Jul 20, 1995 - Based on the synthetic lethal pheno- type of double mutations affecting the assembly of the lipid-linked core-oligosaccharide and the OTase.
The EMBO Journal vol.14 no.20 pp.4949-4960, 1995
STT3, a highly conserved protein required for yeast oligosaccharyl transferase activity in vivo
Romain Zufferey, Roland Knauer1, Patricie Burda, Igor Stagljar2, Stephan te Heesen, Ludwig LehIel and Markus Aebi3 Mikrobiologisches Institut, ETH Zurich, CH-8092 Zurich, Switzerland and 'Institut fur Zellbiologie der Universitat Regensburg, D-93053 Regensburg, Germany 2Present address: Institut fur Molekularbiologie, Universitat Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland 3Corresponding author
N-linked glycosylation is a ubiquitous protein modification, and is essential for viability in eukaryotic cells. A lipid-linked core-oligosaccharide is assembled at the membrane of the endoplasmic reticulum and transferred to selected asparagine residues of nascent polypeptide chains by the oligosaccharyl transferase (OTase) complex. Based on the synthetic lethal phenotype of double mutations affecting the assembly of the lipid-linked core-oligosaccharide and the OTase activity, we have performed a novel screen for mutants in Saccharomyces cerevisiae with altered N-linked glycosylation. Besides novel mutants deficient in the assembly of the lipid-linked oligosaccharide (alg mutants), we identified the STT3 locus as being required for OTase activity in vivo. The essential STT3 protein is -60% identical in amino acid sequence to its human homologue. A mutation in the STT3 locus affects substrate specificity of the OTase complex in vivo and in vitro. In stt3-3 cells very little glycosyl transfer occurs from incomplete lipid-linked oligosaccharide, whereas the transfer of full-length Glc3Man9GlcNAc2 is hardly affected as compared with wild-type cells. Depletion of the STT3 protein results in loss of transferase activity in vivo and a deficiency in the assembly of OTase complex. Keywords: endoplasmic reticulum/glycosylation/oligosaccharyl transferase/Saccharomyces cerevisiae/transmembrane protein
Introduction N-linked glycosylation is an essential modification of secreted proteins in eukaryotic cells. The process is initiated in the endoplasmic reticulum (ER) and follows a highly conserved pathway (Kornfeld and Kornfeld, 1985; Tanner and Lehle, 1987; Herscovics and Orlean, 1993). A core oligosaccharide, Glc3Man9GlcNAc2, is first assembled on the lipid carrier dolichyl pyrophosphate and then transferred en bloc to selected asparagine residues of nascent polypeptide chains. The acceptor asparagine residue is located within the consensus sequence AsN-XSer/Thr, where X can be any amino acid except proline K Oxford University Press
(Gavel and Von Heijne, 1990). This central reaction in the N-linked glycosylation process is catalysed by the oligosaccharyl transferase (OTase) complex. Purified mammalian and avian OTase consists of ribophorin I, ribophorin II and OST48 (Kelleher et al., 1992; Kumar et al., 1994). In the yeast Saccharomyces cerevisiae, the three corresponding homologous proteins Ostlp (Silberstein et al., 1995), Swplp (te Heesen et al., 1993) and Wbplp (te Heesen et al., 1992) were shown to be essential for OTase activity in vivo and in vitro, but purification of the yeast OTase activity revealed that additional polypeptides of 34 kDa (Kelleher and Gilmore, 1994; Knauer and Lehle, 1994; Pathak et al., 1995b), 16 and 9 kDa (Kelleher and Gilmore, 1994) co-fractionated with the activity. The specific functions of these polypeptides are not yet known; however, Wbplp was postulated to be directly involved in catalysis and the binding of the lipid-linked oligosaccharide substrate (Breuer and Bause, 1995; Pathak et al., 1995a). Current models suggest that the initial steps of the oligosaccharide biosynthesis, up to the synthesis of Man5GlcNAc2, occur on the cytoplasmic side of the ER membrane. Nucleotide-activated sugars serve as donors for these reactions. After the translocation of the lipidlinked Man5GlcNAc2 from the cytoplasmic to the lumenal side of the membrane (a reaction which is poorly understood), additional mannose and glucose residues are added, whereby dolichyl phosphate-linked monosaccharides serve as donors (Komfeld and Komfeld, 1985; Kukuruzinska et al., 1987; Tanner and Lehle, 1987; Herscovics and Orlean, 1993). Genetic tools have been applied to characterize this biosynthetic pathway at the ER membrane in higher eukaryotic cells lines (Cummings, 1992) as well as in S.cerevisiae (Kukuruzinska et al., 1987; Tanner and Lehle, 1987; Herscovics and Orlean, 1993). In S.cerevisiae, alg (asparagine linked glycosylation) mutants deficient in the assembly of the lipid-linked oligosaccharide were isolated (Huffaker and Robbins, 1982, 1983; Runge et al., 1984; Runge and Robbins, 1986). ALG genes required for the early steps in the assembly pathway are essential, whereas mutations in ALG genes necessary for late reactions do not show a detectable growth phenotype. This is due to the relaxed specificity of the OTase complex: incomplete oligosaccharides are transferred to asparagine residues of secretory proteins, albeit with a reduced efficiency (Turco et al., 1977; Staneloni et al., 1980; Trimble et al., 1980; Murphy and Spiro, 1981; Sharma et al., 1981; Munoz et al., 1994). The pathway of the assembly of the lipid-linked oligosaccharide is far from being understood. Only a limited number of the biosynthetic enzymes are characterized and little is known about the spatial organization of this pathway at the ER membrane (Komfeld and Komfeld, 4949
R.Zufferey et aL
1985; Abeijon and Hirschberg, 1992; Herscovics and Orlean, 1993). In order to study the process of N-linked glycosylation of secretory proteins in the ER in more detail, yeast genetic techniques were applied to identify individual biosynthetic steps and to elucidate the structural organization of the pathway. In this report, the identification of novel genes required for the N-linked glycosylation process is described. In particular, we show that the essential, highly conserved STT3 protein is required for assembly and function of OTase activity.
Results A screen for mutants deficient in the pathway of N-linked glycosylation Mutations in the OTase component Wbplp may result in a low activity of this enzyme and a temperature-sensitive phenotype. Since reduced OTase activity as measured in extracts of wbpl strains is not directly affected by the temperature, it was concluded that the temperaturesensitive phenotype of wbpl mutant cells is the consequence of the decreased OTase activity (te Heesen et al., 1992). Combinations of the wbpl mutation with alg mutations (deficient in the assembly of the lipid-linked oligosaccharide) result in synthetic lethality, thereby defining a novel phenotype of incomplete biosynthesis of dolichyl-linked carbohydrates (Stagljar et al., 1994). To identify novel loci involved in the pathway of N-linked glycosylation in the ER, we took advantage of this phenotype and screened for mutations synthetically lethal in combination with the wbpl mutation. For that purpose, we applied the red/white sectoring assay (Kranz and Holm, 1990). This screen was successfully used for the isolation of new components of the nuclear pore complex or cell cycle regulatory genes (Bender and Pringle, 1991; Costigan et al., 1992; Wimmer et al., 1992). Four yeast wbpl strains (relevant genotype wbpl-l ade2 ade3 ura3 or wbpl-2 ade2 ade3 ura3) (see Table IV) were transformed with the YCp5O-derived plasmid pCHl 122(WBPl ) which contains the ADE3-, WBPI- and URA3-loci. When grown on complete YPD medium at 23 or 30°C, these cells may lose the plasmid, because wbpl mutations have a temperature-sensitive phenotype at 37°C only. This plasmid loss is visualized by a sectored appearance of resulting colonies (Kranz and Holm, 1990). Mutations synthetically lethal with wbpl should result in the inability to lose the plasmid-encoded WBPI wild-type allele and therefore lead to a uniformly red colour of the corresponding colony. After UV mutagenesis, 314 000 individual colonies were screened and 271 independent mutant strains were found to have a uniformly red phenotype on complete YPD medium at 23°C or 30°C. In addition, they were unable to grow on medium containing 5-fluoro-orotic acid (5FOA). Such conditions are selective for cells not carrying the URA3 plasmid, because in the presence of the URA3 locus 5FOA is converted to a toxic uracil analogue (Boeke et al., 1991). Due to the plasmid requirement of the mutants, the plasmid-encoded URA3 locus gave a 5FOAsensitive phenotype. This 5FOA sensitivity provided additional evidence that the mutant strains required the plasmid for growth. Complementation analysis based on 5FOA sensitivity at 23°C allowed us to classify 186 mutant strains into 25 complementation groups; 85 mutant strains
4950
Table I. Mutations synthetically lethal with wbpl
Complementation group
No. of alleles
Locus
Phenotype reference
1
5
ALG 3
2
11
ALG 5
3 4
6 13
ALG 6 ALG 8
5
5
ALG 9
6
12
ALG 10
Huffaker and Robbins (1983) Huffaker and Robbins (1983) Runge et al. (1984) Runge and Robbins (1986) accumulation of Man6 (P.Burda et al., in preparation) accumulation of Man9Glc2 (P.Burda et al., in
7
6
STT 3
8
5
SLW 4
preparation) Yoshida et al. (1992), this report accumulation of Man9Glc3 (G.Reiss et al., in preparation)
The different mutant strains were classified into complementation groups based on the 5FOA-sensitive phenotype. All mutations result in an underglycosylation of CPY. The loci defined by the different groups were identified by transformation of representative strains with ALG3, ALGS, ALG6 or ALG8-containing plasmids (groups 1-4), by analysis of the accumulation of lipid-linked oligosaccharide (groups 5, 6 and 8) or by the isolation of the wild-type locus (group 7).
