEPD medium, and the final OD reached in the stationary phase was taken as an estimation of their ability to survive or resist osmotic stress under these conditions. As shown in Fig. 6B, sho1, msb2, and msb2 sho1 mutants grew similarly to wt cells in 1.5 M NaCl (OD₆₂₀ = 8 to 9). The msb2 mutation also had no effect on the ssk1 background, although the growth seen with ssk1 cells was lower (OD₆₂₀ = 5.8 to 6). The SHO1 mutation increased the osmosensitivity of ssk1 cells, and deletion of MSB2 in this background aggravated this phenotype (OD₆₂₀ = 2.32 for sho1 ssk1; OD₆₂₀ = 1.85 for ssk1 msb2 sho1). Finally, the hog1 mutant showed the most severe phenotype (OD₆₂₀ = 1). Results were qualitatively similar in experiments using 2 M NaCl. A reconstituted ssk1 msb2 sho1 strain in which an Msb2-HA construct was integrated in the LEU2 locus showed, as expected, a sho1 ssk1 phenotype and was able to grow in high osmolarity medium (data not shown). These results suggest a role for Msb2 growing under conditions of high osmolarity that is only evident when both (Ssk1-dependent and Sho1-dependent) putative branches of the HOG pathway are impaired. Under these restrictive conditions, the cell morphology was significantly altered. The morphology of hog1 cells has been described before (1) and consists of frequent chained cells that fail to separate. wt cells and sho1 or msb2 mutants gave rise to slightly elongated cells under conditions of high osmolarity (2 M) (not shown). In contrast, the triple ssk1 msb2 sho1 mutant generated rounded, huge, and also frequently multibudded cells (Fig. 6C). These effects were also observed with some cells of the sho1 ssk1 mutant but were aggravated in ssk1 msb2 sho1 cells. Under these conditions, the cells had a severe cell polarity defect, as revealed by actin staining, and cytoskeleton polarization was clearly lost (Fig. 6C). These results suggest the crucial role that the Sho1 and Msb2 protein complexes play in maintaining the polarization of the actin cytoskeleton under conditions of hyperosmotic stress.

To further characterize this phenotype, we checked the state of Hog1 activation under conditions of an osmotic stress challenge in these mutants during exponential growth. Cells were grown to an OD₆₂₀ of 1 and subjected to 0.8, 1, 1.5, or 2 M NaCl for 10 min; cells were then collected and processed to quantify Hog1 activation. Osmotic stress induced Hog1 activation in all strains tested and to similar extents. Surprisingly, this was also found to be the case with the triple ssk1 msb2 sho1 mutant (Fig. 6D), indicating that deletion of these three putative upstream components of the pathway does not block the transmission of the signal. In fact, this experiment showed that the solute concentration threshold needed to activate the pathway was even slightly lower for the triple mutant, as 1 M NaCl activated the pathway in this mutant but not so strongly in the rest of strains (Fig. 6D). We confirmed this result by measuring the intracellular glycerol concentration following an osmotic shock in a time course assay. Basal glycerol levels were close to 0.02 μM/mg (dry weight) for wt, hog1, or ssk1 msb2 sho1 cells. However, accumulation was evident at 4 and 6 h after the challenge for the wt and the triple mutant (reaching values of 0.16 to 0.18 μM/mg) but almost absent for the hog1 mutant (Fig. 6E). In addition, translocation of an Hog1-GFP fusion to the nucleus was found to occur upon a shift to high osmolarity medium (Fig. 6F).

These experiments indicated that Hog1 activation is not sufficient to sustain growth under conditions of high osmolarity when the Msb2/Sho1/Ssk1 proteins are absent and collectively implicate the Msb2/Sho1 complex in sustaining growth under conditions of high osmolarity by a mechanism that is, apparently, not dependent on Hog1 phosphorylation.

Msb2 is involved in invasion on solid media.

