THE JOURNAL OF BIOLOGICAL Vol. No. 5, Q and in U.S.A ... · Prepro-a-factor Has a Cleavable Signal Sequence ... Emily A. Evans$, and Giinter Blobel ... 2.5-p1/25-p1 reaction of ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 263, No. 13, Issue of May 5, pp. 6209-6214,1988 Printed in U.S.A.

Prepro-a-factor Has a Cleavable Signal Sequence (Received for publication, October 14,1987)

M. Gerard Waters#, Emily A. Evans$, and Giinter Blobel From the Laboratory of Cell Bwbgy, Howard Hughes Medical Institute, The Rockefeller University, New York, New York 10021

MATa Saccharomyces cerevisiae secrete a small pep- tide mating pheromone termed a-factor. Its precursor, prepro-a-factor, is translocated into the endoplasmic reticulum and glycosylated at three sites. The glyco- sylated form is the major product in a yeast in vitro translation/translocation system. However, there is another translocated, nonglycosylated product that contains a previously unidentified modification. Con- trary to previous results suggesting that the signal sequence of prepro-a-factor is not cleaved, amino-ter- minal radiosequencing has identified this product as prepro-a-factor without its signal sequence, that is, pro-a-factor. The translocated, glycosylated proteins are also processed by signal peptidase. Moreover, we have found that both purified eukaryotic and prokar- yotic signal peptidase can process prepro-a-factor. Experiments using a yeast secretory mutant (secl8) blocked in transport from the endoplasmic reticulum to the Golgi indicate that the protein is also cleaved in vivo. Finally, characterization of the Asn-linked oligo- saccharide present on pro-a-factor in the yeast in vitro system by use of specific glucosidase and mannosidase inhibitors indicates that they have had the three ter- minal glucoses and probably one mannose removed. Therefore they most likely consist of ManeGlcNAct structures, identical to those found in the endoplasmic reticulum in vivo.

The yeast Saccharomyces cereuisiae can exist in two distinct haploid mating types, termed a and a, which can fuse to form a/a diploids (for review see Sprague et al., 1983). Yeast cells of the a mating type secrete a mating pheromone, termed a- factor, that acts on a cells to cause metabolic changes and arrest of the cell cycle at G1. Likewise a cells secrete a pheromone, called a-factor, that induces similar changes by a cells. This cell cycle arrest is a prerequisite to mating.

a-Factor, a 13-amino acid peptide (Stotzler et al., 1976), is synthesized as a larger precursor (Kujan and Herskowitz, 1982; Emter et al., 1983) termed prepro-a-factor. The precur- sor, encoded by the gene MFal (Kurjan and Herskowitz, 1982), is 165 amino acids long. At the amino terminus is a typical signal sequence followed by a pro region that has three sites for Asn-linked glycosylation. The carboxyl half of the molecule consists of four a-factor units separated from each other and from the pro segment by “spacer” regions.

Translocation of prepro-a-factor into the yeast endoplasmic

* This work was supported in part by National Institutes of Health Grant GM-27155 ( t o G. Blobel). The costa of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “Ocluertisemnt’’ in accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Supported by National Institutes of Health Grant GM-07982. To whom correspondence should be addressed.

8 Supported by the Howard Hughes Medical Institute.

reticulum (ER)’ results in glycosylation at the three Asn-X- Thr sites in the pro region (Emter et al., 1983; Julius et al., 1984a). After transit to the Golgi, the cores are extended to form complex chains (Julius et al., 1984a; Julius et al., 198413). Subsequently, the a-factor units are excised from the precur- sor by a series of proteases. First, a membrane-bound endo- protease, the product of the KEX2 gene, releases the four “spacer-a-factor” moieties (Julius et al., 1984b; Achstetter and Wolf, 1985). This proteolytic event probably occurs in the Golgi apparatus (Julius et al., 1984a). Next, either late in the Golgi or in the secretory vesicles (Julius et al., 1984a), the spacer regions are removed by a membrane-bound dipeptidyl aminopeptidase which is the product of the STEl3 gene (Julius et al., 1983). In addition, a carboxypeptidase B-like activity, the product of the KEXl gene, removes two amino acids from the carboxyl terminus of each a-factor unit, re- sulting in the mature peptide (Dmochowska et al., 1987).

Several laboratories have been using an in vitro system derived from yeast to study the process of translocation into the ER. In these systems translocation of prepro-a-factor does not result in any lower M, products that would be indicative of signal sequence cleavage (Hansen et al., 1986; Rothblatt and Meyer, 1986a; Rothblatt and Meyer, 1986b; Waters and Blobel, 1986; Waters et al., 1986). These obser- vations supported the suggestion, also based on the absence of a lower M, product in uiuo, that the signal sequence of prepro-a-factor was not removed by signal peptidase upon translocation into the ER (Julius et al., 1984a). In this work, however, we show that the signal sequence of prepro-a-factor is processed by the signal peptidase present in yeast micro- somes in vitro and in the yeast ER in uiuo. The precursor can also be processed by purified Escherichia coli or canine signal peptidase.

