HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 276, Issue 29, 27731-27739, July 20, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/29/27731    most recent M101986200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Grimme, S. J. Articles by Orlean, P. Search for Related Content PubMed PubMed Citation Articles by Grimme, S. J. Articles by Orlean, P. Social Bookmarking                      What's this?

The Essential Smp3 Protein Is Required for Addition of the Side-branching Fourth Mannose during Assembly of Yeast Glycosylphosphatidylinositols^*

Stephen J. Grimme, Barbara A. Westfall, Jill M. Wiedman, Christopher H. Taron, and Peter Orlean

From the Department of Biochemistry, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801

Received for publication, March 5, 2001, and in revised form, May 15, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The major glycosylphosphatidylinositols (GPIs) transferred to protein in mammals and trypanosomes contain three mannoses. In Saccharomyces cerevisiae, however, the GPI transferred to protein bears a fourth, 1,2-linked Man on the 1,2-Man that receives the phosphoethanolamine (EthN-P) moiety through which GPIs become linked to protein. We report that temperature-sensitive smp3 mutants accumulate a GPI containing three mannoses and that smp3 is epistatic to the gpi11, gpi13, and gaa1 mutations, which normally result in the accumulation of Man₄-GPIs, including the presumed substrate for the yeast GPI transamidase. The Smp3 protein, which is encoded by an essential gene, is therefore required for addition of the fourth Man to yeast GPI precursors. The finding that smp3 prevents the formation of the Man₄-GPI that accumulates when addition of EthN-P to Man-3 is blocked in a gpi13 mutant suggests that the presence of the fourth Man is important for transfer of EthN-P to Man-3 of yeast GPIs. The Man₃-GPI that accumulates in smp3 is a mixture of two dominant isoforms, one bearing a single EthN-P side branch on Man-1, the other with EthN-P on Man-2, and these isoforms can be placed in separate arms of a branched GPI assembly pathway. Smp3-related proteins are encoded in the genomes of Schizosaccharomyces pombe, Candida albicans, Drosophila melanogaster, and Homo sapiens and form a subgroup of a family of proteins, the other groups of which are defined by the Pig-B(Gpi10) protein, which adds the third GPI mannose, and by the Alg9 and Alg12 proteins, which act in the dolichol pathway for N-glycosylation. Because Man₄-containing GPI precursors are normally formed in yeast and Plasmodium falciparum, whereas addition of a fourth Man during assembly of mammalian GPIs is rare and not required for GPI transfer to protein, Smp3p-dependent addition of a fourth Man represents a target for antifungal and antimalarial drugs.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Glycosylphosphatidylinositols (GPIs)¹ with the conserved core structure H₂N-CH₂-CH₂-PO₄-6Man1,2Man1,6Man1,4GlcN1,6Ins phospholipid are made by all eukaryotes. Many GPIs are transferred to the COOH terminus of secretory glycoproteins and serve as membrane anchors, but others remain protein-free (1-3). In Saccharomyces cerevisiae, mannoproteins can be transferred from their GPI anchor to cell wall -glucan (4, 5). The formation of GPIs is essential for growth of yeasts, Leishmania mexicana, and the bloodstream form of Trypanosoma brucei and for embryonic development in mammals (6-10).

GPIs are preassembled on an inositol phospholipid in a stepwise pathway, the enzymes of which are localized in membranes of the endoplasmic reticulum (2, 3, 11, 12). During its assembly, the glycan core of the GPI precursor can be decorated with side branches. In yeast and mammals, the first and second mannoses can be modified with phosphoethanolamine (EthN-P) (11-23), and late stage GPI precursors made in wild type S. cerevisiae and Plasmodium falciparum cells late-stage GPI precursors bear a fourth, 1,2-linked Man on the third, 1,2-linked Man of the glycan core (24-26). Yeast mutants deficient in EthN-P addition to Man-2 and Man-3 and in GPI transfer to protein accumulate GPIs with four mannoses (21, 22, 27, 28). In contrast, in mammalian cells, the largest glycan headgroups characterized contain only three mannoses (13-18), although it has been suggested that minor amounts of Man₄ species may be formed (13, 18, 29). Protein-bound GPIs in mammals usually contain three mannoses, but a fourth, 1,2-linked Man has been detected on the GPI anchor of the murine Thy-1 glycoprotein (30) and human renal membrane dipeptidase (31).

We report that the S. cerevisiae Smp3 protein, a member of a family of potential dolichyl phosphate (Dol-P) Man-utilizing mannosyltransferases, is a candidate for the enzyme that adds the fourth Man during yeast GPI assembly. Temperature-sensitive smp3 mutations prevent the formation of Man₄-GPIs and result in the accumulation of Man₃-GPIs. Addition of the fourth mannose is therefore required to generate the Man₄-GPIs that are the most likely substrates for the yeast GPI transamidase. Because the SMP3 gene is essential, and because addition of a fourth Man seems to be of much greater relative importance in GPI assembly in yeast than it is in mammalian cells, addition of the fourth Man represents a potential target for antifungal drugs.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- [2-³H]myo-Inositol (specific activity, 15-20 Ci/mmol) and [1-³H]ethanolamine (specific activity, 10-30 Ci/mmol) were obtained from American Radiolabeled Chemicals (St. Louis, MO). Tran³⁵S-label mixture (specific activity, 1 Ci/mmol) and phosphatidylinositol-specific phospholipase C were from ICN Biochemicals, and jack bean and Aspergillus satoi -mannosidases from Oxford GlycoSciences (Oxford, United Kingdom). Silica gel 60 HPTLC 5631/5 plates were supplied by Altech (Deerfield, IL).

Yeast Strains and Media-- The yeast strains used in this work are listed in Table I, and the construction of those strains made specifically for this study is detailed below. YPD and SD medium were prepared as described in Ref. 32, and YPGal medium had the same composition as YPD but contained 2% (w/v) galactose instead of glucose. SGlyYE, SGlcYE, and SGalYE were as described in Ref. 22, except that SGlcYE contained 2% (w/v) glucose.

