INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The cell wall preserves the osmotic integrity of fungal cells and determines cellular morphology during various developmental processes. In the budding yeast Saccharomyces cerevisiae, the cell wall must exert its protective role during budding, mating, sporulation, and pseudohyphal growth, showing different morphogenetic patterns that obey diverse environmental conditions. The rapid growth of buds or mating projections, for instance, suggests that the accurate dynamics that rule the development of such an apparently rigid structure may allow a certain flexibility, at least in growing areas, without altering the protective function of the cell wall.

The budding yeast cell wall is basically constituted by highly mannosylated proteins (mannan) and three kinds of polysaccharide chains: (i) the predominant, linear 1,3--glucan, (ii) a minor, highly branched 1,6--glucan component, and (iii) chitin. Glucans make up about half of the dry weight of the cell wall, while chitin accounts for only 1 to 2%. Although cells can survive when deprived of the bulk of chitin (45), this N-acetylglucosamine polymer is essential for the formation of a proper primary septum during cytokinesis and for the characteristic ring-like structures at the mother-bud neck that leave a scar on the mother cell wall after cell division (45). Chemical treatment of isolated yeast cell walls yields an alkali-soluble fraction, rich in 1,3--glucan, and a major acid- and alkali-insoluble -glucan fraction that also retains most of the 1,3--glucan (8). This insolubility has been explained in terms of its covalent linkage to chitin (15), and the nature of this linkage has been found to be a 1,4--glycosidic bond (26).

Careful biochemical approaches have provided evidence that the different components of the yeast cell wall are covalently linked in vivo as macromolecular complexes, constituting what has been called the "flexible building block," in which mannoproteins would be linked by the remnant of their glycosylphosphatidylinositol (GPI) anchor moieties to 1,6--glucan, which in turn is bound to 1,3--glucan core chains (23, 27). According to the model proposed by these authors, chitin could bind either 1,3- or 1,6--glucan, thus completing the hypothetical block. Several of the connections between these components through various types of glycosidic linkages are now known in detail (26, 27). A stoichiometric relationship of the cell wall components in such a basic unit is hard to assess and probably varies in different areas of the wall and under different morphogenetic programs.

A vast array of valuable information has been gathered on the genetic and biochemical characterization of the enzymes responsible for cell wall biosynthesis, including glucan and chitin synthases, chitinases, and exo- and endo--glucanases (for recent reviews, see references 29 and 36). However, the involvement of these enzymes in the construction and maintenance of the cell wall in vivo remains mostly obscure. In particular, little is known about which enzymes are responsible for cross-linking between cell wall polymers. Cross-linking among mannoproteins, -glucans, and chitin may indeed account for the plastic properties of the cell wall during different developmental processes. A critical step for obtaining further insight into the mechanisms involved in cell wall assembly will be to identify these cross-linking enzymes, determine how these proteins work, and find out how the cell wall architecture is modified during the yeast cell cycle.

With the yeast genome sequenced, candidates can be sought from among predicted proteins that bear sugar-binding consensus sites or display a certain degree of homology with known transglycosidases from other organisms. Here we report the existence of a novel family of three putative yeast glycosidases. Characterization of strains with knockouts of these genes suggests their participation in cell wall assembly. Cell cycle expression studies and localization of the gene products at different development stages suggest a specific temporal and spatial role for each gene in cell wall construction.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains and growth media. All experiments were performed with the S. cerevisiae FY1679 strain (MATa/ ura3-52/ura3-52 his3200/HIS3 leu21/LEU2 trp163/TRP1 GAL2/GAL2) and haploid derivatives. To assess the mating type of the segregants, they were mated to strains 1783 (MATa ura352 leu2-3,112 trp1-1 his4 can1) and 1784 (MAT ura352 leu2-3,112 trp1-1 his4 can1). The FY1679 ygr189c::KanMX4 disruptant strain was obtained previously (40). Strains 192-37d (MATa chs1 leu2 trp1 ura3 chs2::LEU2) and HVY244 (MAT ura3 his1 leu2 chs3::LEU2) were kindly provided by A. Durán. For routine cultures, S. cerevisiae was grown on YED (2% yeast extract and 2% glucose) or YEPD (YED plus 2% peptone). When required, glucose was replaced by other carbon sources in the same proportion. To induce sporulation, diploid cells were grown for 24 h in solid presporulation medium (5% glucose, 1% yeast extract, 3% meat extract) and then for 5 days in solid sporulation medium (1% potassium acetate plus sufficient amounts of histidine, tryptophan, leucine, adenine, and uracil). To synchronize cultures with mating pheromone, -factor (Sigma) was added to exponentially growing cells to a final concentration of 10 µg/ml and incubated for 3 h. After this time, cells were collected, washed, resuspended in fresh YEPD, and incubated so as to resume growth. Both shmoo formation and culture synchronicity in the first cycle after release from pheromone-induced arrest were followed by phase-contrast microscopy. The Escherichia coli strain used as the plasmid host was DH5 (supE44 lacU169 (80 lacZM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1). For selective growth, bacteria were grown on Luria-Bertani (LB) medium containing 100 mg of ampicillin/liter.