EXPERIMENTAL PROCEDURES

Materials-Dr. David Julius (Columbia University, NY) kindly provided the SP6 in vitro transcription vector containing the MFal gene. The tripeptide acceptor for oligosaccharidyl transferase, CH&O-Asn-Phe-Thr-NH2, was a gift from Dr. Howard Kaplan (Uni- versity of Texas, Houston). Purified E. coli signal (leader) peptidase was the generous gift of Dr. William Wickner (UCLA). Dr. Randy Schekman (University of California at Berkeley) kindly supplied the yeast strain SEY 5186 (secld suc2-A9 ura3 leu2 MATa) . Dr. Dennis Shields (Albert Einstein College of Medicine, NY) kindly provided the anti-a-factor antiserum. Unless otherwise noted, materials were from the sources described (Waters et al., 1986).

Yeast Tra~slotwns/Transbcatwns-The method was as described previously (Waters et al., 1986) except that the protease inhibitor mix was omitted, nuclease-treated yeast microsomal membranes were used at 5 eq/25-p1 reaction, and 200 ng of prepro-a-factor mRNA was used per 25 pl translation. The protease protection and Endo H digestion experiments were also done as reported previously (Waters and Blobel, 1986). Where indicated Asn-Phe-Thr oligosaccharide

The abbreviations used are: ER, endoplasmic reticulum; PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate; Endo H, endo-8-N-acetylglucosaminidase H.

6209

6210 Prepro-&-factor Has a Cleavable Signal Sequence

acceptor peptide was added from a 10 mM stock in water to yield a final concentration of 1.6 mM, and deoxynojirimycin (Genzyme) was added from a 1 mM stock in water to a final concentration of 100 p ~ .

Radbseguencing-Yeast in vitro translation/translocations were performed as described above except that 10 eq of nuclease-treated yeast microsomal membranes and 400 ng of prepro-a-factor mRNA were used per 25 pl-reaction. In addition, the "energy mix" (Waters and Blobel, 1986), which normally contains the 20 amino acids except methionine, was prepared without either alanine or proline but con- tained the other 19 amino acids. The proteins were labeled by either 2.5-p1/25-p1 reaction of [3H]alanine (Du Pont-New England Nuclear, L - [ ~ - ~ H ] , 85 Ci/mmol, concentrated 10-fold by lyophilization to 20 pg/ml) or [3H]proline (Du Pont-New England Nuclear, ~-[2,3,4,5,- 3H], 109 Ci/mmol, concentrated 10-fold by lyophilization to 22 pg/ ml) instead of [35S]methionine. The reactions were separated by sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis (SDS- 12%PAGE). The proteins were transferred to polyvinylidine difluo- ride membrane (Millipore Immobilon Transfer Membrane) according to the procedure of Matsudaira (1987) except that the gel was soaked twice for 10 min in transfer buffer prior to transfer. The blotted proteins were visualized by autoradiography, the relevant bands ex- cised, and the membrane-bound samples were submitted for auto- mated gas phase Edman degradation (Rockefeller University Protein Sequencing Facility). The sequenator samples were solubilized in 100 pl of 90% acetonitrile/lO% water for 1 h at room temperature and counted in 5 ml of Hydrofluor scintillation fluid (National Diagnos- tics).

Eukaryotic S i g d Peptidose Assay-The 2O-pl assays contained 5 pl of purified signal peptidase complex (Evans et al., 1986) in 200 mM sodium phosphate, pH 6.8,250 mM potassium acetate, pH 7.4, 1 mM dithiothreitol, 0.4% octaethylene glycol mono-n-dodecyl ether (Nik- kol), 0.4 mg/ml phosphatidylcholine, or the same buffer without signal peptidase complex, 1 pl of 8 mg/ml phosphatidylcholine, 4 pl of 500 mM triethanolamine acetate, pH 8, 1% Nikkol, 5 pl of [%]methio- nine-labeled prepro-a-factor synthesized in wheat germ (Waters et al., 1986), and 1 pl of 1% SDS. Incubation was for 90 min at 25 "C.

Prokaryotic Signal (Leader) Peptidase Assay-5 pl of [%]methio- nine-labeled prepro-a-factor synthesized in wheat germ (Waters et al., 1986) was mixed with 5 pl of purified E. coli signal (leader) peptidase (Zwizinski and Wickner, 1980) (2 mg/ml) in 1 : lO polybuffer 74 (Pharmacia LKB Biotechnology Inc.)/l% 8-octyl glucoside or the same buffer without signal peptidase and incubated for 60 min at 37 "C.

