(Received for publication, December 31, 1996, and in revised form, February 17, 1997)
From the Department of Biological Chemistry and the Molecular Biology Institute, UCLA School of Medicine, Los Angeles, California 90024-1737
Ribosome binding to the endoplasmic reticulum has been traditionally studied using an in vitro assay in which potential ribosome receptors have been purified, incorporated into synthetic liposomes, and tested for activity. One such receptor (180 kDa; "p180") has been shown to bind ribosomes with high affinity in such a system when purified to homogeneity. This result has been challenged by data generated in other laboratories, and as a result, doubt has lingered as to the authenticity of p180 as a ribosome receptor. The contribution of the major difference between these studies, the lipid composition of the liposomes used in the in vitro assays, was assessed when identical fractions of rough endoplasmic reticulum-specific membrane proteins were incorporated into liposomes composed of only phosphatidylcholine (as used in other laboratories), a 50:50 mix of phosphatidylcholine and phosphatidylserine (as used in our original studies), or lipids derived from canine pancreatic microsomes (as a physiologically relevant control). The presence of PS was found to be crucial for the incorporation into and ribosome binding activity of p180 in liposomes. These observations are compatible with published studies on the importance of acidic phospholipids in ribosome binding to intact microsomes and reconcile the apparently conflicting in vitro results surrounding the assignment of p180 as a ribosome receptor.
Translocation of proteins across the rough endoplasmic reticulum (ER)1 represents the first step in the secretion of proteins from the cell. From the outset, models describing the process of protein translocation have specified the functional need for the binding of ribosomes to the ER membrane (1). Through the in vitro assay devised by Borgese et al., ribosome binding to intact membranes has been shown to be saturable, salt-labile, and protease-sensitive (2, 3). These findings stimulated considerable research into the identification of a proteinaceous receptor for ribosomes on the cytoplasmic surface of the rough ER. Two of these proteins, p34 (4) and p180 (5), when purified to homogeneity and incorporated into synthetic liposomes, were able to bind ribosomes in a salt-dependent, protease-sensitive manner.
p34 was identified as an unglycosylated protein that was enriched in fractions incorporated into phosphatidylcholine (PC) liposomes that bound ribosomes. p34 was readily cross-linked to ribosomes from these liposomes (6), although quantification indicated that 61 p34 molecules were needed to bind one ribosome (4). Subsequent work has shown that anti-p34 antibodies were able to inhibit both ribosome binding and translocation across the ER (7). Most recently, the primary sequence of p34 was deduced, showing the protein to be a type II membrane protein with a large cytosolic domain. It contains 4.5 tandem leucine-rich repeats (23-24 amino acids in length) and an abundance of charged amino acids (8).
p180 was first identified as a soluble proteolytic fragment that
inhibited ribosome binding to intact microsomes (5). Antibodies produced against the fragment enabled the identification of a 180-kDa
cytosolically disposed integral membrane protein that, when purified
and incorporated into liposomes (50:50 mix of phosphatidyl serine (PS)
and PC), promoted protease-sensitive, high affinity (10-20
nM) ribosome binding. The number of ribosomes bound was virtually identical to the number of p180 molecules incorporated into
the liposomes. Antibodies and Fab fragments against p180 were able to
inhibit ribosome binding and translocation in intact microsomes, and
depletion studies demonstrated that removal of p180 from detergent
extracts by monoclonal antibody affinity chromatography produced
liposomes that were incapable of binding ribosomes or translocating
proteins. The readdition of p180 to prior to reconstitution restored
the bulk of activity (9). This type of study enabled an exploration of
ribosome binding against a background of the native microsomal lipids
and the entire collection of endogenous rough ER-specific proteins.
Studies have been published by Nunnari et al. (10) and
Collins and Gilmore (11), whose conclusions question the role of p180
as a ribosome-binding protein. The basis for these interpretations appears valid; in both cases, it was found that fractions of detergent extracts of RM containing p180 did not possess ribosome binding activity when incorporated into egg PC liposomes. Logically, both groups came to the conclusion that an as yet to be determined membrane
protein is the true ribosome receptor. The recent postulation that the
subunit of the Sec61p complex functions as ribosome receptor during
the translocation reaction (12) is addressed under
"Discussion."
