PROTEASOME INHIBITION COMPROMISES DIRECT RETENTION OF CYTOCHROME P450 2C2 IN THE ENDOPLASMIC RETICULUM

Abstract

To determine whether protein degradation plays a role in the endoplasmic reticulum (ER) retention of cytochromes P450, the effects of proteasomal inhibitors on the expression and distribution of green fluorescent protein chimeras of CYP2C2 and related proteins was examined. In transfected cells, expression levels of chimeras of full-length CYP2C2 and its cytosolic domain, but not its N-terminal transmembrane sequence, were increased by proteasomal inhibition. Redistribution of all three chimeras from the reticular ER into a perinuclear compartment and, in a subset of cells, also to the cell surface was observed after proteasomal inhibition. Redistribution was blocked by the microtubular inhibitor, nocodazole, suggesting that redistribution to the cell surface followed the conventional vesicular transport pathway. Similar redistributions were detected for BAP31, a CYP2C2 binding chaperone; CYP2E1 and CYP3A4, which are also degraded by the proteasomal pathway; and for cytochrome P450 reductase , which does not undergo proteasomal degradation; but not for the ER membrane proteins, sec61 and calnexin. Redistribution does not result from saturation of an ER retention “receptor” since in some cases protein levels were unaffected. Proteasomal inhibition may, therefore, alter ER retention by affecting a protein critical for ER retention, either directly, or indirectly by affecting the composition of the ER membranes.

Introduction

Microsomal cytochromes P450 (CYPs) represent a class of integral endoplasmic reticulum (ER) membrane proteins whose retention in the ER is not well understood. Although it is well established that the N-terminal signal/anchor sequence of CYPs mediates the retention, no known ER retention signals are present in their primary sequences [1–3]. It is likely that multiple mechanisms are involved in retention of CYPs, since some CYPs are excluded from further transport and are retained in the ER directly (direct retention), whereas other CYPs are transported from the ER in a COPII-dependent manner and subsequently retrieved back to the ER [1, 3–5]. In addition, several microsomal P450s have been detected in compartments other than the ER, for example, human CYP2E1 has been detected in Golgi and plasma membranes [6–8] and CYP51 is transported through the Golgi complex to the acrosomal membrane in sperm [9]. The underlying mechanism of such unconventional localization for a small fraction of certain CYPs is not known, although the presence of the CYPs at the cell surface is mostly associated with pathological conditions, and can result in production of anti-CYP autoantibodies [6, 10, 11].

Microsomal CYPs are responsible for drug metabolism, and drug-induced expression of these proteins leads to a strong proliferation of endoplasmic reticulum membranes. Despite substantially elevated levels, the CYPs remain restricted to the smooth ER compartment [12–15]. If the inducer is removed, levels of CYPs and ER membrane decrease by mechanisms that are not well understood. Degradation of CYPs appears to involve several CYP-specific pathways [16]. CYP2E1 and CYP3A4 have been shown to be degradated via the proteasomal pathway [17, 18], whereas other CYPs are degradated by a lysosomal pathway [16].

ER-associated degradation (ERAD) mediated by proteasomes represents an important part of the ER quality control system responsible for removal of unwanted and misfolded proteins and its activity is correlated with many metabolic disorders [19, 20]. Degradation of newly synthesized misfolded proteins destined for compartments other than the ER probably involves their transport to a specialized ER quality control compartment (ERQC), however the mechanism and site of degradation of integral ER-specific proteins is less clear [21–23]. Some ER membrane proteins appear to accumulate in a membrane-free cytosolic structure called the aggresome, which is formed following dislocation from the ER membrane [21]. Other ER membrane proteins undergo proteasomal destruction from a membrane-containing location with or without undergoing transport to a specialized compartment(s) [22, 24, 25]. Quality control operates at several distinct checkpoints along the secretory pathway, so that degradation of some proteins requires functional COPII-dependent ER-Golgi transport, while others are degraded in a membrane compartment independent of COPII [26–28].