Yeast genetics and phenotypic analyses. Tetrad analyses were performed by standard micromanipulation procedures. Markers of the segregants were verified on SD (20 g of glucose/liter, 1.67 g of yeast nitrogen base without amino acids/liter, 5 g of ammonium sulfate/liter, and the appropriate amount of amino acids) plates lacking a particular amino acid. The sensitivity or resistance of the segregants to Geneticin (encoded by the KanMX4 module) was studied on YEPD plates containing 200 mg of Geneticin/liter. Calcofluor white (fluorescent brightener 28; Sigma) and Congo red (Merck) sensitivities were tested by spotting cells onto plates. Cells were grown overnight in YED or SD-Ura and adjusted to an optical density at 600 nm (OD₆₀₀) of 0.13 (approximately 2 × 10³ cells per µl). Five microliters of samples plus two serial 1/10 dilutions were deposited on the surfaces of solid media supplemented with various concentrations of Calcofluor white or Congo red. Growth was monitored after 2 to 3 days at 28°C.

Molecular biology techniques. Standard molecular biology techniques for DNA manipulations and bacterial transformations were used as described elsewhere (43). Restriction enzymes were provided by Boehringer-Mannheim. In the construction of YEL040w and YLR213c deletant strains, the complete open reading frame (ORF) was deleted, except for the start and stop codons. Disruptions were performed by the SFH (short flanking homology) PCR technique (50), which allows the replacement of the target ORF by a selection marker. For YEL040W deletion, the HIS3 marker from plasmid pFA6a-His3MX6 (see Table 2) was used, and for YLR213c deletion, we used the LEU2 marker from the YEplac181 plasmid (see Table 2). The oligonucleotides devised for this purpose are listed in Table 1. S1 primers bear 41 to 42 nucleotides homologous to the upstream sequence of the target ORF plus 18 to 19 nucleotides of the marker module. S2 primers bear 41 nucleotides downstream from the target ORF plus 19 nucleotides of the marker module. The SFH deletion cassettes were obtained by using the Expand High Fidelity PCR System (Boehringer Mannheim). ORF replacements were identified by PCR with the help of diagnostic primers which bind either outside the target ORF (V2 and V3) or within the selection marker (V1 and V4). PCRs were carried out with the primer pairs V1-V2 and V3-V4 using Biotaq DNA Polymerase (BIOLINE).

RNA analysis. Total RNA was isolated from exponentially growing cells in YEPD medium by the acidic phenol method as described previously (2). Samples were also taken from YEPG medium (YEPD in which glucose had been replaced by galactose) under the same conditions and from sporulation medium after 24 h of incubation. Samples were run in a morpholinepropanesulfonic acid (MOPS)-formaldehyde agarose gel and transferred by capillarity to a Nytran nylon membrane (Schleicher & Schuell) as described previously (43). The transferred RNA was cross-linked to the membrane by a 5-min exposure to UV light at 254 nm and hybridized with [-³²P]dCTP (Amersham)-labeled probes. A 1.2-kb SpeI/EcoRV DNA fragment containing the YGR189c coding region, a 1.3-kb DraI/SpeI DNA fragment containing the YEL040W coding region, and a 1.4-kb BamHI/ClaI DNA fragment containing the YLR213c coding region were used as probes to detect their corresponding transcripts. As a reference for normalization of the procedure, a 1.6-kb BamHI/HindIII DNA ACT1-containing fragment was used as a control probe. A Fluor-S MultiImager (Bio-Rad) was used to quantify the radioactive signal on autoradiograms.