could not be attributed to complementation groups and are not dealt with further here. Twenty-three complementation groups were shown to be dependent on a wild-type WBPI locus; the remaining two groups required the ADE3 locus for growth (Bender and Pringle, 1991). As expected, mutations in one WBPI-dependent complementation group were located in the chromosomal wbpl locus itself, inactivating the essential WBPI function (data not shown). In order to classify the 22 remaining groups, representative mutants were chosen and screened for defects in the endoplasmic N-linked glycosylation pathway. To visualize such defects, carboxypeptidase Y (CPY) expression was studied in the mutant cells. CPY is a highly expressed vacuolar peptidase containing four N-linked oligosaccharides (Hasilik and Tanner, 1978). Perturbation in the ER process of N-linked glycosylation is reflected by the altered mobility of CPY in SDS-PAGE (te Heesen et al., 1992; Stagljar et al., 1994). Mutations in eight complementation groups (listed in Table I) led to an underglycosylation of CPY, whereas mutants of the remaining 14 groups did not show any alterations in CPY processing (data not shown). We assume that these mutants do not tolerate the underglycosylation of proteins induced by the wbpl mutation but affect processes required after N-linked glycosylation of polypeptides in the ER. These mutants are not characterized further in this report.
Complementation groups synthetically lethal with wbpl define known ALG loci and novel genes The screening procedure that we applied is based on the known synthetic phenotype of alg wbpl double mutants. Beside novel ALG loci, we expected to identify novel
STT3 is essential for OTase activity
alleles of already known ALG genes. In order to test this prediction, members of the different complementation groups were transformed with different plasmids containing the known, non-essential ALG loci, either ALG3 (S.te Heesen and M.Aebi, unpublished results), ALG5 (te Heesen et al., 1994), ALG6 (G.Reiss, S.te Heesen, J.Zimmerman, P.W.Robbins and M.Aebi, unpublished results) or ALG8 (Stagljar et al., 1994). In these experiments, we took advantage of the fact that the synthetic lethality observed in all eight complementation groups was rescued when the strains were grown in medium containing sorbitol. This allowed us to isolate double mutant strains which did not contain the WBPI ADE3 URA3-carrying plasmid pCH 1122(WBP1). Transformation of these strains with a WBPI-carrying plasmid resulted in normal growth on medium without sorbitol at all temperatures, whereas the empty vector did not support growth on such media at 30°C. Introduction of the ALG3 locus into double mutant strains representing complementation group 1 resulted in growth on medium lacking sorbitol at 30°C but not at 37°C. This phenotype is expected for the wbpl single mutant and thus indicates the complementation of the synthetic lethal phenotype. We conclude that complementation group 1 contained mutants in the ALG3 locus. This conclusion was supported by the alg3-specific underglycosylation pattern of CPY observed in group 1 strains (data not shown). Similar experiments showed that complementation group 2 represents ALG5, group 3 ALG6 and group 4 ALG8 (data not shown), whereas the analysis of lipid-linked oligosaccharides of mutant strains from group 5 and 6 revealed accumulation of Man6GlcNAc2 and Glc2Man9GlcNAc2, respectively (P.Burda, S.te Heesen and M.Aebi, manuscript in preparation). Mutants in these groups are supposed to be defective in biosynthetic steps of the lipid-linked oligosaccharide. Therefore, the two loci identified by these complementation groups were termed ALG9 and ALG1O, respectively. Mutant strains in the two remaining groups 7 and 8 accumulated normal Glc3Man9GlcNAc2 lipidlinked oligosaccharide (for group 7, see below), as did wild-type cells. Complementation analysis revealed that these strains were not defective in one of the known components of the OTase complex [WBPI (te Heesen et al., 1992), SWPI (te Heesen et al., 1993), OSTJ (Silberstein et al., 1995), OST2 (R.Gilmore, personal communication), OST3 (R.Knauer, R.Zufferey, S.te Heesen, M.Aebi and L.Lehle, manuscript in preparation) or OST4 (N.Dean, personal communication)]. Complementation group 7 was analysed further and, for reasons discussed below, the locus represented by this complementation group was termed STT3. The remaining group 8 was named SLW4 (synthetic lethal with wbpl). We have analysed the function of the SIT3 locus in more detail.
The STT3 locus encodes an essential, highly conserved transmembrane protein One representative mutant strain (carrying the stt3-3 allele) from complementation group 7 was selected and backcrossed twice to the wild-type strain SS328. The stt3-3 mutation is synthetic lethal in combination with the wbpl-2 mutation when grown at 30°C but not at 23°C, enabling us to isolate stt3-3 wbpl-2 double mutants in the absence of complementing plasmid DNA. The stt3-3 mutation
Vector OST2 SNvP1 STT3
-mCPY -2
1
2
3
4
Fig. 1. Underglycosylation of CPY in stt3-3 cells and complementation by the S7T3 locus. Yeast strain YG149 was transformed with the plasmids YEp352 (Vector, lane 1), the SWPIcarrying plasmid YEp352(SWP-PstI-BamHI) (SWPI, lane 2), the OST2 isolate YEp352 (OST2) (OST2, lane 3) and the STT3-carrying plasmid pSTT3 (ST73, lane 4). Cells were labelled for 1 h with [35S]methionine and [35S]cysteine, extracts were prepared and CPYspecific antibodies were used for immunoprecipitation. The position of mature CPY (mCPY) and the glycoforms lacking one (-1) or two (-2) oligosaccharides are indicated.
itself had no detectable phenotype at 30°C, led to slightly reduced growth at 37°C and was fully temperaturesensitive at 39°C. We isolated the S7T3 locus by complementation of the synthetic lethal phenotype of the stt3-3 wbpl-2 double mutant at 30°C. The stt3-3 wbpl-2 strain YG89 was transformed with a plasmid pool (Stagljar et al., 1994) containing partially digested yeast chromosomal DNA ligated into the vector YEp352; _104 transformants were selected at 23°C and tested for growth at 30 and 37°C. Fifteen transformants were able to grow at 30°C, five of which also grew at 37°C. These five transformants contained the WBPI locus on the plasmid and were not characterized further. Plasmid DNAs from the remaining temperature-resistant clones were recovered in Escherichia coli, retested for the ability to restore growth at 30°C of strain YG89, grouped according to the restriction pattern and tested for their ability to suppress either the underglycosylation of CPY in an stt3-3 or wbpl-2 strain. One group of suppressing plasmids was able to partially restore glycosylation deficiency in the wbpl-2 single mutant and sequence analysis revealed that they contained the previously isolated SWPI (suppressor of wbpl) locus (te Heesen et al., 1993). Two other plasmids contained common insert sequences which were also able to partially restore the glycosylation deficiency in the wbpl-2 but not the stt3-3 single mutant. Sequence analysis of this wbpl suppressor showed that it encodes the 16 kDa subunit (Ost2p) of the OTase complex (Kelleher and Gilmore, 1994; R.Gilmore, personal communication). One plasmid was recovered that was able to restore growth of the wbpl-2 stt3-3 double mutant at 30°C and at the same time restored glycosylation of CPY in the stt3-3 strain (Figure 1). Deletion analysis located the complementing activity on a 2.7 kb fragment. This DNA fragment was sequenced and a search in the database showed that the complementing DNA encoded the STIf3 locus (S.Yoshida et al., unpublished results). The ST3 locus was identified previously in a screen for temperature-sensitive mutants with an increased sensitivity towards the protein kinase C inhibitor staurosporine (Yoshida et al., 1992). Subsequently, the ST3 locus was isolated by complementation 4951
R.Zufferey et al.