ERAD and quality control of proteins entering the ER involve the action of different chaperones that assist in protein folding, transport and degradation [23, 29]. Interestingly, both cytosolic and ER localized chaperones influence the mobility of proteins in the ER membranes, affecting their stability and ER retention [30]. Recently we have shown that the integral ER membrane protein BAP31, a chaperone which is known to play a role in ER retention and transport regulation, interacts with CYP2C2. Downregulation of BAP31 expression results in relocalization of CYP2C2 into a perinuclear compartment and increases the levels of CYP2C2 in the ER [31]. This suggested that the interaction with BAP31 may play a role in the quality control of CYP and increase its degradation, but it also raised the questions of whether the altered location of CYP2C2 resulted in decreased degradation or whether the increased levels of CYP2C2 resulted in altered localization. If direct retention of CYP2C2 in the ER depends on a receptor-mediated interaction, inhibition of degradation of exogenously expressed CYP2C2 with consequently increased CYP2C2 levels could lead to saturation of the ER receptor resulting in transport out of the ER. If this is the case, direct inhibition of CYP2C2 degradation should result in the altered localization of the CYP2C2 similar to that observed with downregulation of BAP31. The present results show that CYP2C2 is degraded via the proteasomal pathway and that in the presence of proteasomal inhibitors, this protein can undergo transport out of the ER reaching the plasma membrane. However, this relocalization does not appear to be the result of accumulation of CYP2C2 since similar altered distribution is observed for other ER proteins without an accompanying increase in the levels of the proteins in the ER.

Materials and methods

Materials

Tissue culture materials were purchased from Invitrogen, and the proteasomal inhibitors MG132, ALLN and lactacystin were obtained from Calbiochem. Antibodies against green fluorescent protein (GFP), BAP31, GP73 (Golgi Protein 73 kD), asialoglycoprotein receptor 1, calnexin, and sec61α were from Santa Cruz Biotechnology, Inc. and Cy5 conjugated anti-rabbit and anti-mouse antibodies were from Invitrogen. The chemiluminescence western blotting detection kit was obtained from Pierce Chemical Co.

Plasmid constructions

The construction of chimeras CYP2C2/GFP, C1-29/GFP, OEC/GFP (transmembrane domain of epidermal growth factor receptor fused to CYP2C2 with its 21-amino acid N-terminal signal anchor deleted), EEO (truncated epidermal growth factor receptor with its cytosolic domain deleted), C1-21/GFP, CYP2E1/GFP, CYP2C2/YN (CYP2C2 tagged at the C-terminus with 154 N-terminal amino acids of yellow fluorescent protein YFP) and CPR/YFP (NADPH-P450 reductase tagged at the C-terminus with YFP) have been described [32–35]. The chimera CYP3A4/GFP was constructed by amplifying CYP3A4 cDNA present in a pUC vector (kindly supplied by F. Gonzalez) using a 5’ primer with a Bgl II site and a 3’ primer with a Kpn I site, digesting the fragment with Bgl II-Kpn I, and inserting it into the vector EGFP-N1 digested with the same enzymes.

Cell culture and transfection

COS1 and HepG2 cell culture and transfection with Lipofectamine 2000 reagent were conducted as described [36]. To obtain HepG2 cells stably expressing either CYP2C2/GFP or CYP2E1/GFP, cells were transfected with the respective plasmid DNAs, and fluorescence-positive cells expressing the GFP chimera were selected by 4 rounds of fluorescence activated cell sorting at 2-week intervals.

Flow cytometry and fluorescent microscopy

Flow cytometry and confocal fluorescent microscopy studies were done as previously described [31]. Briefly, for flow cytometry, cells were collected by trypsinization, washed with Dulbecco’s minimum essential medium and resuspended in 1 ml of the same medium. Cells were analyzed with a Coulter Epics XL-MCL flow cytometer as described [31]. For microscopy, cells were grown on coverslips placed in 6-well plates and analyzed as described [37]. Treatment with MG132 (10 µM) was continued for 18 h, unless indicated otherwise. For immunofluorescent detection of proteins, primary antibodies were used at dilutions of 1:50 and Cy5 secondary antibodies at 1:100 as recommended by manufacturers. Alexa Fluor conjugated wheat germ agglutinin was used at a dilution of 1:250. Fixed cells were imaged with a Zeiss LSM510 confocal microscope, as described [31].

Western blotting

Proteins were separated by SDS-PAGE and electrophoretically blotted to nitrocellulose. Blots were probed with the antibodies at the dilutions recommended by manufacturers (1:1000 for primary antibodies and 1:2000–1:4000 for secondary antibodies). Western blot analysis was performed using a chemiluminescence detection kit from Pierce Chemical Co. Quantification of the relative proteins levels was done using Image J on scanned western images.

Subcellular fractionation

HepG2 cells stably expressing CYP2C2/GFP were lysed in a buffer containing 0.25 M sucrose in 5 mM Hepes-KOH, pH 6.8 by homogenization with 32 strokes (pestle B) in a Dounce homogenizer and cellular lysates were layered over a discontinuous sucrose gradient, as previously described [37, 38]. After centrifugation 1-ml fractions were collected and proteins in each fraction were detected by SDS-PAGE and western blotting.

