ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS

Vol. 338, No. 1, February 1, pp. 1–6, 1997Article No. BB969812

A Dual Affinity Tag on the 64-kDa Nlt1p Subunit Allows theRapid Characterization of Mutant Yeast OligosaccharylTransferase Complexes

Rahul Pathak and Barbara Imperiali1

Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125

Received August 8, 1996, and in revised form November 15, 1996

first purified from canine pancreas cells as a complexOligosaccharyl transferase catalyzes the glycosyla- of three polypeptides—ribophorin I, ribophorin II, and

tion of selected asparagine residues of nascent polypep- OST48 (1, 2). Transferase complexes with up to sixtide chains as they are translocated into the lumen of subunits have since been isolated from a number ofthe endoplasmic reticulum. To date, this enzyme has eukaryotic sources (3–9). Although the sizes of somebeen purified from a number of eukaryotic organisms. of the subunits of the various transferase complexesPurification of transferase activity has yielded polypep- are heterogeneous, sequence homology shows that thetide complexes of three to six subunits depending on three subunits of the canine enzyme are conservedthe source organism. Here we present the purification throughout evolution from yeast to mammals.of an affinity-tagged version of the enzyme complex The enzyme has been purified from Saccharomycesfrom a membrane protein fraction of the yeast Saccha- cerevisiae as a complex of four to six polypeptides (3–romyces cerevisiae. A yeast strain was created in which

5). The genes encoding five potential subunits of thethe essential 64-kDa glycoprotein Nlt1p subunit of theenzyme have now been cloned. The essential NLT1 oroligosaccharyl transferase was modified by the addi-OST1 gene encodes the 64-kDa glycoprotein subunittion of a 22-residue carboxy-terminal affinity tag; theof the enzyme (10, 11). The 64-kDa Nlt1p is variablytag included both an 8-residue FLAG epitope and a 6-glycosylated in three or four positions in vivo, has oneresidue histidine motif. Facile purification of the oligo-potential carboxy-terminal transmembrane helix, andsaccharyl transferase was achieved using affinity chro-shows significant homology to the luminal domain ofmatography media specific for each segment of the tag.the ribophorin I subunit of the mammalian enzymeThe enzyme was purified as a heteromeric complex of

five subunits in agreement with previously reported (10, 11). Conditional lethal mutations in the gene causecharacterizations of the yeast transferase. Yeast strains defects in protein glycosylation in vivo (11). The essen-bearing affinity-tagged enzyme subunits allow the tial WBP1 gene encodes the 45-kDa glycoprotein sub-rapid characterization of native and mutant trans- unit of the enzyme (12). Wbp1p is glycosylated in twoferase complexes. q 1997 Academic Press positions in vivo, has one potential carboxy-terminal

Key Words: N-linked glycosylation; oligosaccharyl transmembrane helix, and shows significant overall ho-transferase; affinity tagging; affinity purification; Sac- mology to the mammalian OST48 subunit (2, 5, 13).charomyces cerevisiae. Like mutations of the OST1 gene, conditional lethal

mutations of the WBP1 gene also cause protein glyco-sylation defects; in addition, depletion of the WBP1gene product yields a significant reduction in trans-

Oligosaccharyl transferase is a multimeric protein ferase activity and subsequent cell growth arrest (12).complex resident in the rough endoplasmic reticulum The essential SWP1 gene was isolated as a high-copymembrane of eukaryotic cells; the molecular composi- suppressor of one conditional lethal mutation in thetion and the specific functions of the enzyme subunits WBP1 gene (14). The 32-kDa Swp1p is not glycosylated,have yet to be conclusively defined. The enzyme was has three potential carboxy-terminal transmembrane

helices, and shows significant homology to the amino-terminal half of the 64-kDa ribophorin II mammalian1 To whom correspondence should be addressed. Fax: (818) 564-

9297. subunit (3, 5, 14). Depletion of the SWP1 gene product

10003-9861/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

AID ABB 9812 / 6b29$$$241 12-29-96 20:36:42 arca

2 PATHAK AND IMPERIALI

also yields cell growth arrest and a significant reduc- the native enzyme. The intact multi-subunit OTtagwas readily purified on affinity chromatography mediation of transferase activity (14).