of the temperature-sensitive phenotype of the stt3 mutants and the corresponding sequence was deposited in the GenEMBL database (accession number D28952). To demonstrate that the isolated ST3 locus does not encode a high copy number suppressor of the stt3-3 mutation, we marked the wild-type SiT3 locus in a wbpl strain by sitedirected integration of the HIS3 gene without inactivating the S7T3 function. This strain was crossed to an stt3-3 wbpl-2 double mutant and after sporulation, segregants were analysed. stt3-3 wbpl double mutants do not grow at 30°C, whereas S7T3:HIS3 wbpl strains grow at this temperature. All 24 segregants which did grow at 30°C were His', demonstrating that the cloned DNA encoded the locus affected in the stt3 mutation. Based on the DNA sequence, the mature S7T3 protein has a calculated molecular mass of 78 kDa. The S7T3 ORF encodes a putative signal sequence of 30 amino acids (Von Heijne, 1986), a highly hydrophobic N-terminal domain which spans two-thirds of the putative mature protein and a hydrophilic C-terminal domain. Two putative N-linked glycosylation sites are found in the C-terminal part of the S773 protein; one site was located in the Nterminal domain. The primary sequence analysis strongly suggests that Stt3p is a transmembrane protein with a hydrophilic C-terminal domain (Figure 2A). In order to define on which side of the membrane this C-terminal domain is located, we fused the HJS4 polypeptide to the C-terminus of Stt3p and expressed this fusion protein in his4 cells. A cytoplasmic-oriented histidinol dehydrogenase allows these cells to grow on medium containing histidinol, whereas a lumenal orientation does not (Sengstag et al., 1990). As a control, an internal DNA fragment of the S7iT3 locus (positions 1577-1828), resulting in a deletion of a predicted, most C-terminally located transmembrane domain (positions 454-537) was also fused to the HIS4 protein. Such a deletion should result in an inversion of the location of the reporter protein. As shown in Figure 2B, the HIS4-fusion to the full-length protein does not support growth on histidinol, suggesting a lumenal orientation of the C-terminus of Stt3p, whereas the intemal deletion switches the localization and therefore supports growth on histidinol-containing medium. By immunoprecipitation, we have confirmed that the full-length fusion protein is expressed and that the HIS4 protein is glycosylated due to its lumenal orientation (Figure 2C) (Sanglard et al., 1993; te Heesen et al., 1994). These experiments establish that Stt3p is a transmembrane protein with a C-terminal, lumenally oriented, hydrophilic domain. A search for homologous proteins in the databases showed that Stt3p is highly homologous (56% identity over a region of 700 amino acid residues) to both a murine and human transmembrane protein with unknown function (G.Hong et al. unpublished results). A similar high degree of identity was found to the Caenorhabditis elegans T12A2.2 gene product (Wilson et al., 1994), an open reading frame identified within the C.elegans sequencing project (Figure 3). It is remarkable that the identity between the four proteins is not restricted to a single domain but rather is extended over the whole protein. Sequence analysis of the homologous proteins predicts a very similar structure as proposed for the yeast protein. In particular, a signal sequence of 30 amino acids is also 4952
100
A
200
300
400
600
500
700
:-.8
STT3
STT3A
WBP1
Vector
B
Tunicamycin: on
--
2OOkDa --
I
= N
4_
+ ..
_
18kDa --
CE
76kDa
Fig. 2. Analysis of the STT3 protein. (A) Hydropathy analysis of the primary Stt3p sequence according to Kyte and Doolittle (1982), using a window of 19 amino acids. The bar marks the deletion from position 453 to 538 used in the topological analysis. (B) Growth on histidinolcontaining medium of strain FC2a transformed with the plasmids YEp352 (vector), pTH47-X2, expressing His4p at the C-terminus of Wbplp (WBPI), pSTT3-HIS4, expressing His4p at the C-terminus of Stt3p (STT3) and pSTT3A-HIS4 expressing His4p at the C-terminus of the internally deletion version of Stt3p (STT3A). (C) Immunoprecipitation of metabolically labelled Stt3p-His4p fusion from FC2a cells transformed with the plasmid pSTT3-HIS4. Cells were labelled in the presence (+) or absence (-) of tunicamycin. The size of molecular weight markers is indicated. Note that this fusion protein, which does not support growth on histidinol medium, is fully glycosylated. Glycosylation of the fusion protein occurs predominantly at 'cryptic' sites of the HIS4 protein.
found in the mammalian Stt3p homologues. As expected, these signal sequences show the least conservation of amino acid residues. We tested whether the STT3 protein is essential for the vegetative growth of yeast. One STT3 locus was inactivated in a diploid yeast strain by replacing a large part of the S773 coding sequence with the selectable marker HIS3 in a homozygous his3 diploid strain. This diploid strain was sporulated and tetrad analysis was performed. Only tetrads with two viable colonies were recovered and all the viable colonies had an intact S7T3 locus (marked by the Hisphenotype) (data not shown). However, transformation of the heterozygous diploid strain with a 2 ,um-derived S7T3carrying plasmid and subsequent tetrad analysis allowed the recovery of tetrads with four viable spores (data not shown), indicating that the S1T3 locus encodes an essential product.
_D.
31stlisE"1
STT3 is essential for OTase activity
S.5, I .4~~~. n, fl II' 1. '§ ml't m. T
IIa E.
E
M
3
X$|mS -
'4.
FE
I
:-Um.mml
E§T ^Mbes N lDl :-
s~~~~~
..
'.''D
follallol-A
I..
IZ
t@E1|1
' '-S z
"I
I
'-'I.
01,11MIN ivislii
F.,
;_p
7
.S. I.
H. -,A
E
ds,
3'
E:
-
.
El .3!|I;E- 5
;
^- @E..1
Mi 3 r f
I
3 .;lI
-Io''i EZL"
I I
..i
*
I W11717
I
.II
-i
.Z;"
MEm
ME
..
m
'.Iv u
-F-1. ..,I -.1- --I-' E
@
-
li
I I.-.,
imlaso
Nk7jltTj-j
imamn I
El,:. 1.
04 a
'I
; MIN i 0
I.
:E..: ..
6w
III
;z :: ;. .', ;" 3 ;.z --,
IIN loll ;
e.
..
....
1.
P
3:3
H!.
Fig. 3. Sequence identity between the primary amino acid sequence of Stt3p (yeast), the putative Celegans T12A2.2 gene product (accession number U13019, CELT12A2_7) (Celegans) and the murine transmembrane protein (accession number L43260) (Mtus musculus). Identical amino acids in at least two of the three sequences are highlighted. The sequence of the human transmembrane protein (accession number L38961) (98% identical to the murine sequence) was omitted for clarity. The sequences were aligned using the DNASTAR software (Clustal method).
The stt3-3 mutation affects glycosylation of secretory proteins To address the function of the S713 protein, we first analysed the phenotype of the stt3-3 mutation. We had based our screening strategy on the concept that mutations which affect the pathway of N-linked glycosylation in the ER have a cumulative effect on the glycosylation of secretory proteins when combined with the wbpl mutation. Therefore, we examined the expression of CPY in the st3 wbpl double mutant (Figure 4). As reported earlier, the wbpl-2 mutation leads to the underglycosylation of CPY. CPY molecules lacking one, two or three N-linked oligosaccharide chains were precipitated from wibpl extracts. These hypoglycosylated species are visualized as distinct bands migrating faster in SDS-PAGE than mature CPY, the difference in the molecular mass being -2 kDa (Figure 4, lane 1). A similar, but less severe underglycosylation of CPY was observed in the stt3-3 single mutant: besides mature CPY, molecules migrating with the same mobility as the ones lacking one or two oligosaccharide chains were detected in this strain (Figure 4, lane 2). The stt3-3
wbpl-2
-3 Wt
wbpl-2 mCPY
4:.4IP
-1
4W
3
-
-2 -3 -4
4
Fig. 4. The stt3-3 mutation affects CPY processing. The different strains derived from one tetrad were labelled for I h at 23°C with l35SJmethionine and [35SIcysteine. The relevant genotype of the strains is indicated above the lanes. The positions of mature CPY (mCPY) and the different glycoforms lacking up to all four N-linked oligosaccharides (-I to -4) are given. Strains: YGI50 (lane 1; wbpl-2), YG149 (lane 2. stt3-3), YG148 (lane 3; wt), YG147 (lane 4,
otlbpl-2 .stt3-3).