The smp3 strain KH8 (33) was back-crossed twice to strain YMW2 (34) to generate the smp3-1 and smp3-1D strains. The smp3-2 and gaa1-2 strains were isolated in a screen for mutants that require the GPI1 gene for growth,² conducted using the vectors and strains described in Ref. 34. Both primary isolates were back-crossed twice to strain YMW1. The cross of smp3-1 to gpi1-8H yielded viable smp3-1/gpi1 haploids. An smp3/gpi11 strain was generated by crossing smp3-1 with strain gpi11-pPIG-F (22).

smp3/gpi13 "double mutants" were created by crossing haploid smp3-1 or smp3-2 strains with a haploid gpi13::KAN^R-pGAL-GPI13 strain (22) and allowing the resulting diploids to sporulate. Asci were dissected, and ascospores were allowed to germinate on YP-galactose medium supplemented with 0.25 M KCl. The genotypes of segregants from tetratype tetrads were determined from the segregation of selectable markers and temperature sensitivity among these strains and confirmed by the thin-layer chromatographic profiles of their [³H]inositol-labeled lipids. The smp3/gpi13-pGAL-GPI13 double mutants grew more slowly at 25 °C on solid YPGal medium than the single mutants unless osmotic support was provided.

smp3/gaa1 double mutants were generated in a genetic cross of the smp3-1 and gaa1-2 strains. The smp3/gaa1 segregants from tetratype tetrads were distinguishable from the smp3-1 and gaa1-2 single mutation because, unlike the single mutants, they failed to grow at 37 °C on solid YPD medium supplemented with 0.25 M KCl. The presence of both the smp3 and gaa1 mutations in these strains was confirmed by introducing the SMP3 gene on a plasmid; at nonpermissive temperature, the double mutants showed the profile of [³H]inositol-labeled lipids characteristic of gaa1, not smp3 (see Fig. 2A).

smp3-1 was crossed into an ethanolamine-auxotrophic psd1/psd2 strain background (RYY51) (35) to generate a strain capable of enhanced incorporation of exogenous ethanolamine into lipid. Media for growth of the smp3-1/psd1/psd2 strain were supplemented with 2 mM ethanolamine and 2 mM choline.

A heterozygous smp3::Kan^R/SMP3 disruptant, generated by the Saccharomyces Deletion Project, was obtained from Research Genetics (Huntsville, AL). Plasmid pSMP3-426, in which SMP3 is expressed behind its native promoter, was made by using polymerase chain reaction to amplify a genomic DNA fragment containing SMP3 and 401 nucleotides immediately upstream of the start codon and 336 nucleotides immediately downstream of the SMP3 coding region, after which the DNA fragment was cloned into the 2µ plasmid pRS426 (36). Plasmid pGAL-SMP3, in which expression of SMP3 is under the control of the glucose-repressible GAL10 promoter, was made by using polymerase chain reaction to amplify the SMP3 coding region from genomic DNA and cloning the resulting DNA fragment between the GAL10 promoter and the GAL7 termination region of plasmid pMW20 (34), yielding plasmid pGAL-SMP3. Haploid smp3::Kan^R strains complemented by pSMP3-426 or pGAL-SMP3 were generated by introducing these plasmids into the smp3::Kan^R/SMP3 strain, allowing the diploids to sporulate, and recovering kanamycin-resistant smp3::Kan^R-pSMP3-426 or pGAL-SMP3 strains from the tetrads that yielded four viable segregants. Both plasmids therefore expressed functional Smp3p.

A gpi10::LEU2/GPI10 diploid was generated by replacing 95% of the GPI10 coding region with the selectable marker LEU2 using the strategy previously used to disrupt the GPI11 gene (22). Plasmid pGPI10-426, in which GPI10 is expressed behind its native promoter, was made by using polymerase chain reaction to amplify a genomic DNA fragment containing GPI10, 465 nucleotides immediately upstream of the start codon, and 337 nucleotides immediately downstream of the GPI10 coding region, after which the DNA fragment was cloned into the 2µ plasmid pRS426. This plasmid was introduced into gpi10::LEU2/GPI10 diploids, which were then allowed to sporulate. The resulting asci were dissected, and gpi10::LEU2-pGPI10-426 segregants were recovered from tetrads that gave four viable segregants, indicating that pGPI10-426 expressed functional Gpi10p.

Radiolabeling of Lipids-- For [³H]inositol labeling, logarithmically growing cells were resuspended at 10 A₆₀₀ units/ml in inositol-free synthetic medium, shifted as appropriate to nonpermissive temperature for 20 min, and then labeled with 15 µCi of [³H]inositol for 90 min. For [³H]inositol labeling of GPIs in the smp3::Kan^R-pGAL-SMP3 strain, cultures were grown in SDGlyYE medium, cells were then resuspended in SDGlcYE medium for 16 h to repress SMP3 expression, and cultures were then pulse-labeled with [³H]inositol for 2 h as detailed in Ref. 22. In the case of the smp3/gpi13-pGAL-SMP3 strain, GPI13 expression was likewise repressed by incubation in SDGlcYE medium at 25 °C, after which portions of the culture were either maintained at 25 °C or shifted to 37 °C for 20 min before pulse labeling with [³H]inositol (22). For [³H]ethanolamine labeling, strains were grown in SD medium supplemented with choline and ethanolamine and then washed and radiolabeled with 50 µCi of ethanolamine in SD medium lacking ethanolamine. Radiolabeled lipids were extracted and separated by TLC using chloroform/methanol/water (10:10:2.5 by volume) as solvent (22).

Characterization of Glycan Headgroups-- [³H]inositol-labeled lipids from 500 A₆₀₀ units of smp3-1 cells labeled at 25 °C were purified by two rounds of preparative TLC. Size analyses of the neutral glycan headgroup of the de-acylated lipid, jack bean, or A. satoi 1,2-mannosidase sensitivity determinations and positioning of EthN-P side-branches were carried out following protocols described in Refs. 22 and 37. A Man₄-GlcNAc standard was obtained from the Man₄-GPI that accumulates in the gpi7-deleted strain (21, 22), and a Man₃-GlcNAc standard was prepared isolated after jack bean -mannosidase treatment followed by dephosphorylation of the deacylated gpi7 lipid.

³⁵S Labeling and Immunoprecipitation of Gas1p-- Pulse labeling with Tran³⁵S-label mixture, chasing with unlabeled methionine and cysteine, and immunoprecipitation of Gas1p were carried out as described previously (38).