Results

CYP2C2 is degraded by the proteasomal pathway and the signal for degradation is present in the cytosolic domain

To test whether CYP2C2 is degraded via the proteasomal pathway, we analyzed the effect of a selective proteasomal inhibitor, MG132, on the expression of CYP2C2/GFP in transfected COS1 cells. The presence of GFP at the C-terminus has been previously shown to have no effect on CYP2C2 localization or enzymatic activity [32].

Since different microsomal CYPs exhibit different half-lives, ranging from about 4 h for CYP2E1 [17] to 10 h or more for other forms [39], the levels of CYP2C2/GFP after treatment with MG132 were examined at several different time points. In transiently transfected COS1 cells treated with MG132, the level of CYP2C2/GFP detected by western analysis was little changed after 4 and 8 h of treatment, but was substantially increased after 18 h (Fig. 1A). A similar increase was detected after treatment for 18 h with two other proteasomal inhibitors, ALLN and lactacystin (Fig. 1A). On the other hand, there was no increase after treatment with the calpain inhibitor MDL (not shown). These results are consistent with degradation of CYP2C2 via the proteasomal pathway.

Fig. 1. Effect of proteasome inhibitors on the expression levels of CYP2C2/GFP.

A. COS1 cells were transfected with CYP2C2/GFP, and 24 h later cells were treated for the indicated time with 10 µM MG132. Treatments with 100 µM ALLN or 25 µM lactacystin (Lact.) were for 18 h. Numbers below the blot indicate the average fold-increase in the protein level from two independent experiments. A representative blot from two experiments is shown for the ALLN and lactacystin studies. B. COS1 cells were either mock-transfected or transfected with CYP2C2/GFP encoding full-length 2C2, OEC/GFP encoding the cytosolic domain of 2C2, or C1-29/GFP , containing the N-terminal transmembrane domain of 2C2. Numbers below the blot indicate the mean fold- increase (with standard error) in the proteins levels after MG132 treatment from three independent experiments. C. HepG2 cells stably expressing CYP2C2/GFP were treated for the indicated times with 10 µM MG132. Numbers below the blot indicate the average fold-increase in the protein levels from two independent experiments. A–C. Cellular lysates were analyzed by SDS-PAGE using 40 µg of total proteins from COS1 lysates and 100 µg of HepG2 lysates followed by western blotting using HRP-conjugated anti-GFP antibody. Picomolar and femtomolar sensitivity detection solutions were used for the COS1 and HepG2 samples, respectively. D. COS1 cells were transfected with CYP2C2/GFP, C-29/GFP, or OEC/GFP, and after 18 h of treatment with 10 µM MG132 or vehicle, cells were collected and the fluorescence intensities of the cells were determined by flow cytometry. The black line shows the fluorescence intensities of control cells and the red lines fluorescence intensities of cells treated with MG132.

Proteasomal degradation of integral ER membrane proteins can be mediated by one of 3 possible pathways, ERAD-C, ERAD-M and ERAD-L, depending on whether the signal for degradation is located in the cytosolic, the transmembrane, or the luminal domain, respectively [40]. Since microsomal CYPs have only a very short, if any, luminally exposed N-terminal domain they must be degraded through either the ERAD-C or ERAD-M pathways. To distinguish between these two possibilities, we analyzed the effect of MG132 on the levels of chimeric GFP proteins containing either the N-terminal signal/anchor sequence of CYP2C2 (C1-29/GFP) or the catalytic cytosolic domain (amino acids 21–490) fused to the transmembrane domain of epidermal growth factor receptor (EGFR), (OEC/GFP). The EGFR sequence acts as an ER targeting signal and we have shown previously that OEC/GFP is directly retained in the ER and has catalytic activity similar to the native CYP2C2 and CYP2C2/GFP [2, 32]. To compare the effect of MG132 on the levels of these chimeric proteins, fluorescence intensities were measured by flow cytometry and protein levels by western analysis using antibodies against GFP. Treatment of transfected COS1 cells with MG132 for 18 h induced similar substantial increases in the average fluorescence intensity in cells (Fig. 1D) and the amount of protein detected by western blotting (Fig. 1B) for both full-length CYP2C2/GFP and OEC/GFP. In contrast, only modest changes in the level of C1-29/GFP were observed (Fig. 1B, D). These data are consistent with the presence of a signal for proteasomal degradation in the cytosolic domain of CYP2C2 and degradation of CYP2C2 via the ERAD-C pathway.