In Vivo L a b e l i n g and Immunoprecipitation-SEY 5186 (seclb MATa) cells were grown in SD media (0.67% Bacto yeast nitrogen base without amino acids, 2% dextrose, Hicks and Herskowitz, 1976) containing 20 pg/ml uracil and 30 pg/ml leucine a t 22 "C to an OD, of 1. The cells were collected by centrifugation, washed with an equal volume of fresh growth media, and resuspended at 2 ODm/ml. Four 1.25-ml aliquots were prepared in glass culture tubes, two of which received 30 pl of 1 mM tunicamycin (20 pg/ml final; Boehringer Mannheim), and all were incubated at 22 "C for 15 min. Two of the cultures (one without and one with tunicamycin) were shifted to 37 "C while the others were kept a t 22 "C and incubated for 15 min. 25 pl of [3sS]methionine (Du Pont-New England Nuclear, 10 Ci/ml) were added to each culture and incubation (at 22 and 37 "C) continued for 30 min. Trichloroacetic acid was added to 5% and incubated for 15 min at room temperature. The precipitates were collected by centrifugation at about 2500 X g for 10 min, the supernatants removed by aspiration, the pellets resuspended in 50 pl of lysis buffer (50 mM Tris-C1, pH 7.5,l mM EDTA, 1% SDS), and 2 pl of an unlabeled a/ a diploid yeast extract (termed S100-G25, Waters and Blobel, 1986) was added. 100 mg of 1 N HCI washed 0.5-mm glass beads (Sigma, Type V) were added and the tubes vortexed rigorously for 2 min followed by incubation at 100 "C for 3 min. 1 ml of 0 'C immunopre- cipitation buffer (50 mM Tris-C1, pH 7.5, 150 mM NaCI, 0.1 mM EDTA, 0.5% Tween 20) with 1 mM phenylmethylsulfonyl fluoride was added and mixed. After centrifugation as above, 900 pl of the supernatants were transferred to 1.5-ml microcentrifuge tubes, 2 pl of anti2a-factor antiserum were added and incubated at 0 "C for 2 h. 50 pl of 3.6% (w/v) protein-A-Sepharose CL-4B beads (Pharmacia LKB Biotechnology Inc.) in immunoprecipitation buffer were added and the incubation continued for 2 h more with mixing. The beads were then washed at 4 "C with 250 pl of the following solutions: twice with immunoprecipitation buffer, twice with urea wash buffer (100 mM Tris-C1, pH 7.5, 200 mM NaCI, 2 M urea, 0.5% Tween 20), once with 1% @-mercaptoethanol, and once with 0.1% SDS. The samples

were prepared for electrophoresis by addition of SDS-PAGE loading buffer and incubation at 100 "C for 3 min.

RESULTS

Identifieation of a Translocated, Nonglycosylated Product- A number of laboratories (Hansen et al., 1986; Rothblatt and Meyer, 1986; Waters and Blobel, 1986) have developed in vitro systems from yeast capable of translating, translocating, and glycosylating prepro-a-factor. Translation in our system without added mRNA resulted in low background protein synthesis (Fig. 1, lane 1) . Addition of in vitro transcribed prepro-a-factor mRNA resulted predominantly in the synthe- sis of the corresponding protein which migrated on SDS- 12%PAGE at M , = 19,000 (Fig. 1, lane 2). However, there was another minor mRNA-dependent product that migrated at M, = 20,000 (Fig. 1, lane 2). The presence of nuclease- treated yeast microsomes during translation did not signifi- cantly increase background translation (Fig. 1, lane 3). Pre- pro-a-factor synthesis in the presence of yeast microsomes resulted in three additional polypeptides of slower mobility, and the minor product at M , = 20,000 increased (Fig. 1, lane 4 ) . This M , = 20,000 product is also evident in the system of Hansen et al. (1986). Only the M, = 19,000 primary translation product was sensitive to digestion by externally added trypsin (Fig. 1, lanes 5 and 6 ) . However, if the protease digestion was done in the presence of detergent to disrupt membrane integ- rity, then all the prepro-a-factor mRNA-specific products were degraded (Fig. 1, lane 7). The new band at about M , = 29,000 is a degradation product from a high M, translation product (Waters and Blobel, 1986). These data indicate that all the products except the primary translation product were sequestered within microsomes. Endo H treatment of a trans- location reaction resulted in digestion of the three high M, products with production of lower M, forms (Fig. 1, lanes 8 and 9), showing that these were core-glycosylated, probably containing 1, 2, or 3 Asn-linked oligosaccharides each. The M , = 20,000 form did not appear to be affected by Endo H and therefore did not contain Asn-linked oligosaccharide larger than Man5GlcNAc2 (Tarentino and Maley, 1974). What

32-

27.

24.

20. 19-

Translocation n n n a

I I H I : S E E I I + +

Protease Endo H

"

.K l- a,

L

I + I + I + + I + +

1 2 3 4 5 6 7 8 9 1 0

FIG. 1. Identification of a translocated, nonglycosylated de- rivative of prepro-a-factor. As described under "Experimental Procedures," yeast translations were done in the absence (lanes I and 3) or presence (lanes 2 and 4-10) of prepro-a-factor mRNA and absence (lanes 1 and 2 ) or presence (lanes 3-10) of yeast nuclease- treated microsomal membranes (Memb). The reactions were then either prepared for electrophoresis (lanes 1-4 and IO) or subjected to protease (lanes 5-7) or Endo H (lanes 8 and 9 ) digestion protocols. Translocation was also done in the presence of the oligosaccharide acceptor peptide, Asn-Phe-Thr, to inhibit glycosylation (lane IO). SDS-12%PAGE was used. M, is shown at the left. Tryp, trypsin.

