Identification of a cholesterol-regulated 180-kDA microsomal protein in rat hepatocytes

Eur. J. Biochem. 188, 123-129 (1990) 0 FEBS 1990

Identification of a cholesterol-regulated 180-kDa microsomal protein in rat hepatocytes Diego HARO, Pedro F. MARRERO, Jose AYTE and Fausto G. HEGARDT Unit of Biochemistry, School of Pharmacy, University of Barcelona, Spain

(Received July 3, 1989) - EJB 89 0818

The immunoprecipitation by antibodies to 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase of extracts of [3 5S]methionine-pulse-labelled isolated hepatocytes, followed by electrophoresis and fluorography, showed the presence not only of 97-kDa HMG-CoA reductase, but also of another protein of 180 kDa. Boiling the immunoprecipitates both in the presence and in the absence of 2-mercaptoethanol, followed by SDS/ polyacrylamide gel electrophoresis both in the presence and in the absence of 8 M urea, was not found to change the ratio of 180-kDa/97-kDa proteins. These facts suggest that the 180-kDa protein is not an aggregated form of HMG-CoA reductase. A different batch of antibodies obtained from a newly purified HMG-CoA reductase fully titrated the reductase activity, but did not immunoprecipitate the 180-kDa protein, showing that there is no cross-reactivity between these proteins. The 1 SO-kDa polypeptide is a glycoprotein of N-linked high-mannose oligosaccharide chains, which is not processed on the Golgi system. The apparent molecular mass of the carbohydrate is 16 kDa.

The incubation of rat hepatocytes with sterols produces, on the one hand, a decrease in the rate of synthesis, and on the other hand, an acceleration in the turnover rate of the 180-kDa protein. In addition, mevalonate is known to decrease its rate of synthesis. The carbohydrate-free 164-kDa protein was found to degrade only a tenth as fast as the glycoprotein and, furthermore, the degradation was no longer accelerated by sterols. These results support the notion that the 180-kDa protein is not a modified form of 97-kDa reductase, but probably a different protein related to cholesterol metabolism, and also that the N-linked, high-mannose chains, which are bound to the glycoprotein, are required for rapid and controlled degradation of the protein.

The biosynthesis of cholesterol starting with acetyl-CoA in several cell lines and liver from animals is regulated pri- marily by an alteration in the activity of 3-hydroxy-3- methylglutaryl-CoA (HMG-CoA) reductase and HMG-CoA synthase. The sequential activity of the two enzymes catalyzes the formation of mevalonate, which is a precursor not only of cholesterol but also of isoprenoid compounds such as ubiquinone, dolichol and isopentenyl-tRNA. The whole pathway is subject to feedback suppression by the sterol end pro- ducts of the pathway, which mainly consist of cholesterol and 25-hydroxy-cholesterol. These effects can be observed not only in animal livers that have been fed cholesterol but also in cultivated cell lines that have been incubated with exo- genous cholesterol and 25-hydroxy-cholesterol [l]. Similarly, cells incubated with mevalonate are found to repress both enzymes [2].

Although the feedback control of sterols is mostly exerted upon HMG-CoA reductase and HMG-CoA synthase, other enzymes of this pathway are also feedback-regulated by cholesterol, these being squalene synthetase [3, 41, prenyl- transferase [5], and other enzymes which take part in the reactions that convert squalene to cholesterol [6, 71.

Correspondence to F . G . Hegardt, Unidad de Bioquimica, Facultad de Farmacia, Plaza Pi0 XI1 sin, E-08028 Barcelona, Spain

Abbreviations. HMG-CoA, 3-hydroxy-3-methylglutaryl-coenzyme A; endo H, endoglycosidase-H; reductase, HMG-CoA reductase.

Enzymes. HMG-CoA reductase (EC 1.1 .I .34); endoglycosidase H (EC 3.2.1.96).

In order to explore the molecular mechanism of this regulation we, along with others, have performed experiments using isolated hepatocytes incubated in the presence of mevalonate and sterols [8 - 101. In this system, reductase synthesis and degradation are regulated by mevalonate and cholesterol. During the course of these studies, we noted that antibodies to HMG-CoA reductase immunoprecipitated, in addition to the reductase, another polypeptide of 180 kDa [ll]. In this paper this protein is characterized as a glycoprotein with N-linked oligosacharides. Its regulation by cholesterol is also reported, and the influence of the presence or absence of carbohydrates on is turnover rate is measured. Unlike other N-linked glycoproteins [12], the presence of carbohydrates makes the protein more unstable compared to the deglycosylated protein.

We also discuss the possibility of the 180-kDa protein being related to cholesterol metabolism, but different from the HMG-CoA reductase and HMG-CoA synthase.

