Apolipoprotein B Is Intracellularly Associated with an ER-60 Protease Homologue in HepG2 Cells*

Two ALLN (N-acetyl-leucyl-leucyl-norleucinal)-sensitive endoplasmic reticulum (ER)-localized proteases (ER-60 and ER-72) were recently purified from rat liver. We used an antibody to rat ER-60 to investigate the possible role of this protease in apolipoprotein B (apoB) degradation. First, immunoprecipitation and immunoblotting experiments with the anti-rat ER-60 antibody suggested that HepG2 cells contain a homologue of ER-60 with an approximate molecular mass of 58–60 kDa. The ER-60 homologue was mostly associated with the luminal contents of HepG2 microsomes. Evidence from co-immunoprecipitation and cross-linking experiments appear to suggest that the ER-60 homologue in HepG2 cells is associated with apoB intracellularly. A small pool of apoB was recovered when HepG2 lysates were subjected to immunoprecipitation with anti-rat ER-60 antibody followed by a second immunoprecipitation with anti-apoB antibody. Furthermore, cross-linking of permeabilized cells with dithiobis(succinimidylpropionate) further demonstrated association of apoB with the ER-60 homologue in HepG2 cells. Three polypeptides with molecular masses of 78, 66, and 50 kDa were consistently found to be associated with apoB as well as the 58-kDa ER-60 homologue. The 78-kDa protein associated with both apoB and ER-60 appeared to represent immunoglobulin heavy chain-binding protein (BiP) based on immunoprecipitation with a monoclonal antibody. Cross-linking and immunoblotting experiments suggested the association of the 78-kDa BiP with both the 58-kDa ER-60 homologue as well as the 550-kDa apoB. In summary, the data suggests that HepG2 cells contain a 58-kDa protein which is homologous to the rat liver ER-60 in size, antigenecity, and intracellular localization. The ER-60 homologue in HepG2 cells appears to be closely associated with apoB, as well as other proteins possibly representing ER chaperones such as BiP. We hypothesize that the ER-60 homologue may be involved in the degradation of apoB in the ER lumen of HepG2 cells.

Two ALLN (N-acetyl-leucyl-leucyl-norleucinal)-sensitive endoplasmic reticulum (ER)-localized proteases (ER-60 and ER-72) were recently purified from rat liver. We used an antibody to rat ER-60 to investigate the possible role of this protease in apolipoprotein B (apoB) degradation. First, immunoprecipitation and immunoblotting experiments with the anti-rat ER-60 antibody suggested that HepG2 cells contain a homologue of ER-60 with an approximate molecular mass of 58 -60 kDa. The ER-60 homologue was mostly associated with the luminal contents of HepG2 microsomes. Evidence from co-immunoprecipitation and cross-linking experiments appear to suggest that the ER-60 homologue in HepG2 cells is associated with apoB intracellularly. A small pool of apoB was recovered when HepG2 lysates were subjected to immunoprecipitation with anti-rat ER-60 antibody followed by a second immunoprecipitation with anti-apoB antibody. Furthermore, cross-linking of permeabilized cells with dithiobis(succinimidylpropionate) further demonstrated association of apoB with the ER-60 homologue in HepG2 cells. Three polypeptides with molecular masses of 78, 66, and 50 kDa were consistently found to be associated with apoB as well as the 58-kDa ER-60 homologue. The 78-kDa protein associated with both apoB and ER-60 appeared to represent immunoglobulin heavy chain-binding protein (BiP) based on immunoprecipitation with a monoclonal antibody. Cross-linking and immunoblotting experiments suggested the association of the 78-kDa BiP with both the 58-kDa ER-60 homologue as well as the 550-kDa apoB.
In summary, the data suggests that HepG2 cells contain a 58-kDa protein which is homologous to the rat liver ER-60 in size, antigenecity, and intracellular localization. The ER-60 homologue in HepG2 cells appears to be closely associated with apoB, as well as other proteins possibly representing ER chaperones such as BiP. We hypothesize that the ER-60 homologue may be involved in the degradation of apoB in the ER lumen of HepG2 cells.
