The global transcriptional response to transient cell wall damage in Saccharomyces cerevisiae and its regulation by the cell integrity signaling pathway.

In the yeast Saccharomyces cerevisiae, environmental stress conditions that damage the cell wall lead to activation of the so-called "compensatory mechanism," aimed at preserving cell integrity through a remodeling of this extracellular matrix. Here we used DNA microarrays to investigate the molecular basis of this response to two agents that induce transient cell wall damage; namely Congo Red and Zymolyase. Treatment of the cells with these two agents elicited the up-regulation of 132 and 101 genes respectively, the main functional groups among them being involved in cell wall construction and metabolism. The main response does not occur until hours after exposure to the cell wall-perturbing agent. In some cases, this response was transient, but more sustained in others, especially in the case of the genes involved in cell wall remodeling. Clustering of these data together with those from the response to constitutive cell wall damage, revealed the existence of a cluster of co-regulated genes that was strongly induced under all conditions assayed. Those genes induced by cell wall damage showed an enrichment in DNA binding motifs for Rlm1p, Crz1p, SBF (Swi4p/Swi6p), Msn2p/Msn4p, Ste12p, and Tec1p transcription factors, suggesting a complex regulation of this response together with the possible involvement of several signaling pathways. With the exception of PHO89 and FKS2, none of the genes induced by Congo Red was up-regulated in a slt2 strain. Moreover, characterization of the transcriptional response to Congo Red in a rlm1 mutant strain revealed that only a few genes (i.e. PHO89, FKS2, YLR042C, and CHA1) were induced at least partially independently of the transcription factor Rlm1p, the rest being totally dependent on this transcription factor for their activation. Our findings consistently demonstrate that the cell integrity signaling pathway regulates the cell wall damage compensatory response, mainly through transcriptional activation mediated by Rlm1p.

The cell wall is an external envelope that surrounds yeast cells. This structure is essential for maintaining cell morphology and to protect cells from the external environment by preserving their osmotic integrity. In the budding yeast Saccharomyces cerevisiae, the cell wall is composed primarily of ␤-1,3 glucan, ␤-1,6 glucan, chitin, and mannoproteins (1,2,3). These components are linked to each other as macromolecular complexes in which ␤-1,6 glucan acts as a cross-linker, being attached to the major components, ␤-1,3 glucan and mannoproteins, and occasionally to chitin (4,5). ␤-1,3 glucan is covalently linked to chitin to make up the inner cell wall, which is the main element responsible for the mechanical strength of this structure. Mannoproteins form the outer cell wall layer. Two main classes of proteins are coupled covalently to cell wall polysaccharides: GPI-dependent cell wall proteins (GPI-CWPs), 1 which are generally linked to ␤-1,3 glucan through a ␤-1,6 glucan chain, and PIR proteins (PIR-CWPs), which are directly linked to ␤-1,3 glucan (6,7). For some of these cell wall mannoproteins, several functions have been characterized, including adhesion and cell wall construction (for a review see Ref. 3), although in many cases their precise function is unknown.
The cell wall cannot be a static structure since it needs to adapt to its surrounding growth conditions as the cell increases in size, and it also needs to be remodeled during morphogenetic processes such as mating, sporulation, or pseudohyphal growth (2,8,9). Moreover, in its attempt to survive the cell wall can change in composition and/or structure in response to environmental stress (3,10). The existence of this cellular response, which has been called "compensatory mechanism," clearly illustrates the dynamic nature of the cell wall. This cell wall salvage response involves: (i) a remarkable increase in the chitin content; (ii) changes in the association between cell wall polymers (while only 2% of CWPs are linked directly to chitin in a wild-type cell, this linkage is 20-fold more abundant in gas1 cells) (11); (iii) an increase in the bulk of CWPs; and (iv) a transient redistribution of ␤-1,3 glucan synthase complex throughout the cell (12). All these mechanisms are involved in the repair of the cell wall under stress conditions.
