Endoplasmic Reticulum-bound Ribosomes Reside in Stable Association with the Translocon following Termination of Protein Synthesis*

In current views, translation-coupled ribosome binding to the endoplasmic reticulum (ER) membrane is transient, with association occurring via the signal recognition particle pathway and dissociation occurring upon the termination of protein synthesis. Recent studies indicate, however, that ribosomal subunits remain membrane-bound following the termination of protein synthesis. To define the mechanism of post-termination ribosome association with the ER membrane, membrane-bound ribosomes were detergent-solubilized from tissue culture cells at different stages of the protein synthesis cycle, and the composition of the ribosome-associated membrane protein fraction was determined. We report that ribosomes reside in stable association with the Sec61α-translocon following the termination stage of protein synthesis. Additionally, in vitro experiments revealed that solubilized, gradient-purified ribosome-translocon complexes were able to initiate the translation of secretory and cytosolic proteins and were functional in assays of signal sequence recognition. Using this experimental system, synthesis of signal sequence-bearing polypeptides yielded a tight ribosome-translocon junction; synthesis of nascent polypeptides lacking a signal sequence resulted in a disruption of this junction. On the basis of these data, we propose that in situ, ribosomes reside in association with the translocon throughout the cycle of protein synthesis, with membrane release occurring upon translation of proteins lacking topogenic signals.

In current views, translation-coupled ribosome binding to the endoplasmic reticulum (ER) membrane is transient, with association occurring via the signal recognition particle pathway and dissociation occurring upon the termination of protein synthesis. Recent studies indicate, however, that ribosomal subunits remain membrane-bound following the termination of protein synthesis. To define the mechanism of post-termination ribosome association with the ER membrane, membrane-bound ribosomes were detergent-solubilized from tissue culture cells at different stages of the protein synthesis cycle, and the composition of the ribosomeassociated membrane protein fraction was determined. We report that ribosomes reside in stable association with the Sec61␣-translocon following the termination stage of protein synthesis. Additionally, in vitro experiments revealed that solubilized, gradient-purified ribosome-translocon complexes were able to initiate the translation of secretory and cytosolic proteins and were functional in assays of signal sequence recognition. Using this experimental system, synthesis of signal sequence-bearing polypeptides yielded a tight ribosometranslocon junction; synthesis of nascent polypeptides lacking a signal sequence resulted in a disruption of this junction. On the basis of these data, we propose that in situ, ribosomes reside in association with the translocon throughout the cycle of protein synthesis, with membrane release occurring upon translation of proteins lacking topogenic signals.
In mammalian cells, ribosomes comprise a common pool that is distributed between the cytosol and the endoplasmic reticulum (ER) 1 membrane (1). Ribosome binding to the ER membrane occurs following recognition of the nascent polypeptide by the signal recognition particle (SRP), interaction of SRP with its receptor at the ER membrane, and subsequent transfer of the ribosome-nascent polypeptide complex (RNC) to the ER translocation machinery, referred to as the translocon (2)(3)(4). The translocon then performs additional nascent polypeptide recognition functions necessary for the formation of a tight ribosome-translocon junction and subsequent protein translocation (5)(6)(7)(8).
The experimental studies leading to the identification of SRP provided clear evidence that secretory protein synthesis could be initiated on free, cytosolic ribosomes and the resulting RNC complexes transferred to the ER membrane, yielding a membrane-bound ribosome fraction (9 -11). With the SRP pathway functioning as the input pathway for ribosome association with the ER, it has long been presumed that the termination of protein synthesis yields the release of ribosomal subunits from the ER membrane to the cytosol, thereby completing a cycle of ribosome binding and release (12,13). However, in contrast to the very well documented mechanism of SRP-mediated RNC targeting, there is a paucity of experimental evidence in support of termination-elicited ribosome release from the ER. In fact, recent reports have indicated that a substantial fraction of the membrane-associated ribosome pool remains bound following the termination of translation (14,15). Since membranebound and cytosolic ribosomes are known to comprise a common pool, how then is ribosome release from the ER membrane effected? We have previously demonstrated in in vitro studies that membrane-bound ribosomes are capable of reinitiating translation and have proposed that topogenic signals in the nascent polypeptide determine the compartmental fate of these ribosomes (15). For example, membrane-bound ribosomes engaged in the synthesis of cytosolic proteins are released from the membrane, whereas ribosomes translating secretory proteins remain in association with the membrane and can access translocation sites in an SRP receptor-independent manner (15). It is unclear from these studies, however, whether upon termination ribosomes remain in direct association with the translocon or are transferred to a distinct binding site, from which they can later reaccess the translocation machinery.
