Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Signal transduction wikipedia , lookup
Cell growth wikipedia , lookup
Extracellular matrix wikipedia , lookup
Cellular differentiation wikipedia , lookup
Tissue engineering wikipedia , lookup
Organ-on-a-chip wikipedia , lookup
Cell culture wikipedia , lookup
Cell encapsulation wikipedia , lookup
4449 Journal of Cell Science 112, 4449-4460 (1999) Printed in Great Britain © The Company of Biologists Limited 1999 JCS0763 Although calnexin is essential in S. pombe, its highly conserved central domain is dispensable for viability Aram Elagöz1, Mario Callejo1, John Armstrong2 and Luis A. Rokeach1,* 1Département de biochimie, Université de Montréal, CP 6128, succ. Centre-ville, Montréal, 2School of Biological Sciences, University of Sussex, Falmer, Brighton BN1 9QG, UK Québec H3C 3J7, Canada *Author for correspondence (e-mail: [email protected]) Accepted 22 September; published on WWW 17 November 1999 SUMMARY In mammalian cells, the calnexin/calreticulin chaperones play a key role in glycoprotein folding and its control within the endoplasmic reticulum (ER), by interacting with folding intermediates via their monoglucosylated glycans. This lectin activity has been mapped in mammalian calnexin/calreticulin chaperones to the central region, which is a highly conserved feature of calnexin/calreticulin molecules across species. The central domain has also been implicated in Ca2+ binding, and it has been proposed to be involved in the regulation of calcium homeostasis in the ER. Herein, we show that although the Schizosaccharomyces pombe calnexin is essential for viability, cells lacking its 317amino-acid highly conserved central region are viable under normal growth conditions. However, the central region appears to be necessary for optimal growth under high ERstress, suggesting that this region is important under extreme folding situations (such as DTT and temperature). The minimal length of calnexin required for viability spans the C-terminal 123 residues. Furthermore, cells with the central domain of the protein deleted were affected in their morphology at 37°C, probably due to a defect in cell wall synthesis, although these mutant cells exhibited the same calcium tolerance as wild-type cells at 30°C. INTRODUCTION containing the Glc1Man9GlcNAc2 glycan). Monoglucosylated glycoproteins arise from the consecutive action of the glucosidase I and II, or by reglucosylation of Glc0Man9GlcNAc2-containing proteins by UDPGlc:glycoprotein glucosyltransferase (GT; reviewed in Parodi, 1998). A distinctive feature of GT is that, in vitro, it can only reglucosylate unfolded proteins (Sousa and Parodi, 1995). These observations have led to one of the current models on quality control of glycoprotein folding in the ER, in which calnexin/calreticulin and GT constitute key elements (Trombetta and Helenius, 1998, and references therein). According to this model, the mammalian calnexin/calreticulin chaperones interact with monoglucosylated, newly synthesized proteins exclusively through their oligosaccharide moieties in a manner similar to lectins, without peptide contacts. Numerous publications have supported this model (reviewed in Trombetta and Helenius, 1998). However, peptide contacts also appear to be important for the interaction of calnexin with certain substrates. For instance, Ware et al. (1995) have shown that once the glycoprotein-calnexin complex is formed in vitro, the glycans can be removed without disrupting the interaction. Thus, these authors proposed a two-step model for calnexinligand interaction, where the first contact occurs through the Glc1Man9GlcNAc2 oligosaccharide, while the second step is peptide-mediated, probably through exposed hydrophobic patches, in a mechanism akin to other chaperones such as BiP (Ware et al., 1995; Williams, 1995). Moreover, a number of Membrane-bound calnexin and its soluble homologue calreticulin define a new family of molecular chaperones that have been implicated in the folding of glycoproteins in the ER of mammalian cells, such as the cystic fibrosis transductance regulator (CFTR) and the T-cell receptor (reviewed in Bergeron et al., 1994; Williams, 1995; Trombetta and Helenius, 1998; Parodi, 1998). The lumenal domain of calnexin shares extensive similarity with calreticulin. Both proteins contain a highly conserved central region comprising two series of tandemly repeated motifs. Motif 1 (I-DPD/EA-KPEDWDD/E) is involved in Ca2+-binding (Wada et al., 1991; Tjoelker et al., 1994). Ca2+ may be important for calnexin function since chelators disrupt its interaction with ligands in vitro (Bergeron et al., 1994; Le et al., 1994; Loo and Clarke, 1994; Vassilakos et al., 1998). In mammalian calnexin, Motif 2 (G-W- -P-I-NPY), along with Motif 1, were shown to be required for binding glycoproteins through their N-linked glycans (Vassilakos et al., 1996, 1998). The use of the α-glucosidase inhibitors castanospermine and 1-deoxynojirimycin in vivo (reviewed in Bergeron et al., 1994; Williams, 1995; Trombetta and Helenius, 1998; Parodi, 1998) and purified glycans in vitro (Ware et al., 1995; Vassilakos et al., 1996; Vassilakos et al., 1998), demonstrated that the calnexin/calreticulin family displays a lectin-like activity with unique selectivity for monoglucosylated glycoproteins (i.e. Key words: BiP, Chaperone, Protein folding, Yeast genetics, Schizosaccharomyces pombe 4450 A. Elagöz and others publications reported that calnexin/calreticulin can bind nonglycosylated proteins (reviewed in Williams, 1995; Jannatipour et al., 1998). Also, castanospermine and tunicamycin (a potent inhibitor of glycosylation) reduced but did not eliminate the binding of certain glycoproteins to calnexin/calreticulin (Pipe et al., 1998; Keller et al., 1998). Hence, the calnexin/calreticulin chaperones appear to have multiple modes of interaction with their ligands that may vary for different proteins. Calnexin and calreticulin are not the only chaperones involved in glycoprotein folding. The calnexin/calreticulin family was shown to cooperate with other chaperones such as BiP, PDI and Erp57, in the conformational maturation of folding intermediates (e.g. Kim and Arvan, 1995; Oliver et al., 1997; Elliott et al., 1997; Zapun et al., 1998; Hammond and Helenius, 1994; Tatu and Helenius, 1997). Schizosaccharomyces pombe encodes the basic components involved in the quality control of glycoprotein folding found in the mammalian ER (Jannatipour and Rokeach, 1995; Parlati et al., 1995; Fernández et al., 1996; Fanchiotti et al., 1998). Therefore, this fission yeast represents an ideal model organism for the genetic analysis of the mechanisms of glycoprotein folding and its quality control. We and others have isolated cnx1+, the S. pombe calnexin homologue, and have shown that it encodes a protein essential for viability (Jannatipour and Rokeach, 1995; Parlati et al., 1995). Like its mammalian counterparts, Cnx1p is a type I ER-membrane protein containing the characteristic highly conserved central region including motifs 1 and 2, and a cytosolic tail (Jannatipour and Rokeach, 1995; Parlati et al., 1995). As could be expected, Cnx1p was shown to bind Ca2+ (Parlati et al., 1995; our unpublished results). Although a lectin activity for S. pombe calnexin has not been explored as yet, the presence of four copies each of motifs 1 and 2 may suggest that the central region might also be involved in glycan binding. Like the mammalian GT enzyme, Gpt1p in S. pombe was shown to reglucosylate unfolded proteins in vitro, but its presence is dispensable for cell life under normal growth conditions (Fernández et al., 1994, 1996). From these observations it was possible to infer the existence in the fission yeast of a GT-independent mechanism for quality control of protein folding (Jannatipour et al., 1998). We have recently shown that Cnx1p associates with newly synthesized molecules of the glycoprotein acid phosphatase, independently of glucose trimming and reglucosylation by GT (Jannatipour et al., 1998). Thus, in spite of the essentiality of Cnx1p for fission yeast viability, the glucose trimming and reglucosylation cycle do not appear to be indispensable for protein folding in S. pombe. This notion is further supported by the fact that fission yeast cells genetically depleted of both Gpt1 and glucosidase II (Gls2p) are viable (Fanchiotti et al., 1998). Based on the sequence conservation and the role of the central region in mammalian calnexin, it could be predicted that this domain might encode the essential function(s) for S. pombe viability. However, as described