Although we are now beginning to understand some aspects of the regulation of cell wall construction, exactly how all these processes are regulated is still largely unclear. In addition to the activation of the SLT2-MAPK pathway in response to several environmental stimuli, such as high temperature (13), hypo-osmotic shock (14) and mating pheromones (15), this pathway becomes also activated in response to cell wall damage. For example, treatments with cell wall-perturbing agents such Congo Red and Calcofluor White or Zymolyase, which degrades the ␤-1,3 glucan network, lead to activation of the Slt2p MAP kinase (16,17). Additionally, mutant cells with a weakened cell wall, such as fks1⌬, gas1⌬, mnn9⌬, and others, show a constitutive activation of this pathway (18,19). Several cell membrane proteins (Mid2p, Mtl1p, and Wsc1-4p) (16,20,21), should act as sensors for cell wall damage, activating the MAP kinase cascade (22). The phosphorylated Slt2p finally activates several transcription factors, leading to changes in gene expression. Currently, two transcription factors have been identified as targets of Slt2p: the MADS-box transcription factor Rlm1p (23)(24)(25) and SBF (26). This latter is a heterodimeric complex of two proteins, Swi4p and Swi6p, that are involved in the activation of gene expression at the G 1 /S transition (27). Signaling through Rlm1p regulates the expression of many genes involved in cell wall biogenesis (28). Other works have also involved the Sho1p-Kss1p (29) and the HOG1-MAP kinase pathways (3,30) in the control of the cell integrity. More recently, using DNA macroarray technology we obtained a comprehensive view of the cellular response to constitutive cell wall damage by comparing the transcriptome of mutant cells deleted in several genes important for the construction of the cell wall (KRE6, GAS1, KNR4, FKS1) with the expression profiles of wild-type cells (19). In addition to confirming the involvement of the SLT2-MAP kinase pathway, this genomewide analysis also led to the identification of the Calcineurin/ Crz1p signaling pathway and the regulatory machinery from the general stress cellular response as regulators of the response to cell wall damage.
Although the genomic characterization of the transcriptional response to cell wall mutations has allowed us to decipher relevant molecular aspects of the cell wall compensatory mechanism (19), important clues regarding the kinetics and regulation of the response were lacking. In order to obtain additional insight into these aspects, we used an alternative approach involving the characterization of the transcriptional yeast response to the presence of two drugs, Congo Red and Zymolyase, which are known to cause cell wall damage and activate the Slt2p-MAP kinase pathway (17). The global transcriptional response to transient cell wall damage demonstrates that the larger fraction of genes induced are devoted to cell wall remodeling in order to ensure cell integrity under cell wall damageinducing growth conditions. The direct role of the cell integrity pathway in the response was analyzed using mutant strains in some elements of this pathway. The expression profiles of strains deleted in SLT2 and RLM1 disclosed the essential contribution of these proteins and their corresponding signaling pathway to the regulation of the response.
Culture Conditions-For routine cultures, S. cerevisiae was grown on YEPD (1% yeast extract, 2% peptone, 2% glucose). When necessary, Geneticin (200 mg/liter) (Invitrogen) was added. Yeast cells were grown overnight at 24°C to an optical density of 0.8 -1 (OD 600 ). The culture was refreshed to 0.2 OD and grown at 24°C for 2 h 30 min. Next, the culture was divided into two parts. One part was allowed to continue growing under the same conditions (the non-treated culture) while the other one was supplemented with Congo Red to a final concentration of 30 g/ml. The severity of the treatment was calibrated to preserve more than 90% of cell viability. The mechanism of action of this drug is not known, although its interference with proper cell wall assembly has been well documented (31). At this time, cells (ϳ5 ϫ 10 8 ) were collected by centrifugation (t ϭ 0 h sample). Then, for the wild-type strain cultures cells were collected by centrifugation at 0.5, 2, 4, and 6 h of growth, frozen at Ϫ80°C, and processed for RNA extraction (see below). For the rlm1⌬ and slt2⌬ strains, cells were collected at 4 h. For the Zymolyase experiments, the procedure was identical but the medium was supplemented with 5 units/ml of Zymolyase100T (ICN Biomedicals Inc, Aurora, OH), and the times studied were 0 and 2 h.
Preparation of Yeast Extracts and Western Blotting-Cells were grown, treated with Congo Red, and collected as described above. Lysis, collection of proteins, fractionation by SDS-polyacrylamide gel electrophoresis and transfer to nitrocellulose membranes, as well as the detection of phosphorylated Slt2p, using an anti phospho-p44/p42 MAPK (Thr 202 /Tyr 204 ) antibody (New England Biolabs, Beverly, MA) were performed as described by Martín et al. (32). Mouse anti-actin monoclonal antibody (ICN Biomedicals Inc) was used at a dilution of 1:2000 to monitor the levels of actin.