EXPERIMENTAL PROCEDURES

Chemicals

Chemicals were obtained from the following sources : [35S]methionine (1000 Ci/mmol), [3H]mannose (40.6 Ci/ mmol) and 1251-labelled protein A (30 Ci/mg) from Amer- sham; formalin-fixed Staphylococcus aweus (Pansorbin) from Calbiochem; tunicamycin, leupeptin, antipain, phenylmethylsulfonyl fluoride, bovine serum albumin (A-7906)

124

dithiothreitol and endoglycosidase H (endo H) from Sigma; Incubation with sterols other samples of endo H, produced by expression of the cloned gene, and aprotinin were obtained from Boehringer; col- lagenase from Cooper Biochemicals Ltd; SDS, acrylamide, bisacrylamide; N,N,N',N'-tetramethylethylenediamine, and ammonium persulfate from Bio-Rad. Phosphatidylcholine and cholesterol were from Sigma; 25-hydroxy-cholesterol was from Steraloids Inc. All other chemicals were of the highest quality from commercial sources.

Animals

Rats, fed ad libitum were housed under 12-h light/dark cycles. Rat hepatocytes were prepared 6 h after the period of darkness as described in [8].

Immunoprecipitation

HMG-CoA reductase antibodies were raised in rabbit against the purified carboxy-terminal proteolyzed domain of 52 kDa that contained the catalytic site. This enzyme had been taken from rats fed on 5% cholestyramine. The antibodies and control y-globulines were purified by three successive ammonium sulfate precipitations at 40% saturation. The immunoprecipitation of HMG-CoA reductase was carried out as described in [8]. Control immunoprecipitation with non- immune rabbit y-globulines was carried out in the routine manner.

Electrophoresis

Sodium dodecyl sulfate/polyacrylamide (7.5%) gel electrophoresis (SDS/PAGE) was carried out as according to Laemmli [ 131.

Metabolic labeling

Rat hepatocytes (5 x lo6) were incubated in a final volume of 1.5 ml incubation medium composed of Hank's buffer that had been supplemented with amino acids and cofactors (as described by Eagle [14]), without methionine, but with 5 mM glucose, 1.5% bovine serum albumin and 20 mM Tricine pH 7.4 to which 100 pCi [j5S]methionine was added. In the experiments in which cells had been incubated with tunicamycin, 3 pl of a stock solution of 1.25 mg/ml in 25 mM NaOH was added to 1.5 ml fresh medium, at the time of the addition of the radionucleids.

Incubation with [3H]mannose (0.5 mCi) was carried out as in the case of the [35S]methionine, but [35S]methionine was substituted by unlabelled methionine (0.12 pM).

Enzymatic treatment of immunoprecipitates

For their treatment with endo H after the final wash, the immunocomplexes of 2 x106 rat hepatocytes were eluted from Pansorbin by incubation with 20 p1 buffer A (2% SDS, 30 mM dithiothreitol, 100 mM Tris pH 6.8) for 45 min at room temperature. The sample was first boiled for 2 min in a water batch at 100°C and then changed to a water bath at room temperature. Aliquots of endoglycosidase H were added to 100 pl 45 mM sodium citrate buffer pH 5.5, 0.2% SDS, 1 mM dithiothreitol, 100 pM leupeptin, 10 pM antipain, 15 pg/ml aprotinin, 0.2 mM phenylmethylsulfonyl fluoride, and incubated overnight at 24°C. Then 15 pl buffer B (15% SDS, 10% 2-mercaptoethanol in 62.5 mM Tris/HCl pH 6.8) was added to the samples prior to electrophoresis.

Phosphatidylcholine (either 100 mg or 100 mg plus 40 mg cholesterol and 4 mg 25-hydroxycholestero1) in chloroform was dried under nitrogen before 10 mlo.9% NaCl was added. Samples were sonicated for 15 min and centrifuged at 100000 x g for 45 min. The supernatants containing the dispersions were then removed and stored at 4°C under nitrogen. The concentration of phosphatidylcholine in the dispersions was measured using the method described by Beutler [15], and that of sterols using the method of Siedel [16].

Measurement of the rate of synthesis of 180-kDa protein

The rate of synthesis of the 180-kDa protein was determined by preincubating isolated hepatocytes (2 x lo6 cells) at 37 "C in 1.5 ml methionine-free minimum (Eagle's medium) in the presence of effectors. [35S]Methionine (50 FCi) was then added, as described in the appropriate legends. After 20 min, the cells were diluted with 3 ml ice-cold medium containing 2 mM unlabelled methionine, they were then pelleted and washed in the same medium and the immunoprecipitation was carried as described [ll].

Measurement of the turnover rate of the 180-kDa protein

In order to measure the turnover rate of the protein, cells (5 x lo6 cells/ml) were preincubated for 60 min in the presence of [35S]methionine. They were then washed in medium that had been supplemented with 2 mM unlabelled methionine, and following this, resuspended in this medium. The cells were then incubated at 37°C and 2 x lo6 cells were removed at each time and the 180-kDa protein isolated through immunoprecipitation. SDSjPAGE was carried out according to Laemmli [13] as described in [l 11.

