The genetic interaction network of CCW12, a Saccharomyces cerevisiae gene required for cell wall integrity during budding and formation of mating projections

Background Mannoproteins construct the outer cover of the fungal cell wall. The covalently linked cell wall protein Ccw12p is an abundant mannoprotein. It is considered as crucial structural cell wall component since in baker's yeast the lack of CCW12 results in severe cell wall damage and reduced mating efficiency. Results In order to explore the function of CCW12, we performed a Synthetic Genetic Analysis (SGA) and identified genes that are essential in the absence of CCW12. The resulting interaction network identified 21 genes involved in cell wall integrity, chitin synthesis, cell polarity, vesicular transport and endocytosis. Among those are PFD1, WHI3, SRN2, PAC10, FEN1 and YDR417C, which have not been related to cell wall integrity before. We correlated our results with genetic interaction networks of genes involved in glucan and chitin synthesis. A core of genes essential to maintain cell integrity in response to cell wall stress was identified. In addition, we performed a large-scale transcriptional analysis and compared the transcriptional changes observed in mutant ccw12Δ with transcriptomes from studies investigating responses to constitutive or acute cell wall damage. We identified a set of genes that are highly induced in the majority of the mutants/conditions and are directly related to the cell wall integrity pathway and cell wall compensatory responses. Among those are BCK1, CHS3, EDE1, PFD1, SLT2 and SLA1 that were also identified in the SGA. In contrast, a specific feature of mutant ccw12Δ is the transcriptional repression of genes involved in mating. Physiological experiments substantiate this finding. Further, we demonstrate that Ccw12p is present at the cell periphery and highly concentrated at the presumptive budding site, around the bud, at the septum and at the tip of the mating projection. Conclusions The combination of high throughput screenings, phenotypic analyses and localization studies provides new insight into the function of Ccw12p. A compensatory response, culminating in cell wall remodelling and transport/recycling pathways is required to buffer the loss of CCW12. Moreover, the enrichment of Ccw12p in bud, septum and mating projection is consistent with a role of Ccw12p in preserving cell wall integrity at sites of active growth. The microarray data produced in this analysis have been submitted to NCBI GEO database and GSE22649 record was assigned.


Background
The fungal cell wall is a dynamic structure that acts as a permeability barrier for large molecules and is responsible for osmotic stability and cell shape. In Saccharomyces cerevisiae, the cell wall adds up to 30% of the total cell dry weight. It is mainly composed of polysaccharides (85%) and highly glycosylated mannoproteins (15%) [1]. The vegetative cell wall has a layered ultrastructure with an inner layer of β-1,3and β-1,6-glucan and chitin, and an outer layer of mannoproteins. β-1,3glucan is synthesized at the plasma membrane and constitutes the principal cell wall polysaccharide to which other components are cross-linked [2,3]. Chitin is synthesized by plasma membrane chitin synthases (CHS1-3) [4]. It is abundant at the bud neck and the primary septum between mother and daughter cells. In addition, chitin is a minor component of the lateral cell wall. Most of the chitin at the bud neck is linked to non-reducing ends of β-1,3-glucan, whereas the chitin of the lateral cell wall is mainly attached to β-1,6-glucan [4,5]. Soluble cell wall proteins (SCWs) are non-covalently associated with the carbohydrate meshwork [6]. In addition, Pir proteins (proteins with internal repeats; PIR-CWPs) are covalently linked to β-1,3-glucan, most likely via ester linkages between the γ-carboxyl group of glutamate residues, arising from glutamine, and glucose hydroxyl groups [7,8]. Moreover, at the outer layer of the β-1,3-glucan meshwork highly branched β-1,6-glucan is present to which mannoproteins are linked via a remnant of their glycosylphosphatidylinositol (GPI) anchor (GPI-anchored covalently linked cell wall proteins; GPI-CWPs) [4].
GPI-CWPs are either structural proteins or enzymes involved in cell wall biogenesis. They include the flocculins, the agglutinins, the CRH family, the TIR family, the SED1-SPI1 and GAS families [9,10]. Interestingly, the GPI-CWP Ccw12p cannot be assigned to one of these families. Loss of Ccw12p results in reduced growth rate and increased sensitivity to the cell wall perturbing agents Calcofluor White (CW) and Congo Red (CR), suggesting that Ccw12p is required for maintenance of cell wall stability [7,11]. In accordance, the amount of cell wall chitin is significantly increased in ccw12Δ cells, a feature of mutants with defects in cell wall integrity [12]. Electron microscopy of ccw12Δ cells revealed random deposition of cell wall material [11]. Most likely, the additional cell wall substance is reinforcing the inner glucan-chitin layer. Further, mutant ccw12Δ is especially sensitive to the aminoglycoside antibiotic hygromycin B, probably due to enhanced cell wall permeability [13]. Very similar phenotypes have been described for mutants that lack other GPI-CWPs involved in cell wall synthesis and organization. For example, the absence of Crh1p and Crh2p, which are required for the cross-linking of chitin to glucan, induces hypersensitivity to CR and CW [14].
Global transcriptional changes in response to deletion of genes crucial for cell wall biogenesis and remodelling (such as GAS1, FKS1, KRE6 and MNN9), exposure to cell wall perturbing agents (e.g., Zymolyase, CW and CR) and hyper-activation of the protein kinase C (PKC1)-dependent Cell Wall Integrity (CWI) pathway have been characterized [15][16][17][18]. These analyses revealed a major role for the CWI pathway in activating a compensatory mechanism to counteract different cell wall stress conditions and to prevent cell lysis [15,16,19]. The CWI pathway is activated by the plasma membrane sensor proteins Wsc1-3p, Mid2p and Mtl1p [20,21]. In response to wall perturbations the sensors stimulate exchange activity of the guanine nucleotide exchange factor Rom2p that is recruited to the plasma membrane via phosphatidylinositol 4,5-bisphosphate (PI4,5P2) [22]. Rom2p activates the small GTPase Rho1p that in turn activates among others Pkc1p [23]. Pkc1p turns on a mitogen-activated protein (MAP) kinase cascade consisting of the MAP kinase kinase kinase Bck1p/Slk1p, a pair of redundant MAP kinase kinases (Mkk1p and Mkk2p), and the MAP kinase Mpk1p/Slt2p. Slt2p phosphorylation triggers a number of cellular responses, one of which is the transcriptional activation of a variety of genes that have been implicated in cell wall assembly and structure [23]. Notably, genes involved in chitin biogenesis and cross-linking, GFA1, CHS3 and CRH1, were found to be up-regulated in many cell wall mutants [15]. Compensatory chitin deposition was demonstrated to be a crucial trait during cell wall damages and entirely dependent on the activity of Chs3p [24] and on the proteins controlling its transport and activation, Chs4-7p [25,26].
To learn more about the role of Ccw12p, we determined its sub-cellular localization and performed Synthetic Genetic Analysis (SGA) for mutations leading to growth defects when combined with a CCW12 deletion. In addition, we analysed global transcriptional changes in response to deletion of CCW12. Our findings suggest that both, Ccw12p and the CWI pathway are necessary to ensure cell wall integrity, especially during bud emergence and formation of the mating projection.