4953
R.Zufferey et aL
wbpl-2 double mutant expressed heavily underglycosylated CPY molecules (Figure 4, lane 4); the major bands represent glycoforms lacking three or all four N-linked oligosaccharides. A similar underglycosylation due to the stt3-3 mutation was also found in the analysis of the two glycoproteins Wbp 1 p and Ost 1 p (data not shown; see also Figure 9). The observed underglycosylation of CPY in stt3-3 mutant cells is very similar to that observed in strains with a deficiency in the assembly pathway of the lipidlinked oligosaccharide [algS, alg6 or alg8 mutant strains (te Heesen et al., 1994; Stagljar et al., 1994)] or in strains with a defective OTase complex [wbpl (te Heesen et al., 1992) or osti mutant strains (Silberstein et al., 1995)]. In addition, as found for different alg mutations, stt3-3 has a cumulative effect on glycosylation in combination with wbpl. We conclude that the stt3-3 mutation affects the pathway of N-linked glycosylation in the ER and, in combination with the wbpl mutation results in a severe underglycosylation of proteins. This underglycosylation is most likely the cause for the synthetic lethal phenotype at 300C.
The stt3-3 mutation affects the substrate specificity of the OTase in vitro and in vivo The phenotype of underglycosylation of CPY and other secretory proteins was observed previously either in mutant strains with reduced OTase activity (te Heesen et al., 1992, 1993; Silberstein et al., 1995) or in strains with a deficiency in the biosynthesis of the lipid-linked oligosaccharide (alg mutants) (Stagljar et al., 1994). alg mutants have the additional phenotype in that they accumulate the intermediate in lipid-linked oligosaccharide biosynthesis used as a substrate in the affected reaction (Huffaker and Robbins, 1982, 1983; Runge et al., 1984; Runge and Robbins, 1986; Verostek et al., 1993a,b). To identify the cause for the underglycosylation of secretory proteins in the stt3-3 strain, we have analysed the lipid-linked oligosaccharides accumulating in such cells (Figure 5). We observed the presence of full-length lipid-linked oligosaccharides in wild-type, wbpl-l and stt3-3 cells. In particular, complete lipid-linked Glc3Man9GlcNAc2 was extracted from stt3-3 cells. The pattern of lipid-linked oligosaccharides in stt3-3 cells appeared similar to that of wbpl cells. We conclude that in the stt3-3 strain underglycosylation of secretory proteins is not due to a defect in a specific step of the lipid-linked oligosaccharide biosynthesis as it is observed in different alg mutants. Next we tested whether the OTase activity was affected in stt3-3 strains and compared the result with the activity obtained from wild-type cells. To measure the in vitro OTase activity, we used two different lipid-linked oligosaccharide substrates: the full-length lipid-linked Glc3Man9GlcNAc2 as well as lipid-linked chitobiose. It has been shown previously that the OTase complex transfers chitobiose with high efficiency to peptides in vitro (Harford and Waechter, 1979; Sharma et al., 1981). Compared with wild-type activity, extracts prepared from stt3-3 cells transferred the full-length oligosaccharide with a slightly reduced activity to the acceptor peptide but very little transfer activity could be detected when lipid-linked chitobiose substrate was used (Table II). In contrast, wbpl extracts had the same reduced activity irrespective of the
4954
1000
S~ ~ M5N2
[dpml 1500
M9N2
MSN2
[dpm]I]
G3M9N2
2000 1500
-
lI A
~~G3M9N2
R
1000
-
500 0
0
Marker
A
[dpml 3000
-
2000
-
1000
-
0
B
10 20 30 40 50 60 70
[mini
wild-type
[dpmn]
G3M9N2
3000
G3M9N2
2000-
. I
0 I
I
I
~ ~ ~ ~ 10 0
I
0 io 20 30
C
-
10 20 30 40 50 60 70
wbpl-l
40io I
I
6Io
io I
I
J
i0 D
10 20 30 40 50 60
7'0[min]
stt3-3
Fig. 5. Analysis of lipid-linked oligosaccharides in wild-type and mutant strains. Lipid-linked oligosaccharides were labelled in vivo, extracted and analysed by HPLC. (A) Marker: a mixture of oligosaccharides extracted from alg3-1 (M5N2 and G3M9N2) and Aalg5 (M9N2) strains. (B) Wild-type strain SS328. (C) wbpl-l strain MA7-B. (D) stt3-3 strain YG176. The position of Glc3Man9GlcNAc2 (G3M9N2), Man9GlcNAc2 (M9N2), Man5GlcNAc2 (M5N2) and mannose (M) are indicated.
Table II. In vitro activity of OTase in wild-type, stt3-3 and wbpl-l cell extracts Strain
Glycosyl transfer to peptide (%) from Dol-PP-GlcNAc2 Dol-PP-GlcNAc2Man9Glc3
SS328 (wild-type) YG176 (stt3-3) MA7-B (wbpl-I)
100 8 21
100 72 24
The strains indicated were grown to mid-log phase in YPD complete medium, extracts prepared and OTase activity was determined using the indicated lipid-linked oligosaccharides as a substrate. The value obtained from wild-type extracts was taken as 100%. Two independent extracts of each strain were prepared and each extract was assayed several times. Corresponding mean values are given.
substrate used. The S7IT3 protein which was not detected as a component of the purified OTase complex (Kelleher and Gilmore, 1994; Knauer and Lehle, 1994) seems to be required for the transfer of chitobiose by the OTase complex in vitro. We conclude that the stt3-3 mutation results in a sensitivity of the OTase towards suboptimal substrate as measured in vitro. To confirm this conclusion by in vivo experiments, we took advantage of the leaky phenotype of the alg3-1 mutation. alg3-1 mutant strains have no detectable growth phenotype and are assumed to be defective in the first lumenal mannosyl transferase which transfers mannose from Dol-P-Man to the lipid-linked Man5GlcNAc2 (Huffaker and Robbins, 1983; Herscovics and Orlean, 1993). However, a detailed analysis of protein-bound oligosaccharides in alg3-1 strains showed that besides the suboptimal Man5GlcNAc2 oligosaccharide, complete Glc3Man9GlcNAc2 is also transferred to protein. It was postulated that alg3-1 is a leaky alg3 allele (Verostek
STT3 is essential for OTase activity
stt3-3 Endo H:
-
+
-
hou
3tt3-3 alg3-1 aIg3-1
wt +
-
%
after shift to glucose 3 6 9 12
+
mCPY
-1 = AWORM -2= -3=
mCPY -1= -2 -3 -4
-4
_... 1
2
3
4
5
6
7
8
Fig. 6. The stt3-3 mutation affects substrate specificity of OTase. The different strains derived from one tetrad were labelled for 1 h at 23°C with [35S]methionine and [35Slcysteine. The relevant genotype of the strains is indicated above the lanes. The positions of mature CPY (mCPY) and the different glycoforms lacking up to all four N-linked oligosaccharides (-I to -4) are given. Samples were treated with endo H (+) or left untreated (-). Note that the protein-bound oligosaccharides in the alg3-1 stt3-3 strain are fully endo H-sensitive in contrast to alg3-1 alone. Strains: YG167 (lanes 1, 2; stt3-3), YG168 (lanes 3, 4; wt), YG169 (lanes 5, 6; alg3-1), YGI70 (lanes 7, 8; stt3-3 alg3-1).
et al., 1993a,b). The transfer of both oligosaccharides in alg3-1 is confirmed here for CPY. As shown in Figure 6 (lanes 5 and 6), CPY-bound oligosaccharides are partially sensitive towards endo H. Protein-bound Man5GlcNAc2 are endo H-resistant oligosaccharides, whereas transferred Glc3Man9GlcNAc2 lead to endo H-sensitive carbohydrate modifications. An alg3-1 stt3-3 double mutant should therefore allow the influence of the stt3-3 mutation on the utilization of Man5GlcNAc2 and Glc3Man9GlcNAc2 to be assayed in vivo. We generated alg3-1 stt3-3 double mutants by conventional genetic techniques and observed that these double mutants have a severe growth defect at 30°C and are temperature-sensitive at 37°C, whereas the two single mutants are temperature-resistant: the alg3-1 and the stt3-3 mutation synthetically interact. Inspection of the glycosylation capacity of alg3-1 stt3-3 cells by CPY immunoprecipitation showed that the combination of the two mutations results in a severe underglycosylation (Figure 6, lane 7). Interestingly, oligosaccharides transferred to CPY in the stt3-3 alg3-1 double mutant are fully sensitive to cleavage by endoglucosidase H (Figure 6, lane 8), whereas the oligosaccharides transferred to CPY in the alg3-1 single mutant are, as shown above, mostly endo H-resistant. The oligosaccharides in the stt3-3 strain are all endo-H sensitive (Figure 6, lane 2). Due to the stt3-3 mutation, only complete Glc3Man9GlcNAc2 was transferred to protein, while the suboptimal substrate Man5GlcNAc2 was not transferred. The in vivo observations confirm our in vitro results and suggest that the stt3-3 mutation alters the substrate recognition of the OTase complex towards suboptimal lipid-linked oligosaccharide substrates. This conclusion was further supported by the finding that a combination of the stt3-3 mutation with deletions in either the ALG3 (accumulation of Man5GlcNAc2; M.Aebi et al., manuscript in preparation), ALG5 (accumulation of Man9GlcNAc2; Runge et al., 1984) or ALG8 (accumulation of Glc2Man9GlcNAc2; Runge and Robbins, 1986) results in the inviability under all conditions tested (data not shown).