Construction of GFP fusions. In order to create fusion proteins of both Ygr189c and Yel040w with the green fluorescent protein from Aequorea victoria (GFP), we took advantage of a unique SpeI site that exists in frame in both genes. The GFP cassette (733 bp) was synthesized by PCR using oligonucleotides N-GFP and C-GFP (see Table 1). The cassette was flanked by two SpeI restriction sites, which were used to insert it in frame into both genes. A triplet coding for proline before and after the GFP-coding sequence was included. The pFA6b-GFP(S65T)-KanMX6 plasmid (Table 2) was used as a template in PCR.

Site-directed mutagenesis of YGR189c. Plasmid pJV89E, containing the YGR189c ORF (Table 2), was used as a template for PCR. Primers were designed to achieve the changes D136N and E138Q. Briefly, two PCRs were run in parallel using the primers A1-A2 and A3-A4 (see Table 1). Both the A1-A2 (442 bp) and A3-A4 (324 bp) PCR products were utilized as overlapping templates in a second PCR with A1 and A4 as external primers. Thus, a final product of 742 bp was generated, corresponding to a mutated internal fragment of the ORF. Finally, a 481-bp MamI/SpeI fragment from this PCR product was verified by sequencing and subcloned into MamI/SpeI-cleaved pJV89E, thus replacing the wild-type sequence with the mutant sequence.

Confocal-microscopy techniques. Cells were grown overnight in YED (bearing pJV40G) or SD-Ura (bearing pJV89G) and then transferred to fresh medium. After 3 h of incubation, they were harvested by gentle centrifugation, washed twice with phosphate-buffered saline (PBS), and finally resuspended in PBS. For chitin-staining experiments, wheat germ agglutinin-tetramethyl rhodamine isocyanate (WGA-TRITC) (Molecular Probes, Eugene, Oreg.) was used according to the manufacturer's instructions. Samples were observed with an Eclipse TE-300 (Nikon, Tokyo, Japan) microscope attached to a Bio-Rad (Hampstead, United Kingdom) MRC1024 confocal system. Experiments to monitor GFP fusion protein localization over time were performed as above, but samples were mounted on thin SD agar or YED layers, as previously described (20), and time lapse confocal images were taken. When required, cultures were previously synchronized with -factor (Sigma) as described above.

Cell wall fractionation experiments. To collect cell walls, 100-ml cultures in YED medium were grown to an OD₆₀₀ of 0.8 (approximately 10⁷ cells per ml). Cells were washed twice with phosphate buffer, pH 8.5, resuspended in 1 ml of the same buffer, broken in a Fastprep fp120 (Bio 101) with the aid of 1-mm-diameter glass beads according to the manufacturer's instructions (cell breakage was verified by phase-contrast microscopy), and centrifuged. The resulting pellet was washed five times with phosphate buffer, pH 8.5, resuspended in the same buffer containing 0.02% sodium azide, and treated with pronase (Boehringer Mannheim) for 16 h at 35°C with gentle shaking. After centrifugation of this material, the pellet thus obtained was washed twice with phosphate buffer, resuspended in 0.5 ml of the same buffer, heated at 100°C for 10 min, and finally washed again twice. This was considered to be the whole protein-depleted cell wall glucan. To extract the alkali-soluble fraction, this residue was resuspended in 1 ml of 1 M KOH and heated at 60°C for 30 min. The supernatant was then neutralized with acetic acid and precipitated with 2 ml of ethanol. The pellets corresponding to both the soluble and insoluble fractions were washed three times with phosphate buffer. Then the insoluble fraction was resuspended in 1 ml of 0.5 M acetic acid and maintained at 100°C for 3 h in order to extract an alkali-insoluble, acid-soluble fraction. The supernatant was neutralized with KOH, and this fraction was precipitated with 2 ml of ethanol. Both the acid-soluble and -insoluble fractions were washed three times with phosphate buffer. Finally, the pellets of all three (alkali-soluble; alkali-insoluble, acid-soluble; and alkali-insoluble, acid-insoluble) fractions were allowed to dry in a vacuum system, and their dry weights were determined.

