The Proteasome Participates in Degradation of Mutant (cid:1) 1 -Antitrypsin Z in the Endoplasmic Reticulum of Hepatoma-derived Hepatocytes*

of mutant (cid:1) 1 -antitrypsin ( (cid:1) 1 -AT) Z in the endoplasmic reticulum (ER) 1 -AT-deficient 1 phenyl- methylsulfonyl fluoride. The radiolabeled cell lysates were subjected clarification and immunoprecipitation, immunoprecipitates analyzed SDS-PAGE/fluorography exactly as previously to trichloroacetic acid precipitation and scintillation counting of the trichloroacetic precipitates to that there was equivalent in- corporation comparison.

Because retention of mutant ␣ 1 -antitrypsin (␣ 1 -AT) Z in the endoplasmic reticulum (ER) is associated with liver disease in ␣ 1 -AT-deficient individuals, the mechanism by which this aggregated glycoprotein is degraded has received considerable attention. In previous studies using stable transfected human fibroblast cell lines and a cell-free microsomal translocation system, we found evidence for involvement of the proteasome in degradation of Chem. 275, 25015-25022) found that degradation of ␣ 1 -ATZ in a stable transfected murine hepatoma cell line was inhibited by tyrosine phosphatase inhibitors, but not by the proteasomal inhibitor lactacystin and concluded that the proteasome was only involved in ER degradation of ␣ 1 -ATZ in nonhepatocytic cell types or in cell types with levels of ␣ 1 -AT expression that are substantial lower than that which occurs in hepatocytes. To examine this important issue in further detail, in this study we established rat and murine hepatoma cell lines with constitutive and inducible expression of ␣ 1 -ATZ. In each of these cell lines degradation of ␣ 1 -ATZ was inhibited by lactacystin, MG132, epoxomicin, and clasto-lactacystin ␤-lactone. Using the inducible expression system to regulate the relative level of ␣ 1 -ATZ expression, we found that lactacystin had a similar inhibitory effect on degradation of ␣ 1 -ATZ at high and low levels of ␣ 1 -AT expression. Although there is substantial evidence that other mechanisms contribute to ER degradation of ␣ 1 -ATZ, the data reported here indicate that the proteasome plays an important role in many cell types including hepatocytes.
The classical and most common form of ␣ 1 -antitrypsin (␣ 1 -AT) 1 deficiency is a relatively unique genetic disease in that it is associated with injury to one tissue, pulmonary emphysema, by a loss-of-function mechanism and injury to another tissue, chronic hepatitis/hepatocellular carcinoma, by a gain-of-function mechanism. Many studies have provided evidence that emphysema results from lack of the anti-elastase activity of ␣ 1 -AT in the lung (reviewed in Refs. 1 and 2). Liver disease is due to toxic effects of aggregated ␣ 1 -ATZ retained in the ER of liver parenchymal cells. The gain-of-function mechanism is most clearly demonstrated by experiments in mice transgenic for human ␣ 1 -ATZ. These mice develop liver injury and hepatocellular carcinoma despite the fact that they have their own endogenous anti-elastases (3)(4)(5).
The mutant Z allele of ␣ 1 -AT is characterized by a single nucleotide substitution, which results in the replacement of glutamate 342 by a bulky lysine residue (1,2). The studies of Carrell and Lomas (6,7) have shown that this substitution renders the ␣ 1 -AT molecule more susceptible to polymerization and that highly ordered aggregates accumulate in the ER of liver cells.