above, in the fission yeast the other key elements in the calnexin/calreticulin cycle do not seem to be necessary for viability under standard growth conditions. In this study, we wished to delimit the Cnx1p region required for cell viability and assess the relevance of the highly conserved calnexin/calreticulin central domain. Our results showed that the 317-amino-acid (aa) central region is dispensable under normal growth conditions but is required in situations of high-folding stress in the ER. Thus, overall our observations further suggest that the mechanism of glycoprotein folding in S. pombe might be different from that in mammalian cells. MATERIALS AND METHODS Strains and medium The S. pombe strain SP6089 was used for all plasmid shuffling experiments (see below for construction and genotype). S. pombe strain SP556 (h+ ade6-M216 ura4-D18 leu1-32) was used as a wild-type control in various experiments. S. pombe transformations were performed by the lithium acetate procedure, and genomic DNA extraction was as previously described (Moreno et al., 1991). Strains were grown at 30°C (except where indicated) in EMM minimal medium supplemented with nutrient requirements (Moreno et al., 1991). The E. coli strain Top10, from TOPO XL PCR Cloning Kit (Invitrogen Co. CA), was used for cloning PCR products and the E. coli strain AP401 (lon::mini tetR ara− ∆lac-pro nalA argEam rifR thiI [F′ pro AB lacIq Z M15]) was used for plasmid amplification and isolation. Construction of the cnx1::his3 haploid strain SP6089 To construct a haploid S. pombe deleted of entire Cnx1p coding sequences, a plasmid containing a 4.5-kb PstI-PstI genomic fragment encompassing the cnx1 gene was linearized with NcoI and progressive bi-directional digestion was made with ExoIII. The his3 gene was ligated to create the plasmid pSPCA3282, in which the cnx1 coding sequences (955-2805, according to EMBL U13389) were deleted, and this was used in the subsequent steps to obtain cnx∆ strains. Disruption of cnx1 in the haploid strain was obtained by simultaneous transformation of the haploid S. pombe strain 248 (h− his3-D1 ade6M216 ura4-D18 leu1-32; Burke and Gould, 1994); with the PstI-PstI fragment from plasmid pSPCA3282 (containing the cnx1 deletion marked with his3) and the plasmid pSPCA3261 containing Cnx1p coding sequences in the vector pREP42 (see below). Transformants were selected for histidine prototrophy, and selected clones were analyzed by Southern blotting and PCR to ascertain correct disruptive integration. Strain SP6089 had the correct ∆cnx1::his3 complemented by episomal copies of cnx1 (plasmid pSPCA3261). The genotype of strain SP6089 is: h− his3-D1 ade6-M216 ura4-D18 leu1-32 ∆cnx1::his3[pSPCA3261]. DNA manipulation and analysis Procedures used for DNA manipulation and analysis (purification, digestion, electrophoresis, transformation, etc.) were as previously described (Sambrook et al., 1989). The DNA sequences of both strands of mutant cnx1-containing clones were determined by the dideoxy chain-terminating method using a T7 DNA Polymerase Kit (Amersham Pharmacia Biotech, Inc. NJ). Polymerase chain reaction Site-directed mutagenesis experiments were done by polymerase chain reaction (PCR) using the overlap extension method, as described (Higuchi and Krummel, 1988), with Pfu DNA polymerase, using the manufacturer’s conditions (Stratagene, La Jolla, CA). In oligonucleotide sequences (see below), the NdeI restriction site is underlined, the BamHI site is shown in bold characters, the corresponding nucleotide sequence of the ER retention signal ADEL is shown in italics, and stop codons are doubly underlined. Nucleotide numbers correspond to the position on cnx1 sequence (Jannatipour and Rokeach, 1995) as in the database (EMBL accession number U13389). Oligonucleotides used for the construction of each mutant (Fig. 1B) were: primers A, 5′-CCA CCC AAC ACG TGC ATA TGA AGT ACG GAA AG-3′ (1049-1080 nt) and B, 5′-CGG GAT CCT TAC CCA ATT TCA GGA GTC TCG ATG-3′ (2532-2511 nt) Cells without the calnexin/calreticulin central region are viable 4451 for lumenal (#3) mutant; primers A and C, 5′-CGG GAT CCT TAA AGC TCG TCA GCC CCA ATT TCA GGA GTC TCG ATG-3′ (25322511 nt) for lumenal+ADEL (#16); primers A and D, 5′-CGG GAT CCT TAG CGG GTT CAT CAA ATT CAT AAG-3′ (1395-1376 nt) for N-terminal (#9); primers A and E, 5′-CGG GAT CCT TAA AGC TCG TCA GCG GGT TCA TCA AAT TCA TAA G-3′ (1395-1376 nt) for N-terminal+ADEL (#10) mutant; primers A and F, 5′-TTG CTT TTC CAT AGG ATC AGC AAG TGA TCC CCG-3′ (1137-1117 nt) were used to amplify the first 5′ region of the mini-cnx fragment, then primers G, 5′-ATG GAA AAG CAA TCA ATG CAT G-3′ (22772398 nt) and H, 5′-CGG GAT CCG GCT TTT AAC AGA GTC GCT AC-3′ (2932-2913 nt) were used to amplify the 3′ part. Finally, external A and H primers were used to amplify the whole insert of mini-cnx (#11). For the central region mutant (#12), primers A and F were first used to amplify the 5′ region. Primers I, 5′-ATA AAT GAA CCA GAA AAG GAT TTG-3′ (1396-1419 nt) and J, 5′-CGG GAT CCT TAG GAC TCT TGC TTA GAA AGT AAT TC-3′ (2376-2453 nt) were used to amplify the 3′ region, and finally external primers A and H were used to amplify the whole insert. To add the ADEL sequence to the central region mutant (#12+ADEL), primer K, 5′CGG GAT CCT TAA AGC TCG TCA GCG GAC TCT TGC TTA GAA AGT AAT TC-3′ (2376-2453 nt) was used. Primers A, F and I, and H were used for the construction of cnx-N-terminal (#14) mutant, and primers A, F and I, B for luminal-N-terminal (#15) mutant. The various cnx1 mutants were constructed by cloning the PCR products (see above) in pCR-XL-TOPOTM (Invitrogen Co. CA) following the manufacturer’s conditions. The construction of deleted-cnx (#4) fragment was carried out using in vitro mutagenesis system Altered Sites II™ (Promega Co., WI), as described by the manufacturer. Briefly, DEL-CNX1 5′-GAA TTT GAT GAA CCA TGG ATG AAC CAG AAA AGG-3′ (1381-1414 nt), primer was used to create a NcoI site at position 2942 nt (the restriction site is in bold italic characters, and the mutagenized bases are underlined) on wild-type cnx1, then the internal 981-bp NcoI-NcoI fragment was removed. Plasmids pREP41 is an S. pombe expression multicopy vector containing the LEU2 marker and the ars1 origin of replication. Expression in this plasmid is under the control of the thiamine repressible nmt41 promoter (induction ratio 25×), flanking a polylinker site (Maundrell, 1993). In the case of pREP42, the ura4 marker is used for selection instead of LEU2. The full-length and the various cnx1 mutant genes were inserted in pREP41 digested with NdeI and BamHI. The plasmids constructed with pREP41 were designated as follows: pSPCA3220 (#2; full-length cnx1), pSPCA3221 (#3), pSPCA3222 (#4), pSPCA7092 (#9); pSPCA7093 (#10), pSPCA7094 (#11), pSPCA7102 (#12), pSPCA7096 (#13), pSPCA7097 (#14), pSPCA7100 (#15) and pSPCA7095 (#16). The construct containing full-length wild-type cnx1 (#2) in pREP42 was designated pSCPCA3261. Constructs containing inserts #17 and #18 were cloned in pREP42 and the resulting plasmids were designated pSPCA7178 and pSPCA7180, respectively. These last plasmids were constructed by replacing the ClaI-BamHI fragment pSPCA7136 (pREP42-#16) and replaced with the ClaI-BamHI of constructs #12 and #13. All S. pombe cultures in this study were grown in EMM medium under induced conditions (absence of thiamine). Plasmid shuffling experiments S. pombe SP6089 strain containing wild-type cnx1+ inserted on pREP42 (ura4 marker) was transformed with 2 µg of pREP41 (LEU2 marker) containing either full-length wild-type cnx1+ or mutant cnx1. Transformants were grown for 6 days at 30°C in 5 ml liquid EMM supplemented with adenine (Ade) and uracil (Ura) to chase the ura4containing pREP42 plasmid, then cells were plated onto solid EMM+Ade+Ura. After 3-4 days at 30°C, colonies were replica-plated onto EMM+Ade+Ura and EMM+Ade-Ura. Cells which shuffled wild-type cnx1 are Leu+/Ura− and contain mutant cnx1. In the case of pSPCA7178 (#17) and pSPCA7180 (#18), the plasmid shuffling was carried out against pREP41-cnx1+ (pSPCA3220) in EMM+Ade+Leu. Spheroplast and membrane preparations S. pombe wild-type, lumenal and lumenal+ADEL strains were grown at 30°C in 250 ml EMM+Ade+Ura containing 0.5% glucose, to an OD595 of 0.5. Deleted-cnx and mini-cnx strains were grown under same conditions, except that 1 l of medium was used. For each strain, 109-1010 cells were harvested by centrifugation, washed in 1 ml of citrate buffer I (20 mM citrate phosphate, pH 5.6, 40 mM EDTA), and resuspended in 5 ml of citrate buffer II (50 mM citrate phosphate, pH 5.6, 