We tested the effect of the MSB2 mutation on morphogenesis by the use of different solid media. On SLAD (nitrogen starvation) medium (30), mutants showed no obvious differences (not shown). When assayed on YPS (sucrose carbon source) medium, we found msb2 cells to be less filamentous, as evidenced by the morphology of the colony border (Fig. 7A). This was not observed with the sho1 mutant but was evident with the msb2 sho1 mutants, indicating the dominance of the MSB2 deletion. On YPM (mannitol carbon source) medium (50), msb2 mutants showed a clear difference compared to wt cells; whereas wt cells generated filamentous colonies, the absence of Msb2 partially suppressed this phenotype (Fig. 7B). This was also observed with the sho1 mutant and the double msb2 sho1 mutant. In addition, deletion of MSB2 and/or SHO1 also suppressed such effects in an ssk1 background. Thus, the Sho1 and Msb2 proteins partially block filamentous growth on both media. We also tested the effects of MSB2/SHO1 mutations under conditions of hypoxia, as it has been recently shown that these conditions require a different signaling pathway than aerobic-normoxic conditions (21, 84). This situation probably better mimics the environment found during host infection, and it was therefore of interest to determine the role of Msb2 under these conditions. During hypoxia in an atmosphere containing 0.2% oxygen and 6% CO₂, colonies of msb2 and sho1 mutants in different backgrounds contained fewer hyphi compared to wt strain colonies. In contrast, the ssk1 mutant was hyperfilamentous under these conditions (Fig. 7C). Deletion of SHO1 or MSB2 or both in the ssk1 background suppressed the invasion, as assayed by the standard plate washing procedure. Results are shown for 30°C, but essentially the same results were observed at 37°C (not shown). Although filamentation and invasion are different morphogenetic programs, we tested whether msb2 mutants would be defective in filamentation. We assayed the behavior of the msb2 mutant at 37°C in YEPD medium supplemented with 1%, 5% (not shown), or 10% serum. Cells were able to form normal filaments under these conditions, and no differences were observed regarding the morphology and length of filaments, indicating that Msb2 is not required for the dimorphic transition (Fig. 7D). In addition, deletion of MSB2 did not alter the filamentation pattern of sho1, ssk1, or ssk1 sho1 mutants (data not shown). Collectively, these results indicate that Msb2 is an important mediator of invasion on certain solid media.

FIG. 7.

Effect of Msb2 on morphological transitions. Stationary-phase cultures of the indicated strains were collected from exponentially growing cells, washed with PBS, and counted. (A and B) A total of 50 CFU was spread on YPS (sucrose) (A) or YPM (mannitol) (B) plates and incubated at 30°C for 7 to 12 days before the colonies were photographed. (C) Cells were grown in YPS medium under conditions of hypoxia (6% CO₂, 0.2% O₂). On the indicated days after plating (D+2 and D+2), photographs of the border colony morphology were taken. Colonies were also eventually washed and photographed. (D) A total of 10⁵ cells of the indicated strains/ml were inoculated in liquid YEPD supplemented with 1 and 10% serum or left unsupplemented and were incubated at 37% for 3 h.

Msb2 is not involved in resistance to oxidative stress or virulence.

Given the involvement of the HOG pathway in C. albicans in oxidative stress resistance and virulence (2, 5), we checked whether Msb2 could play a significant role in this process. However, the absence of MSB2—either alone (wt background) or in combination with other mutations (ssk1 and sho1)—did not significantly alter the resistance of these strains to hydrogen peroxide or menadione. Activation of Hog1 in msb2 mutants in response to hydrogen peroxide treatment was detectable and similar to that seen with wt cells, whereas such activation was absent in ssk1 backgrounds, confirming previous results that indicate that Ssk1 is the main upstream element involved in the transmission of the oxidative stress signal (11) (not shown). We also tested the virulence of mutants in the upstream components of the HOG pathways in a mouse systemic infection model (Fig. 8). msb2 mutants were found to be as virulent as wt cells (at least according to the mean survival time parameter). In contrast, sho1 mutants displayed reduced virulence (40 to 60% survival after 21 days). These differences were not the result of altered expression of the URA3 marker, as determined by qRT-PCR (data not shown).

FIG. 8.

Role of Msb2 in Candida albicans virulence. Survival curves of BALB/c mice infected systemically with 10⁶ cells of the indicated strains of C. albicans.

DISCUSSION

In S. cerevisiae, adaptation to high osmolarity depends on the activation of the Hog1 MAPK (34, 66, 71). In hog1 mutants, osmotic stress activates the mating pheromone response by a process that is partially dependent on the Sho1 membrane protein. MSB2 was identified as a suppressor of this cross-talk activity (determined using a FUS1-LACZ gene reporter [65]) and was shown later to encode a signaling mucin that is situated upstream of Sho1 and physically interacts with it (16).