Wbpl p -21 = ,X
-2-
Fig. 7. Depletion of Stt3p alters CPY and Wbplp glycosylation. Strain YG157 was shifted from galactose to glucose medium, resulting in the repression of Stt3p synthesis. Samples were taken at the time indicated, labelled for 1 h, lysed and used for CPY- (top) or Wbplp(bottom) specific immunoprecipitation. The positions of the mature proteins and the glycoforms lacking N-linked oligosaccharides are given. CPY and Wbplp contain four and two N-linked oligosaccharides, respectively. * marks the position of a protein nonspecifically precipitated by the anti-Wbplp serum.
Loss of OTase activity in Stt3p-depleted cells The essential function of the ST13 protein allowed us to study the effect of Stt3p depletion on the process of Nlinked glycosylation in vivo. We constructed a strain with a deletion of the S7T3 locus but containing a hybrid gene in which the S7T3 coding sequence was placed downstream of the GAL] promoter (Johnston and Davis, 1984). This strain is able to grow on galactose- but not on glucose-containing medium. We tested the effect of the Stt3p-depletion on the processing of the two glycoproteins CPY and Wbplp. Cells were grown in galactose medium, washed and then cultivated in glucose-containing medium. Aliquots of the culture were withdrawn, the cells were pulse-labelled, fragmented and immunoprecipitation of the two proteins CPY and Wbplp performed (Figure 7). It is evident that the glycosylation of both proteins is affected by the depletion of Stt3p. CPY lacking up to all four Nlinked oligosaccharides and Wbplp molecules without the two N-linked sugars are detected after a 12 h depletion. We note that depletion of Wbplp (te Heesen et al., 1992), Swplp (te Heesen et al., 1993) and Stt3p all result in a similar underglycosylation of CPY. This underglycosylation of CPY was not due to a defect in the assembly of the lipid-linked oligosaccharide substrate in Stt3p-depleted cells. In contrast, we observed predominantly lipid-linked Glc3Man9GlcNAc2 (Figure 8), as opposed to wild-type cells, no intermediates in the synthesis of the lipid-linked oligosaccharide were isolated from Stt3p-depleted cells. The same phenotype was observed upon depletion of Wbplp, a component of the purified OTase complex. Therefore, we measured OTase activity in extracts derived from these depleted cells. As substrates, both full-length lipid-linked oligosaccharide as well as dolichyl-PP-chitobiose were used. For both substrates, a continuous decrease of OTase activity was observed upon depletion of the ST3 protein (Table III). The loss of activity was not due to a general defect in the function of the ER, because other ER-resident enzymatic
4955
R.Zufferey et al.
[dpm]
M5N2
M
G3MqN2
500.
M5N2
M
[dpmj
M wt Astt3 stt3-3
G3M9N2
3000-
2501
.
-
1000-
1oo0o
A
o
wild-type
o10 do jo do
4o so
60 70o galactose
B
wild-type
so do9otminl glucose
[dpml
2500
2500
2000
201
iooo-1
500
500~
1O
C
do
30 40 50 6so 70
GAL-WBP1
0
galactose
[dpml
D
10
20
30
GAL-WBP1
60 ;0lmlnl
40
glucose
[dpml
1000-I 750
-
_-
4wm.w
1
15001 1000-
Osti p
-Wbpl p - Ost3p
__m~ [dpml
-
2
-
SwpIp
3
Fig. 9. Affinity purification of OTase complex from different cells. The OTase complex from strain YG157, grown on galactosecontaining medium (wt, lane 1), strain YG157, grown for 15 h in glucose-containing medium (Astt3, lane 2) and strain YG176 (stt3-3, lane 3) was purified and analysed by SDS-PAGE. In each lane, the amount equivalent to 0.028 A280 was applied and visualized with silver stain. The position of Ostlp (60/62 kDa), Wbplp, Ost3p (34 kDa) and Swplp are given. Due to the underglycosylation in Astt3 and stt3-3 strains, hypoglycosylated forms of Ostlp and Wbplp are detected. Note the absence of OstIp and Ost3p in the complex derived from Stt3p-depleted cells.
500 250
O
E
GAL-STT3
galactose
F
1o 2o 3o 40 5o o iO lminj GAL-S1T3
glucose
Fig. 8. Depletion of Stt3p and Wbplp results in the accumulation of lipid-linked Glc3Man9GlcNAc2 oligosaccharides. The three strains SS328 (wild-type, A and B), 45-C3 (depletion of Wbplp, C and D) and YG157 (depletion of Stt3p, E and F) were grown in galactosecontaining medium (A, C, E) or shifted from galactose- to glucosecontaining medium (B, D, F). 12 h after the shift, cells were labelled using [3H]mannose, lipid-linked oligosaccharides were extracted and analysed. The positions of Glc3Man9GlcNAc2 (G3M9N2), Man5GlcNAc2 (M5N2) and mannose (M) are indicated. Table III. In vitro activity of OTase in Stt3p-depleted cell extracts Time after shift to glucose (h) 0 3 6 9 12 15 22
Glycosyl transfer to peptide (%) from Dol-PP-GIcNAc2
Dol-PP-GlcNAc2Man9Glc3
100 75 42 23 12 3 0
100 81 50 25
20 9 4
Strain YG157 was grown in complete medium containing galactose to mid-log phase, washed and used to inoculate minimal medium with 2% glucose as carbon source. Cultures were grown at 30°C for the time indicated, extracts were prepared and OTase activity was measured. The value obtained at the shift from galactose to glucose was taken as 100%.
activities such as Dol-P-Gluc Synthase, Dol-P-Man Synthase or Dol-PP-GlcNAc-forming enzyme were not affected by the depletion (data not shown). In addition, translocation and the secretion process are not affected by the Stt3p depletion as visualized by the normal sorting of underglycosylated CPY to the vacuole (Figure 7). However, the loss of OTase activity was due to a disappearance
4956
of OTase complex in depleted cells (Figure 9). Affinity purification of the OTase showed that in undepleted cells a complex which consists of four proteins [OstIp, Wbplp, Ost3p (34 kDa) and Swplp] can be purified (Knauer and Lehle, 1994) (Figure 9). Upon depletion of Stt3p, Ostlp and Ost3p disappear from the purified complex, whereas Wbplp (which becomes underglycosylated) and Swplp are still present. This decrease of the OTase complex coincides with the loss of OTase activity as measured in vivo and in vitro. Although not a component of the purified complex, the ST3 protein is required for the assembly or the stability of this complex in vivo.