1.2 M sorbitol). Cells were spheroplasted with 25 mg of NovoZym (Sigma Chemicals Co.) by a 45-minute incubation at 37°C. Spheroplasts were spun at 500 g for 5 minutes at 4°C, washed twice in citrate buffer II containing 1.2 M sorbitol. The pellet was resuspended in 2 ml of lysis buffer (0.1 M sorbitol, 20 mM Hepes, pH 7.5, 20 mM potassium acetate, pH 7.4) containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride: PMSF), 10 mM iodoacetamide: IAA), 300 µg/ml pepstatin A, 300 µg/ml leupeptine, 300 mg/ml phenanthroline). Lysates were prepared in a Potter-Elvejhem homogenizer and cleared at 1000 g by spinning for 5 minutes at 4°C. The supernatant was spun at 15,000 g for 15 minutes at 4°C in an SW 50.1 ultracentrifuge (Beckman Instruments, Fullerton, Ca). The supernatant from this spin was centrifuged at 100,000 g for 1 hour at 4°C in the SW 50.1 ultracentrifuge, and the resulting pellet was resuspended in 2.5 ml of ice-cold lysis buffer. Membrane extraction S. pombe microsomal membranes (as described above) were treated for 15 minutes at 4°C by mixing with 1 vol. of either 1 M NaCl, 0.2% SDS, 0.2 M sodium carbonate, pH 11.5, or 3% Triton X-100. Membrane lysates were spun at 80,000 g for 1 hour at 4°C in an SW 50.1 ultracentrifuge, then the pellet from this spin was resuspended in 0.1 ml of 3× Laemmli’s sample buffer (Laemmli, 1970) (P fraction). Proteins in the supernatant fraction were treated for 30 minutes at 4°C in the presence of 6% ice-cold tricholoroacetic acid (TCA) and spun at 2,000 g for 45 minutes at 4°C. The pellet was washed twice in icecold 80% acetone and dissolved in 0.1 ml in sample buffer (S fraction). Before SDS-PAGE, samples were boiled for 5 minutes. Cell-wall resistance to lytic enzymes 20 ml cultures of wild-type or mutant cnx1 cells were grown in liquid EMM+Ade+Ura medium at 30°C to an OD595 of 0.2-0.5. Cells were harvested, washed in 10 ml of citrate buffer I (20 mM citrate phosphate, pH 5.6, 40 mM EDTA, pH 8.0), spun and resuspended in 5 ml of citrate buffer II (50 mM citrate phosphate, pH 5.6, 0.1 M sorbitol). 2.5 mg of NovoZym (Sigma Chemicals Co.) was added and cells were incubated at 37°C. Cell wall degradation was monitored by measuring cell density at OD595 and taking samples every 20 minutes. Cell viability was also followed at the same time, by plating an appropriate amount of cells on solid EMM+Ade+Ura. Confocal microscopy Confocal immunofluorescence using anti-Cnx1p or anti-BiP antibodies (diluted 1:100) was carried out essentially as previously described (Pidoux and Armstrong, 1993; Jannatipour et al., 1998). Briefly, cells were grown for 8 hours in EMM at 37°C before fixation and labeling with anti-Cnx1p and anti-BiP antibodies. Calcofluor staining of cell wall Exponentially growing cells were washed once with 1× PBS (130 mM NaCl, 2.5 mM KCl, 5 mM Na2HPO4, 1.8 mM KH2PO4, pH 7.4) and resuspended in 100 µl of the same buffer containing 20 mg/ml of Calcofluor white (Sigma, F-3543). The cells were subsequently washed in 1× PBS and placed on a microscope slide and covered with a coverslip. Cells were observed under UV light or Nomarski interference. 4452 A. Elagöz and others Fig. 1. S. pombe calnexin (cnx1) mutants ER lumen Cytosol C.terminal used in this work and viability of A N.terminal (86aa) Conserved central-domain (317aa) (52aa) TM (48aa) phenotypes. (A) Schematic representation of 1 1 1 12 2 2 2 100 aa conserved amino acids of the central domain NH2 COOH of Cnx1p. The amino acid sequences of 121 QYEVNpEEGL nCGGAYLKLL aepthgemsn sidYrIMFGP DKCGvndRVH calnexin homologues were aligned as 172 FIFKHKNPlT GEYSEKHLds rpasllkpgi tnlytlivkp dqtfevring described in Materials and Methods. The 221 dvvrqgslfy dfiPPVLPPV EIyDPEDiKP aDWvDePeIP DPNAVKPDDW highly conserved central region extends 271 DEDAPrmIPD pDAvKPEdWL EDEPlYIPDP EAqKPEDWDD EEDGDWiPse from residues 121 to 437. A consensus was made with the sequences aligned (see 319 IiNPKCiEgA GCGEWKpPWI RNPNYRGpWs PPMIpNPefi GeWYPRKIPN Materials and Methods), with similarities 371 PDYFDddhps hfgPlygvGf ELWTMQpnIr FsNiyVghsi edaerlgNET and identities denoted as follows: bold 421 FLpKLKaere llskqEs letters for 100% identity; capital-underlined Viability* B letters for 75%, capital letters for 50%, small pREP41 (no insert) wild type (#2) SP caps for 25%, and 0% identical residues + (A) were given by lower case letters. Open lumenal (#3) + (B) 2+ boxes, putative Ca -binding regions (motif ADEL lumenal +ADEL (#16) + (C) 1); double-headed arrows, conserved deleted_cnx (#4) + oligosaccharide-binding domains (motif 2; N.terminal (#9) (D) see Tjoelker et al., 1994; Vassilakos et al., ADEL N.terminal+ ADEL (#10) (E) 1998; reviewed by Trombetta and Helenius, (G) mini_ cnx (#11) (F) (H) + 1998); TM, the transmembrane domain; the (I) central region (#12) (J) ‘tree’, the N-glycosylation site; black boxes, ADEL #12+ADEL (#13) potential phosphorylation sites by protein (K) cnx-N.terminal (#14) kinase C (PKC); black bars, potential phosphorylation site by casein kinase II lumenal -N.terminal (#15) (CK2); SP, the signal peptide; N.terminal, lumenal -C.terminal (#17) C.terminal, indicate deletion of the 86-aa N ADEL lumenal -C.terminal (#18) terminus or 123-aa C terminus, respectively, of mature Cnx1p. (B) Plasmid shuffling experiments to study the cnx1 mutants were carried out as described in Materials and Methods. Viability*, cells which exchanged mutant cnx1 against wild-type cnx1+and contain only mutant cnx1. Primers used for the construction of cnx1 mutants by the overlapping PCR strategy are shown by arrows Immunoprecipitations and immunoblotting Immunoblots and immunoprecipitations were carried out essentially as previously described (Jannatipour et al., 1998). Protein sequence alignments Protein sequence alignments were realized by using the Block Maker multi-alignment program from BCM Search Launcher (http://dot.imgen.bcm.tmc.edu:9331), Human Genome Center (Baylor College of Medicine, Houston, TX). The sequences aligned to obtain the consensus were: A. thaliana (U08135), C. elegans (Z22181), Canis familiaris (P24643), D. melanogaster (U30466), Glycine max (U20502), H. sapiens (P27824), Mus musculus (P35564), Rana rugosa (D78590), Heliantus tuberosus (Z35108), S. cerevisiae (U12980) and S. pombe (U13389). RESULTS Cells deleted of the highly conserved central region of calnexin/calreticulin are viable Disruption of the cnx1 is lethal in S. pombe (Jannatipour and Rokeach, 1995; Parlati et al., 1995). In order to identify the Cnx1p sequences required for cell viability, we undertook a deletion approach. The central region is a highly conserved feature in the calnexin/calreticulin family of chaperones (see Fig. 1A). Because in mammalian cells, this segment was shown to encompass the ligand (glycan)-binding and the Ca2+ domains (Vassilakos et al., 1996, 1998; Tjoelker et al., 1994), it could be expected that since Cnx1p contains four copies of the repeated motifs 1 and 2, the deletion of this region would be lethal. To test this assumption, we constructed mutant #4 (deleted-cnx; Fig. 1B), and assessed the importance of this region by plasmid shuffling in the strain SP6089, in which genomic cnx1 was deleted and replaced with his3 (see Materials and Methods). Viability in this strain SP6089 is ensured by complementation with the plasmid pSPCA3261, bearing cnx1+ under the control of the nmt1 promoter (see Materials and Methods). The construct encoding full-length, wild-type Cnx1p (#2) was used in this experiment as a control (see Fig. 1B), and conferred viability as expected. Surprisingly, the cells harboring construct #4 were viable in the absence of a wild-type copy of cnx1. Western blot, Southern blot and PCR analyses confirmed that the viable phenotype of strain #4 was not due to gene conversion between wild-type and mutant genes, nor to integration of the wild-type sequences into the genome (not shown). From these results it is then possible to conclude that even though deletion of cnx1 is lethal, the calnexin/calreticulin highly conserved region is not essential for viability in S. pombe when grown under standard laboratory conditions (minimal medium, 30°C). In order to delimit the Cnx1p essential sequences and to evaluate whether the central region by itself can confer viability, a set of cnx1 deletion mutants was constructed and transformed into strain SP6089 (see Fig. 1B). For certain constructs encoding soluble proteins (#9, #10, #12, #13, #15, #17 and #18), versions with or without the ADEL ER-retention signal (Pidoux and Armstrong, 1992) were made in order to ensure that lack of viability was not due to escape of