Bgs3p, a putative 1,3-beta-glucan synthase subunit, is required for cell wall assembly in Schizosaccharomyces pombe

EUKARYOTIC CELL, Feb. 2003, p. 159–169 Vol. 2, No. 11535-9778/03/$08.00�0 DOI: 10.1128/EC.2.1.159–169.2003Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Bgs3p, a Putative 1,3-�-Glucan Synthase Subunit, Is Required for CellWall Assembly in Schizosaccharomyces pombe

Victoria Martín, Blanca García, Elena Carnero, Angel Duran, and Yolanda Sanchez*Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, and Departamento de Microbiología

y Genetica, Universidad de Salamanca, 37007 Salamanca, Spain

Received 31 July 2002/Accepted 9 October 2002

�-Glucans are the main components of the fungal cell wall. Fission yeast possesses a family of �-glucansynthase-related genes. We describe here the cloning and characterization of bgs3�, a new member of thisfamily. bgs3� was cloned as a suppressor of a mutant hypersensitive to Echinocandin and Calcofluor White,drugs that interfere with cell wall biosynthesis. Disruption of the gene is lethal, and a decrease in Bgs3p levelsleads to rounded cells with thicker walls, slightly reduces the amount of the �-glucan, and raises the amountof �-glucan polymer. These cells finally died. bgs3� is expressed in vegetative cells grown in different conditionsand during mating and germination and is not enhanced by stress situations. Consistent with the observedexpression pattern, Bgs3-green fluorescence protein (GFP-Bgs3p) was found at the growing tips duringinterphase and at the septum prior to cytokinesis, always localized to growth areas. We also found GFP-Bgs3pin mating projections, during the early stages of zygote formation, and at the growing pole during ascosporegermination. We conclude that Bgs3p localization is restricted to growth areas and that Bgs3p is a glucansynthase homologue required for cell wall biosynthesis and cell elongation in the fission yeast life cycle.

Fission yeast, with its well-defined cylindrical shape, pro-vides an ideal system to study cell morphogenesis. In optimalconditions, cell growth is achieved in a polarized fashion by tipelongation along the main axis, which remains constant indiameter. At the onset of mitosis, polarized cell growth abatesand septum deposition occurs in the middle of the cell, fol-lowed by medial fission (37). Starvation induces haploid cells ofopposite mating types to mate in pairs, forming diploid zy-gotes. The zygotes then undergo meiosis to form four roundedhaploid spores, which germinate in a polarized fashion whennutrient conditions improve (49). The job of the cell wall is topreserve the osmotic integrity of cells and determine cellularmorphology during tip elongation, septation, mating, sporula-tion, and germination growth (18). In all of these morphoge-netic changes, continuous cell wall polymer synthesis is re-quired for viability. The fact that such processes lackhomologous counterparts in mammalian cells makes cell wallbiosynthetic enzymes an attractive, and in some cases essential,antifungal target (14).

The Schizosaccharomyces pombe cell wall is an extracellularmatrix with a layered organization consisting of an outer layerof glycoproteins (9 to 14% �-galactomannan) and an innerlayer of carbohydrate polymers. This carbohydrate layer ismainly composed of 1,3-�-glucan (48%) interwoven into afibrillar network with 1,3-�-glucan (18 to 28%), although it alsocontains some 1,6-�-glucan (2%) (27). Among the polysaccha-rides, 1,3-�-glucans are the most prevalent, and it is generallyaccepted that they are the main structural components respon-sible for cell wall rigidity (27). In situ localization of three

different types of �-glucan has indicated that 1,6-�-branched1,3-�-glucan and 1,6-�-glucan are located throughout the glu-can layer (less-dense layer), whereas the linear 1,3-�-glucan islocated exclusively in the primary septum of dividing cells (16).

The enzyme complex that catalyses the synthesis of 1,3-�-glucan chains is 1,3-�-glucan synthase (GS). GS activity hasbeen characterized mainly in Saccharomyces cerevisiae (33).Based on studies of mutants resistant and hypersensitive to GSinhibitors, gene disruption experiments, and biochemical char-acterization, the following model has been proposed (reviewedin reference 4 and 10). S. cerevisiae GS is composed of at leasttwo proteins: (i) the putative catalytic subunit, a large-molec-ular-size (�200-kDa) polypeptide with 16 transmembrane do-mains encoded by two genes, FKS1 and FKS2 (11, 17, 32), and(ii) a small GTP-binding subunit, Rho1p, which stimulates GSactivity in its prenylated form (12, 40). FKS1 expression is cellcycle regulated and more abundant during vegetative growth,whereas FKS2 expression is calcineurin dependent and impor-tant for efficient sporulation (32, 50). A third FKS homologue,FKS3, is dispensable and is not necessary for growth or sporu-lation (A. Ram, cited as a personal communication in refer-ence 45). Genes with strong homology to FKS1 have beenidentified from the fungal pathogens Candida albicans (GSC1,GSL1, and GSL2), Aspergillus nidulans (fksA), Aspergillus fu-migatus (AfFks1), Cryptococcus neoformans (CnFKS1), Para-coccidioides brasiliensis (PbFKS1), and Pneumocystis carinii(Pcgsc-1) (for a review, see reference 10).

At least four genes sharing homology with 1,3-�-GS catalyticsubunits have been identified in S. pombe. cps1�/bgs1�, bgs2�,bgs3� (the present study), and bgs4� (cosmid c1840). Of these,Cps1/Bgs1p is presumably involved in the assembly of theseptum-polymer 1,3-�-glucan, although no catalytic activity hasbeen directly demonstrated to date. Certain mutations in bgs1�

confer hypersensitivity to cyclosporine and papulacandin B, aninhibitor of 1,3-�-GS (20), whereas other mutants are unable

* Corresponding author. Mailing address: Instituto de Microbi-ología Bioquímica, CSIC/Universidad de Salamanca, Edificio Depar-tamental, Room 231, Campus Miguel de Unamuno, 37007 Salamanca,Spain. Phone: 34-923-121589. Fax: 34-923-224876. E-mail: [email protected].

159

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

to assemble the division septum (23, 26). It has recently beenshown that Bgs1p is localized to the cell division site in amanner dependent on the actomyosin ring and the septation-inducing network (25). In addition, Bgs1p has been localized tothe sites of growth (6). Bgs2p is a sporulation-specific GSrequired for proper ascospore wall maturation (24, 29). bgs2�sporulating diploids show a defect in GS activity and fail toassemble the cell wall properly, resulting in a failure to developviable ascospores (29). The presence of two more genes en-coding 1,3-�-GS isoforms suggests that each may perform adistinct function. Alternatively, even though Bgs1p seems to beimportant for septation and Bgs2p is necessary for sporulation,it is possible that these other two isoforms perform overlappingroles in cell wall assembly. GS activities are regulated in acomplex manner by the activities of the Rho family GTPasesand protein kinase C (reviewed in reference 1). Under- oroverproduction of these regulators dramatically affects S.pombe morphology and cell wall integrity. In contrast, almostnothing is known about how the various components are or-ganized into a functional unit subsequent to their synthesis.

Here we report the characterization of bgs3�. We presentevidence that bgs3�, which encodes a �-GS homologue, isessential for cell growth. Bgs3p is always found in the growthareas and cannot be substituted by overproduction of any ofthe other GS homologues, suggesting a unique role for Bgs3pin the biosynthesis of 1,3-�-glucans where new cell wall depo-sition is necessary.

MATERIALS AND METHODS

Strains, media, and chemicals. The genotypes of the S. pombe strains used inthe present study are listed in Table 1. Complete yeast growth medium (YES),selective minimal medium (MM) supplemented with the appropriate require-ments, and sporulation medium (SPA) have been described elsewhere (34).Echinocandin LY280949 (Ech; Lilly Company) (41) was stored at �20°C in astock solution (2.5 mg/ml) in 50% ethanol and was added to the medium after anautoclaving step at the corresponding final concentration. Calcofluor White (Cf)was prepared (15 mg/ml) in water with a few drops of 10 N KOH, filter sterilized,and added as described above to MM medium or to YES medium, the latterpreviously buffered with 50 mM potassium hydrogen phthalate (pH 6.1). Generalyeast cultures and genetic manipulations were as described previously (34).Escherichia coli DH5� was used for plasmid propagation. E. coli CJ236 andMV1190 were used for in vitro site-directed mutagenesis. Luria-Bertani and2xYT media (43) were supplemented with 50 �g of ampicillin or 25 �g ofkanamycin/ml when appropriate.