RNA Isolation-Total RNA was isolated from exponentially growing cells (5 ϫ 10 8 ) by the "mechanical disruption protocol" using the RNeasy MIDI kit (Qiagen, Hilden, Germany), following the instructions of the manufacturer. RNA concentrations were determined by measuring absorbance at 260 nm. RNA purity and integrity were assessed using RNA Nano Labchips in an Agilent 2100B Bioanalyzer (Agilent Technologies, Palo Alto, CA) following the manufacturer's instructions.
cDNA Synthesis and Chip Hybridization-cDNA was synthesized from 25-30 g of total RNA by reverse transcription, using the CyScribe TM First-Strand cDNA Labeling Kit, incorporating Cy3-dUTP or Cy5-dUTP (Amersham Biosciences) into the cDNA corresponding to each sample to be compared. Labeled cDNA was dried in a vacuum trap and used as a hybridization probe after resuspension in 45 l of hybridization solution (50% Formamide, 6ϫ SSC, 0.5% SDS, 5ϫ Denhardt's, 20 g of poly(A) (P-9403, Sigma) and salmon sperm (Invitrogen, 100 g/ ml). The printed microarrays, including the complete set of 6306 ORFs coded by the S. cerevisiae genome, used in this study were provided by the Microarray Centre of the University Health Network, Toronto, Canada (www.microarrays.ca). Slides were prehybridized in prehybridization solution (6ϫ SSC, 0.5% SDS, 1% bovine serum albumin (A-7906, Sigma) for 1 h and were then hybridized overnight with labeled probe at 42°C in a water bath. Before scanning, the chips were washed and dried in a Lucidea TM SlidePro Automated Hybridization Station (Amersham Biosciences). For each condition tested, comparison of treated and non-treated samples, the total RNA from two different cultures was analyzed, and, in addition, for each sample two different hybridizations were performed, including fluorochrome swapping in order to minimize transcriptional changes due to technical variability. This therefore corresponded to four DNA microarrays analyzed for each condition.
Microarray Image Analysis-Microarrays were scanned with a GenePix 4000B scanner (Axon Instruments, Union City, CA) at a resolution of 5 m (PMT values ranging from 550 to 700 and laser power 100%). GenePix Pro 4.0 analysis software (Axon Instruments, Union City, CA) was used to locate spots in the microarray with the appropriate grid and to obtain the two Cy3/Cy5 image TIFF files. All images were further processed using GenePix 4.0 software according to the manufacturer's instructions.
Data Processing and Statistical Methods-Flagged spots and spots with an average intensity minus background below the mean of the background (considering this as the median of the local background for each spot) for all the non-flagged spots in any of the channels (Cy3 or Cy5) were not retained for further analysis. Within this group, the spots showing in one channel a value of intensity minus background higher than 5 times the mean of the background for all spots in the microarray for that channel were recovered. The reproducibility of replicates in each microarray (two spots per ORF) was analyzed by creating a normal distribution log 2 (R A ϫ 1/R B ), where R A and R B are the ratios of replicated spots after background subtraction and normalization. Replicates that exceeded the average by more than Ϯ3 S.D. were discarded. Data analysis was accomplished using the GeneSight 4.0 software package (BioDiscovery Inc, El Segundo, CA). Locally weighed linear regression (lowess) analysis was performed as a normalization method to remove the intensity-dependent deviation in the log 2 (ratio) values. Significance analysis of the results was conducted using a Student's t test (GeneSight). Genes with p values of less than 0.05 were considered to be significantly differentially expressed with respect to the treatment condition. The Sotarray Server from GEPAS (Gene Expression Pattern Analysis Suite) (gepas.bioinfo.cnio.es/tools.html); Bioinformatics Unit, CNIO, Spain) was used to perform the SOTA clustering and to visualize the resulting tree.
Promoter Analysis-For each gene induced at least 2-fold under the different conditions tested, 800-bp upstream from their start codon (ATG) were analyzed using the MEME algorithm (meme.sdsc.edu/ meme/website/intro.htm), the Regulatory Sequence Analysis Tools (rsat.ulb.ac.be/rsat) and Reduce (lognormal.bio.columbia.edu/reduce/), to identify possible common DNA motifs. Regulatory Sequence Analysis Tools were also used to search for specific sequence motifs.