Synthetic genetic interactions detected by mutant screening in a ccw12Δ background
Our previous electron microscopic studies and biochemical analyses of the cell wall of ccw12Δ mutants suggested a crucial structural role for Ccw12p during vegetative cell growth [11,12]. To gain further insight into the role of Ccw12p, we screened the Euroscarf collection of yeast mutants for mutations causing a growth defect when combined with a CCW12 deletion.
Haploid deletion mutants in 4869 genes marked by the resistance locus for G418 were robotically arrayed and crossed with a ccw12::nat R strain resistant to nourseothricin. The resulting diploid cells were selected, sporulated, and haploid double mutants scored for growth on medium containing both G418 and nourseothricin (see Methods for details). This selection procedure was performed twice to minimize technical errors and false positives. Double mutants showing no growth or very slow growth phenotypes were selected. We identified 129 mutants with putative genetic interactions with ccw12Δ. To verify these interactions, all identified mutant strains were sporulated, double mutant spores were selected and analysed for a synthetic phenotype. In this way, 21 bona fide synthetic interactions of ccw12Δ were confirmed. Nine of these double mutants did not grow (synthetic lethal = sl) whereas 12 mutants displayed very slow growth phenotypes (synthetic sick = ss; doubling times >10 hrs) (Table 1). Interestingly, 10 of the identified genes (BCK1, BRE5, CHS3, CHS5, CHS7, EDE1, MID2, SHE4, SLA1 and SLT2) have been described previously as CCW12 genetic interactors validating the quality of our screening procedure [27][28][29][30][31]. Moreover, the identification of 11 new interactors underlines the sensitivity of the screening. The 21 candidate genes were arranged into functional categories according to the Saccharomyces Genome Database (SGD):

Cell wall synthesis and regulation
Three of the identified genes are elements of the CWI pathway: the upstream sensor MID2 (ss), the MAP kinase kinase kinase BCK1 (sl) and the MAP kinase SLT2 (sl). Moreover, downstream targets of CWI pathway involved in chitin deposition were found: CHS3 (ss), a N-acetylglucosamine transferase responsible for chitin synthesis at the bud neck and in the lateral cell wall as well as for chitin deposition in response to cell wall stress; CHS4 (ss), a direct activator of Chs3p activity [32] anchoring Chs3p to the septin ring via Bni4p [33]; CHS5 (sl), a component of the exomer complex involved in the polarized transport of Chs3p in specialized vesicles [34]; and CHS7 (ss), a membrane protein controlling Chs3p export from the ER [35][36][37]. As a compensatory response, increased accumulation of chitin has been observed in ccw12Δ cells [12]. Moreover, deletion of chitin biosynthesis genes in a ccw12Δ background causes lethality, further proving the essential interconnection of Ccw12p and chitin in cell wall stability. Furthermore, we identified another target of the CWI pathway: MNN10 (sl), a Golgi mannosyltransferase involved in processing of N-linked glycans and septum formation [38,39]. Interestingly, simultaneous deletion of HOG1 and CCW12 resulted in a severe growth defect (ss). This result is consistent with previous findings identifying Hog1p, a MAP kinase involved in osmoregulation, as modulator of the response to cell wall damage [15][16][17]40,41].

Cell polarity and vesicular transport
Three genes required for endocytosis and polarization of the secretory pathway, EDE1 (ss) (key endocytic protein involved in a network of interactions with other endocytic proteins), SHE4 (ss) (regulator of myosin function), and SLA1 (sl) show severe synthetic phenotypes in combination with ccw12Δ. SLA1 is crucial for NPFXDmediated endocytosis that is required for the function of the CWI pathway sensor Wsc1p, thus connecting endocytosis and cell wall integrity [42]. In addition, five genes involved in cell polarity caused a synthetic phenotype: ARC18 (sl) (motility and integrity of cortical actin patches), SAC6 (sl) (organization and maintenance of the actin cytoskeleton), PFD1 (sl) (biogenesis of actin and alpha-and gamma-tubulin), BRE5 (ss) (vesicular transport between the endoplasmic reticulum and Golgi compartments) and FEN1 (ss) (membrane biogenesis).

Other genes
Finally, genes belonging to other functional categories were found: WHI3 (ss), a cell size regulator required for passage through START; SRN2 (ss), a component of the endosomal sorting complex required for transport-I (ESCRT-I) which is involved in ubiquitin-dependent sorting of proteins into the endosome; PAC10 (sl), a cochaperone promoting efficient protein folding, and YDR417C (ss), a gene of unknown function potentially coding for a GPI anchored protein.