Prepro-a-factor Has a Ckauable Signal Sequence 621 1

then is the modification on the M, = 20,000 translocated nonglycosylated form of prepro-a-factor?

Asn-linked glycosylation can be inhibited in vitro by the addition of small peptides containing the Asn-X-Ser(Thr) glycosylation consensus site (Lau et al., 1983). These peptides are recognized by oligosaccharyltransferase and become gly- cosylated (Welpy et al., 1983), competing with the sites on prepro-a-factor. Translocation in the presence of an Asn- Phe-Thr peptide inhibited glycosylation of prepro-a-factor and increased the production of the M, = 20,000 form and the once glycosylated M, = 24,000 form (Fig. 1, lane 10). This result suggests that the glycosylated products also had the modification present on the M, = 20,000 form. This is sup- ported by the observation that after Endo H digestion, which removes all of the oligosaccharide except one GlcNAc (Tar- entino and Maley, 1974), the products approached the elec- trophoretic mobility of the M, = 20,000 form, not that of the M, = 19,000 primary translation product (Fig. 1, bne 9). However, it is possible that in the absence of glycosylation the modification resulting in the M , = 20,000 form was enhanced, that is, either the protein was glycosylated or it was modified resulting in the M, = 20,000 product.

Radwsequencing-It was noted previously (Hansen et al., 1986; Waters and Blobel, 1986) that when low membrane concentrations were used during translocation the M, = 20,000 form was a major product. In contrast, high membrane concentrations resulted predominantly in the M, = 32,000 glycosylated form. This behavior was attributed to a limiting pool of dolichol-linked oligosaccharides a t low membrane concentration (Hansen et al., 1986). In contrast, the efficient production of the M, = 20,000 form, even at low membrane concentration, suggested to us that the modification was the result of a catalytic process not limited by small pool size of a co-substrate, as glycosylation was. One such catalytic proc- ess that accompanies translocation is cleavage of the signal sequence (Blobel and Dobberstein, 1975).

We therefore performed amino-terminal radiosequencing of the in vitro products to determine if the signal sequence was cleaved. Prepro-a-factor was synthesized in the presence of 13H]alanine or [3H]proline and cotranslationally translo- cated into yeast microsomes. The products of the transloca- tion reaction were separated on SDS-l2%PAGE, blotted to polyvinylidine difluoride membrane, and subjected to gas phase protein sequencing directly off the membrane. Fig. 2A shows an aliquot of each reaction used for radiosequencing. Fig. 2B shows the 3H counts/min released at each cycle for the alanine- or proline-labeled M, = 19,000 primary transla- tion product, M, = 20,000 translocated nonglycosylated form and M, = 32,000 glycosylated form. The sequencing data for the M. = 19,000 primary translation product (Fig. 2B) agrees with the amino terminus predicted from the DNA sequence (Fig. 2C) (Kurjan and Herskowitz, 1982). The initiator me- thionine was present and therefore had not been removed by methionine aminopeptidase. This is in accordance with the rules describing removal of this residue (Tsunasawa et al., 1985; Flinta et al., 1986) which are based on the identity of the amino acid following methionine. In this case, the pres- ence of arginine in the second position precluded methionine aminopeptidase activity.

Both the M, = 20,000 and M, = 32,000 products yielded sequence data (Fig. 2B) that unambiguously located the amino terminus at position 20 of the precursor. Therefore, a 19- amino acid signal sequence was removed by signal peptidase (Fig. 2C). We refer to the M, = 20,000 form as pro-a-factor and the M, = 32,000 form as glycosylated pro-a-factor. Cleav- age between amino acid 19 and 20 corresponds to the predicted

cleavage site based on amino acid sequences recognized by signal peptidase (Perlman and Halvorson, 1983; von Heijne, 1983). These results confirm the original assumption (Kurjan and Herskowitz, 19821, based on the gene sequence, that the signal sequence of prepro-a-factor is cleaved.

It is interesting that pro-a-factor, which has a predicted molecular mass of 16,634 daltons, runs as a M , = 20,000 polypeptide in our SDS-lB%PAGE system. The reason for this is unknown, but may be due to the SDS binding and/or folding properties of the protein. It is probably this anomalous electrophoretic behavior which led to the previous belief that prepro-a-factor had an uncleaved signal sequence (Julius et al., 1984a). Julius et al. (1984a) pointed out, however, that if cleavage had no affect on electrophoretic mobility, it could have occurred and gone undetected. Indeed, we have found that in an SDS-18% gel system cleavage has no affect on migration (data not shown).