the mutant Cnx1p proteins from the ER. The functionality of the ADEL retention signal was verified by western blotting on supernatants from cultures (see Fig. 2G). The synthesis of all the Cells without the calnexin/calreticulin central region are viable 4453 Fig. 2. Localization of Cnx1p mutants and cell-wall sensitivity. Membranes from wildtype (A), deleted-cnx (C) and mini-cnx (E) cells were prepared as described in Materials and Methods and extracted with Tris-buffered saline, pH 7.5, alone (Mock), or supplemented with either 0.1% SDS, 1.5% Triton X-100, 0.1 M sodium carbonate, pH 11.5, or 0.5 M NaCl (high salt). This was followed by centrifugation (1 hour, 100,000 g) resulting in pellet (P) and supernatant (S) fractions. The fractions were separated by 10% SDS-PAGE for wild-type fractions and by 14% SDS-PAGE, containing 4 M urea, for deleted-cnx and mini-cnx extracts. Fractions were than analyzed by immunoblotting using polyclonal anti-Cnx1p antibodies. Note that the mini-Cnx1p and deleted-Cnx1p proteins repeatedly appear as doublet bands on highresolution gels, the highest mobility species probably being degradation products. Arrows indicate the position of mutant Cnx1p bands. Asterisk, diffusion of Cnx1p from #4 and #11 extracts was repeatedly observed. Confocal immunofluorescence on wild-type (B), deleted-cnx (D) and mini-cnx (F) cells was carried out with anti-Cnx1p antibodies as described in Materials and Methods. Labeling of the nuclear envelope and reticular structures through the cell and under the plasma membrane is apparent in each case. (G) Secreted and intracellular lumenal and lumenal+ADEL Cnx1p (lanes 1-7) were analyzed by western blotting with anti-Cnx1p antibodies. Molecular mass markers in kDa are indicated (MW). (H) Sensitivity to lytic enzymes (NovoZym 234) on cells grown at 37°C was monitored as the decrease in OD595, and converted to percentage of intact cells. The time required to reach 50% lysis is indicated as t1 (55 minutes) for mini-cnx, t2 (90 minutes) for deleted-cnx, and t3 (120 minutes) for wild type (values are the average of 3 experiments). Sensitivity to hygromycin B was done by drop test (as described in Fig. 6) using 50 µg/ml of the antibiotic at 30°C, on exponentially growing cells. Cnx1p mutant proteins was confirmed by western blotting with anti-Cnx1p polyclonal antibodies (not shown), and the ability of the different constructs to support viability was assessed, as above, by plasmid shuffling against pSPCA3261 encoding fulllength Cnx1p. As shown in Fig. 1B, in agreement with Parlati et al. (1995), the lumenal constructs with or without the ADEL retention sequence (#3 and #16) conferred viability. However, constructs including the central region without the sequences encoding the first 86 aa of the mature Cnx1p (#12-15) did not exchange with pSPCA3261 expressing full-length Cnx1p. This would suggest that although the central region is conserved in all calnexin/calreticulin molecules, it does not encode a function(s) sufficient for viability. On the other hand, the mini-cnx construct (#11) encoding the signal peptide, the last 52 aa of the lumenal domain, the transmembrane domain (TM) and the 48-aa cytosolic domain, conferred a viable phenotype. Constructs #9 and #10 encoding the first N-terminal 86 aa of the mature Cnx1p did not shuffle against wild type, indicating that this region alone is not sufficient for viability. Nevertheless, the first N-terminal 86 aa might have a role, such as in the folding of lumenal Cnx1p when the central region is present, as constructs #12-15 did not support viability. In this vein, Peterson and Helenius (1999) reported that a deletion of the 83 N-terminal aa of mammalian calreticulin inhibits its ability to bind to substrates, most likely by disrupting its structural integrity. Finally, constructs #17 and #18, which represent lumenal versions of Cnx1p deleted of the last 52 aa, did not support viability. Taken together, these results suggest that Cnx1p encodes at least two distinct domains: (1) a non-essential central domain, which in mammalian cells was demonstrated to bind folding glycoproteins (via their glycans) and Ca2+ (see Bergeron et al., 1994; Williams, 1995; Trombetta and Helenius, 1998; Parodi, 1998; Vassilakos et al., 1998); and (2) an essential domain for cell viability, mapping within the last 123 residues of the fission yeast molecule. By cell fractionation studies we confirmed that wild-type Cnx1p and the mutants #4 and #11 are membrane-bound (see Fig. 2A,C,E), and that constructs #3 and #16 are soluble proteins (not shown). In addition, confocal microscopy with anti-Cnx1p antibodies showed that the mutant proteins localized to the ER (Fig. 2B,D,F). Moreover, confocal microscopy with anti-BiP antibodies presented the typical S. pombe ER staining pattern (not shown), as previously reported for cnx1+ cells (Pidoux and Armstrong, 1992; Jannatipour et al., 1998). Mutants supporting viability were studied further. The growth rates of deleted-cnx and mini-cnx cells are reduced at 37°C but are similar to wild type at 30°C As a first step to assess how the different cnx1 mutants affect the physiology of S. pombe, we determined the growth rates of cells bearing the constructs lumenal (#3), lumenal+ADEL (#16), deleted-cnx (#4) and mini-cnx (#11). For comparison, we also determined the growth rates of the strain SP556 4454 A. Elagöz and others A 6 30ºC 5 C g time in min 3 m OD595n 4 2 SP556 lumenal (#3) lumenal+ADEL (#16) deleted-cnx (#4) mini-cnx (#11) wt (#2) 1 800 600 400 200 0 0 B 500 1000 1500 2000 0 2500 Time (min) 5 30 32 34 37 ºC 37ºC 4 wt (#2) deleted-cnx (#4) mini-cnx (#11) D 2 1 mini-cnx (#11) deleted-cnx (#4) wt (#2) g time in min m OD595n 3 800 600 Temp. shift + Glycerol 400 200 0 0 0 1000 2000 3000 4000 5000 6000 Time (min) 7000 37 30 ºC Fig. 3. Cells deleted of the central domain show temperature-sensitive reversible growth rates. Cells were cultured to late log phase and diluted into 25 ml of fresh EMM+Ade+Ura medium to OD595 of 0.02 and growth rates were determined. (A) Growth curves of SP556, wild-type (#2), lumenal (#3), lumenal+ADEL (#16), deleted-cnx (#4), and mini-cnx (#11) cells grown at 30°C. (B) Growth curves for cells grown at 37°C for 70 or 115 hours. (C) Generation (g) times for wild-type (#2), deleted-cnx (#4) and mini-cnx (#11) cells were calculated on the exponentialphase portions of the growth curves and determined at temperatures ranging from 30°C to 37°C. (D) Generation times for wild-type (#2), deleted-cnx (#4), and mini-cnx (#11) cells were determined after a temperature shift from 37° to 30°C. Late log-phase cells, cultured at 37°C, were diluted into fresh EMM+Ade+Ura medium to an OD595 of 0.02, and the growth temperature was shifted down to 30°C. Generation times were also calculated in the presence of 4% (approx. 0.4 M) glycerol at 30°C and 37°C (dotted lines). Cell growth was monitored by measuring OD595. The values represent the average of three experiments. carrying an intact genomic copy of cnx1+ and the strain carrying episomal copies of full-length cnx1+ (#2). As shown in Fig. 3A, all the strains grew at about the same rate at 30°C without a major lag phase. However at 37°C, while strains #3 and #16 grew at essentially the same rate as SP556 and (not shown) strain #2, deleted-cnx (#4) and mini-cnx (#11) cells grew at considerable slower rate, reaching stationary phase after 3000 minutes and 6000 minutes, respectively (see Fig. 3B). Moreover, at 37°C, the lag phase for #2 was 100 minutes, but 500 minutes for deleted-cnx and 1500 minutes for mini-cnx cells. A prolongation of the lag phase may correspond to a reversible adaptation of the organism to the stress conditions or may represent the time required for the selection of mutant populations of cells able to grow at 37°C. To discriminate between these two possibilities, we reasoned that if the first hypothesis was correct, the cells growing at 37°C when shifted back to 30°C would grow at the same rate as wild type (#2). On the other hand, if the second possibility was true, we expected that these mutant cells would grow at an altered rate when shifted down to 30°C. As shown in Fig. 3D, the growth rates of strains #4 (deleted-cnx) and #11 (mini-cnx) after the shift-down (i.e. from 37°C to 30°C) were practically indistinguishable from those carrying wild-type cnx1+ (#2). Moreover, when these mutant strains that were shifted down to 30°C were recultured at 37°C, they exhibited the same lag phase and growth rates as in the first experiment (not shown). In addition, when these mutant cells grown at 37°C were plated on solid medium they appeared homogenous without papillae, thus excluding the possibility of selection of mutated populations. To further explore the hypothesis of adaptation of strains #4 and #11 to the temperature stress, we measured the growth rates at intermediate temperatures. As shown in Fig. 3C, while the generation time for the wild-type strain remained practically invariant between 30°C and 37°C, that of strains #4 and #11 increased significantly with temperature. Furthermore, Cells without the calnexin/calreticulin central region are viable 4455 Fig. 4. S. pombe BiP/Cnx1p levels and co-immunoprecipitation in different strains. (A) Exponentially growing S. pombe cells expressing wild-type or mutant Cnx1p were incubated at 30°C and 37°C, with or without 4% (approx. 