RESULTS

Immunoprecipitation of the 180-kDa protein

Immunoprecipitation with anti-(HMG-CoA reductase) antibodies of extracts from rat hepatocytes, which had been incubated with [35S]methionine for 1 h, showed through SDS/ PAGE that, in addition to the HMG-CoA reductase (apparent molecular mass 104 kDa), another protein of 180 kDa was present. Several authors have reported [17,18] the occurrence of a protein of about 170 kDa that is found in immunoprecipitates of several cell lines treated with anti- (HMG-CoA reductase) antibodies. Chin et al. [17] among others were of the opinion that this 170-kDa was an aggregate of HMG-CoA reductase molecules which was produced when samples were either boiled in the presence of SDS, or electrophoresed in gels devoid of 8 M urea. In order to clarify this point, some aliquots taken from immunoprecipitates of rat hepatocytes were boiled and some were left unboiled. They were all then processed by SDS/PAGE in the presence of 8 M urea, whereupon no change in the aggregated state was found (Fig. 1).

Ness et al., using an immunoblotting technique, reported on the interconvertible character of the 104-kDa protein and the 180-kDs HMG-CoA reductase through sulfhydryl-disulfide forms [19]. In order to prove or disprove this assertion, hepatocytes isolated from rats which had been fed a normal chow diet were labelled with [35S]methionine and immunoprecipitated with anti-reductase antibodies. Subsequently,

125

Fig. 1. The effect of boiling on immunoprecipitates of HMG-CoA reductase. Rat hepatocytes were obtained through standard procedures. After a 60-min pulse of 100 pCi/ml (0.12 pM) [35S]methionine, the cells were solubilized and aliquots corresponding to 5 x 1 O6 cells were immunoprecipitated with different amounts of antibodies (lanes 2 and 6, 2.5 pg; lanes 3 and 7, 5 pg; lanes 4 and 8, 10 pg anti-reductase antibodies). The immunoprecipitates obtained were dispersed in buffer containing 10% 2-mercaptoethanol. Some samples were boiled (lanes 1-4) for 3 min and others left unboiled (lanes 5 - 8 ) ; all were processed by SDSjPAGE in 8 M urea and fluorography. Lanes 1 and 5 were control serum antibodies (10 pg). Numbers to the left of the panel correspond to the molecular mass of the following markers: 205 kDa, myosin; 116 kDa, p-galactosidase; 97.4 kDa, phosphoryl- ase h; 45 kDa egg albumin

immunoprecipitates were boiled with 2-mercaptoethanol, treated with iodoacetic acid to prevent the formation of disulfide bonds, and then processed through SDSjPAGE and fluorography. There was no change in the ratio of 180-kDa protein/lOCkDa reductase, irrespective of whether the thiol groups were present or not in the processing media (data not shown). At this point, we came to the conclusion that the 180- kDa protein was not a disulfide-bridge dimer of 104-kDa HMG-CoA reductase.

Immunoprecipitation of HMG-CoA reductase with a diflerent batch of antibodies

During the course of our experiments, we again purified rat liver HMG-CoA reductase, and a different preparation of antibodies was elicited in rabbit. When we carried out experiments as described above, we observed that these antibodies did not immunoprecipitate the 180-kDa protein (Fig. 2, lane 2) from rat hepatocytes, at variance with results with the old antibodies in which the occurrence of 180-kDa protein was clearly demonstrated (Fig. 2, lane 3). Titration experiments showed that this set of antibodies fully immunoprecipitated the HMG-CoA reductase activity as did the antibodies commonly used (Fig. 2 B).

This experiment showed that there is no cross-reactivity between HMG-CoA reductase and the 180-kDa protein; moreover, that the 180-kDa protein does not have HMG-CoA reductase activity as antibodies fully titrate the HMG-CoA reductase activity while the 180-kDa protein is not immunoprecipitated. The production of this batch of antibodies clearly shows that HMG-CoA reductase and the 180- kDa protein are different proteins.

02 04 06 08

‘6-Globulins (rng)

Fig. 2. Immunoprecipitation of the 104-kDa HMG-CoA reductase and the 180-kDa protein by different batches of anti-(HMG-CoA reductase) antibodies. (A) Rat hepatocytes were obtained through standard procedures. After a 60-min pulse of 100 pCi/ml (0.12 pM) [35S]methionine, the cells were solubilized and aliquots corresponding to 5 x lo6 cells were immunoprecipitated with different batches of anti-(HMG-CoA reductase) antibodies and processed by SDSjPAGE and fluorography. Lane 1, control serum antibodies; lane 2 new batch of anti-(HMG- CoA reductase) antibodies; lane 3, the batch of antibodies used throughout all other experiments. Arrows show the position of the 180-kDa protein (1 80) and the HMG-CoA reductase (Rd). Numbers to the left of the panel correspond to the positions of the molecular mass markers as in Fig. 1 (plus bovine serum albumin 66 kDa), which were run in parallel to the unknown samples. (B) HMG-CoA reductase purified to apparent homogeneity was titrated for HMG-CoA reductase activity using the new batch (0) and the commonly used batch of anti-(HMG- CoA reductase) antibodies (0) . Non-immune serum antibodies taken as control are also represented (V)