Chitin determination in cell walls. Chitin levels were measured as described previously by Kapteyn et al. (24).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The S. cerevisiae genome has three ORFs that code for homologous putative glycosidases. In previous studies, the YGR189c gene was sequenced (1) and its deletion showed it to be nonessential (40). A search in the databases yielded two predicted protein sequences from the S. cerevisiae genome that showed significant degrees of similarity with the deduced YGR189c gene product (Fig. 1A): the YEL040w (29% identity and 46% similarity in 408 amino acids) and YLR213c (34% identity and 49% similarity in 227 amino acids) gene products. In turn, Ylr213c and Yel040w share 39% identity and 56% similarity along 324 amino acids. All three proteins have an N-terminal secretion signal for their incorporation into the secretory pathway. Furthermore, the proteins encoded by YGR189c and YEL040w, but not YLR213c, also display a potential carboxy-terminal domain for GPI anchor attachment (5, 14), a characteristic of many glycoproteins that are targeted to the cell wall, and a C-terminal Ser- and Thr-rich region, a typical feature of several heavily O-glycosylated cell wall or periplasmic proteins.

View larger version (63K): [in this window] [in a new window]   FIG. 1.   (A) Amino acid alignment of homologous Ygr189c, Yel040w, and Ylr213c sequences. Identical and conserved residues are boldfaced and are indicated by asterisks and dots, respectively. The putative catalytic domain similar to that of the 1,3-1,4--glucanases from Bacillus licheniformis is boxed. (B) Amino acid alignment of homologous sequences within the proposed catalytic domain of bacterial endo--1,3-1,4-glucanases (48), corresponding to Ygr189c, Yel040w, Ylr213c, Kre6, and Skn1 from S. cerevisiae, B. licheniformis 1,3-1,4--glucanase (BACLIC), and A. thaliana xyloglucan endotransglycosidase (TCH4). A single asterisk identifies the proposed catalytic nucleophile (18, 21), and two asterisks indicate the position of the proposed general acid-base residue (49). Amino acids that are identical in all the sequences are boxed. Boldface indicates residues of conserved properties with respect to the B. licheniformis domain in more than three sequences. Numbers correspond to the amino acid positions in the protein sequences.

Deletion of CRH1/YGR189c and CRH2/YEL040w causes additive sensitivity to Calcofluor white and Congo red. To explore the roles of YGR189c, YEL040w, and YLR213c in cell wall construction, a search for cell wall-specific phenotypes was carried out on FY1679-derived strains with deletions of these genes. Haploid strains with the YGR189c gene deleted had been constructed in the FY1679 background by incorporation of a Kan^r marker (40). yel040w and ylr213c deletant strains were constructed by replacing the ORF with the HIS3 and LEU2 markers, respectively (see Materials and Methods). We found that deletion of either YGR189c or YEL040w led to sensitivity to the cell wall-binding dyes Calcofluor white and Congo red (Fig. 2), both of which are known to interfere with proper assembly of cell wall components. When spotted onto Congo red and Calcofluor white plates, ylr213 strains did not show any sensitivity to these compounds (Fig. 2). We shall refer to YGR189c and YEL040w (previously designated UTR2, for transcript with undetermined function [33]) as CRH1 and CRH2, respectively (for Congo red hypersensitive), below. The YLR213c gene will be referred to as CRR1, for CRH related.

View larger version (55K): [in this window] [in a new window]   FIG. 2.   Sensitivities to Congo red and Calcofluor white of strains bearing all the possible combinations of deletions of the CRH1, CRH2, and CRR1 genes. Cells were grown in YED, and 1/10 dilution series of each strain were spotted onto YED plates containing the indicated amounts of Congo red and Calcofluor white.

crh1 crh2

View larger version (72K): [in this window] [in a new window]   FIG. 3.   (A) Complementation studies on a crh1 crh2 deletion strain growing on YED plates containing Congo red at the amounts indicated and transformed with various plasmids harboring the wild-type (WT) CRH1 or CRH2 gene in a pRS416-based centromeric vector, a YEp352-based episomic vector, or the pCM190-based overexpression vector (see Table 2). (B) Resistance to Congo red induced by overexpression of CRH2 in a wild-type background. (C) A mutant crh1-136N,138Q allele does not complement the sensitivity to Congo red in the crh1 crh2 strain.

