Ribosomal regulation of polytopic topogenesis

2545

Journal of Cell Science 113, 2545-2555 (2000) Printed in Great Britain © The Company of Biologists Limited 2000 JCS1333

Role of the ribosome in sequence-specific regulation of membrane targeting and translocation of P-glycoprotein signal-anchor transmembrane segments Jian-Ting Zhang*, Ernest Han and Yang Liu Department of Pharmacology and Toxicology, IU Cancer Center and Walther Oncology Center, Indiana University School of Medicine, Indianapolis, IN 46202, USA *Author for correspondence (e-mail: [email protected])

Accepted 25 April; published on WWW 22 June 2000

SUMMARY It is thought that the topology of a polytopic protein is generated by sequential translocation and membrane integration of independent signal-anchor and stop-transfer sequences. Two well-characterized cell-free systems (rabbit reticulocyte lysate and wheat germ extract) have been widely used to study the biogenesis of secretory and membrane proteins, but different results have been observed with proteins expressed in these two different systems. For example, different topologies of Pglycoprotein (Pgp) were observed in the two systems and the cause was thought to be the source of ribosomes. To understand how the ribosome is involved in dictating membrane translocation and orientation of polytopic proteins, individual signal-anchor sequences of Pgp were dissected and examined for their membrane targeting and translocation in a combined system of wheat germ ribosomes (WGR) and rabbit reticulocyte lysate (RRL). Addition of wheat germ ribosomes to the rabbit

reticulocyte lysate translation system can enhance, reduce, or have no effect on the membrane targeting and translocation of individual Pgp signal-anchor sequences, and these effects appear to be determined by the amino acid residues flanking each signal-anchor. Ribosomes regulate the membrane targeting and translocation of Pgp signal-anchors in a polytopic form differently from the same signal-anchors in isolation. Furthermore, we demonstrated that ribosomes regulate the membrane targeting and translocation of each signal-anchor cotranslationally and that this activity of ribosomes is associated with the 60S subunit. Based on this and previous studies, we propose a mechanism by which ribosomes dynamically dictate the membrane targeting and translocation of nascent polytopic membrane proteins.

INTRODUCTION

more than one topological form, and only one particular transmembrane form (CtmPrP) is responsible for neurodegenerative diseases (Hegde et al., 1998a; Lopez et al., 1990; Yost et al., 1990). Colicin Ia is another example with two topological structures, which interchange when the protein changes between closed and open channel states (Slatin, 1994). Two well-established cell-free systems (rabbit reticulocyte lysate and wheat germ extract) have been widely used to study the topogenesis of membrane proteins in general. However, different topologies were found when proteins were expressed in these two different systems (Lopez et al., 1990; Spiess et al., 1989), which was suggested to be due to the existence of a cytoplasmic factor in rabbit reticulocyte or differences in the ribosomes of these systems (Lopez et al., 1990; Spiess et al., 1989). Using Pgp as a model protein to investigate the biogenesis of polytopic proteins, we also showed that Pgp formed different topologies when translated in rabbit reticulocyte lysate and in wheat germ extract (Wang et al., 1997; Zhang and Ling, 1995). This observation was shown to be due to the difference between the ribosomes of these two systems. It is possible that mammalian and wheat germ ribosomes differ in directing the formation of a polytopic topology.

P-Glycoprotein (Pgp) is a membrane transporter that spans the membrane multiple times (polytopic) and is responsible for multidrug resistance observed in cancer cells (Gottesman et al., 1995; Gottesman and Pastan, 1993; Ling, 1997). Pgp consists of a tandem repeat of a unit containing a transmembrane (TM) domain followed by an ATP-binding domain. Thus, it belongs to the ATP-binding cassette (ABC) or traffic ATPase membrane transport superfamily (Doige and Ames, 1993; Higgins, 1992). Recent extensive studies on the topological folding and biogenesis of Pgp suggested that it might have more than one topological structure (Bibi and Beja, 1994; Cianfriglia et al., 1996; Poloni et al., 1995; Romagnoli et al., 1999; Skach et al., 1993; Zhang and Ling, 1991). This feature of multiple topologies, although still controversial (Kast et al., 1995; Loo and Clarke, 1995), may be very important to the understanding of the functional mechanism of Pgp in transporting multiple unrelated substrates. This notion of multiple topological forms relating to function was exemplified by prion protein (PrP) and by bacterial toxin colicin Ia. It was shown that PrP has

Key words: Topogenesis, P-glycoprotein, Ribosome, Signal-anchor

2546 J.-T. Zhang, E. Han and Y. Liu It is thought that each signal-anchor and stop-transfer sequence is very important for the topogenesis of a polytopic protein (e.g. Pgp) (Blobel, 1980; Hartmann et al., 1989; Lipp et al., 1989; Skach and Lingappa, 1993; Wessels and Spiess, 1988). In this study, we dissected and examined each individual putative signal-anchor sequence of Pgp in order to further understand how ribosomes direct the topological folding of polytopic membrane proteins using a combined system of wheat germ ribosome and rabbit reticulocyte lysate. Surprisingly, we found that ribosomal regulation of membrane targeting and translocation of Pgp signal-anchors was specific to the sequence. The elements in Pgp signal-anchor sequence responsible for this observation appear to be the amino acid residues flanking each signal-anchor. Furthermore, the membrane targeting and translocation of signal-anchors in a polytopic protein were regulated by ribosomes differently from the signal-anchors in isolation. Finally, we demonstrated that ribosomes regulate the membrane targeting and translocation of each signal-anchor cotranslationally and that this activity of ribosomes is associated with the 60S ribosomal subunit.

MATERIALS AND METHODS Materials pGEM-4z plasmid, SP6 RNA polymerase, RNasin, ribonucleotides, RQ1 DNase, rabbit reticulocyte lysate (RRL) and pGEM-T Easy vector were obtained from Promega. Dog pancreatic microsomal membranes were prepared as previously described (Walter and Blobel, 1983). [35S]methionine and Amplify were purchased from New England Nuclear and Amersham Corp., respectively. m7G(5′)ppp(5′)G cap analog was obtained from Pharmacia LKB Biotechnology Inc. Peptide N-glycosidase F (PNGase F) and restriction enzymes were obtained from Boehringer Mannheim, New England Biolabs, or Promega. All other chemicals were obtained from Sigma or Fisher Scientific. Isolation of wheat germ ribosomes (WGR) and WGR subunits Wheat germ ribosomes (WGR) and the 40S and 60S subunits were prepared using the method as previously described with modifications (Montesano and Glitz, 1988; Treadwell et al., 1979). Briefly, 15 g wheat germ was ground together with 15 g acid-washed sea sand for 10 minutes in a mortar on ice and then mixed with 150 ml of buffer A (90 mM KCl, 4 mM magnesium acetate, 2 mM CaCl2, 6 mM βmercaptoethanol, 10% glycerol). After 15 minutes of centrifugation at 15,000 rpm in a Beckman JA25.5 rotor, the supernatant was decanted and filtered through two layers of cheesecloth. The filtrate was then mixed with 0.1 volume of 0.1 M magnesium acetate and 0.1 volume of 1 M Tris-HCl, pH 7.8, before being centrifuged again for 15 minutes. The upper seven-eighths of the supernatant was collected and layered over a 1.5-ml pad of 20% sucrose in buffer A and centrifuged in a Beckman 55Ti rotor for 2 hours at 54,000 rpm. The resulting pellet was resuspended in buffer B (20 mM Hepes-KOH, pH 7.6, 150 mM KCl, 5 mM magnesium acetate, 6 mM βmercaptoethanol, 10% glycerol), divided into portions and stored at −70°C. To isolate 40S and 60S ribosomal subunits, 0.13 ml of the ribosome preparation was layered onto a 12 ml 15%-30% (w/v) linear sucrose gradient in buffer C (40 mM KCl, 80 µM magnesium acetate, 3 mM β-mercaptoethanol, 10 mM Tris-HCl, pH 7.8) and centrifuged for 4 hours at 40,000 rpm in a Beckman SW41Ti rotor at 10°C. Fractions of 80-130 µl each were collected and the OD at 260 nm of each fraction was determined. Fractions containing 40S and 60S subunits