126

20 40 60 90 20 40 60 90 Time of Chase h n )

Fig. 3. The effect ofsterols on the turnover rate of the 55S-lubelled total ceN protein and the 180-kDa protein in isolated rat hepatocytes. Rat hepatocytes were obtained through standard procedures. The hepatocytes were incubated for 45 rnin with dispersions of phosphatidylcholine (720 pg/ml), then pelleted and incubated again for 45 min in phosphatidylcholine. Following this, the hepatocytes were labelled by incubation with 100 pCi/ml (0.12 pM) [35S]methionine. After a 60-min pulse, the cells were switched to 5 ml medium containing 2 mM unlabelled methionine and either dispersions of phosphatidylcholine alone (720 pg/ml) (0) or plus the sterol mixture (860 pM cholesterol and 86 pM 25-hydroxycholesterol) (0). At the indicated chase time, the cells were solubilized and aliquots precipitated in 10% tricholoroacetic acid (A) and immunoprecipitated with non-immune and anti-reductase y-globulins (B). The immunoprecipitates were processed through electrophoresis in SDSjPAGE (7.5%) and then fluorography, and the radioactivity of the 180-kDa protein, after being scraped from the gel, was measured in a liquid scintillation counter. The percentage of radioactivity at zero time (1.7 pCi/106 cells in A, or 23000 cpm/106 cells in (B) is represented on the ordinate

Influence of cholesterol on the rate of synthesis and the turnover rate of the 180-kDa protein

Through immunoprecipitation of a biologically radio- labelled protein, we investigated whether cholesterol affected the rate of synthesis of the 180-kDa protein. In order to do so, we incubated rat hepatocytes with either phosphatidylcholine on its own, or phosphatidylcholine plus a mixture of sterols, for 90 min. Following this, a 20-min pulse of [35S]methionine was given, and the immunoprecipitation carried out as described in Experimental Procedures. A measurement of the radioactivity bound to the total protein was also taken. The result of this experiment shows that, when incubated with sterols dispersed in phosphatidylcholine, the rate of synthesis of the 180-kDa protein decreases by 22%.

In order to determine the influence of cholesterol in the degradation of this protein, isolated hepatocytes were incubated for 90 rnin with phosphatidylcholine, then pulse-labelled for 60 min with [35S]methionine. After switching the cells to a medium containing unlabelled methionine (chase) and phosphatidylcholine either alone or plus the sterol mixture, the cells were pelleted and processed through solubilization, immunoprecipitation, SDSjPAGE and fluorography; the radioactivity of the 180-kDa band was measured using a liquid scintillation counter. Fig. 3A shows the turnover rate of the total trichloroacetic-acid-precipitable protein, which was found to be the same, irrespective of the presence or absence of cholesterol. The turnover rate of the 35S-labelled 180-kDa protein however, was found to be different (Fig. 3B). In the cells incubated with phosphatidylcholine alone, the 180-kDa protein declined with a half-life of about 67 min, which was similar to the rate of turnover of the 180-kDa protein incu-

Fig. 4. The effect of endo H and tunicumycin on the 180-kDa protein. Isolated rat hepatocytes were pulsed for 60 min with [35S]methionine in the presence (lanes 4-6), or absence (lanes 1-3 ) of tunicamycin (2.5 pg/ml). The soluble extracts of the cells were immunoprecipitated with anti-reductase antibodies. Some immunoprecipitates were incubated with 30 mU endoglycosidase H (lanes 3 and 6) for 24 h and resolved by SDSjPAGE and fluorography. Lanes 1 and 4 represent control antiserum. Arrows show the position of glycosylated (1 80) and deglycosylated 180-kDa protein (164), and the HMG-CoA reductase (Rd). Figures on the left correspond to the positions of the molecular mass markers described in Figs. 1 and 2

bated without phosphatidylcholine (70 min). However, in hepatocytes that had been incubated with phosphatidylcholine plus cholesterol the 180-kDa protein declined with half-life of 46 min. These results, together with the fact that mevalonate inhibits the rate of synthesis of the 180-kDa protein [8], demonstrated that it is probably a protein related to cholesterol metabolism.

The 180-kDa protein is a glycoprotein of the endoplasmic reticulum

The 180-kDa protein immunoprecipitated with antibodies HMG-CoA reductase is a glycoprotein with N-linked high- mannose carbohydrates. Incubation of the immunoprecipitates with endo H is found to produce a decrease in the molecular mass apparently corresponding to 16 kDa (Fig. 4, lane 3). When cells are incubated in the presence of tunicamycin (an antibiotic which prevents the formation of N-acetylglucos- aminyl-asparaginyl-linked units), an analogous decrease of the molecular mass is produced (Fig. 4, lane 5) . Incubation with endo H of immunoprecipitates which had been isolated after tunicamycin treatment of rat hepatocytes, produced a complete transformation of the 180-kDa band to the 164-kDa band (Fig. 4, lane 6). Fig. 4 also shows a slight decrease in the molecular mass of the HMG-CoA reductase as a result of the endo H treatment (compare lanes 2 and 3). This effect confirms the attachment of the N-linked high-mannose chains to HMG-CoA reductase as described by Liscum et al. [29].