0.4 M) glycerol. Cell extracts were immunoprecipitated with anti-Cnx1p antibodies (lanes 1-6 and 9-14), with pre-immune rabbit serum (lane 7), or with rabbit pre-immune without proteinA-Sepharose (lane 8). The immunoprecipitated material was resolved by 12% SDS-PAGE and immunoblotted with anti-BiP antibodies. Molecular mass markers in kDa are indicated (MW). Note that S. pombe BiP appears as a doublet, with an upper band being the glycosylated form (approx. 10% of molecules), and the lower band the non-glycosylated form (Pidoux and Armstrong, 1993). The lower part of A at 30°C and 37°C, is the anti-Cnx1p western blotting of the immunoprecipitated material. The asterisk indicates that the deleted-Cnx1p and mini-Cnx1p bands correspond to a different part of the same gel. (B) Wild-type cnx1+ or mutants #4 and #11 were grown at various temperatures and BiP and Cnx1p were detected by western blotting with antiBiP or anti-Cnx1p antibodies on 5 µg of protein extracts. (C) Cells from wild-type cnx1+ or mutants #4 and #11 adapted at 37°C were shifted down to 30°C. Extracts were made from cells grown at both temperatures and the BiP and Cnx1p were detected by western blotting with ant-BiP or anti-Cnx1p antibodies. (D) Mutant strains #4 and #11 were cultured at 30°C or 37°C in the presence (+) or absence (−) of glycerol, as indicated. Western blot analyses to detect Cnx1p were done with 15 µg of protein extract instead of 5 µg using anti-Cnx1p antibodies. (E) Accumulation of BiP was quantified (Scion Image software, NIH) in different cnx1 mutants (#3, #16, #4 and #11) with respect to wild type (#2) at 30°C and at 37°C. 5 µg of protein extracts were used to quantify BiP levels by western blotting. The ratio with respect to wild-type cnx1+ (#2) is shown as a bar graph, and represents the average of two experiments at each temperature. the slow growth phenotype of cells #4 and #11 was reverted by retransformation of episomal copies of wild-type cnx1+ (#2) into these strains (not shown). Taken together, these results support the notion of a gradual adaptation of the physiology of the deleted-cnx and mini-cnx cells to the temperature stress, the mini-cnx strain being more affected than the deleted-cnx strain. 52 amino acids of the lumenal region of mini-Cnx1p are sufficient to form a complex that includes BiP We have previously shown that Cnx1p and BiP are found in a complex, under normal and ER-stress conditions, that may be involved in the folding of glycoproteins and non-glycosylated proteins (Jannatipour et al., 1998). Moreover, KAR2/BiP and CNE1 (the calnexin homologue) were shown to genetically interact in S. cerevisiae (Brodsky et al., 1999). To examine whether the calnexin mutants interact with BiP, we carried out immunoprecipitations with polyclonal anti-Cnx1p antibodies. After separation of the immunoprecipitated extracts by SDS- PAGE, the presence of BiP was analyzed by western blotting using polyclonal anti-BiP antibodies. As shown in Fig. 4A (lanes 1-6), BiP coprecipitated with full-length and mutant Cnx1p, including deleted-cnx and mini-cnx. This interaction is specific since no BiP was precipitated from SP556 total protein extracts with or without pre-immune serum (Fig. 4A, lanes 7 and 8). Moreover, since identical results were obtained after 2 hours incubation with cycloheximide, it seems unlikely that mini-Cnx1p could bind as a misfolded substrate to BiP. The decrease in the amount of BiP coprecipitated with miniCnx1p (#11) could be the consequence of the lower amounts of BiP (see Fig. 4E, lane 5) and/or lesser efficiency of miniCnx1p precipitation due the loss of epitopes in this mutant. As shown in Fig. 4A (lower panel of lanes 1-8) the anti-Cnx1p western blotting of the immunoprecipitates detected lower amounts of mini-Cnx1p. Overall, these observations suggest that these 52 aa of the lumenal region of the mini-Cnx1p are sufficient for the formation of a complex containing both Cnx1p and BiP. 4456 A. Elagöz and others Fig. 5. Cells depleted of the calnexin’s highly conserved central region exhibit S. pombe morphology that is reversible by 1.2 M sorbitol. Late-log phase S. pombe cells, grown at 30°C, were diluted into fresh EMM+Ade+Ura liquid medium and grown for 20 hours at 30°C. For cultures at 37°C, cells were grown for 20 hours for wild type, 30 hours for deleted-cnx, and 70 hours for mini-cnx. Phase-contrast micrographs of cells were then done on cells grown at 30°C (A), 37°C (B) or at 37°C in the presence of 1.2 M sorbitol (C). Deleted-cnx and mini-cnx strains were transformed with wild-type cnx1+, and grown in EMM+Ade at 37°C for 20 hours (D). Arrows indicate the vesicles appearing at 37°C. Phase-contrast microscopy was done with a Zeiss optical microscope, at a 2500× magnification on the film. Bar, 10 µm. The calnexin levels in deleted-cnx and mini-cnx strains drop at 37°C but BiP levels remain similar to those in cnx1+ cells To obtain clues at the molecular level regarding the mechanism allowing adaptation of deleted-cnx and mini-cnx cells to temperature stress (see Fig. 3), we performed western blot analyses with anti-BiP antibodies with extracts of strains #2, #4 and #11 grown at the exponential phase at 30°C, 32°C, 34°C and 37°C. Fig. 4B shows the results of these experiments. In strain #2 (full-length Cnx1p), the calnexin levels remain unchanged at different temperatures since its expression is driven by an heterologous promoter (nmt1). However, in strains #4 (deleted-cnx) and #11 (mini-cnx) the calnexin levels dropped considerably at 37°C (see arrows in Fig. 4B). Western blotting with anti-Cnx1p antibodies using 15 µg of cell extract instead of 5 µg confirmed the presence, albeit diminished, of the deleted-Cnx1p and mini-Cnx1p in cells grown at 37°C (see Fig. 4D, lanes 3 and 4). The reduction in the levels of mutant Cnx1p at elevated temperatures could be due to instability of these proteins in vivo or to proteolytic degradation in vitro. However, the second possibility could be excluded since the levels of wild-type Cnx1p (#2) remained unchanged under these conditions (see Fig. 4B). As expected for a heat-shock protein, BiP synthesis was increased at 37°C in the cnx1+ strain (see Fig. 4B). Interestingly, at 30°C the BiP levels were higher in mutants #4 and #11 than in wild type, but remained unchanged at different temperatures. The BiP and Cnx1p levels were practically restored when the cells adapted at 37°C were shifted down and cultured at 30°C (see arrows in Fig. 4C). To explore the possibility that an induction of BiP expression could compensate for the reduction of Cnx1p levels in strains #4 and #11 during adaptation at 37°C, we quantified by western blotting the BiP levels in the various cnx1 mutants at 30°C and 37°C. As shown in Fig. 4E, at 30°C and compared to wild type (#2), BiP accumulation is about 1.8-fold higher in the lumenal (#3) and lumenal+ADEL (#16), 1.6-fold in the deleted-cnx (#4) mutant and about 1.4-fold in the mini-cnx cells (#11). Interestingly, the BiP levels did not further increase to compensate for the drop in Cnx1p levels in strains #4 and #11 cultured at 37°C (Fig. 4D, lanes 1-4; and see below). This may suggest that perhaps other chaperones compensate for the reduction in the overall chaperone efficiency in the cnx1 mutants. Deleted-cnx and mini-cnx cells exhibit temperaturedependent altered morphology We assessed whether the various cnx1 mutations affected the morphology of S. pombe cells microscopically. Lumenal and lumenal+ADEL cells presented no visible difference in their morphology (not shown). In contrast, the deleted-cnx and minicnx cells grown at 37°C had an aberrant round shape morphology and the majority contained large vesicles (Fig. 5B, see arrow). The rounded shape of these cells suggested that the synthesis of a component(s) of their cell wall could be affected. A