Phenotypic analysis of individual single mutants in genes identified by SGA
On the basis of computational analyses of large scale experiments EDE1, SHE4, SLA1 and SAC6 are predicted to be involved in cell wall organization, whereas PFD1, WHI3, SRN2, PAC10, FEN1 and YDR417C are not linked to cell wall synthesis, signalling or compensating mechanisms [43]. To further characterize the latter genes, we assayed CW sensitivity of the corresponding deletion mutant strains. CW interferes with the assembly of compensatory cell wall chitin. Therefore, CW sensitivity is a common feature of mutants with cell wall defects. Cells were spotted on solid rich medium containing 25 μg/ml CW, and incubated for 2 days at 30°C ( Figure 1A). Due to the previously described phenotype [11], the ccw12Δ mutant was included as a control for growth inhibition. Growth of pfd1Δ and ydr417cΔ was not affected by CW ( Figure 1A). In contrast, pac10Δ, whi3Δ and fen1Δ cells displayed intermediate, and srn2Δ cells high CW sensitivity.
Moreover, we analysed the CWI pathway in the CW sensitive deletion mutants, pac10Δ, whi3Δ, fen1Δ and srn2Δ, since constitutive activation of the CWI pathway is a characteristic phenotypic trait of mutants that directly affect cell wall structure. We addressed CWI pathway induction by measuring phosphorylation of the MAP kinase Slt2p. Slt2p activity increases as a result of tyrosine/threonine phosphorylation of the protein.
Protein extracts were analysed by immunoblotting using α-phospho-p44/p42 MAP kinase antibodies, which detect the dually-phosphorylated form of Slt2p. The MAP kinase Slt2p was not activated in wild-type (wt) and pac10Δ, whi3Δ, fen1Δ and srn2Δ mutant strains under normal growth conditions ( Figure 1B). However, cell wall stress conditions (CW treatment and increased temperature) resulted in hyper-activation of Slt2p in the  mutants in comparison to wt cells ( Figure 1C). Taken together, these results suggest that Pac10p, Srn2p, Whi3p and Fen1p are not directly involved in cell wall assembly but rather affect cell wall stress response. These proteins are involved in crucial processes such as protein folding, sorting into endosomes, cell cycle regulation and membrane lipid synthesis ( Table 1). The corresponding mutants display reduced fitness with respect to various stress conditions (e.g., growth at 41°C; hygromycin B, ethanol treatment; http://www.yeastgenome.org/). Thus, our results suggest that in pac10Δ, whi3Δ, fen1Δ and srn2Δ cells hyper-activation of the CWI pathway in response to cell wall stress conditions is an indirect effect, although a direct regulation cannot be ruled out completely.

Global transcriptional responses of the ccw12Δ mutant
In order to comprehensively assess the effect of the cell wall damage caused by CCW12 depletion, we examined the ccw12Δ transcriptome using DNA microarray technology, as described in Methods. Yeast strains were grown to an OD 600 of 1 and total RNA was used to analyse expression levels of all S. cerevisiae genes present on the Affymetrix GeneChip ® Yeast Genome S98 array (Additional file 1). Deletion of CCW12 resulted in transcriptional changes (fold change ≥2 or ≤ 0.5) of 473 genes (8% of the yeast genome) (Additional file 2). Interestingly, 91% of these genes were up-regulated, while only 9% showed reduced expression. Among the 431 up-regulated genes, 94% had 2 to 3-fold increased transcript levels, and 6% were more than 4-fold enhanced. In contrast, 41 genes were down-regulated, with 85% showing a 2 to 3-fold transcriptional reduction and 15% being reduced more than 4-fold. Transcriptome data were verified for selected genes using quantitative RT-PCR (qRT-PCR) ( Table 1,  Table 2 and Table 3) and Northern blot analysis (Additional file 3). Conformity between microarray, qRT-PCR and Northern blot data was obtained for the genes analysed. In the case of ECM13, transcript levels detected by qRT-PCR were less marked, possibly due to different gene regions experimentally verified. However, all experimental approaches point out significant induction of ECM13 transcription in ccw12Δ (Table 2 and Additional file 3). To further analyse the transcriptional profile of the ccw12Δ mutant, differentially expressed genes were grouped into functional categories according to the MIPS nomenclature (http://mips.helmholtz-muenchen.de). We expressed the number of up-and down-regulated transcripts in each of the categories as percentage of the total number of differentially expressed genes and compared this functional classification with the functional catalogues of genes described in MIPS. As a result the major remodelling of gene expression is accentuated on a genomic scale ( Figure 2). Transcripts encoding products involved in central metabolism, generation of energy, cell rescue and biogenesis of cellular components are significantly overrepresented in mutant ccw12Δ (p-value < 0.05; p-values were obtained using hypergeometric distribution statistics; default setting in FunCatDB tools, MIPS). In contrast, transcripts encoding products involved in protein synthesis and transcription   were underrepresented (although with a p-value > 0.05), consistent with the fact that cell proliferation is impaired by mutations affecting cell wall biogenesis [44]. At present it cannot be completely excluded that part of the observed responses, especially related to transcription and protein biosynthesis, are due to the reduced growth rate of the ccw12Δ mutant (approximately 25% less than wt). Among the differentially expressed genes, 7% (32 out of 472 genes) are directly involved in the construction and remodelling of the cell wall ( Table 2). Most of these genes showed up-regulation while only two, MUC1 and TIR1, have reduced expression. These findings indicate a strong bias towards remodelling of the cell wall in the ccw12Δ mutant to compensate for defects. Furthermore, the DNA microarray data revealed that some of the characteristic target genes of the CWI and HOG pathways [45,18,46] were affected in the ccw12Δ mutant. As shown in Table 2, we found regulation of BCK1, coding for the MAPKKK of the pathway; SLT2, the terminal kinase; MLP1, encoding a putative Ser/Thr kinase with similarity to Slt2p; and SDP1, whose product is a dual-specificity MAP kinase phosphatase acting on and regulating the phosphorylated/active form of Slt2p [23]. SLT2 levels increased about 2.2 times, in agreement with a similar modulation reported for other cell wall mutants such as kre6Δ, gas1Δ or mnn9Δ [15]. Downstream targets of the CWI pathway included genes involved in chitin synthesis (RCR1, CSI2, CHS2, CHS3 and CHS6) [47] and β-1,6-glucan synthesis (KRE9) [48] as well as mannoprotein-encoding genes, such as Pir3p, the putative GPI-anchored Fit2p and Ylr194cp, and GPI-anchored aspartic proteases (yapsines YPS3, YPS5 and YPS6), whose role in the cell wall construction has been inferred from genetic studies [49]. Further, we found transcriptional induction of PBS2, the MAPKK gene of the HOG pathway.
Another group of up-regulated genes is involved in the biosynthesis of protein-linked glycans, including MNS1, which encodes the ER mannosidase I, and ROT2, encoding a luminal α-glucosidase II of the ER. Mns1p and Rot2p are involved in the trimming of Nlinked glycans in the ER [50,51]. Similarly, up-regulation was found for OCH1 and YUR1, coding for Golgi α-1,6and α-1,2-mannosyltransferases, respectively, both being involved in the maturation of N-linked oligosaccharides [51]; and for KTR3 and KTR4, encoding α-1,2-mannosyltransferases involved in the outer chain synthesis of the N-glycans and in the elongation of O-linked oligosaccharides [51].
Beside the alterations in the expression levels of glycosyltransferases and glycosidases, up-regulation of ECM4-13-15-17-21-31 was observed. These genes encode proteins that are poorly characterized or have unknown functions. However, due to the corresponding mutant phenotypes these genes are thought to play a role in different aspects of cell wall construction.