Recent research has identified further subunits of as a complex of five polypeptides—Nlt1p-tag, Wbp1p,Ost3p, Swp1p, and Ost2p—even though only one sub-the yeast oligosaccharyl transferase complex that have

no apparent homologs in the purified mammalian en- unit carried the affinity tag. All five subunits appearedto be represented in identical stoichiometric amountszyme complex. The hydrophobic 34-kDa subunit is en-

coded by the OST3 gene (15). Unlike the other oligosac- including the nonessential Ost3p. The quaternarystructure of the enzyme complex was stable under vig-charyl transferase genes, OST3 is not essential for the

vegetative growth of the yeast cell; however, deletion orous detergent and ionic strength conditions. The ac-cessible purification of the OTtag complex provides anof the gene yields glycosylation defects similar to those

that result from mutations in the NLT1/OST1, WBP1, experimental system for the study of the molecularcomposition and ultimately the biochemical mecha-and SWP1 genes (15). Purification of the wild-type en-

zyme indicated that the 34-kDa subunit might have nism of this fascinating enzyme.been underrepresented in the multimeric complex (3,5). An additional 16-kDa subunit was observed in one MATERIALS AND METHODSpurification of the enzyme (3), but not in two subse-

Plasmids. Standard molecular biological and polymerase chainquent purifications (4, 5). That subunit is encoded byreaction techniques were used (21). Plasmid pYES2 (Invitrogen) wasthe essential OST2 gene (16). As with the other essen- modified by inserting a sequence encoding a 22-amino-acid affinity

tial oligosaccharyl transferase genes, conditional muta- tag including an 8-residue FLAG epitope and a 6-residue polyhisti-tions in the OST2 locus yield protein glycosylation de- dine tag for carboxy-terminal fusion to yeast genes. The resultant

pYES2(FLAG-6His) vector included the GAL1,10 promoter upstreamfects, and overexpression of the OST2 gene product alsoand the cyc1 terminator downstream of the polycloning region andsuppresses a conditional mutation in the WBP1 geneaffinity tags. The plasmid pRS414(CUP1p-FLAG-6His-cyc1) was cre-(16). Ost2p has two or three potential transmembrane ated for expression of FLAG and 61 His-tagged proteins from the

domains and might have a significant cytosolic domain inducible CUP1 promoter. The FLAG-6His-cyc1 sequence from plas-mid pYES2(FLAG-6His) was inserted into the polycloning region ofunlike the other three essential oligosaccharyl trans-plasmid pRS414 (22); the CUP1 promoter was obtained from J. Abel-ferase subunits. In addition, two other genes, OST4son and inserted upstream of the FLAG-6His-cyc1 sequence. Plasmidand STT3, have been associated with transferase activ-pRS414(CUP1p-NLT1-FLAG-6His-cyc1) was created by inserting

ity in vivo (17, 18), and one purification of the enzyme the coding region from the NLT1 gene into the vector pRS414-yielded a 9-kDa protein subunit that has not yet been (CUP1p-FLAG-6His-cyc1) downstream of the CUP1 promoter and

upstream of the FLAG-6His sequence such that the final codon ofcharacterized genetically (3).the NLT1 coding region was fused to the first codon of the affinityTo expand the potential for molecular and mechanis-tag sequence. The affinity tag sequence included a stop codon aftertic studies of the enzyme, we have created a yeast the final codon of the 6His tag. A site sensitive to Factor Xa protease

strain with an affinity-tagged oligosaccharyl trans- was placed upstream of the affinity tag to facilitate its removal inferase. The NLT1 gene encoding the largest subunit of future experiments. The carboxy-terminal sequence of the encoded

Nlt1p-tag fusion protein including the final four residues of the wild-the enzyme was deleted from a S. cerevisiae strain andtype polypeptide was Asn-Val-Thr-Asn-Val-Asp-Ile-Glu-Gly-Arg-replaced with a modified version of the same gene. TheGly-Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys-Gly-His-His-His-His-His-coding region of the modified gene was fused in-frame His-Stop.