classical way to test this hypothesis is to grow cells in a hyperosmotic medium (Fanchiotti et al., 1998). As shown in Fig. 5C, the aberrant morphology of deleted-cnx and mini-cnx cells grown at 37°C was suppressed in the presence of 1.2 M sorbitol (Fig. 5C) or glycerol at 1.2 M (not shown), or by transforming the plasmid encoding wild-type Cnx1p (Fig. 5D). In order to study the cell wall integrity of the cnx1 mutants, exponentially growing cells were treated by NovoZym 234, an enzyme complex able to degrade S. pombe cell wall polymers (Ishiguro et al., 1997). As shown in Fig. 3H, deleted-cnx and mini-cnx cells exhibited increased sensitivity to the lytic enzymes at 37°C when compared to cnx1+ cells. The assembly of the yeast cell wall requires the biosynthesis and transport of glycoproteins and Cells without the calnexin/calreticulin central region are viable 4457 Fig. 6. Calcofluor staining of cnx1 mutants. Late-log phase S. pombe cells, grown at 30°C, were diluted into fresh EMM+Ade+Ura medium and grown for 20 hours at 30°C. For cultures at 37°C, cells were grown for 20 hours for wild type and mini-cnx strain transformed with wt cnx1+, 30 hours for deleted-cnx, and 70 hours for mini-cnx. Exponentially growing cells cultured at 30°C or at 37°C were stained with Calcofluor White (see Materials and Methods) (A,C). Nomarski interference images show the same field of cells stained with Calcofluor (B,D). Arrows indicate vesicles labeled with the dye. Microscopy was done with a Zeiss microscope, at a 1500× magnification on the film. Bars, 10 µm. β-glucans (Orlean, 1997). The antibiotic hygromycin B was reported to affect the viability of yeast cells with mutations affecting the early stages of glycoprotein biosynthesis, in particular at the ER (Dean, 1995; Silberstein et al., 1998). Hence, we tested the effect of this antibiotic on mutants #4 and #11. As it can be seen in Fig. 2H (right panel) at 30°C, the mutant strains showed increased sensitivity, while none of the strains grew at 37°C in the presence of the drug. To further study the cell wall in the cnx1 mutants, we used Calcofluor White. This fluorescent dye stains an unknown cell wall component of S. pombe cells that concentrates in the septum of dividing cells (Robinow and Hyams, 1989). The results of the microscopic observations are presented in Fig. 6. Calcofluor stained the septa of mutants #4 and #11 at 30°C and 37°C (Fig. 6A,C) in a similar manner to wild type (#2) or the mini-cnx cells also containing a plasmid expressing cnx1+. However, the large vesicles observed at 37°C in mutants #4 and #11 (see Figs 5, 6) were also labeled by the dye, thus suggesting that these vesicles might accumulate a Calcofluor-stainable cell-wall component. Taken together, these results suggest that the conserved central domain of Cnx1p may be required, directly or indirectly, for correct cell wall synthesis and morphology at 37°C. The slow growth phenotype of the deleted-cnx and mini-cnx strains can be rescued by glycerol and sorbitol The growth rate of deleted-cnx and mini-cnx cells is reduced at temperatures higher than 30°C (Fig. 3B,C). We observed, however, that these mutant cells grew faster in the presence of glycerol and sorbitol at 1.2 M. Therefore, we wished to examine in more detail the effect of these polyols on the growth of cnx1 mutants. Accordingly, cells cultured exponentially were subjected to the drop test for their ability to grow at 30°C and 37°C on minimal medium (EMM) or EMM supplemented with glycerol or sorbitol. The slow growth of deleted-cnx and mini-cnx cells at 37°C was rescued in the presence of 4% (approx. 0.4 M) glycerol (see Figs 7, 3D). It should be noted that in this case, glycerol does not act as a source of carbon since EMM contains 2% glucose. The same effect was observed in the presence of 0.4 M sorbitol (not shown). Because the osmolality of the growth medium was reported to affect the levels of molecular chaperones and the stability of certain proteins (Kültz et al., 1997), we next analyzed the accumulation of Cnx1p and BiP in the presence of glycerol or sorbitol. As shown in Fig. 4D (lanes 3-6) the levels of deletedCnx1p and mini-Cnx1p at 37°C were restored to those at 30°C when cells were cultured in the presence of 0.4 M glycerol or sorbitol (not shown). Interestingly, the addition of these polyols resulted in the enhancement of coprecipitated BiP with Cnx1p (see Fig. 4A, lanes 9-14). The deleted-cnx and mini-cnx strains cultured at 37°C are more sensitive to DTT, NaCl and Ca2+ The central-domain of the calnexin/calreticulin chaperones was shown to bind Ca2+, and it was proposed that membrane-bound calnexin could participate in the retention of ER-soluble proteins by anchoring a protein-calcium gel via its transmembrane domain (Wada et al., 1991; Tjoelker et al., 1994). As such, it could be expected that strains deleted of the central region of Cnx1p would be less tolerant to increased levels of Ca2+ in the medium. To assess this possibility, we performed drop tests with 10 mM Ca2+, at 30°C and 37°C. The growth of the deleted-cnx and mini-cnx strains remained unaffected in the presence of Ca2+ at 30°C (Fig. 7); however, it was severely inhibited at 37°C. This inhibition was totally rescued by the addition of 0.4 M glycerol (Fig. 7) or 0.4 M sorbitol (not shown). 4458 A. Elagöz and others Fig. 7. Effect of ER-stress on the growth of cnx1 mutants. Serial dilutions of exponentially growing S. pombe cells in liquid MM+Ade+Ura (EMM containing 2% glucose) medium at 30°C, were spotted on the same medium with or without 4% (approx. 0.4 M) glycerol and supplemented with either 10 mM CaCl2,1 mM DTT or 50 mM NaCl. The plates were incubated for 96 hours at 30°C or 37°C; strains used in this experiment are indicated. DTT is a reducing agent that induces protein misfolding in the ER (Braakman et al., 1992). It was of interest to evaluate the response of the various Cnx1p mutants to a DTT challenge. As shown in Fig. 7, in the presence of 1 mM DTT the growth of deleted-cnx and mini-cnx cells was affected at 30°C and completely inhibited at 37°C in the case of minicnx. Thus the deletion of the highly conserved central domain of Cnx1p compromised the cell’s tolerance to the combined effect of the two ER stresses (DTT and temperature; see arrows in Fig. 7). In the presence of NaCl at a concentration as low as 50 mM in EMM, the growth of deleted-cnx and mini-cnx cells was drastically affected at 37°C (see Fig. 7). The inhibitory effect of NaCl suggests that these mutants cells are impaired in their regulation of NaCl homeostasis. The sensitivity to both DTT and NaCl was partially suppressed by the addition of 0.4 M glycerol (see Fig. 7); however, 0.4 M sorbitol did not rescue the growth inhibition of mini-cnx cells at 37°C in the presence of these additives (not shown). DISCUSSION S. pombe cells deleted of the calnexin/calreticulin highly conserved central domain are viable The central region is a salient feature of the calnexin/calreticulin molecules studied thus far. Its roles in mammalian cells in the tethering of folding intermediates of glycoproteins through their oligosaccharide moieties, and in the binding of Ca2+, have been demonstrated (Tjoelker et al., 1994; Vassilakos et al., 1998). Based on these observations we assumed that this 317-aa region containing motifs 1 and 2 would encode the essential function(s) of S. pombe calnexin. Our results showed that under normal culture conditions (minimal medium at 30°C), the fission yeast cnx1 mutants (#4 and #11) lacking the central domain grow at the same rate as wild-type cells. Furthermore, the central region alone could not rescue the non-viability of cnx1∆ cells. These genetic observations may suggest that the interaction of monoglucosylated folding intermediates with the putative lectin determinant of S. pombe calnexin might not be absolutely required for the folding of glycoproteins, under normal growth conditions. Our results are consistent with four lines of evidence supporting this notion: (1) cells deleted of gtp1, the gene encoding glucosyl transferase, are viable at 30°C and 37°C (Fernandez et al., 1998); (2) Cnx1p interacts with acid phosphatase independently of glucose trimming and reglucosylation, i.e. the two pathways that generate monoglucosylated glycoproteins (Jannatipour et al., 1998); (3) cells genetically depleted of GT and glucosidase II (the enzyme that trims G2 glycans to G1 and G0), are viable at 30°C and 37°C (Fanchiotti et al., 1998); (4) an alg6∆/gpt1∆ double mutant strain (Alg6p is the enzyme that adds glucose to the lipid-linked glycan) is viable at 