One interesting observation, arising from unbiased nationwide screening studies of ␣ 1 -AT deficiency in Sweden, indicates that only 10 -15% of deficient individuals develop clinically significant liver disease (8,9). In previous studies we tested the hypothesis that this subgroup of deficient individuals is susceptible to liver injury by virtue of additional unlinked genetic traits or environmental factors that delay degradation of the mutant ␣ 1 -ATZ molecule after it is retained in the ER (10). With the use of fibroblast cell lines from deficient patients with liver disease (susceptible hosts) compared with those from deficient individuals without liver disease (protected hosts), we found that more efficient degradation of retained mutant ␣ 1 -ATZ in the ER correlated with protection from liver disease. These results, therefore, focused our attention on the mechanism by which ␣ 1 -ATZ is degraded in the ER. Subsequent studies showed that lactacystin inhibited ER degradation of ␣ 1 -ATZ in stable transfected human fibroblast cell lines and in a cell-free microsomal system (11,12), indicating that the proteasome and the ubiquitin system were involved in this important quality control mechanism. Degradation of ␣ 1 -ATZ in stable transfected Chinese hamster ovary cells and in primary cultures of human mononuclear phagocytes was also inhibited by lactacystin (13). However, Cabral et al. (14) recently found that lactacystin did not inhibit degradation of ␣ 1 -ATZ in a stable transfected murine hepatoma cell line. Degradation of ␣ 1 -ATZ in this cell line was markedly decreased by tyrosine phosphatase inhibitors. These authors concluded that the proteasome was only involved in degradation of ␣ 1 -ATZ in nonhepatocytic cells or alternatively was only involved in ␣ 1 -ATZ degradation in cell types with lower levels of ␣ 1 -AT biosynthe-sis. Because hepatocytes are the predominant site of synthesis of ␣ 1 -AT and the cells predominantly affected by the pathobiological process of ␣ 1 -AT deficiency-associated liver disease, this is a very important issue. In this study we examined hepatocytes in further detail by generating rat and murine hepatoma cell lines that express ␣ 1 -ATZ. We also examined the role of the relative level of ␣ 1 -AT biosynthesis by generating hepatoma cell lines with regulated expression of ␣ 1 -ATZ.
Cell Lines-Fibroblast cell lines that were transduced with amphotropic recombinant retroviral particles bearing ␣ 1 -ATZ cDNA and have stable constitutive expression of ␣ 1 -ATZ (CJZ12B) as described previously (10). The same approach was also used to establish stable constitutive expression of ␣ 1 -ATZ in the mouse hepatoma cell line Hepa1-6 (Hepa1-6N2Z9B) and the rat hepatoma cell line H11 (H11N2Z1) (15). Our previous studies have shown that the goat anti-human 1-AT antibody does not recognize endogenous murine 1-AT in Hepa1-6 cells (15). The H11 cell line has been shown to be highly differentiated for hepatocytic function, but, because it is extinguished for expression of HNF-1␣ and HNF-4, it does not express endogenous rat ␣ 1 -AT (16). We also established HeLa and Hepa1-6 cell lines with inducible expression of ␣ 1 -ATZ. Full-length ␣ 1 -ATZ cDNA was subcloned into the pTRE plasmid provided by CLONTECH, and HeLa Tet-Off cells were transfected with the resulting pTRE ␣ 1 -ATZ plasmid (HTO/Z). Hepa1-6 cells were transfected with the Tet-On plasmid provided by CLONTECH (Hepa1-6TON4Z2). Candidate colonies were screened by luminometry for inducible expression of a luciferase reporter plasmid after transient transfection. The colony with the optimal profile of low background and highest inducibility was selected for transfection with the pTRE ␣ 1 -ATZ plasmid. Pulse labeling experiments showed dose-dependent, time-dependent, and reversible induction of ␣ 1 -ATZ synthesis in the HTO/Z cell line and in the Hepa1-6TON4Z1 cell line. 2 Metabolic Labeling, Immunoprecipitation, and Analytical Gel Electrophoresis-Cell lines were subjected to pulse-chase radiolabeling as described previously (11,12). For the pulse period, the cells were incubated at 37°C for 2 h in 250 Ci/ml Tran 35 S-label in Dulbecco's modified Eagle's medium lacking methionine. The cells were then rinsed vigorously and incubated in Dulbecco's modified Eagle's medium with excess unlabeled methionine for time intervals up to 6 h as the chase period. At the end of each chase period, the extracellular medium was harvested and the cells were lysed in phosphate-buffered saline, 1% Triton X-100, 0.5% deoxycholic acid, 10 mM EDTA, and 2 mM phenylmethylsulfonyl fluoride. The radiolabeled cell lysates were subjected to clarification and immunoprecipitation, and immunoprecipitates were analyzed by SDS-PAGE/fluorography exactly as described previously (12). Aliquots of the radiolabeled cell lysates were also subjected to trichloroacetic acid precipitation and scintillation counting of the trichloroacetic acid precipitates to ensure that there was equivalent incorporation between cell lines under comparison. Results were quanti- . Cells were then subjected to a pulsechase protocol in the absence or presence of the same inhibitors (pulse: 2 h; chase: 3 h). Cells were then lysed and equivalent aliquots of each cell lysate subjected to immunoprecipitation with anti-human ␣ 1 -AT followed by SDS-PAGE/fluorography. The relative electrophoretic mobility of molecular mass markers is indicated at the right margin. There were no significant differences in incorporation of total trichloroacetic acid-precipitable radioactivity between conditions (data not shown).