Additional support for the role of p180 as a ribosome receptor came from its further characterization, and the demonstration of its ability to bind ribosomes within the context of native membranes both in vivo and in vitro (13). Its primary structure indicates an N-terminal membrane anchor followed by a highly repetitive, positively charged (pI = 11) domain in which a unique decapeptide motif is expressed 54 times without interruption. The C-terminal half is characterized by its low pI (4.3) and its ability to bind ATP. Studies in which the various domains of p180 were expressed in yeast have shown that the N-terminal repeat-containing domain is necessary and sufficient for ribosome binding. This feature was determined by both morphological analysis in vivo and by ribosome binding studies carried out in vitro (13).
Nonetheless, there is still lingering doubt and controversy based in large part on the differences between our data (5, 9) and the results of Nunnari et al. (10) and Collins and Gilmore (11). Since the research groups of these colleagues have carried out pioneering research in the field of protein translocation and the quality and accuracy of their work is beyond question, we sought a rational basis for the discrepancy in our findings. As can be seen below, we hypothesize that the major difference between our results was the choice of phospholipids used in the preparation of proteoliposomes for ribosome binding studies.
Acidic phospholipids (PS and phosphatidylinositol) make up a significant fraction (approximately 15-25%) of the lipids of the rough ER (14, 15). Ribosome binding activity, measured in vitro, was lost when microsomes were digested with a phospholipase C specific for acidic lipids (15) and not with lipases of differing specificity (14). A striking technical difference between our studies and those of Nunnari et al. (10) and Collins and Gilmore (11) was the fact that their test membranes were composed exclusively of phosphatidylcholine; no acidic phospholipids were added. To examine the role of the lipid composition on ribosome binding, identical protein fractions were incorporated into liposomes having differing compositions: pure PC, a 50:50 mix of PC and PS, or a lipid fraction isolated from canine microsomes. The data presented below indicate that lipid composition is not only crucial for activity but can also explain the discrepancies that led to the original questioning of the role of p180 as a ribosome receptor.
Six ml of 100 A280/ml canine pancreatic rough microsomes, prepared as described previously (9), were washed with five volumes of 0.5 M KOAc, 50 mM Tris-HCl, pH 7.5, 5 mM Mg(OAc)2 and pelleted by centrifugation in a Beckman Ti-60 rotor at 113,000 × g for 75 min at 4 °C. The pellet was resuspended in 40 ml of TMG (50 mM Tris-HCl, pH 7.5, 5 mM Mg(OAc)2, 10% glycerol) with 0.5% Nikkol, and pelleted through a cushion of the above solution with 0.5 M sucrose in a Beckman Type 60 Ti rotor at 113,000 × g for 75 min at 4 °C. The pellet was then resuspended in 12 ml of 175 mM KOAc, 1% octylglucoside (OG) in TMG and incubated on ice for 15 min. The suspension was subsequently centrifuged over a cushion of 1.8 M sucrose, 175 mM KOAc, 1% OG in TMG, in a Beckman Ti-80 rotor, at 175,000 × g for 100 min at 4 °C. This pellet was resuspended in 6 ml of 0.7 M KOAc, 1% OG in TMG (HSS buffer), incubated for 15 min on ice, and centrifuged over a cushion of 1.6 M sucrose, 0.7 M KOAc, 1% OG in TMG in a Beckman Ti-80 rotor at 175,000 × g for 2 h. at 4 °C (5). The supernatant was termed the "high salt supernatant (HSS)" for the reconstitution experiments.