30°C. Nevertheless, strains #4 and #11, lacking the highly conserved central domain, exhibited reduced growth rates at 37°C. Because deleted-Cnx1p and mini-Cnx1p are correctly targeted to the ER (Fig. 2), the reduced growth rates of these cells at elevated temperature could be due in part to the observed decrease in Cnx1p levels (see Fig. 4). This is supported by the fact that the presence of 0.4 M glycerol or sorbitol corrected both the growth rates and the levels of miniCnx1p and deleted-Cnx1p. The stabilization of mutant Cnx1 proteins could be the result of general changes in chaperone levels and protein degradation activities induced by osmotic changes produced by these polyols (Kültz et al., 1997; Fernández et al., 1997). Another non-exclusive possibility could be that, in addition, glycerol might act as a chemical chaperone as previously reported in vitro (Gekko and Timasheff, 1981) and in vivo (Sato et al., 1996). In spite of the fact that cells without the central region are viable under standard conditions, this domain appears to be required when cells are cultured under high ER stress such as the combination of DTT and temperature. Our results are congruent with those of Fanchiotti et al. (1998), who showed that gpt1∆/alg6∆ double-mutant cells do not grow at 37°C, and proposed that the interaction between monoglucosylated folding intermediates and calnexin is essential under conditions of extreme ER stress. A possible role of the central domain in cell-wall biosynthesis Cells deleted of the central domain (strains #4 and #11) Cells without the calnexin/calreticulin central region are viable 4459 exhibited anomalous, round-shaped morphology at 37°C, that was suppressed by culturing the cells in the presence of 1.2 M sorbitol or glycerol. Moreover, these cnx1 mutants exhibited increased sensitivity to cell-wall lytic enzymes and to hygromycin B, suggesting that these cells are defective in the biosynthesis of some cell-wall component(s), and that Cnx1p could be involved in this process. Interestingly, in Saccharomyces cerevisiae, the ER glucosidase I (CWH41), glucosidase II (GLS2) and Kar2p/BiP were shown to be involved in β-1-6 glucan synthesis (Simons et al., 1998; Shahinian et al., 1998), one of the cell wall components. Thus it is tempting to speculate that in yeasts the calnexin cycle may play a major role in the biosynthesis of cell wall components, and a secondary one in the folding of glycoproteins, whereas in mammalian cells, which obviously lack a cell wall, the calnexin cycle evolved to be specialized in the folding of glycoproteins. The central domain is not required for Ca2+ homeostasis at 30°C The highly conserved central domain of calnexin/calreticulin has been shown to bind Ca2+ (Wada et al., 1991; Tjoelker et al., 1994), and it was also proposed that calnexin could participate in the retention of the ER soluble proteins by anchoring a protein-calcium gel via its transmembrane domain (Wada et al., 1991). The cells deleted of the central domain were viable in the presence of 10 mM Ca2+ at 30°C. However, the requirement for the Cnx1p central domain was manifested under ER-stress, thus arguing that the ensemble of ER-resident proteins do not provide sufficient Ca2+ buffering capacity to ensure viability under these conditions. The essential function(s) of Cnx1p is found within its 123 C-terminal residues We have shown that the essential function(s) of Cnx1p can be delimited to the 123 C-terminal amino acids. This segment spans the last 52 aa of the lumenal domain, the 23-aa transmembrane domain and the 48-aa cytosolic domain. When the mini-Cnx1p and lumenal-Cnx1p molecules are compared, the overlapping region corresponds to the C-terminal 52 aa of the lumenal domain. Moreover, lumenal versions of Cnx1p deleted of the 52 aa failed to support viability (constructs #17 and #18). Therefore, the essential function(s) of Cnx1p is likely to be located within this 52-aa stretch of the lumenal domain. Although this region displays the same basic organization as other calnexin molecules, no significant conservation at the level of aa sequence can be observed among different species. Our results may explain the observation that, in spite of the high degree of sequence conservation of the central domain and the similarities in the structural organization between mammalian and fission yeast calnexin, the canine homologue failed to rescue the lethal phenotype of cnx1∆ S. pombe cells (Parlati et al., 1995). This lack of complementation might be due to the poor similarity of sequence between these two proteins within the 52-aa segment. We have shown that the 52-aa stretch of Cnx1p is sufficient for the formation of a complex that includes BiP. Therefore, it is tempting to speculate that one of the crucial functions of Cnx1p may reside in the formation of this complex (perhaps along with other chaperones) that might be implicated in protein folding. Alternatively, this 52-aa stretch of the lumenal domain could still contain enough chaperone activity by itself to support viability of S. pombe. Further experiments are required to discriminate between these possibilities, and to explore whether this region encodes a yet to be defined function. The viability of a calnexin-less mammalian cell line has been interpreted as the result of a compensation by its soluble homologue calreticulin (Scott and Dawson, 1995). However, since glucose-trimming inhibitors do not completely obliterate the interaction of glycoproteins with calnexin/calreticulin, nor do they totally inhibit protein secretion, it may be suggested that lectin-independent glycoprotein folding pathways must exist in the ER of higher eukaryotes. Such a pathway might involve protein-protein interactions with calnexin, as it has been previously proposed, or with the assistance of other chaperones like BiP, akin to S. pombe (Williams, 1995; Jannatipour et al., 1998). In this regard, it should be noted that secretion of the fungal glycoprotein cellulase I is not diminished in S. pombe cells deleted of the calnexin central domain (our unpublished results). Complexes containing calnexin and BiP seem to exist in S. cerevisiae, since kar2/BiP mutants are synthetic with the deletion of CNE1/calnexin (Brodsky et al., 1999; Simons et al., 1998). Furthermore, calreticulin was shown to form a stable complex with BiP in tobacco (Crofts et al., 1998), and mammalian calnexin interacts with BiP in the absence of translation (Tatu and Helenius, 1997). Future endeavors should focus on verifying whether the central domain of Cnx1p displays lectin activity, and the nature of a putative chaperone function encoded by the 52-aa Cterminal region of the Cnx1p lumenal domain. We wish to express our appreciation to Dr Kathy Gould for providing the his3 strains and plasmids, and to Drs Guy Boileau, Michel Bouvier and Nabil Seidah for critical reading of the manuscript. We thank Mehrdad Jannatipour and Chai Ezerzer for their help with some of the constructions. Also we thank Dr Robert Nabi for helpful assistance with the microscopy. This work was supported by grants from the Medical Research Council of Canada (L.A.R.). J.A. is a Wellcome Trust Senior Fellow, M.C. is a Doctoral Fellow of the Cystic Fibrosis Foundation of Canada. REFERENCES Bergeron, J. J. M., Brenner, M. B., Thomas, D. Y. and Williams, D. B. (1994). Calnexin: a membrane-bound chaperone of the endoplasmic reticulum. Trends Biochem. Sci. 19, 124-128. Braakman, I., Helenius, J. and Helenius, A. (1992). Manipulating disulfide bond formation and protein folding in the endoplasmic reticulum. EMBO J. 11, 1717-1722. Brodsky, J. L., Werner, E. D., Dubas, M. E., Goeckeler, J. L., Kruse, K. B. and McCracken, A. A. (1999). The requirement for molecular chaperones during endoplasmic reticulum-associated protein degradation demonstrates that protein export and import are mechanistically distinct. J. Biol. Chem. 274, 3453-3460. Burke, J. D. and Gould, K. L. (1994). Molecular cloning and characterization of the Schizosaccharomyces pombe his3 gene for use as a selectable marker. Mol. Gen. Genet. 