In order to clarify further the character of the carbohydrate chains bound to the 180-kDa protein, hepatocytes were in-

127

PSI METH~ON" l j l I l j l i l ; [ 'H]MA"OSE

TUNICAMYCIN - - - - ANTI~ODIES

Fig. 5. The incorporation of radioactive mannose on the 180-kDa protein. Isolated rat hepatocytes were pulsed for 120 min either with [35S]methionine (lanes 1-3) or with [3H]mannose (lanes 4-6) in the presence (lanes 3,6) or absence (lanes 2 , 5 ) of 2.5 bg/ml tunicamycin. The soluble extracts of the cells were immunoprecipitated with anti- reductase antibodies and the immunoprecipitates resolved by SDS/ PAGE and fluorography. Lanes 1 and 4 correspond to the control serum. Arrows show the position of the glycosylated (180) and the deglycosylated 180-kDa protein (164). The position of molecule mass markers are as in Figs. 1 and 2

cubated with [3H]mannose. The fluorography of the immunoprecipitates obtained after the incubation of rat hepatocytes with [3H]mannose showed the 180-kDa protein. However, when the hepatocytes were incubated in the presence of tunicamycin (Fig. 5, lanes 5 and 6) this band was found to disappear. Incubation of the same hepatocytes, however, with [3sS]methionine as a control was found to produce the 180- kDa and 164-kDa bands respectively as shown above (Fig. 5, lanes 2 and 3).

These results suggest that the 180-kDa protein is a glycoprotein with N-linked mannose carbohydrates. Taking an average molecular mass for every N-linked mannose chain as 2 kDa, it is suggested that the 180-kDa protein is composed of eight different N-linked mannose chains.

The presence of the high-mannose carbohydrate chains in the 180-kDa protein suggested that it could be a glycoprotein bound to the endoplasmic reticulum. In order to prove this assertion, a Western blot of microsomes was carried out using anti-(HMG-CoA reductase) antibodies. The results showed that the 180-kDa protein was present in the liver washed- microsomal fraction, but absent in the cytosolic supernatant 100000 x g fraction (data not shown).

In order to determine whether or not the 180-kDa protein can be processed in the Golgi system, we incubated hepatocytes with [35S]methionine for 20 and 60 min and then incubated some aliquots of the immunoprecipitates with endo H. The results are shown in Fig. 6. It can be seen that the radioactive 180-kDa protein was fully transformed into the 164-kDa protein, irrespective of the time for which it was radioactively p~iIsc'C1. As the 180-kDa protein did not acquire resistance to endo H, it can thus be concluded that this protein was not processed by the Golgi system during the period assayed.

Fig. 6. The effect of the pulse time on the resistance of the 180-kDa protein to endo H treatment. Isolated rat hepatocytes were pulsed either for 20 min (lanes 1, 2) or 60 min (lanes 3, 4) with [35S]methionine. The soluble extracts of the cells were immunoprecipitated with anti-reductase antibodies. The immunoprecipitates were incubated for 24 h with either 0.5 mU endoglycosidase H (lanes 2, 4) or buffer alone (lanes 1, 3) and resolved by SDSjPAGE and fluorography as described. Arrows show the position of the glycosylated (1 80) and the deglycosylated 180-kDa protein (1 64). The position of molecular mass markers are shown to the left

Injluence of carbohydrate moiety on the 180-kDa turnover rate

In order to ascertain whether the carbohydrate could influence the degradation of the 180-kDa protein, isolated hepatocytes were incubated with [35S]methionine for 1 h both in the presence and absence of tunicamycin. The hepatocytes were then switched to a medium containing unlabelled methionine and the chase was carried out both in the presence and in the absence of tunicamycin as in the pulse. After solubilization, immunoprecipitation, SDSjPAGE and fluorography, the 180-kDa bands were scraped from the gels and the radioactivity measured using a liquid scintillation counter. As seen in Fig. 7B, the presence of tunicamycin, which inhibits the binding of mannose chains to the 180-kDa reductase, produces a strong stabilization of the protein which declines with a half-life of about 695 min. The hepatocytes which had not been incubated with tunicamycin had a known half-life of 70 min. There was found to be no change, however, in the turnover rate of the total trichloroacetic-acid-precipitable protein (Fig. 7A). It becomes clear therefore that the presence of carbohydrate is responsible for the degradation of the 180-kDa protein, because in its absence the turnover rate is strongly diminished.