FIG. 2.
Effect of lactacystin on the fate of ␣ 1 -ATZ in a human fibroblast cell line. The CJZ12B cell line was preincubated for 1 h at 37°C in serum-free medium alone (control) or supplemented with lactacystin (30 M). Cells were then subjected to a pulse-chase protocol in the absence or presence of lactacystin (pulse: 2 h). Cell lysates (IC) and extracellular fluid (EC) were analyzed as described in the legend to Fig. 1. The mobility of molecular mass markers is shown at the right margin. The inset to the right shows quantitative data from densitometric scanning of fluorograms. These data are reported as percentage of ␣ 1 -ATZ remaining by comparison with time 0, which is arbitrarily set at 100%. The solid line with diamonds is for control, and the dashed line with squares is for lactacystin. fied by densitometric scanning of fluorograms using the ImageQuant software (Molecular Dynamics, Sunnyvale, CA). After correcting for background, all values were expressed as percentage of ␣ 1 -ATZ by semi-log regression analysis.
Preparation of Soluble and Insoluble Fractions from Cell Lysates-Cell lysates were passed through a 25-gauge needle 10 times on ice (17). Insoluble material was recovered by centrifugation at 16,000 ϫ g for 15 min. Pellets were solubilized in 50 l of 50 mM Tris-HCl, pH 6.8, 5% SDS, 10% glycerol with 1 min of sonication and then 10 min of boiling. This protocol has been shown to provide specific separation of ␣ 1 -AT mutants into soluble and insoluble fractions (18).

Effect of Proteasome Inhibitors on Degradation of ␣ 1 -ATZ in
Human Fibroblast Cell Lines-In previous studies we showed that lactacystin inhibits degradation of ␣ 1 -ATZ in human fibroblast cell lines genetically engineered for expression of ␣ 1 -ATZ (11). Here we examined the effect of several additional proteasome inhibitors to determine whether the effect of lactacystin was generalizable and whether other proteasome inhibitors were more or less effective. The CJZ12B human fibroblast cell line was incubated for 1 h at 37°C with each inhibitor and then subjected to pulse-chase radiolabeling in the presence of each inhibitor. The highest dose of each inhibitor that did not affect cell viability or incorporation into trichloroacetic acid-precipitable radioactivity was selected for further study. Fig. 1 shows the results for ␣ 1 -ATZ present in intracellular lysates after a pulse of 2 h and a chase of 3 h. There was a marked increase in the 52-kDa ␣ 1 -ATZ polypeptide in cells treated with lactacystin, MG132, epoxomicin, and clasto-lactacystin ␤-lactone. At the optimal dose for each inhibitor, there was no significant difference in the effectiveness of lactacystin, MG132, or epoxomicin. Clasto-lactacystin ␤-lactone was slightly less effective in this cell line.
A pulse-chase experiment in the CJZ12B cell line with multiple time points is shown in Fig. 2. The results show that, in the absence of inhibitor (control, top panel), ␣ 1 -ATZ was synthesized as a 52-kDa polypeptide, which disappeared from intracellular over 1-2 h and was almost completely degraded by 4 -6 h of the chase period with negligible amounts secreted into the extracellular fluid. A similar amount of ␣ 1 -ATZ was present in the intracellular contents at time 0 after lactacystin treatment, but only began to disappear from intracellular contents by 4 -6 h. There was no evidence for secretion of ␣ 1 -ATZ into the extracellular fluid after lactacystin treatment. The halftime for disappearance of ␣ 1 -ATZ was 82 min (r 2 ϭ 0.82) in the absence and 219 min (r 2 ϭ 0.80) in the presence of lactacystin. These data indicate that degradation of ␣ 1 -ATZ in genetically engineered human fibroblasts was decreased, but not completely abrogated, by several different proteasome inhibitors. There is no evidence that one of these inhibitors, lactacystin, mediated an increase in secretion of ␣ 1 -ATZ, even when lower doses (as low as 10 M) and longer preincubation periods (as long as 2 h) were used (data not shown).