242, 169-176. Crofts, A., Lebrogne-Castel, N., Pesca, M., Vitale, A. and Denecke, J. (1998). BiP and calreticulin form an abundant complex that is independent of endoplasmic reticulum stress. Plant Cell 10, 813-823. Dean, N. (1995). Yeast glycosylation mutants are sensitive to aminoglycosides. Proc. Natl. Acad. Sci. USA 92, 1287-1291. Elliott, J. G., Oliver, J. D. and High, S. (1997). The thiol-dependent reductase ERp57 interacts specifically with N-glycosylated integral proteins. J. Biol. Chem. 272, 13849-13855. Fanchiotti, S., Fernandez, F., D’Alessio, C. and Parodi, A. J. (1998). The UDP-Glc:Glycoprotein glucosyltransferase is essential for 4460 A. Elagöz and others Schizosaccharomyces pombe viability under conditions of extreme endoplasmic reticulum stress. J. Cell Biol. 143, 625-635. Fernandez, F., D’Alessio, C., Fanchiotti, S. and Parodi, A. J. (1998). A misfolded protein conformation is not a sufficient condition for in vivo glucosylation by the UDP-Glc:glycoprotein glucosyltransferase. EMBO J. 17, 5877-5886. Fernández, F. S., Jannatipour, M., Hellman, U., Rokeach, L. A. and Parodi, A. J. (1996). A new stress protein: Synthesis of Schzosaccharomyces pombe UDP-Glc:glycoprotein glucosyltransferase mRNA is induced by stress conditions but the enzyme is not essential for cell viability. EMBO J. 15, 705-713. Fernández, F. S., Trombetta, S. E., Hellman, U. and Parodi, A. J. (1994). Purification to homogeneity of UDP-glucose:glycoprotein glucosyltransferase from Schizosaccharomyces pombe and apparent absence of the enzyme from Saccharomyces cerevisiae. J. Biol. Chem. 269, 30701-30706. Fernández, J., Soto, T., Vicente-Soler, J., Cansado, J. and Gacto, M. (1997). Osmo-stress-induced changes in neutral trehalase activity of the fission yeast Schizosaccharomyces pombe. Biochim. Biophys. Acta 1357, 41-48. Gekko, K. and Timasheff, S. N. (1981). Mechanism of protein stabilization by glycerol: preferential hydration in glycerol-water mixtures. Biochem. 20, 4667-4676. Hammond, C. and Helenius, A. (1994). Folding of VSV protein: sequential interaction with BiP and calnexin. Science 266, 456-458. Higuchi, R. and Krummel, B. S. R. K. (1988). A general method of in vitro preparation and specific mutagenesis of DNA fragments: study of protein and DNA interactions. Nucleic Acids Res. 16, 7351-7367. Ishiguro, J., Saitou, A., Duran, A. and Ribas, J. C. (1997). cps1 a Schizosaccharomyces pombe gene homolog of Saccharomyces cerevisiae FKS genes whose mutation confers hypersensitivity to cyclosporin A and papulacandin B. J. Bacteriol. 1977, 7653-7662. Jannatipour, M., Callejo, M., Parodi, A. J., Armstrong, J. and Rokeach, L. A. (1998). Calnexin and BiP interact with acid phophatase independently of glucose trimming and reglucosylation in S. pombe. Biochemistry 37, 17253-17261. Jannatipour, M. and Rokeach, L. A. (1995). The Schizosaccharomyces pombe homologue of the chaperone calnexin is essential for viability. J. Biol. Chem. 270, 4845-4853. Keller, S. H., Lindstrom, J. and Taylor, P. (1998). Inhibition of glucose trimming with castanospermine reduces calnexin association and promotes proteasome degradation of the α-subunit of the nicotinic acetylcholine receptor. J. Biol. Chem. 273, 17064-17072. Kim, P. S. and Arvan, P. (1995). Calnexin and BiP act as sequential molecular chaperones during thyroglobulin folding in the endoplasmic reticulum. J. Cell Biol. 128, 29-38. Kültz, D., Garcia-Perez, A., Ferraris, J. D. and Burg, M. B. (1997). Distinct regulation of osmoprotective genes in yeast and mammals. J. Biol. Chem. 272, 13165-13170. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. Le, A., Steiner, J. L., Ferrel, G. A., Shaker, J. C. and Sifers, R. N. (1994). Association between calnexin and a secretion-incompetent variant of human α1-antitrypsin. J. Biol. Chem. 269, 7514-7519. Loo, T. W. and Clarke, D. M. (1994). Prolonged association of temperaturesensitive mutants of human P-glycoprotein with calnexin during biogenesis. J. Biol. Chem. 269, 28683-28689. Maundrell, K. (1993). Thiamine-repressible expression vectors pREP and pRIP for fission yeast. Gene 123, 127-130. Moreno, S., Klar, A. and Nurse, P. (1991). Molecular genetic analysis of fission yeast Schizosaccharomyces pombe. Meth. Enzymol. 194, 795-823. Oliver, J. D., van der Wal, F. J., Bulleid, N. J. and High, S. (1997). Interaction of the thiol-dependent reductase ERp57 with nascent glycoproteins. Science 275, 86-88. Orlean, P. (1997). Biogenesis of yeast wall and surface components. In Cell Cycle and Cell Biology (ed. J. R. Pringle, J. R. Broach and E. W. Jones), pp. 229-362. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. Parlati, F., Dignard, D., Bergeron, J. J. M. and Thomas, D. Y. (1995). The calnexin homologue cnx1+ in Schizosaccharomyces pombe, is an essential gene which can be complemented by its soluble ER domain. EMBO J. 14, 3064-3072. Parodi, A. J. (1998). Reglucosylation of glycoproteins and quality control of glycoprotein folding in the endoplasmic reticulum of yeast cells. Biochim. Biophys. Acta 1426, 287-295. Peterson, J. R. and Helenius, A. (1999). In vitro reconstitution of calreticulinsubstrate interactions. J. Cell Sci. 112, 2775-2784. Pidoux, A. L. and Armstrong, J. (1992). Analysis of the BiP and identification of an ER retention signal in Schizosaccharomyces pombe. EMBO J. 11, 1583-1591. Pidoux, A. L. and Armstrong, J. (1993). The BiP protein and the endoplasmic reticulum of Schizosaccharomyces pombe: fate of the nuclear envelope during cell division. J. Cell Sci. 105, 1115-1120. Pipe, S. W., Morris, J. A., Shah, J. and Kaufman, R. J. (1998). Differential interaction of coagulation factor VIII and factor V with protein chaperones calnexin and calreticulin. J. Biol. Chem. 273, 8537-8544. Robinow, C. F. and Hyams, J. S. (1989). General cytology of fission yeast. In Molecular Biology of the Fission Yeast (ed. A. Nasim, P. Young and B. F. Johnson), pp. 273-330. London: Academic Press Limited. Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Sato, S., Ward, C. L., Krouse, M. E., Wine, J. J. and Kopito, R. R. (1996). Glycerol reverses the misfolding phenotype of the most common cystic fibrosis mutation. J. Biol. Chem. 271, 635-638. Scott, J. E. and Dawson, J. R. (1995). MHC class 1 expression and transport in a calnexin-deficient cell line. J. Immunol. 155, 143-148. Shahinian, S., Dijkgraaf, G. J., Sdicu, A.-M., Thomas, D. Y., Jakob, C. A., Aebi, M. and Bussey, H. (1998). Involvement of protein N-glycosylation and processing in the biosynthesis of cell wall β-1,6-glucan of Saccharomyces cerevisiae. Genetics 149, 843-856. Silberstein, S., Schlenstedt, G., Silver, P. A. and Gilmore, R. (1998). A role for the DnaJ homologue Scj1p in protein folding in the yeast endoplasmic reticulum. J. Cell Biol. 143, 921-933. Simons, J. F., Ebersold, M. and Helenius, A. (1998). Cell wall 1,6-β-glucan synthesis in Saccharomyces cerevisiae depends on ER glucosidases I and II and the molecular chaperone BiP/Kar2p. EMBO J. 17, 396-405. Sousa, M. and Parodi, A. J. (1995). The molecular basis for the recognition of misfolded glycoproteins by the UDP-Glc:glycoprotein glucosyltransferase. EMBO J. 14, 4196-4203. Tatu, U. and Helenius, A. (1997). Interactions between newly synthesized glycoproteins, calnexin and a network of resident chaperones in the endoplasmic reticulum. J. Cell Biol. 136, 555-565. Tjoelker, L. W., Seyfried, C. E., Eddy, R. L., Jr., Byers, M. G., Shows, T. B., Calderon, J., Scheiber, R. B. and Gray, P. W. (1994). Human, mouse, and rat calnexin cDNA cloning: Identification of potential calcium binding motifs and gene localization to human chromosome 5. Biochemistry 33, 3229-3236. Trombetta, E. S. and Helenius, A. (1998). Lectins as chaperones in glycoprotein folding. Curr. Opin. Struct. Biol. 8, 587-592. Vassilakos, A., Cohen-Doyle, M. F., Peterson, P. A., Jackson, M. R. and Williams, D. B. (1996). The molecular chaperone calnexin facilitates folding and assembly of class 1 histocompatibility molecules. EMBO J. 15, 1495-1506. Vassilakos, A., Michalak, M., Lehrman, M. A. and Williams, D. B. (1998). Oligosaccharide binding characteristics of the molecular chaperones calnexin and calreticulin. Biochemistry 37, 3480-3490. Wada, I., Rindress, D., Cameron, P. H., Ou, W. J., Doherty II, J. J., Louvard, D., Bell, A. W., Dignard, D., Thomas, D. Y. and Bergeron, J. J. M. (1991). SSRgamma and associated calnexin are major calcium binding proteins of the endoplasmic reticulum membrane. J. Biol. Chem. 266, 19599-19610. Ware, F. E., Vassilakos, A., Peterson, P. A., Jackson, M. R., Lehrman, M. A. and Williams, D. B. (1995). The Molecular Chaperone Calnexin Binds Glc1Man9GlcNAc2 Oligosaccharide as an Initial Step in Recognizing Unfolded Glycoproteins. J. Biol. Chem. 270, 4697-4704. Williams, D. B. (1995). The Merck Frosst award lecture 1994: Calnexin, a molecular chaperone with a taste for carbohydrate. Biochem. Cell Biol. 73, 123-132. Zapun, A., Darby, N. J., Tessier, D. C., Michalak, M., Bergeron, J. J. M. and Thomas, D. Y. (1998). Enhanced catalysis of ribonuclease B folding by the interaction of calnexin or calreticulin with ERp57. J. Biol. Chem. 273, 6009-6012.