Post-translational degradation of apoB has been shown to modulate the intracellular levels of newly-synthesized apoB molecules (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19). Recent evidence suggest that apoB degrada-tion may occur in the cytosol by the proteasome (19) as well as in the ER 1 lumen by an unidentified ER protease(s) (17,18). The identity of the ER-associated protease involved in apoB degradation has remained elusive, however, some characteristic features of this degradative system have recently been documented. The ER-associated protease appears to be responsible for fragmentation of apoB into a number of distinct degradation intermediates including an abundant 70-kDa fragment (12,16). The activity of the ER-associated protease is also inhibitable by ALLN in a dose-dependent manner (12,16). Intraluminal degradation of secretion-competent apoB associated with nascent HDL-like and LDL-like lipoprotein particles in the secretory pathway is also ALLN-sensitive and may be mediated by a putative ER-localized protease (17). Work by Ginsberg and co-workers (18) has recently shown that the luminal degradative process is also DTT sensitive.
The ER lumen contains a number of ER-resident proteins including molecular chaperones and proteases. Recently, two ER-associated proteases were purified to homogeneity and characterized (20,21). One of these proteases, the ER-60 protease was first purified from the ER of rat liver, and was shown to be a cysteine protease (20,22). ER-60 has 98% homology in amino acid sequence to rat phosphoinositide-specific phospholipase C (20). This protease has been implicated in degradation of human lysozyme (23). ER-60 was shown to be chemically cross-linked to misfolded mutant lysozyme. It was also found to degrade the reduced and denatured form of lysozyme, but not the native form in vitro (23).
In the present report, we demonstrate that HepG2 cells contain a homologue of rat liver ER-60, an ER-localized cysteine protease. Evidence from co-immunoprecipitation, chemical cross-linking, and a combination of immunoprecipitation and immunoblotting experiments suggest that apoB-100 is associated intracellularly with the ER-60 homologue in HepG2 cells. It is thus postulated that the ER-60 homologue may be involved in a degradative system affecting apoB in the ER. Such a protease may be responsible for the ER luminal protease activity fragmenting apoB into distinct degradation intermediates.

EXPERIMENTAL PROCEDURES
Materials-HepG2 cells (ATCC HB 8065) were obtained from American Type Culture Collection. Fetal bovine serum (certified grade) and cell culture media were obtained from Life Technologies (Toronto, Ontario, Canada). Monoclonal anti-BiP was obtained from StressGen (Vic-* This work was supported by an operating grant from the Heart and Stroke Foundation of Ontario (to K. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  toria, British Columbia, Canada). Rabbit anti-ER-60 polyclonal antibody was developed from purified rat ER-60 (20).
Metabolic Labeling and Permeabilization of HepG2 Cells-Cultured HepG2 cells were maintained as described (24). Confluent HepG2 cultures were incubated with methionine/cysteine-free minimal essential medium for 120 min, pulse-labeled (15 min to 2 h, see figure legends) with 100 Ci/ml [ 35 S]methionine ϩ cysteine (Pro-Mix TM , NEN Life Science Products), washed 3 times and chased for 10 min, in culture medium supplemented with 10 mM methionine and 2 mM cysteine. The cells were then washed and incubated in CSK buffer (0.3 M sucrose, 0.1 M KCl, 2.5 mM MgCl 2 , 1 mM sodium-free EDTA, 10 mM PIPES, pH 6.8) containing 50 g/ml digitonin for 10 min at room temperature. Digitonized cells were washed three times in CSK buffer and then incubated in CSK buffer under various conditions as described under figure legends. The cells were solubilizied in a solubilization buffer (phosphate-buffered saline containing 1% Nonidet P-40, 1% deoxycholate, 5 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 100 KIU/ml Trasylol, 0.1 mM leupeptin, 0.5 M ALLN). Cell extracts were centrifuged in a microcentrifuge at 14,000 rpm for 10 min, and the supernatant was subjected to immunoprecipitation.
Subcellular Fractionation and Isolation of Total Microsomes-Total microsomes were isolated from intact cells (4 ϫ 60-mm dishes) essentially as described (3,5,17,25). In some experiments, total microsomes were extracted with sodium carbonate and fractionated by ultracentrifugation as described (3,5,17,25) to isolate a luminal fraction and a membrane-enriched fraction. The membrane and luminal fractions were then subjected to immunoprecipitation as described below.