Exogenous expression of high levels of CYP2C2/GFP in transiently transfected COS1 cells can activate the ER stress response and lead to apoptosis [41] which might underlie the proteasomal degradation of CYP2C2/GFP in these cells. In contrast, apoptosis is not induced by expression of CYPs in HepG2 cells. We, therefore, also analyzed the effect of MG132 on the expression of CYP2C2/GFP in a HepG2 cell line that stably expresses lower levels of CYP2C2/GFP. Similarly to COS1 cells after treatment with MG132, a relatively slow increase of CYP2C2/GFP levels was observed in stable HepG2 cells, with about a 2-fold increase after 8 h treatment and a 5-fold increase after 18 h (Fig. 1C).

These results established that inhibition of proteasomal degradation in either transiently transfected COS1 cells or stable HepG2 cells induces a substantial increase in levels of CYP2C2/GFP.

CYP2C2 redistributes from the ER to a juxtanuclear compartment and the plasma membrane

To determine if inhibiton of proteasomal degradation altered the cellular distribution of CYP2C2/GFP, we analyzed the subcellular localization of CYP2C2/GFP in control cells and in cells treated with MG132 by confocal microscopy. In both stably transfected HepG2 cells (Fig. 2 B,D,F) and transiently transfected COS1 cells (Fig. 2A,C,E), treatment with MG132 induced a strong increase of the level of CYP2C2/GFP fluorescence in a perinuclear location and in 20% to 30% of the cells, strong fluorescence at the nuclear and plasma membranes was also observed (Fig. 2 C–F). A similar pattern of altered localization was observed in cells treated with two other proteasome inhibitors, ALLN and lactacystin (not shown). Interestingly, redistribution to a perinuclear compartment and, in about 20% of the cells to the cell surface, was also detected with the chimera C1-29/GFP in COS1 cells (Fig. 2G,I), for which the level of expression was only modestly changed by MG132 treatment (Fig. 1B).

Fig. 2. Subcellular localization of CYP2C2/GFP in cells treated with MG132.

COS1 cells transiently transfected with CYP2C2/GFP (A,C,E) and with C1-29/GFP (G,I) or stably transfected HepG2 cells expressing CYP2C2/GFP (B,D,F,H,J) were treated with 10 µM MG132 or vehicle (Control) for 18 h as indicated. Panels H and J, show cells that were additionally treated with nocodazole (ND). Following fixation, cells were analyzed by confocal microscopy. Scale bars= 5 µm.

Microtubules are required for normal ER-Golgi transport [42], so to test whether redistribution of CYP2C2/GFP in cells treated with MG132 involves conventional transport out of the ER, HepG2 cells stably expressing CYP2C2/GFP were treated with the microtubular inhibitor, nocodazole. Consistent with previous results in COS1 cells [3, 32], nocodazole treatment did not alter the typical ER distributon of CYP2C2/GFP in HepG2 cells (Fig. 2H). However, in cells treated with MG132 and nocodazole, rather than the perinuclear and plasma membrane pattern observed with MG132 alone, fluorescence was detected in a punctuate pattern (Fig. 2J) which is a typical distribution of proteins in the secretory pathway after inhibition of microtubules. These results are consistent with transport of CYP2C2/GFP out of the ER in a normal microtubule-dependent process.

Effect of inhibiton of proteasomal degradation on the localization of other ER membrane proteins

It is possible that the redistribution of CYP2C2/GFP after treatment with MG132 resulted from a gross reorganization of the ER membranes rather than transport of CYP2C2/GFP to the plasma membrane and perinuclear region. We, therefore, looked at the distribution of several other integral ER membrane proteins in HepG2 cells stably expressing CYP2C2/GFP (Fig. 3). Immunofluorescent staining showed that endogenous calnexin and sec61α were distributed throughout the ER as expected in control cells, and their distribution was not significantly affected by treatment with MG132, even in cells in which CYP2C2/GFP was present at the plasma membrane (Fig. 3). Although treatment with MG132 resulted in some accumulation of calnexin and sec61α in a perinuclear region, these proteins were mostly retained in a reticular ER, unlike CYP2C2/GFP, indicating that the structure of the ER was retained in these cells. Furthermore, the Golgi protein, GM130, distributed in a pattern typical for Golgi in both control and MG132-treated cells, (see below, Fig. 4C) indicating that the Golgi compartment was not disrupted by MG132. Thus, the redistribution of CYP2C2/GFP after MG132 treatment does not result from gross reorganization of the secretory membrane compartment.