Cloning of the bgs3� gene. ehs2-1 mutants (strain GI1) were transformed witha fission yeast genomic library constructed in plasmid pDB248 (15); 30,000 Leu�

transformants were selected and replated on medium containing 4 �g of Ech/mlto select for plasmids conferring drug resistance. The screening yielded eighttransformants resistant to Ech; six of them were also resistant to Cf. Six plasmids(each from one transformant) rescued the Ech- or Cf-hypersensitive phenotype.They all contained the same 12.5-kb insert. One plasmid, pBG1, was chosen forrestriction analysis (Fig. 1). Fragments of the insert were subcloned into pAL-KS(20) and assayed for the ability to complement drug hypersensitivity when rein-troduced into strain GI1. Activity was localized to an approximately 9.8-kbSpeI-SpeI fragment subcloned in pAL-KS (pBG17). To identify a smaller activesubclone, it was necessary to replace the ClaI site by a SacII site by site-directedmutagenesis, followed by digestion of the plasmid created with SacII and reli-gation to obtain an insert of �8 kb in plasmid pJR33 (Fig. 1). The DNAsequence of the fragment contained in pJR33 was performed with custom-synthesized primers purchased from Isogen. Sequencing was performed on aABI 377 sequencer (Applied Biosystems, Inc.) by using the Taq DyeDeoxyterminator cycle sequencing kit, as supplied by the manufacturer. The DNAsequence reported here has been deposited in the EMBL database under acces-sion number AJ249371. To check whether bgs3� was the gene affected by theehs2-1 mutation, we subcloned a SalI-ClaI fragment from pBG1 into the SalI-SmaI sites of the integrative vector pJK148 carrying the leu1� marker. Theplasmid obtained was linearized with EcoRI and transformed into the PN22strain. Integrants were isolated and tested for stability. A cross between theSalI-ClaI integrant strain and the ehs2-1 strain revealed 4PD, 12NPD, and 20T,out of 36 tetrads, indicating that they were different loci.

Gene disruption of bgs3�. The bgs3::ura4� disruption construct was a three-step process. The 5-noncoding region of the bgs3� open reading frame (ORF)nucleotides (nt) �1970 to �25 was obtained via PCR by inserting the ApaI andSalI sites, one at each end, and was ligated into the same sites of the SK vectorto yield pJR34. The 3-flanking region of the bgs3� ORF (nt 5481 to 7470) wasobtained by PCR, inserting the PstI and BamHI sites as described above, and wascloned into the same sites of pJR34 to yield pJR35. A 1.7-kb SalI-PstI fragmentcontaining the ura4� gene was cloned into the same sites of pJR35, creatingpBG40. bgs3� gene disruption was accomplished by using the 5.7-kb fragmentfrom pBG40 cut with ApaI and BamHI and transforming the MS75 diploidstrain. Transformants were replica plated five consecutive times on YES mediumin order to eliminate the cells that had not integrated the construct. The correctintegration was then analyzed by PCR by using the following oligonucleotides:M13 (5-CTGGTGGCCTTAGGTA-3) in the ura4� gene, B53 (5-GGTTATTAAAGCAAATTGCAC-3) upstream from nt �2235 and therefore external tothe disruption cassette, and B11 (5-GGTGAATCATAAGCATCC-3) in thebgs3� gene. Correct integrations were also confirmed by genomic Southernblotting. In both cases, tetrad analysis of the heterozygous diploid disclosed twoviable (Ura�) and two unviable spores, indicating that the bgs3� gene is essentialfor viability.

Shutoff and overexpression of the bgs3� gene. For shutoff and overexpression,we made use of the regulatable nmt promoters (13, 31). To overexpress bgs3�, anmt3X-bgs3�-Kanr strain was constructed by fusion of the nmt promoter to thebgs3� ORF by a PCR-based method with the primers 5-GGAACTCTTTGGATAACAAAGCGACATTGGGAACGTGGCAATTCTGTCGTGGTTTA

TABLE 1. Strain list

Strain Genotypes Source

PN22 h� leu1-32 S. MorenoGI1 h� leu1-32 ehs2-1 This workMS38 h� leu1-32 ade6M210 ura4D-18 his3D1 This workMS75 h�/h� leu1-32/leu1-32 ade6M210/ade6M216 ura4D-18/ura4D-18 his3D1/his3D1 This workMS300 h�/h� leu1-32/leu1-32 ade6M210/ade6M216, ura4D-18/bgs3::ura4� his3D1/his3D1 This workMS301 h� leu1-32 ade6M210 ura4D-18 his3D1 41 nmt-bgs3�-ura4� This workMS302 h� leu1-32 ade6M210 ura4D-18 his3D1 GFP-bgs3�-ura4� This workBG30 h� leu1-32 nmt-bgs3�-kan This workMS37 h� leu1-32 his3D1 This workHVP365 h90 leu1-32 H. ValdiviesoSM11 h� leu1-32 cdc10-129 S. MorenoPN35 h� leu1-32 ura4D-18 cdc25-22 S. MorenoPPG313 h� leu1-32 ade6M216 orb6-25 P. PerezPPG324 h� leu1-32 tea1-50 P. PerezMS170 h� leu1-32 cps8-188 J. Ishiguro

160 MARTIN ET AL. EUKARYOT. CELL

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

TAAAGTGAATCTTGGTGAATTCGAGCTCGTTTAAAC-3, 5-GAATTTATGGTAGCATCATCGCTAAGGATTTGAATAGAAGAAGATGGCGAATGCTCACCCTTTTTATAATCCATCATGATTTAACAAAGCGACTATA-3, and the pFA6a-kanMX6-P3nmt1 modules described elsewhere (3). Forshutoff, a ura4�-41nmt1-bgs3� strain was constructed by using plasmidpMS54 and the one-step gene replacement. pMS54 is an SK-based vectorcontaining a 2-kb fragment with 5 bgs3� sequences (cloned in the SpeI siteof the multiple cloning site), the ura4� marker fragment cloned in BamHI-PstI, and the 41nmt1 promoter (cloned in PstI-SalI), followed by a 3-kbfragment of bgs3� ORF (nt �3 to �2670) cloned in the SalI-KpnI sites. ASacI-KpnI fragment was used to transform a diploid strain (MS75), andcorrect integration was analyzed by PCR and Southern blotting. We thenselected for haploids in MM without thiamine and uracil.

Epitope tagging. For epitope tagging, we used a 115-bp NotI-NotI DNAfragment containing three repeats of the hemagglutinin (HA) epitope (48).pJR33 was modified to create a NotI site at nt 5478 of the bgs3� coding sequence(pBG5), allowing in-frame insertion of the HA epitope at amino acid Glu-1826of the Bgs3p (pBG11). pBG11 fully rescued the lethality of the bgs3� strains.

Strains containing a chromosomal copy of bgs3� tagged with the green fluo-rescence protein (GFP-bgs3�) were obtained by one-step gene replacement.First, the GFP was fused to the N terminus of Bgs3p to make pMS60. pMS60 wasobtained by introducing a NotI site at amino acid �2 of the Bgs3p (pJR33) andcloning the GFP in that site. pMS60 fully rescued the lethality of the bgs3�strains. Then, a SacI-KpnI fragment from pMS60 containing the bgs3� promoterand the GFP-Bgs3p was subcloned into the PstI-KpnI sites of pMS54 (describedabove) to yield pMS65. A SacI-KpnI fragment of pMS65 was used to transforma haploid strain MS38, and correct integration of the fusion protein was analyzedby PCR and Southern blotting.

Cell wall analysis. For enzymatic lysis of cell suspensions, early-logarithmic-phase cells grown in MM at 28°C (with or without thiamine) were collected,washed with 50 mM citrate phosphate buffer (pH 5.6), suspended in a smallvolume of the same buffer, and adjusted to an optical density at 600 nm (OD600)of 4.0. Cell suspensions were treated with Novozyme 234 (200 �g/ml) at 28°Cwith shaking. The OD600 was monitored at the indicated times and then nor-malized relative to the absorbance of a control sample of each strain withoutenzyme at each time point.

For labeling and fractionation of cell polysaccharides, exponentially growingcultures incubated at 28°C in MM (with or without thiamine) were diluted withthe same medium, supplemented with D-[U-14C]glucose (3 �Ci/ml), and incu-bated overnight. Total glucose incorporation was monitored by measuring ra-dioactivity in trichloroacetic acid-insoluble material. Cells were harvested in theearly logarithmic phase, supplemented with unlabeled cells as the carrier, washedtwice with 1 mM EDTA, and broken with glass beads in a Fast-Prep, with three20-s spins at a speed setting of 6. Cell walls were purified by repeated washing

and differential centrifugation (once with 1 mM EDTA, twice with 5 M NaCl,and twice with 1 mM EDTA) at 1,000 g for 5 min. Finally, the purified cellwalls were heated to 100°C for 5 min. Cell wall samples were extracted with 6%NaOH for 60 min at 80°C and neutralized with acetic acid. Precipitation of thegalactomannan fraction from the neutralized supernatant was performed withthe Fehling reagent, adding unlabeled yeast mannan as the carrier. Other sam-ples of the cell wall suspension were incubated with Zymolyase 100T (250 �g ofenzyme and a cell volume equivalent to 150 �l of the initial cell culture; Seika-gaku Kogyo Co.) in 50 mM citrate phosphate buffer (pH 5.69) for 36 h at 30°C.Samples without enzyme were included as controls. After incubation, sampleswere centrifuged and washed with the same buffer. One milliliter of 10% tri-chloroacetic acid was added to the pellets, and their radioactivity was measured.The pellets were considered as the 1,3-�-glucan fraction, and supernatants wereconsidered as the �-glucan-plus-galactomannan fraction. �-Glucan was also cal-culated as the radioactivity remaining after subtraction of galactomannan and1,3-�-glucan from total cell wall incorporation. All determinations were carriedout in duplicate, and data for each strain were calculated from three independentcultures.