Amino Acid Sequence Analyses-- The amino acid sequences of other eukaryotic proteins resembling S. cerevisiae Smp3p (NP_014792), Gpi10p (NP_011373), Alg9p (NP_014180), and Alg12p (NP_014427) were identified in Psi-BLAST, BLASTp, or tBLASTn searches (39, 40) of the NCBI, Sanger Center, and Stanford DNA Sequencing and Technology Center data bases. Each new amino acid sequence was used as probe in a Psi-BLAST search of all S. cerevisiae proteins, and the new protein was then provisionally designated the counterpart of the S. cerevisiae protein with which it gave the alignment with the lowest E value and to which it showed the highest level of amino acid identity and similarity. The GenBank^TM accession numbers of the Schizosaccharomyces pombe Smp3p, Gpi10p, Alg9p, and Alg12p counterparts are Q09837, T41079, T50116, and T39659, respectively. The accession numbers of the Drosophila Smp3p, Gpi10p, Alg9p, and Alg12p counterparts are AAF47201, AAF47795, AAF56419, and AAF54441, respectively. The accession numbers of the human Smp3p, Gpi10p, Alg9p, and Alg12p counterparts are BAB14263, NP_004846 (Pig-B), BAB15154, and AAH01729 respectively. The Candida albicans proteins are encoded by the following contigs of genomic DNA from C. albicans strain SC5314, sequenced by the Stanford DNA Sequencing and Technology Center: CaSmp3p by nucleotides 5967-7460 of contig6-2467, CaGpi10p by nucleotides 14358-15803 of contig6-2493, CaAlg9p by nucleotides 40345-42024 of contig6-2488, and CaAlg12p by nucleotides 38706-40457 of contig6-2478. Sequence data for C. albicans were obtained from the Stanford DNA Sequencing and Technology Center web site. Sequencing of C. albicans was accomplished with the support of the National Institute of Dental and Craniofacial Research and the Burroughs Wellcome Fund. Amino acid sequences were aligned using the CLUSTAL W program (41), and an unrooted phylogenetic tree was generated from that alignment using the DRAWTREE option of the PHYLIP program (42), including positions with gaps and not correcting for multiple substitutions or using branch lengths.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Identification of Smp3p as a Candidate GPI Mannosyltransferase-- Because the fourth Man is added to the yeast GPI precursor in the endoplasmic reticulum (24), and because the three core mannoses added to the GPI precursor in the endoplasmic reticulum are donated by Dol-P-Man (2, 3, 11, 43), the fourth Man on yeast GPIs might also be expected to originate from Dol-P-Man. Two lines of evidence suggested that the Smp3 protein might be a Dol-P-Man-utilizing mannosyltransferase involved in GPI assembly. First, the S. cerevisiae genome encodes several proteins similar to Pig-Bp (19, 20, 43, 44), which is required for addition of the 1,2-linked Man to mammalian GPIs (43). Of these, Gpi10p is the functional homolog of Pig-Bp (19, 20), and Alg9p and Alg12p are nonessential proteins involved in mannosyl transfer to the dolichol-linked precursor in N-glycosylation (45). The SMP3 gene encodes a fourth related protein. The temperature-sensitive smp3-1 mutant had previously been isolated in a screen for yeast mutants that stably maintained a heterologous plasmid (33), and its availability allowed us to test it for a defect in GPI anchoring. Genetic evidence implicating Smp3p in GPI anchoring came from the fact that we isolated an allele of SMP3 in a screen for mutants that require the GPI1 gene for growth.² Gpi1p, which is necessary for growth at 37 °C but not 25 °C, is a subunit of the protein complex that catalyzes the first step in GPI assembly, the formation of GlcNAc-PI (38, 46). The premise behind our synthetic lethality screen was that an additional mutation affecting GPI assembly or transfer to protein might be lethal when combined with the gpi1 mutation. One temperature-sensitive strain isolated in this way was complemented by a centromeric plasmid (47) containing SMP3 that we recovered from a genomic yeast DNA library. We confirmed by integrative transformation that the cloned SMP3 gene was closely linked to the locus conferring temperature sensitivity, and we refer to this mutation as smp3-2.

The temperature-sensitive growth phenotype of both the smp3-1 and smp3-2 alleles was osmotically remediable; inclusion of either 0.25 M KCl or 1 M sorbitol in solid medium restored the ability of these strains to grow at 37 °C. Although SMP3 was shown to be an essential gene (33), we tested whether haploid smp3 disruptants might be capable of vegetative growth if given osmotic support. Heterozygous smp3::Kan^R/SMP3 diploids were allowed to sporulate, and the resulting asci were dissected onto solid YPD medium or onto YPD medium supplemented with 0.6 M KCl or 1.0 M sorbitol. When ascospores were dissected onto YPD medium, only the two kanamycin-sensitive, wild type segregants gave rise to colonies, whereas the smp3::Kan^R segregants either failed to germinate or failed to divide. When dissected onto osmotically supported YPD medium, the smp3::Kan^R segregants completed 4-8 rounds of cell division but then ceased further growth. We conclude that Smp3p is essential for vegetative growth.

smp3 Mutants Accumulate a Candidate GPI Precursor-- To establish whether smp3 mutants are defective in GPI synthesis, we tested whether they accumulate an intermediate in the GPI synthetic pathway. Accumulation of a precursor provides a more sensitive indication of a GPI assembly defect in late-stage yeast GPI anchoring mutants than testing for defects in GPI transfer to protein can. For example, the gpi10-1 mutant accumulates a Man₂-containing GPI precursor but incorporates normal amounts of [³H]inositol into proteins (19).

The smp3-1 and smp3-2 mutants, a wild type strain, and smp3 mutants harboring a centromeric library plasmid containing the SMP3 gene were pulse-labeled with [³H]inositol at 25 and 37 °C, after which labeled lipids were extracted from the cells and separated by TLC. Both strains accumulated a major [³H]inositol-labeled lipid (3-1), as well as a minor, less polar species (3-2), suggesting that they have a defect in GPI synthesis (Fig. 1A, lanes 3, 4, 7, and 8). Accumulation of lipid 3-1 was highest at 25 °C in the smp3-1 strain, whereas smp3-2 accumulated somewhat more at 37 °C. Lipid accumulation by both smp3 strains was abolished when the SMP3 gene was introduced into these strains on a centromeric plasmid (Fig. 1A, lanes 5, 6, 9, and 10), strongly suggesting that their GPI synthetic defect is due to a mutation in the SMP3 gene.