Cleavage of Prepro-a-factor by Heterologous Signal Pepti- dases-Fig. 3 shows that both purified eukaryotic signal pep- tidase (lanes 3 and 4 ) and purified prokaryotic signal (leader) peptidase (lanes 5 and 6) can cleave the signal sequence of prepro-a-factor. Furthermore, the appearance of a M, = 20,000 translocated nonglycosylated form of prepro-a-factor after translocation into E. coli inverted membrane vesicles (Fecycz and Blobel, 1987) or microsomes from Neurospora crassa (Addison, 1987) suggests that the signal sequence is cleaved in these systems. In addition, Candida species are capable of cleaving prepro-a-factor in an in vitro translation/ translocation system developed from that organism?

Prepro-a-factor Is Cleaved in Vivo-We were interested in determining whether the signal sequence was removed from prepro-a-factor in yeast in vivo. Since translocation of prepro- a-factor and transport from the endoplasmic reticulum is rapid (Julius et al., 1984a), we anticipated difficulty in detect- ing pro-a-factor or glycosylated pro-a-factor in wild type cells. We therefore used a secretory mutant, secl8, that is blocked in transport from the ER to the Golgi at the nonpermissive temperature (Novick et al., 1980) to "trap" the polypeptides in the ER. These mutant cells were incubated at either the permissive (22 "C) or the nonpermissive (37 " C ) temperature in the absence or presence of tunicamycin to inhibit glycosyl- ation (Duskin and Mahoney, 1982). The cells were then labeled with [35S]methionine, and polypeptides were immu- noprecipitated with anti-a-factor antiserum. At the permis- sive temperature without tunicamycin (Fig. M, lane 3) no cross-reactive material was seen in the M , = 19,000-32,000 range, presumably due to the rapid transit of a-factor through the secretory pathway (Julius et al., 1984a). If tunicamycin was present at the permissive temperature, however (Fig. 4A, lane 41, a significant amount of a-factor cross-reactive mate- rial was detected. This material migrated at M, = 20,000, corresponding to the cleaved form, pro-a-factor. This material probably accumulated without imposing the secretory block because tunicamycin slows the rate of secretion of a-factor from yeast cells (Julius et al., 1984a). When the secretory block was imposed by incubation at 37 "C in the absence of tunicamycin (Fig. 4.4, lane 5) a M, = 32,000 glycosylated form accumulated which co-migrated with glycosylated pro-a-fac- tor produced in vitro, which was shown to have undergone signal sequence cleavage. The presence of tunicamycin at the nonpermissive temperature (Fig. 4A, lane 6) efficiently inhib- ited glycosylation, resulting in accumulation of a M, = 20,000 form corresponding to the cleavage product pro-a-factor. These results suggest that the signal sequence of prepro-a- factor is cleaved in vivo because the immunoprecipitable

* T. Rapoport, personal communication.

6212 Prepro-a-factor Has a Cleavable Signal Sequence

20- - 19- -

1 2

B 300 19 kD - Alanine 19 kD - Proline

200

h 0

100

0 1 2 3 4 5 6 7 8 9 10111213141516 1 2 3 4 5 6 7 8 910111213141516

20 kD - Alanine 20 kD - Proline A Ala Pro 1000 - 32- 800 -

600 - 0 h

400 - 200 -

0 1 2 3 4 5 6 7 8 910111213141516 1 2 3 4 5 6 7 8 910111213141516

1200 1 32 kD - Alanine 3000 1 32 kD - Proline

1 2 3 4 5 6 7 8 910111213141516 1 2 3 4 5 6 7 8 910111213141516

CYCLE CYCLE

C met arg phe PRO ser ile phe thr ALA val leu phe ALA ALA ser ser ALA leu prepro 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

ALA ALA PRO val asn thr thr thr glu asp glu thr ALA gln ile PRO ALA prepro 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Pro 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6

FIG. 2. The signal sequence of prepro-a-factor is removed upon translocation in vitro in yeast. Prepro-a-factor was labeled with ('Hlalanine or [3H]proline and translocated in vitro. A, aliquots of the reaction were subjected to SDS-12%PAGE. M, is shown at the left. B, the polypeptides were electroblotted to polyvinylidine difluoride membrane, visualized by autoradiography, excised, and subjected to automated ban degradation. The counts/min released at each cycle for the indicated polypeptides is shown. The increasing background level in the M, = 19,OOO alanine sequence was probably due to an air leak in the sequenator during this determination. C. primary amino acid sequence of prepro-a-factor and of the cleavage product, pro-a-factor. The residues used for radiolabeling are cqpitalued.

Prepro-a-factor Has a Cleavable Signal Sequence 6213

Markers I( Canine E. coli

Memb - + - + - + SPase

32- _I_

1 2 3 4 5 6 FIG. 3. Purified eukaryotic and prokaryotic signal pepti-

dase can process prepro-a-factor. Prepro-a-factor translations in the absence (lane 1 ) or presence (lane 2 ) of yeast microsomes (Memb) are shown as markers. Prepro-a-factor synthesized in a wheat germ lysate was incubated in the absence (lane 3 ) or presence (lane 4 ) of purified canine signal peptidase complex or in the absence (lane 5 ) or presence (lane 6 ) of purified E. coli signal (leader) peptidase. SDS- 12%PAGE was used. M, is shown a t the left. SPase, signal peptidase.

