Reassessment of the subcellular localization of p63

The secretory pathway of eukaryotic cells is subdivided into a series of different membrane compartments of defined molecular composition (Palade, 1975). Newly synthesized exocytic proteins enter this pathway at the rough endoplasmic reticulum (ER) and are subsequenlty transported to the Golgi apparatus through a recently identified ER-Golgi intermediate compartment (Saraste and Kuismanen, 1984; Tooze et al., 1988; Schweizer et al., 1990; Hauri and Schweizer, 1992; Lotti et al., 1992).

The ER is the largest continous endomembrane system in the cell and fulfills a wide variety of functions. These include synthesis, translocation, glycosylation, folding, assembly, and processing of secretory and membrane proteins, lipid synthesis and metabolism, as well as sorting and degradation of proteins, and regulation of intracellular calcium. Morphologically and biochemically the ER has been subdivided into three distinct but continuous (Fawcett, 1981) subcompartments, the smooth ER, the rough ER (containing bound ribosomes), and the nuclear envelope (Palade, 1975; and references therein). While the overall protein and lipid composition of smooth and rough ER membranes is very similar (Hinmann and Philipps, 1970; Colbeau et al., 1971; Sharma et al., 1978), a certain number of proteins such as the ribophorins (Kreibich et al., 1978), the docking protein (Hortsch and Meyer, 1985; Hortsch et al.,1985), and the signal sequence receptor now called the translocon-associated protein complex (Vogel et al., 1990; Hartmann et al., 1993) are mainly restricted to the rough ER.

The ER-Golgi intermediate compartment (ERGIC) is a complex tubulo-vesicular membrane system localized primarily at the cis side of the Golgi apparatus but also extending out into the cell periphery. It is best defined by antibodies to ERGIC-53, a 53 kDa non-glycosylated type I transmembrane protein (Schweizer et al., 1988; Schindler et al., 1993) or its putative rat homologue p58 (Lahtinen et al., 1992). The small GTP-binding protein rab2p (Chavrier et al., 1990; Sodeik et al., 1993) and beta-COP, a major component of nonclathrin-coated vesicles, are also associated with the ERGIC (Duden et al., 1991; Oprins et al., 1993). The ERGIC is the site at which ER-to-Golgi protein transport is blocked at 15°C (Schweizer et al., 1990), and it is most likely equivalent to the ‘budding compartment’ of mouse coronavirus (Tooze et al., 1988; Krijnse-Locker et al., 1994) and to the ‘pre-Golgi vacuoles’ of Semliki Forest virus-infected cells (Saraste and Kuismanen, 1984). However, it is still unclear whether the ERGIC shares membrane continuities with the rough ER as has been postulated on the basis of studies of the mouse hepatitis virus budding compartment (Krijnse-Locker et al., 1994), or whether it is a separate transport intermediate (Plutner et al., 1992; Saraste and Kuismanen, 1992; Lippincott-Schwartz, 1993; Banfield et al., 1994).

Recently, a novel 63 kDa marker protein (p63) was identified using a monoclonal antibody (mAb) approach (Schweizer et al., 1993a,b). p63 is a non-glycosylated, reversibly palmitoylated transmembrane protein with a type II orientation. By indirect immunofluorescence mAbs against p63 labeled an extended ER-like membranous network in Vero cells. However, the p63 pattern did not include prominent staining of the outer nuclear membrane as is typically found for rough ER markers, and it showed partial overlap with the ERGIC marker protein ERGIC-53. Unlike ERGIC-53, which follows a cycling pathway (Lippincott-Schwartz et al., 1990; Hauri and Schweizer, 1992), the distribution of p63 was insensitive to organelle perturbants such as low temperature and brefeldin A (Schweizer et al., 1993a), suggesting that p63 is a resident protein. An immunoelectron microscopy analysis of p63 revealed isolated labeling of tubulo-vesicular structures close to the nucleus and the Golgi apparatus. However, the interpretation of these findings was limited by the low labeling intensity achieved with the anti-p63 mAb. From these collective findings, p63 was suggested to have a stable ER-Golgi intermediate localization.

To achieve a more precise and quantitative localization of p63, we have produced a rabbit polyclonal antiserum against the purified protein. This new anti-p63 antibody has been used to reassess the subcellular distribution of p63 as determined by confocal microscopy, immunoelectron microscopy, and a biochemical analysis of rough and smooth membranes derived from rat liver. The results presented in this paper demonstrate that p63 is a resident protein of the rough ER and not of ER-Golgi intermediate structures.