Comparison of the ccw12Δ transcriptome with transcriptional profiles from other cell wall-related mutants
In order to compare transcriptional changes in the ccw12Δ mutant with transcriptomes from previous studies investigating responses to constitutive cell wall damage caused by deletions of cell wall-related genes, such as GAS1, FKS1, KRE6 and MNN9, and transient cell wall damage induced by CR or Zymolyase treatment [15,16], data were organized by hierarchical clustering (see Methods). As shown in Figure 3, two specific gene clusters could be identified.
Cluster I includes 20 genes that were highly induced in the majority of the mutants/conditions analysed. These genes are directly related to cell wall compensatory responses: the CWI pathway MAP kinase SLT2 and its homolog MLP1 (YKL161c); the known CWI pathway downstream targets PRM5; the PIR-CWP PIR3; and four genes coding for GPI-anchored proteins (YLR194c, YPS3, SVS1 and YJL171c). In addition, cluster I includes genes whose absence causes increased stress sensitivity: FLC1, coding for a putative flavin adenine dinucleotide (FAD) transporter required for uptake of FAD into endoplasmic reticulum and cell wall maintenance via disulphide bond formation [52]; PRY2, encoding a protein of unknown function (null mutant has a decreased resistance to hyperosmotic stress); SUR1, coding for a mannosylinositol phosphorylceramide (MIPC) synthase involved in sphingolipid biosynthesis (null mutant is sensitive to caffeine); FBP26, encoding a fructose-2,6-bisphosphatase (null mutant has a decreased resistance to hyperosmotic stress); and YKE4, coding for a zinc transporter (null mutant has decreased resistance to CW). Except for SUR1, all of these genes are activated in a strain with hyper-activated PKC1 [53], and thus can be considered as targets of the CWI pathway ( Figure 3, right column). Accordingly, promoter analyses of these genes revealed the presence of one or more putative binding sites for the transcription factors of the CWI pathway, RLM1 and/or SWI4 (data not shown). Cluster II contains genes which are moderately upregulated under most of the analysed conditions (64 genes). Many of them are also induced upon PKC1 overexpression (Figure 3). This cluster includes functional categories like heat shock response and metabolism, comprising 8 genes connected to sugar, glycoside, polyol and carboxylate catabolism (TPS1, ARA1, KGD1, YJR096W, ACO1, IDH1, CIT1 and IDH2) and 6 genes involved in the tricarboxylic-acid pathway (CIT2, KGD1, ACO1, IDH1, CIT1 and IDH2).
To further verify the importance of the CWI pathway for ccw12Δ cells, we analysed the activation of Slt2p. As positive controls for Slt2p activation, we used cell extracts from gas1Δ cells and from wt cells incubated at 37°C for three hours. Compared to the wt, the ccw12Δ mutant displayed a 3.85-fold higher level of phosphorylated Slt2p (Figure 4). Addition of 1 M Sorbitol or 0.5 M KCl for osmotic stabilization greatly reduced the phosphorylation of Slt2p in ccw12Δ cells.
In summary, loss of CCW12 activates the CWI pathway which in turn up-regulates the expression of a distinct subset of genes that is also induced in all the other cell wall stress conditions analysed so far. Moreover, this pathway ensures cell viability of mutant ccw12Δ as emerged in the SGA screen.
Ccw12p is crucial for cell wall stability of forming daughter cells During vegetative growth, the CWI pathway is essential for bud formation [23]. Since ccw12Δ mutants depend on the CWI pathway (see above) and produce buds that are abnormally round [54], we speculated that Ccw12p could play a role in maintaining cell wall integrity during bud growth. Thus, we analysed the mutant for abnormalities during vegetative growth. When grown in YPD medium, buds of ccw12Δ cells appeared rounder when compared to buds of wt cells. Bud morphology was further monitored by determining the relation between the length of the long and short cell axis. In contrast to wt cells (axes ratio of 1.2 ± 0.1), the long/short axis ratio was diminished in ccw12Δ cells (axes ratio 0.8 ± 0.1). Furthermore, 16% of ccw12Δ cells showed cell lysis during bud emergence as compared to only 3% of wt ( Figure 5A, a  and b). Addition of 1 M Sorbitol to the growth medium could partially prevent cell lysis (10%) and round bud morphology (axes ratio 1.1 ± 0.1) of ccw12Δ cells ( Figure  5A, c). Moreover, when ccw12Δ cells were shifted from osmotically stabilized medium (YDP plus 1 M Sorbitol) to YPD (hypotonic shock), buds showed pronounced lysis. While 41% of the mutant cells underwent cell death mainly in early stages of bud formation, only 8% of wt cells were affected ( Figure 5A, d). To a minor extent, cells lysed after completion of cytokinesis, but, in that case lysis was mainly restricted to daughter cells ( Figure 5A, e and f).
To further study the role of Ccw12p in bud formation, we analysed its sub-cellular localization. For this purpose we constructed a plasmid encoding a Ccw12p-GFP fusion protein. The GFP-coding sequence was inserted into the CCW12-HA gene [11] at bp +129 without altering the N-terminal signal sequence and the GPI attachment site. Transcription of GFP fusion was under the control of the endogenous CCW12 promoter. The Ccw12p-GFP fusion protein was completely functional since it was able to complement all described mutant phenotypes, i.e. the slow growth phenotype (wt T d = 1.7 hrs; ccw12Δ T d = 2.2 hrs; ccw12Δ plus Ccw12p-GFP T d = 1.8 hrs) and the CW hyper-sensitivity (data not shown). Fluorescence microscopy revealed a uniform GFP signal around the cell surface confirming the presence of Ccw12p in the plasma membrane/cell wall. Interestingly, Ccw12p-GFP appeared more concentrated at the presumptive budding site on the cell surface of the mother cell shortly before bud emergence ( Figure  5B, a). Concomitant with bud emergence, Ccw12p-GFP strongly accumulated at the periphery of small to medium-sized buds ( Figure 5B, b and c). When buds reached a larger size, the protein appeared in the lateral cell wall of daughter cells and labelled the septum region with an asymmetric localization towards the daughter side. Concurrent with mother/daughter cell separation, Ccw12p-GFP was still detected on the daughter side of the cells (Figure 5B, d). Thereafter, the Ccw12p-GFP signal vanished, suggesting a specific I II Figure 3 Hierarchical clustering of genes differentially expressed in ccw12Δ. The transcriptional response of the ccw12Δ mutant was compared to transcriptomes of other cell wall mutants (gas1Δ, fks1Δ, mnn9Δ and kre6Δ) and cell wall stress conditions (Zymolyase and Congo Red) [15,16]. Each column represents a different condition. Each row represents the ratio of expression for each gene as it is indicated in the colour scale. The clustering tree was built using MeV software from TIGR. function related to the stability of the growing cell wall and the septum. Interestingly, a similar distribution was found for upstream sensors and other components of the CWI pathway, further underlining the crucial role of monitoring correct assembly and structural capacity of the newly synthesized cell wall [55].
In summary, the cell lysis phenotype of ccw12Δ mutants and the localization of Ccw12p-GFP are supporting an essential function of Ccw12p for cell wall integrity predominantly during the formation of daughter cells.