Discussion In this report we describe the isolation of novel mutants deficient in the assembly of the lipid-linked oligosaccharide. Our screening procedure was based on the observation that the OTase mutation wbpl in combination with alg mutations results in a synthetic lethal phenotype (Stagljar et al., 1994). Indeed, our screening yielded new alleles of defined ALG loci but in addition, novel mutants deficient in the assembly of the lipid-linked oligosaccharide were also isolated (ALG9 and ALGIO). This suggests that our screening method favours the isolation of mutants defective in the assembly of the lipid-linked oligosaccharide and will allow a detailed characterization of this complex pathway. In contrast, in a complementary screen for novel mutations synthetic lethal with a deletion in the ALG5 locus, we identified predominantly mutations affecting OTase components (G.Reiss and M.Aebi, manuscript in preparation). Also in this screen, we identified the Si7T3 locus as a major complementation group. This is not unexpected since, as shown above, the stt3 mutation is synthetic lethal with different alg mutations. The S7T3 locus was first identified by mutations resulting in a sensitivity towards the protein kinase C (PKCI) inhibitor staurosporine (Yoshida et al., 1992). PKCJ is known to be the central element in the regulation of cell integrity via regulation of cell wall biosynthesis in
S1T3 is essential for OTase activity S.cerevisiae (Levin and Bartlett-Heubusch, 1992; Paravicini et al., 1992; Levin and Errede, 1995). A deletion of PKCI results in rapid cell lysis and an altered composition of the cell wall (Levin and BartlettHeubusch, 1992; Paravicini et al., 1992; Roemer et al., 1994). Mutations which alter cell wall biosynthesis (either by affecting f-glucan synthesis or 0-linked protein glycosylation) have a synthetic lethal phenotype in combination with the pkcl mutation (Roemer et al., 1994). We now show that an stt3 mutation affects the N-linked glycosylation process. It is likely that a deficiency in N-linked glycosylation, required for cell wall biogenesis (Klis, 1994) also results in a synthetic effect in combination with the inactivation of PKCJ. Indeed, we could show that the stt3-3 mutation results in a synthetic phenotype in combination with the Apkcl mutation. It is therefore not surprising that S7T3 was detected in a screen for staurosporine-sensitive mutants (Yoshida et al., 1992). The SiT3 protein is a highly conserved transmembrane protein with a hydrophilic, C-terminal domain oriented towards the lumen of the ER. The experiments presented in this report led us to propose that Stt3p is required for the process of N-linked glycosylation in the ER and essential for OTase activity in vivo. This conclusion is based on the following experimental evidence: (i) a mutation in the S7T3 locus (stt3-3) can affect the substrate specificity of the OTase complex both in vitro and in vivo; (ii) depletion of Stt3p results in a loss of OTase activity and the accumulation of the lipid-linked substrate of this enzyme. Loss of activity was also observed upon depletion of known OTase components, namely Wbplp (te Heesen et al., 1992) and Swplp (te Heesen et al., 1993); (iii) depletion of Stt3p specifically affects N-linked glycosylation but not other functions in the ER, as for example translocation or transport to the Golgi compartment. Though our in vivo data strongly suggest that Stt3p is required for OTase activity, the purification of this enzyme did not reveal a protein of 80 kDa molecular weight, at least not in stoichiometric amounts (Kelleher and Gilmore, 1994; Knauer and Lehle, 1994). In the following, we discuss two different hypotheses to explain this discrepancy. The STT3 protein is a component of the OTase activity but only required in substochiometric amounts for OTase activity in vitro (substrate recognition hypothesis) Since we showed that a mutation in the S7T3 protein can result in a high sensitivity of the OTase complex towards suboptimal lipid-linked oligosaccharide substrate in vivo and in vitro, this implies that Stt3p is involved in the recognition and/or binding of the lipid-linked oligosaccharide. The mutant stt3 protein no longer recognizes a suboptimal substrate, resulting in the severe synthetic lethality of the stt3-3 mutation with different alg mutations and the greatly reduced OTase activity from stt3-3 cells using the chitobiose substrate. However, this hypothesis does not explain why a depletion of Stt3p results in the assembly of incomplete OTase complex.
The STT3 protein is required for the assembly of the OTase complex in vivo (OTase assembly
hypothesis) Depletion of this assembly factor would result in an incomplete OTase complex. Phenotypically, the depletion of such an assembly factor is indistinguishable from the depletion of a structural component of the OTase complex. We observed that depletion of Stt3p led to a loss of OTase
activity. Within the framework of the OTase assembly hypothesis, the stt3-3 mutation would result in the assembly of a suboptimal OTase complex. This incorrectly assembled complex would no longer be able to accept
incomplete lipid-linked oligosaccharides as substrates, resulting in the synthetic phenotype of stt3-3 alg double mutants and the reduced affinity of the transferase in extracts of stt3-3 cells towards lipid-linked chitobiose. Additional in vivo functions of Stt3p are not directly assayed in our experiments. It is for example possible that Stt3p mediates the coupling of the OTase complex to the translocation machinery. Irrespective of the function of the SJT3 protein, the degree of conservation of this protein between yeast and higher eukaryotes is surprising. Structural components (Wbplp and Ostlp) of the OTase complex have a much lower degree of identity between the yeast and the mammalian homologue (between 25 and 30%; Silberstein et al., 1992, 1995). The high degree of conservation of Stt3p might be indicative for multiple functions of this protein in the pathway of N-linked glycosylation. Since Stt3p is highly conserved not only in the lumenal, hydrophilic domain but also in the hydrophobic membrane-spanning domain, an essential function for this part of the protein is suggested. In view of the two hypothetical roles of the STT3 protein discussed above, this hydrophobic domain may be involved in the binding of the lipid carrier dolichol (substrate recognition hypothesis) or in the assembly of the OTase complex consisting of transmembrane proteins with one or more hydrophobic membrane-spanning regions (OTase assembly hypothesis). Consistent with the essential function of Stt3p for OTase activity, mammalian SJT3 protein co-purified with the translocation apparatus (E.Hartmann and T.Rapoport, personal communication). However, a more detailed genetic and biochemical analysis of Stt3p will be required to clarify its role in the process of N-linked glycosylation.
Materials and methods Yeast strains and media Yeast strains used are listed in Table IV. Standard yeast media (Guthrie and Fink, 1991) were used.
Isolation of mutants synthetically lethal with wbpl alleles The strains YG135, YG136, YG137 and YG138 were grown at 30°C in liquid minimal medium lacking uracil to an OD546 of 3.5-4. These strains harbour the pCH 1122WBPI plasmid which contains the NruISpeI WBPI locus cloned into the NruI site of CHI 122 (Kranz and Holm, 1990). I X 103 cells were plated onto YPD agar plates and irradiated with UV light (7 mJ/cm2) (UV Stratalinker 2400, Stratagene). A survival rate of 15-25% was obtained. The plates were incubated in the dark for 5-6 days at 23°C (YG136 and 138) or 30°C (YG135 and 137). Nonsectoring colonies were streaked onto YPD plates, retested for the nonsectoring phenotype and finally checked for growth on SFOA-containing minimal medium. For complementation analysis of the mutant strains, individual strains of the two mating types were mated on YPD plates and
4957
R.Zufferey et aL Table IV. Yeast strains used in this study Strain
Genotype
Reference
SS328 SS330 FC2a MA7-B 45-C3 YG89 YG91 YG135 YG136 YG137 YG138 YG147 YG 148 YG149 YG150 YG156 YG157 YG167 YG168 YG169 YGI70 YG176 YG294 YG323 YG324
MATa ade2-101 ura3-52 his3A200 lys2-801 MATa ade2-101 ura3-52 his3A200 tyrl MATa ura3-52 leu2-3 leu2-1 12 trpl-I his4-401 HOLI-l MATa ade2-101 ura3-52 his3A200 lys2-801 wbpl-l MATa ade2-101 tyrl his3A200 Awbpl::HIS3 URA3:GALl-WBPl MATx ade2-101 ade3 ura3-52 his3A200 leu2 lys2-801 wbpl-2 stt3-3 MATa ade2-101 ura3-52his3A200 Aalg5::HIS3 MATa ade2-101 ade3 ura3-52 his3A200 lys2-801 wbpl-2 [pCHI122WBPI MATa ade2-101 ade3 ura3-52 his3A200 leu2 wbpl-l [pCHI 122WBPl] MATa ade2-101 ade3 ura3-52 his3A200 leu2 wbp 1 -2 [pCH 1122WBP I MATa ade2-101 ade3 ura3-52 his3A200 lys2-801 wbpl-1 [pCHI 122WBPI} MATa ade2-101 ade3 ura3-52 his3A200 lys2-801 stt3-3 wbpl-2 MATa ade2- 101 ura3-52 his3A200 leu2 MATa ade2-101 ade3 ura3-52 his3A200 lys2-801 leu2 tyrl stt3-3 MATa ade2-101 ura3 his3A200 tyrl wbp 1-2 MAToa/a Astt3::HIS3/STT3 ura3-52/ura3-52 lys2-801/LYS2 tyrl/TYRI his3-A200/his3A200 MATa ade2-101 lys2-801 URA3:GALI-STT3 Astt3::HIS3 his3A200 MATa ade2-101 ade3 ura3-52 leu2 tyrl lys 2-801 stt3-3 MATa ade2-101 ura3-52 his3 tyrl lys 2-801 MATa ade2-101 ura3-52 leu2 alg3-1 MATa ade2-101 ade3 ura3-52 his3 alg3-1 stt3-3 MATa ade2-101 ade3 ura3-52 his3A200 leu2 tyrl stt3-3 MATa ade2-101 ura3-52 his3A200 tyrl stt3-3 wbpl-2 MATa ade2-101 ura3-52 his3A200 lys2-801 wbpl-l STT3:HIS3 MATa ade2-101 ura3-52 his3A200 lys2-801 wbpl-l STT3:HIS3
Vijayraghavan et al. (1989) Vijayraghavan et al. (1989) Sengstag et al. (1990) te Heesen et al. (1992) te Heesen et al. (1992) This study te Heesen et al. (1992) This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study This study
replica-plated on 5FOA minimal medium selective for the corresponding diploid strains.