Fig. 3. Cellular localization of ER resident proteins in MG132-treated HepG2 cells stably expressing CYP2C2.

Stably transfected HepG2 cells expressing CYP2C2/GFP were treated with 10 µM MG132 or vehicle (Control) for 18 h. Following fixation and permeabilization, cells were immunostained with anti-calnexin, anti-sec61 or anti-BAP31 antibodies as indicated, visualized with Cy5-conjugated secondary antibodies, and analyzed by confocal microscopy. Cy5 fluorescence is shown in the left panels, GFP fluorescence in the middle panels, and an overlay of the first two panels in the right panels (Merge). Bars, 5 µm.

Fig. 4. Colocalization of CYP2C2/GFP with plasma membrane markers.

Row A. HepG2 cells stably expressing CYP2C2 were treated with MG132 for 18 h and following fixation, the cellular plasma membrane was stained with Alexa Fluor conjugated wheat germ agglutinin. Row B. COS1 cells were co-transfected with CYP2C2/GFP and truncated EGFR receptor EEO, and 24 h after transfection, cells were treated with MG132 for 18 h and immunostained with an antibody against the extracellular domain of EGFR, followed by Cy5-conjugated secondary antibody. Row C. HepG2 cells stably expressing CYP2C2/GFP and treated with MG132 were immunostained with anti-GM130 antibody, followed by Cy5 conjugated secondary antibody. D. HepG2 cells stably expressing CYP2C2/GFP were treated with 5 µg/ml anisomycin for 18 h and after fixing the cells, nuclei were stained with propidium iodide (PI). Rows A–D. In the first panel staining with Alexa Fluor, propidium iodide, or immunostaining is shown, in the second panel, GFP fluorescence is shown, and in the third panel an overlay of the first two panels is shown (Merge). Bars, 5 µm.

Interestingly, treatment with MG132 resulted in redistribution of BAP31, an ER membrane chaperone which interacts with CYP2C2 [31], similar to that observed for CYP2C2/GFP, with 20–30% of the cells exhibiting cell surface localization (Fig. 3). There was a striking colocalization of BAP31 and CYP2C2/GFP in the plasma membrane and perinuclear regions in these cells although in contrast to CYP2C2/GFP, BAP31 distribution in an ER pattern was also easily detected. Whether the redistribution of BAP31 results from its interaction with CYP2C2 or is coincidental will require further study.

CYP2C2 is present at the plasma membrane

To confirm that the CYP2C2/GFP was located near the plasma membrane in cells treated with MG132, colocalization of the CYP2C2/GFP with cell surface markers was examined. The distribution of CYP2C2/GFP fluorescence overlapped with that of fluorescent-stained wheat germ agglutinin in stably transfected HepG2 cells (Fig. 4A) and with epidermal growth factor receptor truncated by deletion of its cytosolic domain (Fig. 4B) which is transported to the cell surface [2, 32]. These experiments indicate that a fraction of the CYP2C2/GFP is redistributed to the plasma membrane after MG132 treatment.

The perinuclear region in which the fluorescent proteins accumulate superficially resembles localization in the Golgi, but there was little overlap between CYP2C2/GFP and GM130, a Golgi marker protein, in the perinuclear region of MG132-treated stably transfected HepG2 cells expressing CYP2C2/GFP (Fig. 4C). This result suggests that the perinuclear structure in which CYP2C2/GFP accumulates may represent an ER quality control compartment (ERQC).

In cells undergoing apoptosis, ER membrane can replace shedding plasma membrane and ER markers are detected at the cell surface [43, 44]. To test whether apoptosis was induced by MG132 treatment and the resulting ER movement to the plasma membrane could underly the redistribution of CYP2C2/GFP, HepG2 cells were treated with a known apoptosis inducer, anisomycin [45]. In cells treated with anisomycin, the hallmarks of apoptosis, strong chromatin condensation and fragmentation were observed, but CYP2C2/GFP remained exclusively in the ER and localization at the plasma membrane was not detected (Fig. 4D). Similar results were observed with two other inducers of apoptosis, thapsigargin and tunicamycin (not shown) so it is clear that redistribution of CYP2C2/GFP in MG132 treated cells is not a consequence of apoptosis.