In view of findings demonstrating that bound ribosomes can regulate translocon gating (6,16), that the nascent polypeptide exit site on the ribosome directly abuts the translocon (17)(18)(19) and more recent findings that the translocon serves as the site for malfolded protein export from the ER (20), it is clear that insights into the regulation of ribosome-translocon interactions are crucial to the understanding of both ribosome exchange and bidirectional protein traffic across the ER. Toward this end, we report that in intact cells membrane-bound ribosomes remain in direct association with the Sec61␣-translocon throughout termination and reinitiation of protein translation. Evidence is also provided demonstrating that the disassembly of the ribosome-translocon junction occurs upon synthesis of protein lacking a topogenic signal. These results indicate that the physiological process of ribosome exchange on the ER membrane is driven by the compartmental fate of the translated protein and thus is not mechanistically coupled to the termination of protein synthesis.  , MI). Puromycin, cycloheximide, Triton X-100, and Triton X-114 were from Sigma. Digitonin and dodecylmaltoside were from Calbiochem. DHPC was obtained from Avanti Polar Lipids (Alabaster, AL). Nucleasetreated reticulocyte lysate was from Promega (Madison, WI). Cell culture reagents and proteinase K were from Invitrogen.
Cell Culture and Labeling-Human Jurkat T cells were cultured in RPMI supplemented with 10% fetal bovine serum, 100 units/ml penicillin and 100 g/ml streptomycin at 37°C with 5% CO 2 . Cells were kept within a density range of 0.3-1 ϫ 10 6 cells/ml by subculturing every 2-3 days. For labeling with [ 32 P]orthophosphate, cells at a density of 0.5 ϫ 10 6 cells/ml were starved in serum-and phosphate-free RPMI for 40 min at 37°C, [ 32 P]orthophosphate was added to a final concentration of 100 Ci/ml, and cells were labeled for 40 min.
Optimization of Cell Lysis Conditions-Jurkat cells were sedimented, washed with ice-cold phosphate-buffered saline, and resuspended at 1 ϫ 10 7 cells/ml in lysis buffer containing 2% detergent (digitonin, dodecylmaltoside, or DHPC), 400 mM KOAc, 25 mM K-HEPES (pH 7.2), 5 mM Mg(OAc) 2 , 0.2 mM cycloheximide, 40 units/ml RNasin, and 1 mM phenylmethylsulfonyl fluoride. Samples were incubated for 30 min at 4°C and were then centrifuged at 10,000 rpm (7500 ϫ g) for 10 min at 4°C to sediment insoluble matter. Pellets were resuspended in SDS-PAGE sample buffer (5% SDS, 0.5 M Tris, and 100 mM ␤-mercaptoethanol), whereas 150 l of supernatants were layered above 50-l sucrose cushions (0.5 M sucrose, 400 mM KOAc, 25 mM K-HEPES (pH 7.2), 5 mM Mg(OAc) 2 ) in TLA100 tubes. To sediment ribosomes, samples were centrifuged at 80,000 rpm (250,000 ϫ g) for 30 min at 4°C in a TLA100 rotor (Beckman). Pellets were resuspended in SDS-PAGE sample buffer, while the supernatants from both the 10,000 and 80,000 rpm spins were precipitated by the addition of trichloroacetic acid to a final concentration of 10%. Samples were incubated for 15 min on ice and centrifuged at 13,000 rpm (13,000 ϫ g) for 15 min at 4°C. The precipitated supernatants were washed in acetone and resuspended in SDS-PAGE sample buffer. Pellet and precipitated supernatant samples corresponding to the same amounts of cell lysates were then processed for SDS-PAGE. Gels were transferred onto nitrocellulose membranes using a semidry transfer apparatus (Bio-Rad) and transfer buffer (50 mM CAPS (pH 11.0), 20% methanol, and 0.075% SDS). Immunoblots for the indicated proteins were visualized using ECL detection reagent (Amersham Biosciences). Immunoblot films were scanned on an Epson Expression 636, and band densities were quantified using Fujix MacBAS version 2.0 software.