Isolation of the STT3 locus A genomic library (Stagljar et al., 1994), containing partially digested yeast chromosomal DNA ligated into the vector YEp352 (Hill et al., 1986), was transformed into the strain YG89 and transformants were selected on minimal medium lacking uracil at 23°C. Approximately 104 transformants were tested for growth at 30 and 37°C by replica-plating. Plasmid DNA from positive colonies (growing at 30°C but not at 37°C) was isolated by extracting total yeast DNA which was used to transform E.coli strain DHcS to ampicillin resistance. Recovered plasmids were tested for their ability to support growth of strain YG89 at 30°C and to restore normal glycosylation in strain YG 149. The S7T3-carrying plasmid was called pSTT3.
Inactivation of the STT3 locus The 1.2 kb BstEII fragment was excised from the STT3 coding region and replaced by the 1.8 kb BamHI HIS3 cassette (Struhl and Davis, 1981). The resulting plasmid was digested with BamHI and XbaI to direct integration at the S7T3 locus. DNA was used to transform the diploid strain SS328XSS330 to His+. The correct integration was confirmed by Southern blot analysis. Tetrads from two independent transformants were dissected and analysed.
Construction of the Gal-STT3 strain The GALI-JO promoter containing fragment of pBM150 (Johnston and Davis, 1984) was cloned into the EcoRI and BamHI-cleaved vector YIp5 (Rothstein, 1991), yielding the plasmid YIpGal. The 1.6 kb NcoI fragment containing the 5' part of the S7T3 ORF was cloned into the BamHI site of YIpGal and the 3' part was later introduced as a KpnISphI fragment, resulting in YIpGal(STT3). The sequence between the BamHI site of the GAL-Promoter and the ATG start codon of S7T3 appears as follows: 5'-GGATCCATGG-3'. YIpGal(STT3) was linearized with StuI to direct integration at the ura3-52 locus and used to transform the diploid strain YGl 56 (relevant genotype stt3::HIS3/S7T3). One Ura+ transformant was used for sporulation. Tetrad analysis yielded strain YG157.
Integration of HIS3 at the STT3 locus The 3.1 kb ClaI-XbaI fragment of the S773 locus was cloned into the vector pRS303 and subsequently the 3' terminal sequences downstream of the PmlI site (position 1828) were deleted from the construct, resulting in the plasmid pRS303(STT3A). This plasmid was used to integrate the HIS3 locus at the STT3 site in the strain MA7-B.
4958
Determination of the membrane topology of Stt3p For the in-frame fusion of HIS4 to S7T3, the last 867 nucleotides of the S7T3 ORF were amplified by PCR using the two primers 5'GTGACCGCAAATACG-3' and 5'-GTGTGGCCCTCGAGGACTCTCAAG-3'. In the second primer, the stop codon of the S7T3 ORF was replaced by an XhoI restriction site. The WBPI coding sequences were removed from the plasmid pTH47-X2 by cutting with SacI-XhoI and replaced by STT3 sequences (SacI-BspEI from pSTT3 and BspEI-XhoI from the PCR amplification product). The resulting plasmid was named pSTT3-HIS4. Functionality of the fusion protein was determined by transformation of the diploid strain YG 156 (relevant genotype stt3::HIS3/ S7T3) and subsequent tetrad dissection. Complete tetrads containing two His+ segregants were obtained. An in-frame deletion within the S7T3 coding sequence was introduced into plasmid pSTT3-HIS4 by digestion with BspEI, filling in with Klenow polymerase, partial digestion with PmlI and religation (pSTT3AHIS4). This deletion removes 84 amino acids of the Stt3p between position 453 and 538. HIS4-fusion proteins were expressed in strain FC2a and tested for their ability to confer growth on histidinol medium (Sengstag et al., 1990) after 4 days incubation at 23°C. Expression of inactive fusion protein was tested by immunoprecipitation of the fusion protein (Sanglard et al., 1993; te Heesen et al., 1994).
Extraction of lipid-linked oligosaccharides Cells were grown at 30°C in YPD + 20 mg/l adenine sulfate (ade) or YPGal + ade to a density of 0.5-4x 107/ml. For depletion experiments, the cells were grown in YPGal + ade, washed twice with YPD + ade, suspended in YPD + ade and grown at 30°C for 12 h. Control cells were diluted in fresh YPGal + ade. 109 cells were pelleted, washed once with YPO.ID (YPD containing 0.1% glucose). The cells were suspended in 200 ,l YPO. 1 D containing 250 ,Ci D-[2-3H]-mannose (555 GBq/mmol, Amersham) and incubated 12 min at 25°C. The extraction of lipid-linked oligosaccharides was performed according to Lehle (1980). Labelling was stopped by the slow addition of 5 ml chloroform/ methanol 3:2 (CM) during I min under continuous vortexing. After 5 min at 23°C, extracts were centrifuged for 5 min at 1400 g and the supernatant was discarded. Acid-washed glass beads (0.5 ml) were added to the cell pellet, vortexed for 1 min and the pellet was washed twice with 4 ml CM. All washing and extraction steps were performed by vortexing for 1 min. The cell pellet was washed three times with UP(+) [chloroform:methanol:water (CMW), 3:48:47, v/v/v ) containing 4 mM MgC12] and once with UP(-) (CMW, 3:48:47, v/v/v). Lipid-linked oligosaccharides were then extracted with 3X4 ml CMW (10:10:3, v/v/v), the extraction solutions were combined and dried under nitrogen at 37°C. For hydrolysis, the dried lipid-linked oligosaccharides were
STT3 is essential for OTase activity suspended in 35 .l1 I-propanol, followed by the addition of I ml 20 mM HCI and incubated at 100°C for 45 min. To remove the lipids, the hydrolysate was extracted once with 5 ml CM, dried under nitrogen, dissolved in 200-400 pg H20 and filtered through a 0.45 p.m filter (Millipore UFC30HVOO). The oligosaccharides were stored frozen at -20°C. For analysis of the oligosaccharides, a Supelco LC-NH2 column (250X4.6 mm) (Cacan et al., 1993) including an LC-NH2 guard column was first equilibrated by running a gradient of acetonitrile/water (70:30, v/v) to acetonitrile/water (50:50, v/v) over 75 min, 5 min at 50:50 and returning to 70:30 over 5 min using a Merck/Hitachi L-2600A pump. Before injection of the sample, the column was washed with acetonitrile/ water (70:30) for 20 min. The flow rate was 1 ml/min. Oligosaccharide sample (40 g. in water) was injected using an autosampling device (Merck/Hitachi AS-2000). The eluate from the column was mixed continuously with scintillation fluid (FLO-Scint A, Packard) in a ratio of 1:1.5 (eluate:scintillation mix, v/v). Radioactivity was monitored using a flow monitor (FLO-ONE A-525, Packard) equipped with a 0.5 ml flow cell. The counting efficiency for 3H in the flow cell was 22%.
Immunological methods Immunoprecipitation of CPY was performed as described (Rothblatt and Schekman, 1988; te Heesen et al., 1992). For digestion with endo H, 25 p1 of sample buffer were added to the washed immunoprecipitates and heated to 95°C for 5 min in a thermoshaker. Extract (10 p1) was mixed with 10 p1 of endo H buffer (50 mM Na3-citrate, pH 5.5/10 mM NaN3), 1 mU endo H (Boehringer Mannheim) and incubated overnight at 370C.