We considered the possibility that the presence of tunicamycin could inhibit the expression of a proteinase which would alter the rate of degradation of the 180-kDa protein. If this is so, the change in turnover rate of the 180-kDa protein would not be due to the absence of carbohydrate itself but rather to the level of activity of this putative proteinase. In order to exclude this possibility, we first incubated the hepatocytes (as in the experiment described in Fig. 7) but with a suboptimal concentration of tunicamycin (1.5 pg/ml) and continued the experiment as before. There is evidence of co- existence of both the 180-kDa and 164-kDa forms, in a similar fashion to that shown in Fig. 4. When the 180-kDa and 164- kDa bands corresponding to the different incubation periods were removed from the gels, and the radioactivity measured,

128

100, - 0. v

2 5 0 - ?

ki 0

- c

TJ [r

m ul m

A L. 8 ! l a ' tTunicamycin

-

Total cell protein 180 kDa protein I I

20 40 60 90 20 40 60 90

Time of Chase (rnin)

Fig. 7. The effect of tunicamycin on the turnover rate of 35S-labeIled 280-kDa protein in isolated rat hepatocytes. Rat hepatocytes were prepared for experimentation using the standard procedure. Hepato- cytes were incubated for 60 rnin with 0.12 pM [35S]methionine (100 pCi/ml), both in the presence (0 ) and absence (0) of tunicamycin (2.5 pg/ml). Thereafter, the cells were switched to 5 ml medium containing 2 mM unlabelled methionine and either tunicamycin ( 0 ) (2.5 pg/ml) or buffer on its own, (0) and incubated for a further 90 min. After solubilization, the aliquots were processed for precipitation using 10% trichloroacetic acid (A) and for immunoprecipitation using non-immune and anti-reductase y-globulines (B). The immunoprecipitates were electrophoresed in SDSjPAGE and fluorography was carried out. The radioactivity of the 180-kDa protein was measured with a liquid scintillation counter after being scraped from the gel. The percentage of radioactivity at zero time, 1.33 and 0.96 pCi/ lo6 cells in the presencc and absence of tunicamycin respectively (A) or 7000 and 3700 cpm in the presence and absence of tunicamycin respectively (B), is represented on the ordinate

m Y) 0

20 40 60 90 Time of Chase (min)

Fig. 8. The sterol-mediated regulation of the 180-kDaprotein in isolated rut heputocytes incubated both in the presence and in the absence of tunicamycin. Rat hepatocytes were set up as described. The hepatocytes wcrc incubated for 45 min in a suspension of phosphatidylcholine (720 pg/ml), thcn pelleted and incubated again for a further 45 min in phosphatidylcholine. Thereafter, incubation of 0.12 pM [35S]methionine (100 pCi/ml) were carried out both with 2.5 pg/ml tunicamycin (A,A) and without (0,O) in the presence of phosphatidylcholine. Following a 60-min pulse, the cells were switched to 5 ml medium, containing 2 mM unlabelled methionine and either suspensions of phosphatidylcholine (720 pg/ml) alone ( 0 , A ) or plus sterols (860 pM cholesterol and 86 pM 25- hydroxycholesterol) (0,A) and incubated for a further 90 min. At the indicated chase times, the cells were solubilized, and the aliquots immunoprecipitated and processed in SDSjPAGE (7.5%) and fluorographed. The percentage of radioactivity of 35S-labelled protein of the 180-kDa band is represented on the ordinate after having been scraped from the gel and measured on a liquid scintillation counter. The 100% value (at zero time) was 19100 cpm for the incubations in the presence of tunicamycin and of 26900 cpm when the antibiotic was absent

the 180-kDa protein declined with a half-life of 92 rnin and the 164-kDa protein with a half-life of about 605 min. This result therefore excluded the hypothesis that the change in half-life caused by the tunicamycin was due to the effect of this product on a hypothetically sensitive proteinase.

As cholesterol diminishes the stability of the 180-kDa protein, we were interested in knowing whether the carbohydrate- free form was more, or less easily degraded by cholesterol than the glycoprotein. Thus hepatocytes were preincubated for 90 min with phosphatidylcholine. They were then incubated with phosphatidylcholine and [35S]methionine both in the presence and in the absence of tunicamycin. Following this, they were chased in the presence of unlabelled methionine and dispersions of phosphatidylcholine on its own and plus sterols, both in the presence and in the absence of tunicamycin. The radioactivity of the 180-kDa and 164-kDa bands was determined as described. Fig. 8 shows the results obtained. The addition of cholesterol to the cells incubated with tunicamycin was found to change the turnover rate of protein, at variance with the case of cells that have not been incubated with tunicamycin. These results suggest that the absence of carbohydrate not only stabilizes the 180-kDa protein, but also disconnects the degradation effect caused by cholesterol. It appears that carbohydrates trigger the acceleration of the degradation of this cholesterol-regulated protein, both in the presence and in the absence of cholesterol.