DME (4.5 g/l glucose) and RPMI-1640 medium were obtained from GIBCO BRL (Grand Island, NY). Fetal calf serum (FCS) was from Hazelton Biologics (Lenexa, KS); Nusera from Collaborative Biomedical Products (Bedford, MA); DEAE-dextran, chloroquine and protease inhibitors from Sigma Chemical Co. (St. Louis, MO); ECL western blotting reagents from Amersham Corp. (Arlington Heights, IL); Protein A-Sepharose beads from Repligen Corporation (Cambridge, MA); UDP-[¹⁴C]galactose from Du Pont New England Nuclear (Boston, MA); enzyme-grade sucrose from J.T. Baker Inc. (Phillipsburg, NJ); cell culture dishes from Falcon (Becton Dickinson Co., Lincoln Park, NJ); multichamber slides from Nunc Inc. (Naperville, IL); FITC goat anti-mouse IgG, FITC goat anti-rabbit IgG, and Texas Red goat anti-rabbit IgG from Cappel (Westchester, PA); Sprague Dawley male rats from Harlan Industries; and human plasma fibronectin from the New York Blood Center (NY).

Plasmid pECE-p63 was described previously (Schweizer et al., 1994). Plasmid CDMK encoding the entire cathepsin D cDNA sequence followed by SMEQKLISEEDLNSE KDEL (Pelham, 1988) was kindly provided by Dr H. Pelham (MRC Cambridge, UK).

COS cells (African green monkey kidney cells, CRL 1650; American Type Culture Collection, Rockville, MD) were cultured in DME supplemented with 10% FCS, 50 units/ml penicillin, 50 μg/ml streptomycin, and Fungizone at 37°C in a humidified 5% CO₂ atmosphere. Transient transfection of COS cells (grown in eight-well multicham-ber slides for immunofluorescence and 60-mm plates for immunoelectron microsopy) was performed as described (Schweizer et al., 1994). For the expression of p63wt, plasmid pECE-p63 was transfected while plasmid CDMK was used to express a cathepsin D-c-myc-KDEL fusion protein.

Mouse mAb G1/296 against the p63 protein has previously been characterized (Schweizer et al., 1993a). Hybridomas producing a mAb (9E10) raised against a synthetic peptide comprising residues 409-439 of human c-myc (Evan et al., 1985) was obtained from the American Type Culture Collection. Rabbit polyclonal antisera against p58 (Saraste and Svensson, 1991) and mouse mAb ID3 against the protein disulfide isomerase (PDI) tail (KDDDKAVKDEL) were kind gifts from Dr J. Saraste (University of Bergen, Norway) and Dr S. Fuller (EMBL Heidelberg, Germany), respectively. Mouse mAbs against ribophorin II (Hortsch et al., 1986) and ERGIC-53 (Schweizer et al., 1988) were generously provided by Dr D. Meyer (University of California, Los Angeles) and Dr. H.-P. Hauri (Biozentrum Basel, Switzerland), respectively.

A polyclonal antibody against the native p63 protein was produced as follows. p63 protein was affinity-purified from p63wt-transfected COS cells as described by Schweizer et al. (1994), except that pooled fractions were dialyzed against 100 mM phosphate buffer, pH 8.0, containing 0.8% NaCl. A 75 μg sample of purified p63 protein was mixed with complete Freund’s adjuvant and the resulting emulsion was injected into the foodpads of a New Zealand White rabbit. A booster injection with 45 μg of purified antigen in incomplete Freund’s adjuvant was given three weeks later by the same route. Ten days after the booster, serum containing polyclonal antibodies against p63 was harvested. For the detection of p63 in rat, the polyclonal antibody was affinity-purified via p63 blotstrips. For this purpose the p63 from ten 60-mm dishes of p63wt-transfected COS cells was immunoprecipitated with mAb G1/296, separated by preparative SDS-PAGE and transferred to nitrocellulose. After staining with Ponceau S, the nitrocellulose strip containing the p63 band was cut into pieces, blocked with 3% nonfat dry milk powder in phosphate-buffered saline (PBS), and incubated for 3 hours at room temperature with 200 μl of the polyclonal anti-p63 antiserum in blocking solution. After three wash steps with PBS, specific antibodies were eluted with 1 ml 100 mM glycine, pH 2.5 (for 10 minutes at room temperature), and immediately neutralized.