A Markers u 22 22 37 37 Temperature

Memb - + - + - + Tunicamycin

GP- - 0

SDS- 124dPAGE

1 2 3 4 5 6

B GP- 0 0

SDS-18X.PAGE

4M Urea

PP- 0. P- - 0 -

1 2 3 4 5 6 FIG. 4. Prepro-a-factor is cleaved in vivo. A, prepro-a-factor

translations in the absence (lane 1 ) or presence (lane 2 ) of yeast microsomes (Memb) are shown as markers. sed8 cells were incubated at the permissive (lanes 3 and 4 ) or the nonpermissive (lanes 5 and 6 ) temperature, in the absence (lanes 3 and 5 ) or presence (lanes 4 and 6 ) of tunicamycin. The cells were labeled with [3SS]methionine, lysed, immunoprecipitated with anti-a-factor antiserum, and the precipitated material subjected to SDS-IL%PAGE. PP, prepro-a- factor; P, pro-a-factor, CP, glycosylated pro-a-factor. B, the same as A but SDS-lS%PAGE with 4 M urea was used, and the gel was run twice as long as required for the dye front to reach the bottom.

products co-migrate with polypeptides produced in vitro that have been shown to be cleaved.

We have found that use of SDS-lS%PAGE with 4 M urea, instead of SDS-lZ%PAGE results in the expected migration of pro-a-factor, that is, it migrates faster than prepro-a-factor (Fig. 4B, lanes 1 and 2).3 The expected migration of the

We have found (data not shown) that the migration of uncleaved and cleaved products on a 12% gel with 4 M urea is the same as on a 12% gel and that these polypeptides co-migrate on an 18% gel. Therefore, one should be cautious about using any one gel system to discern pro-a-factor from prepro-a-factor.

cleavage product in the urea/l8% gel suggests that the aber- rant mobility in the 12% gel is not due to another unidentified posttranslational modification.

Fig. 4B shows an aliquot of the same material as in Fig. 4A but electrophoresed on a SDS-18% gel with 4 M urea. The a- factor cross-reactive material produced in the presence of tunicamycin (lanes 4 and 6) again co-migrated with pro-a- factor produced in vitro. This characteristic behavior of the protein in the two-gel systems suggests that they are identical and therefore supports our conclusion that the signal sequence is removed in vivo.

Structure of the Oligosaccharides Present on Pro-a-factor in Vitro-Pro-a-factor was not the only uncharacterized poly- peptide present in our in vitro translocations. Often we ob- served an a-factor-specific product that had a higher M, than the fully glycosylated product (Waters and Blobel, 1986, Waters et al., 1986).

Transfer of Glc3Man9GlcNAc2 oligosaccharide cores from dolichol to protein (for review see Kornfeld and Kornfeld, 1985) is followed by trimming to Man8GlcNAc, (Byrd et al., 1982) by specific glucosidases and a mannosidase (for review see Kukuruzinska et al., 1987). This processing occurs in the yeast ER (Esmon et al., 1984). and the enzymes involved have been characterized (Kilker et al., 1981; Saunier et al., 1982; Jelinek-Kelly et al., 1985). We used specific glucosidase and mannosidase inhibitors to test if the high M, product evident in our translocation system was glycosylated pro-a-factor prior to glucose and mannose trimming. Deoxynojirimycin, a glucose analog, is an inhibitor of glucosidase I and I1 (Saunier et al., 1982), the enzymes which remove the 3 terminal glucose residues. Translocation of prepro-a-factor in the presence of this inhibitor resulted in a higher M , glycosylated form (Fig. 5, lane 2). These data suggest that after translocation and transfer of Glc3Man9GlcNAc2 cores in vitro, the cores are trimmed to Man9GlcNAcp. If deoxymannonojirimycin, a man- nosidase inhibitor (Fuhrmann et al., 1984), was present during translocation, a very slight, but reproducible, increase in M , was evident for glycosylated pro-a-factor (data not shown). This result suggests that the 1 mannose residue that is re- moved in vivo (Byrd et al., 1982) is also removed in our in vitro translocation system resulting in Man8GlcNAc2 oligo- saccharide structures.

DISCUSSION

This work shows that prepro-a-factor has a cleavable signal sequence and can be processed by yeast signal peptidase in

DNM - + a

32- 0

20- - 190,

1 2 FIG. 5. The core oligosaccharides on pro-a-factor are

trimmed in vitro. Prepro-a-factor was translocated in the absence (lane 1 ) or presence (lane 2 ) of deoxynojirimycin (DNM) to inhibit glucose removal from oligosaccharide cores. SDS-lZ%PAGE was used. M, is shown at the left.

6214 Prepro-a-factor Has a Cleavable Signal Sequence

uitro and in uiuo. This was formerly thought not to be the case. Furthermore, the signal sequence can be recognized and processed by E. coli, yeast, and canine signal peptidase.