to a sequence encoding a 22-residue carboxy-terminalYeast strains and media. The yeast strains used in the experi-

affinity tag including both an 8-residue FLAG epitope ments are listed in Table I. Standard yeast genetic techniques and(19) and a 6-residue polyhistidine motif for purification yeast media were used (23). Yeast strains were transformed by the

lithium acetate method (24). Yeast strains JD52 and JD53 wereof the enzyme on commercially available affinity chro-obtained from A. Varshavsky. Diploid yeast strain JD52/53 was ob-matography media. Expression of the modified genetained by mating the haploid strains JD52 and JD53 and selectingunder the control of the CUP1 promoter (20) from the zygotes by cell morphology. An NsiI–MscI fragment of the chromo-

yeast metallothionein gene under noninducing condi- somal copy of the NLT1 gene (10) of strain JD52/53 was deleted andtions yielded an appropriate amount of the tagged sub- replaced with the LEU2 gene to create strain RPY27; successful

disruption of the locus was verified by Southern blotting. Leucineunit.prototrophs were selected and sporulated; as expected, only twoYeast strains bearing the OTtag2 complex were via-spores from each tetrad gave rise to viable colonies. Diploid strainble and grew well on standard media; the affinity tag RPY27 was transformed with plasmid pRS414(CUP1p-NLT1-FLAG-

was stable even in high density cultures. The yields of 6His-cyc1) and sporulated; all four spores from several tetrads gaverise to viable colonies. All four colonies from one tetrad were trans-enzyme activity and the biochemical properties of theferred to media selective for leucine and tryptophan prototrophy; twoOTtag complex were indistinguishable from those ofstrains, RPY40 and RPY41, were viable on the media indicatingthe presence of both the chromosomal deletion and the expressionplasmid.2 Abbreviations used: OTtag, affinity-tagged oligosaccharyl trans-

ferase; YPD media, 1% yeast extract, 2% peptone, 2% dextrose; TX- Isolation and solubilization of yeast membranes. Yeast cells weregrown in YPD media and harvested in mid-log phase. The cells were100, Triton X-100; PC, phosphatidylcholine; NTA, nitrilotriacetic

acid. broken with glass beads, and membranes were harvested by centrifu-

AID ABB 9812 / 6b29$$$241 12-29-96 20:36:42 arca

3AFFINITY-TAGGED OLIGOSACCHARYL TRANSFERASE

TABLE I

Yeast Strains

Strain Genotype Source

JD52 MATa leu2-3,112 his3D200 trp1-D63 ura3-52 lys2-801 A. VarshavskyJD53 MATa leu2-3,112 his3D200 trp1-D63 ura3-52 lys2-801 A. VarshavskyJD52/53 MATa/MATa leu2-3,112/leu2-3,112 his3D200/his3D200 trp1-D63/trp1-D63 ura3-52/ura3-52 lys2-801/ This study

lys2-801RPY27 MATa/MATa leu2-3,112/leu2-3,112 his3D200/his3D200 trp1-D63/trp1-D63 ura3-52/ura3-52 lys2-801/ This study

lys2-801 nlt1-D2::LEU2//RPY39 MATa/MATa leu2-3,112/leu2-3,112 his3D200/his3D200 trp1-D63/trp1-D63 ura3-52/ura3-52 lys2-801/ This study

lys2-801 nlt1-D2::LEU2// pRS414(CUP1p-NLT1-FLAG-6His-cyc1)RPY40 MATa leu2-3,112 his3D200 trp1-D63 ura3-52 lys2-801 nlt1-D2::LEU2 pRS414(CUP1p-NLT1-FLAG- This study

6His-cyc1)RPY41 MATa leu2-3,112 his3D200 trp1-D63 ura3-52 lys2-801 nlt1-D2::LEU2 pRS414(CUP1p-NLT1-FLAG- This study

6His-cyc1)

gation at 40 krpm for 60 min in a Beckman Ti45 rotor (5). Membrane against the recombinant fusion protein, and after one more injectiona strong response was seen against the 64-kDa subunit of the purifiedproteins were partially solubilized in 50 mM Hepes, pH 7.5, 500 mM