COS cells were grown in eight-well multichamber slides. The immunofluorescence procedure was that of Schweizer et al. (1988). In brief, formaldehyde-fixed and saponin-permeabilized cells were incubated with mAb G1/296 against p63, followed by goat anti-mouse FITC or with the anti-p63 polyclonal antibody followed by goat antirabbit FITC. For double immunofluorescence, the formaldehydefixed, saponin-permeabilized cells were sequentially incubated with the anti-p63 polyclonal antibody, goat anti-rabbit Texas Red, mAb 9E10, and goat anti-mouse FITC. The specimens were examined with a BioRad MRC 1000 confocal laser scanning microscope system attached to a Zeiss microscope.

HepG2 cells and COS cells were fixed in a mixture of 2% paraformaldehyde and 1% acrolein (final concentrations) in 0.1 M sodium phosphate buffer, pH 7.4, for 2 hours and kept in 1% paraformaldehyde in the same buffer until further processing. Cell samples were embedded in 10% gelatin on ice. Blocks with cells were immersed in 2.3 M sucrose in phosphate buffer at 4°C and ultrathin cryosections were single or double immuno-labeled with 10 nm and 15 nm Protein A-conjugated colloidal gold probes (Slot and Geuze, 1985; Slot et al., 1988). The sections were picked-up from the knives according to the method of Liou and Slot (1994).

All solutions contained a 1:500 dilution of a protease inhibitor cocktail (5 mg/ml benzamidine, and 1 mg/ml each of pepstatin A, leupeptin, antipain and chymostatin in 40% dimethylsulfoxide, 60% ethanol) and 0.17 TIU/ml of aprotinin. One male Sprague-Dawley rat (150 g body weight) was starved overnight before killing. The liver was homogenized in 1 M sucrose (4 ml homogenization buffer per 1 g liver tissue) in a motor-driven tight-fitting teflon homogenizer before an equal volume of 2.5 M sucrose was added. This mixture was centrifuged for 45 minutes at 37,000 rpm in a Ti60 rotor (Beckman Instruments Inc., Palo Alto, CA). The postnuclear supernatant was rehomogenized, diluted with a half volume of water and then centrifuged for 15 minutes at 15,000 rpm in a JA21 rotor (Beckman Instruments Inc., Palo Alto, CA). The postmitochondrial supernatant (PMS) was adjusted to 1.3 M sucrose and discontinuous sucrose gradient centrifugation was performed according to Kreibich et al. (1978). Loose particulate matter that sedimented between smooth (SM) and rough (RM) membranes was collected and referred to as intermediate membranes (IM). Each membrane fraction was sedimented after dilution with 2 volumes of 0.25 M sucrose, 50 mM Tris-HCl, 25 mM KCl, 5 mM MgCl₂, pH 7.4 (0.25 M STKM), by centrifugation for 60 minutes at 40,000 rpm in a Ti60 rotor. The pellets were resuspended in 0.25 M STKM.

Galactosyltransferase activity was determined by monitoring the transfer of [¹⁴C]galactose from UDP-[¹⁴C]galactose to ovomucoid as described (enable and Coggeshall, 1965; Stieger et al., 1988) except that the reaction was done in the presence of 2 mM ATP for 15 minutes. Glucose-6-phosphatase was assayed as described by Aronson and Touster (1974) and released phosphate was measured by the procedure of Chen et al. (1956). Protein was determined with the Biorad protein assay kit using protein standard I (BioRad Laboratories, Richmond, CA), or with the Micro BCA protein assay (Pierce, Rockford, IL).

For the analysis of p63 in COS cells, mock-transfected or p63wt-transfected COS cells were washed with PBS, scraped into 1 ml PBS and centrifuged for 10 minutes at 800 rpm (132 g_av). The pellet was resuspended in PBS containing 40 μg/ml phenylmethylsulfonyl fluoride and the above described protease inhibitor cocktail by passing it five times through a 25-gauge needle connected to a 1 ml syringe. Aliquots corresponding to 20 μg of protein were adjusted to equal volumes of 35 μl. For the detection of p63 in rat liver fractions, aliquots corresponding to 30 μg of protein in 0.25 M STKM were adjusted to 0.1 M Na₂HPO₄ and 1% Triton X-100 in a final volume of 30 μl and solublized for 30 minutes on ice. For the detection of ribophorin II and p58 in rat liver fractions, aliquots corresponding to 30 μg of protein in 0.25 M STKM were adjusted to equal volumes of 30 μl.