DISCUSSION

By using immunoblotting analysis and immunoprecipitation techniques with anti-(HMG-CoA reductase) antibodies, we have found two polypeptides with molecular masses of 10CkDa and 180-kDa. The occurrence of a possible HMG- CoA reductase form with a molecular mass of about 180 kDa has already been reported by various authors. Chin et al., whilst studying immunoprecipitates of UT-1 cells, found the presence of this form together with another of 160 kDa in a proportion of no more than 7% [17]. They considered it an aggregated form of the 97-kDa band that was enhanced through boiling SDS, and disaggregated by urea/SDS. Hardeman et al. [18], studying the synthesis of HMG-CoA reductase in compactin-resistant C-1 00 cells, reported that immunoprecipitates of [35S]methionine-labelled cells display- ed a band of 170 kDa whose origin and significance was not clear. Ness et al. initially reported that in rat liver there was only one form of 104 kDa which was not glycosylated [21]. However they then almost simultaneously reported the existence of two forms 1191 that were interconvertible depending on the level of the redox state of the sample, resulting from the thiol reagents present in the medium. When the sample had been isolated with 10 mM dithiothreitol, only the 104- kDa reductase could be seen; however, when the redox state had been moved to a more oxidized medium, the level of the 200-kDa band increased.

The data reported from other laboratories suggesting that the 180-kDa protein is in fact HMG-CoA reductase are based merely on circumstantial evidence. There is no evidence as to the enzymatic properties of the 180-kDa protein. Further, the results that emerged after boiling the immunoprecipitates of rat hepatocytes, both in the presence and in the absence of 2- mercaptoethanol, and 8 M urea, support the notion that the 180-kDa protein is neither an aggregate nor a dimer form of HMG-CoA reductase. The experiments carried out with new antibodies, showing that they could titrate HMG-CoA re-

129

ductase activity without immunoprecipitating the 180-kDa protein, demonstrated that there is no cross-reactivity between the two proteins and thus that they are different. Research groups working with anti-(HMG-CoA reductase) antibodies can be classified under three headings: (a) those who have never reported the occurrence of this 180-kDa protein [9]; (b) those who have observed this protein accidentally, but have not attempted to characterize it [17, 181; (c) the group of Ness et al. who first did not report this protein [21], but later [19] noticed its occurrence. The apparent cross-reactivity between the 180-kDa protein and HMG-CoA reductase could be explained by the fact that there probably exists, in the course of the HMG-CoA reductase purification, a proteolyzed fraction of the 180-kDa protein (size 52 kDa) that sometimes copurifies with it throughout the whole process. When the mixture of these two proteins is injected in rabbit, it elicits not only anti-(HMG-CoA reductase) antibodies but also anti- (1 80-kDa protein) antibodies. We have observed that, under conditions of proteolysis, a fraction of 35S-labelled protein of 52 kDa appeared in SDSjPAGE which we initially attributed to HMG-CoA reductase. Because of its abundance, it probably corresponds to the proteolyzed fraction of the 180-kDa protein that copurifies with the proteolyzed fraction of HMG- CoA reductase (52 kDa). In addition, the 97-kDa and 180- kDa proteins have different turnover rates, different carbohydrate contents, different indexes of phosphorylation [8] and, moreover, the deglycosylated protein does not have a molecular mass of double that of the 104-kDa protein [17, 221.

The 180-kDa protein is regulated by mevalonate and cholesterol and its synthesis and turnover rate are decreased and increased, respectively, when hepatocytes are incubated with cholesterol. The synthesis of the 180-kDa protein is impaired when rat hepatocytes are incubated with mevalonate [8, 111.

We suggest therefore that the 180-kDa protein is a microsomal protein, related to cholesterol metabolism, that is different from the HMG-CoA reductase and which can be regulated by the cholesterol and mevalonate levels. It can be solubilized by a process of freeze-thawing together with HMG-CoA reductase and, for unknown reasons, can co-pu- rify with it. Anti-(l80-kDa protein) antibodies are elicited at the same time as the anti-(HMG-CoA reductase) antibodies are apparently produced.

This protein is a glycoprotein of N-linked high-mannose oligosaccharide chains; it is sensitive to endo H action which diminishes its apparent molecular mass by 16 kDa. These data are supported by the fact that, when rat hepatocytes are incubated with the inhibitors of glycosylation (tunicamycin) the high molecular mass protein appears with a lower molecular mass (164 kDa) identical to that obtained when the immunoprecipitates have been treated with endo H.

Since the enzymatic steps of N-glycosylation are located in the lumen of the endoplasmic reticulum, it appears that a portion of the 180-kDa protein, present on the luminal sur- face, penetrates endoplasmic reticulum membrane. Further- more, this microsomal protein is not processed on the Golgi system, as over a period of time, similar to that of its half-life, the protein does not acquire resistance to endo H.