Purified p180 was isolated by diluting 2 ml of the HSS with 18 ml of TMG and placing it on a 2-ml column of equilibrated DEAE-Sepharose CL-6B (Pharmacia) at a flow rate of 10 ml/h. After the flow-through was collected, the column was washed with 5 volumes of 0.5 M KOAc, 0.1% OG in TMG, 2 volumes of 100 mM KOAc, 0.1% OG in TMG, and then 5 volumes of 100 mM KOAc, 1% OG in TMG. p180 was eluted by a gradient of 100-400 mM KOAc, 1% OG in TMG (6 volumes) and appeared at approximately the midpoint of the gradient (5). The fractions were analyzed on SDS-PAGE gradient gels (10-15% acrylamide) and stained with Coomassie Blue, and the fractions containing p180 were pooled and diluted with 3 volumes of TMG. This pool was placed over a 1-ml column of equilibrated CM-Sepharose CL-6B (Pharmacia) at a flow rate of 3 ml/h. The column was initially washed with 5 volumes of 100 mM KOAc, 1% OG in TMG and eluted with a gradient of 100-500 mM KOAc, 1% OG in TMG (6 volumes). p180 eluted at approximately 300 mM KOAc. Fractions were analyzed by SDS-PAGE on gradient gels (10-15% acrylamide) and stained with Coomassie Blue. All of the solutions used for the above purification included 110 µM phenylmethylsulfonyl fluoride and 200 µM EGTA.
Antibodies and ImmunodepletionThe two monoclonal anti-p180 antibodies and the one polyclonal anti-p180 antibody were previously described (5, 9). The monoclonal antibody against ribophorin I was described by Hortsch and Meyer (16), and the polyclonal antibody against mammalian Sec61p was the generous gift of Tom Rapoport (Harvard Medical School, Boston, MA) (17). The monoclonal antibody used for the mock depletions was raised against Schizosaccharomyces pombe fatty acid synthetase and does not react with any canine rough ER protein. Immunoblots were visualized by a chemiluminescent system (Renaissance System, DuPont NEN).
Immunodepletions were performed by cross-linking a mixture of the two anti-p180 monoclonals or the control antibody to protein A-agarose (Schleicher and Schuell) by the method of Schneider (18) with the previously described modifications (9). The immunoaffinity matrix was equilibrated with several volumes of HSS buffer. The HSS and an equal volume of matrix were mixed by rotation for 2 h at 4 °C. The matrix was pelleted at 1000 × g for 3 min and washed with one-half volume of HSS buffer. After pelleting the matrix, the wash was combined with the treated HSS and termed "depleted" or "mock-depleted" HSS, depending on the matrix used.
Formation of LiposomesThe canine microsomal lipids (mixed
lipids) were prepared by extracting 1 volume of microsomes with 1.25 volumes of chloroform and 2.5 volumes of methanol by stirring for 60 min at room temperature. The insoluble material was collected by
centrifugation in a Sorvall GSA rotor at 3000 × g for
30 min. The insoluble material was resuspended in 5 volumes of
water/chloroform/methanol (0.8:1:2) for 60 min. The insoluble material
was removed by centrifugation as above, and the two organic phases were
combined. To this combined volume, an equal volume of chloroform/water
(1:1) was added. The phases were separated by centrifugation in a
Sorvall GSA rotor at 3000 × g for 15 min. The organic
(bottom) phase was taken and combined with an equal volume of benzene
and dried down by rotary evaporation. The lipids were resuspended in
400 mM KOAc, 50 mM Tris-HCl, pH 7.5, 0.75%
cholic acid, 2 mM dithiothreitol, 10% glycerol, 0.5 mM EDTA and stored at 80° C (19).
To form the liposomes, an equal mix of PS (from bovine brain, 98% purity determined by TLC; Sigma) and PC (from egg yolk, 99% purity determined by TLC; Sigma), PC alone, or mixed lipids were dried down and resuspended at 20 mg/ml in 100 mM KOAc, 50 mM Tris-HCl, pH 7.5, 1% OG, 10% glycerol, 2 mM dithiothreitol. For the HSS, mock-depleted, depleted, and readdition liposomes, approximately 1 mg of detergent extract was mixed with 100 µl (2 mg) of lipid. For the readded liposomes, 25 µg of p180 was added to the lipids and detergent extract. For the pure p180 liposomes, 25 µg of p180 was mixed with 50 µl (1 mg) of lipid. The samples were then dialyzed against decreasing concentrations of detergent in 50 mM KOAc, 50 mM Tris-HCl, pH 7.5, 1 mM CaCl2: 0.75% OG for 1 h in 100 volumes; 0.25% OG for 3 h in 300 volumes; and no OG for 12 h in 1000 volumes. The samples were removed from the dialysis tubing, and EGTA was added to 2 mM. The liposomes were recovered by centrifugation in a Beckman Ti-80 rotor at 112,000 × g for 20 min at 4 °C and resuspended in approximately 200 µl of membrane buffer (25 mM KOAc, 50 mM Tris-HCl, pH 7.5, 5 mM Mg(OAc)2, 250 mM sucrose, 0.4 mM EGTA) in a Dounce homogenizer. The liposomes were normalized either for protein content by SDS-PAGE on gradient gels (10-15% acrylamide) stained with silver (20) (HSS, mock-depleted, depleted, and readdition) or for their ability to scatter light at 600 nm (no protein and p180).