Effect of Proteasome Inhibitors on Degradation of ␣ 1 -ATZ in Hepatoma Cell Lines-We used amphotropic recombinant retroviral particles to establish hepatoma cell lines with stable expression of ␣ 1 -ATZ. First, we applied this strategy for the murine hepatoma cell line Hepa1-6. This cell line was used by Cabral et al. (14). Antibody to human ␣ 1 -AT does not recognize FIG. 3. Effect of proteasome inhibitors on degradation of ␣ 1 -ATZ in a murine hepatoma cell line. The Hepa1-6N2Z9B cell line was subjected to an experimental protocol identical to that described in the legend to Fig. 1. There were no significant differences in incorporation of trichloroacetic acid-precipitable radioactivity between conditions (data not shown). Lactacyst, lactacystin; Epox, epoxomicin; ␤ lactone, clasto-lactacystin ␤-lactone.
FIG. 4. Effect of MG132 on the fate of ␣ 1 -ATZ in a murine hepatoma cell line. The Hepa1-6N2Z9B cell line engineered for constitutive expression of ␣ 1 -ATZ was preincubated for 1 h at 37°C in serum-free medium alone (control) or supplemented with MG132 (10 M). Cells were then subjected to a pulse-chase protocol in the absence of presence of MG132 (pulse: 2 h). Cell lysates (IC) and extracellular fluid (EC) were analyzed as described in the legend to Fig. 1. The lower band (ϳ50 kDa) is known to be a nonspecific product of immunoprecipitation, as determined by blocking experiments with unlabeled purified ␣ 1 -ATZ and by parallel immunoprecipitation with nonimmune IgG (data not shown). The inset to the right shows quantitative data from densitometric scanning of fluorograms. These data are reported as percentage of ␣ 1 -ATZ remaining by comparison to time 0 which is arbitrarily set at 100%. The solid line with diamonds is for control, and the dashed line with squares is for lactacystin. endogenous murine ␣ 1 -AT expressed in this cell line (data not shown). Once human ␣ 1 -ATZ expression was established in Hepa1-6 (Hepa1-6N2Z9B), we examined the effect of proteasomal inhibitors. The results of a pulse-chase experiment with one time point and multiple inhibitors are shown in Fig. 3. There was a marked increase in the 52-kDa ␣ 1 -ATZ polypeptide after incubation with lactacystin, MG132, epoxomicin, and clasto-lactacystin ␤-lactone. The results of a pulse-chase experiment in Hepa1-6N2Z9B cells with multiple time points and MG132 are shown in Fig. 4. In cells that had been preincubated in the absence of MG132 (control), there was a 52-kDa polypeptide at time 0 intracellular. It disappeared progressively over 1-3 h of the chase period and was only faintly seen at 4 and 6 h of the chase period. Negligible amounts of ␣ 1 -ATZ were secreted into the extracellular fluid. In the presence of MG132, a similar amount of ␣ 1 -ATZ was present at time 0 intracellular with very little disappearance until 4 h of the chase period. There was no increase in secretion of ␣ 1 -ATZ in the presence of MG132. The half-time for disappearance of ␣ 1 -ATZ was 70 min (r 2 ϭ 0.91) in the absence and 224 min (r 2 ϭ 0.92) in the presence of MG132. These data indicate that degradation of ␣ 1 -ATZ in Hepa1-6 cells is reduced by proteasome inhibitors.