Since the proportion of mannose carbohydrate chains present in the 180-kDa protein is apparently 10% by mass, we investigated whether the carbohydrate moiety played a role in its degradation. The current data establishes that the carbohydrate-free protein was degraded at less than a tenth the rate of the glycoprotein, and that its degradation was not accelerated when rat hepatocytes were incubated in the presence of a mixture of cholesterol and 25-hydroxy-cholesterol,

(a variance which occurs when 180-kDa protein contains the mannose carbohydrates). However, the turnover rate of the total protein did not change on incubation with cholesterol.

It is clear therefore that there is a correlation between the absence of carbohydrates and the stability of the protein. It appears that the presence of carbohydrates could be the signal for the movement of the protein to its degradation site, and similarly the absence of carbohydrates could promote a protein conformation not susceptible to degradation. The 180- kDa protein could be a good model for the study of the degradation of endoplasmic reticulum proteins.

Research work at the molecular level will provide infor- mation on the gene encoding the 180-kDa protein, along with the regulation of its expression by cholesterol and mevalonate.

We are very grateful to Mr Robin Rycroft and Miss Sophie Newick for their valuable assistance in the preparation of the English manuscript. This work was supported by Grant PB86-0514 from Comisih Asesora de Investigacibn Cientqica y Ticnica, Spain

REFERENCES 1. Brown, M. S. & Goldstein, J. L. (1980) J. Lipid Res. 21, 505-

517. 2. Mehrabian, M., Callaway, K. A,, Clarke, C. F., Tanaka, R. D.,

Greenspan, M., Lusis, A. J., Sparkes, R. S., Mohandas, T., Edmond, J., Fogelman, A. M. & Edwards, P. A. (1986) J . Biol. Chem. 261, 16249-16255.

3. Faust, J. R., Goldstein, J. L & Brown, M. S. (1979) Proc. Natl Acad. Sci USA 76, 5018 -5022.

4. Slakey, L. L., Craig, M. C., Beytia, E., Brieris, A., Feldbruegge, D. H., Dugan, R. E., Qureshi, A. A,, Subbarayan, C. & Porter, J. W. (1972) J . Bid. Chem. 247, 3014-3022.

5. Clarke, C. F., Tanaka, R. D., Svenson, K., Wamsley, M., Fogelman, A. M. & Edwards, P. A. (1987) Mol. Cell B i d . 7,

6. Gonzalez, R., Carlson, J. P. & Dempsey, M. E. (1979) Arch. Biochem. Biophys. 196, 574-580.

7. Gaylor, J. L. (1981) in Biosynthesis of isoprenoid compounds (Porter, J. W. & Spurgeon, S. L., eds) vol I, pp. 481 - 543, John Wiley and Sons, New York.

8. Marrero, P. F., Haro, D. & Hegardt, F. G. (1986) FEBS Lett.

9. Edwards, P. A., Lan, S-F., Tanaka, R. D. & Fogelman, A. M.

3138-3146.

197,183-186.

(1983) J . Biol. Chem. 258,7272-7275 10. Edwards, P. A. (1975) Biochim. Biophys. Acta. 409, 39-50. 11. Haro, D., Marrero, P. F. & Hegardt, F. G. (1985) Biochem.

Biophys. Res. Commun 132, 598 - 604. 12. Olden, K., Bernard, B. A,, Humphies, M. J., Yeo, T-K., Yeo,

K-T., White, S. L., Newton, S. A., Banes, H. C. & Parent, J . B. (1 985) Trends Biochem. Sci. 10, 78 - 82.

13. Laemmli, U. K. (1970) Nature 227,680-685. 14. Eagle, H. (1959) Science 130, 342. 15. Beutler, H. 0. & Henniger, G. (1981) SwisFood3, 27-29. 16. Siedel, J., Schlumberger, H., Klose, S., Ziegenhorn, J. &

Wahlefeld, A. W. (1981) J . Clin. Chem. Clin. Biochem. 19, 838. 17. Chin, D. J., Luskey, K. L., Faust, J. R., MacDonald, R. J., Brown,

M. S. & Goldstein, J. L. (1982) Proc. Nut1 Acad. Scz. USA 79,

18. Hardeman, E. C., Jenke, S. H. & Sirnoni R. D. (1983) Proc. Natl Acad. Sci. USA 80,1516-1520.

19. Ness, G. C., McCreery, M. J., Sample, C. E., Smith, M. & Pendleton, L. C. (1985) J . Bid. Chem. 260, 16395-16399.

20. Liscum, L., Cummings, R. D., Anderson, R. G. W., de Martino, C. N., Goldstein, J. L. & Brown, M. S. (1983) Proc. Nut1 Acad. Sci. USA 80,7165-7169.

21. Ness, G. C., Sample, C. E., Smith, M., Pendleton, L. C. & Eichler, D. C. (1986) Biochem. J . 233, 167-172.

22. Liscum, L., Luskey, K. L., Chin, D. J., Ho, Y . K., Goldstein, J. L. & Brown. M. S. (1983) J . Biol. Chem. 258.8450-8455.

7704 - 7708.

Documents

Identification of a cholesterol-regulated 180-kDA microsomal protein in rat hepatocytes