To test whether CYP2C2/GFP was integrated into the plasma membrane after treatment with MG132, membrane organelles from HepG2 cells stably expressing CYP2C2/GFP were fractionated by sucrose gradient centrifugation and CYP2C2/GFP and organelle markers were detected by western blotting. In untreated cells, CYP2C2/GFP was predominantly in the ER fraction, co-sedimenting with the ER marker calnexin (Fig. 5A). However, in cells treated with MG132, CYP2C2/GFP was detected not only in ER fractions, but also in fractions at the top of the gradient cosedimenting with the plasma membrane marker asialoglycoprotein receptor 1. Additionally, CYP2C2/GFP was present in regions cosedimenting with the Golgi apparatus, which could represent localization in the Golgi or the ERQC. Cosedimentation of CYP2C2/GFP with the plasma membrane confirms that a fraction of CYP2C2/GFP not only colocalized with the cell surface, but was integrated into the plasma membrane.

Fig. 5. Subcellular fractionation of MG132-treated HepG2 cells and membrane orientation of plasma membrane localized CYP2C2/GFP.

A. HepG2 cells stably expressing CYP2C2/GFP and treated with MG 132 (+) or vehicle (−), were lysed and subcellular organelles were separated by centrifugation in a discontinuous sucrose gradient. Numbers below the blots indicate the number of the collected fractions with the bottom of the gradient being fraction 1. CYP2C2/GFP; calnexin, an ER marker; GP76, a Golgi marker; and asialoglycoprotein receptor 1 (ASGPR1), a plasma membrane marker, were detected in each fraction by western blotting. The blot is a representative experiment from two independent experiments. B. COS1 cells, transiently transfected with CYP2C2/YN, were treated with MG132 for 18 h and following fixation, either non-permeabilized cells or cells permeabilized with 0.1% Triton X-100 were immunostained with anti-GFP antibody and Cy5-conjugated secondary antibody. Bars, 5 µm.

Orientation of Membrane Associated CYP2C2/GFP

To test the transmembrane orientation of cell surface localized CYP2C2/GFP, COS1 cells were transfected with CYP2C2 tagged at the C-terminus with a non-fluorescent half of the YFP (CYP2C2/YN). After treatment with MG132, the chimeric YN proteins were detected with anti-GFP antibody, which recognizes YN, in intact or permeabilized cells. CYP2C2/YN was detected in permeabilized cells and in a fraction of the cells was distributed in a perinuclear and plasma membrane pattern while fluorescence was only weakly detected in nonpermeabilized cells (Fig. 5B). These observations are consistent with the transport of CYP2C2 out of the ER without a change in its membrane topology.

Reversibility of CYP2C2/GFP redistribution after MG132 treatment

If redistribution of CYP2C2/GFP results from nonspecific toxic effects of MG132 that irreversibly damage the cells, then the abnormal distribution of CYP2C2/GFP should not be reversed by terminating MG132 treatment. However, fluorescence of CYP2C2/GFP returned to an ER reticular pattern if stably transfected HepG2 cells were allowed to recover from MG132 treatment for 8 h (Fig. 6). The redistribution of CYP2C2/GFP is therefore, not the result of irreversible toxic effects of MG132 treatment.

Fig. 6. Redistribution of CYP2C2/GFP to the ER following washout of MG132.

HepG2 cells stably expressing CYP2C2/GFP were treated with MG132 (B) or vehicle (Control) (A) for 18 h or were treated for 18 h and then placed in a fresh medium for 8 h (C). Following fixation, GFP fluorescence was detected by confocal microscopy. Bars, 10 µm.

Effect of proteasomal inhibition on other CYPs and cytochrome P450 reductase

To determine if the effects of MG132 were specific to CYP2C2 or more general for other CYPs, we tested the effects of proteasomal inhibition on the levels and distribution of CYP3A4 and CYP2E1 which are known to undergo proteasomal degradation [39]. In COS1 cells treated with MG132, the levels of both CYP2E1/GFP and CYP3A4/GFP increased as analyzed by flow cytometry (Fig. 7B). In the HepG2 cells stably expressing CYP2E1/GFP or transiently transfected with CYP3A4/GFP, treatment with MG132 resulted in redistribution of the CYP proteins to the perinuclear region and in 20% to 30% of cells to the cell surface (Fig. 7A), similar to that observed for CYP2C2/GFP (Fig. 2). The redistribution of CYP3A4/GFP was not as dramatic as that for CYP2C2/GFP and CYP2E1/GFP with perinuclear distribution observed in most cells, but with plasma membrane fluorescence in a smaller fraction of the cells.