were pooled and the Mg2+ concentration was raised to 8 mM by addition of 1 M magnesium acetate. Each pooled fraction was then diluted to 12 ml with buffer D (40 mM KCl, 8 mM magnesium acetate, 3 mM β-mercaptoethanol, 2 mM Tris-HCl, pH 7.8) and centrifuged overnight at 40,000 rpm in a Beckman SW55Ti rotor. The clear and colorless pellet was gently rinsed and resuspended with buffer E (80 mM KCl, 8 mM magnesium acetate, 8 mM βmercaptoethanol, 4 mM Tris-HCl, pH 7.8) and stored at −70°C with the addition of 1 volume of glycerol. Treatment of WGR with high salt was performed using a previously described procedure (Spremulli et al., 1977). Briefly, The ribosome preparation in buffer B was diluted with an equal volume of buffer B containing 1.0 M KCl and incubated on ice for 30 minutes, followed by centrifugation at 42,000 rpm (Beckman TLS55 rotor) for 5 hours. The top three-quarters of the supernatant were dialyzed against buffer B containing 50 mM KCl for 3 hours and stored at –70°C. The remaining supernatant was discarded. The washed ribosome pellet was resuspended in buffer B containing 50 mM KCl and stored at –70°C. Construct engineering Pgp-TM5R, −TM9R and −TM11R of hamster Pgp were constructed by PCR. cDNAs encoding TM5, TM9 and TM11 of hamster Pgp were amplified using two primers each. The primers used for TM5R were 5′-GTCGTGATGGAGGGAGGACAAAAG-3′ (A265a) and 5′-CTTCAAGATCTGTCCAATAGAATA-3′ (B326a); for TM9R, 5′-GTCGTGATGGAGTTTGATAACCCT-3′ (A801a) and 5′-TATTTTCAAACTTCT-3′ (B13); for TM11R, 5′-GTCGTGATGGAGTATGCCCAGAGC-3′ (A951a) and 5′-TGGCTTTGGCA-3′ (B9). The sense primers all contain a Kozak initiation codon ATG (underlined). The PCR products were cloned into pGEM-T Easy vectors. The cDNA inserts in pGEM-T Easy vectors were then released by digestion with EcoRI in combination with BglII (for TM5), HincII (for TM9) or XmnI (for TM11) and purified. The purified inserts were then ligated into a reporter-gene-containing pGEM-4z vector linearized with EcoRI and BglII. For TM9 and TM11, the EcoRI and BglII sites in pGEM-4z were blunted. The reporter in the pGEM-4z vector was derived from an ATP-binding domain of Pgp and was used successfully in our previous studies (Chen and Zhang, 1996; Zhang, 1996; Zhang et al., 1998, 1993). To construct Pgp-TM5R(L), a cDNA fragment encoding TM5 and its flanking amino acids were released from TM5R by digestion with EcoRI and SspI and purified. The cDNA fragment, together with a reporter-encoding cDNA digested with RsaI and HindIII, was then ligated into EcoRI/HindIII-digested pGEM-4z vector. The resulting construct was named Pgp-TM5R(L). To engineer Pgp-3TM11R and Pgp-11TM3R, a two-step PCR was employed to swap domains between the mutant Pgp-TM3R (Zhang et al., 1998) and Pgp-TM11R. In the first PCR, cDNA encoding the amino acid sequence at the amino-terminal side of TM3 and TM11 was amplified separately. The primers used for TM3 were SP6 vector primer and 5′-GATCCCAAAGACTTTGTCACCAAT-3′ (B186a) and for TM11 were 5′-GAAGAAAGCACACATTGGAATGTT-3′ (A190a) and 5′-CACTTTTGCCAACCAG-3′ (B50). The PCR products generated were then purified and used as primers, together with T7 vector primer (for TM3) or SP6 vector primer (for TM11), to amplify TM11 or TM3 with the reporter-coding sequence. The second PCR products were then cloned into pGEM-T Easy vector, followed by subcloning into a vector containing the reporter sequence. Pgp-N2,5R and Pgp-N2,11R were also engineered using a two-step PCR. The first PCR was performed using Pgp-N3 (Zhang et al., 1993) as template and universal SP6 primer and specific primers; for PgpN2,5R, 5′-GGCTTCTTCTAAATTGGTGTTGAGCTC-3′ (B612TM5), and for Pgp-N2,11R, 5′-GCAAGCTCTGGGCATAGGTGTTGAGCTC-3′ (B612TM11), to amplify the amino-terminal domain of Pgp and to create the joint between TM2 and TM5 or TM11. The PCR products generated were then purified and used as primers, together with TM7 vector primer, to amplify TM5 or TM11 with the reporter-

Ribosomal regulation of polytopic topogenesis 2547 encoding sequence. The second PCR products were then cloned into pGEM-T Easy vector, followed by subcloning into a vector containing the reporter sequence. All DNA constructs were sequenced to confirm the correct fusion, truncation, and to eliminate potential PCR-derived mutations. In vitro transcription and translation In vitro transcription was performed as previously described (Zhang and Ling, 1991). Cell-free translations in RRL, proteolysis/membrane protection assay, limited endoglycosidase treatment, isolation of membrane fractions by centrifugation, as well as analysis using SDSPAGE and fluorography, were performed as previously described (Zhang et al., 1993). Translations in RRL or ribosome-depleted RRL with added WGR or WGR subunits were performed as described by Wang et al. (1997). The images were digitized using a HP ScanJet 6100C and Adobe Photoshop 4.0. After calibrated from the background, the density of glycosylated and unglycosylated protein bands was determined using a Metamorph. The net density of each protein band was then used to calculate the effects of WGR on the generation of the glycosylated protein using the following formula: % effect = 100× {[Gly+WGR÷(Gly+Ungly)+WGR] – [Gly−WGR÷ (Gly+Ungly)−WGR]} ÷ [Gly−WGR÷(Gly+Ungly)−WGR] , where Gly and Ungly represent the glycosylated and unglycosylated proteins, respectively. The subscripts –WGR and +WGR represent proteins translated in the absence and presence of wheat germ ribosomes, respectively.