Enzyme preparation and 1,3-�-GS assays were performed basically as de-scribed previously (29).

Western blotting. Bgs3-HAp expressed in S. pombe cells was detected byWestern blotting. Cell extracts were prepared from 108 log-phase cells, sus-pended in 100 �l of lysis buffer (100 mM Tris-HCl [pH 8], 100 mM NaCl, and0.1% Triton X-100 containing 100 �M p-aminophenyl methanesulfonyl fluorideand 1 �g of pepstatin, leupeptin, and aprotinin/ml), and broken with glass beadson a FastPrep FP120 apparatus (Savant; Bio 101) twice, with a 15-s pulse at aspeed of 5. The resulting homogenates were collected, and the glass beads andbulky debris were removed by centrifugation at low speed (5,000 g for 20 s at4°C). Protein amounts were quantified by using the protein assay kit fromBio-Rad. One volume of 2 loading buffer (42) was added to each sample.Samples were denatured by boiling for 5 min, and equivalent amounts of eachsample (50 �g) were subjected to sodium dodecyl sulfate–7.5% polyacrylamidegel electrophoresis. After electrophoresis, proteins were blotted onto an Immo-bilon-P membrane (Millipore) and incubated with anti-HA immunoglobulin(12CA5; Boehringer Mannheim), and immunodetection was accomplished byusing the ECL detection kit (Amersham Corp).

Microscopy. The localization of GFP-Bgs3p was visualized in living cells. ForCf staining of cell walls and septum, 1 ml of cells was harvested, washed once with1 ml of phosphate-buffered saline (PBS), and resuspended in 100 �l of PBS withCf at 20 �g/ml for 5 min at room temperature. For nucleus staining, 1 ml of cellswas fixed with paraformaldehyde (4%), washed three times with PBS (1 ml), andresuspended in 10 �l of PBS containing DAPI (4,6-diamidino-2-phenylindole).Samples were observed under a Leica DM RXA microscope.

For electron microscopy, cells from strain MS301, treated or not treated with

FIG. 1. (A) Restriction map and subcloning of a DNA fragment that complements the ehs2-1 mutant phenotypes. The ClaI* site originally inthe sequence was substituted by a SacII site by mutagenesis. The arrow represents the predicted bgs3� ORF. The ability of subclones tocomplement ehs2-1 mutant phenotypes is indicated as either “�” (rescue) or “�” (no rescue). (B) Complementation of Ech- and Cf-hypersensitivephenotypes by plasmid pJR33. GI1 (h� leu1-32 ehs2-1) cells were transformed with pJR33 or pAL, and transformants were streaked out onto MMplates in the presence or absence of Ech (1 �g/ml) or Cf (1 mg/ml). Plates were incubated at 28°C for 4 days.

VOL. 2, 2003 GLUCAN SYNTHASE IS ESSENTIAL IN S. POMBE 161

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

thiamine for 24 h, were washed and fixed with 3% glutaraldehyde in 0.1 Mpotassium phosphate (pH 7) for 1 h. Cells were embedded in agar and treatedwith 1% osmium tetroxide for 2 h. Fixed cells were washed and dehydratedthrough a graded series of acetone (10 to 100%) and then embedded in Spurrresin. Thin sections were stained with uranyl acetate and Reynold’s lead citrateand examined under a Zeiss EM900 electron microscope. Images were processedwith Adobe Photoshop software.

RESULTS

Cloning of S. pombe bgs3� by complementation of the ehs2-1mutant. To identify the fission yeast genes involved in glucanbiosynthesis, we started a search for mutants hypersensitive tothe cell wall inhibitors Cf and Ech (5). The rationale behindthis approach is that mutants with a weakened cell wall cannotwithstand the additional disturbance caused by these drugs anddie at lower concentrations of the antifungal agents than cellswith a normal cell wall (22).

In the present study, we cloned the gene that complementsthe hypersensitivity of one of these mutants: the ehs2-1 mutant(for Ech hypersensitive). ehs2-1 was unable to grow at 1 �g ofEch or 0.1 mg of Cf/ml (Fig. 1B), whereas the wild-type strainwas able to withstand concentrations of 7.5 �g of Ech and 1.5mg of Cf/ml. In addition, the mutant cells showed a lytic ther-mosensitive phenotype at 37°C, which was suppressed when anosmotic stabilizer (1.2 M sorbitol) was added to the medium(data not shown). The ehs2-1 mutant was transformed with afission yeast genomic library constructed in plasmid pDB248(15). Thirty thousand transformants were screened for comple-mentation of the Ech-hypersensitive phenotype. The screeningyielded eight rescuing plasmids. Six plasmids contained thesame fragment of DNA (see Materials and Methods). One ofthose plasmids, pBG1, containing a 12.5-kb insert was able tocomplement the Ech- and Cf-hypersensitive phenotypes andwas chosen for restriction analysis (Fig. 1A). The smallestsubclone able to complement ehs2-1 was the 8-kb SacII-SpeI(pJR33) (Fig. 1). Sequencing of this fragment revealed a largeuninterrupted ORF of 5,481 nt. The protein sequence pre-dicted from this fragment showed strong homology with the S.cerevisiae FKS1 and FKS2 and with the S. pombe cps1�/bgs1�

and bgs2� genes. We named the gene bgs3� because it was thethird homologue at that time. Recently, a fourth homologuehas appeared in the database (SPCC1840.02C), and we refer toit as Bgs4p. Bgs3p is also related to GSs from other organisms,but it has no particularly close homologue among the knownproteins. Sequence identity values range between the highest(54 and 53% with S. pombe Bgs1p and Bgs4p, respectively) andthe lowest (41 and 42% with C. albicans Fks1p and Gsl1p,respectively). As with the other Fks proteins, Bgs3p predicts an210.8-kDa integral membrane protein with a cytoplasmic Nterminus and 16 transmembrane helices. Bgs3p does not con-tain the proposed UDP-glucose-binding motif QXXRW, al-though it does have a region with limited homology to BscAp,the catalytic subunit of cellulose synthase from Acetobacterxylinum spp. (45).

To determine whether bgs3� was the true ehs2� gene orwhether it was acting as an extragenic multicopy suppressor ofehs2-1, we constructed an integrative plasmid with a fragmentof pBG1 (SalI-ClaI) and integrated it in the genome of awild-type strain. We then crossed the integrant strain (Leu�)with an ehs2-1 mutant of opposite mating type and found a

high proportion of recombinants between the bgs3� locus(Leu�) and the hypersensitive phenotypes defined by theehs2-1 mutation (see Materials and Methods). The results in-dicated that bgs3� is not the structural gene that complementsthe ehs2-1 mutation. Also, the bgs3� gene failed to comple-ment the lytic phenotype at 37°C of the ehs2-1 mutants, insupport of the notion that they are different genes.

The bgs3� gene is essential for germination and vegetativegrowth. To investigate the phenotype resulting from completedeletion of the bgs3� gene, we constructed a diploid strain ofthe genotype bgs3::ura4�/bgs3� in which a copy of bgs3� wasdeleted and replaced by the ura4� gene (described in Materialsand Methods). Tetrad analysis showed two viable and twounviable spores (Fig. 2A), and all of the viable spores producedura-null colonies. Microscopic observation of the unviablespores showed that most of them germinated to form roundedcells that undergo 0 to 3 cell divisions before growth stops (Fig.2B). Similar results were obtained when spores bearing thebgs3::ura4� mutant allele were allowed to germinate in a me-dium containing 1.2 M sorbitol. Therefore, bgs3� is essentialfor cell viability and is also required for germination.