Ccw12p is involved in the mating process
The observed mating and agglutination defects of ccw12Δ mutants indicate a role of CCW12 in conjugation [7]. In accordance with these findings, our transcriptome analysis revealed significant regulation of genes involved in different steps of the mating process (Table 3) and hierarchical cluster analysis further demonstrated that this response is specific for ccw12Δ cells and not observed in other cell wall mutants (data not shown). The gene encoding the mating pheromone α-factor, MF(ALPHA)2, showed a 3-fold reduction and genes coding for cell wall or plasma membrane proteins involved in mating such as FIG1, FIG2 and FUS1 were down-regulated as well. Further, transcription of a set of pheromone-regulated genes (PRM1-3-4-6-7) required for different steps of membrane and nuclear envelope fusion was reduced. Surprisingly, the pheromone-regulated gene PRM5 was significantly activated (2.84-fold). However, PRM5 is induced by the CWI pathway [18] and thus activation by cell wall stress might overcome the repression during mating response. Finally, PRM10, that was proposed to be involved in mating, was up-regulated 3.67 fold, possibly due to the presence of one Rlm1p and two Swi4p binding sites in its promoter (data not shown).  To further substantiate the indicated role of Ccw12p during the mating process, we analysed the response of ccw12Δ cells to mating pheromone. During pheromone treatment, cells arrest in G1 and undergo polarized growth, forming a mating projection that is directed towards the signal source. This chemotropic response is thought to involve the generation of an internal landmark that reflects the direction of incidence of the external pheromone signal and overrides the spatial cues that normally control bud formation [56,57]. Although transcriptome analysis was performed in ccw12Δ MATalpha cells, we further characterized the pheromone response of MATa cells since the mating pheromone α-factor is commercially available, and for both, ccw12Δ MATa and MATalpha cells agglutination and conjugation defects have been demonstrated [7]. However, minor differences cannot be completely excluded. Wt and ccw12Δ MATa cells were treated with 20 μg/ml of α-factor. Budding index ( Figure 6A) and formation of mating projections (shmoos) ( Figure 6B) were monitored up to five hours after α-factor addition. In wt and ccw12Δ cells the percentage of budded cells decreased from 74% and 62%, respectively, to 2-3% after two hours of pheromone treatment. The percentage of budded wt cells remained constant with only a slight increase to 7% of budded cells after five hours ( Figure 6A). Different from that, after 3 hours the number of budded ccw12Δ cells rapidly increased to reach 62% five hours after α-factor addition ( Figure 6A) suggesting an earlier release from the G1 arrest.
Mating projections of ccw12Δ cells were rounder and less polarized than those of wt cells ( Figure 7A, a and b). Moreover, methylene-blue vital-staining showed that about 10% of the ccw12Δ cells undergo mating-induced cell death (versus 1% of wt cells) ( Figure 7A, c). In addition, some shmoo-forming cells are still attached to dead daughter cells ( Figure 7A, d-e). Further, 23% of ccw12Δ mutant cells that are exiting the pheromone block undergo cell lysis during bud emergence (versus 5% of wt cells) (Figure 7A, f). These data support a role for Ccw12p in stabilizing the cell wall during shmoo formation and during budding upon re-entry into the mitotic cell cycle.
The mating projection is a crucial structure during mating since it serves to concentrate the proteins involved in signalling (pheromones and pheromone receptors), cell adhesion (agglutinins) and fusion (Fus2p) in the area of future cell contact and fusion [58]. Interestingly, components of the CWI pathway, such as the cell wall integrity sensor Mid2p [59], accumulate in the same area suggesting a crucial role for cell wall construction and sensing during this process. Therefore, we used the GFP fusion to monitor Ccw12p localization during α-factor treatment. As shown in Figure 7B, Ccw12p-GFP is enriched in mating projections with a gradient starting from the tip of the shmoo. The same localization of Ccw12p-GFP was observed when shmoo formation was triggered in MATalpha cells by the presence of MATa cells (data not shown). In combination our data reinforce the notion that Ccw12p plays a general role in cell wall stability at the shmoo apex.