Global Analysis of the Transcriptional Response to Transient Cell Wall Damage Caused by Congo Red Treatment in Yeast-
The aim of this study was to characterize, in a comprehensive way, the response of the yeast Saccharomyces cerevisiae to the presence of transient cell wall damage and how this response is regulated. To achieve this, we exposed yeast cells, in the logarithmic growth phase, to 30 g/ml of Congo Red (CR) and DNA microarrays were used to analyze the gene expression of 6306 open reading frames encoded by the yeast genome. We characterized the pattern of gene expression at time 0, 2, 4, and 6 h after exposure of the yeast to CR. The time-course experiments were performed by comparing the abundance of mRNA at each time in cells growing in the presence of CR relative to yeast cells growing in the absence of the drug (see "Experimental Procedures" for details). A complete data set of the genes induced in our study is shown in Table I. The analysis revealed that 132 genes were induced at least 1.8-fold at one or more of the time points tested (67 genes after 2 h, 58 genes after 4 h, and 75 genes after 6 h).
We then distributed the significant responsive genes into functional categories according to YPD/SGD classification. As shown in Table I, the main functional groups of up-regulated genes belong to those involved in cell wall organization and biogenesis, metabolism and the generation of energy, together with unknown genes. From the large group of 34 cell wallrelated genes, four different sets can be highlighted. The first subgroup comprises 15 genes out of the 60 -70 genes encoding GPI-cell wall proteins (GPI-CWPs) so far identified in the yeast genome (34). Of special relevance are proteins such Cwp1p, which has been implicated in cell wall strengthening (35); Pst1p, which is secreted from protoplasts under conditions of cell wall regeneration (36); and Crh1p and Crh2p, two putative transglycosidases involved in the cross-linking between yeast cell wall polymers (37,38). A second set in this group includes PIR1, PIR2/HSP150, PIR3, and PIR4/CIS3, encoding for PIR-CWPs. The products encoded by these genes are directly linked to ␤-1,3 glucan (6). Deletion of all four genes results in swollen cells that are sensitive to Calcofluor White and Congo Red (6) while overexpression of PIR2 confers resistance to the antifungal protein osmotin (39). A third subgroup encompasses four genes involved in chitin synthesis: namely, GFA1, encoding the glutamine-fructose-6-phosphate amidotransferase, an enzyme involved in the first steps of the chitin synthesis pathway (40); CHS3, coding for the enzyme responsible for the synthesis of the bulk of chitin, both at the lateral cell wall and at the ring between mother and daughter cells (41); CHS1, which contributes to repairing mechanisms during cytokinesis (42); and PCM1, which is involved in the biosynthesis of N-acetylglucosamine (43). The presence of these genes is consistent with the finding that cell wall damage causes an increase in chitin deposition (44). The fourth set of cell wall related genes upregulated under our conditions contains those involved in the biosynthesis and degradation of ␤-1,3 and ␤-1,6 glucan. This group includes FKS1, FKS2, KRE6, KRE11, and genes encoding exoglucanase (EXG1 and EXG2) (45) and endoglucanase/ glucanosyl transferase activities (BGL2) (46). Finally, KTR2, encoding for a mannosyltransferase of the KRE2 family (47), and PSA1, coding for the mannose-1-phosphate guanyltransferase (48), both required for the biosynthesis of mannoproteins, were also up-regulated.
The induction of the functional cluster of genes related to the metabolism of carbohydrates, amino acids, and sphingolipids was not unexpected. In particular, active remodeling of the cell wall would require a considerable degree of carbohydrate mobilization. Prolonged treatment with CR (6 h) also leads to the activation of genes involved in phosphate metabolism.
Interestingly, we found a cluster of genes related to the SLT2-MAP kinase pathway that were clearly induced at 2-4 h after addition of the drug (Table I), the gene most strongly induced within this cluster being MLP1 (YKL161C). The role of MLP1 in the cell integrity pathway has not been yet elucidated, but it is known to encode a protein kinase-like homologous to Slt2p that interacts with the Rlm1p transcription factor (24).