Proteins were separated on 8% SDS-polyacrylamide minigels (BioRad Laboratories, Richmond, CA) using the Laemmli (1970) system, and transferred to nitrocellulose membranes according to the method of Towbin et al. (1979). For the immunoreaction, the nitro-cellulose sheet was blocked with either 3% nonfat dry milk powder in PBS in the presence (affinity-purified anti-p63 polyclonal antibody) or absence (anti-p63 polyclonal antibody, and anti-ribophorin II mAb) of 0.1% Tween 20, or with 5% nonfat dry milk powder in PBS containing 1% Tween 20 (anti-p58 polyclonal antibody), incubated with the antibody of choice in blocking solution followed by the appropriate horseradish peroxidase-conjugated secondary antibody (Amersham Corp., Arlington Heights, IL). For the development the ECL detection system (Amersham Corp., Arlington Heights, IL) was used according to the manufacturer’s directions.

Quantification of autoradiograms was carried out by means of a Molecular Dynamics Personal Densitometer (Sunnyvale, CA).

p63 protein affinity-purified from p63wt-transfected COS cells was used as an antigen to produce a rabbit polyclonal antibody. To verify that the polyclonal antibody specifically reacts with the p63 protein, COS cells were examined by immunoblotting (Fig. 1). In this experiment equal amounts of protein (20 μg per lane) of homogenates from either mock-transfected COS cells (lane1) or COS cells transfected with the human p63wt cDNA (lane 2) were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with the anti-p63 polyclonal antibody. In mock-transfected COS cells, a single protein with an apparent molecular mass of 63 kDa was detected (lane 1). As seen in lane 2, this signal was greatly increased when p63wt-transfected COS cells were probed.

We next analyzed the new antibody by indirect immuno-fluorescence on saponin-permeabilized COS cells using a confocal laser scanning microscope. Fig. 2B shows that the fluorescence pattern of the polyclonal antibody with its extended ER-like membrane structure was indistinguishable from that obtained with the anti-p63 mAb G1/296 (Fig. 2A). With both antibodies the staining pattern lacked the strong labeling of the outer nuclear membrane which is typically found for rough ER proteins.

To examine the relationship of the p63 positive structures and the rough ER in more detail, double immunofluorescence was performed. To this end, COS cells were transfected with plasmid CDMK bearing a fusion gene that encodes an ER marker protein (cathepsin D-c-myc-KDEL) consisting of the lysosomal enzyme cathepsin D, the c-myc epitope recognized by the monoclonal antibody 9E10 and the C-terminal ER-retention signal KDEL (Pelham, 1988); the expressed protein has previously been demonstrated to be efficiently retained in the ER (Pelham, 1988). Forty-three hours posttransfection the cells were double-stained for endogenous p63 using the polyclonal rabbit antiserum (Fig. 2C) and for the transfected cathepsin D-c-myc-KDEL with mAB 9E10 (Fig. 2D). The overall distribution of the two proteins was very similar except that the p63 pattern seemed somewhat less reticular. The only striking difference observed was in the staining of the outer nuclear membrane. While in cells stained for cathepsin D-c-myc-KDEL the nuclear envelope was strongly labeled and visible as a bright continuous line surrounding the nucleus, the same structure was either not stained or appeared as scattered dots when cells were labeled with the p63 antiserum.

Taken together, these data demonstrated that the newly generated polyclonal rabbit antiserum is a valid tool for analyzing the p63 protein. In addition, it became obvious that the precise localization of the p63 protein cannot be resolved at the level of light microscopy.

A previous attempt to localize p63 by immunoelectron microscopy was limited by the very low labeling intensity obtained with the mAb. Initial experiments with the anti-p63 polyclonal serum indicated that very specific and strong labeling for p63 could be achieved. It was therefore possible to carry out a detailed ultrastructural analysis to unequivocally establish the subcellular localization of the p63 protein.

First, the distribution of endogenous p63 was determined in HepG2 cells (Fig. 3). For this purpose ultrathin cryosections were labeled with the polyclonal antiserum against p63, followed by a second antibody and Protein A-gold. Almost all of the labeling was found over the rough ER (Fig. 3A, C), with some gold particles also present in the perinuclear cisternae (Fig. 3B). The Golgi apparatus was devoid of label (Fig. 3B). An equivalent distribution was found in untransfected Vero and COS cells (not shown). While the overall labeling in Vero cells was comparable to that in HepG2 cells, the staining was weaker in the COS cell line.