Next we established expression of ␣ 1 -ATZ in the rat hepatoma cell line H11. This cell line is highly differentiated for hepatocytic function, but completely lacks endogenous ␣ 1 -AT expression because expression of transcription factors HNF-1␣ and HNF-4 is extinguished (15). At the end of a pulse-chase experiment with a 2-h pulse and a 3-h chase in H11N2Z1 cells, there was a marked increase in ␣ 1 -ATZ remaining intracellular after treatment with lactacystin, MG132, epoxomicin, and clasto-lactacystin ␤-lactone (Fig. 5). When MG132 was examined in a pulse-chase experiment with multiple time points in Fig. 6, it was again shown to mediate a decrease in degradation of ␣ 1 -ATZ. In control, the 52-kDa ␣ 1 -ATZ began to disappear at 2 h of the chase period. In cells treated with MG132, ␣ 1 -ATZ began to disappear at 3 h and the rate of disappearance was less at each subsequent time point. MG132 did not mediate an increase in secretion of ␣ 1 -ATZ into the extracellular fluid. The half-time for disappearance of ␣ 1 -ATZ was 100 min (r 2 ϭ 0.95) in the absence and 153 min (r 2 ϭ 0.92) in the presence of MG132.
Effect of Proteasome Inhibitors on Degradation of ␣ 1 -ATZ in Cells with High Levels of ␣ 1 -ATZ Biosynthesis as Compared with Those with Low Levels of ␣ 1 -ATZ Synthesis-To address this issue we established cell lines with inducible (regulated) expression of ␣ 1 -ATZ. First, we used the tetracycline-regulated expression system (Tet-Off) to establish regulated expression of ␣ 1 -ATZ in HeLa cells. Using this system, the tetracycline analogue doxycycline suppresses expression of the target gene. To determine whether expression of ␣ 1 -ATZ in this system is dependent on the concentration of doxycycline and whether we could establish conditions of low and high ␣ 1 -ATZ expression for proteasome inhibitor experiments, we examined the effect of different concentrations of doxycycline in our first cell line HTO/Z (Fig. 7a). The HTO/Z cell line was incubated at 37°C for 48 h with doxycycline in several different concentrations and then subjected to pulse-labeling for 30 min. The results of immunoprecipitation/SDS-PAGE/fluorography on cell lysates show that there was abundant 52-kDa ␣ 1 -ATZ polypeptide in the absence of doxycycline, progressively decreasing as the concentration of doxycycline increased. The 52-kDa ␣ 1 -ATZ polypeptide was undetectable in cells treated with doxycycline at 1.0 ng/ml. Expression of ␣ 1 -ATZ in this system is also timedependent and reversible. 3 Pulse-chase experiments show that there is intracellular retention of ␣ 1 -ATZ in this cell line in a manner that recapitulates the defect in ␣ 1 -AT deficiency, 3   viding further evidence for the validity of this model system. Next, we examined several additional concentrations of doxycycline to establish a concentration that reduced ␣ 1 -ATZ levels to ϳ100-fold lower, but still detectable. In Fig. 7b we found that concentrations of doxycycline from 0.001 to 0.1 ng/ml mediated a progressive decline in ␣ 1 -ATZ, reaching ϳ100-fold lower only at 0.1 ng/ml. This concentration of doxycycline was then selected for low levels of ␣ 1 -ATZ expression in the next experiment.
The HTO/Z cell line at high (no doxycycline) or low (0.1 ng/ml doxycycline for 48 h) levels of ␣ 1 -ATZ was subjected to pulsechase radiolabeling after treatment in the absence or presence of proteasome inhibitor epoxomicin (Fig. 8). At high levels of ␣ 1 -ATZ (Fig. 8a), in the absence of epoxomicin (control), ␣ 1 -ATZ was synthesized as a 52-kDa polypeptide at time 0 intracellular. It was retained for 1 h and then progressively disappeared over 2-3 h of the chase period to an undetectable level at 4 h.