NaCl, 10 mM MgCl2, 140 mM sucrose, 1.0% Nonidet P-40, and 0.1 oligosaccharyl transferase complex. Ascitic fluid from an inducedtumor was collected using standard procedures (25) and used as amM 4-(2-aminoethyl)benzoylfluoride. The solubilized fraction was

clarified by centrifugation at 40 krpm for 60 min. The supernatant polyclonal antiserum specific for Nlt1p. Antiserum specificity en-abled recognition of the glycoprotein from a crude lysate of yeastwas carefully separated from a glassy pellet and an amorphous tur-

bid pellet. proteins.Analytical methods. The protein composition of various biochemi-Purification of the OTtag complex. Oligosaccharyl transferase ac-

cal fractions was visualized by SDS–PAGE (26) on 4–20% acryl-tivity was monitored as described previously (5). Twenty millilitersamide gradient gels followed by silver staining (27). Broad-rangeof solubilized membrane proteins isolated from 40 g RPY40 yeastmolecular weight markers (New England Biolabs) were used to esti-cells was gently agitated with 6 ml of concanavalin A–agarose (Vec-mate the apparent molecular weight of each subunit. For immu-tor) at 47C for 12 h. The agarose medium was then pelleted by low-noblotting, proteins were separated by SDS–PAGE and then trans-speed centrifugation and the supernatant removed and discarded.ferred to nitrocellulose (28). The membranes were blocked with non-The medium was then washed with 80 ml of the solubilization bufferfat dried milk and then probed with antiserum specific for Nlt1pcontaining 0.5 mg/ml PC (Avanti Polar Lipids). Enzyme activity was(above), Wbp1p (14), Swp1p (unpublished results), Ost2p (16), or theeluted batchwise by treating the medium three times with 750 mM

FLAG epitope (Eastman Kodak). The membranes were then probeda-methylmannopyranoside in 80 ml of the same buffer. The combinedwith alkaline phosphatase-conjugated secondary antibodies (anti-eluted fractions were then adjusted to 15 mM imidazole and 1 mM

Nlt1p, anti-Wbp1p, anti-Swp1, and anti-FLAG) or with horseradishb-mercaptoethanol and applied to 6 ml of Ni-NTA–agarose mediumperoxidase-conjugated secondary antibodies (anti-Ost2p). Alkaline(Qiagen) by batch incubation. The medium was then applied to a 2-phosphatase-conjugated secondary antibodies were visualized withcm-diameter column and washed with 40 ml of 20 mM Hepes, pH5-bromo-4-chloro-3 *-indolylphosphate and nitroblue tetrazolium7.5, 150 mM NaCl, 10 mM MgCl2, 140 mM sucrose, 1.0% TX-100, and(Pierce); horseradish peroxidase-conjugated secondary antibodies0.5 mg/ml PC. Bound proteins were eluted with 10-ml applicationswere visualized by enhanced chemiluminescence (Amersham).of 20 mM Mes, pH 5.0, 150 mM NaCl, 10 mM MgCl2, 140 mM sucrose,

1.0% TX-100, and 0.5 mg/ml PC; eluted fractions were immediatelyadjusted to 50 mM Hepes, pH 7.5. Active fractions were pooled and RESULTSapplied to 1.0 ml M2–agarose (Eastman Kodak) batchwise. The me-dium was washed batchwise three times with 10 ml 50 mM Hepes, To facilitate the in vivo and in vitro examinationpH 7.5, 150 mM NaCl, 10 mM MgCl2, 140 mM sucrose, 1.0% TX-100, of the yeast oligosaccharyl transferase, a yeast strainand 0.5 mg/ml PC buffer and then eluted batchwise with 0.05 mg/ bearing a tagged version of the largest enzyme subunitml FLAG peptide (Eastman Kodak) in 5.0 ml of the same buffer.

was created. Two affinity tags were fused to the car-Approximately 10% of the total activity in the initial fraction of solu-boxy-terminus of the NLT1 gene. The eight-residuebilized yeast membrane proteins was recovered after three steps of

affinity purification. FLAG epitope enables immunopurification of theProduction of polyclonal antiserum against Nlt1p. The DNA en- tagged gene product with the monoclonal antibody M2

coding amino acids 23 through 460 of the yeast Nlt1p (10) was sub- medium (19), and the six-histidine motif enables bind-cloned into vector pJC20HC (Carl Parker, personal communication) ing of the subunit to immobilized nickel cations (29).to create vector pJC-NLT1-HC expressing the luminal domain of