Cross-Linking of Cells with DSP-To cross-link HepG2 cells, the cells were normally first permeabilized with digitonin as described above. Permeabilized cells were incubated with DSP (1 mM) for 30 min on ice, and the cross-linking reaction was stopped by the addition of 2 mM glycine and incubation for 15 min on ice. Cross-linked cells were washed with phosphate-buffered saline, solubilized in solubilization buffer as above, and subjected to immunoprecipitation.
Immunoprecipitation and Immunoblotting-Immunoprecipitation was performed using Immunoprecipitin (Life Technologies) as described previously (12,26). In procedures where monoclonal antibodies were employed they were first bound to Affi-Gel beads (Bio-Rad) then incubated with the samples overnight at 4°C and centrifuged to pellet the beads. Immunoprecipitates were washed three times with wash buffer and were prepared for SDS-PAGE (12,26).
Immunoblotting was performed by SDS-PAGE analysis of cell lysates of normal or cross-linked cells, followed by electrophoretic transfer of proteins onto nitrocellulose membranes using a Bio-Rad Semi-Dry blotter. Gels used for immunoblotting experiments were minigels (8 ϫ 5 cm) and were either 10% resolving (for ER-60) or 4.5% resolving (for apoB). The membranes were washed, blocked with a 5% solution of fat-free dry milk powder, and then incubated with a primary antibody. After washing, the membranes were subsequently incubated with a second antibody conjugated to peroxidase. Detection of peroxidase activity was carried out using an AEC Turbo substrate (Dimension Lab Inc., Mississauga, Canada).

Detection of an ER-60 Protease Homologue in HepG2 Cells-
The ER-60 protease has been previously detected in rat liver (20). We first attempted to detect the presence of an ER-60 homologue in HepG2 cells. Two approaches were used. First, intact HepG2 cells were pulsed for 1 h, chased briefly to label intracellular proteins, and then immunoprecipitated with a polyclonal antibody against purified rat ER-60. As shown in Fig. 1A (lanes 3 and 4), we consistently detected a protein with an approximate size of 58 kDa following immunoprecipitation with the anti-rat ER-60 antibody. In contrast, non-immune rabbit serum (lanes 1 and 2) did not immunoprecipitate a similarly sized band from HepG2 cells, suggesting the specificity of the immunoprecipitation with the anti-rat ER-60 antibody. The 58-kDa band immunoprecipitated from HepG2 cells thus resembles rat ER-60 in both size and antigenecity. In a second approach, we solubilized HepG2 cells and separated the total cell extract by SDS-PAGE and immunoblotted with either non-immune rabbit serum or with rabbit anti-rat ER-60 antibody. Fig. 1B, lanes 5 and 6, shows the immunoblot of HepG2 cell lysate probed with non-immune rabbit serum. No bands were detectable having a size in the 58-kDa range. However, immunoblotting with the rabbit anti-rat ER-60 antibody (Fig. 1B, lanes 7 and 8) resulted in the detection of a protein with an approximate molecular size of 58 kDa. Thus, the immunoprecipitation and immunoblotting data both suggest the presence of a protein in HepG2 cells which resembles the rat ER-60 in size and antigenecity. Hence, we refer to this protein as an ER-60 protease homologue. Subcellular fractionation of HepG2 cells was also carried out to investigate the intracellular localization of the ER-60 homologue. Intact HepG2 cells were pulsed for 1 h, chased briefly, and then homogenized to isolate a total microsomal fraction. The microsomes were fractionated further by sodium carbonate extraction to prepare a membrane pellet and a soluble luminal content. The microsomal fractions were then immunoprecipitated with the anti-rat ER-60 antibody. As shown in Fig. 1C, the 58-kDa ER-60 homologue was readily detectable in the lumen of microsomes (lanes 7-9), although some signal was also detectable in the membrane fraction (lane 10). The data suggest that the ER-60 protease homologue in HepG2 cells is present in the microsomal fraction and appears to be predominately luminal in nature with some possible association with the microsomal membrane. Immunoblotting of the same microsomal fractions showed similar results (data not shown), further confirming the observations made by immunoprecipitation.