To further characterize the terminal phenotype of the bgs3-null mutants, bgs3::ura4� spores were germinated in rich liquidmedium for 12 h and stained with Cf and DAPI to visualize cellwalls and nuclei. The bgs3::ura4� germinating spores wererefringent, lemon shaped, and arrested with 1 or 2 nuclei. Incells with one nucleus, only one of the poles was stained withCf, whereas in spores bearing two nuclei a structure similar toa septum was detected in the middle. However, no spores wereseen with Cf staining on both poles (Fig. 2B). It seems that

FIG. 2. (A) Bgs3p is essential for cell viability. Tetrads from abgs3::ura4�/bgs3� strain were dissected on YES medium and incu-bated at 28°C for 4 days. (B) Terminal phenotype of the bgs3-nullmutants. Spores prepared from the bgs3::ura4�/bgs3� strain were in-oculated in YES medium for 12 to 16 h, fixed, and stained with DAPIand Cf to visualize nuclei and the cell wall, respectively. The arrowsindicate Cf-stainable material.

162 MARTIN ET AL. EUKARYOT. CELL

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

bgs3-null spores fail to elongate and resume growth as wild-type spores do. They are able to undergo mitosis and assemblestructures similar to a septum, but they are incapable of main-taining longitudinal growth and polarity.

It remained possible that the phenotype associated with thenull mutant might be a peculiarity associated with spore ger-mination. To investigate Bgs3p function during vegetativegrowth, we constructed a bgs3� gene under control of thethiamine-regulatable and reduced-expression-rate nmt1 pro-moter P41 nmt (13). This construct was integrated into thegenome of a wild-type diploid strain (MS75), the endogenousbgs3� promoter being replaced by the P41 nmt promoter (seeMaterials and Methods). We selected for haploids (nmt-bgs3�)in MM without uracil and thiamine. When bgs3� was ex-pressed in the absence of thiamine, this strain was perfectlyviable, showing a generation time of 3 h at 28°C, similar to thestrain used as a control. Cells displayed a normal cell morphol-ogy (Fig. 3A). In contrast, when thiamine was added to repressbgs3� expression, we observed changes in cell morphology. At13 h after thiamine addition, the percentage of rounded cellsincreased (Table 2); these cells were also shorter and wider.Whereas 100% of the cells grown in MM without thiamine wasdistributed in the 7- to 14-�m length range, the rounded cellswere distributed between 6 and 10 �m. Later on in the timecourse, at 24 h after thiamine addition, the growth rate sloweddown and finally stopped, and the cells became rounded andrefringent, with most of them showing a monopolar pattern ofgrowth (Fig. 3A). These experiments indicated that Bgs3p isnecessary to maintain cell growth and morphology.

Bgs3p depletion causes an unbalanced cell wall composi-tion. To visualize cell wall material, mutant cells depleted forBgs3p (grown in MM plus thiamine for 24 h) were stained withCf (Fig. 3A). The cells exhibited fluorescence throughout thecell wall, but the signal was more intense at one of the poles,with some cells having a cap of Cf-stainable material. The samecells grown in MM without thiamine showed fluorescencemainly at the septum (Fig. 3A). We used electron microscopyto examine cells of strain MS301 that had been in MM underrepressed or derepressed conditions for 24 h. At least 20 sec-tions of cells grown under each condition were examined. Thecell wall ultrastructure of cells grown in the absence of thia-mine displayed a thinner-electron transparent layer sur-rounded by an electron-dense layer similar to that describedpreviously in wild-type cells (Fig. 3B) (16). In contrast, manycells grown in the presence of thiamine (bgs3� off) showed verythick walls; although the three-layer structure was maintained,the electron transparent layer containing mainly �- and �-glu-cans was strongly thickened (Fig. 3B). Some cells showed a cellwall enlargement at the poles. To identify this material moreprecisely, cell wall constituents were isolated and characterizedafter cells were grown in the presence of radioactive glucoseduring the repression of the nmt promoter as described above(Fig. 3C). For each time point, the cells were fed with 14C-labeled glucose 12 h before cell wall analysis (see Materials andMethods). As shown in Fig. 3C, incorporation of radioactiveglucose into the cell wall of nmt-bgs3� in the absence of thia-mine (time zero) was similar to that of wild type but wasconsiderably higher than that of the wild type after 36 h in thepresence of thiamine (from 36 to 48% of total glucose incor-porated; Fig. 3C). A slight decrease in the amount of �-glucan

was detected 24 h after thiamine addition (77%), whereas theamount of �-glucan was �2-fold higher compared to wild-typecells. This effect was even more remarkable after 36 h, whenthe amount of �-glucan changed slightly but the amount of�-glucans was threefold higher. Additionally, the amount ofgalactomannan showed a modest decrease, leading to a drasticalteration in the �/� glucan ratio. These changes were solelydue to the bgs3� shutoff because thiamine addition per se didnot cause major alterations in the cell wall composition ofwild-type S. pombe (Fig. 3C). These findings suggest that thecarbohydrate synthesis of bgs3� shutoff mutant cell walls(rounded cells) was reduced in �-glucan but enhanced in�-glucan and that �-glucan was the main component of thiswall. To confirm the above results, we performed a cell wallresistance test by using the Novozyme enzymatic complex, reg-ularly used to obtain S. pombe protoplasts. In this experiment,Novozyme 234 was used as a source of �-glucanases. Cells ofstrain MS301 (nmt-bgs3�) grown for 24 h in MM with orwithout thiamine as described above were incubated in thepresence of Novozyme (200 �g/ml) at 28°C (described in Ma-terials and Methods). When bgs3� was repressed (with thia-mine), the cells were considerably more resistant to the de-grading enzymatic complex than in the absence of thiamine(bgs3� on) (Fig. 3D). These results indicate that the bgs3�

mutation alters the pattern of cell wall synthesis, rendering thecells more resistant to degradation, probably as a response ofa survival compensatory mechanism.

We next measured the specific activity of 1,3-�-GS in vitrobefore and after thiamine addition and found that it was notappreciably decreased (data not shown). In this scenario, al-though �-glucan polymer is diminished, other GSs would con-tinue to produce cell wall glucan, masking the lack of functionof bgs3�. It could also be speculated that the activity measuredin vitro (radioactivity incorporation into a linear 1,3-�-glucanpolymer) would not correspond to the activity of the enzyme invivo since this enzyme could be involved in the synthesis ofbranched 1,3-�-glucan.

Overexpression of bgs3� is lethal. To examine the pheno-types arising from the overexpression of bgs3�, the nmt1 pro-moter was integrated in the genome in front of the bgs3� geneinitiator methionine (ATG) (see Materials and Methods). Itwas found that overexpression of bgs3� was toxic to the cellsand that growth was not restored when an osmotic stabilizer(1.2 M sorbitol) was added to the medium (data not shown).We examined cell morphology in nmt-bgs3� cells 20 h afterinduction (promoter on). Some cells had an asymmetric shapein which one end of the cell had swollen abnormally to producetadpole- or round-bottom flask-shaped cells (data not shown).To examine the deposition of cell wall material, these cellswere stained with Cf. Cell wall stainable material was observedin engrossed septa and in cells that had lost their integrity.After 30 h of derepression, most of the cells showed a terminalphenotype, becoming spherical and with multiple septa (datanot shown).

We also measured the level of cell wall components and1,3-�-GS activity in cells in which the expression of Bgs3p wasupregulated (nmt1-bgs3� on, after 22 h in the absence of thi-amine). Neither the level of �-glucan nor GS activity wassignificantly altered (data not shown). To rule out the possi-bility that limiting conditions of other factors involved in glu-

VOL. 2, 2003 GLUCAN SYNTHASE IS ESSENTIAL IN S. POMBE 163

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

can synthesis might be unable to properly activate the bgs3�

gene in vitro, we overexpressed the regulatory subunit of GS inS. pombe, encoded by the rho1� gene (2), together with bgs3�.Overexpression of rho1� (under control of the nmt1 promoter)increased wild-type GS activity �2-fold, although no furtherincrease in activity was observed after the overexpression ofboth genes (data not shown).

Expression of Bgs3p under different growth conditions. Atthis point we sought to determine how the gene was regulated

FIG. 3. Lethal phenotype of the P41 nmt-bgs3� shutoff mutants. Cells grown at 28°C in MM were supplemented with thiamine to repress thenmt promoter. (A) Cell morphology of P41 nmt-bgs3� grown without thiamine (�T) or with thiamine (�T) for 24 h. Nomarsky (upper panel) andfluorescence micrographs of Cf-stained cells (lower panel) are shown. (B) Electron microscope sections of P41 nmt-bgs3� cells grown withoutthiamine (�T, upper left panel) or with thiamine (�T, upper right and lower panels) as described for panel A. Bars (0.6 and 0.4 �m).Magnification, 12,000 (upper left and lower panels) and 20,000 (upper right panel). (C) Composition of the cell wall in the P41 nmt-bgs3�

shutoff mutants. The relative levels of [14C]glucose radioactivity incorporated into each cell wall polysaccharide (white, �-glucan; gray, �-glucan;black, galactomannan) are shown for the wild type (PN22) with thiamine or the nmt-bgs3� mutant (MS301) at different times after thiamineaddition. Values are the means of three independent experiments with duplicate samples. Standard deviations for the total carbohydrate valuesare shown. (D) Susceptibility to Novozyme (200 �g/ml), a cell wall-degrading enzymatic complex, of P41 nmt-bgs3� shutoff mutant cells grown inthe presence of thiamine (�T) or in the absence of thiamine (�T) for 24 h. The OD600 was monitored at the indicated times, with the absorbanceof a control untreated sample of the strain grown previously (�T or �T) at each time point taken as 100%.