Ribosome BindingPreparation of tritium-labeled HeLa cell ribosomes and stripped rough microsomes as well as ribosome binding assays were performed (final volume, 50 µl) as described previously (9). Briefly, liposomes incubated with labeled ribosomes were mixed with heavy sucrose, overlaid with a cushion, and floated by centrifugation. Binding was assessed by quantification of radiolabel associated with the floated fraction.
To
initially analyze the ribosome binding activity of liposomes composed
of different lipids, an extract containing rough ER-specific proteins
was incorporated into liposomes. The starting material for these
liposomes was HSS originating from a detergent extraction of rough
microsomes. The HSS was derived from a pellet of ribosomes and their
associated proteins that was obtained by centrifugation of a medium
salt (500 mM KOAc)-detergent extract (see "Experimental
Procedures"). To remove the ribosome-associated membrane proteins
from the ribosomes, the pellet was resuspended in high salt (700 mM KOAc) and detergent, and the ribosomes were repelleted.
The HSS contains most of the rough ER-specific membrane proteins
including p180, ribophorins I and II, Sec61p, docking protein
(SR), SSR, and the signal peptidase complex (Fig. 1, lane 1, and data not shown). Workers in the field agree that
ribosome receptors are among this collection of ribosome-associated
membrane proteins and are associated with the ribosomes in low salt and dissociated in high salt (5, 9, 12).
Incorporation of this extract into liposomes regardless of composition
(pure PC, PS/PC, or microsomal lipids) resulted in ribosome binding
characterized in Fig. 2. One difference observed was
that saturation was not achieved with the pure PC liposomes, although
the rate of increase in ribosomes bound versus ribosomes added did level off somewhat at higher ribosome concentrations. As a
control, liposomes were formed that did not contain any proteins. These
liposomes were isolated, normalized based on scattering of visible
light, and assayed for ribosome binding (Fig. 3,
panels A-C, open squares). Interestingly, even in the
absence of protein, PC liposomes had a much higher capacity to bind
ribosomes than the other liposomes in a manner that was not saturable
(Fig. 3A, open squares). This elevated background
may have resulted from an interaction of the positively charged choline
moiety with negatively charged ribosomes and would be consistent with
the inability of the pure PC liposomes to exhibit saturation of
ribosome binding activity. Regardless, one point is clear; lipid
composition of liposomes is relevant when examining ribosome binding,
and the results obtained with mixtures containing acidic phospholipids more closely resemble those using native microsomal lipids.
Effect of p180 Depletion on Ribosome Binding to Liposomes of Varying Composition
To further examine its role in ribosome
binding, p180 was depleted from ribosome-associated membrane protein
fractions by affinity chromatography using two monoclonal antibodies.
As a control, the HSS was treated with a monoclonal that does not react to any mammalian ER proteins. As can be seen from both the protein profiles and immunoblots, greater than 90% of p180 was removed, and
there was no significant difference in the presence or amount of other
proteins (Fig. 1, lanes 2 and 3). There was no
change in the level of two specific control proteins, ribophorin I and Sec61 (Fig. 1). As an additional control, p180 was purified from the
HSS by salt gradient elution of two consecutive steps of ion exchange
chromatography (Fig. 1, lane 4) so that it could be added back to the depleted extract before its incorporation into liposomes ("readdition control"). This fraction was previously characterized by immunoblotting (9).
The protein profiles of the liposomes show that with all three lipid
mixtures, the mock and p180-depleted extracts were incorporated to
similar extents (Fig. 4, panels A-C,
lanes 3 and 4). However, the resulting activities
were quite different. With the PS/PC and mixed lipid liposomes, the
depletion of p180 caused a 90 and 80% reduction in ribosome binding
activity, respectively, compared with mock-depleted controls (Fig. 3,
B and C). The p180-depleted PC liposomes showed
only a 30% reduction of activity under similar circumstances (Fig.