RESULTS The Nin-Cout membrane translocation of TM3 and TM4 of Pgp in RRL can be reduced in the presence of WGR Previously, we have shown that TM3 of a truncated Chinese hamster Pgp molecule (Pgp-N3) translated in RRL could have two membrane orientations, as shown in Fig. 1C (Zhang et al., 1993). The orientation with an extracellular glycosylation reporter had an additional oligosaccharide chain attached to the reporter and thus had a slower mobility on SDS-PAGE than

the one with a cytoplasmic reporter. Orientations of the glycosylation reporter can also be determined by its resistance (extracellular or rough microsome (RM) lumenal location) or sensitivity (cytoplasmic location) to protease digestion. Hence, the glycosylation of the reporter domain serves as an indicator for the orientation of the carboxyl terminus of the protein. We have also shown previously that production of the Pgp-N3 topology with an extracellular glycosylation reporter was dramatically reduced by addition of wheat germ ribosomes (WGR) to the RRL translation system (Wang et al., 1997; Zhang and Ling, 1995) (see also Fig. 1A, lanes 1-2). A possible explanation for this observation is that TM3 of Pgp synthesized in the presence of WGR cannot function as a signal-anchor sequence to reinitiate efficiently the Nin-Cout membrane translocation. To test this hypothesis, we used a construct that has only TM3 (mutant Pgp-TM3R) followed by a carboxylterminal glycosylation reporter. We also used a construct that has only TM4 (Pgp-TM4R). TM3 and TM4 in these constructs have been shown previously to be able to function as a signalanchor sequence to initiate de novo Nin-Cout membrane translocation in an RRL translation system (Zhang et al., 1998). As shown in Fig. 1A, translation of Pgp-TM3R and PgpTM4R in RRL both generated glycosylated proteins (indicated by asterisks in lanes 3 and 5), which represent proteins with Nin-Cout membrane orientation (Fig. 1C) (see also Zhang et al., 1998). However, addition of WGR to the translation drastically reduced the production of these glycosylated proteins with NinCout membrane orientation (Fig. 1A, lanes 4 and 6) (see also Fig. 1B). Thus, we conclude that TM3 and TM4 of Pgp in the presence of WGR cannot function as signal-anchor sequences to initiate successful membrane targeting and translocation in the Nin-Cout orientation in the RRL translation system. The Nin-Cout membrane translocation of different signal-anchor TM segments of Pgp was affected differently by WGR The above interesting observation brings us to the question of whether the effect of WGR is specific to TM3 and TM4 of Pgp

Fig. 1. Effect of WGR on membrane translocation and orientation of Pgp-N3, Pgp-TM3R and Pgp-TM4R. (A) Effect of WGR on the membrane translocation of Pgp-N3, Pgp-TM3R and Pgp-TM4R. Translation of Pgp-N3, Pgp-TM3R (mutant) and Pgp-TM4R was performed with RRL in the absence (lanes 1,3,5) or presence (lanes 2,4,6) of WGR. Membrane-associated proteins were isolated by centrifugation and analyzed by SDS-PAGE. Protein products indicated by asterisks represent proteins with a glycosylated reporter. (B) Quantification of the effect of WGR on the production of the glycosylated proteins. The effect of WGR on the production of glycosylated proteins in A was determined as described in Materials and Methods. (C) Topological folding of Pgp-N3, Pgp-TM3R and Pgp-TM4R. Transmembrane (TM) segments are shown as boxes. Oligosaccharide chains are indicated by branched symbols. The carboxyl-terminal reporter is represented by a solid ellipse. The positions of marker proteins (kDa) in A are shown.

2548 J.-T. Zhang, E. Han and Y. Liu since these two segments are involved in generating the alternative topologies for Pgp (Zhang et al., 1998, 1993, 1995). To answer this question, we engineered the glycosylation reporter to TM1, TM5, TM7, TM9 and TM11 of hamster Pgp to generate Pgp-TM1R, -TM5R, -TM7R, -TM9R and -TM11R (Fig. 2A). These TM segments were expected to function as putative signal-anchor sequences to initiate Nin-Cout membrane translocation, according to the sequential model of membrane targeting and translocation by signal-anchor and stop-transfer sequences for polytopic membrane proteins (Blobel, 1980). To determine whether WGR affected the Nin-Cout membrane translocation of these putative signal-anchor TM segments, we first needed to demonstrate that these TM segments can function as signal-anchor sequences to initiate the Nin-Cout membrane translocation. Previously, we have shown that TM7 failed to initiate the Nin-Cout membrane translocation (Han and Zhang, 1998). Similar findings were obtained with TM9 (data not shown). Thus, TM7 and TM9 could not be used for this study.

TM1, on the other hand, could function as a signal-anchor sequence to initiate the Nin-Cout membrane translocation (Han and Zhang, 1998) and thus was used for this study. TM5 and TM11 also have signal-anchor activity to initiate Nin-Cout membrane translocation, and generated products with glycosylated reporters in the presence of RM (Fig. 2B, compare lanes 1 and 4 with 2 and 5, respectively). This observation was confirmed by endoglycosidase PNGase F treatment (Fig. 2B, lanes 3 and 6) and proteinase K digestion of the membraneassociated proteins, which produced membrane-protected and glycosylated reporters (Fig. 3A, lanes 1-2 and 3B, lanes 7-8). We then tested whether the Nin-Cout membrane translocation of TM1, TM5 and TM11 was affected by WGR. As shown in Fig. 2C, generation of the glycosylated Pgp-TM1R (compare lane 1 with 2) in RRL was decreased in the presence of WGR, although the extent of decrease was not as great as that of PgpTM3R and Pgp-TM4R (Fig. 1A, lanes 3-6) (see also Fig. 2D). Interestingly, generation of the glycosylated protein of Pgp-

Fig. 2. Effect of WGR on membrane translocation and orientation of Pgp-TM1R, Pgp-TM5R, Pgp-TM11R and Pgp-TM5R(L). (A) Schematic representation of Pgp-TM1R, Pgp-TM5R, Pgp-TM11R and Pgp-TM5R(L). Partial amino acid sequence at the amino-terminal end and flanking the TM segment (numbered boxes) of each fusion construct are shown using single letter code. The branched symbol indicates the potential glycosylation sites. Glycosylation reporter is shown by a solid oval symbol. Note that the reporter in Pgp-TM5R is slightly smaller than that in other constructs. (B) Nin-Cout membrane translocation of Pgp-TM5R, Pgp-TM11R and Pgp-TM5R(L). Translation of Pgp-TM5R, Pgp-TM11R, and Pgp-TM5R(L) was performed in RRL in the absence (lanes 1,4,7) or presence (lanes 2,5,8) of RM and nascent proteins were analyzed by SDS-PAGE. Samples in lanes 3,6,9 represent endoglycosidase PNGase F treatment of membrane-associated proteins. (C) Effect of WGR on Nin-Cout membrane translocation of Pgp-TM1R, Pgp-TM5R, Pgp-TM11R and Pgp-TM5R(L). Translations of Pgp-TM1R, Pgp-TM5R, PgpTM11R, and Pgp-TM5R(L) were performed in RRL in the absence (lanes 1,3,5,7) or presence (lanes 2,4,6,8) of WGR. Membrane-associated proteins were isolated by centrifugation and analyzed by SDS-PAGE. The protein products indicated by asterisks represent proteins with a glycosylated reporter. (D) Quantification of the effect of WGR on the production of the glycosylated proteins. The effect of WGR on the production of glycosylated proteins in C was determined as described in Materials and Methods. (E) Schematic representation of the folding of Pgp-TM1R, Pgp-TM5R, Pgp-TM11R and Pgp-TM5R(L) with a glycosylated reporter. Note that Pgp-TM1R has three additional oligosaccharide chains in the loop linking TM1 and the glycosylation reporter. The positions of marker proteins (kDa) in B and C are shown.