View larger version (47K): [in this window] [in a new window]   Fig. 1.   smp3 mutants accumulate a candidate GPI precursor. A, accumulation of [3H]inositol-labeled lipids and correction of this phenotype by SMP3. Cells were labeled with [3H]inositol at 25 or 37 °C, and radiolabeled lipids were extracted, separated by TLC, and detected by fluorography. Lanes 1 and 2, lipids from wild type cells; lane pairs 3-4 and 7-8, lipids from smp3-1 and smp3-2 cells, respectively; lane pairs 5-6 and 9-10, lipids from smp3-1 and smp3-2 cells harboring a centromeric plasmid expressing SMP3 (pSMP3) behind its native promoter, respectively. Odd-numbered lanes contain lipids labeled at 25 °C, and even-numbered lanes contain lipids radiolabeled at 37 °C. B, depletion of Smp3p leads to accumulation of lipid 3-1. The smp3-pGAL-SMP3 strain was shifted to SGlcYE medium to repress SMP3 expression as described under Experimental Procedures, after which cultures were labeled with [3H]inositol. A control culture incubated in SGalYE was incubated and radiolabeled in parallel. Radiolabeled lipids were extracted, separated by TLC, and detected by fluorography. Lane 1, culture shifted to glucose; lane 2 culture shifted to galactose. C, mild base sensitivity and phosphatidylinositol-specific phospholipase C resistance of lipid 3-1. Samples of lipids from smp3-1 cells labeled with [3H]inositol at 25 °C were submitted to mild base hydrolysis in methanolic NH3 (lane 2), mock-treated with aqueous methanol (lane 1), incubated with phosphatidylinositol-specific phospholipase C (lane 4), or mock-treated (lane 3). Remaining lipids were extracted and separated by TLC. D, radiolabeling of lipid 3-1 with [3H]ethanolamine. The psd1/psd2 (WT) (lane 1) and smp3-1/psd1/psd2 (s3) (lane 2) strains were radiolabeled with [3H]ethanolamine, and the smp3-1/psd1/psd2 strain was labeled in parallel with [3H]inositol (s3) (lane 3). Radiolabeled lipids were separated on the same TLC plate. o- indicates the origin of the chromatogram.

To show that the lipid accumulation phenotype of the temperature-sensitive smp3 mutants can be mimicked by depleting Smp3p in an smp3-disrupted strain, we isolated a haploid smp3::Kan^R strain that was complemented by a plasmid-borne copy of SMP3 under the control of the glucose-repressible GAL10 promoter. A culture of smp3::Kan^R-pGAL-SMP3 cells was incubated in repressing medium for 16-48 h to allow them to become depleted of Smp3p, after which they were pulse-labeled with [³H]inositol. The glucose-repressed cells accumulated a lipid with the same chromatographic mobility as lipid 3-1 accumulated (Fig. 1B, lanes 1 and 2), indicating that lipid 3-1 indeed accumulates as a consequence of the loss of Smp3p function. Lipid 3-2 was not unambiguously resolved in this experiment. Further lipid labeling experiments were done with temperature-sensitive smp3 strains.

Lipid 3-1 is sensitive to mild base hydrolysis but resistant to hydrolysis by phosphatidylinositol-specific phospholipase C (Fig. 1C). The former property indicates the lipid has ester-linked fatty acyl chains, and the latter is consistent with the presence of an acyl chain esterified to the inositol; both are features of late-stage yeast GPI precursors (19, 20, 22-24). Lipid 3-1 could also be radiolabeled with [³H]ethanolamine, consistent with it being a GPI with one or more EthN-P moieties (Fig. 1D, lane 2).

Epistasis Relationships of smp3-- To obtain genetic evidence that lipid 3-1 is made in the GPI assembly pathway, we tested whether its formation is abolished in a mutant defective in the first step in GPI synthesis, and whether, in turn, smp3 blocks the formation of late-stage and "complete" GPI precursors. Double mutants were constructed between smp3 and (i) gpi1, which blocks GlcNAc-PI synthesis, the first step in GPI assembly (38); (ii) gpi11 (22), which is defective in the yeast counterpart of human Pig-Fp (48) and accumulates two Man₄-GPIs; (iii) gpi13-pGAL-GPI13, in which depletion of Gpi13p blocks addition of phosphoethanolamine to Man-3 and leads to accumulation of a Man₄-GPI bearing EthN-P on its first Man (22); and (iv) gaa1, which is blocked in GPI transfer to protein and accumulates the complete GPI precursor CP2 (28).

The smp3-1/gpi1 mutant did not accumulate lipid 3-1 at either 25 or 37 °C (Fig. 2A, lanes 1-4), indicating that 3-1 accumulation is dependent on formation of GlcNAc-PI and placing the smp3 block downstream of gpi1 in the GPI assembly pathway. smp3, however, is epistatic to gpi11, gaa1, and gpi13. In the smp3-1/gpi11 strain, formation of Man₄-GPIs 11-1 (which bears two EthN-Ps, one of which is on Man-3) and 11-2 (which bears a single EthN-P on Man-2; Fig. 2A, lane 5) was almost completely prevented (Fig. 2A, lanes 6 and 7). In the smp3-1/gaa1 double mutant, accumulation of the complete precursor CP2, a Man₄-GPI with three EthN-Ps (19) (Fig. 2A, lane 8), was essentially abolished (Fig. 2A, lanes 9 and 10). We noted that the smp3/gaa1 double mutant had a more severe growth defect than strains harboring the smp3-1 or gaa1-2 mutations alone, because, in contrast to the single mutants, the smp3/gaa1 strain did not grow at 37 °C on solid YPD medium supplemented with 0.25 M KCl. Strikingly, the smp3-1 and smp3-2 mutations inhibited the formation of the Man-Man-Man-(EthN-P)Man-GPI (lipid 13-1) that accumulates in Gpi13p-depleted strains (22) and did so in a temperature-sensitive manner. After shifting the double mutants to 37 °C, formation of lipid 13-1 was reduced in the smp3-1/gpi13-pGAL-GPI13 strain and abolished in smp3-2/gpi13-pGAL-GPI13 (Fig. 2B, lanes 2, 3, 5, and 6), and lipid 3-1 accumulated as it does in smp3 single mutants. The smp3 mutation therefore blocks the formation of the Man₄-GPIs that accumulate in strains defective in each of three essential proteins that act late in the GPI anchoring pathway.

View larger version (34K): [in this window] [in a new window]   Fig. 2.   Epistasis relationships of smp3. A, the smp3-1, gpi1, gpi11, and gaa1 mutants (lanes 1, 2, 5, and 8, respectively) and smp3-1/gpi1 (lanes 3 and 4), smp3-1/gpi11 (lanes 6 and 7), and smp3-1/gaa1 (lanes 9 and 10) double mutants were labeled with [3H]inositol at 25 or 37 °C after a 20-min shift of the culture to 37 °C, and radiolabeled lipids were extracted, separated by TLC and detected by fluorography. Positions of lipids 3-1, 11-1, 11-2, and CP2 are indicated, and their structures are described in the text. B, smp3/gpi13-pGAL-GPI13 strains were incubated for 16 h at 25 °C in SGlcYE medium, and each culture then divided into two portions, one of which was maintained at 25 °C, whereas the other was shifted to 37 °C for 20 min before pulse labeling with [3H]inositol. Radiolabeled lipids were then extracted and separated by TLC. Lanes 2 and 3, [3H]inositol-labeled lipids that accumulate in the smp3-1/gpi13-pGAL-GPI13 strain at 25 and 37 °C, respectively. Lanes 5 and 6, lipids radiolabeled in the smp3-2/gpi13-pGAL-GPI13 strain at 25 and 37 °C, respectively. Lipids radiolabeled in the gpi13-pGAL-GPI13 strain after shift to glucose-containing medium are displayed in lane 1, and lipids labeled in the gaa1 and smp3-2 mutants at 37 °C are separated in lanes 4 and 7, respectively. The asterisk indicates the position of a trace, aberrant lipid in lanes 5 and 6.