A small amount of the mature 56-kDa ␣ 1 -ATZ polypeptide was secreted into the extracellular fluid. In the presence of epoxomicin (Fig. 8a, lower panel), a similar amount of ␣ 1 -ATZ was present at time 0 intracellular, but in this case there was very little disappearance during the chase period. The half-time for disappearance of ␣ 1 -ATZ was 108 min (r 2 ϭ 0.95) in the absence and 193 min (r 2 ϭ 0.93) in the presence of epoxomicin. The results also indicate that epoxomicin did not mediate an increase in secretion of ␣ 1 -ATZ under these conditions. At 100fold lower levels of ␣ 1 -ATZ (Fig. 8b), in the absence of epoxomicin (control), ␣ 1 -ATZ was again synthesized as a 52-kDa polypeptide, which progressively disappeared over 2-3 h of the chase period to undetectable levels at 4 h and with a small amount of 56-kDa mature ␣ 1 -ATZ secreted. At this lower level of ␣ 1 -ATZ expression, the presence of epoxomicin (Fig. 8b, lower  panel) was associated with a significant decrease in disappearance of ␣ 1 -ATZ from the intracellular contents, but no change in the amount of ␣ 1 -ATZ found in the extracellular fluid. The half-time for disappearance of ␣ 1 -ATZ was 127 min (r 2 ϭ 0.93) in the absence and 279 min (r 2 ϭ 0.99) in the presence of epoxomicin. These results indicate that epoxomicin inhibited degradation of ␣ 1 -ATZ to a similar extent in cells with high levels as compared with the same cells with 100-fold lower levels of ␣ 1 -ATZ expression. Inhibition of degradation by epoxomicin was not associated with an increase in secretion of ␣ 1 -ATZ. The results also indicate that there was no major difference in the fate of ␣ 1 -ATZ, degradation or secretion, in the HTO/Z cell line engineered for high levels as compared with 100-fold lower levels of ␣ 1 -ATZ expression.
We also examined the effect of proteasome inhibitors on degradation of ␣ 1 -ATZ in a hepatoma cell line with inducible (regulated) expression of ␣ 1 -ATZ. Here we used the Hepa1-6 cell line and engineered it for Tet-On expression of ␣ 1 -ATZ. The Hepa1-6 TON4Z1 cell line only expresses ␣ 1 -ATZ in the presence of doxycycline. 3 We examined the effect of proteasome inhibitor lactacystin on ␣ 1 -ATZ in this cell line under conditions of maximal expression (1 g/ml doxycycline for 48 h) in Fig. 9. The results show that, in the absence of lactacystin (control), ␣ 1 -ATZ was synthesized as a 52-kDa polypeptide at time 0 intracellular, retained for 1 h, and then it underwent progressive disappearance between 2 and 4 h of the chase period. In the presence of lactacystin, a similar amount of the 52 kDa ␣ 1 -ATZ polypeptide was present at time 0 intracellular, but there was very little disappearance of the polypeptide over the duration of the 4-h chase period, indicating that lactacystin mediates an inhibition of ␣ 1 -ATZ degradation in hepatoma cells with relatively high levels of ␣ 1 -ATZ expression. The half-time for disappearance of ␣ 1 -ATZ was 98 min (r 2 ϭ 0.95) in the absence and 215 min (r 2 ϭ 0.94) in the presence of lactacystin. Lactacystin also mediated an inhibition of ␣ 1 -ATZ degradation in Hepa1-6 TON4Z1 cells with ϳ30-fold lower levels of ␣ 1 -ATZ expression, as directed by a lower concentration of doxycycline (data not shown).
Effect of Proteasome Inhibitors on the Solubility of ␣ 1 -ATZ-Next we examined the detergent-soluble and -insoluble fractions from HTO/Z cells treated with MG132 to determine whether ␣ 1 -ATZ formed more insoluble aggregates when the proteasome was inhibited. HTO/Z cells were subjected to a pulse-chase experiment in the absence or presence of MG132 in several different concentrations, and then the cell lysates were analyzed after separation into soluble and insoluble fractions (Fig. 10). The results show that MG132 mediated a concentrationdependent increase in ␣ 1 -ATZ in soluble fraction and a dramatic increase in insoluble ␣ 1 -ATZ. The accumulation of ␣ 1 -ATZ in the insoluble fraction mediated by MG132 was already apparent at time 0 of the chase period (after a 1-h preincuba- tion and 1-h pulse in the presence of MG132). Interestingly, two distinct ϳ48and ϳ38-kDa cleavage products of ␣ 1 -ATZ also appeared in the insoluble pellet in the presence of MG132. Densitometric analysis of the ϳ52-, ϳ48-, and ϳ38-kDa ␣ 1 -ATZ polypeptides in the insoluble fraction taken together showed that MG132 mediated an 7.4-fold increase in insoluble ␣ 1 -ATZ. These results indicate that there was a marked increase in insoluble aggregates of ␣ 1 -ATZ when the proteasome was inhibited and that insoluble aggregates of ␣ 1 -ATZ underwent two distinct proteolytic cleavage reactions when the proteasome was inhibited. DISCUSSION A detailed elucidation of the mechanisms by which mutant aggregated ␣ 1 -ATZ is degraded in the ER is essential for understanding how the quality control apparatus of the ER works in general and for understanding the specific issue of how a subgroup of ␣ 1 -AT-deficient individuals become susceptible to liver injury and carcinogenesis. Previous studies have shown that there is a lag in the disposal of this mutant protein in genetically engineered human fibroblasts from "susceptible" deficient patients (10). Moreover, this lag in the ER disposal/ quality control mechanism appears to be specific, i.e. it affected the disposal of two polymerogenic mutants of ␣ 1 -AT, but not a model unassembled membrane protein (19).