Ribosomal regulation of polytopic topogenesis 2549 Fig. 3. Proteinase K treatment of Pgp-TM5R, PgpTM5R(L), Pgp-TM11R, Pgp-3TM11R and Pgp11TM3R. Membrane-associated nascent PgpTM5R (A) and Pgp-3TM11R, Pgp-11TM3R, PgpTM11R and Pgp-TM5R(L) (B) were isolated by centrifugation and treated with proteinase K in the absence (lanes 1,2,4,5,7,8,10,11) or presence (lanes 3,6,9,12) of Triton X-100. The membraneprotected fragments were pelleted by centrifugation and analyzed by SDS-PAGE. Samples in lanes 2, 5,8,11 were further treated with endoglycosidase following proteinase K digestion. The positions of marker proteins (kDa) are shown.

TM11R was increased (Fig. 2C, compare lane 5 with 6) while that of Pgp-TM5R was not significantly changed (Fig. 2C, compare lane 3 with 4) in the presence of WGR (see also Fig. 2D). Thus, all three possibilities (no effect, enhancement and reduction) for the WGR effect on the membrane translocation of putative Pgp signal-anchor TM segments exist, and this effect may be TM/signal-anchor specific. It is noteworthy that Pgp-TM1R has four glycosylation sites at the C-terminal side of TM1 and therefore glycosylation of all sites resulted in a much bigger glycosylated protein than Pgp-TM3R, Pgp-TM5R and Pgp-TM11R (see Fig. 2E). The length and size of the reporter sequence do not influence WGR effects on membrane translocation of nascent signal-anchor TM sequences expressed in RRL The reporter in Pgp-TM5R is slightly shorter than those in

other constructs (see Fig. 2A). To determine whether the longer reporter sequence influences the effect of WGR on the Nin-Cout membrane translocation of the signal-anchor TM segment, we constructed Pgp-TM5R(L) with a reporter that is bigger than that in Pgp-TM5R (Fig. 2A), but exactly the same as in PgpTM3R. As shown in Fig. 2B, Pgp-TM5R(L) could initiate the Nin-Cout membrane translocation to generate a glycosylated protein in the presence of RM (Fig. 2B, compare lane 7 with 8) as demonstrated by endoglycosidase treatment (Fig. 2B, lane 9). Proteinase K digestion showed that the reporter of PgpTM5R(L) was indeed located inside the RM lumen and was glycosylated (Fig. 3B, lanes 10-12). Addition of WGR to the translation of Pgp-TM5R(L) did not change significantly the production of proteins with Nin-Cout membrane orientation (Fig. 2C, compare lane 7 with 8), similar to that of TM5R (Fig. 2C, lanes 3-4) (see also Fig. 2D). Hence, we conclude that the size and sequence of the reporter do not interfere with the effect

Fig. 4. Importance of amino acid sequences flanking TM segments in ribosomal effect on TM membrane translocation. (A) Schematic representation of Pgp-TM3R, Pgp-TM11R, Pgp11TM3R and Pgp-3TM11R constructs. The amino acid sequences flanking TM3 and TM11 are indicated by solid and dashed lines, respectively, with partial amino acid sequences shown in single letter code. The numbered boxes and the branched symbol indicate the TM segment and potential glycosylation sites, respectively. Glycosylation reporter is shown by a solid oval symbol. (B) Nin-Cout membrane translocation of Pgp-11TM3R and Pgp-3TM11R. Translation of Pgp11TM3R and Pgp-3TM11R was performed in RRL in the absence (lanes 1,4) or presence (lanes 2,5) of RM and nascent proteins were analyzed by SDS-PAGE. Samples in lanes 3,6 represent endoglycosidase PNGase F treatment of membrane-associated proteins. (C) Effect of WGR on Nin-Cout membrane translocation of Pgp-11TM3R and Pgp-3TM11R in comparison with Pgp-TM3R and Pgp-TM11R. Pgp-TM3R, Pgp-11TM3R, Pgp-TM11R and Pgp-3TM11R were translated in RRL in the absence (lanes 1,3,5,7) or presence (lane 2,4,6,8) of WGR. Membrane-associated proteins were isolated by centrifugation and analyzed by SDS-PAGE. The protein products indicated by asterisks represent proteins with a glycosylated reporter. (D) Quantification of the effect of WGR on the production of the glycosylated proteins. The effect of WGR on the production of glycosylated proteins was determined as described in Materials and Methods.

2550 J.-T. Zhang, E. Han and Y. Liu of WGR on the generation of Nin-Cout orientation of Pgp signal-anchor TM segments. Amino acid sequences flanking signal-anchor TM segments are important for the WGR effect Next we were interested in determining whether the sequence flanking each signal-anchor TM segment is responsible for the TM/signal-anchor-specific ribosomal effect. To address this issue, we swapped the sequence at the amino-terminal side of Pgp-TM3R, of which the Nin-Cout membrane translocation was decreased, and Pgp-TM11R, of which the Nin-Cout membrane translocation was enhanced, by the presence of WGR in the RRL translation system. These fusion constructs (named Pgp11TM3R and Pgp-3TM11R) are shown schematically in Fig. 4A. Translation of both Pgp-11TM3R (Fig. 4B, lanes 1-2) and Pgp-3TM11R (Fig. 4B, lanes 4-5) generated glycosylated proteins in the presence of RM, indicating the production of proteins with Nin-Cout membrane orientation. This was confirmed by endoglycosidase treatment (Fig. 4B, lanes 3 and 6) and by proteinase K digestion of membrane-associated Pgp11TM3R and Pgp-3TM11R (Fig. 3B, lanes 1-6). Addition of WGR to the translation of Pgp-11TM3R produced slightly more glycosylated protein than the translation in the absence of WGR (Fig. 4C, compare lane 3 with 4). On the other hand, addition of WGR to the translation of Pgp-3TM11R produced slightly less glycosylated protein than the translation in the absence of WGR (Fig. 4C, compare lane 7 with 8). These are opposite to the WGR effect on Pgp-TM3R (Fig. 4C, compare lane 1 with 2) and Pgp-TM11R (Fig. 4C, compare lane 5 with 6), respectively (see also Fig. 4D). The effect of WGR on the generation of glycosylated proteins of Pgp-TM3R and PgpTM11R appears to be more dramatic (Fig. 4C, compare lanes 1 and 5 with 2 and 6, respectively). This observation suggests that the amino acid sequence flanking each signal-anchor TM segment is likely responsible, at least in part, for TM/signalanchor-specific WGR effects.