crh1 crr1

crh2 crr1

crh1 crh2 crr1

The solubility of -glucan in alkali is altered in crh1 crh2 mutants. Chemical fractionation of isolated cell walls (7) from wild-type, crh1, crh2, and crh1 crh2 double-mutant strains was carried out to detect possible modifications in the distribution of cell wall polymers. Cell walls from various clones of each strain were isolated, digested with pronase to eliminate mannoproteins, and treated with hot alkali to obtain an alkali-soluble fraction. The insoluble residue was subjected to an acid treatment, leading to the isolation of a minor acid-soluble fraction and a residual acid-insoluble, alkali-insoluble fraction. Glucan insolubility in alkali is due to its binding to chitin (15, 26), so variations in the normal proportions of soluble and insoluble -glucan may reflect either an altered proportion of glucan and chitin in the cell wall or an abnormal degree of cross-linking between these polymers. The average percentage (dry weight) of each fraction obtained from these experiments is shown in Fig. 4; the whole pronase-treated polysaccharide residue is considered 100%. According to these data, the alkali-soluble fraction increased in the crh1 strain compared to the wild type. In contrast, no significant variations were found in the crh2 strain. However, simultaneous deletion of CRH1 and CRH2 clearly showed a more severe phenotype than that conferred by single disruption of CRH1: the alkali-soluble fraction was almost doubled in the crh1 crh2 strain compared to that in the wild type, in detriment to the alkali-insoluble fraction. These data prompted us to measure the levels of chitin in these strains. As shown in Fig. 4, no significant variations in the content of chitin were observed that could explain the differences in the level of alkali-soluble glucan between the wild-type strain and the crh1 or crh1 crh2 knockout strain.

View larger version (39K): [in this window] [in a new window]   FIG. 4.   Determination of the alkali-insoluble-acid-soluble, alkali-soluble, and alkali-insoluble-acid-insoluble fractions of cell wall glucan in wild-type, crh1, crh2, and crh1 crh2 cells. The whole glucan fraction that remained after protease treatment of purified cell walls was considered 100%. a, the chitin content for each strain, in micrograms of glucosamine per milligram (dry weight) of cells, is given.

Transcriptional regulation of the CRH1, CRH2, and CRR1 genes. At this point we wished to know how these genes were regulated at different stages of the yeast life cycle. To achieve this goal, Northern blot hybridization experiments were carried out on total RNA extracted from cells grown under different conditions. Under normal growth conditions, CRH2 showed a level of expression approximately sixfold higher than that of the CRH1 gene. By contrast, the CRR1 transcript was found at very low levels, about 40 times lower than those of the CRH1 transcript (data not shown). When glucose was replaced by galactose under standard growth conditions, the expression of the CRH1 gene approximately doubled (Fig. 5A). To investigate whether the deletion of these genes would cause transcriptional activation of any of the other counterparts, we determined the mRNA levels of each gene in single-deletant backgrounds. As expected, the corresponding transcripts were not detected in strains with deletions of these genes (Fig. 5A). Neither the deletion of CRH1 nor that of CRH2 led to a significant variation in the expression of the other two genes of the family. However, the expression of CRH2 was up-regulated in a crr1 background.

View larger version (39K): [in this window] [in a new window]   FIG. 5.   Northern blot analysis for the study of the expression of the CRH1, CRH2, and CRR1 genes. (A) Relative mRNA levels for CRH1, CRH2, and CRR1. mRNA expression levels in a wild-type background when cells were grown in glucose-rich medium (open bars) were assigned a value of 1 for each hybridization assay. The amount of RNA was normalized to that of ACT1 mRNA (a constitutively expressed gene in S. cerevisiae). mRNA levels for each gene in a wild-type background when cells were grown in galactose-based YEG medium (solid bars), in a crh1 strain grown in YED (stippled bars), in a crh2 background (horizontally striped bars), and in a crr1 strain (diagonally striped bars) are also shown. (B) Northern blot analysis of CRH1, CRH2, and CRR1 expression during the cell cycle in synchronized haploid cells after release from pheromone-induced arrest (see Materials and Methods). A relative value of 1 was assigned to RNA levels for each transcript at time 0. The exposure times for each film were as follows: 2 h for ACT1, overnight for CRH1 and CRH2, and 10 days for CRR1. Solid line, CRH1; shaded line, CRH2; dashed line, CRR1.