Analysis of Ribosome-associated Membrane Proteins-Jurkat cells (0.5 ϫ 10 6 cells/ml) were either treated with 0.2 M pactamycin or 0.2 mM puromycin for 15 min at 37°C or were untreated. Alternatively, cells were incubated in serum-free RPMI for 24 h at 37°C. Regardless of treatment, cells were then incubated in the presence of 0.2 mM cycloheximide for 5 min at 37°C. Cells were permeabilized to remove cytoplasmic contents as previously described (14) and were then suspended in digitonin-based lysis buffer (described above). If cells had been labeled with 32 P, phosphatase inhibitors (0.4 mM Na 3 VO 4 , 5 mM NaF) were included. Following a 30-min incubation on ice, cell lysates were centrifuged at 10,000 rpm for 10 min at 4°C, and supernatants were layered onto 10-ml 15-50% linear sucrose gradients containing 400 mM KOAc, 25 mM K-HEPES (pH 7.2), 5 mM Mg(OAc) 2 , 0.1% digitonin, 0.2 mM cycloheximide, and 10 units/ml RNasin. Gradients were centrifuged at 35,000 rpm for 3 h at 4°C in an SW41 rotor (Beckman) and were manually fractionated by puncturing tube bottoms. Absorbance values at 260 nm were then determined for the gradient fractions. Proteins in each gradient fraction were precipitated with trichloroacetic acid and processed for SDS-PAGE and immunoblotting. Alternatively, fractions containing either monosomes or polysomes were pooled and diluted 1:1 with 400 mM KOAc, 25 mM K-HEPES (pH 7.2), 5 mM Mg(OAc) 2 , and 0.1% digitonin. The samples were then layered above a 1.5 M sucrose cushion in the same buffer, and ribosomes were pelleted by centrifugation at 100,000 rpm for 90 min at 4°C in a TLA100.3 rotor. To isolate the ribosome-bound membrane protein fraction, ribosomes were resuspended in 1 ml of 2% Triton X-114, 500 mM KOAc, and 25 mM K-HEPES (pH 7.2) and incubated on ice for 30 min. Samples were incubated at 37°C for 5 min to induce phase separation and centrifuged at 3000 rpm at room temperature. The upper aqueous phase containing soluble proteins was discarded, and the lower detergent phase enriched in membrane proteins was extracted twice with 500 mM KOAc and 25 mM K-HEPES (pH 7.2), washed sequentially with acetone and metha-nol, and then suspended in SDS-PAGE sample buffer. Samples were resolved on 10 -18% linear polyacrylamide gels and silver-stained.