The effect of WGR on membrane translocation is different for a signal-anchor TM sequence in a polytopic form The above studies suggested that the membrane translocation and orientation of each individual signal-anchor TM segment translated in RRL was affected differently by WGR and this difference appeared to be determined by the amino acid sequences flanking the signal-anchor TM segment. To determine if WGR has the same effect on the membrane translocation of these signal-anchor TM segments in the presence of other TM segments (i.e. in a polytopic form), we studied two new constructs, Pgp-N2,5R and Pgp-N2,11R (Fig. 5), in comparison with Pgp-N3 (Fig. 1). These constructs contain both TM1 and TM2 followed by TM5, TM11 or TM3, respectively, with a C-terminal glycosylation reporter. We used TM5 and TM11 in comparison with TM3 because their membrane targeting and translocation in isolation was affected differently by ribosome (Nin-Cout translocation of TM3 was decreased, that of TM5 was not changed significantly, and that of TM11 was dramatically increased by WGR). As shown in Fig. 5A, translation of Pgp-N2,5R and Pgp-N2,11R in the absence of RM generated proteins of approx. 43 kDa (lanes 1 and 6, indicated by arrowheads). In the presence of RM, the translated protein shifted to approx. 54 kDa (Fig. 5A, lanes 2 and 7) due to the addition of four oligosaccharide chains, which were demonstrated by endoglycosidase treatment (data not shown). Thus, both Pgp-N2,5R and Pgp-N2,11R have their C-terminal reporter located in the RM lumen and glycosylated together with the loop linking TM1 and TM2 (Fig. 5C, model I). This was confirmed by the protection of glycosylated reporters in Pgp-N2,5R and Pgp-N2,11R from proteinase K digestion by membranes (indicated by arrows, Fig. 5A, lanes 3-5 and 8-10). When Pgp-N2,5R and PgpN2,11R were translated in RRL in the presence of WGR, an additional protein with a size approx. 3 kDa smaller than the major form was produced (indicated by arrowheads, Fig. 5B,

Fig. 5. Effect of ribosomes on membrane translocation of Pgp TM segments in a polytopic form. (A) Translation and treatment of Pgp-N2,5R and PgpN2,11R. Translation of Pgp-N2,5R (lanes 1-5) and Pgp-N2,11R (lanes 6-10) was performed in RRL in the absence (lanes 1,6) or presence (lanes 2,7) of RM and nascent proteins were analyzed by SDS-PAGE. Membrane-associated nascent proteins were isolated by centrifugation and treated with proteinase K in the absence (lanes 3,4,8,9) or presence of Triton X-100 (lanes 5,10). The membrane-protected fragments after digestion were pelleted and analyzed by SDS-PAGE. Samples in lanes 4,9 were further treated with endoglycosidase following proteinase K digestion. (B) Effect of WGR on Nin-Cout membrane translocation of Pgp-N2,5R and Pgp-N2,11R. Pgp-N2,5R and Pgp-N2,11R were translated in RRL in the absence (lanes 1,3) or presence (lanes 2,4) of WGR. Membrane-associated proteins were isolated by centrifugation and analyzed by SDS-PAGE. The protein products indicated by asterisks represent proteins with a glycosylated reporter (model I structure in C). (C) Possible topological folding for Pgp-N2,5R and Pgp-N2,11R. Model I structure has four oligosaccharide chains (branched symbols) with the glycosylation reporter in RM lumen whereas the model II structure has the reporter located in cytoplasm together with TM5 or TM11 and thus has only three oligosaccharide chains.

Ribosomal regulation of polytopic topogenesis 2551 lanes 2 and 4). Thus, it is very likely that the membrane translocation ability of TM5 and TM11 in Pgp-N2,5R and Pgp-N2,11R was reduced by adding WGR to the translation in RRL. This effect resulted in the generation of proteins with TM5 and TM11 and their C-terminal reporter in the cytoplasm (Fig. 5C, model II). This observation is similar to that of PgpN3 (Fig. 1A, lanes 1-2), but different from the observation with individual TM5 and TM11 where their Nin-Cout membrane translocation ability was not changed or was enhanced by WGR, respectively (Fig. 2C). Therefore, the effect of WGR on the membrane translocation of signalanchor TM segments translated in RRL not only depends on the sequence flanking each signal-anchor TM segment but also on the complexity of the membrane protein. The WGR activity affecting the membrane translocation of Pgp signal-anchor TM sequences is associated with the 60S subunit To rule out the possibility that the activity of WGR affecting

Fig. 7. Direct competition between WGR and RRL ribosome. (A) Translation of Pgp-TM3R in RRL in the presence of increasing amount of WGR. Pgp-TM3R was translated in RRL in the absence (lane 1) or presence of various amount of WGR (lanes 2-6). Membrane-associated proteins were then isolated by centrifugation and analyzed by SDS-PAGE. The protein product indicated by asterisk represents the protein with a glycosylated reporter. (B) Quantification of the effect of WGR on the production of the glycosylated proteins. The effect of WGR on the production of glycosylated proteins was determined as described in Materials and Methods. (C) Orientation of Pgp-TM11R translated in the absence of RRL ribosome. RRL translation system was stripped of ribosomes and used for translation of Pgp-TM11R without (lane 1) or with added RRL ribosome (RRR) (lane 2) and WGR (lane 3). The membrane-associated protein was isolated by centrifugation and analyzed by SDS-PAGE. Protein products indicated by an asterisk represent proteins with a glycosylated reporter.

Fig. 6. The effect of ribosomes on the membrane translocation of Pgp TM segments is associated with the 60S subunit. (A) Effect of WGR subunits on Nin-Cout membrane translocation of Pgp-TM11R. Pgp-TM11R was translated in RRL in the absence (lane 1) or presence of WGR (lane 2), 40S subunits (lane 3), 60S subunits (lane 4), combined 40S and 60S subunits (lane 5), salt-extracted WGR pellet (lane 6) or supernatant (lane 7). Membrane-associated proteins were isolated by centrifugation and analyzed by SDS-PAGE. The protein product indicated by asterisk represents the protein with a glycosylated reporter. (B) Fractionation of 40S and 60S of WGR. The subunits of WGR were isolated as described in Materials and Methods. The distribution of 40S and 60S subunits of WGR after centrifugation is shown relative to the position of undissociated 80S WGR.

the membrane translocation of Pgp signal-anchor TM sequences in RRL translation system is due to co-isolation of contaminating proteins with WGR, WGR was treated with high salt and the extracted fraction was separated from WGR by centrifugation. The extracted fraction was dialyzed to remove salt before being tested for its activity to affect the membrane translocation of Pgp signal-anchor TM segments in the RRL system. As shown in Fig. 6A, the activity to enhance Nin-Cout membrane translocation of Pgp-TM11R (indicated by *) appeared to remain associated with the WGR (lane 6) and it was not associated with the salt extract (lane 7). Thus, the activity affecting the membrane translocation of Pgp signalanchor TM segments is associated with WGR and not due to a protein coisolated with WGR. We next tested which subunit of WGR has the activity affecting the membrane translocation of Pgp signal-anchor TM sequences. The small (40S) and large (60S) subunits of WGR were dissociated and separated by centrifugation on a linear sucrose gradient (Fig. 6B). The pooled fractions containing

2552 J.-T. Zhang, E. Han and Y. Liu

Fig. 8. Model of ribosomal regulation of topology for monotopic and polytopic membrane proteins. Ribosome with a nascent peptide (A) docks onto the protein conducting channel (translocon) with the signal-anchor sequence translocating into the channel in Nin-Cout (B) or Nout-Cin (B′) orientations. For polytopic proteins, the ribosome releases from the translocon after both TM1 and TM2 translocate into the translocon and redocks after the third TM (signal-anchor) sequence emerges from the ribosome (E). However, the ribosome may not redock and thus result in an alternative structure (E′).