TABLE 2. Percentage of rounded cells of strain nmt-bgs3� grownin MM with or without thiamine

Strain (treatment)% Rounded cellsa at:

13 h 15 h 17 h 19 h 21 h 23 h

nmt-bgs3� (�thiamine) 6.72 11.7 8.93 14.56 31.36 48.23nmt-bgs3� (�thiamine) 0.36 0.1 0.3 0.2 0.17 0.2

a Two hundred cells were scored in each case.

164 MARTIN ET AL. EUKARYOT. CELL

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

during changes in the growth conditions of the cells and atdifferent stages of the yeast life cycle. Previous studies address-ing other GS isoforms had suggested that Bgs2p would beabsent during vegetative growth and would be strongly inducedunder sporulation conditions (29) and that Bgs1p would bepresent in vegetative growth and that its levels would notfluctuate appreciably in the cell cycle (25). Nothing is knownabout Bgs4p levels of expression. To achieve this goal, Westernblot experiments were carried out on protein extracts obtainedfrom the MS38 wild-type strain bearing an epitope-tagged ver-sion of Bgs3p (pBG11; see Materials and Methods) and grownunder different conditions. The anti-HA monoclonal antibodyspecifically recognized a single polypeptide with a molecularmass of ca. 200 kDa (Fig. 4A) in cells growing exponentially inglucose-rich medium (YES). Neither a parallel treatment ofthe cells with 20 mM CaCl2 nor a heat shock treatment for 4 hat 39°C led to any significant variation in the amount of Bgs3p.Bgs3p was also present in extracts from cells grown to satura-tion (stationary-phase cells) and in cells growing in low-glucosemedium (0.1% glucose, 3% glycerol). Similar results were ob-

tained with extracts from MM or from MM without nitrogenfor 10 h (Fig. 4A). In summary, the levels of Bgs3p wereessentially unaffected by the presence of different media orgrowth conditions.

We next analyzed protein samples from cells from strainHVP365 (h90) carrying Bgs3HAp incubated under sporulationconditions for 0, 3, 6, 9, and 24 h. After Western blot analysis,Bgs3p was found to be expressed in mating at about the samelevel as in the vegetative cycle; it was almost undetectableduring sporulation and appeared again in germinating cells(the sample was taken after 6 h of germination in rich medium)(Fig. 4B).

We therefore considered the possibility that the level ofBgs3p might be regulated in a cell cycle-dependent manner. Toaddress this issue, we prepared a synchronous population of S.pombe cells by arrest and release of the cdc25-22 mutant bear-ing the epitope-tagged version of Bgs3p. We found that thelevels of Bgs3p did not vary appreciably along the course of thisexperiment (data not shown).

Bgs3-GFPp fusion localizes to growing areas of the cell wallunder different developmental conditions. With a view to de-termining the intracellular distribution of Bgs3p in vivo, weadded the GFP sequence to the N terminus of the bgs3� gene(see Materials and Methods). Cells of strain MS302, with theGFP-tagged version of the bgs3� gene integrated under its ownpromoter, were fully viable and grew with a generation timesimilar to that of wild-type cells. The cells were grown in YESmedium, and fluorescence was monitored by microscopy. Asshown in Fig. 5A, Bgs3p mainly localized to two regions: celltips and the medial region. In a newly born cell, just aftercytokinesis, Bgs3p still localized to the new end at which theseptum had previously been formed (Fig. 5A, arrow 2). It thenmoved to the old end. Upon NETO (new-end take off [37]),Bgs3p was seen at the new end and thereafter persisted at bothends (Fig. 5A, arrowhead 4). At mitosis, cells were simulta-neously visualized for GFP and Cf. Bgs3p localized to theregion of the cell that overlapped with Cf staining (Fig. 5B). Inseptating cells Bgs3p staining seemed to be double in width.This observation suggested staining of the plasma membrane,in agreement with the predicted behavior of a membrane pro-tein like the other members of the S. pombe GS family (25).

To further characterize the cell cycle-regulated presence ofBgs3p at the growing tips, GFP-Bgs3p was expressed in the cellcycle thermosensitive mutants cdc10-129 and cdc25-22. Whencultured at 36°C, cdc10-129 cells were arrested in G1, and cellgrowth was polarized to one of the tips, whereas cdc25-22 cellswere arrested at the end of G2, with both ends of the cellactively growing (38). pMS60 was transformed into these cells,after which the cultures were incubated for 3 to 4 h at therestrictive temperature and visualized. GFP-Bgs3p accumu-lated at the only growing tip in cdc10-129 cells (Fig. 5Cb),whereas in cdc25-22 mutant cells it appeared at both tips (Fig.5Cc). In both cases, staining was also seen in patches inside thecells. This could be a consequence of the ectopic expression ofGFP-Bgs3p from a plasmid.

We also examined the localization of GFP-Bgs3p at therestrictive temperature in three polarity mutants transformedwith pMS60. tea1-50 mutant cells mislocalize one of the twogrowing poles and become branched (30); orb6-25 mutantsbecome spherical (47), and cps8-188 mutant cells are depolar-

FIG. 4. Expression levels of Bgs3p under different growth condi-tions. Total cell extracts (50 �g) were prepared from wild-type cellscontaining the bgs3�-HA on a plasmid (pBG11), run on an SDS–7.5%PAGE gel, and immunoblotted with anti-HA antibody. The band cor-responds to a polypeptide of ca. 200 kDa (Bgs3p) and was absent in acontrol sample without pBG11. (A) Total proteins were obtained fromcultures grown at 28°C to mid-log phase in YES under the indicatedconditions. For the “minus (�) nitrogen” lane, proteins were obtainedfrom cells grown at 28°C in MM without nitrogen for 12 h. For low-glucose medium, glucose (3%) was substituted by glycerol (3%) in theYES medium, leaving 0.1% glucose. (B) Protein levels of Bgs3p inwild-type (h90) cells carrying the Bgs3HAp incubated under sporula-tion conditions for 0, 3, 6, 9, and 24 h. The germination sample wasobtained from a 3-day sporulation culture of the same strain incubatedin YES rich medium for 6 h.

VOL. 2, 2003 GLUCAN SYNTHASE IS ESSENTIAL IN S. POMBE 165

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

ized due to changes in the actin structure and distribution atthe restrictive temperature (36°C) (19). In orb6 and tea1 mu-tants, after 5 h at 36°C GFP-Bgs3p was still localized to the newgrowing tips (Fig. 5Ce and f). In cps8 mutants, Bgs3p localiza-tion at the growing tips was much broader and often distrib-uted across the whole of the cell surface (Fig. 5Cd). Theseresults suggested that correct actin localization is essential forthe correct localization of Bgs3p. However, cells treated withLatrunculin A (a poison that depolymerizes actin) for 10 minor 1.5 h show a Bgs3p localization similar to that seen in cellswithout treatment (data not shown). This suggests that eventhough Bgs3p requires polarized secretion to be delivered,actin is not required to maintain Bgs3p localization once it hasbecome attached to the membrane.

The fact that Bgs3p is always localized to the growing tipsand septum prompted us to examine other situations in whichnew cell wall material is synthesized, such as the sexual differ-entiation process and germination. Upon nitrogen starvation,diffusible mating pheromones induce the polarized cell growthof cells of the opposite mating type toward one another (36).To examine Bgs3p localization during the conjugation process,a homothallic h90 strain (HV365) was transformed with plas-mid pMS60 and starved for nitrogen. GFP-Bgs3p staining be-came concentrated at the projection tip during pheromone-induced polarization, and the initial stages of cell fusion (Fig.6A). GFP-Bgs3p appeared to be localized in a double-ring-likestructure, probably associated with the cell membrane projec-tion. Immediately before karyogamy, cell wall degradation oc-curs, and the cell wall expands at the fusion point in order toform the zygote; apparently, at this point Bgs3p localizationappears broader. Immediately before both nuclei fused, GFP-

Bgs3p fluorescence was less intense but was still present onboth sides of the fusion neck (Fig. 6A). In S. pombe, diploidzygotes enter meiosis and sporulation immediately, a processrequiring the formation of a new cell wall around the devel-oping ascospores (49). Careful scrutiny of the sporulation pro-cess revealed that the fluorescence signal was completely ab-sent from the spore periphery and the membrane sacs in whichthe spore wall later forms (data not shown). Finally, asci weredigested with helicase to remove the ascus envelope, and as-cospores carrying the plasmid pMS60 were selected for leucineprototrophy and allowed to regenerate in liquid MM for 6 to10 h. As expected, GFP-Bgs3p was always found toward thegrowing tip, localized in a polar fashion (Fig. 6B).