ApoB Is Associated Intracellularly with the ER-60 Protease ]protein labeling mixture, chased for 10 min, and subsequently subjected to immunoprecipitation either with non-immune rabbit serum or with a rabbit anti-rat ER-60 antibody (Ab). Immunoprecipitates were analyzed by SDS-PAGE and fluorography. Lanes 1 and 2, immunoprecipitation with non-immune rabbit serum. Lanes 3 and 4, immunoprecipitation with rabbit anti-rat ER-60 antibody. B, intact cells were solubilized and the cell lysate was analyzed by SDS-PAGE, followed by immunoblotting either with non-immune rabbit serum (left blot) or with rabbit anti-rat ER-60 antibody (right blot) as described under "Experimental Procedures." Lanes 5 and 6, immunoblotting with non-immune rabbit serum; lanes 7 and 8, immunoblotting with rabbit anti-rat ER-60 antibody. C, intact HepG2 cells were homogenized and then a total microsome fraction was isolated. The microsomes were treated with sodium carbonate and then centrifuged to separate the luminal contents (lanes 7-9) and the membrane (lane 10). Note that the luminal contents were divided into three fractions (lanes 7-9) and then immunoprecipitated and analyzed on SDS-PAGE due to the large volume of this fraction.
Homologue-First, the association of apoB with the ER-60 homologue was examined by pulse-chase labeling of HepG2 cells followed by co-immunoprecipitation of apoB and ER-60 homologue. Fig. 2A (lane 1) shows the immunoprecipitation of apoB. Goat anti-human apoB antibody detected the full-length apoB as well as a large number of shortened apoBs which mostly represent nascent apoB chains since the cells were pulsed for a long period of time but chased briefly. Interestingly, the apoB immunoprecipitate included a band migrating at around 58 kDa which may possibly represent the ER-60 homologue coimmunoprecipitating with apoB. A 58-kDa band was consistently detected in the apoB immunoprecipitate in several experiments. The ER-60 antibody immunoprecipitated a 58-kDa protein ( Fig. 2A, lane 2) similar to that observed in Fig. 1A.
We also attempted to co-immunoprecipitate apoB and ER-60. In this approach, labeled HepG2 cells were first immunoprecipitated with the anti-rat ER-60 antibody, the immunoprecipitate was eluted and then subjected to a second immunoprecipitation with monospecific goat anti-human apoB antibody. A small amount of apoB was detected ( Fig. 2A, lane 3) in the ER-60 immunoprecipitate. This further suggests that the ER-60 homologue in HepG2 cells is associated with a small pool of apoB.
Experiments were also performed to demonstrate the association of apoB with the ER-60 homologue in CHAPS-solubilized HepG2 cells. Solubilization in CHAPS appears to provide a less stringent condition allowing for a better detection of intracellular protein-protein interactions (28), and has been previously employed to study intracellular association of ER chaperones with secretory and membrane proteins. As shown in Fig. 2B, immunoprecipitation of apoB from CHAPS-solubilized HepG2 cells resulted in the detection of a number of additional proteins, which may represent proteins genuinely associated with apoB or nonspecifically immunoprecipitated. Probing CHAPS-solubilized HepG2 cells with the anti-rat ER-60 antibody again detected a 58-kDa protein, however, other bands were also detected in the ER-60 immunoprecipitate including a protein migrating at 550 kDa. Thus, the data from CHAPS-solubilized HepG2 cells appears to support the notion that apoB and the ER-60 homologue may be intracellularly associated.
To confirm the observations in Fig. 2, A and B, we performed experiments in which 35 S-labeled HepG2 lysates were first immunoprecipitated with either non-immune rabbit serum or non-immune goat serum. The immunoprecipitates were then dissociated and re-immunoprecipitated with either rabbit antirat ER-60 antibody or with goat anti-human apoB, respectively. As shown in Fig. 2C, no detectable bands were observed in either co-immunoprecipitation experiments, suggesting that the small amount of apoB detected in Fig. 2A, lane C, is not due to nonspecific co-immunoprecipitation. Further confirmation for the specific association of the ER-60 protease homologue and apoB came from the experiment in Fig. 2D. In this experiment, HepG2 lysates were first immunoprecipitated with either non-immune rabbit serum or with rabbit anti-rat ER-60 antibody. Both immunoprecipitates were dissociated and reimmunoprecipitated with goat anti-human apoB antibody. The immunoprecipitates were then fractionated by SDS-PAGE and immunoblotted with goat anti-human apoB antibody. As shown in Fig. 2D, no detectable band was observed when the HepG2 lysates were first immunoprecipitated with non-immune rabbit serum (lane 1). However, immunoprecipitation with rabbit anti-rat ER-60 antibody (lane 2) clearly resulted in the detection of a 550-kDa protein when immunoblotted with goat antihuman apoB antibody. These observations thus elaborate our earlier observations suggesting the intracellular association of apoB and ER-60 protease homologue in HepG2 cells.