40S and 60S subunits were used to determine their effect on the Nin-Cout membrane translocation of Pgp-TM11R. As shown in Fig. 6A, the activity to enhance the Nin-Cout membrane translocation of Pgp-TM11R appears to remain with the 60S but not the 40S subunit (compare lanes 2, 3, 4 and 5). Thus, the 60S subunits, and not the 40S subunits, of WGR contain the activity that affects the membrane translocation of Pgp signal-anchor TM segments. The WGR effect on membrane translocation of Pgp signal-anchor TM segments in RRL was due to direct competition with RRL ribosomes for translation and membrane targeting The results in Fig. 6 indicate that the effect of WGR on membrane translocation of Pgp signal-anchor sequences may be due to direct competition of WGR with RRL ribosomes. To test this hypothesis, different amounts of WGR were added to the translation of Pgp-TM3R in RRL. As shown in Fig. 7A,B, the ratio of glycosylated to unglycosylated Pgp-TM3R decreased with increasing amounts of WGR added to the translation. This observation suggests that a direct competition of RRL by WGR likely affected the Nin-Cout membrane insertion by the signal-anchor sequence TM3. To further confirm the above conclusion, we stripped ribosomes from the RRL translation system by centrifugation and supplemented the system with WGR for translation. As shown in Fig. 7C, no translation products were produced from Pgp-TM11R in the ribosome-stripped RRL system (lane 1). Addition of the removed RRL ribosomes back to the stripped translation system restored its translation ability and both the Nin-Cout and Nout-Cin membrane orientations were produced for Pgp-TM11R (Fig. 7C, lane 2). However, addition of WGR to the ribosome-stripped RRL translation system produced proteins of Pgp-TM11R with only the glycosylated Nin-Cout membrane orientation (Fig. 6C, lane 3). No unglycosylated Pgp-TM11R was produced by WGR in the stripped RRL system. Thus, the effect of WGR on the orientation of Pgp signal-anchor TM sequences in the RRL translation system is likely due to their direct role in translating and targeting the nascent protein to RM.

DISCUSSION In this study, we investigated the role of ribosome in directing the membrane targeting and translocation of individual Pgp putative signal-anchor sequence (TM segments) using a combined system of RRL and WGR. We demonstrated that ribosomes regulate the membrane translocation and orientation of Pgp signal-anchor TM segments cotranslationally and that this activity of ribosomes is associated with the 60S subunit. The WGR regulation of membrane translocation and orientation of each Pgp signal-anchor TM segment was different depending on the flanking amino acid residues of the signal-anchor sequence and how many TM segments were present. The effect of WGR on the generation of glycosylated fusion proteins of putative Pgp signal-anchor sequences could also be due to their effect on the glycosylation of the reporter. However, this possibility is highly unlikely. Firstly, all four potential sites in the reporter of Pgp-TM1R were fully glycosylated in the presence of WGR (Fig. 2B). No partially glycosylated proteins were found. This is supported by the observation that translation of Pgp-TM11R by WGR in a ribosome-stripped RRL system did not generate any unglycosylated protein. Furthermore, no protected reporter lacking oligosaccharide chains was found after proteinase K digestion (Zhang and Ling, 1995), suggesting that there was no unglycosylated reporter in the RM lumen. These observations argue against the possibility that WGR inhibits glycosylation. Secondly, wheat germ extract is widely used for topology and biogenesis studies of many membrane and secretory proteins, and no change in glycosylation status of the nascent protein has been reported due to the use of wheat germ extract (e.g. see Lopez et al., 1990; Spiess et al., 1989). Thirdly, if WGR affected the glycosylation enzymes, it would be expected to either inhibit or enhance glycosylation and not the both. The results in this study suggested that the membrane orientation of TM11 is enhanced whereas that of TM3 and TM4 is decreased by the presence of WGR in the RRL translation system. Lastly, glycosylation occurs in RM lumen whereas WGR is in cytoplasm. The effect of WGR on glycosylation can

Ribosomal regulation of polytopic topogenesis 2553 only happen by affecting the translocation of reporters into the RM lumen and making them available to glycosylation in the RM lumen. Therefore, the change in production of glycosylated proteins by WGR is likely due to the effect of ribosomes on the membrane translocation and orientation of nascent proteins and not on the glycosylation per se. Since microsomes used in cell-free expression studies were generally derived from dog pancreatic endoplasmic reticulum, it is possible that TRAM (translocating chain-associated membrane protein) or other proteins such as ribosome receptors in the protein-conducting channel do not match perfectly with WGR to support membrane translocation of nascent proteins. Thus, the Nin-Cout membrane translocation of Pgp signal-anchor TM segments may be decreased by addition of WGR. However, this possibility is less likely because in the case of Pgp-TM11R the Nin-Cout membrane translocation was enhanced by the existence of WGR in the translation system, which argues for a favored match. It is more likely that mammalian and plant ribosomes contain different proteins that may be involved in regulating the interaction between nascent peptides, ribosomes and the protein-conducting channel, which in turn regulates the membrane translocation and orientation of nascent peptides. Previously, prion and artificially engineered proteins translated in RRL and WGE have been shown to form different topologies (Lopez et al., 1990; Spiess et al., 1989). It has been suggested that a cytoplasmic factor in RRL or the difference in ribosomes of these two systems may be responsible for the difference in generating different topologies (Lopez et al., 1990; Spiess et al., 1989). We have also observed that the NinCout membrane translocation of TM3 of CFTR could be reduced by addition of WGR (unpublished observation). Thus, the effect of WGR on the membrane translocation of TM segments is not only specific to Pgp sequences, but also to TM segments of other membrane proteins. From this study using the unique combination of WGR and RRL translation system, we conclude that ribosomes cotranslationally regulate the translocation and orientation of membrane proteins and this regulation is signal-anchor TM segment and its flanking sequence specific. The effect of WGR on topological formation of Pgp in RRL system is likely due to competition of WGR with RRL ribosome for translating Pgp signal-anchor TM sequences. That is, the nascent proteins translated by WGR behave differently from those translated by RRL ribosomes. It should be noted, however, that the effect of WGR could also be due to the formation of hybrid ribosomes with subunits from WGR and RRL. Addition of the large subunit (60S) of WGR alone to the RRL translation system affected the membrane translocation of Pgp signal-anchor TM segment, which demonstrates that the 60S ribosome contains the activity to regulate membrane protein translocation and hybrid ribosomes are formed. This is not surprising because the 60S subunit of ribosome is thought to interact with the protein conducting channels on ER membranes. Our results strongly suggest that ribosomes are involved in regulating the translocation and orientation of membrane proteins and that this regulation is specific to the sequence of nascent proteins. Based on our observation in this study, we propose a mechanism for the regulation of the membrane translocation of a nascent signal-anchor TM segment by ribosomes. As shown in Fig. 8, a signal-anchor sequence has