In Vitro Transcription and Translation-The generation of mRNA transcripts coding for pPL86, with and without N-terminal luciferase extensions, and Rluc94 has previously been described (15). Translations were conducted as described (23), except that a postribosomal supernatant reticulocyte lysate-derived fraction was used. The postribosomal supernatant was obtained by two sequential centrifugations of reticulocyte lysate (80,000 rpm, 30 min, 4°C) in a TLA100 rotor. Protein synthesis activity was restored to the ribosome-depleted lysate by the addition of 2 pmol of reticulocyte ribosomes, 2 eq of floated RM, or 2 pmol of detergent-solubilized ribosome-translocon complexes. Translation reactions were conducted for 30 min at 25°C and were stopped by the addition of 1 mM cycloheximide. When the binding of free ribosomes to RM was investigated, 2 eq of RM were added at this point, and the samples were incubated at 25°C for 15 min. To investigate protease accessibility of the nascent polypeptides, proteinase K was added to a final concentration of 100 g/ml, and samples were placed on ice for 30 min and then prepared for SDS-PAGE as described (23).

Translation-independent Association of Ribosomes with the
Translocon-The present study was performed to define the physiological basis of ribosome exchange on the ER membrane, as occurs in the intact cell, and to determine the membrane binding site for ER-associated ribosomal subunits following the termination of protein synthesis. In order to identify the binding partners of membrane-bound ribosomes present in intact cells, both prior to and after the termination of protein synthesis, it was necessary to utilize an experimental system in which actively translating ribosomes could be readily distinguished from inactive ribosomes. Thus, these studies were conducted in tissue culture cells (Jurkat), whose monosome and polysome populations are clearly distinguishable upon velocity sedimentation. The use of tissue culture cells necessitated the development of methodologies that would allow efficient solubilization of the ER membrane without disruption of the binding interactions between ribosomes and ribosome-associated membrane proteins (RAMPs) (24). The relative abilities of three detergents (digitonin, dodecylmaltoside, and DHPC) to fulfill this criterion are described in Fig. 1. When added to cells, all of the indicated detergents were capable of solubilizing the ER membrane, as determined by the liberation of ER components into the supernatant following low speed (7500 ϫ g) centrifugation (Fig. 1A). As expected, following a high speed (250,000 ϫ g) centrifugation step in which ribosomes are sedimented, the ER luminal chaperone GRP94 remained soluble, regardless of the detergent employed (Fig. 1B). However, translocon components Sec61␣ and ribophorin I, proteins that have been designated as RAMPs in previous studies (24,25), were separated from ribosomes to different degrees and in a detergent-specific manner.
In the presence of DHPC, ribosome-RAMP interactions were greatly disrupted (Fig. 1B), whereas in the presence of dodecylmaltoside, ribosome-Sec61␣ interactions were stable, al-though ribosome-ribophorin I interactions were disrupted (Fig.  1B). In contrast, digitonin was able to maintain the majority of ribosome-RAMP interactions and thus was utilized in subsequent studies.
We next employed experimental conditions in which cellular, ER-associated ribosomes could be identified as either actively translating polysomes or as post-termination monosomes. In actively metabolizing cells, most membrane-bound ribosomes are active in protein translation and reside within polysomes ( Fig. 2A). In examining the RAMP fraction derived from ERassociated cellular polysomes, we observed that the majority of the Sec61␣ was recovered in association with these polysomes, whereas a significant fraction (approximately 30%) of cellular ribophorin I and the entirety of GRP94 were found at the top of the gradient, not associated with ribosomes ( Fig. 2A). To obtain post-termination monosomes, cells were treated under a variety of experimental conditions to promote polysome breakdown. Jurkat cells undergo apoptosis upon prolonged serum starvation, with the induction of apoptosis resulting in an inhibition of the initiation stage of protein translation (26,27). We verified these prior findings in demonstrations that 24 h of serum starvation caused a decrease in the levels of translation-ally active polysomes and an accompanying increase in the levels of inactive 80 S monosomes (Fig. 2B). Surprisingly, following apoptosis-elicited polysome breakdown, Sec61␣, the primary site of ribosome binding, remained in association with the monosome fraction (Fig. 2B). If, following termination, membrane-bound ribosomes had diffused away from the translocon, the resulting "vacant" Sec61 complexes would have been recovered in a free form at the top of the gradient (Fig. 2, A and B). Interestingly, a slightly smaller form of Sec61␣ was observed following serum deprivation, suggesting that the induction of apoptosis is accompanied by the proteolysis of Sec61␣.