The affinity tags were placed at the carboxy-terminusNlt1p with a carboxy-terminal six-histidine affinity tag. The fusionof the subunit to minimize interference with the func-protein was expressed in Escherichia coli BL21(DE3)pLysS cells (No-

vagen) from the T7 promoter. Induced cells from a 500-ml culture tion of the enzyme. The 64-kDa glycoprotein Nlt1p sub-were lysed in 6 M guanidine–HCl, and the lysate was applied to Ni- unit, which assembles into the enzyme complex in theNTA–agarose beads. The pure recombinant protein was eluted with membrane of the rough endoplasmic reticulum, ap-30 mM EDTA and 150 mM imidazole in 6 M urea. Swiss–Webster

pears to have an extensive luminal domain followed bymice were innoculated at 2-week intervals with 10 mg of protein each(25). After three rounds of injection, a strong response was seen a potential transmembrane hydrophobic domain and a

AID ABB 9812 / 6b29$$$241 12-29-96 20:36:42 arca

4 PATHAK AND IMPERIALI

short cytoplasmic tail of a few amino acids (10, 11). In detergent concentrations to minimize the presence ofany loosely interacting polypeptides that might not befact, of the five cloned yeast oligosaccharyl transferase

genes, only Ost2p might have a significant domain on essential for oligosaccharyl transferase activity. Theoverall activity yield of the purification of OTtag wasthe cytoplasmic side of the endoplasmic reticulum

membrane (16). The NLT1-FLAG-6His fusion was ex- comparable to that of the purification of the native en-zyme (5).pressed on a low-copy yeast episomal plasmid (22) un-

der the control of the promoter from the yeast CUP1 To date, five polypeptides have been consistently en-riched in the final fraction (Fig. 1). Four of the pre-gene (20). The CUP1 promoter provided inducible ex-

pression of the tagged gene product with CuSO4; the viously reported oligosaccharyl transferase subunitswere conclusively identified by Western blotting. Thebackground expression from the promoter in YPD me-

dia with no added CuSO4 gave expression of the tagged tagged Nlt1p subunit migrated as a pair of glycoformsat roughly 70 kDa; both glycoforms were detected byNlt1p comparable to that of the wild-type NLT1 gene.

One chromosomal copy of the NLT1 gene was deleted antibodies specific for the recombinant Nlt1p and forthe FLAG epitope (Fig. 2, lanes a and b). The glycosyl-from a diploid yeast strain and replaced with the se-

lectable LEU2 marker. The resultant strain RPY27 ated Wbp1p was detected at 45 kDa and Swp1p wasdetected at 30 kDa by the appropriate antisera (Fig. 2,was heterozygous for a deletion of all but 135 bp of the

coding region of the essential NLT1 gene; sporulation lanes c and d). The recently described Ost2p (16) wasconclusively detected at 16 kDa both by silver-stainedof this strain yielded viable colonies from only two

spores of each tetrad dissected. Strain RPY27 was SDS–PAGE (Fig. 1) and by Western blotting (Fig. 2,lane e). Ost2p was detected in one previously reportedtransformed with the plasmid expressing the tagged

NLT1 gene product to yield strain RPY39. When strain characterization (3) of the native enzyme complex butnot in two other characterizations (4, 5). Based on theRPY39 was sporulated and dissected, several tetrads

yielded viable colonies from all four spores indicating clear enrichment of Ost2p in the purification of theaffinity-tagged oligosaccharyl transferase complex, wethat the NLT1-FLAG-6His fusion complemented the

deletion of the essential NLT1 gene. Haploid strains concluded that Ost2p is a bona fide member of themulti-subunit enzyme. Cross-reactivity of each poly-RPY40 and RPY41 were isolated from one tetrad and

harbored both the chromosomal deletion of the NLT1 clonal antibody to other proteins in the purified fractionwas minimal.gene and the expression plasmid.

Yeast strain RPY40 grew well on YPD media both A fifth polypeptide was detected at 34 kDa and hason plates and in culture. Western blotting analysisshowed that the tagged Nlt1p was stable even in highdensity cultures. Two 70-kDa glycoforms (5) of Nlt1p-FLAG-6His were present when whole cell yeast ex-tracts were probed with anti-FLAG M2 monoclonal an-tibody; the two 70-kDa glycoprotein bands were re-vealed by probing the same fraction with anti-Nlt1pantiserum—no untagged 64-kDa Nlt1p was present.The oligosaccharyl transferase activity of solubilizedmembrane protein fractions from strain RPY40 wasassessed as described previously (5) and found to beindistinguishable from similar fractions prepared fromyeast strains expressing the native Nlt1p subunit (datanot shown). Additionally, strain RPY40 showed compa-rable growth rates to the strain expressing the wild-