Complementation of the bgs3� lethal phenotype by other GShomologues. To test for possible genetic complementation be-tween bgs3� and the other GS homologous genes, we decidedto express the GS homologous genes under control of theirown promoters in a bgs3� background. Plasmids pCP1 (cps1�/bgs1�), pMS5 (bgs2�-HA), pJG1 (bgs4�), and pAL-KS (emptyvector) were transformed in the diploid strain bgs3�/bgs3�

(MS300). Diploids bearing each plasmid were obtained, butafter sporulation we were not able to rescue any progeny of thebgs3::ura4� genotype (data not shown). This lack of comple-mentation could be a consequence of the genes being ex-pressed in specific situations because the experiments wereperformed in multicopy plasmids under control of their ownpromoters. To rule out this possibility, we decided to expressthe GS homologues in plasmids under control of the mediumlevel nmt promoter P41 nmt (13). pJG22 (pREP41X-cps1�/bgs1�), pMS24 (pREP41X-bgs2�-HA), pJG18 (pREP41X-bgs4�), and pREP41X (empty vector) were used for the

FIG. 5. Localization of Bgs3p in S. pombe cells. (A) Cells expressing functional GFP-tagged Bgs3p from a chromosomal integration werevisualized in vivo under the fluorescence microscope. Bgs3p was localized to the medial region during cytokinesis (arrowheads 1 and 2) and to thegrowing ends during monopolar and bipolar growth (arrowheads 3 and 4). (B) The localization of cell wall material (Cf staining) and GFP-Bgs3pis shown in cells at different stages of cytokinesis. The cells represent different stages of cell wall deposition, the top one being the least advancedand the bottom being the most advanced. (C) Fluorescence micrographs of different S. pombe strains transformed with plasmid pMS60(GFP-Bgs3p) and grown in MM at 28°C and then for 4 h at 36°C. (a) Wild type; (b) cdc10-129; (c) cdc25-22; (d) cps8-188; (e) orb6-25; (f) tea1-50.

166 MARTIN ET AL. EUKARYOT. CELL

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

complementation experiments. We also expressed the bgs2�

gene (sporulation-specific GS) under control of the bgs3� pro-moter, pMS62. None of the plasmids was able to rescue thehaploid bgs3� mutant. As a control, we used pBG16(pREP41X-bgs3�), a plasmid that was able to rescue bgs3�lethality.

Conversely, overexpression of bgs3� under control of thenmt promoter (pREP3X-bgs3�) or under its own promoter(pBG11) did not rescue the growth defect of a cps1-12 mutantstrain (20) nor the lytic phenotype of a cwg1-1 mutant at therestrictive temperature (37°C), i.e., a point mutation in thebgs4� gene (42; J. C. Ribas, unpublished data). These resultssuggested that ectopic expression of cps1�, bgs2�, and bgs4�

failed to support the lack of function of Bgs3p and that bgs3�

was not able to suppress the defect associated with mutationsin other GS homologues.

DISCUSSION

Cf interacts with 1,4-�-glucans, such as chitin and cellulose,and probably also with 1,3-�-glucans to form fluorescent com-

plexes. This interaction disorganizes polymer assembly (35).Echs are naturally occurring lipopeptide antibiotics that inhibit1,3-�-glucan synthesis (9). Therefore, by searching for mutantshypersensitive to both compounds, we attempted to raise thenumber of mutants affected in the assembly of �-glucans (5).bgs3� was cloned indeed by complementation of the Ech-hypersensitive phenotype of one such mutant, ehs2-1 (to bedescribed elsewhere), and also complemented hypersensitivityto Cf. We soon found that bgs3� was not the wild-type allele ofthe ehs2-1 mutant. This is not a rare event in S. pombe cloningsince there are no centromeric plasmids available and thelibraries are based on multicopy plasmids. In certain specificsituations a higher number of copies of the gene may be toxicto cells, although currently we do not have any data to confirmwhether this is indeed the case.

S. pombe has four GS-related genes, bgs1� to bgs4�, and todate the two best characterized display essential functions indifferent situations. Thus, bgs1� is essential for septum assem-bly (25, 26), although it may be required for cell wall synthesisin other events (6) and bgs2� is required for cell wall sporematuration (24, 29). It is of interest to know whether the othertwo genes bgs3� and bgs4� play independent essential func-tions or whether they share overlapping nonessential individ-ual functions, as is the case for FKS1 and FKS2 in S. cerevisiae.

Several lines of evidence indicate that bgs3� is an essentialgene that could be involved in cell wall assembly during elon-gation. First, Bgs3p is essential for spore viability and lethalitywas not rescued in the presence of 1.2 M sorbitol. bgs3::ura4�

germinating spores were refringent, lemon shaped, and ar-rested with one or two nuclei but failed to elongate. In addi-tion, all our attempts to obtain a bgs3� haploid strain, either bydirect transformation of haploid wild-type cells or by maintain-ing bgs3� haploids with episomal copies of bgs3� and laterselecting for bgs3� haploids without the bgs3� plasmid, wereunsuccessful. Second, when the bgs3� promoter was substi-tuted by the regulatable and reduced expression promoter P41nmt in the chromosome, cells grown in the presence of thia-mine (promoter off) were shorter and wider and finally died.The terminal phenotype of these cells was similar to the onepreviously seen in the bgs3� spores. Thus, Bgs3p appears toplay a role in the survival of spores and vegetative cells and isprobably important for the maintenance of cell shape and theestablishment of cell polarity.

Cellular localization of Bgs3p. Bgs3p was expressed duringvegetative growth, and its levels did not vary significantlythroughout the cell cycle. It was expressed with no significantvariations, in different media and different growth conditions,and its levels were not enhanced by stress situations. Thus,protein levels do not appear to determine Bgs3p function dur-ing the mitotic cycle. Instead, other factors, such as its intra-cellular localization and/or activation of the enzymatic com-plex, might be important for regulating the function of Bgs3p.However, we did detect some changes during sexual develop-ment, Bgs3p was present in the early steps (3 and 6 h) ofmating, after which the protein disappeared in sporulation (12and 24 h in SPA) and reappeared during germination.

Bgs3p localization changed at different stages of the cellcycle. Bgs3p localizes to the growing poles during interphaseand to the division septum during cytokinesis. It was also as-sociated with the cell tip projection “shmoo,” with the fusion

FIG. 6. GFP-Bgs3p localization during pheromone-induced conju-gation steps (A) and in germination (B). The homothallic h90 strain(HV365) was transformed with plasmid pMS60 and starved for nitro-gen. (A) A series of ordered paired panels, each pair showing GFP-Bgs3p fluorescence (top) and Cf images (bottom) of the same cells, areseen. Bgs3p was concentrated at the projection tip during pheromoneinduced polarization and early fusion (columns a and b). Next, cell wallremodeling takes place during this stage, and Bgs3p localization be-came broader and more diffuse (columns c and d); finally, stainingalmost disappears (column e). (B) shows GFP-Bgs3p fluorescence inhaploid spores bearing plasmid pMS60 after 4 h of germination. Bgs3plocalization is predominantly polarized to the growing tip.

VOL. 2, 2003 GLUCAN SYNTHASE IS ESSENTIAL IN S. POMBE 167

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

neck at early steps of zygotes formation and with the growingpole early in germination. All of these regions correspond tothe sites at which cell surface structures are growing or beingactively remodeled.

The predicted bgs3� product is a membrane-bound enzyme,and this could be related to the kind of staining pattern that weobserved with the fusion protein GFP-Bgs3p at the growingtips, especially during cytokinesis. Previous and ongoing workhas suggested that proteins such as Mok1p, Rho1p, Rho2p,Pck2p, and Pck1p, which are involved in �- and �-glucan syn-thesis, are localized to the growing sites and are intimatelyassociated with the actin cytoskeleton (1, 21, 44). Bgs3p local-izes similar to the actin patches during the cell cycle (28), andthe results reported here with the cdc10-109 and cdc25-22mutants suggest that the position of Bgs3p is dependent onpolarized growth. In polarity mutants such as tea1-50 and orb6-25, Bgs3p localizes to the new growing surface, whereas in anactin point mutant cps8-188, Bgs3p is not properly localized,displaying fluorescence all over the cell membranes. This ob-servations suggest that Bgs3p localization depends on the F-actin cytoskeleton.

The location of GFP-Bgs3p during the early stages of pri-mary septum formation differs from the Bgs1p one. In the caseof Bgs1p, colocalization experiments clearly indicate that thecell wall material of the primary septum, stained with Cf, lagsbehind the GFP-Bgs1p signal (6). In contrast, with Bgs3p theGFP signal at the septum level is always behind the corre-sponding signal of the Cf-stained material. This observationfavors the hypothesis that Bgs3p may be involved in the syn-thesis or assembly of cell wall material corresponding to thesecondary septum. The simultaneous observation of Bgs1p andBgs3p labeled with different GFP tags will be required toclarify this point.