We have characterized the role that Msb2 plays in the fungal pathogen C. albicans by the analysis of mutants deleted from this gene either alone or in combination with other signaling elements. We present evidence that Msb2 participates in the biosynthesis of the cell wall and in the invasion of solid surfaces and propose on the basis of two main lines of experimental data that this occurs through the activation of the Cek1 MAPK. First, deletion of MSB2 results in sensitivity to two compounds that affect the construction of the fungal cell wall, namely, Congo red and caspofungin, in similarity to what has been observed in studies of cek1 mutants (23, 77). Congo red is known to bind chitin as well as to inhibit β-(1,3)-glucan synthesis in vitro (80), and resistance or sensitivity to this compound has been associated with the activation or deactivation of the Cek1 pathway (77). msb2, sho1, msb2 sho1, and cek1 mutants are also sensitive to caspofungin, a compound that inhibits glucan synthesis (41), and to zymoliase, a glucanase-enriched preparation, suggesting a relation of the Cek1 pathway to the glucan moiety of the cell wall. Second, deletion of MSB2 blocks Cek1 activation under conditions (Congo red and caspofungin challenge) that involve substantial cell wall stress. As is consistent with this, ssk1 and hog1 mutants, which hyperactivate Cek1 (1, 77), are more resistant to these compounds, and this effect is mostly suppressed by MSB2 and/or SHO1 deletions. This role of Msb2 in Cek1 activation supports previous data from our group showing that the Hog1 and Cek1 MAPKs play complementary roles in cell wall biogenesis (23) and is in close agreement with recent data in S. cerevisiae, in which the FG and HOG pathways mutually regulate each other (96). While the final effects of Msb2/Sho1 deletion on the cell wall are evident, the triggering molecular mechanisms are presently unknown. In S. cerevisiae, however, cleavage of the Msb2 protein by the Yps1 aspartyl protease is required for signaling (90), as this step releases an inhibitory domain present in the heavily glycosylated ScMsb2 extracellular domain (18). Also, defects in glycosylation (pmi40-101 mutants) activate the invasion MAPK pathway (17). Finally, the O-mannosyl transferase PMT4 is involved in Msb2 glycosylation, and tunicamycin inhibition of Msb2 glycosylation results in activation of the FG pathway MAPK Kss1 in a pmt4 background (96). Therefore, Msb2 seems to behave as a membrane sensor which connects the FG and HOG pathways (96). Such a role may be similar to that played in a C. albicans homolog. CaMsb2 shows little primary similarity to ScMsb2 but shares a similar overall organization, which consists of a signal peptide (amino acids [aa] 1 to 24), a transmembrane domain (aa 1297 to 1319), and internal repeats that lie in the external domain of the protein. It could also be the case that CaMsb2 behaves as a sensor monitoring cell wall damage, either by detecting altered glycosylation (in response to tunicamycin) or through indirect independent mechanisms (use of Congo Red, zymoliase, or caspofungin). Interestingly, increased P-Cek1 basal levels are found in C. albicans pmt1 and pmt4 mutants (10), supporting the notion that these Pmts may be involved in Msb2 glycosylation.

Which are the final effectors of this signaling pathway? One attractive candidate is the repressor protein Sko1, as this protein mediates caspofungin resistance, and Sko1 levels are downregulated in hst7 mutants (73); Msb2/Sho1 could ultimately influence Psk1, a kinase that regulates Sko1 (73). It should be also noted that recent transcriptomal analyses indicate that the CEK1 pathway represses the chitinase-encoding gene CHT2 (36), suggesting that cek1 mutants could have a lower chitin content. This could contribute to the increased sensitivity of msb2 and sho1 mutants to caspofungin, as these two polymers play compensatory roles in the cell wall (72) and as their synthesis is coordinated in response to cell wall damage (28, 92). Genetic analysis of the he Cek1 and the Sko1 pathway may clarify this point.

We also demonstrate that Cek1 becomes activated upon receiving an upstream stimulatory signal from the Cdc42 GTPase. In C. albicans, activated alleles of the Cdc42 GTPase are able to activate the Cek1-MAPK pathway via the Hst7 MAPK kinase and this process requires both the Sho1 and Msb2 proteins (but not the Cla4 or Cst20 kinases). As deletion of CST20 (but not CLA4) suppresses the lethality associated with these alleles in C. albicans (89), activation of this cascade is not the primary mechanism of cellular death in these mutants, suggesting that Cdc42 interacts with other cellular targets, as occurs in S. cerevisiae (39). Although we do not provide evidence here for a direct (i.e., physical) interaction of Sho1 with Msb2, our studies suggest that the two proteins act together in cooperation with Cdc42 to promote activation of the Cek1 pathway and we have experimental support for the idea of a Sho1-Cdc42 interaction. It must be emphasized that the phenotypes of sho1 and msb2 are not similar, with Sho1 having more drastic effects on Cek1 activation and Msb2 having more drastic effects on morphogenesis and invasion. This could mimic the S. cerevisiae situation, where there are both Sho1-dependent and Sho1-independent roles for Msb2 (87). One of the main functions of a putative Msb2/Sho1 complex would be to attach the Cdc42 kinase to the membrane to promote invasion on solid surfaces via P-Cek1, in close agreement with the phenotypes of cek1 and hst7 (15, 42) (hyphally defective) mutants, ssk1 (9) and hog1 (1) mutants, or cpp1 (31) (hyphally enhanced) mutants. Our results do not eliminate the possibility, however, that another MAPK may also contribute to invasion under these conditions. An attractive candidate is Mkc1, the cell integrity pathway MAPK (3, 62), as the route involving this MAPK promotes polarized growth in S. cerevisiae (60) and mediates invasion in C. albicans (45). Such cross-talk occurs and Cdc42^G12V can efficiently mediate the activation of the cell integrity Mkc1 MAPK when the Msb2/Sho1 proteins are absent. A finding that must be considered in this context is that while caspofungin activates Cek1, equinocandines also activate the cell integrity pathway via the Mkc1 MAPK (61), although this occurs independently of the putative Msb2/Sho1 complex.