Initial studies of the ER degradation of ␣ 1 -ATZ indicated that the ubiquitin-dependent proteasomal system was involved. Degradation of ␣ 1 -ATZ in genetically engineered human fibroblast cell lines and in a cell-free microsomal translocation system was inhibited by lactacystin (11). Degradation of ␣ 1 -ATZ in the cell-free system was shown to be dependent on ATP and, more importantly, a polyubiquitinated calnexin-␣ 1 -ATZ complex was shown to be a degradative intermediate (11). Subsequent studies by Novoradovskaya et al. (13) have shown that degradation of ␣ 1 -ATZ is also inhibited by lactacystin in transfected Chinese hamster ovary cells and in primary cultures of human mononuclear phagocytes. Using an experimen- In each case separate wells were incubated at 37°C for 1 h in the absence or presence of 10 M epoxomicin (epox). The cells were then subjected to pulse-chase radiolabeling in the absence or presence of epoxomicin (pulse: 30 min) and the individual samples analyzed exactly as described in the legend to Fig. 1. The volume of samples from the low-expressing cells used for immunoprecipitation was 8 times greater than the volume for corresponding samples from high-expressing cells. The film for panel B (low expression) was subjected to fluorography for 7 days as compared with 1 day for panel A (high expression). The insets at the right show quantitative data from densitometric scanning of these fluorograms. These data are reported as percentage of ␣ 1 -ATZ remaining by comparison to time 0, which is arbitrarily set at 100%. The solid line with diamonds is for control, and the dashed line with squares is for lactacystin. IC, cell lysates; EC, extracellular fluid. tal approach in which the degradative machinery in the reticulocyte lysate of the cell-free system is fractionated and reconstituted with purified components, we have recently found evidence for at least three different pathways in the degradation of ␣ 1 -ATZ, including ubiquitin-dependent and -independent proteasomal mechanisms and at least one nonproteasomal mechanism (12). Subsequent studies have suggested that autophagy may constitute one of the nonproteasomal mechanisms (15) and have substantiated the concept that there are multiple pathways involved in ER degradation of ␣ 1 -ATZ.
In the most recent study of this issue, Cabral et al. (14) found that lactacystin did not inhibit degradation of ␣ 1 -ATZ in a stable transfected murine hepatoma Hepa1-6 cell line. Degradation of ␣ 1 -ATZ in this cell line was decreased by tyrosine phosphatase inhibitors. Lactacystin did inhibit degradation of ␣ 1 -AT Null Hong Kong, a truncated mutant, in a separate stable transfected Hepa1-6 cell line. Taken together, these results indicated that the proteasome did not play a role in degradation of ␣ 1 -ATZ in hepatoma cell line even though the proteasome was active in these cells, that an entirely separate mechanism for degradation existed, and that this distinct mechanism was specific for ␣ 1 -ATZ. Taking into consideration the previous results indicating involvement of the proteasome in degradation of ␣ 1 -ATZ in genetically engineered fibroblasts (11,13), in primary cultures of human macrophages (13), and in the cell-free microsomal system (11), these authors concluded that the proteasomal mechanism was cell type-specific, either for nonhepatocytic cell types and/or cell types with lower levels of endogenous ␣ 1 -AT expression than hepatocytes (14).