In contrast to the transcriptional induced response, 111 genes were repressed under the CR conditions studied, most of them (93 genes) being repressed after prolonged treatments (6 h). One of these genes was SWI6 that encode for one of the transcription factors of the SBF complex. The data concerning expression for the repressed set of genes are shown in Table II,   In order to validate the expression profiles obtained by microarray analysis, real-time quantitative RT-PCR was performed. For this analysis, a subset of genes including different levels of induction was selected. The same RNAs used in the microarray experiments to analyze the expression profile after 4 h of CR treatment were used for the Q-RT-PCR assays. As shown in Table III, in general we found a very good correlation between both data sets, supporting the validity of the microarray experiments. Additional detailed information of the complete data sets of microarray experiments is available on the World Wide Web at www.ucm.es/info/mfar/.
Kinetics of the Cellular Response to Cell Wall Damage-In contrast to many other environmental stresses that activate rapid cellular responses, the response to cell wall damage does not occur until hours after exposure to the cell wall-perturbing agent. The global transcriptional response 30 min after the addition of CR revealed only a small set of genes that were significantly activated, including YLR194C, YIL117C (PRM5), YDR055W (PST1), and YJL171C, all of them being highly up-regulated at longer times. Detailed analysis of the CR temporal expression pattern (Table I) revealed that the global response to cell wall stress was transient in some cases, but more sustained in others. About 50% of the up-regulated genes were induced at 2 h after the start of drug treatment, 34% of them, maintaining a sustained response until 6 h. In contrast, 34% were only induced at 2 h, and 31% at 2 to 4 h. In addition, one group of genes (12%) was induced only at 4 h and, finally, a set of about 38% of differentially expressed genes was induced late in the response (6 h).
Global Analysis of the Transcriptional Yeast Response to Zymolyase Treatment-To test whether other types of cell wall damage might lead to a similar transcriptional response, we next analyzed the expression profile of yeast cells growing in the presence of 5 units/ml of Zymolyase100T for 2 h. Zymolyase treatment affects cell wall integrity owing to the presence of ␤-1,3 glucanase and mannanase activities in its composition. The response to this drug included the up-regulation of 101 genes (Table I) (49), SPI1 encoding a GPI-CWP, ECM13 (50); and PAU6 coding for a PIR-CWP. In common with both drug treatments was the up-regulation of genes involved in the Slt2p-signaling pathway, indicating a major role of this mechanism under these conditions. Some of the significant differences between both stimuli were found in the groups of genes related to metabolism and in genes of unknown function (Table I).
Promoter Analysis of the Up-regulated Genes-We reasoned that the identification of multiple transcription factor binding sites in the promoters of genes commonly induced under conditions eliciting cell wall damage should give an idea of the inputs coming from different regulatory signaling cascades involved in the regulation of this response. Using the RSAT program (51), we analyzed DNA binding motifs in the 800 bp non-coding upstream sequences of genes that were up-regulated at least 2-fold by CR at all the time points studied, and by Zymolyase. The algorithm used by this program localizes regulatory motifs that appear more frequently than would be expected if they had a random distribution. We also searched for DNA sequences that matched DNA motifs annotated in the databases (52). The most important results from this analysis can be seen in Table IV, showing the percentages of induced genes containing the corresponding DNA consensus motifs compared with their random appearance in the genome. Under all the conditions studied, we found the highest enrichment in genes carrying motifs for Rlm1p, suggesting a major role for the cell integrity pathway mediated by the MAP kinase Slt2p in the regulation of the cell wall compensatory transcriptional response. Thus, as shown in Fig. 2, the levels of the active dually phosphorylated form (Thr 190 -Tyr 192 ) of Slt2p were clearly increased in response to treatment with Congo Red at the different times. Additionally, the significant enrichment in genes bearing motifs for the transcription factors Crz1p (53),  II Classification into functional families of ORFs whose transcripts are down-regulated at least 2-fold in any of the different Congo Red treatments assayed ORFs whose transcripts are repressed at least 2-fold are included. The timing of repression for each gene is indicated in parentheses (see ''Experimental Procedures''). Procedures"). The clustering algorithm arranges the genes according to their similarity in expression under the different conditions, and hence genes with similar expression patterns are clustered together (57). In the present work, this clustering identified two significant groups of co-regulated genes, containing 91 (cluster 1) and 42 genes (cluster 2) respectively (Fig. 3). Within cluster 1, two different sub-clusters were visualized: subcluster 1a contained genes induced under transient conditions and was mainly enriched in cell wall-related genes and genes of unknown function. In contrast, subcluster 1b mainly contained genes involved in cellular metabolism, which were preferentially up-regulated in cell wall mutants and by Zymolyase treatment. Finally, with the exception of the first 6 genes those included in cluster 2 were common to all cell wall damage conditions investigated, both constitutive, due to cell

TABLE IV Promoter analysis of induced genes (ratioϾ2)
The percentages of genes with at least one promoter binding site for the transcription factor indicated are included together with its random appearance in the yeast genome. wall defects, and transient -caused by CR and Zymolyase treatments. Within cluster 2, the twenty genes included in subcluster 2b were strongly induced under all cell wall damage conditions and can therefore be considered the "core" of the transcriptional compensatory response to cell wall damage. It is likely that this group of genes would be essential for adaptation of the yeast to cell wall stress. We next compared our expression profiles with other previously published DNA microarray datasets relevant to our experiments. Around 35-40% of the genes up-regulated under our conditions were induced in cells treated with dithiothreitol (58), a drug that may affect the cell wall by altering cell wall disulfide linkages, or caspofungin (59), a drug that inhibits the activity of ␤-1,3 glucan synthase (see Supplemental Data in Table V at www.ucm.es/info/mfar/).