Fig. 7. Effect of MG132 treatment on the expression level and localization of CYPs 2E1, 3A4 and CPR.

A. HepG2 cells stably expressing CYP2E1/GFP, or transiently transfected with CYP3A4/GFP and CPR/YFP, as indicated were treated with MG132 or vehicle (Control) and the subcellular localization of fluorescent proteins was analyzed by confocal microscopy. B. Flow cytometry analysis of the expression levels of COS1 cells transfected with CYP2E1/GFP, CYP3A4/GFP or CPR/YFP. The black line shows the fluorescence intensities of control cells and the red one fluorescence intensities of MG132-treated cells. Western analysis of CPR/YFP expression levels in transiently transfected COS1 cells treated with vehicle (−) or MG132 (+) for 18 h is shown in the insert in the CPR panel. Forty µg of total protein from cellular lysates was loaded on each lane, and the blot was probed with anti-GFP antibody. Numbers below the blot indicate average relative levels of the CPR/YFP protein from two independent experiments.

We also examined the redox partner of CYPs, cytochrome P450 reductase (CPR). Treatment with MG132 had only modest effects on the levels of CPR/YFP as analyzed by flow cytometry or western analysis (Fig. 7B), consistent with a report that CPR is not degraded by the proteasome [39]. Nevertheless, treatment with MG132 resulted in redistribution of CPR/YFP (Fig. 7A), although similarly to CYP3A4/GFP, the redistribution was less dramatic than that observed for CYP2C2/GFP (Fig. 2) so that plasma membrane localization was observed in a smaller fraction of cells. These results show that redistribution after MG132 treatment can occur for a protein whose level is not substantially increased, and interestingly not only for a CYP, but for a CYP associated protein.

Discussion

In these studies, a dramatic redistribution of CYP2C2 from the ER to a perinuclear compartment and the cell surface was observed after treatment with MG132 in both transiently transfected COS1 cells and HepG2 cells stably expressing a low level of CYP2C2/GFP. Interestingly, CYP2E1, CYP3A4, and, to a smaller degree CPR, exhibited similar redistribution after inhibition of proteasomal degradation. CPR forms a complex in the ER with CYPs which is required for CYP activity so that its redistribution could be either as a complex with CYPs or by an independent parallel mechanism. Prolonged treatment of cells with proteasome inhibitors can induce an ER overload, leading to a change in the morphology of the ER and other compartments [46, 47]. However, the redistribution of these drug metabolizing enzymes is not the result of a collapse and reorganization of the ER in the cells treated with proteasomal inhibitors, since the integral ER proteins, sec61 and calnexin, largely remained in a normal ER pattern. Reversal of the redistribution when the proteasomal inhibitors are removed, indicates that an irreversible toxic response to the inhibitor does not mediate the redistribution. Further, the redistribution of CYP2C2 was blocked by the microtubular inhibitor, nocodazole, which suggests that the transport from the ER uses the conventional pathway of secretory and plasma membrane proteins, which is dependent on microtubules. The data support the idea that the enzymes in the microsomal drug metabolism complex, CYPs and CPR, share a common mechanism of ER retention that is disrupted by inhibition of proteasomal degradation.

The initial impetus for these experiments was to test whether the increased levels of CYP2C2 in the ER after inhibition of its degradation could saturate an ER retention “receptor” and result in transport and redistribution of the protein. However, the levels of CPR/YFP and C1-29/GFP were little changed after MG132 treatment, while these proteins were redistributed similarly to CYP2C2/GFP. Further, if saturation resulted in the redistribution of CYP2C2/GFP, then ER localization in addition to the perinuclear and cell surface localization should have been observed, but there was little or no CYP2C2/GFP detected in the ER after MG132 treatment. The results, thus, indicate that increased accumulation of proteins in the ER and saturation of the ER retention mechanism does not explain the redistribution.

The perinuclear compartment where most of the CYP accumulates probably represents the ERQC. Such accumulation has been shown to be microtubule-dependent [22] which would be consistent with the observed dispersal of this structure after treatment with nocodazole. Calnexin is known to redistribute partially to the ERQC when proteasomal degradation is inhibited [22, 48]. In cells treated with MG132, a fraction of calnexin, and to a lesser extent sec61α, accumulated in the perinuclear compartment, partially overlapping the distribution of CYP2C2/GFP. In the subfractionation studies, CYP2C2/GFP and calnexin were also detected at densities near those for the Golgi marker, GP76, after MG132 treatment which may represent the ERQC. These data are consistent with accumulation of CYP2C2/GFP in the ERQC after proteasomal inhibition. In the quality control process, misfolded or unnecessary proteins are selected either for proteasomal destruction or for rescue and further transport from the ER [24]. The ERQC compartment appears to be connected to a conventional secretory pathway so that not all proteins in the ERQC are degraded, but some are targeted for further transport [22, 49]. This is consistent with our results showing that a fraction of the CYP2C2/GFP can reach the plasma membrane.