Treatment of cells with pharmacological inhibitors of translation, such as puromycin, a compound that promotes premature termination, or pactamycin, an inhibitor of the initiation stage of translation, efficiently elicited polysome breakdown (Fig. 2, C and D). Consistent with the results obtained following serum deprivation, Sec61␣ was found in association with monosomes arising from pharmacological inhibition of protein synthesis. These results corroborate earlier in vitro findings obtained in isolated rough microsomes and thus support the hypothesis that upon termination, membrane-bound ribosomes remain in association with the translocon (28). Significantly, these data comprise the first experimental evidence that this phenomenon occurs in intact cells.
Analysis of Translocon Composition as a Function of Ribosome Translation State-Having demonstrated that upon termination, membrane-bound ribosomes remain in association with Sec61␣, we next examined whether the total complement of RAMPs associated with translationally active polysomes and post-termination monosomes varied as a function of translation. In these experiments, polysomes from control cells and 80 FIG. 1. Detergent-specific solubilization of endoplasmic reticulum proteins. A, cells were solubilized using digitonin, dodecylmaltoside, or DHPC and were subjected to low speed centrifugation (7500 ϫ g) to sediment insoluble matter. Pellets and supernatants were then resolved by SDS-PAGE, and the percentages of the indicated proteins present in the supernatants were quantified by immunoblotting. B, supernatants derived from low speed centrifugation were subjected to high speed centrifugation (250,000 ϫ g) to sediment ribosomes, and the percentages of the indicated proteins in the ribosomal pellets were determined as in A.

FIG. 2. Translocon components remain ribosome-associated throughout the protein translation cycle. Cells were left untreated (A), serum-starved for 24 h (B)
, or treated with puromycin (C) or pactamycin (D). All cells were then treated with cycloheximide to preserve the remaining polysome structure, permeabilized to remove cytosolic contents, and subsequently solubilized with digitonin. Lysates were centrifuged through 15-50% sucrose gradients, and the absorbance of the fractions at 260 nm was determined. Proteins present within each fraction were analyzed by immunoblotting. The identities of the 60 and 80 S peaks were verified by determination of the presence of 18 and 28 S rRNAs (data not shown). Note that the y axis scales differ among the graphs.
S monosomes from pactamycin-treated cells were selected as highly enriched populations of translationally active and posttermination ribosomes, respectively. Equivalent amounts of ribosomes from each population were purified and subsequently extracted with Triton X-114 to enrich for the bound integral membrane protein fraction. As shown in Fig. 3A, the composition of the ribosome-associated membrane protein fraction displayed little variation with the translation state of the associated ribosome. It should also be noted that the posttermination ribosomes recovered from the pactamycin-treated cells lack nascent polypeptides, and thus the nascent polypeptide does not appear to perform a function in regulating translocon composition. As can be readily seen from the data presented in Fig. 3A, the three main heterooligomeric complexes of the translocon, the Sec61 complex (Sec61␣, -␤, and -␥), the oligosaccharyltransferase complex (ribophorin I, ribophorin II, and OST48), and the TRAP complex (TRAP␣, -␤, -␥, and -␦) all are present at nearly equivalent levels in the two ribosome populations.