The expression profiles of cells harboring activated alleles of PKC1 (PKC1-R398A) and RHO1 (RHO1-Q68H) (55) as well as MKK1 (28), components of the MAP kinase Slt2p cell integrity pathway, share many genes (50% of the global response; 75% within the cell wall-related cluster) in common to our profiles (see supplemental data in Table V on the web site for details), indicating a major regulatory role for this pathway in the response. Interestingly, 58 and 45% of the cell wall genes induced under our experimental conditions were also induced in cells overexpressing the transcription factor Tec1p (56), and also after long treatments with pheromones (55), respectively, suggesting additional levels of regulation for genes involved in the transcriptional response to cell wall damage through the pseudohyphal growth and mating MAP kinase pathways. In contrast, although analysis of the promoters of the co-regulated genes suggests a role for the Calcineurin-dependent pathway in the response to cell wall damage, only 25 genes out of those activated in the presence of CR and Zymolyase were found to be up-regulated by the Ca 2ϩ /calcineurin-dependent pathway (53).
Transcriptional Response to Cell Wall Damage in slt2 and rlm1 Mutant Strains-Our current findings strongly support the notion that the regulation of the transcriptional response to cell wall damage mainly depends on the Slt2p-cell integrity pathway. We therefore addressed the involvement of the Slt2p pathway on the regulation of this response by characterizing the transcriptional profiles of slt2 and rlm1 deletant strains grown for 4 h in the presence of CR. Interestingly, we found that the presence of SLT2 is essential for the transcriptional activation induced by CR, except for two genes, PHO89 (2.8-fold) and FKS2 (2-fold). PHO89 encodes a phosphate/Na ϩ co-transporter involved in phosphate uptake under high-phosphate growth conditions (60), while FKS2 codes for the alternative subunit of ␤-1,3 glucan synthase. With respect to the transcriptional profile of the rlm1 strain, surprisingly we found that almost the whole transcriptional response to Congo Red depends on Rlm1p, except for PHO89, FKS2, YLR042C (encoding a GPI-CWP), and CHA1 (encoding for serine-threonine ammonia lyase activity (61)). These results were further verified by Q-RT-PCR. PHO89 was induced in the rlm1 mutant to levels similar to those of the wild-type strain (WT-3.6-fold/rlm1-3.1 fold), while the others showed a partial dependence on Rlm1p for their up-regulation (YLR042C (WT-2.6-fold/rlm1-1.8-fold), FKS2 (WT-3.2-fold/rlm1-1.8-fold) and CHA1 (WT-3.6-fold/rlm1-1.8-fold)). A complete data lists on the experiments described in this section is available on line at www.ucm.es/info/mfar/.