Downregulation and upregulation of Bgs3p causes severecellular defects. In the long term, Bgs3p depletion causes celldeath concomitantly with several interesting phenotypes. (i)Cells are shorter and wider, showing a terminal refringentphenotype, although they do not lyse. (ii) Cells have thickerwalls, as shown by a higher incorporation of glucose in vivo,and by electron microscopy. (iii) Finally, the cell wall compo-sition of Bgs3p-depleted cells is slightly reduced in �-glucancontent but very enhanced in �-glucan and, as a consequence,is more resistant to degradation by Novozyme than is the wildtype.

The modest reduction in the �-glucan content could accountfor the contribution of Bgs3p to cell wall glucan synthesis,although here we failed to distinguish which type of glucan wasreduced. It has recently been shown that mutations in Fks1paffect the cell wall contents of both 1,3-�- and 1,6-�-glucan inS. cerevisiae (8). Therefore, as discussed above, our bgs3 shutoffmutant could be altered in the 1,3-�-glucan polymer, whichmight serve as a 1,6-�-glucan or a 1,3-�-glucan acceptor. Thehigh levels of �-glucan seen in the bgs3 shutoff mutant could bea consequence of improper or deficient attachment of thispolymer, which in this situation becomes upregulated andforms a highly thickened cell wall, or could be part of a generalcompensatory response that occurs to retain at least partial cellwall integrity. Compensatory mechanisms have been widelydescribed in S. cerevisiae cell wall mutants (39), but have notbeen addressed in S. pombe. Alternatively, Bgs3p could be a

glucan transporter. The fact that the reaction mechanism of1,3-�-glucan synthesis has not been elucidated, the lack ofclear similarities with other polymerizing �-glycosyltrans-ferases (7), and a predicted membrane topology that is remi-niscent of bacterial and eukaryotic transport proteins (11) to-gether suggest the possibility that Bgs3p (as well as othermembers of the GS family) may be a GS component involvedin the transport of 1,3-�- and/or 1,6-�-glucan across the plasmamembrane.

Overexpression of bgs3� did not elicit any general increasein cell wall biosynthesis and was not correlated with any in-crease in the �-glucan content or in the GS activity. This resultwas not surprising and indicates that the enzyme probablyfunctions in a multiprotein complex and that a higher amountof one of the subunits is not sufficient to generate a higheramount of the enzyme product. A similar situation has beendescribed for the chitin synthase complex in S. cerevisiae, inwhich overproduction of CHS3 (the catalytic subunit of thecomplex which makes most of the chitin found in the cell)is not translated into an increase in the amount of chitin. Inthis sense, transporters, activators, and the necessary proteinsfor the localization of the catalytic subunit must be overpro-duced simultaneously for a rise in the amount of chitin to occur(46).

From a common structure to different functions in vivo.Addressing the biological relevance of the different GS cata-lytic subunit homologues in fission yeast is difficult due tounknown substrate specificities and our poor knowledge of thearchitecture and the linkages between different polymers.Cps1p/Bgs1p does play an essential role in the assembly of theprimary septum (23, 25, 26), although recent localization datacould open a new role for Bgs1p in cell wall elongation (6).Bgs2p is required for ascospore wall maturation (24, 29) andbgs4� is the structural gene of the cwg1-1 mutant isolated as aGS-defective conditional mutant and is essential for viability(42; Ribas, unpublished).

Bgs3p is important for cell wall extension during polargrowth and for the assembly of the secondary septum, althoughits depletion does not promote cell lysis. Therefore, Bgs3p maybe responsible for some transglycosylation reaction involved inthe biosynthesis of the 1,6-�-branched 1,3-�-glucan, the maincell wall component distributed throughout the less-denselayer of the cell wall, including the secondary septum.

The results from complementation studies suggest that ec-topic expression of cps1�/bgs1�, bgs2�, and bgs4� failed torescue the viability of bgs3� cells and that the overexpressionof bgs3� was not able to suppress the defect associated withmutations in other GS homologues. All of the observationsdescribed above, together with the essential role of each bgsgene, suggest that different GS activities perform essentialnonoverlapping functions in the synthesis or transport of dif-ferent �-glucan polymers. Further studies on the biochemicalspecificity of each isoform will be required to unravel theprecise function of each GS activity or gene.

ACKNOWLEDGMENTS

We thank H. Valdivieso and C. Roncero for critical reading of themanuscript. We also thank C. Belinchon for help with microscopy andJ. C. Ribas and P. Perez for plasmids, strains, and helpful discussions.

168 MARTIN ET AL. EUKARYOT. CELL

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from

V.M. and E.C. acknowledge support from a fellowship granted bythe MEC (Spain). This work was supported by grants BIO98-0814,BIO2001-1663 from the Comision Interministerial de Ciencia y Tec-nología (Spain), and CS17/01 from the Junta de Castilla y Leon and bya contract with Lilly SA (Spain).

REFERENCES

1. Arellano, M., P. M. Coll, and P. Perez. 1999. Rho GTPases in the control ofcell morphology, cell polarity, and actin localization in fission yeast. Microsc.Res. Technol. 47:51–60.

2. Arellano, M., A. Duran, and P. Perez. 1996. Rho1 GTPase activates the(1–3)�-D-glucan synthase and is involved in Schizosaccharomyces pombemorphogenesis. EMBO J. 15:4584–4591.

3. Bahler, J., J.-Q. Wu, M. S. Longtine, N. G. Shah, A. McKenzie III, A. B.Steever, A. Wach, P. Philippsen, and J. R. Pringle. 1998. Heterologousmodules for efficient and versatile PCR-based gene targeting in Schizosac-charomyces pombe. Yeast 14:943–951.

4. Cabib, E., J. Drgonova, and T. Drgon. 1998. Role of small G proteins in yeastcell polarization and wall biosynthesis. Annu. Rev. Biochem. 67:307–333.

5. Carnero, E., J. C. Ribas, B. García, A. Duran, and Y. Sanchez. 2000. Schizo-saccharomyces pombe ehs1p is involved in maintaining cell wall integrity andin calcium uptake. Mol. Gen. Genet. 264:173–183.

6. Cortes, J. C. G., J. Ishiguro, A., Duran, and J. C. Ribas. 2002. Localizationof the (1,3)�-D-glucan synthase catalytic subunit homologue Bgs1p/Cps1pfrom fission yeast suggests it is involved in septation, polarized growth,mating, spore wall formation and spore germination. J. Cell Sci. 115:4081–4096.

7. Dhugga, K. S. 2001. Building the wall: genes and enzyme complexes forpolysaccharide synthases. Curr. Opin. Plant Biol. 4:488–493.

8. Dijkgraaf, G. J., M. Abe, Y. A. Ohya, and H. Bussey. 2002. Mutations inFks1p affect the cell wall content of �-1,3- and �-1,6-glucan in Saccharomycescerevisiae. Yeast 19:671–690.

9. Douglas, C., J. A. Marrinan, W. Li, and M. B. Kurtz. 1994. A Saccharomycescerevisiae mutant with echinocandin-resistant 1,3-�-D-glucan synthase J. Bac-teriol. 176:5686–5696.

10. Douglas, C. M. 2001. Fungal �(1,3)-D-glucan synthesis. Med. Mycol. 39:55–66.11. Douglas, C. M., F. Foor, J. A. Marrinan, N. Morin, J. B. Nielsen, A. M. Dahl,

P. Mazur, W. Baginsky, W. Li, M. El-Sherbeini, J. A. Clemas, S. M. Man-dala, B. R. Frommer, and M. B. Kurtz. 1994. The Saccharomyces cerevisiaeFKS1 (ETG1) gene encodes an integral membrane protein which is a subunitof 1,3-�-D-glucan synthase. Proc. Natl. Acad. Sci. USA 91:12907–12911.

12. Drgonova, J., T. Drgon, K. Tanaka, R. Kollar, G.-C. Chen, R. A. Ford,C. S. M. Chan, Y. Takai, and E. Cabib. 1996. Rho1p, a yeast protein at theinterface between cell polarization and morphogenesis. Science 272:277–279.

13. Forsburg, S. L. 1993. Comparison of Schizosaccharomyces pombe expressionsystems. Nucleic Acids Res. 21:2955–2956.

14. Georgopapadakou, N. H. 2001. Update on antifungals targeted to the cellwall: focus on �-1,3-glucan synthase inhibitors. Exp. Opin. Investig. Drugs10:269–280.