Because the hepatocyte is an extremely important, if not the most important, site of synthesis of ␣ 1 -ATZ with respect to the development of liver disease, in this study we sought to examine in further detail the involvement of the proteasome in ER degradation of ␣ 1 -ATZ in cells of hepatocytic lineage. The results show that lactacystin, MG132, epoxomicin, and clastolactacystin ␤-lactone all inhibit degradation of ␣ 1 -ATZ in several different types of genetically engineered hepatoma cell lines, including the murine hepatoma Hepa1-6 used by Cabral et al. and a rat hepatoma cell line H11, which has the advantages of being highly differentiated for hepatocytic function, but lacking endogenous expression of ␣ 1 -AT. Degradation of ␣ 1 -ATZ in hepatoma cell lines with constitutive and inducible expression of ␣ 1 -ATZ was blocked to an equivalent extent by proteasomal inhibitors. Finally, studies in HeLa and Hepa1-6 cell lines with inducible expression of ␣ 1 -ATZ showed that the proteasome was involved in degradation of ␣ 1 -ATZ at both high and 30 -100-fold lower levels of expression. These results suggest to us that the lack of involvement of the proteasome in degradation of ␣ 1 -ATZ in the Hepa1-6 cell line generated by Cabral et al. (14) is cell line-specific, perhaps reflecting a type of adaptation. This hypothesis by no means diminishes the importance of the observations of Cabral et al. or the importance of tyrosine phosphatases in the quality control mechanism. There is now ample evidence for multiple mechanisms/ pathways in the ER quality control apparatus and for cellular "adaptation." In fact, gene expression profile analysis has shown marked changes in expression of many genes in yeast cells that accumulate misfolded proteins (20 -23). Moreover, if a cellular adaptation mechanism is truly applicable, then the results of Cabral et al. raise the interesting possibility that the adaptation is specific for ␣ 1 -ATZ, a polymerogenic mutant, and not for ␣ 1 -AT Null Hong Kong, a mutant that is truncated and not likely to be polymerogenic.
Three other results of this study deserve comment. First, there is a marked increase in the formation of insoluble aggregates of ␣ 1 -ATZ and two distinct degradation products appear exclusively in the insoluble fraction when the proteasome is inhibited. These degradation products could theoretically be generated in, and/or accumulate in, the ER or the cytoplasm. Several lines of evidence make it more likely that they are generated in and localize to the ER. ␣ 1 -ATZ has not been detected outside the ER lumen in intact cells (15) or in the supernatant of cell-free mammalian translocation reactions (11) and proteasome inhibitors do not induce aggresomes in cells that express mutant ␣ 1 -ATZ (15). Similar degradation products are generated processively in the lumen of microsomal vesicles that have translocated wild type ␣ 1 -ATZ in a cell-free microsomal translocation reaction (11), suggesting the existence of an endoluminal proteolytic system that recognizes wild type or mutant ␣ 1 -ATZ when it is retained in the ER for a prolonged period of time. Second, there was no evidence for a significant increase in secretion of ␣ 1 -ATZ in fibroblasts, hepatoma cells, or HeLa cells in the presence of lactacystin, MG132, epoxomicin, or clasto-lactacystin ␤-lactone at doses optimal for inhibition of degradation of ␣ 1 -ATZ. Using the lower doses of lactacystin or longer periods of preincubation with lactacystin described by Novoradovskaya  The HTO/Z cell line was incubated at 37°C for 1 h in the absence or presence of MG132 in several different concentrations as indicated at the bottom. The cells were then subjected to pulse-chase radiolabeling in the absence or presence of MG132 in the same concentrations (pulse; 60 min) and the cell lysates separated into soluble and insoluble fractions exactly as described under "Experimental Procedures." An equivalent aliquot from the starting material was used for immunoprecipitation from soluble and insoluble fractions. Immunoprecipitates were subjected to analytical gel electrophoresis exactly as described above. The migration of the novel ϳ48and 38-kDa insoluble proteolytic fragments is indicated by the arrows at the right margin. observed a lesser degree of inhibition of degradation, but no increase in secretion of ␣ 1 -ATZ in fibroblasts, hepatoma cells, and HeLa cells at low and high levels of expression (data not shown). Third, secretion of ␣ 1 -ATZ did not increase when the level of its synthesis was decreased. This was shown by modulating the level of synthesis 30 -100-fold in tetracycline-regulated cell lines. This result is noteworthy because the rate of polymerization of ␣ 1 -ATZ decreases at lower concentrations in purified systems (6). The results in this report therefore suggest that other factors working in concert with polymerization play a role in the fate of ␣ 1 -ATZ when it accumulates in the ER.