Previous studies have indicated that several translocon com-ponents are phosphoproteins and that phosphorylation state can regulate their interactions with ribosomes (29 -31). Furthermore, phosphorylation of RM has been demonstrated to have functional consequences for protein translocation (30). To determine whether the phosphorylation state of translocon components varies as a function of the translation state of associated ribosomes, cells were labeled with [ 32 P]orthophosphate, and polysome-and monosome-associated membrane protein fractions were isolated. In these analyses, we observed that two components of the RAMP fraction, TRAP␣ and Sec61␤, were recovered as phosphoproteins and, furthermore, that their phosphorylation state did not vary with the ribosome translation state (Fig. 3B). Hence, it appears that phosphorylation of translocon components does not serve a regulatory role in coupling translocon composition to the ribosome translation state. Functional Analysis of Soluble Ribosome-Translocon Complexes-Thus far, our analyses have focused on the characterization of membrane proteins that remain bound to active ribosomes (polysomes) and post-termination ribosomes (monosomes) following solubilization with digitonin. However, there are proteins that participate in the translocation reaction, such as TRAM and the signal peptidase complex (32,33), that are not found in association with ribosomes following solubilization of the ER with digitonin (24,25). We therefore wanted to determine whether, following the reinitiation of translation by membrane-bound ribosomes, the core RAMPs identified in Fig.  3A were responsible for the nascent polypeptide recognition events that contribute to the regulation of ribosome-membrane interactions (15,28). To address this question, ribosome-translocon complexes were purified from canine RM, as depicted in Fig. 4. We chose this experimental system to ensure purity of the final ribosome-translocon preparation. By initially floating the RM though high salt-containing (350 mM KOAc) sucrose solutions and then pelleting the membranes in a subsequent centrifugation step, any free ribosomes present in the RM fraction were removed. The purified RM were then solubilized with digitonin, and the ribosome-translocon complexes were subsequently purified by velocity sedimentation on linear sucrose gradients. Importantly, the complement of RAMPs bound to these RM-derived ribosomes was identical to that bound to  1-4) and polysomes from untreated cells (lanes 5-8) were obtained by pooling gradient fractions containing the appropriate ribosome peaks as shown in Fig. 2, D and A, respectively. Ribosomes were then pelleted, and RAMPs were extracted with Triton X-114, processed for SDS-PAGE on a 10 -18% gradient gel, and silver-stained. Lanes containing the same relative amounts of RAMPs were extracted from equal amounts of ribosomes. B, cells were labeled with [ 32 P]orthophosphate, and RAMPs bound to monosomes and polysomes were then obtained and processed by SDS-PAGE as described for A. Proteins were visualized by autoradiography, and the indicated radioactive bands were identified by molecular weight. Jurkat cell-derived ribosomes (compare Figs. 3A and 4). These ribosome-translocon complexes were then used in in vitro translation reactions with mRNAs encoding truncated polypeptides either possessing (preprolactin) or lacking (luciferase) an ER signal sequence. The ability of the soluble, ribosome-bound translocon to perform the essential nascent polypeptide recognition event was then assessed using a protease protection assay. We hypothesized that if recognition occurred, the soluble translocon would remain bound to the ribosome and thereby yield protection of the nascent polypeptide from exogenous proteases; conversely, if recognition did not occur, the translocon would detach from the ribosome, and the nascent polypeptide would become accessible to protease. This experimental approach has previously been validated in studies demonstrating that digitonin-soluble translocons can bind to free RNCs bearing secretory polypeptides to protect them from proteolytic degradation (19,34). While assays would be preferred that directly assess whether translocons remained bound to ribosomes following translation, such analyses were hindered by the low percentage of translationally active ribosomes present in the purified preparations (data not shown).
We first assessed the topology of nascent pPL86 polypeptides within the translocon to verify that in our experimental system, the addition of protease to the luminal side of the translocon would not degrade a nascent polypeptide contained within. Toward this end, mRNAs were prepared that encoded for pPL86 with N-terminal luciferase extensions of varying length. Free ribosomes carrying these nascent polypeptides have previously been shown to target to the ER membrane in an SRP-dependent manner (15). Upon in vitro translation in the presence of RM, pPL86-mers lacking extensions were protected from protease, in agreement with numerous reports ( Fig  5A, lanes 1 and 2) (7, 35). However, while the nascent pPL86 polypeptides possessing N-terminal extensions displayed protease sensitivity, the digestion products were of the same approximate size as pPL86 (Fig. 5A, lanes 3-8, light arrows). Thus, the topology of pPL86 within the translocon appears to be a hairpin, with the N-terminal extensions breaching the ribosome-translocon junction and accessing the cytosol. This result is similar to the previously described hairpin topology of nascent vesicular stomatitis virus glycoprotein upon RNC targeting to the ER (36).