two options for inserting into the protein-conducting channel (Nin-Cout, stage B, or Nout-Cin, stage B′) as it emerges from the ribosome (stage A). Orientation of a bitopic membrane protein (with a single signal-anchor TM segment) is determined by the charged amino acid sequence flanking the TM (Beltzer et al., 1991; Hartmann et al., 1989; Parks and Lamb, 1991; Sipos and von Heijne, 1993), and it has been proposed that the positive charges flanking the signal-anchor TM sequence of the nascent peptide may interact with the negative charges on the cytoplasmic domains of the protein-conducting channel (Akita et al., 1990). In this study, we observed that ribosomes translating the nascent peptide may also be involved in regulating the orientation of bitopic membrane proteins. Ribosomes facilitate the generation of one orientation or another depending on the amino acid sequence flanking the signal-anchor TM of the nascent protein. How this is accomplished is currently unknown. Whether the charged amino acids flanking signal-anchor TM segments are involved in the ribosome’s regulation remain to be tested. It is, however, tempting to speculate that the process of ribosomal regulation of membrane translocation of nascent proteins likely occurs in the initial ribosome-docking step (Fig. 8, stage A to B or B′). At this step, mammalian and plant ribosomes may behave differently, depending on the sequence of the nascent protein, and thus target the protein to membranes with different orientations (e.g. more Nin-Cout orientation for TM11 and less for TM3 and TM4 translated by WGR compared to mammalian ribosomes). At this step, other trans-acting factors may also be involved in regulating the membrane orientation of nascent proteins, possibly by interacting with ribosome and the nascent protein (Hegde and Lingappa, 1999). In the absence of these trans-acting factors, prion is synthesized exclusively in a transmembrane form (CtmPrP form), associated with the development of neurodegenerative diseases. Involvement of these factors in the membrane translocation of signal-anchor TM segments could allow for enormous variation and diversity in topogenesis of nascent proteins (Hegde and Lingappa, 1999). Previously, it has been shown that ribosomes synthesizing a polytopic protein with three TM segments are bound to the protein-conducting channel at all times (Mothes et al., 1997). If this is true, we would not have observed the model II structure with Pgp-N3, Pgp-N2,5R and Pgp-N2,11R (Fig. 5C). In a previous study, we also showed that TM3 and TM4 of Pgp, in the presence of their preceding TM segments (TM1 and TM2), could not function as signal-anchor sequences to reinitiate membrane translocation in the wheat germ expression system (Zhang and Ling, 1995). We later demonstrated that the failure of TM3 and TM4 to reinitiate the membrane translocation was likely due to WGR, which could not support their Nin-Cout membrane translocation while translating them (Wang et al., 1997). Apparently with Pgp sequences, ribosome synthesizing Pgp-N3, Pgp-N2,5R or PgpN2,11R were released from the protein-conducting channel after TM1 and TM2 were translocated into the translocon (Fig. 8, stage D). At this step, a lumenal gate would close the protein-conducting channel to prevent leakage between cytoplasm and ER lumen (Liao et al., 1997). While continuing to synthesize the protein, the ribosome will redock the third TM segment as a signal-anchor sequence (e.g. TM3, TM5 or TM11 in Pgp-N3, Pgp-N2,5R and Pgp-N2,11R, respectively),

2554 J.-T. Zhang, E. Han and Y. Liu as it emerges from the ribosome (Fig. 8, stage E). Possibly, at this step WGR has less efficiency than RRL ribosomes to redock and reinitiate the membrane translocation of the third TM segment (signal-anchor sequence) and thus result in the alternative topology (Fig. 8, stage E′-F′). Other previous studies have also indicated the important role of ribosomes in membrane protein biogenesis. Studies on apolipoprotein B (apo B) (Hegde and Lingappa, 1996) suggested that the ribosome and the membrane-conducting channel interact tightly to ensure nascent secretory proteins are shielded from the cytoplasm during transfer into the endoplasmic reticulum. Also, TRAM in the protein-conducting channel has been suggested to be responsible for interacting with the ribosome (Hegde et al., 1998b). The ribosome has also been shown to go through several stages of docking to and releasing from the protein-conducting channel during the synthesis and translocation of an artificial membrane protein (Liao et al., 1997). The ribosome is a very important regulator for the integration of Pgp sequences into membranes (Borel and Simon, 1996). Membrane integration of nascent Pgp sequences could occur only after the synthesis of nascent peptide was complete and the ribosome has been released from nascent peptides. The difference between plant and mammalian ribosomes in translocating nascent membrane proteins provides us with an invaluable system to determine the role of the ribosome in cotranslational regulation of membrane protein translocation and folding. We believe that further characterization and comparison studies between mammalian and plant ribosomes will help elucidate the mechanism by which ribosomes interact with protein-conducting channels and trans-acting factors and direct membrane translocation and orientation of nascent proteins. Future studies using reconstitution of purified ribosomal proteins may help us identify the ribosomal proteins in the large 60S subunit that are responsible for the ribosomal regulation of translocation and orientation of membrane proteins. These studies are currently in progress in our laboratory. We would like to thank members of our laboratory for helpful discussion and suggestions for this study and Dr David Ohannesian for critical comments on the earlier version of this manuscript. This work was supported by the National Institutes of Health grants CA64539 and GM59475. J.T.Z. is a recipient of a Career Investigator Award from the American Lung Association. E.H. is a recipient of a predoctoral fellowship from the US Army Medical Research and Materiel Command (DAMD17-94-J-4080).

REFERENCES Akita, M., Sasaki, S., Matsuyama, S. and Mizushima, S. (1990). SecA interacts with secretory proteins by recognizing the positive charge at the amino terminus of the signal peptide in Escherichia coli. J. Biol. Chem. 265, 8164-8169. Beltzer, J. P., Fiedler, K., Fuhrer, C., Geffen, I., Handschin, C., Wessels, H. P. and Spiess, M. (1991). Charged residues are major determinants of the transmembrane orientation of a signal-anchor sequence. J. Biol. Chem. 266, 973-978. Bibi, E. and Beja, O. (1994). Membrane topology of multidrug resistance protein expressed in Escherichia coli. N-terminal domain. J. Biol. Chem. 269, 19910-19915. Blobel, G. (1980). Intracellular protein topogenesis. Proc. Natl. Acad. Sci. USA 77, 1496-1500. Borel, A. C. and Simon, S. M. (1996). Biogenesis of polytopic membrane