Determination of OTase activity and purification of OTase
complex OTase activity was measured as described (Sharma et al., 1981) using DolPP-GlcNAc2 or DoIPPGIcNAc2Man9GGc3 as glycosyl donor and the hexapeptide YNLTSV as N-glycosylation acceptor. The incubation volume was 0.05 ml and consisted of 10 mM Tris-HCI, pH 7.4, 1.1% Triton X-100, 10 mM MnCI2, 9.5% DMSO, DoIPP-GlcNAc2 (4000 c.p.m.) or DolPPGlcNAc2Man9Glc3 (22 000 c.p.m.), 4 mM hexapeptide and 0.12 mg protein. In the case of DoIPP-GlcNAc2 the reaction was incubated for 30 min and the reaction analysed as described (Knauer and Lehle, 1994). Both assays were linear with respect to time and protein concentration. With DolPPGlcNAc2MangGIc3 as glycosyl donor the reaction was carried out for 10 min and stopped with I ml methanol. The reaction mixture was centrifuged and the precipitate was washed twice with 0.5 ml 50% methanol. The combined supernatants were dried and dissolved in 0.1% TFA and the formed glycopeptide was identified and quantified by HPLC on a Vydac RP1 8 column using the following gradient: 5 min 0.1 % TFA followed by a linear increase from 0.1 % TFA to 60% acetonitrile/0.1% TFA within 55 min; flow rate 1 ml/min. To isolate the OTase complex, membranes and solubilized enzyme fraction were prepared as described previously (Knauer and Lehle, 1994). The complex was isolated therefrom by immunoaffinity chromatography on an anti-Wbplp-Sepharose column (Knauer and Lehle, 1994).
Acknowledgements We thank Reid Gilmore for antibodies, communication of results before
publication and bringing the murine S7T3 homologue to our attention, Neta Dean, Enno Hartmann and Tom Rapoport for communicating results before publication, and Karl Graham for the correction of the manuscript. S.t.H. and M.A. thank Corinne Villers and Rene Cacan for the help with the HPLC system. This work was supported by the Swiss National Science Foundation (grant 3100-040350.94/1 to M.A. and a MD/ PhD grant 31-37141.93 to R.Z), the Deutsche Forschungsgemeinschaft (SFB43) and Fonds der Chemie to L.L.
References Abeijon,C. and Hirschberg,C.B. (1992) Trends. Biochem. Sci., 17, 32-36. Bender,A. and Pringle,J.R. (1991) Mol. Cell. Biol., 11, 1295-1305. Boeke,J.D., LaCroute,F. and Fink,G.R. (1991) Mol. Gen. Genet., 197, 345-346.
Breuer,W. and Bause,E. (1995) Eur J. Biochem., 228, 689-696. Cacan,R., Labiau,O., Mir,A.M. and Verbert,A. (1993) Eur J. Biochem., 215, 873-881. Costigan,C., Gehrung,S. and Snyder,M. (1992) Mol. Cell. Biol., 12, 1162-1178.
Cummings,R.D. (1992) Synthesis of asparagine-linked oligosaccharides:
pathways, genetics, and metabolic regulation. In Allen,H.J. and Kisalius,E.C. (eds), Glycoconjugates, Composition, Structure, and Function. Marcel Dekker, New York, pp. 333-360. Gavel,Y and Von Heijne,G. (1990) Protein Engineering, 3, 433-442. Guthrie,C. and Fink,G.R. (1991) Guide to Yeast Genetics and Molecular Biology. Academic Press, San Diego. Harford,J.B. and Waechter,C.J. (1979) Arch. Biochem. Biophys., 197, 424-435. Hasilik,A. and Tanner,W. (1978) Eur J. Biochem., 91, 567-575. Herscovics,A. and Orlean,P. (1993) FASEB J., 7, 540-550. Hill,J.E., Myers,A.M., Koerner,T.J. and Tzagaloff,A. (1986) Yeast, 2, 163-168. Huffaker,T.C. and Robbins,P.W. (1982) J. Biol. Chem., 257, 3203-32 10. Huffaker,T.C. and Robbins,P.W. (1983) Proc. Natl Acad. Sci. USA, 80, 7466-7470. Johnston,M. and Davis,R.W. (1984) Mol. Cell. Biol., 4, 1440-1448. Kelleher,D.J. and Gilmore,R. (1994) J. Biol. Chem., 269, 12908-12917. Kelleher,D.J., Kreibich,G. and Gilmore,R. (1992) Cell, 69, 55-65. Klis,F.M. (1994) Yeast, 10, 851-869. Knauer,R. and Lehle,L. (1994) FEBS Lett., 344, 83-86. Kornfeld,R. and Kornfeld,S. (1985) Annu. Rev. Biochem., 54, 631-664. Kranz,J.E. and Holm,C. (1990) Proc. Natl Acad. Sci. USA, 87, 66296633. Kukuruzinska,M.A., Bergh,M.L.E. and Jackson,B.J. (1987) Annu. Rev. Biochem., 56, 914-944. Kumar,V., Heinemann,F.S. and Ozols,J. (1994) J. Biol. Chem., 269, 13451-13457. Kyte,J. and Doolittle,R.F. (1982) J. Mol. Biol., 157, 105-132. Lehle,L. (1980) Eur J. Biochem., 109, 589-601. Levin,D.E. and Bartlett-Heubusch,E. (1992) J. Cell Biol., 116, 12211229. Levin,D.E. and Errede,B. (1995) Curr Opin. Cell Biol., 7, 197-202. Munoz,M.D., Hernandez,L.M., Basco,R., Andaluz,E. and Larriba,G. (1994) Biochim. Biophys. Acta, 1201, 361-366. Murphy,L.A. and Spiro,R.G. (1981) J. Biol. Chem., 256, 7487-7494. Paravicini,G. et al. (1992) Mol. Cell Biol., 12, 4896-4905. Pathak,R., Hendrickson,T.L. and Imperiali,B. (1995a) Biochemistry, 34, 4179-4185. Pathak,R., Parker,C.S. and Imperiali,B. (1995b) FEBS Lett., 362, 229234. Roemer,T., Paravicini,G., Payton,M.A. and Bussey,H. (1994) J. Cell Biol., 127, 567-579. Rothblatt,J. and Schekman,R. (1988) Methods Cell Biol., 32, 3-36. Rothstein,R. (1991) Methods Enzymol., 194, 281-301. Runge,K.W. and Robbins,P.W. (1986) J. Biol. Chem., 261, 15582-15590. Runge,K.W., Huffaker,T.C. and Robbins,P.W. (1984) J. Biol. Chem., 259, 412-417. Sanglard,D., Sengstag,C. and Seghezzi,W. (1993) Eur J. Biochem., 216, 477-485. Sengstag,C., Stirling,C., Schekman,R. and Rine,J. (1990) Mol. Cell. Biol., 10, 672-680. Sharma,C.B., Lehle,L. and Tanner,W. (1981) Eur J. Biochem., 116, 101-108. Silberstein,S., Kelleher,D.J. and Gilmore,R. (1992) J. Biol. Chem., 267, 23658-23663. Silberstein,S., Collins,P.G., Kelleher,D.J., Rapiejko,P.J. and Gilmore,R. (1995) J. Cell Biol., 128, 525-536. Stagljar,I., te Heesen,S. and Aebi,M. (1994) Proc. Natl Acad. Sci. USA, 91, 5977-5981. Staneloni,R.J., Ugalde,R.A. and Leloir,L.E (1980) Eur J. Biochem., 105, 275-278. Struhl,K. and Davis,R.W. (1981) J. Mol. Biol., 152, 535-552. Tanner,W. and Lehle,L. (1987) Biochim. Biophys. Acta, 906, 81-99. te Heesen,S., Janetzky,B., Lehle,L. and Aebi,M. (1992) EMBO J., 11, 2071-2075. te Heesen,S., Knauer,R., Lehle,L. and Aebi,M. (1993) EMBO J., 12, 279-284. te Heesen,S., Lehle,L., Weissmann,A. and Aebi,M. (1994) Eur J. Biochem., 224, 71-79. Trimble,R.B., Byrd,J.C. and Maley,F. (1980) J. Biol. Chem., 255, 11892-11895. Turco,S.J., Stetson,B. and Robbins,P.W. (1977) Proc. Natl Acad. Sci. USA, 74, 4411-4414. Verostek,M.E, Atkinson,P.H. and Trimble,R.B. (1993a) J. Biol. Chem., 268, 12104-12115. Verostek,M.F., Atkinson,P.H. and Trimble,R.B. (1993b) J. Biol. Chem., 268, 12095-12103.
4959
R.Zufferey et al. Vijayraghavan,U., Company,M. and Abelson,J. (1989) Genes Del., 3, 1206-1216. Von Heijne,G. (1986) Nucleic Acids Res., 14, 4682-4690. Wilson,R. et al. (1994) Nature, 368, 32-38. Wimmer,C., Doye,V., Grandi,P., Nehrbass,U. and Hurt,E.C. (1992) EMBO J., 11, 5051-5061. Yoshida,S., Ikeda,E., Unoj. and Mitsuzawa,H. (1992) Mol. Gen. Genet., 231, 337-344. Received on June 7, 1995; revised on July 20, 1995
4960