We next examined the localization of p63wt using cryosections of transfected COS cells (Fig. 4). In these preparations the polyclonal antibody gave very abundant labeling due to the overexpression of the p63 protein. The staining pattern in the transfected cells was essentially the same as described for endogenous p63 in HepG2 cells. The vast majority of the label was found over cisternae of the rough ER (Fig. 4A, B). Consistent with previous data (Schweizer et al., 1994), overexpression of p63 did not result in mislocalization to the Golgi (Fig. 4C) and plasma membrane.

The localization of p63 to the rough ER was further confirmed by double immunoelectron microscopy using the lumenal KDEL-containing protein PDI as an established marker for the ER. Double labeling of p63wt-transfected COS cells with the polyclonal antiserum against p63 and mAb ID3 against PDI showed extensive colocalization of the two markers in rough ER elements (Fig. 4B, C). Only occasionally were vesicles at the cis-Golgi face co-labeled with the anti-bodies. In addition, the labeling intensity of the perinuclear cisternae seemed less pronounced for p63 as compared to PDI (not shown). When the gold labeling of p63 and PDI in the per-inuclear membrane and the rough ER was quantitated in double-labeled sections, the p63/PDI ratio was about three times higher for the rough ER than for the perinuclear cisternae. A complete statistical analysis, however, was not possible due to variations in p63 and PDI expression between individual cells.

To define more precisely the relationship of p63 and the intermediate compartment marker protein ERGIC-53, HepG2 cells were double-labeled with mAb G1/93 (against ERGIC-53) and the anti-p63 polyclonal antiserum (Fig. 3B, C). In general, the two staining patterns did not overlap. ERGIC-53 was mainly present in tubulo-vesicular membrane profiles at the cis face of the Golgi apparatus (Fig. 3B) while p63 was hardly detected in the cis-Golgi region. In contrast, some membrane elements of the rough ER showed co-labeling for p63 and ERGIC-53 (Fig. 3C).

In summary, the immunoelectron microscopy data demonstrated that p63 is predominantely localized in the rough ER and not in ER-Golgi intermediate membranes.

To obtain independent evidence for the localization of p63 in the rough ER, cell fractionation studies were carried out using rat liver as the tissue source. The postmitochondrial supernatant (PMS) was fractionated according to the procedure of Kreibich et al. (1978) to obtain a rough ER fraction (RM) that was relatively free of smooth intracellular membranes (SM). As shown in Table 1, the Golgi marker, galactosyltransferase, was enriched in the SM fraction (5.7-fold over the PMS) and only 4% was recovered in rough membranes. In contrast, the RM fraction contained 52% of the glucose-6-phosphatase activity. This enzyme is a marker for the ER but is not restricted to the rough portion of the organelle. We next examined the various fractions for the content of a rough ER-specific protein, ribophorin II (Kreibich et al., 1978). The immunoblot shown in Fig. 5 and its quantification in Table 1 revealed that ribophorin II was almost quantitatively recovered in rough membranes and barely detected in the smooth membranes. In the RM fraction ribophorin II was enriched 7.3-fold over the PMS and more than 70-fold over SM. A very similar result was obtained when the distribution of p63 was analyzed by immunblotting using affinity-purified antibodies (Fig. 5 and Table 1 for quantification). p63 was almost exclusively detected in RM where it was enriched 6.7-fold over the PMS and 67-fold over SM. The distribution of p63 was clearly different from that obtained with the intermediate compartment marker p58 (Saraste and Svenson, 1991; Lahtinen et al., 1992) (Fig. 5, and Table 1 for quantification). Immunoblots probed with anti-p58 polyclonal antibodies showed that 49% of the protein was recovered in SM, 28% in intermediate membranes (IM) and only 8% in the RM fraction.

Taken together, the rat liver fractionation data demonstrated that p63 cofractionates with the rough ER marker ribophorin II but is largely separated from p58, a marker protein of the intermediate compartment.

In the present study we have established the subcellular distribution of p63 by combining biochemical and morphological approaches, and by using a novel anti-p63 polyclonal antibody. Both immunoelectron microscopy and subcellular fractionation clearly demonstrated that p63 is localized in the rough ER. This result is in contradiction to previous data (Schweizer et al., 1993a) that suggested a post-ER/pre-Golgi localization for p63. The postulated ER-Golgi intermediate localization of p63 was mainly based on an overlapping distribution found for p63 and the intermediate compartment marker protein ERGIC-53 in Vero cells by indirect immunofluorescence and immunoelectron microscopy. On cryosections of Vero cells p63 labeling was associated with tubulovesicular structures close to the nucleus and Golgi, some of which were positive for ERGIC-53. A major limitation with this ultrastructural analysis was the extremly low labeling intensity achieved with the anti-p63 mAb used in the study. The low level of labeling prevented a quantitative determination of the extent of overlap between p63 and ERGIC-53, and might explain why p63 staining was not detected in the expanded ER.