smp3 Is Blocked in Addition of a Fourth Man to GPI Precursors-- The finding that smp3 blocks the accumulation of Man₄-GPIs is consistent with the notion that the mutation prevents the addition of the fourth, -1,2-linked mannose to GPIs and led to the prediction that lipid 3-1 is a Man₃-GPI. We tested this by determining the structure of the glycan headgroup of lipid 3-1. [³H]Inositol-labeled 3-1 was isolated by preparative TLC, and its neutral glycan headgroup was obtained following deacylation, HF-dephosphorylation, and re-N-acetylation. Portions of this material were submitted to size analysis by HPTLC without, or after, digestion with -mannosidases. The full-size glycan from lipid 3-1 has a mobility corresponding to that of Man₃-GlcNAc-Ins (Fig. 3A, lane 2). JBM treatment converted the 3-1 glycan to GlcNAc-Ins, whereas A. satoi 1,2-mannosidase converted it to Man₂-GlcNAc-Ins (Fig. 3B, lanes 2-5). These results indicate that the glycan of lipid 3-1 contains three -linked mannoses, the outermost of which is in -1,2 linkage. smp3 is therefore indeed defective in the addition of the fourth, -1,2-linked mannose to the GPI precursor.

View larger version (50K): [in this window] [in a new window]   Fig. 3.   Characterization of the glycan headgroup of lipid 3-1. Strain smp3-1 was radiolabeled with [3H]inositol at 25 °C, and [3H]inositol-labeled lipid 3-1 was isolated by preparative TLC and deacylated with mild base. A, size analysis of the neutral glycan. The headgroup was re-N-acetylated with acetic anhydride and then dephosphorylated with 50% aqueous HF, and the glycan was submitted to HPTLC (lane 2). Lanes 3 and 4 contain Man4-GlcNAc-[3H]Ins (M4) and Man3-GlcNAc-[3H]Ins (M3) size standards prepared from the Man4 GPIs that accumulate in the gpi7 mutant (22). Lane 1 displays a series of NaB[3H]4-reduced Glc2 (G2)-Glc6 (G6) oligomers. B, -mannosidase sensitivity and positioning EthN-P side branches. HF-dephosphorylated glycans were incubated JBM (HF/JBM) (lane 3) or with A. satoi 1,2-mannosidase (HF/1,2M) (lane 5). Mock-incubated controls for the digestions are shown in lanes 2 and 4. A sample of deacylated, re-N-acetylated headgroup was treated first with JBM and then with HF (JBM/HF) (lane 6). Glycans were separated by HPTLC and detected by fluorography. M0, Ins, M1, M2, M3, and M4 indicate the positions of GlcNAc-Ins, inositol, Man-GlcNAc-Ins, Man2-GlcNAc-Ins, Man3-GlcNAc-Ins, and Man4-GlcNAc-Ins, respectively. The mobilities of the GPI glycans relative to the [3H]Glc oligomers correspond to those previously published (27).

The finding that lipid 3-1 could be radiolabeled with [³H]ethanolamine indicated the presence of one or more EthN-Ps on this Man₃-GPI. To determine the position of its EthN-P side-branch(es), the deacylated 3-1 headgroup was first treated with JBM and then dephosphorylated and re-N-acetylated. This succession of treatments yielded two major glycans, one migrating as Man₂-GlcNAc-Ins, the other as Man-GlcNAc-Ins (Fig. 3B, lane 6), indicating that lipid 3-1 is a mixture of structural isoforms. The recovery of Man-GlcNAc-Ins indicates that the GPI it originated from must have borne an EthN-P moiety on its first mannose and therefore that only one EthN-P was present on the original GPI. Because the GPI that gave rise to Man₂-GlcNAc-Ins comigrated with the one that yielded Man-GlcNAc-Ins, this component of lipid 3-1 must likewise bear only one EthN-P side-branch. We conclude that lipid 3-1 consists predominantly of a mixture of two Man₃-GPI isoforms, one bearing EthN-P on Man-2 and one bearing EthN-P on its first Man. Because traces of material with the mobility of Man₃-GlcNAc-Ins were also present in the sample in Fig. 3B, lane 6, it is possible that the lipid 3-1 mixture also contained a small amount of a Man₃-GPI with a single EthN-P moiety on Man-3, although the presence of this material could also be explained by incomplete JBM digestion.

smp3 Strains Have a Partial Defect in GPI Anchoring of Protein-- We tested whether the smp3 mutants are defective in GPI attachment to protein by examining GPI anchor-dependent processing of Gas1p, a standard procedure for detecting a GPI anchoring defect in yeast (49). A block in GPI attachment to Gas1p in turn prevents maturation of a 125-kDa form of the protein in the Golgi and causes Gas1p to remain in a core-glycosylated, 105-kDa form (49). The two forms of Gas1p are detected by pulse labeling smp3 cells with [³⁵S]methionine at 25 or 37 °C and then performing a chase during which cultures are maintained at 25 or 37 °C, after which ³⁵S-labeled Gas1p is immunoprecipitated. In the wild type control strain, only the 125-kDa form of Gas1p was seen after the chase period (Fig. 4, lanes 1 and 2), whereas in the gpi1 control, only the 105-kDa protein was present after a 60-min chase at 37 °C (Fig. 4, lane 3). The smp3-2 mutant showed a partial defect in Gas1p processing at 25 °C (Fig. 4, lanes 4-6), as did the gpi1 mutant at 25 °C (38). Gas1p maturation remained incomplete at 37 °C in smp3-2 (Fig. 4, lanes 7-9). The partial block in Gas1p processing suggests that Smp3p is required for efficient transfer of GPIs to protein.