This finding has a number of implications for the biosyn- thesis of a-factor. It has been proposed that the signal se- quence of prepro-a-factor, if uncleaved, would anchor the precursor in the membrane and may facilitate or be required for proteolytic processing and sorting leading to excision and secretion of a-factor (Julius et al., 1984a). Since the signal sequence is cleaved, and pro-a-factor lacks any other signifi- cant hydrophobic stretch, the precursor is probably not asso- ciated with the membrane. This is in agreement with the finding of Hansen et al. (1986), using an in vitro translocation system, that the translocated form of prepro-a-factor is alkali- extractable and therefore not an integral membrane protein. Thus, proteolytic processing and sorting must not require integral membrane association. Furthermore, the proposal that the prepro region might remain associated with the plasma membrane thereby performing some cellular function (Julius et al., 1984a) is unlikely given signal sequence cleavage.

Fusions to the prepro segment of prepro-a-factor have been used by many researchers to express and secrete homologous and heterologous proteins from yeast. The fusions contain the amino half of prepro-a-factor, including the segments needed for proteolytic excision of the protein of interest. These in- clude yeast invertase (Emr et al., 1983), human epidermal growth factor (Brake et al., 1984), human interferon-a1 (Singh et al., 1984), a consensus a-interferon and &endorphin (Bitter et al., 1984; Zsebo et al., 1986), mouse interleukin-2 (Miyajima et al., 1985), calf prochymosin (Smith et al., 1985), calcitonin (Zsebo et al., 1986), somatomedin-C ( E r s t , 1986), a fragment of human atrial natriuretic peptide (Vlasuk et al., 1986), and somatostatin (Green et al., 1986). Since signal sequence cleav- age is thought to be a co-translational event (Blobel and Dobberstein, 1975), it is likely that the signal sequence of prepro-a-factor is cleaved prior to synthesis and translocation of the carboxyl half of the molecule. If this is the case, it is likely that the signal sequence is also removed from the prepro-a-factor fusion proteins, although this has not been demonstrated.

We have used oligosaccharide-processing inhibitors to elu- cidate the structure of the core oligosaccharides present on pro-a-factor after in vitro translocation and glycosylation. Our data suggest that Glc3Man,GlcNAcz cores are initially transferred and then trimmed to MansGlcNAcz by ER resi- dent glucosidases and a mannosidase. This corresponds to the structure of oligosaccharides present on proteins in the ER in vivo (Esmon et aL, 1984).

The results presented here and previously (Hansen et al., 1986; Rothblatt and Meyer, 1986; Waters and Blobel, 1986) indicate the yeast in vitro translation/translocation system is capable of translation, translocation, signal peptide cleavage, glycosylation, and oligosaccharide-trimming reactions. These are most of the reactions known to be directly involved in secretory and integral membrane protein synthesis and trans- location prior to exit from the ER. Further biochemical analy- ais of this in vitro translocation system, in conjunction with genetic approaches available in yeast, should lead to a more complete understanding of the initial steps in the biogenesis of secretory and integral membrane proteins.

Acknowledgments-We would like to thank Dr. David Julius for the prepro-a-factor transcription vector, Dr. Howard Kaplan for oligosaccharide acceptor peptide, Dr. Dennis Shields for anti-a-factor antiserum, Dr. Randy Schekman for the see18 strain, and Dr. William Wickner for E. coli leader peptidase. We thank Dr. Yves Audigier for noting the usefulness of the urea gel system and Donna Atherton for microsequencing. We are grateful to Drs. Randolph Addison and Tom Rapoport for communicating unpublished results and to William Chirico, Jacques YaDeau, Gary Greenburg, and Raymond Deshaies for helpful discussions.

REFERENCES Achstetter, T., and Wolf, D. H. (1985) EMBO J. 4,173-177 Addison, R. (1987) J. BioL Chem. 262, 17031-17037 Sitter, G. A., Chen, K. K., Banks, A. R., and Lai, P-H. (1984) Proc. Natl. Acad.

Blobel, G., and Dobberstein, B. (1975) J. Cell Biol. 67, 835-851 Brake, A. J., Merryweather, J. P., Coit, D. G., Heberlein, U. A., Masiarz, F. R.,

Natl. Acad. Sci. U. S. A. 81,4642-4646 Mullenbach, G. T., Urdea, M. S., Valenzuela, P., and Barr, P. J. (1984) Proc.

Byrd, J. C., Tarentino, A. L., Maley, F., Atkinson, P. H., and Trimble, R. B.

Dmochowska, A., Dignard, D., Henning, D., Thomas, D. Y., and Bussey, H. (1982) J. BioL Chem. 267,14657-14666

Duskin, D., and Mahoney, W. C. (1982) J. Bwl. Chem. 267,3105-3109 (1987) Cell 60, 573-584

Emr, S. D., Schekman, R., Flessel, M. C., and Thorner, J. (1983) Proc. Natl.