ApoB Can Be Cross-linked to the ER-60 Protease Homologue in Permeabilized HepG2 Cells-We also attempted to cross-link apoB and the ER-60 protease homologue using a commonly used chemical cross-linking agent, DSP. Preliminary experiments with DSP showed that prior permeabilization of HepG2 cells increased the efficiency of DSP-mediated cross-linking of proteins and enhanced the sensitivity of detection of apoB cross-linking to ER-60. We thus performed all our subsequent cross-linking experiments in permeabilized cells. Intact HepG2 cells were pulsed, briefly chased, and then permeabilized with digitonin. Permeabilized cells were then cross-linked with DSP, solubilized, and immunoprecipitated with antibodies FIG. 2. Co-immunoprecipitation of ER-60 and apoB. A, near confluent cells were pulsed for 2 h with [ 35 S]protein labeling mixture, chased for 10 min, and immunoprecipitated with goat anti-apoB antibody or rabbit anti-rat ER-60 antibody (Ab). In the third lane, labeled cell extracts were first immunoprecipitated with anti-rat ER-60 antibody and then the immunoprecipitates were eluted and subjected to a second immunoprecipitation with goat anti-apoB antibody. Immunoprecipitates were analyzed by SDS-PAGE and fluorography. B, intact cells were pulsed and chased as in A and solubilized in a CHAPS (1%) solubilization buffer. Cell lysates were then immunoprecipitated with anti-apoB antibody or anti-rat ER-60 antibody, and analyzed by SDS-PAGE and fluorography. C, intact cells were pulsed and chased as in A and solubilized in the solubilization buffer containing Nonidet P-40 and deoxycholate. Cell lysates were then immunoprecipitated with either non-immune rabbit serum (lane 1) or non-immune goat serum (lane 2). The immunoprecipitates were dissociated and re-immunoprecipitated either with rabbit anti-rat ER-60 antibody (lane 1) or with goat antihuman apoB antibody (lane 2), respectively. The second immunoprecipitates were then analyzed by SDS-PAGE and fluorography. D, intact cells were solubilized and the cell lysate was immunoprecipitated either with non-immune rabbit serum (lane 1) or with rabbit anti-rat ER-60 antibody (lane 2). Both immunoprecipitates were then dissociated and re-immunoprecipitated with goat anti-human apoB antibody. The second immunoprecipitates were analyzed by SDS-PAGE and immunoblotted with goat anti-human apoB antibody as described under "Experimental Procedures." against apoB and ER-60. The immunoprecipitates were treated with a high concentration of DTT to dissociate cross-linked proteins before analysis by SDS-PAGE and fluorography. Fig.  3A shows the apoB immunoprecipitates from control and DSPcross-linked permeabilized HepG2 cells. As shown in Fig. 3A (lane 1) goat anti-human apoB antibody consistently recovered from non-cross-linked cells both the full-length apoB, as well as a number of nascent chains. Cross-linking of HepG2 cells resulted in the recovery of apoB, its nascent chains, as well as a few other proteins (Fig. 3A, lane 2). Two bands with approximate sizes of 58 and 78 kDa became more visible in the apoB immunoprecipitates when cells were cross-linked with DSP (observed in several independent experiments).
When probed with the anti-rat ER-60 antibody (Fig. 3A, lane  3), a single 58-kDa protein band was detected in control cells (as expected). A strong band was also visible with an approximate size of 300 -400 kDa. This band has been previously observed when immunoprecipitation is performed with Immu-noprecipitin. It is detected in some immunoprecipitation experiments (when a preclearing step with Immunoprecipitin is not performed) and is the result of nonspecific binding of the unknown protein with the Immunoprecipitin (the protein can be recovered by incubating the HepG2 lysate with Immunoprecipitin without any added antibody) (data not shown).