3A). This result may indicate the existence of an additional
ribosome-binding protein. Such a protein is either able to bind
ribosomes when incorporated into PC liposomes but not when incorporated
into PS/PC or mixed lipid liposomes, or, alternatively, it can become
incorporated only into the pure PC liposomes.
The readdition of p180 to the depleted HSS restored most of the ribosome binding activity in the PS/PC and the mixed lipid liposomes (Fig. 3, B and C, triangles). The readdition of p180 to pure PC liposomes had no effect on activity (Fig. 3A, triangles). An analysis of the protein profiles showed that additional p180 was not incorporated into the PC liposomes when p180 was restored to the depleted extracts (Fig. 4A, lane 5). This finding, i.e. that p180 is incorporated into PC liposomes within the context of the HSS but cannot be incorporated when it is depleted and then readded to the HSS, was surprising and gave the first indication as to what might have led to discrepancies in published studies (10, 11). The protein components should be identical, since p180 was specifically removed from the HSS by the monoclonal antibodies and no other proteins were seen to elute with p180 from the antibody column (9). The only compositional difference is in the lipid component of the proteoliposomes. These data can be explained by the two following postulates. The first is that ribosome binding activity in pure PC liposomes is mediated by a high level of binding to the lipid itself and that under these circumstances, as suggested above, a protein other than p180 contributes to the binding reaction. The second is that p180, when purified away from endogenous lipids, is not able to be incorporated into liposomes in the absence of acidic phospholipids. As part of the HSS, p180 may still be associated with acidic phospholipids from the original extract. Upon affinity purification, however, the acidic phospholipid is lost and must be replaced. An inability of affinity-purified p180 to become incorporated into pure PC liposomes would provide reasonable support for this hypothesis.
Purified p180 Is Not Incorporated into Pure PC LiposomesEqual amounts of pure p180 (Fig. 1, lane 4) were added to the detergent and lipid-containing mixtures used for liposome construction. Unlike the PS/PC and mixed lipid liposomes, p180 incorporation into pure PC liposomes was barely detectable (Fig. 4, panels A-C, lane 6). Densitometric quantification indicated that levels of p180 incorporated into pure PC liposomes were only 5-10% compared with the liposomes with other lipids. These liposomes were normalized based on their scattering of visible light, and approximately equal quantities of liposomes were recovered for analysis. These data provide the most direct support for our hypothesis that differences in p180's ability to bind ribosomes observed between research groups can be explained by differences in lipid composition of the liposomes used in the in vitro assays.
Previous studies involving reconstitution of ribosome binding into liposomes have utilized lipids derived from a variety of sources and with varying compositions (5, 10, 11, 21, 22). In this study, we demonstrate that substantially different results are obtained depending on lipid compositions. The presence of acidic phospholipids is necessary for p180-mediated ribosome binding activity, consistent with the need for acidic phospholipids for ribosome binding to intact membranes (14, 15). Our results gave identical results both for PS/PC-containing liposomes and for liposomes composed of the naturally occurring microsomal lipids, thereby substantiating the use of PS/PC as a suitable substitute for endogenous lipids. In contrast, in liposomes composed of pure phosphatidylcholine, many anomalous results were obtained. In addition to higher levels of background ribosome binding to protein-free pure PC liposomes, the loss of p180 had a significantly smaller effect on ribosome binding activity, and once depleted from a detergent extract of ER membrane proteins, readded p180 could barely become incorporated. These findings provide a reasonable explanation for many of the results obtained in other laboratories (10, 11).
In the published studies of Nunnari et al. (10), pure PC liposomes were used in the reconstitution of ribosome binding activity. In their work, two further items are worth noting that may have a bearing on their observations. Although the fractions of ER membrane proteins used in the construction of liposomes were analyzed, what actually was incorporated was not. Thus, as our studies point out, although p180 was undoubtedly present in the starting material, it probably did not become incorporated into the pure PC liposomes that were ultimately tested for binding activity. Additionally, immunoblot analysis of the fractions obtained following DEAE fractionation showed p180 to be degraded to a smear of immunoreactive products (Ref. 10, Fig. 3C), even prior to incorporation attempts.