FIG. 1. Purification of the OTtag complex. OTtag was purified ontype protein. a series of affinity chromatography media. As an initial purification

To establish a method for the accessible isolation of step, solubilized yeast membrane proteins were separated with thelectin concanavalin A immobilized on agarose. The glycoprotein com-a stable, active oligosaccharyl transferase, the OTtagplexes were then applied to Ni-NTA–agarose and eluted with a pHcomplex was purified by a series of three chromato-step gradient. Even after only two steps of purification, the oligosac-graphic steps. Two of the chromatography media used charyl transferase complex is visible (lane d); the apparent molecular

in the purification, Ni-NTA–agarose and M2–agarose, weights of the five subunits are 70, 45, 34, 30, and 16 kDa. OTtagwere specific for each segment of the affinity tag; the was then separated from the contaminating proteins by purification

on an immobilized monoclonal antibody specific for the FLAG epitopethird, concanavalin A–agarose, was specific for the oli-of the enzyme complex. The protein content of each fraction in thegosaccharides of the polypeptide complex and had pre-purification was visualized by SDS–PAGE followed by silver stain-viously been shown to be useful for the purification of ing. (a) Molecular weight standards, (b) solubilized yeast membrane

the native enzyme (3, 5). The purification was carried proteins, (c) concanavalin A eluate, (d) Ni-NTA eluate, and (e) M2(anti-FLAG) eluate.out under vigorous conditions including high salt and

AID ABB 9812 / 6b29$$$241 12-29-96 20:36:42 arca

5AFFINITY-TAGGED OLIGOSACCHARYL TRANSFERASE

STT3 and OST4 have been identified as genes essentialfor proper transferase function in vivo (17, 18).

To address the molecular composition of the enzyme,the largest subunit was affinity tagged. The resultingOTtag complex was purified by a sequence of affinitychromatography steps. In accord with the purificationof the native yeast enzyme, the four large subunits of70, 45, 34, and 30 kDa were observed in the purifiedcomplex; the size of the 70-kDa polypeptide is consis-FIG. 2. Identification of the subunits of purified OTtag. The sub-tent with that expected of the tagged Nlt1p subunit.units of the OTtag complex were identified by Western blotting. M2

(anti-FLAG) eluate treated with (a) anti-Nlt1p polyclonal antiserum, Western blotting identified three of the subunits as the(b) anti-FLAG monoclonal antibody, (c) anti-Wbp1p polyclonal anti- products of the NLT1-FLAG-6His, WBP1, and SWP1serum, (d) anti-Swp1p polyclonal antiserum, and (e) anti-Ost2p poly- genes.clonal antiserum. (Note: The anti-Ost2p is weaker and exhibits some

A polypeptide at 16 kDa was present in the purifiednonspecific binding; however, the molecular weight of the intenseOTtag complex, and this polypeptide was conclusivelyband corresponds to Ost2p. A fivefold excess sample was applied to

the gel for the Ost2p analysis to compensate for the weaker signal.) identified as Ost2p by Western blotting. This observa-tion updates the previously reported description of afour-subunit oligosaccharyl transferase complex fromthis laboratory (5) and confirms the results of anotherbeen identified as Ost3p based on molecular weight. previously reported purification of a six-subunit com-Unlike the other four characterized subunits of the plex (3). It is likely that the 16-kDa subunit of thetransferase, Ost3p is not essential for the vegetative enzyme was not observed on the Coomassie blue-growth of the yeast cell (15). However, Ost3p appeared stained linear SDS–PAGE gels used to characterizeto be present in the enzyme complex at a level compara- the native complex (5). Here, using silver-stained gra-ble to the other subunits and was not depleted through dient SDS–PAGE gels Ost2p was consistently ob-the purification of the enzyme. No other polypeptides served in the purified OTtag complex, and Westernwere detected at apparent stoichiometric amounts rela- blotting of the native complex purified according to (5)tive to the known transferase subunits. When the puri- confirms the presence of Ost2p in the enzyme (data notfied enzyme was assayed by SDS–PAGE at high pro- shown). The OST2 gene is essential for viability andtein concentrations, a few low-molecular-weight poly- for protein glycosylation (16), and the 16-kDa subunitpeptides were detected. One of these might be the should be considered an integral part of the enzyme.9-kDa oligosaccharyl transferase subunit previously The final 34-kDa subunit was identified as Ost3pdescribed (3). The affinity-tagged purification verifies by molecular weight. The nonessential 34-kDa Ost3pthe molecular composition of the transferase and shows subunit did not appear to be underrepresented in thethat the subunits form a stable quaternary structure in vivo complex, nor did it appear to be depleted duringthat is maintained in the membrane and in a detergent the independent purification of OTtag describedsolubilization. herein. Since this subunit appears to be present in thepurified OTtag complex in an amount comparable to