Phenotypic Analysis of Mutant Strains in Genes Involved in the Compensatory Mechanism-Next, we wished to further characterize how the response to cell wall damage is linked to cell wall remodeling by measuring the sensitivity to Congo Red of BY4741-derived strains with deletions in the genes induced in response to CR that belong to the functional clusters of cell wall construction and signaling. The results are shown in Fig.  4. Strains deleted in the genes encoding regulatory proteins displayed the most severe phenotypes. Mutant strains deleted in SLT2, PTP2, or SWI4 were unable to grow in the presence of 100 g/ml of Congo Red, while those deleted in MSG5, CRZ1, and SWI6 were sensitive to the drug. Regarding the genes involved in cell wall construction, the most severe phenotype was due to the absence of FKS1, encoding ␤-1,3 glucan synthase. Deletion of CRH1 (encoding a putative GPI-CWP transglycosidase) or EXG1 (coding for an exo-glucanase) rendered these strains sensitive to CR, while those deleted in PIR2 or BGL2 were slightly sensitive. The rest of the strains were insensitive to the presence of the drug. DISCUSSION Stressful conditions that cause damage to the cell wall, a structure that is essential for cell survival, leads to cellular responses that guarantee the osmotic integrity of the cell and therefore cell survival through the remodeling of this extracellular matrix (10). In a previous work, we identified a specific group of 80 genes that are up-regulated in response to persistent cell wall damage caused by mutations in cell wall-related genes, and we showed that at least three signaling pathways, including the general stress response mediated by Mns2p/ Msn4p, the Ca 2ϩ /calcineurin and the Pkc1p-Slt2p pathways, were involved in the cell wall compensatory response. Here we have expanded our understanding of the cell wall compensatory mechanism by investigating the transcriptomic response to transient cell wall damage caused by CR and Zymolyase, two drugs that specifically alter the integrity of the cell wall, and we show that this response is regulated by the Slt2p-MAP kinase signaling pathway. The kinetics of the "compensatory response," described here for the first time, indicates the existence of a transient response for one cluster of genes and a sustained response for another set of genes. The genes that exhibit transient increases must be involved in the transition to the new environment; for instance, the genes included in the "signal transduction" cluster, which show a transient response peaking at 2-4 h. Genes whose expression changes to a new level and remains altered probably encode proteins with a persisting role under the new conditions. Interestingly, the largest set of genes with a sustained response is included in the family of genes involved in cell wall construction and remodeling.
The transcriptional profile of the response to cell wall damage clearly reflects the necessities of remodeling of the cell wall under these circumstances. The alterations in the cell wall caused by CR could be compensated by changes in the balance among the different cell wall components or by changes in the type of association between them. It is very likely that the group of cell wall-related genes up-regulated under the condi-  fks1, gas1, knr4, kre6, and mnn9) (19), represented in each column, is shown. Each row represents the expression ratios for a particular gene under the conditions listed. Clustering was obtained by the Sotarray Server from GEPAS (gepas.bioinfo.cnio.es/tools.html). Red represents expression ratios higher than 1, and green indicates ratios lower than 1. The degree of color saturation represents the amount of the expression ratio, as indicated by the scale bar. White denotes missing values.
tions studied here would play a direct role in compensating for transient cell wall damage. Although the function of many of these proteins remains unknown, they must play a role in cell wall strengthening and remodeling. An interesting issue is that despite the high transcriptional activation of some of these genes encoding CWPs their deletion mostly leads to mild phenotypes of sensitivity to CR. It is probable that the simultaneous deletion of several genes working in a coordinated fashion would be required for the complete impairment of cell wall remodeling that takes place under cell wall damage conditions. Alternatively, this may be explained, at least in part, by the existence of redundant functions for some of these CWPs. For example, the simultaneous deletion of CRH1 and its homologous CRH2 renders this strain unable to grow in CR (37). On the other hand, double disruptants of dfg5 (up-regulated under our conditions) and dcw1, coding for homologous mannosidases, are synthetically lethal (62).