Consistent with our results on ribosome binding, the chemical cross-linking approach used by Collins and Gilmore (11) demonstrated that p180 was a major ER membrane component that was proximal enough to ribosomes to be detected by this method. Moreover, their data indicate a stoichiometry of ribosomes to p180 of 1:1. Nonetheless, these authors failed to find that fractions containing p180 bound ribosomes when incorporated into pure PC liposomes. Data were presented demonstrating the presence of p180 in the starting material used for incorporation, but a profile of what actually was incorporated into the liposomes was not shown. An explanation wholly consistent with these observations is the same as above; i.e. purified p180 does not become incorporated into pure PC liposomes.
In both cases, however, ribosome binding was detected in other fractions: fractions containing smaller proteins and ones that bind to ion exchangers. Our results presented here are consistent with these results, since we also found binding activity in pure PC liposomes that was not mediated by p180. A good candidate based on size and charge (8) would be p34, whose activity, coincidentally, was also characterized in pure PC liposomes (21). Based on the results of our p180 depletion studies, p34 probably has little if any activity in PS/PC or endogenous lipid liposomes. The nature of the active factor in pure PC liposomes, unfortunately, remains within the realm of speculation, since neither Nunnari et al. nor Collins and Gilmore have gone on to publish an identification of the binding component in their detergent extracts. An alternative scenario would be one in which, together with p180, another protein is depleted that does not significantly affect ribosome binding but is needed for p180's incorporation into PC liposomes and not into the other liposomes. This is doubtful, however, since additional proteins were not detected when p180 was eluted from the antibody affinity column (data not shown).
Interestingly, we observed that p180 was incorporated into pure PC liposomes when complete detergent extracts were used as a starting material but once purified by various chromatography steps, it was not. A plausible explanation would be that during purification of detergent extracts, in contrast to solubilized whole microsomes, the requisite acidic lipids are lost either from the mixture or from a specific association with p180. This interpretation is supported by the fact that both purified and readded p180, when mixed with detergents and an acidic phoshpholipid, enabled the formation of active liposomes.
More recently, Kalies et al. (12) have proposed that the subunit of the mammalian Sec61p is the major binding partner for
ribosomes in the ER membrane. As can be seen from the data presented
here (Fig. 1), immunoblotting shows that there is no reduction in the
amount of the
subunit of the Sec61 complex in extracts that have
totally lost their ability to bind ribosomes when incorporated into
PS/PC or ER lipid-containing liposomes. Moreover, our studies and those
of others show that ribosome binding is exquisitely sensitive to
proteases (2, 9, 22). Sec61
remains intact after digestion with
significantly higher levels of protease (12) than those previously
shown to totally eliminate ribosome binding to either intact membranes
or liposomes (2, 5, 9, 23, 24).
The role of p180 in translocation is not addressed in this paper, and
it remains an issue for a separate and thorough analysis. Previously,
depletion of p180 was shown to significantly reduce translocation
activity in proteoliposomes (10), whereas recent results show that only
the Sec61p and docking protein (SR) complexes are needed for the
translocation of certain preproteins (17). An acceptable hypothesis,
consistent with both sets of data would be that p180 may not be
absolutely essential for the translocation of preproteins in a highly
reconstituted system (22) but may have a stimulatory or regulatory role
in liposomes made from more complex mixtures of ER-derived proteins and
in intact membranes (10). Preliminary support for such a hypothesis
comes from Northern analysis of p180 showing that its highest
expression occurs in tissues with significant levels of secretory
activity, e.g. pancreas, liver, and
placenta.2 These tissues also have an
abundant and extensive rough ER network throughout the cytoplasm. Since
it has been known for a number of years that not all bound ribosomes
are engaged in protein synthesis (25-27), our current working
hypothesis is that p180 maintains the high level of bound ribosomes
found on the ER membranes of actively secreting cells and tissues and
thus may perform an accessory or ancillary role in the translocation
process. Regardless, we agree with the recent statement from the
Rapoport lab with respect to the binding of ribosomes to the ER
membrane: "Our findings suggest that the targeting process is more
complex than previously anticipated" (28).