Another important conclusion derived from our work is that Msb2 contributes to omosensitivity in C. albicans under conditions of defined genetic backgrounds. As previously described, double sho1 hog1 mutants show a more drastic growth defect under conditions of osmotic stress compared to hog1 mutants (77), suggesting a role for Sho1 in osmotic stress independent of the MAPK. We demonstrate here that deletion of MSB2 in a mutant with the two upstream branches impaired (sho1 ssk1 mutants) seriously compromises growth under conditions of high osmolarity (i.e., in the presence of sorbitol and sodium chloride). In S. cerevisiae, the simultaneous deletion of SSK1 and SHO1 results in a failure to activate the Hog1 MAPK (64) upon osmotic stress challenge. However, and in sharp contrast to the results of experiments with S. cerevisiae, CaHog1 phosphorylation is not blocked in an ssk1 msb2 sho1 mutant. Furthermore, CaHog1 is efficiently translocated to the nucleus (as determined using a GFP-Hog1 fusion [4]) and leads to accumulation of glycerol, a compatible solute in this organism (81). It has been recently shown that CaHog1 is regulated by the MAPKKK Ssk2 only under conditions of stress (13). Ssk2—but not Ste11—appears to be essential for the activation of the Hog1 MAPK in response to osmotic and oxidative stress and its translocation to the nucleus, suggesting that the putative Sho1/Msb2-Ste11 complex has no apparent role in signaling for the HOG pathway. Our results clearly support this model and indicate that other as-yet-undefined upstream components feed into the Ssk2 MAPKKK. It is interesting that in S. cerevisiae, two proteins, Hkr1 and Msb2, are involved in growth under conditions of high osmolarity (87). Hkr1, however, seems to be specialized with respect to the HOG pathway, whereas Msb2 is involved in the FG pathway (67). Therefore, although both proteins interact with Sho1, they are specialized for specific responses. The existence of a C. albicans Hkr1 homolog could partially account for the differences in phenotypes attributed to Sho1 and Msb2.

Importantly, we demonstrate that activation of Hog1 in C. albicans is not sufficient to sustain growth under conditions of high osmolarity. Different explanations, such as differences in transcriptional responses between ssk1 msb2 sho1 and hog1 mutants (in terms of quality or quantity of genes differentially expressed and/or levels attained) or even an inconveniently developed (spatial or temporal) adaptive response, can be invoked to explain this behavior. In S. cerevisiae, for example, Msb2 and Sho1 are localized to regions of polarized growth (16, 71), where they promote Hog1 activation. A putative Msb2/Sho1 complex could be involved in the maintenance of the cellular polarity necessary for growth under restrictive conditions in C. albicans. A more novel mechanism explaining the role of Msb2 in osmotic stress could involve differential localization. In S. cerevisiae, Msb2 is cleaved by the Yps1 aspartyl protease (90) to activate the MAPK pathway. This process mimics the situation with the MUC1 mammalian homologue (49), whose cytoplasmic tail associates with β-catenin (93) and localizes to the nucleus, promoting gene expression. It is tempting to speculate that CaMsb2 could be similarly processed, as a similar cleavage domain exists (approximately 170 aa in a C-terminal domain) in the protein, with aspartyl proteases (37, 58) being testable candidates.

In conclusion, our results support a model in which Msb2 would be involved in invasive growth and cell wall biosynthesis in C. albicans via the Cek1 MAPK. Our results also provide experimental support for a functional specialization of this branch in filamentous fungi, as recently suggested (43, 44). Such a scenario is in agreement with recent results that show how changes in turgor in S. cerevisiae activate the HOG pathway via the SLN1 and not the SHO1 branch (74) and that in C. albicans Ssk2 is the only MAPKKK signaling to Hog1 (13). They also indicate that the signaling mucin Msb2 is involved in growth under conditions of high osmolarity in C. albicans (and maybe other fungi) by a mechanism that is independent of the activation of the Hog1 MAPK.