View larger version (38K): [in this window] [in a new window]   Fig. 4.   smp3 is partially defective in Gas1p processing. smp3-2 and wild type cells were pulse-labeled at 25 or 37 °C, and gpi1 cells were labeled at 37 °C with Tran[35S]-label mixture. Samples of the smp3 cultures were taken at the end of the pulse labeling period (time 0, lanes 4 and 7). The radioactivity in the labeled cultures was then chased with unlabeled amino acids for 30 or 60 min at 25 °C (lanes 5 and 6) or 37 °C (lanes 8 and 9). 35S-Labeled Gas1p was then immunoprecipitated from extracts of the cells, separated by SDS-polyacrylamide gel electrophoresis, and detected by fluorography. Lanes 1 and 2, [35S]Gas1p immunoprecipitated from the wild type control strain after a 60-min chase. Lane 3, [35S]Gas1p from gpi1 cells after a 60-min chase.

Neither smp3-1, smp3-2, nor the smp3-pGAL-SMP3 strain shifted to glucose showed a discernible defect in [³H]inositol incorporation into protein. The incompleteness of the Gas1p processing block and the lack of effect on [³H]inositol incorporation into protein may reflect a degree of leakiness of the smp3 mutation or incomplete repression of SMP3 expression. Alternatively, a Man₃-GPI may be transferred to protein (see under "Discussion"), although if this is the case, these aberrant protein-bound GPIs do not allow the cells to grow under nonpermissive conditions.

Smp3p-related Proteins Form a Subgroup of a Family of Candidate Dol-P-Man-dependent Mannosyltransferases-- Searches of sequence data bases revealed that the genomes of S. pombe, C. albicans, Drosophila melanogaster, and Homo sapiens encode proteins resembling S. cerevisiae Smp3p. The GenBank^TM accession numbers of these protein sequences are given under "Experimental Procedures." The new Smp3ps show amino acid sequence similarity to the Alg9, Alg12, and Pig-B(Gpi10) proteins but are unlikely to be the sequence and functional homologs of these putative mannosyltransferases because the S. pombe, D. melanogaster, and C. albicans genomes all encode obvious counterparts of the latter proteins.

To show that the Smp3p-related proteins are indeed a separate subgroup of the Alg9p/Alg12p/Pig-Bp(Gpi10p) family, we generated a CLUSTAL W alignment of the amino acid sequences of 20 of these proteins and used it to produce an unrooted phylogenetic tree using the DRAWTREE option of the PHYLIP program (42). This analysis confirmed that the Smp3p-related proteins represent a separate group, as do the proteins designated the counterparts of Alg9p, Alg12p, and Pig-Bp(Gpi10p) (not shown). The Smp3 proteins are also distinguishable from other members of the family of putative mannosyltransferases because they have a variation on an amino acid sequence motif (19) that characterizes these proteins. Thus, the sequence ³¹³HQEXRF in the five Smp3-related proteins is HKEXRF in all Alg9, Alg12, and Pig-B(Gpi10) proteins. (The first amino acid in these blocks is numbered according to its position in S. cerevisiae Smp3p, and X denotes a less conserved amino acid.) The family of candidate mannosyltransferases seems to include only four easily recognizable relatives; our searches of eukaryotic protein sequences using various Smp3p, Alg9p, Alg12p, and Pig-Bp(Gpi10p) sequences as probes have so far detected only proteins with obvious amino acid sequence similarity to one or other of these four proteins.

We have not established whether the human Smp3p counterpart is the functional equivalent of yeast Smp3p, but we note that the presence of a human gene for such a protein is consistent with the fact that traces of GPIs that have been speculated to bear a fourth Man have been detected in lipid extracts of mammalian cells (13, 29).

Smp3p and Gpi10p Do Not Substitute for One Another in Vivo-- If the Gpi10 and Smp3 proteins are mannosyltransferases, then they transfer mannose to closely related GPI structures to form -1,2-mannosidic linkages. We therefore tested whether these enzymes might exhibit any cross-specificity for each other's acceptor if expressed at high levels in stains deficient in the other protein. However, overexpression of GPI10 behind its native promoter on a 2µ plasmid neither restored the ability of smp3-1 strains to grow at 37 °C nor had a discernible effect on the lipid accumulation phenotype of smp3-1. Further, smp3 null mutants could not be rescued by overexpression of GPI10, and conversely, SMP3 did not restore viability to gpi10 disruptants when present on a 2µ vector. Smp3p and Gpi10p therefore cannot substitute for one another in vivo under the conditions used, indicating that these putative mannosyltransferases have a high degree of specificity for their respective acceptor GPIs in vivo.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The key findings reported here are that the essential Smp3 protein is required for addition of the fourth Man to yeast GPI precursors; that in the absence of Smp3p function, transfer of EthN-P to the third GPI Man is severely, if not completely blocked; and that Smp3p function is needed for efficient transfer of GPI precursors to yeast protein.

Function and Essential Role of Smp3p-- The simplest explanation for the GPI assembly defect in smp3 mutants and for the fact that Smp3p resembles other proteins implicated in the transfer of Man from Dol-P-Man to manno-oligosaccharides is that Smp3p is the mannosyltransferase that adds the fourth, -1,2-linked Man during assembly of yeast GPIs. Assuming that the primary biochemical defect in conditional or null smp3 mutants is in the addition of the fourth Man to GPIs, then this mannosyl residue itself must be critical, either for completion of the substrate(s) of the transamidase complex that adds GPIs to protein, or, if not for formation of transfer-competent GPIs per se, then for some subsequent function of the protein-bound GPI. We discuss these possibilities below.

The evidence to date strongly indicates that the GPI(s) that are normally transferred to protein in yeast all bear a fourth, -1,2-linked mannose residue. Thus, protein-bound GPIs in yeast bear at least four mannoses (50), and the gaa1 and gpi8 mutants, which are defective in GPI transfer to protein, accumulate Man₄-GPIs, the most prominent of which is the complete precursor Man-(EthN-P)Man-(EthN-P)Man-(EthN-P)Man-GlcN-(acyl-Ins)PI (27, 28). Because smp3 is epistatic to gaa1, as well as to gpi11 and gpi13, which both accumulate Man₄-GPIs, Smp3p may be required for the formation of all GPI transamidase substrates, or at least enough "Man₄-complete precursors" to maintain viability. The latter possibility explains the fact that the Smp3p-deficient strains we have tested show a partial defect in GPI transfer to protein. Despite the strong lipid accumulation phenotype in these strains, they might nonetheless be leaky and still capable of making Man₄ transamidase substrates and attaching them to protein, although at levels too low to support cell growth under nonpermissive conditions. Another possibility, namely that the apparent leakiness of the smp3 deficient strains is due to the transfer of Man₃GPIs to protein, is considered below.