DISCUSSION that of the essential OTtag subunits, it is likely to bepresent in the in vivo complex at the same stoichiomet-The oligosaccharyl transferase has been purified

from the yeast S. cerevisiae by several groups with rea- ric amount. If it was underrepresented in the purifica-tion of the native complex (3, 5), then that underrepre-sonable but not complete agreement in the observed

subunit composition of the enzyme. Three independent sentation would more likely have been due to a deple-tion of the polypeptide in vitro rather than apurifications of the enzyme yielded multimeric poly-

peptide complexes with subunits of 64, 45, 34, and 30 substoichiometric participation in the in vivo complex.Although this subunit is not essential for viability, itskDa (3–5); those subunits are encoded by the NLT1/

OST1, WBP1, OST3, and SWP1 genes, respectively loss affects the glycosylation state of several proteinsin vivo and reduces the measured oligosaccharyl trans-(10–12, 14, 15). The 34-kDa subunit is not essential

for the growth of the yeast cell (15), and it might be ferase activity of membrane extracts in vitro (15).Therefore, loss of the Ost3p subunit during purificationdepleted during the purification of the enzyme (3, 5).

One of the three purifications also included low-molecu- would reduce but not eliminate enzyme activity.No other polypeptides were observed at apparentlar-weight subunits with weights of 16 and 9 kDa (3)

that were not observed in the other two (4, 5); the 16- stoichiometric amounts relative to the known oligosac-charyl transferase subunits. However, under high pro-kDa subunit is encoded by the OST2 gene (16). In addi-

tion to those polypeptides identified as transferase sub- tein loading conditions several additional polypeptideswere observed, particularly at low molecular weights.units by direct biochemical purification of the enzyme,

AID ABB 9812 / 6b29$$$241 12-29-96 20:36:42 arca

6 PATHAK AND IMPERIALI

2. Silberstein, S., Kelleher, D. J., and Gilmore, R. (1992) J. Biol.One of these might have been the 9-kDa subunit identi-Chem. 267, 23658–23663.fied in a purification of the native enzyme (3). Another

3. Kelleher, D. J., and Gilmore, R. (1994) J. Biol. Chem. 269,might have been the product of the OST4 gene which12908–12917.has a predicted weight of 3.6 kDa (17). It is possible

4. Knauer, R., and Lehle, L. (1994) FEBS Lett. 344, 83–86.that the products of the OST4 and STT3 genes are5. Pathak, R., Hendrickson, T. L., and Imperiali, B. (1995) Bio-present in the complex at substoichiometric amounts, chemistry 34, 4179–4185.

and it is also possible that they are required for the in 6. Kumar, V., Heinemann, F. S., and Ozols, J. (1994) J. Biol. Chem.vivo function of the enzyme but not the actual catalysis 269, 13451–13457.of the glycosylation reaction. 7. Breuer, W., and Bause, E. (1995) Eur. J. Biochem. 228, 689–

696.The OTtag system provides a setting for the compre-8. Kumar, V., Korza, G., Heinemann, F. S., and Ozols, J. (1995)hensive study of the oligosaccharyl transferase en-

Arch. Biochem. Biophys. 320, 217–223.zyme. OTtag has initially been used to refine the under-9. Kumar, V., Heinemann, F. S., and Ozols, J. (1995) Biochem. Mol.standing of the molecular composition of the enzyme

Biol. Int. 36, 817–826.by tagging the Nlt1p subunit of the enzyme. The short10. Pathak, R., Parker, C. S., and Imperiali, B. (1995) FEBS Lett.tag is equally effective when placed at the carboxy-

362, 229–234.terminus of the Wbp1p subunit (unpublished results).