Does the yeast cell respond to different cell wall defects in a specific manner? Although there are differences in the transcriptional profiles induced by different stimuli leading to cell wall damage, the main core of the response is similar. Comparison of the transcriptional profiles under conditions of constitutive damage resulting from mutations at different steps of cell wall construction (gas1⌬, knr4⌬, fks1⌬, kre6⌬, mnn9⌬) (19) with the transient damage caused by Congo Red and Zymolyase indicates that, besides some differences between them, the genes included in the functional groups involved in cell wall remodeling (cell wall biogenesis, morphogenesis, signal transduction, and stress) under transient cell wall damage conditions are also up-regulated in at least one of the mutants previously characterized. It should be noticed that the transcriptomic profile of the gas1 mutant was the most similar to those identified in response to CR, reflecting the fact that this mutant exhibits the most pronounced cell wall defects (63). The hierarchical clustering of the transcriptional profiles of transient and constitutive cell wall damage permitted us to define a cluster of 20 genes that are strongly induced under all the conditions assayed, and would represent the main transcriptional fingerprint for cell wall stress. This cluster includes genes involved in cell wall biogenesis (CWP1, SED1, PIR3, FIG. 4. Analysis of sensitivity to Congo Red. Several yeast strains deleted in genes up-regulated in the microarray experiments were tested for their ability to grow on solid medium in the presence of 100 g/ml of Congo Red (see "Experimental Procedures" for details). Data in each column correspond to different experiments and therefore wild-type BY4741 and slt2 (sensitive to CR) strains were included for comparison. Strains sensitive to the drug are underlined. CRH1, KTR2, GFA1, PST1, and YLR194C), signaling (SLT2, MLP1), metabolism (YPL088W, YPS4, FBP26), stress (HSP12, PRM5), and genes with unknown function (SRL3, YIL023C, YLR414C, YHR097C, YAL053W). The inclusion of this latter group in the cluster is very interesting since they can now be annotated as being involved in the compensatory response to cell wall damage.
Characterization of the expression profiles of cells deleted in SLT2 growing in the presence of CR (described here) indicates that the compensatory response activated in the presence of transient cell wall damage is totally regulated through the MAP kinase Slt2p, except for PHO89 and (although to a lesser extent) for FKS2. Interestingly, up-regulation of both genes in the presence of CR is dependent on the Crz1p transcription factor, 2 indicating that these genes are regulated through the Calcineurin pathway. Moreover, our data indicate that nearly all the response is mediated through Rlm1p, the MADS-box transcription factor acting downstream from the MAP kinase Slt2p, with the exception of PHO89, FKS2, YLR042C, and CHA1, which should be regulated by other different transcription factors. Previous reports have shown that the induction of gene expression under artificial conditions and by an increase in temperature is dependent on R1m1p, with the exception of the FKS2 gene (28). Our results further underscore the importance of Rlm1p-dependent regulation in the response to temperature shift to the transcriptional response to cell wall damage. Furthermore, our experiments also indicate that in the absence of Rlm1p there is no induction of an alternative transcriptional response. This result is surprising looking at the phenotype of a strain deleted in RLM1. In contrast to a mutant in SLT2, which is unable to grow under cell wall damage conditions, deletion of RLM1 does not lead to defects in cell integrity; and even rlm1 strains are resistant to treatments with Zymolyase or Calcofluor White (24). These phenotypes suggest that Rlm1p is not the only transcriptional targets of Slt2p. However, our findings indicate that Rlm1p mediate the majority of the effects of Slt2p on transcriptional activation under conditions of cell wall damage and suggest that the possible functional link between the cell integrity pathway and the transcription factors of other signaling pathways must be regulated by a Rlm1p-dependent circuit. As previously suggested, Slt2p/Mpk1p may have additional targets that do not affect gene expression (28).
Therefore, in contrast to the possible participation of many signaling pathways in the regulation of the response to cell wall damage as inferred from the promoter analysis of co-induced genes under the conditions studied here, the transcriptional profiles of slt2 and rlm1 mutant strains indicate that the cell integrity pathway, through SLT2 and RLM1, is the main element responsible for this regulation. This could be explained by: 1) the existence of different signaling pathways regulating the set of genes related to the cell wall compensatory response depending on the stimuli; 2) the existence of additional signaling pathways regulating the response to cell wall damage in a Rlm1p-dependent way. For example, several of the Kss1p-regulated genes may participate in cell wall remodeling during filamentous growth, being regulated specifically by Tec1p under these conditions and by Rlm1p under conditions of cell wall damage. Thus, it has been shown by Gasch et al. (58) that genes involved in the environmental stress response (ESR) can be regulated by different transcription factors, depending on the specific environmental insult. Some genes of the ESR are induced through Msn2p/Msn4p in response to heat shock but through Yap1p in response to H 2 O 2 (58). In accordance, despite the presence of several Crz1p binding sites in some of the promoters of the genes related to cell wall construction that are up-regulated in the presence of CR (see Supplementary Table V for details), none of them (except FKS2) is dependent on Crz1p. 2 The real participation of other signaling pathways, additionally to the cell integrity pathway, in the regulation of the compensatory response to cell damage must await further investigation.