Four bands were recovered from cross-linked cells when immunoprecipitated with rabbit anti-rat ER-60 antibody (50, 58, 66, and 78 kDa) (Fig. 3A, lane 4). Among these four bands, the 58-kDa band appeared to represent the ER-60 homologue itself. The ER-60 homologue was consistently found to cross-link to the protein species at 50, 66, and 78 kDa in several experiments, two of which are shown in Fig. 3B. In some immunoprecipitates, a small amount of the 550-kDa apoB was also detected after cross-linking of permeabilized HepG2 cells and immunoprecipitation with the anti-ER-60 antibody. However, long exposures of the fluorograph were needed to visualize the apoB bands, suggesting the cross-linking of only a very small amount of apoB with ER-60 homologue.
To further confirm the association of apoB and ER-60, permeabilized and cross-linked HepG2 cells were solubilized and immunoprecipitated with the goat anti-human apoB antibody. The apoB immunoprecipitates were then analyzed by SDS-PAGE and the gel was further immunoblotted with the anti-rat ER-60 antibody (Fig. 3C). As shown in Fig. 3C, a 58-kDa protein was clearly detectable in the apoB immunoprecipitate from cross-linked HepG2 cells. We also detected the ER-60 protease homologue in the apoB immunoprecipitates from control cells (Fig. 3C, lane 1) upon immunoblotting, but crosslinking with DSP consistently resulted in an increase in the abundance of the 58-kDa ER-60 protease homologue (Fig. 3C,  lane 2).
Finally, to examine the possibility of nonspecific immunoprecipitation in the experiments above, we performed an experiment in which control cells and cells cross-linked with DSP were immunoprecipitated with non-immune goat serum (Fig.  3D). The immunoprecipitates were then analyzed by SDS-PAGE and immunoblotted with rabbit anti-rat ER-60 antibody. No bands were detected on the blot, suggesting that the detection of a 58-kDa protein in Fig. 3C was not due to nonspecific immunoprecipitation.
ApoB and ER-60 Protease Homologue Are Both Associated with a 78-kDa Protein Resembling BiP-As described above, immunoprecipitation of apoB and ER-60 from DSP-crosslinked HepG2 cells consistently resulted in the detection of a 78-kDa protein. Considering the size homology of this protein species with the molecular chaperone, BiP, we examined whether this protein is indeed BiP by immunoprecipitating cross-linked HepG2 cells with a monoclonal antibody against BiP. Non-immune mouse serum did not detect the 78-kDa protein (Fig. 4A, lane 1). The mouse anti-BiP antibody detected a 78-kDa protein as expected, as well as a number of other proteins (Fig. 4A, lane 3). The cross-linking of BiP with several intracellular proteins is clearly expected considering the role of BiP in the folding of many secretory proteins. Interestingly, among the proteins cross-linked to BiP was a band at 550-kDa protein comparable in size to apoB (Fig. 4D, lane 2). A 58-kDa protein resembling ER-60 in size was also detectable in the BiP immunoprecipitate.
To further investigate intracellular association with BiP, HepG2 cells were cross-linked and then immunoprecipitated with non-immune mouse serum bound to Affi-Gel beads (Fig.  4B, lane 1) or mouse anti-BiP antibody bound to Affi-Gel beads (Fig. 4B, lane 2). The immunoprecipitates were analyzed on two gels (a 10% gel for blotting with ER-60 antibody and a 4.5% gel for blotting with apoB antibody). The gels were then immu-

FIG. 3. Cross-linking of ER-60 and apoB in permeabilized HepG2 cells.
A, near confluent cells were pulsed for 2 h with [ 35 S]protein labeling mixture and chased for 10 min with excess cold methionine and cysteine. Cells were then permeabilized with digitonin, and crosslinked with DSP. Cross-linked cells were solubilized and subjected to immunoprecipitation with anti-apoB antibody or anti-rat ER-60. Immunoprecipitates were treated with sample buffer containing 400 mM DTT to reduce the cross-links and then analyzed by SDS-PAGE and fluorography. B, HepG2 cells were pulsed, permeabilized, and crosslinked as in A, and immunoprecipitated with rabbit anti-rat ER-60 antibody. The immunoprecipitates were treated as in A and then analyzed by SDS-PAGE and fluorography. The figure shows the results using duplicate dishes with (lanes 2 and 4) or without (lanes 1 and 3) cross-linking with DSP. C, HepG2 cells were permeabilized and crosslinked as in A. Control and cross-linked cells were immunoprecipitated with goat anti-apoB antibody bound to Affi-Gel. The immunoprecipitates were treated as in A and then analyzed by SDS-PAGE, followed by immunoblotting with rabbit anti-rat ER-60 antibody. Lane 1, immunoprecipitation of non-cross-linked cells. Lane 2, immunoprecipitation of cross-linked cells. D, HepG2 cells were permeabilized and cross-linked as in A. Control and cross-linked cells were immunoprecipitated with non-immune goat serum. The immunoprecipitates were treated with sample buffer to reduce the cross-links, and then analyzed by SDS-PAGE, followed by immunoblotting with rabbit anti-rat ER-60 antibody. Lane 3, non-cross-linked cells. Lane 4, cross-linked cells.