Emter, O., Mechler, B., Achstetter, T., Muller, H., and Wolf, D. H. (1983)

Ernst, J. F. (1986) DNA 6,483-491 Esmon, B., Esmon, P. C., and Schekman, R. (1984) J. BioL Chem. 269,10322-

Sci. U. S. A. 81,5330-5334

Acad. Sci. U. S. A. 80,7080-7064

Biochem. Bwphys. Res. Commun. 116,822-629

10327 Evans, E. A., Gilmore, R., and Blobel, G. (1986) Proc. Natl. Acad. Sci. U. S. A.

Fecycz, I. T., and Blobel, G. (1987) Proc. Natl. Acad. Sei. U. S. A. 84, 3723-

Flinta, C., Persson, B., Jornvall, H., and von Heijne, G. (1986) Eur. J . Bioehem.

Fuhrmann, U., Bause, E., Legler, G., and Ploegh, H. (1984) Nature 307, 755-

83,581-585

3727

164,193-196

7SR Green. R.. Schaber. M. D.. Shields. D.. and Kramer. R. (1986) J. BioL Chem. $Si, 7558-7565 ' I , , , . _

Hansen, W., Garcia, P. B., and Walter, P. (1986) Cell 46, 397-406 Hicks, J. B., and Herskowitz, I. (1976) Genetles 83,245-258 Jelinek-Kelly, S., Akiyama, T., Saunier, B., Tkacz, J. S., and Herscovics, A.

Julius, D., Blair, L., Brake, A., Sprague, G., and Thorner, J. (1983) Cell 32, (1985) J. Blol. Chem. 260,2253-2257

Julius, D., Schekman, R., and Thorner, J. (1964a) Cell 36,309-318 Julius, D., Brake, A,, Blair, L., Kunisawa, R., and Thorner, J. (1984b) Cell 37,

Kilker, R. D., Saunier, B., Tkacz, J. S., and Herscovics, A. (1981) J. Biol. Chem.

Kornfeld, R., and Kornfeld, S. (1985) Annu. Reu. Biochem. 64,631-664 Kukuruzinska, M. A,, Bergh, M. L. E., and Jackson, B. J. (1987) Annu. Reu.

Kujan, J. and Herskowitz, I. (1982) Cell 30, 933-943 Lau, J. T.'Y., Welpy, J. K., Shenhagamurthi, P., Naider, F., and Lennarz, W.

839-852

1075-1089

266,5299-5304

Bkxhem. 66,915-944

J. (1983) J. BioL Chem. 268.15255-15260 Matsudaira, P. (1987) J. BWL Chem. 262,10035-10038 Miyajima, A., Bond, M. W., Otsu, K., Arai, K., and Arai, N. (1985) Gene (Amst.) 2-7 165-1Gl

Novick, P., Field, C., and Schekman, R. (1980) Cell 21,205-215 Perlman, D., and Halvorson, H. 0. (1983) J. Mol. Biol. 167,391-409 Rothblatt, J. A., and Meyer, D. (1986a) CeU 44,619-628 Rothblatt. J. A,. and Mever. D. (1986b) EMBO J. 6,1031-1036

-..( "I "*

Saunier, B., Kiiker, R. D., Jr., Tkacz, J. S., Quaroni, A., and Herscovics, A.

SinKh, A., Lugovoy, J. M., Kohr, W. I., and Perry, L. J. (1984) Nucleic Acids (1982) J. Biol. Chem. 267.14155-14161

Smith, R. A., Duncan, M. J., and Moir, D. T. (1985) Science 229,1219-1224 Sprague, G. F., Jr., Blair, L., and Thorner, J. (1983) Annu. Reu. Microbial. 37,

Ris. 12,8927-8938

Stotzler, D., Kiltz, H-H., and Duntze, W. (1976) Eur. J. Biochern. 69,397-400 Tarentino, A. L., and Maley, F. (1974) J. BioL Chem. 249,811-817 Tsunasawa, S., Stewart, J. W., and Sherman, F. (1985) J. Biol. Chem. 260,

Vlasuk, G. P., Bencen, G. H., Scarborough, R. M., Tsai, P-K., Whang, J. L., 5382-5391

Maack, T., Camargo, M. J. F., Kirsher, S. W., and Abraham, J. A. (1986) J. Riol Chem. 261.4789-4796

623-660

"" ~

von Heijne, G. (1983) Eur. J. hiochem. 133, 17-21 Waters, M. G., and Blobel, G. (1986) J. Cell. Biol. 102,1543-1550 Waters, M. G., Chirico, W. J., and Blobel. G. (1986) J . Cell Blol 103, 2629-

~ ~ " I ~ ~~

2636 Welpy, J. K., Shenbagamurthi, P., Lennarz, W. J., and Naider, F. (1983) J.

Biol. Chem. 268, 11856-11863 Zsebo, K. M., Lu, H-S., Fieschko, J. C., Goldstein, L., Davis, J., Duker, K.,

Suggs, S. V., Lai, P-H., and Bitter, G. A. (1986) J. Biol. Chem. 261, 5858- 5865

Zwizinski, C., and Wickner, W. (1980) J. Bwl. Chem. 266, 7973-7977