Crh1- and Crh2-GFP fusions localize at chitin-rich points of the cell wall under different developmental conditions. With a view to determining the subcellular localization of the Crh1 and Crh2 polypeptides in vivo, we constructed plasmids encoding Crh1-GFP and Crh2-GFP fusion proteins. The GFP-coding sequence was cloned into the ORFs of the CRH1 and CRH2 genes without altering either the N-terminal secretion signal or the GPI attachment site (see Materials and Methods). Transcription of both GFP fusions was driven by their own gene promoters. crh1 and crh2 strains harboring episomic plasmids including GFP fusions were obtained, and fluorescence was followed by confocal microscopy.

View larger version (124K): [in this window] [in a new window]   FIG. 6.   Cellular localization by confocal microscopy of Crh1-GFP and Crh2-GFP. (A through F) Crh1-GFP visualized in crh1 cells bearing the pJV89G plasmid during vegetative growth. (A and B) Images of the same cells 45 min apart. The fluorescent patch in panel A marks the site of bud emergence. (C through F) Cells from the same transformant after release from mating pheromone-induced arrest in different stages of the budding cycle: cells with incipient budding (C), with medium-sized buds (D), with large buds (E), and after separation (F). (G) A double homozygous crh1 crh2 diploid strain transformed with pJV89G and allowed to sporulate, showing Crh1-GFP localization in ascospores. (H through Q) Time lapse localization of Crh2-GFP in crh2-deleted cells bearing the pJV40G plasmid. The pictures show a sequence of images taken every 10 min on a growing bud from the stage of bud emergence (H and I), every 20 min during bud growth (J through M), and during the stages of cytokinesis and cell separation at intervals of 10 min, except for the last one (30 min) (N through Q). (R through W) Confocal analysis of six slides (0.5-µm width) of budding cells expressing Chr2-GFP. Red color, chitin-rich areas detected by WGA-TRITC staining; green color, localization of the protein. (X through Y) Colocalization of Crh2-GFP and WGA-TRITC-stained chitin at the bud scar that remains from the last cytokinetic event in both diploid (X) and haploid (Y) backgrounds. (Z) Colocalization of Crh1-GFP and WGA-TRITC-stained chitin at the bud scar. (a through e) crh2 strain transformed with pJV40G in the process of mating to untransformed cells of the opposite mating type. The sequence in panels a through c shows the localization of Crh2-GFP in the mating projection and throughout the process of cell fusion (time lapse between pictures, 15 min). In panel d, a mature zygote is shown, with Crh2p marking the cytokinesis site. The image in panel e depicts the same zygote after 90 min, when it had developed a second diploid bud. A residual Crh2-GFP mark is apparent at the birth site of the first diploid, showing the bipolar polarity pattern characteristic of diploids. (f) Localization of Crh2-GFP in a chs2::LEU2 crh2 background. White arrowheads indicate accumulation of GFP fusion proteins at incipient bud sites and bud necks. Yellow arrowheads indicate accumulation of fluorescence at the shmoo. Arrows point to GFP-Crh localization at bud scars.

crh1 crh2 chs2

crh1 crh2 crr1 chs2

crh1 crh2 chs3

crh1 crh2 crr1 chs3

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A novel family of yeast cell wall-related proteins. Our knowledge of the yeast cell wall has increased in the past few years. Some of the linkages between yeast cell wall polymers (-1,3-glucan, -1,6-glucan, chitin, and mannoproteins) have been described (23, 26, 27), but little is known about the involvement of particular enzymes in the cross-linking mechanisms. Proteins such as Bgl2 or Gas1 have been postulated to exert such a function (13, 37), but no molecular evidence has been provided that such a role is exerted in vivo. To date, the only yeast cell wall enzyme for which a glycosyltransferase activity has been biochemically characterized is Bgl2. This protein was found to function as a -1,3-glucanosyl transferase (13), although it had previously been described as an exo- (25) and endo--1,3-glucanase (35). Here we present a family of homologous yeast cell wall-related proteins, namely, Crh1, Crh2, and Crr1. They show a potential N-terminal signal peptide for their integration into the secretory pathway, and, in addition, Crh1 and Crh2 display in their sequences a C-terminal consensus for GPI anchoring and Ser- and Thr-rich areas, both features characteristic of cell wall proteins (5). In a previous systematic analysis, Hamada et al. (14) have shown the functionality of the GPI-anchoring sequence present in both Crh1 and Crh2 proteins. By means of fusion of the C-terminal domain of each protein to a reporter protein including a secretion signal, -galactosidase, and a hemagglutinin (HA) epitope, these authors showed that Ygr189c (Crh1) and Yel040w (Crh2) fusion proteins were incorporated into the cell wall and released after treatment with laminarinase (14). These data indicate that Crh1 and Crh2 are cell wall proteins covalently attached to the cell wall glucan. Outside the Ser- and Thr-rich regions, all three proteins show significant similarities to bacterial endo--1,3-1,4-glycosidases and plant xyloglucan endotransglycosidases, including a hypothetical catalytic domain. This domain has been perfectly characterized in bacterial endo--1,3-1,4-glycosidases, in which glutamates at positions 134 and 138 has been demonstrated to act as a nucleophile and a general acid-base residue, respectively, in the double-displacement catalytic mechanism for retaining glycosidases (49). In the context of the yeast cell wall, where polysaccharide contents are high, it is likely that proteins with putative glycosidase activity (like Crh1, Crh2, and Crr1) would function as glucanosyl transferases. In plants, endotransglycosidase activity has been theorized to participate in cell wall expansion during growth by cutting and rejoining xyloglucan chains that cross-link microfibrils of cellulose in the primary cell wall (9). If it is assumed that this sort of cell wall-biosynthetic mechanism might also work in the yeast cell (3), members of the Crh family could be involved in cross-linking between cell wall polymers by means of transglycosylation reactions.