proteins: membrane segments assemble within translocation channels prior to membrane integration. Cell 85, 379-389. Chen, M. and Zhang, J. T. (1996). Membrane insertion, processing, and topology of cystic fibrosis transmembrane conductance regulator (CFTR) in microsomal membranes. Mol. Membr. Biol. 13, 33-40. Cianfriglia, M., Poloni, F., Signoretti, C., Romagnoli, G., Tombesi, M. and Felici, F. (1996). Topology of MDR1-P-glycoprotein as indicated by epitope mapping of monoclonal antibodies to human MDR cells. Cytotechnology 19, 247-251. Doige, C. A. and Ames, G. F. (1993). ATP-dependent transport systems in bacteria and humans: relevance to cystic fibrosis and multidrug resistance. Annu. Rev. Microbiol. 47, 291-319. Gottesman, M. M., Hrycyna, C. A., Schoenlein, P. V., Germann, U. A. and Pastan, I. (1995). Genetic analysis of the multidrug transporter. Annu. Rev. Genet. 29, 607-649. Gottesman, M. M. and Pastan, I. (1993). Biochemistry of multidrug resistance mediated by the multidrug transporter. Annu. Rev. Biochem. 62, 385-427. Han, E. S. and Zhang, J. T. (1998). Mechanism involved in generating the carboxyl-terminal half topology of P-glycoprotein. Biochemistry 37, 1199612004. Hartmann, E., Rapoport, T. A. and Lodish, H. F. (1989). Predicting the orientation of eukaryotic membrane-spanning proteins. Proc. Natl. Acad. Sci. USA 86, 5786-5790. Hegde, R. S. and Lingappa, V. R. (1996). Sequence-specific alteration of the ribosome-membrane junction exposes nascent secretory proteins to the cytosol. Cell 85, 217-228. Hegde, R. S. and Lingappa, V. R. (1999). Regulation of protein biogenesis at the endoplasmic reticulum membrane. Trends Cell Biol. 9, 132-137. Hegde, R. S., Mastrianni, J. A., Scott, M. R., DeFea, K. A., Tremblay, P., Torchia, M., DeArmond, S. J., Prusiner, S. B. and Lingappa, V. R. (1998a). A transmembrane form of the prion protein in neurodegenerative disease. Science 279, 827-834. Hegde, R. S., Voigt, S., Rapoport, T. A. and Lingappa, V. R. (1998b). TRAM regulates the exposure of nascent secretory proteins to the cytosol during translocation into the endoplasmic reticulum. Cell 92, 621-631. Higgins, C. F. (1992). ABC transporters: from microorganisms to man. Annu. Rev. Cell Biol. 8, 67-113. Kast, C., Canfield, V., Levenson, R. and Gros, P. (1995). Membrane topology of P-glycoprotein as determined by epitope insertion: transmembrane organization of the N-terminal domain of mdr3. Biochemistry 34, 4402-4411. Liao, S., Lin, J., Do, H. and Johnson, A. E. (1997). Both lumenal and cytosolic gating of the aqueous ER translocon pore are regulated from inside the ribosome during membrane protein integration [see comments]. Cell 90, 31-41. Ling, V. (1997). Multidrug resistance: molecular mechanisms and clinical relevance. Cancer Chemo. Pharmacol. 40, S3-8. Lipp, J., Flint, N., Haeuptle, M. T. and Dobberstein, B. (1989). Structural requirements for membrane assembly of proteins spanning the membrane several times. J. Cell Biol. 109, 2013-2022. Loo, T. W. and Clarke, D. M. (1995). Membrane topology of a cysteine-less mutant of human P-glycoprotein. J. Biol. Chem. 270, 843-848. Lopez, C. D., Yost, C. S., Prusiner, S. B., Myers, R. M. and Lingappa, V. R. (1990). Unusual topogenic sequence directs prion protein biogenesis. Science 248, 226-229. Montesano, L. and Glitz, D. G. (1988). Wheat germ cytoplasmic ribosomes. Structure of ribosomal subunits and localization of N6,N6dimethyladenosine by immunoelectron microscopy. J. Biol. Chem. 263, 4932-4938. Mothes, W., Heinrich, S. U., Graf, R., Nilsson, I., von Heijne, G., Brunner, J. and Rapoport, T. A. (1997). Molecular mechanism of membrane protein integration into the endoplasmic reticulum. Cell 89, 523-533. Parks, G. D. and Lamb, R. A. (1991). Topology of eukaryotic type II membrane proteins: importance of N-terminal positively charged residues flanking the hydrophobic domain. Cell 64, 777-787. Poloni, F., Romagnoli, G., Cianfriglia, M. and Felici, F. (1995). Isolation of antigenic mimics of MDR1-P-glycoprotein by phage-displayed peptide libraries. Int. J. Cancer 61, 727-731. Romagnoli, G., Poloni, F., Flego, M., Moretti, F., Di Modugno, F., Chersi, A., Falasca, G., Signoretti, C., Castagna, M. and Cianfriglia, M. (1999). Epitope mapping of the monoclonal antibody MM12.10 to external MDR1 P-glycoprotein domain by synthetic peptide scanning and phage display technologies. Biol. Chem. 380, 553-559.

Ribosomal regulation of polytopic topogenesis 2555 Sipos, L. and von Heijne, G. (1993). Predicting the topology of eukaryotic membrane proteins. Eur. J. Biochem. 213, 1333-1340. Skach, W. R., Calayag, M. C. and Lingappa, V. R. (1993). Evidence for an alternate model of human P-glycoprotein structure and biogenesis. J. Biol. Chem. 268, 6903-6908. Skach, W. R. and Lingappa, V. R. (1993). Intracellular trafficking of pre(pro) proteins across RER membrane. In Mechanisms of Intracellular Trafficking and Processing of Proproteins (ed. Y. Peng Loh), pp. 19-77. Boca Raton, FL: CRC Press Inc. Slatin, S. L., Qiu, X.-Q., Jakes, K. S. and Finkelstein, A. (1994). Identification of a translocated protein segment in a voltage-dependent channel. Nature 371, 158-161. Spiess, M., Handschin, C. and Baker, K. P. (1989). Stop-transfer activity of hydrophobic sequences depends on the translation system. J. Biol. Chem. 264, 19117-19124. Spremulli, L. L., Walthall, B. J., Lax, S. R. and Ravel, J. M. (1977). Purification and properties of a Met-tRNAf binding factor from wheat germ. Arch. Biochem. Biophys. 178, 565-575. Treadwell, B. V., Mauser, L. and Robinson, W. G. (1979). Initiation factors for protein synthesis from wheat germ. Methods Enzymol. 60, 181-193. Walter, P. and Blobel, G. (1983). Preparation of microsomal membranes for cotranslational protein translocation. Methods Enzymol. 96, 84-93. Wang, C., Chen, M., Han, E. and Zhang, J. T. (1997). Role of ribosomes in reinitiation of membrane insertion of internal transmembrane segments in a polytopic membrane protein. Biochemistry 36, 11437-11443. Wessels, H. P. and Spiess, M. (1988). Insertion of a multispanning membrane

protein occurs sequentially and requires only one signal sequence. Cell 55, 61-70. Yost, C. S., Lopez, C. D., Prusiner, S. B., Myers, R. M. and Lingappa, V. R. (1990). Non-hydrophobic extracytoplasmic determinant of stop transfer in the prion protein. Nature 343, 669-672. Zhang, J. T. (1996). Sequence requirements for membrane assembly of polytopic membrane proteins: molecular dissection of the membrane insertion process and topogenesis of the human MDR3 P-glycoprotein. Mol. Biol. Cell 7, 1709-1721. Zhang, J. T., Chen, M., Han, E. and Wang, C. (1998). Dissection of de novo membrane insertion activities of internal transmembrane segments of ATPbinding-cassette transporters: toward understanding topological rules for membrane assembly of polytopic membrane proteins. Mol. Biol. Cell 9, 853863. Zhang, J. T., Duthie, M. and Ling, V. (1993). Membrane topology of the Nterminal half of the hamster P-glycoprotein molecule. J. Biol. Chem. 268, 15101-15110. Zhang, J. T., Lee, C. H., Duthie, M. and Ling, V. (1995). Topological determinants of internal transmembrane segments in P-glycoprotein sequences. J. Biol. Chem. 270, 1742-1746. Zhang, J. T. and Ling, V. (1991). Study of membrane orientation and glycosylated extracellular loops of mouse P-glycoprotein by in vitro translation. J. Biol. Chem. 266, 18224-18232. Zhang, J. T. and Ling, V. (1995). Involvement of cytoplasmic factors regulating the membrane orientation of P-glycoprotein sequences. Biochemistry 34, 9159-9165.