11. Silberstein, S., Collins, P. G., Kelleher, D. J., Rapiejko, P., andFurther biochemical study of the yeast enzyme will be Gilmore, R. (1995) J. Cell Biol. 128, 525–536.aided by the rapid and amenable purification of the 12. te Heesen, S., Janetzky, B., Lehle, L., and Aebi, M. (1992) EMBOtagged enzyme. Overexpressed subunits of the trans- J. 11, 2071–2075.ferase are readily purified when tagged, and a protease 13. te Heesen, S., Rauhut, R., Aebersold, R., Abelson, J., Aebi, M.,

et al. (1991) Eur. J. Cell Biol. 56, 8–18.sensitive site will allow removal of the tag. In addition,14. te Heesen, S., Knauer, R., Lehle, L., and Aebi, M. (1993) EMBOthe stable and reversible association of the polyhisti-

J. 12, 279–284.dine motif with metal cations (29) enables specific bind-15. Karaoglu, D., Kelleher, D. J., and Gilmore, R. (1995) J. Cell Biol.ing to an immobilized medium through a single site

130, 567–577.of each enzyme complex. Such binding can be used to16. Silberstien, S., Collins, P. G., Kelleher, D. J., and Gilmore, R.examine the subunit interactions under mildly dena- (1995) J. Cell Biol. 131, 371–383.

turing conditions and to assay the substrate binding 17. Chi, J. H., Roos, J., and Dean, N. (1996) J. Biol. Chem. 271,properties of the individual enzyme substrates. Finally, 3132–3140.a plasmid carrying a tagged gene encoding a trans- 18. Zufferey, R., Knauer, R., Burda, P., Stagljar, I., te Heesen, S.,ferase subunit is ideally suited to site-directed muta- et al. (1995) EMBO J. 14, 4949–4960.genesis. After mutations are introduced into the trans- 19. Brizzard, B., Chubet, R., and Vizard, D. (1994) Biotechniques

16, 730–734.ferase gene, the plasmid can be transformed into the20. Butt, T. R., Sternberg, E. J., Gorman, J. A., Clark, P., Hamer,appropriate yeast strain where the mutant complex can

D., et al. (1984) Proc. Natl. Acad. Sci. USA 81, 3332–3336.be rapidly assayed for molecular composition and for21. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecularcatalytic efficiency.

Cloning: A Laboratory Manual, Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, NY.

ACKNOWLEDGMENTS 22. Sikorski, R. S., and Hieter, P. (1989) Genetics 122, 19–27.23. Guthrie, C., and Fink, G. R. (1991) Guide to Yeast Genetics andThis research was supported by NIH Grant GM39334. We thank

Molecular Biology (Abelson, J. R., and Simon, M. I., Eds.), Aca-Markus Aebi for providing antiserum specific for Wbp1p, the cloneddemic Press, San Diego.WBP1 and SWP1 genes, and yeast strain TH451 bearing a deletion in

the chromosomal WBP1 locus. We thank Reid Gilmore for antiserum 24. Schiestl, R. H., and Gietz, R. D. (1989) Curr. Genet. 16, 339–specific for Ost2p. We also thank Carl Parker, Ebrahim Zandi, Chris 346.Trotta, and Christine Davis for useful advice during the course of 25. Harlow, E., and Lane, D. (1988) Antibodies: A Laboratory Man-these experiments and the preparation of the manuscript. Numerous ual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor,plasmids and yeast strains were kindly provided by the laboratories NY.of Carl Parker, Alexander Varshavsky, John Abelson, and Judith

26. Laemmli, U. K. (1970) Nature 227, 680–685.Campbell. Nlt1p polyclonal antibody production was performed by27. Wray, W., Boulikas, T., Wray, V. P., and Hancock, R. (1981)Susan Ou of the Caltech Antibody Facility.

Anal. Biochem. 118, 197–203.28. Towbin, H., Staehelin, T., and Gordon, V. (1979) Proc. Natl.

REFERENCES Acad. Sci. USA 76, 4350–4354.29. Hochuli, E., Dobeli, H., and Schacher, A. (1987) J. Chromatogr.1. Kelleher, D. J., Kreibich, G., and Gilmore, R. (1992) Cell 69, 55–

65. 411, 177–184.

AID ABB 9812 / 6b29$$$241 12-29-96 20:36:42 arca