We next tested the nascent polypeptide recognition properties of the translocon in several different conditions. When synthesized on free ribosomes in the absence of RM, both pPL86 and Rluc94 were protease-sensitive (Fig. 5B, lanes 1 and  2). In contrast, the post-translational addition of RM yielded protection of pPL86, but not Rluc94, consistent with the notion that the SRP targeting machinery selectively directs free, signal sequence-bearing RNCs to the translocon (Fig. 5B, lane 3  versus lane 4). When translation was initiated using membrane-bound ribosomes, pPL86 was again resistant to protease, as the RNC continually remained membrane-bound and accessed the translocon in an SRP-independent manner (Fig. 5B, lane 5 versus lane 6) (15). In contrast, when the synthesis of a cytosolic protein (Rluc94) lacking a signal sequence is initiated on membrane-bound ribosomes, the RNCs detach from the membrane, allowing protease access to the nascent polypeptide (Fig. 5B, lanes 5 and 6) (15). Importantly, similar results were obtained using the soluble, gradient-purified ribosome-translocon complexes as the source of translation activity (Fig. 5B,  lanes 7 and 8). Using this system, the nascent pPL86 was inaccessible to protease, indicating that the soluble translocon recognizes the signal sequence and remains bound to the ribosome. However, upon translation of Rluc94 on soluble ribosome-translocon complexes, the nascent polypeptide was found to be wholly protease-accessible (Fig. 5B, lanes 7 and 8). This result suggests that the translocon had failed to interact productively with the Rluc94 nascent polypeptide and the ribosome-translocon junction had become disrupted, yielding protease sensitivity. As a control, and to allow unequivocal identification of any limit digestion products, translations were conducted using the soluble ribosome-translocon complexes in the absence of exogenously added mRNAs (Fig. 5C). As depicted in Fig. 5C, run-off translation of the soluble complexes,  3 and 4), 60 amino acids (lanes 5 and 6), or 94 amino acids (lanes 7 and 8). Following translation and targeting of RNCs to RM, samples were either untreated or treated with 100 g/ml proteinase K and were then processed for SDS-PAGE. Unproteolyzed nascent polypeptides are indicated by black arrowheads, and primary proteolysis products are indicated by white arrowheads. B, translations of pPL86 or Rluc94 were conducted using free ribosomes (lanes 1-4), membrane-bound ribosomes (lanes 5 and 6), or soluble ribosome-translocon complexes (lanes 7 and 8). Following translation, half of the free ribosomes were allowed to bind to RM (lanes 3 and 4). Accessibility of protease to the nascent polypeptides was then assessed as in A. A ribosome-protected fragment of Rluc94 is not seen due to a paucity of methionines in the C-terminal part of this truncated protein (amino acids 50 -94). The diffuse band present in all lanes represents the nonenzymatic labeling of globin present in the translation mixture. C, translations using soluble ribosome-translocon complexes were conducted in the absence of exogenously added mRNA. A number of high molecular weight proteins were detected (arrowheads) but were degraded to a single diffuse band (asterisk) following proteolysis.
in the absence of added mRNA, yielded several higher molecular weight radiolabeled proteins representing the residual presence of endogenous mRNAs in the ribosome-translocon preparations. Following proteolysis, these proteins were degraded to give a diffuse band of ϳ40 -60 amino acids (asterisk). The radiolabeled band, indicated by the asterisk in Fig. 5B, lane 8, thus does not represent specific degradation products of either pPL86 or Rluc94. We propose, then, that the translocon components identified in Fig. 3A are functional in nascent polypeptide recognition and determine whether, upon reinitiation of translation, a membrane-bound ribosome remains in association with the translocon or alternatively is released from the membrane to continue translation in the cytosol. DISCUSSION In the present study, we have investigated the interactions of ER-associated ribosomes with components of the membrane during the initiation, elongation, and termination stages of translation and report the following conclusions. 1) Following termination, ribosomes remain in stable association with the translocon. 2) The protein composition of the translocon does not vary when bound ribosomes transition between translationally active and post-termination states.