If the presence of the fourth Man is indeed obligatory for a GPI to serve as a transamidase substrate, then our results suggest a specific requirement for this residue, namely, that the fourth mannose is important or necessary for transfer of the "bridging EthN-P" to Man-3 of the GPI precursor. Thus, Gpi13p-deficient mutants, which are blocked in the addition of EthN-P to Man-3, accumulate a GPI with the structure Man-Man-Man-(EthN-P)Man-GlcN-(acyl-Ins)PI (lipid 13-1 in Ref. 22; Fig. 2B, lane 1), suggesting that the acceptor GPI recognized by Gpi13p is this Man₄GPI. Our finding that the smp3-2 mutation abolishes the formation of any detectable lipid 13-1 in the smp3-2/gpi13-pGAL-GPI13 strain (Fig. 2B, lane 6) in turn raises the possibility that Gpi13p requires its acceptor GPI to bear a fourth mannose. We note that if Gpi13p could act on a Man₃-GPI before Smp3p does, then a single smp3 mutant might be expected to show an accumulation of trimannosyl GPIs bearing two or more EthN-Ps, which was not observed.

Although the earliest consequence of an Smp3p deficiency is a severe impairment in the formation of Man₄-GPIs, including the presumed acceptor for addition of the bridging EthN-P, it is possible that the fourth Man plays its critical role after GPI transfer to protein. Thus, EthN-P may be added to Man-3 of Man₃-GPIs in smp3 mutants, and such Man₃-GPIs may, in turn, be transferred to protein. This would also explain the apparent leakiness of smp3 mutants with respect to GPI transfer to protein. Lipid 3-1 contains traces of material that could have originated from (EthN-P)Man-Man-Man-GlcN-(acyl-Ins)PI (Fig. 3B, lane 6), a species that could itself serve as GPI transamidase substrate and that would accumulate in the smp3/gaa1 double mutant. However, if (EthN-P)Man-Man-Man-GlcN-(acyl-Ins)PI were to receive additional EthN-P side-branches on Man-1 or Man-2 or both and be converted to the Man₃ counterpart of complete precursors, then such species should have been detectable in significant amounts in some of our lipid radiolabeling experiments as polar GPIs with chromatographic mobilities distinct from those of the known yeast Man₄-GPIs. Traces of polar lipids that are potential aberrant GPIs were detectable in extracts of certain smp3 strains (Fig. 2B, lanes 5 and 6, asterisk) or became visible after prolonged exposure of TLC plates to x-ray film, but the amounts of these species are too small to verify whether they are Man₃-GPIs. Importantly, however, if such trace lipids are indeed Man₃-GPIs capable of being transferred to protein, they would be predicted to be present at much higher levels in the smp3/gaa1 double mutant; however, no new polar lipids accumulated when this transamidase-defective strain was shifted to nonpermissive temperature (Fig. 2A, lane 10). We note too that the smp3/gaa1 double mutant has a more severe growth defect than either smp3 or gaa1 alone, suggesting that Smp3p-dependent addition of the fourth Man is necessary for the generation of optimal GPI transamidase substrates.

Although our results provide little support for the notion that Man₃-GPIs bearing two or more EthN-Ps are formed, they do not exclude the possibility that (EthN-P)Man-Man-Man-GlcN-(acyl-Ins)PI can be transferred to protein in smp3 mutants. However, even if this occurred and some or all of the GPI anchors on yeast protein were to contain only three mannoses, this would not be sufficient for growth, as indicated by the inviability of smp3 disruptants and the conditional lethality of the smp3-1 and smp3-2 strains. The inability of protein-bound Man₃-GPIs to support growth in turn would imply that the fourth mannose, which is normally present on all yeast GPI anchors (50), fulfills another, essential function after anchor transfer to protein. One possibility is that this additional Man is necessary for incorporation of certain mannoproteins into the yeast cell wall upon creation of a cross-link between a Man in the GPI and -1,6-glucan (4, 5). The fourth Man, however, may not form the attachment point for all mannoproteins, because in some cell wall proteins, the linkage to -glucan may be through the reducing end of a core GPI Man (5). Another possible role for the fourth mannose on protein-bound GPIs, consistent with the partial Gas1p processing defect of smp3-2, is that the fourth mannose may be required for efficient transport of GPI anchored proteins from the endoplasmic reticulum.

Although we cannot yet pinpoint the essential role of Smp3p, the simplest reason why this protein is necessary is the one implied by the earliest discernible biochemical consequence of a block in the addition of the fourth Man to yeast GPI precursors, namely, an inability to form Man-Man-Man-(EthN-P)Man-GlcN-(acyl-Ins)PI, the presumed acceptor for EthN-P transfer to Man-3. This in turn implies that addition of the fourth Man is critical for addition of the bridging EthN-P to Man-3 in yeast. The incomplete block in GPI transfer to protein in smp3 mutants must then be attributed to the leakiness of these strains and their consequent ability to make and transfer some Man₄-complete precursor to protein, even though these mutants accumulate Man₃-GPIs lacking the bridging EthN-P. The gpi10-1 strain provides a precedent for this possibility; it has a strong lipid accumulation phenotype yet exhibits neither an apparent GPI anchoring defect nor temperature sensitivity for growth (19). Nonetheless, Gpi10p is an essential protein because it is required for addition of the third mannose during GPI precursor assembly (19, 20).