crh1 crh2

crh1 crh2

CRH1, CRH2, and CRR1 are regulated differently during the yeast life cycle. The fact that the cell wall is a dynamic structure in the different developmental programs of the yeast life cycle implies that the biosynthesis and assembly of cell wall components must be accurately regulated. Although Crh1 and Crh2 might have an overlapping function, their different patterns of expression and protein localization suggest a specific role for each protein of the family in different stages of the yeast life cycle. While CRH2 transcript levels are high and stable throughout the mitotic cycle, CRH1 expression is less abundant but reaches two maximum points: a clear G₁ peak and a second peak at M/G₁. The late G₁ phase is associated with early stages of budding, a crucial point in the construction of the cell wall, which has to be remodeled for bud emergence and growth. Consistent with this, several proteins involved in the biosynthesis of cell wall components and cell wall construction, among which are FKS1, KRE6, CSD2, and GAS1, have been shown to be specifically expressed at this stage (19). Probably, periodic expression of genes involved in cell wall construction helps the cell to successfully pass through this stage (19).

Crh1 and Crh2 localize at polarized growth sites. Here we provide evidence that both Crh1 and Crh2 are localized at the cell surface. To our knowledge, this is the first time that a detailed localization pattern has been reported for a cell wall protein by means of fusion to GFP under different developmental conditions. In general, the localization of these proteins is reminiscent of the distribution of cell wall chitin, which is mainly concentrated in the ring at the bud neck and the subsequent bud scar. Nevertheless, the distribution and timing of Crh1- and Crh2-GFP fusion proteins during the mitotic cycle revealed differences between them, consistent with their different transcription profiles and with the possibility of these two genes performing their common function at different cell cycle stages. Crh1-GFP is located at the site of bud emergence and, later, at the mother-daughter constriction in large-budded cells, consistent with the peaking of CRH1 transcripts at the equivalent cell cycle stages. On the other hand, the expression of the CRH2 gene did not vary significantly during the cell cycle, and consequently Crh2-GFP was detectable throughout the entire budding process. Crh2-GFP was concentrated at the chitin-rich bud neck, although it was also detectable throughout the lateral cell wall. At the time of cell separation, the Crh2-GFP signal in the lateral wall was more conspicuous in the daughter than in the mother cell. Interestingly, chitin has been reported to be absent from the lateral walls of daughter cells until the time of septum formation (45). Moreover, both Crh1- and Crh2-GFP fusion proteins frequently marked the bud scar remaining from the previous budding event, but not older scars, as determined by simultaneous staining of chitin with rhodamine-conjugated WGA. It has been shown by others that the chitin participating in the linkages to both -1,3- and -1,6-glucan is synthesized by CSIII (26, 27). This chitin is found at the base of an emerging bud and, in a dispersed form, is interspersed through the whole cell wall. On the basis of Crh1 localization, a suggestive hypothesis might be the involvement of this protein in glucan-chitin assembly early at the budding site and later at the septum formed between mother and daughter cells at cytokinesis. In the case of Crh2, its timing and localization indicate a role for this protein in assembly during the whole budding process, not only at the neck but also in the lateral cell wall.