noblotted with either rabbit anti-rat ER-60 antibody or goat anti-human apoB antibody. Immunoblotting with anti-ER-60 antibody detected a 58-kDa protein in the BiP immunoprecipitate but not in the immunoprecipitate from non-immune serum. Similarly, immunoblotting with anti-human apoB antibody detected a 550-kDa protein in the BiP immunoprecipitate but not in the immunoprecipitate from non-immune serum. The data thus confirms the association of both the ER-60 protease homologue as well as apoB with the 78-kDa BiP protein.
In another experiment (Fig. 4C), control and DSP-crosslinked HepG2 cells were immunoprecipitated with rabbit antirat ER-60 antibody (lanes 1 and 2, respectively). Again, a single 58-kDa band was detected in control cells whereas a 58-kDa band as well as three other proteins with molecular sizes of 78, 66, and 50 kDa were observed in the immunoprecipitate from cross-linked cells. To examine the identity of the 78-kDa protein, DSP-cross-linked cells were first immunoprecipitated with rabbit anti-rat ER-60 antibody and the immunoprecipitate was dissociated in a high DTT buffer (see "Experimental Procedures"). The sample was then re-immunoprecipitated with mouse anti-BiP antibody bound to Affi-Gel. As shown in Fig. 4C, lane 3, the 78-kDa band was recovered following the second immunoprecipitation with the BiP antibody. Interest-ingly, a second band around 50 kDa was also recovered in this experiment. Since the BiP antibody is raised against a KDEL sequence and has been reported to recover a 50-kDa protein previously (StressGen Technical Bulletin), it is possible that the 50-kDa protein may also represent a KDEL-bearing protein. Overall the cross-linking of a 78-kDa protein with both apoB and the 58-kDa ER-60 suggest intracellular association of BiP with both apoB and ER-60 protease homologue. DISCUSSION Recent evidence from our laboratory suggest that apoB may be degraded both when associated with the ER membrane as well as when assembled into secretion competent lipoprotein particles (17). The kinetics of degradation of membrane-bound apoB and lipoprotein-associated apoB appeared to differ, with the membrane-bound apoB being degraded very rapidly (17). These findings raised the possibility that more than one proteolytic system may be involved in the degradation of different apoB pools. The degradation of lipoprotein-associated apoB in the lumen of the ER suggested that ER-resident proteases may be involved in this degradation. In the present report, we present evidence for association of apoB with an ER protease, ER-60. The evidence from co-immunoprecipitation and chemical cross-linking experiments were consistent in showing an intracellular association of apoB with an ER-60 protease homologue found in HepG2 cells. To our knowledge, this is the first report showing the presence of an ER-60-like protease homologue in HepG2 cells. It is also interesting that this protease is intracellularly associated with apoB, raising the possibility of its involvement in the apoB degradation process.
Cross-linking experiments in permeabilized HepG2 cells showed that the ER-60 protease homologue is intracellularly associated with other proteins with molecular masses of 50, 66, and 78 kDa. The 78-kDa protein which consistently crosslinked to ER-60 appeared to represent BiP. In addition, immunoprecipitation of BiP from CHAPS-solubilized HepG2 cells as well as cross-linking experiments showed that BiP may also be associated with apoB. These findings raise the intriguing possibility of a three way association between apoB, ER-60, and BiP in the ER lumen. Whether association of ER-60 and apoB with BiP is an indication of a regulatory process for targeting of apoB for degradation remains to be demonstrated and requires further investigation. It would not be surprising if apoB associates with BiP in the lumen of the ER since the folding and biogenesis of the extremely large apoB polypeptide would be expected to involve the action of intraluminal molecular chaperones. ApoB is already known to associate with microsomal triglyceride transfer protein, protein disulfide isomerase, and possibly calnexin (29,30) during its biogenesis in the ER. Patel and Grundy (30) also showed association of apoB41 with KDEL-bearing proteins including a 78-kDa protein resembling BiP in COS cells.