Discussion
In this work we provide evidence that Ccw12p has a specific role in maintaining the stability of the newly synthesized cell wall especially during bud emergence and formation of the mating projection. Ccw12p localizes to areas of active cell wall synthesis such as forming buds ( Figure 5B). In the absence of Ccw12p, buds are abnormally round [54] and cells tend to lyse during bud emergence and development ( Figure 5A, panel b; 16% of ccw12Δ vs. 3% of wt cells). Moreover, when ccw12Δ cells are subjected to hypotonic shock buds lyse more frequently in early stages of development ( Figure 5A, panel d; 41% of ccw12Δ cells vs. 8% of wt cells). Even after cytokinesis, daughter cells tend to lyse ( Figure 5A, panel e). In agreement with the observed lysis phenotype, Ccw12p is specifically localized at the daughter side of the septum ( Figure 5B, panel d). During mating, cell lysis is observed when the shmoo is forming ( Figure 7A, panel c; 10% of ccw12Δ cells vs. 1% of wt cells) and Ccw12p is specifically localized to the mating projection ( Figure 7B), further supporting a local function for cell wall stability during remodelling. The CWI pathway plays a crucial role in cell wall stability especially during bud and shmoo formation. Mutants in this pathway show some of the phenotypic traits that are characteristic for ccw12Δ [23]. We could demonstrate, that in ccw12Δ cells the pathway is hyperactivated ( Figure 4) and that osmotic stabilization suppresses CWI pathway hyper-activation ( Figure 4) as well as the growth defect of ccw12Δ [12]. In addition, SGA and transcriptome analyses further demonstrate the interdependence between components of the CWI pathway and CCW12 (Tables 1 and 2). SGA revealed synthetic lethality of the CCW12 deletion in combination with mutants of the key regulatory kinases of the CWI pathway, BCK1 and SLT2 (Table 1). In addition, combinatory deletions with chitin synthesis genes (Table  1; [60]), and MID2, one of the main sensors of the CWI pathway, (Table 1; [29]) displayed severe synthetic growth phenotypes. Surprisingly, no direct involvement of the sensors WSC1-3 [23] was found. Wsc1p normally localizes to sites of new cell surface growth as similarly observed for Ccw12p-GFP during the budding process. However, we could demonstrate an interaction with SLA1. It is known that Sla1p, a component of the actinand clathrin-based endocytic machinery can serve as an NPFX (1,2) D-specific endocytic adaptor and that the NPFX (1,2) D-Sla1p system is responsible for directing Wsc1p into an endocytic and recycling pathway necessary to maintain yeast cell wall polarity [42]. In turn, inhibition of Wsc1p endocytosis causes defects in polarized deposition of cell wall material and increased sensitivity to perturbation of cell wall synthesis. In our analysis, SLA1 deletion conferred synthetic lethality in combination with ccw12Δ (Table 1), similar to what was found in independent screens performed combining deletions of SLA1 and genes involved in glucan and chitin synthesis [60,30]. Moreover, the SLA1 transcript is 2.69-fold increased when CCW12 is absent (Table 1). These results suggest that NPFX (1,2)  could be explained by a redundant function of one of the other WSC sensors. In contrast, deletion of SLA1 might not only affect WSC1 but other WSC-family members as well.
Transcriptional profiling of the ccw12Δ mutant revealed predominantly cell wall remodelling activities. Members of all categories of cell wall related genes are regulated: CWI pathway signalling, structural components, chitin/glucan synthesis and glycosylation with many of them under the control of PKC1-mediated signalling cascade. Interestingly, despite the high transcriptional activation of some of these genes, their deletion mostly generates mild or no cell wall phenotypes and only a few of them conferred synthetic lethality when deleted in combination with CCW12. This may either be explained by functional redundancy or the need to repress several cell wall functions and/or components in order to cause severe cell wall damage.
In addition to the CWI pathway, the high-osmolarity glycerol (HOG) and the SLN1-SKN7 pathway contribute to cell wall integrity, The HOG pathway is modulated by various plasma membrane sensors [61]. One of which is the kinase Sln1p, the initiating member of a two-component type of phospho-relay cascade that negatively regulates the HOG pathway. In addition, Sln1p activates the transcription factor Skn7p which in turn triggers the transcription of specific target genes such as OCH1 and NCA3 [13]. CWI and HOG pathway are closely interconnected suggesting a cooperative role for both pathways controlling cell integrity [15][16][17]40,41]. Likewise, it was suggested that the CWI and the SLN1-SKN7 pathway may function in parallel to protect cells from lysis [13]. Interestingly, SGA revealed the interdependence between CCW12 and HOG1, the MAP kinase of the HOG pathway (Table 1). Further, transcription of OCH1 and NCA3 is significantly increased in the ccw12Δ mutant (2.69-and 3.36fold increase respectively; Additional file 2) confirming previous data that suggested Ccw12p to be a modulator of Sln1p activity [13]. Thus, different signalling pathways are involved to ensure cell wall integrity in the absence of Ccw12p. However, the CWI pathway and cell wall remodelling genes play a dominant role as demonstrated by the preponderance of CWI related genes identified in the SGA screen.
We further addressed the question whether different cell wall defects induce a similar response in yeast cells. Although there is some variation in the regulation of individual genes under different conditions, their functional classification indicates a common principle. We found significant overlap in the transcriptional response to CCW12 deletion as compared to that in other cell wall mutants and treatment with cell wall perturbing agents such as Zymolyase or CR [15,16]. Notably, under all tested conditions we found 20 genes to be strongly induced, which could constitute the core response for cell wall stress (Figure 3, Cluster I). With the exception of SUR1, all the genes in this cluster are CWI pathway targets, reinforcing the notion that survival under cell wall stresses is mediated by different aspects regulated by CWI pathway.
In a previous work, a marked decrease in the mating and agglutination ability of ccw12Δ strain was detected [7]. Here we demonstrate that ccw12Δ cells exhibit diminished α-factor sensitivity and earlier release from the pheromone-induced G1 arrest ( Figure 6). The mating-related phenotypes could be due to a combination of different factors: i) during pheromone treatment, the absence of Ccw12p at the shmoo apex results in mating induced cell death of~10% of the cells ( Figure 7A). As a result, mating frequency is significantly reduced. ii) The ccw12Δ expression profile revealed down-regulation of genes involved in the mating process which is not detected in other cell wall mutants. iii) ccw12Δ cells show an early release from the pheromone-induced G1arrest. The time cells display mating projections and the recovery time from pheromone arrest can greatly influence the mating efficiency. It is well known that in the absence of a successful mating event, G1-arrested cells re-enter the mitotic cycle through a recovery process that involves down-regulation of the mating mitogenactivated protein kinase cascade mainly through Msg5p phosphatase activity on the MAP kinase, Fus3p [62]. Furthermore, recovery is regulated by POG1, coding for a putative transcriptional activator [63]. Overexpression of POG1 inhibits α-factor-induced G1-arrest and transcriptional repression of the CLN1 and CLN2 genes. This loss of transcriptional repression occurs through SCB/MCB promoter elements and requires Bck1p, known to up-regulate Swi4-dependent cell-cycle box (SCB)/MluI cell-cycle box (MCB) promoter elements during vegetative growth [64]. The G1-cyclin Cln2p, in addition to driving the G1-to S-phase transition [65], when over-expressed blocks the ability of cells to arrest in the presence of α-factor, primarily through an effect on Ste20p, an activator of the mating MAPK cascade [66]. In agreement with these data, we detected a 3.14fold increase for POG1 and a 2.18-fold increase for CLN2 in ccw12Δ cells, explaining the earlier G1-release.