15. Hirano, T., Y. Hiraoka, and M. Yanagida. 1988. A temperature sensitivemutation of the Schizosaccharomyces pombe gene nuc2� that encodes anuclear scaffold-like protein blocks spindle elongation in mitotic anaphase.J. Cell Biol. 106:1171–1183.

16. Humbel, B. M., M. Konomi, T. Takagi, N. Kamasawa, S. A. Ishijima, and M.Osumi. 2001. In situ localization of �-glucans in the cell wall of Schizosac-charomyces pombe. Yeast 18:433–444.

17. Inoue, S. B., N. Takewaki, T. Takasuka, T. Mio, M. Adachi, Y. Fujii, C.Miyamoto, M. Arisawa, Y. Furuichi, and T. Watanabe. 1995. Characteriza-tion and gene cloning of 1,3-�-D-glucan synthase from Saccharomyces cer-evisiae. Eur. J. Biochem. 231:845–854.

18. Ishiguro, J. 1998. Genetic control of fission yeast cell wall synthesis: thegenes involved in wall biogenesis and their interactions in Schizosaccharo-myces pombe. Genes Genet. Syst. 73:181–191.

19. Ishiguro, J., and W. Kobayashi. 1996. An actin point-mutation neighboringthe “hydrophobic plug” causes defects in the maintenance of cell polarityand septum organization in the fission yeast Schizosaccharomyces pombe.FEBS Lett. 392:237–241.

20. Ishiguro, J., A. Saitou, A. Duran, and J. C. Ribas. 1997. cps1�, a Schizosac-charomyces pombe gene homolog of Saccharomyces cerevisiae FKS geneswhose mutation confers hypersensitivity to cyclosporin A and papulacandinB. J. Bacteriol. 179:7653–7662.

21. Katayama, S., D. Hirata, M. Arellano, P. Perez, and T. Toda. 1999. Fissionyeast �-glucan synthase Mok1 requires the actin cytoskeleton to localize thesites of growth and plays an essential role in cell morphogenesis downstreamof protein kinase C function. J. Cell Biol. 144:1173–1186.

22. Klis, F. M. 1994. Cell wall assembly in yeast. Yeast 10:851–869.23. Le Goff, X., A. Woollard, and V. Simanis. 1999. Analysis of the cps1 gene

provides evidence for a septation checkpoint in Schizosaccharomyces pombe.Mol. Gen. Genet. 262:163–172.

24. Liu, J., X. Tang, H. Wang, and M. Balasubramanian. 2000. Bgs2p, a 1,3-�-glucan synthase subunit, is essential for maturation of ascospore wall inSchizosaccharomyces pombe. FEBS Lett. 478:105–108.

25. Liu, J., X. Tang, H. Wang, S. Oliferenco, and M. K. Balasubramanian. 2002.The localization of the integral membrane protein Cps1p to the cell divisionsite is dependent on the actomyosin ring and the septation-inducing networkin Schizosaccharomyces pombe. Mol. Biol. Cell 13:989–1000.

26. Liu, J., H. Wang, D. MacCollum, and M. K. Balasubramanian. 1999. Drc1p/Cps1p, a 1,3-�-glucan synthase subunit, is essential for division septumassembly in Schizosaccharomyces pombe. Genetics 153:1193–1203.

27. Manners, D. J., and M. T. Meyer. 1977. The molecular structures of someglucans from the cell wall of S. pombe. Carbohydr. Res. 57:189–203.

28. Marks, J., and J. S. Hyams. 1985. Localization of F-actin through the celldivision cycle of Schizosaccharomyces pombe. Eur. J. Cell Biol. 39:27–32.

29. Martin, V., J. C. Ribas, E. Carnero, A. Duran, and Y. Sanchez. 2000. bgs2�,a sporulation-specific glucan synthase homologue is required for properascospore wall maturation in fission yeast. Mol. Microbiol. 38:308–321.

30. Mata, J., and P. Nurse. 1997. tea1 and the microtubular cytoskeleton areimportant for generating global spatial order within the fission yeast. Cell89:939–950.

31. Maundrell, K. 1993. Thiamine-repressible expression vectors pREP andpRIP for fission yeast. Gene 123:127–130.

32. Mazur, P., N. Morin, W. Baginsky, M. El-Sherbeini, J. A. Clemas, J. B.Nielsen, and F. Foor. 1995. Differential expression and function of twohomologous subunits of yeast 1,3-�-D-glucan synthase. Mol. Cell. Biol. 15:5671–5681.

33. Mol, P. C., H.-M. Park, J. T. Mullins, and E. Cabib. 1994. A GTP-bindingprotein regulates the activity of (1,3)-�-glucan synthase, an enzyme directlyinvolved in yeast cell wall morphogenesis. J. Biol. Chem. 269:31267–31274.

34. Moreno, S., A. Klar, and P. Nurse. 1991. Molecular genetic analysis of fissionyeast Schizosaccharomyces pombe. Methods Enzymol. 194:795–823.

35. Nicholas, R. O., D. W. Williams, and P. A. Hunter. 1994. Investigation of thevalue of �-glucan-specific fluorochromes for predicting the �-glucan contentof the cell wall of zoopathogenic fungi. Mycol. Res. 98:694–698.

36. Nielsen, O., and J. Davey. 1995. Pheromone communication in the fissionyeast Schizosaccharomyces pombe. Semin. Cell Biol. 6:95–104.

37. Nurse, P. 1994. Fission yeast morphogenesis. Mol. Biol. Cell 5:613–616.38. Nurse, P., P. Thuriaux, and K. Nasmyth. 1976. Genetic control of the cell

division cycle in the fission yeast Schizosaccharomyces pombe. Mol. Gen.Genet. 146:167–178.

39. Popolo, L., T. Gualtieri, and E. Ragni. 2001. The yeast cell-wall salvagepathway. Med. Mycol. 39:111–121.

40. Qadota, H., C. P. Python, S. B. Inoue, M. Arisawa, Y. Anraku, Y. Zheng, T.Watanabe, D. E. Levin, and Y. Ohya. 1996. Identification of yeast Rho1p GT-Pase as a regulatory subunit of 1,3-�-glucan synthase. Science 272:279–281.

41. Radding, J. A., S. A. Heidler, and W. W. Turner. 1998. Photoaffinity analogof the semisynthetic echinocandin LY303366: identification of echinocandintargets in Candida albicans. Antimicrob. Agents Chemother. 42:1187–1194.

42. Ribas, J. C., M. Díaz, A. Duran, and P. Perez. 1991. Isolation and charac-terization of Schizosaccharomyces pombe mutants defective in cell wall (1–3)�-D-glucan. J. Bacteriol. 173:3456–3462.

43. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, N.Y.

44. Sayers, L. G., S. Katayama, K. Nakano, H. Mellor, I. Mabuchi, T. Toda, andP. J. Parker. 2000. Rho-dependence of Schizosaccharomyces pombe Pck2.Genes Cells 5:17–27.

45. Thompson, J. R., C. M. Douglas, W. Li, C. K. Jue, B. Pramanik, X. Yuan,T. H. Rude, D. L. Toffaletti, J. R. Perfect, and M. Kurtz. 1999. A glucansynthase FKS1 homolog in Cryptococcus neoformans is single copy and en-codes an essential function. J. Bacteriol. 181:444–453.

46. Trilla, J. A., A. Duran, and C. Roncero. 1999. Chs7p, a new protein involvedin the control of protein export from the endoplasmic reticulum that isspecifically engaged in the regulation of chitin synthesis in Saccharomycescerevisiae. J. Cell Biol. 145:1153–1163.

47. Verde, F., D. J. Wiley, and P. Nurse. 1998. Fission yeast orb6, a Ser/Thrprotein kinase related to mammalian rho kinase and myotonic dystrophykinase, is required for maintenance of cell polarity and coordinates cellmorphogenesis with the cell cycle. Proc. Natl. Acad. Sci. USA 95:7526–7531.

48. Wilson, I. A., H. L. Niman, R. A. Houghten, M. L. Chereson, M. L. Connolly,and R. A. Lerner. 1984. The structure of an antigenic determinant in aprotein. Cell 37:767–778.

49. Yamamoto, M., Y. Imai, and Y. Watanabe. 1997. Mating and sporulation inSchizosaccharomyces pombe, p. 1037–1106. In J. R. Pringle, J. R. Broach, andE. W. Jones (ed.), The molecular and cellular biology of the yeast Saccharomy-ces, vol. 3. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.

50. Zhao, C., U., S. Jung, P. Garret-Engele, T. M. Roe, S. Cyert, and D. Levin.1998. Temperature-induced expression of yeast FKS2 is under the dualcontrol of protein kinase C and calcineurin. Mol. Cell. Biol. 18:1013–1022.

VOL. 2, 2003 GLUCAN SYNTHASE IS ESSENTIAL IN S. POMBE 169

on Decem

ber 21, 2015 by guesthttp://ec.asm

.org/D

ownloaded from