The fraction of newly-synthesized apoB associated with the ER-60 homologue appears to be quantitatively very low. In several experiments performed it appears that only a small fraction of newly-synthesized apoB pool could be recovered in association with the 58-kDa ER-60 homologue. This finding is not surprising and should be expected since it is unlikely that a large portion of the apoB pool would be associated with the ER-60 homologue at any given time. In addition, a major fraction of the newly-synthesized apoB pool is normally found on the cytosolic face of the ER membrane and is unlikely to associate with the ER-60 homologue which appears to be luminal in nature. Presumably only the apoB pool that is fully-translocated and is present in the ER lumen would be targeted by a luminal protease such as the ER-60 protease homologue.
Although we do not have direct evidence implicating the FIG. 4. Intracellular association of ER-60 protease and apoB with the 78-kDa BiP. A, near confluent cells were pulsed for 2 h with [ 35 S]protein labeling mixture and chased for 10 min with excess cold methionine and cysteine. Cells were then permeabilized with digitonin (50 g/ml) for 10 min and washed three times with CSK buffer. Permeabilized cells were cross-linked as in Fig. 3A. Cross-linked cells were solubilized and subjected to immunoprecipitation with non-immune mouse serum bound to Affi-Gel (lane 1), goat anti-human apoB antibody (lane 2), and mouse anti-BiP antibody bound to Affi-Gel (lane 3). Immunoprecipitates were treated with sample buffer containing 400 mM DTT to reduce the cross-links and then analyzed by SDS-PAGE and fluorography. B, HepG2 cells were permeabilized and cross-linked as in A. Permeabilized cells were immunoprecipitated either with non-immune mouse serum bound to Affi-Gel (lane 1 in both blots) or with mouse anti-BiP antibody bound to Affi-Gel (lane 2 in both blots). The immunoprecipitates were analyzed by SDS-PAGE, followed by immunoblotting either with rabbit anti-rat ER-60 antibody (left blot) or with goat anti-human apoB antibody (right blot). C, HepG2 cells were pulsed, chased, permeabilized, and cross-linked as in A. Cross-linked cells were solubilized and immunoprecipitated with rabbit anti-rat ER-60 antibody (lanes 1-3). The immunoprecipitates were treated with sample buffer to reduce the cross-links. Lane 1, non-cross-linked cells; lane 2, cross-linked cells. The sample in lane 3 was re-immunoprecipitated with mouse anti-BiP antibody bound to Affi-Gel, before analysis by SDS-PAGE and fluorography.
ER-60 protease homologue in apoB degradation, we hypothesize that the ER-60 protease is involved in intraluminal degradation of apoB in the ER, based on the evidence showing apoB-ER-60 association and the cysteine protease activity of ER-60 (20,21). Recent evidence from our laboratory 2 and others (18) suggest the existence of an ER-localized proteolytic system distinct from the ubiquitin-proteasome system. Ginsberg and co-workers (18) reported that even after apoB is translocated into the ER lumen, it can potentially be degraded by a luminal DTT-sensitive degradative system that may be responsible for degrading the N-terminal 70-kDa apoB fragment. We have previously reported that degradation of apoB in permeabilized HepG2 cells can generate specific apoB fragments in the ER including an abundant 70-kDa intermediate (12,16,17,26). More recent experiments in our laboratory support the hypothesis that apoB degradation may occur in two steps. 3 Our data in the present report demonstrating the association of apoB with an ER-resident protease, ER-60, supports the above notion that proteases other than the proteasome may be involved in the apoB degradation process. It is thus reasonable to hypothesize that the ER-60 protease homologue detected in HepG2 cells may be involved in a luminal proteolytic system that degrades fully-translocated apoB. The luminal process appears to be distinct from the proteasome-mediated degradative process (19) operating on the cytosolic side of the ER membrane, which degrades the bulk of apoB co-translationally. Efforts are underway to further characterize the ERassociated degradation pathway and confirm its subcellular localization and the involvement of the ER-60 protease in this pathway. It would also be interesting to investigate the role of the ER-associated protease in degradation of nascent apoBcontaining lipoprotein particles in the lumen of the ER, and the possible association of the ER-60 protease homologue with these lipoprotein particles.