Conclusions
Transcriptome and synthetic genetic analyses revealed that in ccw12Δ cells a cell wall compensatory mechanism is activated that integrates several major regulatory systems mainly under control of the CWI pathway-signalling module. Further, we demonstrate for the first time that Ccw12p is mainly localized in areas of newly synthesized cell wall such as emerging buds and mating projections.
Our data strongly suggest that Ccw12p has a specific role in reinforcing the expanding cell wall. It will be a challenging future task to unravel the molecular mechanism of Ccw12p action.

Yeast strains and plasmids
Standard procedures were used for all DNA manipulations [67]. All cloning and transformations were made in E. coli strain DH5α. Oligonucleotide sequences are available upon request.
Plasmid pCCW12-GFP: a 741 bp DNA fragment corresponding to the coding sequence of Aequorea victoria GFP was amplified by PCR and subcloned into the BamHI restriction site of pCCW12 [11]. Correct orientation of the insert was verified by restriction digestion and DNA sequencing.
Yeast cells were grown in synthetic complete (SD) or YPD medium at 30°C. Haploid deletion mutants were available from the deletion project consortium EURO-SCARF. These strains were arrayed in duplicate on fiftyone 384-format plates using a colony picker. Arrays were propagated at 30°C on YPD or YPD supplemented with 200 μg/ml G418 (Invitrogen) and 100 μg/ml clon-NAT (Werner BioAgents).