In contrast, the rabbit polyclonal antiserum used in the present study showed strong and specific labeling of p63 when ultrathin frozen sections were probed. In an extensive ultrastructural analysis, p63 was predominantly detected in membranes of the rough ER. Identical results were obtained for three different cell types as well as for endogenous and overexpressed p63. Colabeling of p63 and ERGIC-53 in HepG2 cells revealed significant overlap only in the rough ER. This finding suggests that the overlap in the p63 and ERGIC-53 distributions observed in previous double immunofluorescence experiments (Schweizer et al., 1993a) was most likely due to ERGIC-53 present in the rough ER and not p63 in the ER-Golgi intermediate compartment. The presence of ERGIC-53 in the ER is not unexpected, since the protein was postulated to recycle via the ER (Lippincott-Schwartz et al., 1990; Hauri and Schweizer, 1992).

An unexplained finding of the previous and present study is a difference in the immunofluorescence pattern between classical ER markers and p63. While the outer nuclear membrane was either not labeled or appeared as scattered dots when the anti-p63 antibodies were used, the nuclear ring was always strongly and continuously labeled with antibodies to the established ER markers. This difference was a major reason why p63 was not considered to have an ER localization in the initial studies. A difference in the labeling intensity of the perinuclear cisternae between p63 and a known ER marker (PDI) was also seen by immunoelectron microscopy. However, variations in the expresssion levels of the two proteins prevented a complete statistical analysis, which would be necessary to unequivocally establish this point. Therefore we can only speculate that the observed distribution for p63 may point to a subcompartmentalization of the rough ER.

Subcellular fractionation of rat liver confirmed the rough ER localization of p63 established by immunoelectron microscopy. The p63 protein together with ribophorin II was almost exclusively localized in the rough microsomal fraction. Only 4% of the Golgi enzyme galactosyltranferase and 8% of the intermediate compartment marker p58 were recovered in this fraction. At present it is not clear why p63 was separated from the other ER marker proteins, including PDI and ribophorin II, and cofractionated with ERGIC-53 when Vero cells were fractionated by a combination of Percoll and Metrizamide gradients (Schweizer et al., 1993a). Recently, co-fractionation of p63 with different markers of the ER was reported in Vero cells when a velocity-controlled Nycodenz step gradient was used for separation (Füllekrug et al., 1994). Interestingly, Chen et al. (1991, 1994) have described a novel rat antigen with similar biochemical properties to p63.

The antigen, a 65 kDa membrane protein (RER-65), was identified by a mAb and localized to the rough ER. It was reported, however, that RER-65 could not be detected in skeletal muscle, intestinal epithelium and pancreas, while p63 was clearly present in these tissues when assayed by northern blotting (skeletal muscle, pancreas; A. Schweizer, J. Rohrer and H.-P. Hauri, unpublished data) or immunofluorescence (human intestinal cell line Caco-2; Schweizer et al., 1993a). In addition, seven peptides of a tryptic digest of RER-65 showed strong homology to p63, while two peptides had no homology at all. It will be of considerable interest, therefore, to directly compare the complete primary sequences of p63 and RER-65. Previous work (Schweizer et al., 1994) has established that all three domains of p63 contribute to the proper intracellular localization of this protein. Further, the retention of p63 correlated with the formation of large Triton X-100-insoluble oligomers, suggesting that self-association of p63 protein may serve a major mechanism for retention. This aggregation could also explain how p63 is specifically restricted to the rough portion of the ER.

We thank Dr S. Fuller, Dr H.-P. Hauri, Dr D. Meyer and Dr J. Saraste for kindly providing antibodies against PDI, ERGIC-53, ribophorin II and p58, respectively. Plasmid CDMK was a kind gift from Dr H. Pelham. J. Griffith is acknowledged for her excellent preparative work in the immunoelectron microscopy. This work was supported by United States Public Service grant CA08759. J. Rohrer was the recipient of a Damon Runyon-Walter Winchell cancer postdoctoral fellowship. A. Schweizer was supported by a W.M. Keck fellowship.