A small fraction of CYP2E1 or CYP2D6 has been reported to localize at the cell surface. In the case of CYP2E1, evidence was presented that the the topology of protein transported to the cell surface was reversed with the catalytic domain on the extracellular side of the membrane [8]. For CYP2D6, both the native and reversed topology have been proposed [10, 50, 51]. Interestingly, in nonpermeabilized cells, the YFP tag at the C-terminus of CYP2C2/YN was not detected by antibodies indicating that the CYP2C2/YN retained its normal topology after MG132 treatment. Further, CYP2C2/GFP was not glycosylated, based on its size, which would have been expected, if its topology was reversed with the catalytic domain on the luminal side of the ER membrane [37]. While under some conditions the topologies of CYPs are inverted at the plasma membrane [50, 51], presumably due to inversion of topology during insertion into the ER membrane leading to transport to the cell surface, transport to the cell surface after MG132 treatment does not require topology inversion for CYP2C2.

It was recently shown that a fraction of misfolded proteins can be transported from the ER and that this transport is increased by inhibition of proteasomal degradation [28] suggesting a competition between transport and degradation pathways. The transport was dependent on ER exit signals present in the proteins although an exit signal is not sufficient for transport since the yeast plasma membrane protein Yor1P was retained in the ER after inhibition of proteasomal degradation even though it contains an ER exit motif [52]. CYP2C2 as an integral ER membrane protein does not contain a functional exit motif, but does contain the sequence DXE that is the best defined ER exit signal [53]. This sequence is usually near the C-terminal transmembrane domain of a protein while the CYP2C2 DXE motif is located in the middle of the protein rather than near the N-terminal transmembrane peptide and, therefore, is unlikely to be functional. In addition, a similar redistribution is observed for C1-29/GFP in which the DXE motif has been deleted. In contrast to the studies of Kincaid and Cooper [28], our studies with CYP2C2 suggest that an ER exit sequence is not strictly required for transport from the ER after inhibition of proteasomal degradation.

Similar redistribution from the ER to a perinuclear compartment and cell surface was detected for another integral ER membrane protein, BAP31. Unlike CYP2C2, BAP31, as a chaperone, apparently is capable of shuttling between the ER and other compartments of the secretory pathway, although the mechanisms or signals responsible for its ER retention/retrieval are not clear [54–56]. Recent studies show that BAP31 shuttles between peripheral ER and the ERQC and is also involved in retrotranslocation of proteasomal substrates, consistent with its role in ERAD [57, 58]. Interestingly, downregulation of BAP31 also results in the redistribution of CYP2C2 to a compartment resembling the ERQC [31]. BAP31 may, therefore, redistribute as a complex with CYP2C2 when proteasomal degradation is inhibited. In a similar way, another ER membrane chaperone, calnexin, can be transported to the plasma membrane by binding to its substrates which are destined for the cell surface [59].

The transport of CYPs out of the ER after proteasomal inhibition is probably an indirect effect of the inhibition and not the result of decreased degradation of the CYPs. Inhibition of proteasomal degradation could inhibit the proteolytic activation of a protein that is required for ER retention of CYPs or inhibit the degradation of a protein that then abnormally accumulates and facilitates the ER exit of the CYPs. On the other hand, changes in membrane composition may underlie the redistribution. Proteasomal degradation is known to activate many transcription factors which regulate ER membrane composition. In yeast, transcription of a gene regulating membrane lipid composition is increased by an ER membrane bound transcription factor, which is activated by proteasomal degradation [60, 61]. In mammalian cells, proteasomal degradation regulates sterol biosynthesis, also affecting membrane composition [62]. Thus, inhibition of proteasomal degradation can affect the composition of the ER membrane, which could then affect the mobility and distribution of ER proteins. Proliferation of the ER membranes, occurring during induction of P450 expression by drugs, also results in changes in membrane composition [14, 63]. So, it is possible that inhibition of proteasomal degradation affects ER retention of CYPs by indirectly modifying the ER membrane composition. Additional studies will be required to identify the critical target of proteasomal degradation that mediates the redistribution of CYPs after inhibition of the degradation.