SGA screening
The method described in [27] was used. Briefly, the query strain ERccw12Δ was pinned onto a fresh YPD plate at a density of 384 cells per plate. Then, the deletion mutant array was pinned on top of the query cells. The resulting diploids were selected on YPD containing G418 (200 μg/ml) and clonNAT (100 μg/ml). Arrays were then pinned onto sporulation medium. After a 9day incubation at 23°C spores were pinned onto haploid selection medium (SD supplemented with histidine, arginine and canavanine) to select for growth of MATa spore progeny. For 2 days meiotic progenies carrying the G418 resistance locus were selected on medium containing G418 at 30°C (SD supplemented with histidine, arginine, canavanine, G418). The G418 and Nat resistant cells were selected on medium containing G418 and clonNAT for 2 days at 30°C (SD supplemented with histidine, arginine, canavanine, G418, nourseothricin). Colony size was scored by eye. Each screen was done in duplicate.
To confirm putative interactions spores were diluted in sterile water, spotted onto solid haploid selection medium and incubated for 2 days at 30°C (selection of the MATa progeny). Double mutants were then selected as described above. According to colony growth double mutants were scored as synthetic sick (ss) or synthetic lethal (sl) after 2 days of incubation at 30°C. Microarray analysis of differential gene expression SEY6210 and ccw12Δ (MEY12A) strains were cultured in YPD supplemented with adenine (39 mg/l) to an OD 600 of 1. Total RNA was extracted according to [71]. For each strain two pools of three independent RNA preparations were obtained, representing two biological replicates each. Concentration of total RNA was measured at 260 nm and sample quality was checked using RNA Nano Labchips in a Bioanalyzer 2100B (Agilent Technologies, Palo Alto, CA). Double stranded cDNA was synthesized from 5 μg of total RNA using "Onecycle cDNA Synthesis Kit" (Affymetrix, Santa Clara, CA). After cDNA purification using the "GeneChip Sample Cleanup Module" (Affymetrix), this DNA was used as template for the in vitro transcription using the "IVT Labeling kit" (Affymetrix) to obtain the biotin labelled cRNA. The obtained cRNA was fragmented and hybridized to the Affymetrix GeneChip ® Yeast Genome S98 array for 16 h at 45°C. Hybridized microarrays were washed and stained with a streptavidin-phycoerythrin conjugate in a "GeneChip ® Fluidics Station 450". All these procedures were carried out as suggested by the manufacturer. Hybridized cRNA was finally identified by the fluorescence signal in a "GeneChip ® 3000" scanner. The files generated from the scanning were converted to gene expression signals using the GCOS software (Affymetrix, [72]). Data from the two mutant samples were compared against data from the two SEY6210 samples, obtaining a total of four comparisons. Genes with at least two Present values as "Detection call" and four Increase or Decrease values as "Signal call" were selected for further analysis. The Signal Log Ratio was obtained as the mean of the Signal Log Ratio obtained in each of the four comparisons, and genes with Signal Ratio >2 or < 0.5 were considered for further analysis. Clustering and visualization of the data was performed using the MeV software (TIGR) [73].
Quantitative RT-PCR assays SEY6210 and ccw12Δ (MEY12A) strains were cultured in YPD supplemented with adenine (39 mg/l) to an OD 600 of 1. Total RNA was isolated from 5 × 10 7 cells using the RNeasy MINI kit (mechanical disruption protocol; QIA-GEN, Hilden, Germany). On-column DNase digestion of the samples was performed following manufacturer's instructions. From each strain, two biological replicates were processed. Concentration of total RNA was determined by measuring the absorbance at 260 nm. RNA purity and integrity were assessed using the RNA Nano Labchips in an Agilent 2100B Bioanalyzer (Agilent Technologies, Palo Alto, CA). First strand cDNAs were synthesized from 1 μg of total RNA in 20 μl final volume, using the High Capacity RNA-to-cDNA Master Mix (Applied Biosystems) following the recommendations of the manufacturer. To exclude the presence of genomic DNA, reactions were performed in the absence of reverse transcriptase. Real time quantitative PCR assays were carry out in an ABI 7900HT Fast Real-Time PCR instrument (Applied Biosystems, Foster city, CA) using standard PCR conditions. Duplicates of all reactions were analysed. Quantification of BCK1, CHS3, ECM13, EDE1, FIG1, FUS1, MLP1, PFD1, PRM10, PRM3, SLA1, SLT2, and YPS5 was performed using the "FastStart Universal SYBR Green Master (ROX) (Roche Applied Science)". Based on ORF sequences provided in SGD (http://www. yeastgenome.org/), gene-specific primers were designed using "Primer Express ® Software" (Applied Biosystems). Primer sequences will be provided upon request. For quantification, the abundance of each transcript in the mutant was determined relative to the standard 18S ribosomal RNA and with respect to the wt by using the Comparative Ct Method. Following this approach, the two RNA samples from the mutant were compared to each of the two samples from the wt strain, thus obtaining a total of four comparisons (four ratios). Final data for each transcript was expressed as a media of these four ratios with their corresponding standard deviation.

Test for Calcofluor white (CW) sensitivity
Strains were grown in YPD medium to exponential phase. A total of 10 8 cells were collected and suspended in 0.1 ml of sterile water. 10-fold serial dilutions were prepared. 3 μl of each dilution were spotted onto YPD plates containing 25 μg/ml CW. After two days of incubation at 30°C, growth was scored,

Preparation of cell extracts and Western blot analyses
Yeast strains were grown in YPD medium to 2 × 10 7 cells/ml. In the case of stress induction experiments, log-phase cultures were treated with CW (40 μg/ml for 1 hour) or exposed to elevated temperature (37°C for 3 h). 1 × 10 7 cells were collected and cell pellets frozen in liquid nitrogen. 20 μl of SDS sample buffer (60 mM Tris-Cl pH 6.8, 2% SDS, 10% glycerol, 5% β-mercaptoethanol, 0.01% bromophenol blue) were added to the frozen cells. Samples were boiled for 5 min to extract proteins.

Pheromone treatment
Exponentially growing yeast cells (5 × 10 6 cells/ml YPD) were treated with 20 μg/ml α-factor (GenScript) or mock treated. At indicated times, cells were analysed by phase-contrast microscopy and scored for budding and shmoo formation by counting at least 200 cells after mild sonication.

Fluorescence microscopy
Yeast strains transformed with plasmid pCCW12-GFP were grown under selective conditions. Cells were collected by centrifugation, washed twice with PBS, and incubated on ice for at least 15 min before further processing. The cells were examined as wet mounts using a BX60 microscope (Olympus, Melville, NY) and a DC290 digital photo camera (Eastman Kodak, Rochester, NY) or an Eclipse 90i microscope (Nikon, Tokyo, Japan) equipped with epifluorescence, Nomarski optics, and a Hamamatsu ORCA-ER device camera (Nuhsbaum, McHenry, IL). Cell viability was analysed by methylene blue staining [74]. For DNA staining, 8.3 μg/ml 4,6-diamidine-phenylindole (DAPI) was used.
To determine cell sizes, differential interference contrast pictures of log-phase cells were taken. The ratio of the length of the long and short cell axis was determined. For average values at least 180 cells per strain were surveyed.