3) The translocon performs nascent polypeptide recognition functions necessary for determining the compartmental fate of translationally active membrane-bound ribosomes. On the basis of the data contained herein, we propose that in the intact cell, membranebound ribosomes remain in association with the translocon throughout the cycle of initiation, elongation and termination. Consistent with previous results, membrane-bound ribosomes engaged in the synthesis of cytosolic proteins are released from the translocon. This finding is in agreement with our previous suggestion that during the elongation cycle of protein synthesis, a stable ribosome-translocon junction requires productive translocation of the nascent polypeptide (15,28).
The results obtained by in vivo analysis of membrane components interacting with post-termination ribosomes are complementary to several previous observations. In an early experiment defining the ribosome-binding characteristics of Sec61␣, in vitro treatment of RM with puromycin at physiological salt concentrations did not yield disruption of the ribosome-Sec61␣ interaction (24). Although puromycin treatment does not replicate physiological termination, these data suggested the existence of a stable ribosome-translocon interaction. This concept is supported by experiments investigating ER membrane permeability to small molecules. Upon treatment of permeabilized cells or rough microsomes with puromycin, an increase in the passage of small molecules across the ER membrane was observed, but it was diminished upon the release of ribosomes from the membrane with increasing salt concentrations (16,37). These data support the proposal that following termination, ribosomes remain directly bound to the translocon and prevent the pore from adopting a conformation closed to small molecules; removal of ribosomes from the translocon complex then restores membrane impermeability through closure of the pore and/or translocon disassembly, as has previously been observed (38).
Recent studies using cryoelectron microscopy have defined the interactions between ribosomes and the Sec61 complex at the atomic level (17)(18)(19). The composite structures of ribosomes and attached Sec61 complexes are similar under conditions in which the ribosome is stalled in the process of translocation or lacks a nascent chain. However, we have demonstrated that translocon-associated ribosomes can initiate translation of cytoplasmic proteins and as such will eventually dissociate from the translocon (Fig. 5B). To accommodate these two observations, we suggest that in the absence of binding interactions between the nascent polypeptide and components of the trans-location machinery, ribosomal conformation changes that accompany the elongation cycle of protein translation lead to the eventual dissociation of the ribosome-translocon complex. From current atomic structure models, it can be suggested that movement of ribosomal expansion segment 27 may contribute to this release process (19). As the complex between the translocon and a ribosome translating a cytoplasmic protein exists only transiently (15), capturing the complex to allow identification of its structural details will prove challenging.
An important role for the translocon in mediating the process of protein export, or retrotranslocation, from the ER has emerged in recent years (20). This process, a component of ER-associated degradation, requires that malfolded protein substrates be selected for export and discharged through the translocon to the cytoplasm for proteosomal degradation (39,40). However, our results indicate that in the cell, translocons are occupied by ribosomes under steady-state conditions, and termination of translation does not yield ribosome dissociation from the translocon. How then do translocons become available to participate in retrotranslocation? Forward and reverse translocation of protein substrates through the translocon does not occur simultaneously, since microsomes saturated with bound RNCs have been shown to be retrotranslocation-incompetent (40). Perhaps the onset of retrotranslocation is accompanied by structural transitions within the translocon, having the effect of both displacing inactive ribosomes from the cytoplasmic face of the translocon and preventing free ribosomes from targeting to such a translocon. Irrespective of the mechanism, the data presented herein indicate that accessing a vacant translocon for the purpose of retrotranslocation may be more complex than implied by classical ribosome exchange models.