Implications of the Structural Heterogeneity of Lipid 3-1 for GPI Assembly in Yeast-- The major GPIs that can be radiolabeled in yeast mutants can be arranged in a scheme that implies a branched assembly pathway (22), and each of the two major isoforms of lipid 3-1 can be placed in one of the branches (Fig. 5). The Man-Man-(EthN-P)Man-GlcN-(acyl-Ins)PI isoform of lipid 3-1 (lipid 3-1-1) can be inserted between the Gpi10p and Gpi13p steps in an (EthN-P)Man-1 arm that is defined by a succession of intermediates from Man-(EthN-P)Man-GlcN-(acyl-Ins)PI (19, 20) to complete precursor that all bear EthN-P on Man-1 (19-23). The Man-(EthN-P)Man-Man-GlcN-(acyl-Ins)PI isoform of 3-1 (lipid 3-1-2) is a potential precursor of the Man₄-GPI with EthN-P on Man-2 that accumulates in the gpi11 mutant (22). A straightforward explanation for the formation of the latter two GPIs is that they are intermediates in an (EthN-P)Man-2 arm of a branched pathway. However, no obvious precursor to lipid 3-1-2 has been identified in strains deficient in addition of the third GPI mannose. Thus, although the gpi10 mutant accumulates Man-(EthN-P)Man-GlcN-(acyl-Ins)PI, the likely precursor of lipid 3-1-1 (19, 20), strains with a Gpi10p-deficiency alone have not been reported to accumulate Man-Man-GlcN-(acyl-Ins)PI or (EthN-P)Man-Man-GlcN-(acyl-Ins)PI species that could serve as precursors of lipid 3-1-2. Although an as yet undiscovered Man-2-substituted GPI may be generated earlier, it is also possible that pathway branching occurs after Gpi10p-dependent addition of Man-3. This could involve an "isomerization" in which EthN-P is removed from Man-1 and one is added to Man-2, or it could involve the addition of EthN-P to Man-2 of an unsubstituted Man₃-GPI. Although the latter GPI has not been demonstrated unambiguously, it is possible that lipid 3-2 is indeed Man-Man-Man-GlcN-(acyl-Ins)PI (Fig. 5), but we have so far been unable to resolve sufficient amounts of lipid 3-2 for characterization of its glycan headgroup.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Model for assembly pathway for yeast GPI assembly and site of the smp3 block. Illustrated are eight GPI structures that have been determined in detail (19-23) and the steps that are blocked in GPI anchoring mutants. Dashed arrows indicate the routes of the two putative pathway branches; it is not known whether both pathways lead to protein-bound GPIs or whether some GPIs remain "free." The precursor of lipid 3-1-2 is unknown, but possible intermediates in its synthesis are italicized, and the possibility that lipids 3-1-1 and 3-1-2 can be interconverted is indicated by a double-headed arrow. The structures of the GPI(s) that accumulate in mcd4 have not been determined, but mcd4 is likely to be defective in EthN-P addition to Man-1 (18, 20, 56). M, mannose; G, glucosamine; P, phosphate; E, ethanolamine.

We cannot rule out the formal possibility that the Man-2-substituted GPIs of the (EthN-P)Man-2 pathway are nonphysiological lipids generated only in GPI assembly mutants. However, the formation of Man₃- and Man₄-GPIs lacking EthN-P on Man-1 suggests that addition of EthN-P to Man-1 is not absolutely required for transfer of the third, 1,2-linked mannose, although treatment of cells with an apparent inhibitor of EthN-P addition to Man-1 leads to accumulation of Man-Man-GlcN-(acyl-Ins)PI (20).

Smp3p and Bridging EthN-P Addition as a Targets for Antimicrobial Agents-- Because GPI synthesis is essential in fungi and protozoa (6-9), this process could be targeted by drugs that exploit differences between mammals and microbes in the enzymology of GPI assembly. One proposed target is the addition of the first Man in trypanosomal GPI biosynthesis, which, unlike in mammals, does not require prior inositol acylation of GlcN-PI (51, 52). The identification of a species-specific inhibitor of EthN-P transfer to Man-1 (20, 53) indicates that even the same step in GPI assembly can be inhibited selectively in different organisms, a finding that validates the search for further selective inhibitors.

Our results suggest that both Smp3p-dependent addition of the fourth mannose to GPI precursors and Gpi13p-dependent transfer of the bridging EthN-P to a Man₄-GPI are reactions that could be targeted selectively by antifungal agents. Thus, from the structures and relative abundance of the GPI precursors that can be detected in mammalian cells, it is clear that Man₄-GPI precursors are at best rarely formed and that the bridging EthN-P is readily added to Man-3 of a Man₃-GPI (13-18, 29). Moreover, cell lines deficient in the mammalian counterparts of the GPI transamidase components Gpi8p and Gaa1p accumulate Man₃-GPIs with one to three EthN-P substituents, suggesting that the major mammalian GPI transamidase substrate has three mannoses (54, 55). This is consistent with the fact that many protein-bound GPIs in mammalian cells have only three mannoses (1, 51). The opposite holds in S. cerevisiae; all protein-bound GPIs bear a fourth, -1,2-linked mannose (50), the presumed GPI transamidase subunits are Man₄-GPIs, and the results of the present study indicate that at best, only very small amounts of Man₃-GPIs bearing the bridging EthN-P can be formed. The apparent rarity of free Man₄-GPIs in mammalian cells raises the possibility that addition of a fourth mannose to GPI precursors may be dispensable, in contrast to yeast, whereas Smp3p-dependent addition of Man-4 is essential for viability. Smp3p therefore represents a potential target for antifungal agents, and indeed, a likely Smp3p counterpart with 35% identity and 54% similarity to S. cerevisiae Smp3p is encoded in the genome of the fungal pathogen C. albicans. We note that the GPI biosynthetic intermediates made by P. falciparum also include structures bearing a fourth mannose (25, 26); the transferase that adds this mannose may therefore also be a potential target for antimalarial agents.

The transferases that add the bridging EthN-P to fungal GPIs are also potential drug targets. The Gpi13 and Pig-O proteins, which are required for addition of EthN-P to Man-3 of yeast and mammalian GPI precursors, respectively (22, 23, 29), appear to differ in their specificities for GPI glycans. Thus, Pig-Op transfers EthN-P to a Man₃-GPI, whereas we have presented evidence here that its sequence homolog, Gpi13p, shows strong, possibly absolute specificity for a Man₄-GPI as acceptor for the bridging EthN-P. This difference in acceptor specificity could, in principle, also be exploited in the development of antifungal drugs.

	ACKNOWLEDGEMENTS

We thank Dr. L. Popolo for anti-Gas1p serum; Drs. S. Harashima, D. Voelker, and M. Walberg for strains; and D. Meling for sequence analyses. Sequence data for C. albicans were obtained from the Stanford DNA Sequencing and Technology Center web site. Sequencing of C. albicans was accomplished with the support of the National Institute of Dental and Craniofacial Research and the Burroughs Wellcome Fund.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant GM46220 and by a Helen Corley Petit Professorship from the College of Liberal Arts and Sciences of the University of Illinois at Urbana-Champaign.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Recipient of a Burroughs Wellcome Fund Scholar Award in Pathogenic Mycology. To whom correspondence should be addressed: Dept. of Biochemistry, University of Illinois at Urbana-Champaign, 309 Roger Adams Laboratory, 600 S. Mathews Ave., Urbana, IL 61801. Tel.: 217-333-4139; Fax: 217-244-5858; E-mail: p-orlean@uiuc.edu.

Published, JBC Papers in Press, May 16, 2001, DOI 10.1074/jbc.M101986200

² B. A. Westfall and P. Orlean, unpublished data.

	ABBREVIATIONS

The abbreviations used are: GPI, glycosylphosphatidylinositol; Dol-P-Man, dolichol phosphate mannose; EthN-P, phosphoethanolamine; HPTLC, high performance thin layer chromatography; JBM, jack bean -mannosidase; Ins, inositol.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES