ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 290, No. 1, October, pp. 143-152, 1991

Stable Ornithine Decarboxylase in a Rat Hepatoma Cell Line Selected for Resistance to a-Difluoromethylornithine’ John

L. A. Mitchell,2

Department

Jane A. Hoff,

of Biological Sciences, Northern

and Aviva Illinois

Bareyal-Leyser

University,

DeKalb, Illinois 60115

Received April 25, 1991; and in revised form June 10, 1991

Ornithine decarboxylase (ODC) is extremely unstable in mammalian cells. This unusual characteristic facilitates rapid fluctuations in the activity of this enzyme in response to variations in its biosynthesis. Unfortunately, very little is known about the mechanism or regulation of this ODC-specific proteolytic pathway. This study describes the production and characterization of a variant of the rat hepatoma HTC cell line that is strikingly deficient in this pathway. This cell variant was induced by selection for growth in stepwise increasing concentrations (up to 10 mM) of the irreversible ODC inhibitor, (Ydifluoromethylornithine (DFMO). Resistance to this inhibitor appears to result from a combination of elevated (10X) ODC biosynthesis and inhibited degradation, producing greater than a 2000-fold increase in the level of ODC protein. In these variant cells (DH23b) inhibition of protein synthesis by cycloheximide did not result in rapid loss of enzyme activity or ODC protein determined by radioimmunoassay. Pulse-chase studies with [3SS]methionine confirmed that this enzyme was not preferentially degraded, even when spermidine was added to the media. ODC purified from the variant cells was found to be identical to the control cell enzyme in size, isoelectric point, substrate binding kinetics, and sensitivity to the inhibitor DFMO. Also, as in the control cells, a major fraction of the ODC molecules extracted from DH23b cells was shown to be phosphorylated on a serine residue. The inability to detect physical or kinetic differences between the parent and the variant cell ODC suggests that the unusual stability of ODC in this cell is associated with a defect in a cellular mechanism for ODCo 1991Academic PEW, IN. specific degradation.

tine, spermidine, and spermine (l-3). The initial enzyme in the polyamine biosynthetic pathway, ornithine decarboxylase (ODC)3 (EC 4.1.1.17), is particularly noted for its extremely sensitive regulation. Very rapid increases and decreases in the cellular level of this enzyme are induced by changes in hormones, growth factors, or other signaling agents. Variations in the rate of synthesis of this enzyme, due to transcriptional (4-6) and translational (7-9) controls, result in rapid fluctuations in ODC levels because this protein is extremely unstable in the cell (10, 11). This unusual enzyme instability does not appear to require either lysosomal activity or ubiquitin ligation (1214), and the mechanism of this very short half-life and its control are not well understood. Several characteristics of this enzyme have been investigated for their contribution to this extremely rapid protein turnover, including its propensity for amino-terminal arginylation (15), the presence of “PEST” regions (16), and a particular carboxy-terminal end segment that appears to be destabilizing (17). Additionally, ODC degradation is markedly enhanced by cellular increases in the pathway products, the polyamines. This feedback response may involve the induction of a small, specific regulatory protein, antizyme (18-21); however, the mode of action of this regulator has not been elucidated. The mechanism of rapid ODC degradation appears to be altered somewhat in a variant cell line, HMOA, which was selected from rat hepatoma (HTC) cells through its resistance to the competitive ODC inhibitor Lu-methylornithine (22-24). In these cells the normal ODC half-life of less than 20 min has been extended to between 5 and 10 h (23, 24), resulting in a two- to fivefold elevation in

Mammalian growth processes appear to involve careful management of the synthesis of the polyamines putresr This research was supported by NIH Grants GM-33841 51251. ’ To whom correspondence should be addressed. 0003-9861/91 $3.00 Copyright 0 1991 by Academic Press, All rights of reproduction in any form

and CA-

3 Abbreviations used: ODC, ornithine decarboxylase; DFMO, difluoromethylornithine; DTT, dithiothreitol; PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate; Chaps, 3-[(3-cholamidopropyl)dimethylammino]propanesulfonate; PLP, pyridoxol5-phosphate. 143

Inc. reserved.

144

MITCHELL,

HOFF,

AND TABLE

ODC Activity

Cell line

HTC DH23b

Time after media change” (h) 4 24 24 24 24

BAREYAL-LEYSER I

and Protein in DH23b

Media additions

and HTC

Cells

ODC activity (U/mg -t SD; IZ > 4) 4.6 0.9 24 49.9 172

MGBG, cyclohexylamine’ O-l day off DFMO 3-4 day off DFMO

? 1.6 +_0.65 + 8.0 f 9.4 f 85

ODC protein* (ng/mg + SD; n > 4) 3.6 0.75 19.1 9900 1970

9~ 1.2 f 0.45 + 8.2 f 5300 f 1600

’ ODC assay and protein determinations were performed either 4 or 24 h after the last resuspension of the cells in fresh medium. * ODC protein was determined by radioimmunoassay as described under Experimental Procedures. ’ MGBG (0.25 pM) and cyclohexylamine (0.125 mM) were added to depress cellular levels of spermidine as described previously (11).

the levels of this protein. This variant line does, however, show the normal response of rapid ODC degradation in the presence of exogenously added spermidine or spermine (11, 25-28). It has been suggested that the slow constitutive rate of ODC degradation in this variant may be associated with a delayed degradation of the ODC-antizyme complex (26). As a consequence of the extremely short half-life of ODC protein only very low cellular levels of this enzyme are observed, making it a very difficult protein to study (29). To circumvent this problem investigators have selected for cells overproducing ODC by using stepwise increments of the irreversible inhibitor difluoromethylornithine (DFMO). Mammalian cells resistant to this drug almost always exhibit amplification of ODC gene expression resulting from both increased gene copy number and ODC-mRNA (30-34). ODC-specific protein synthesis has been increased in some of these lines so that it is a few thousand fold greater than normal, and several percent of the total protein synthesized is ODC that is rapidly inactivated by the DFMO. In spite of this extremely high rate of ODC synthesis in these variants selected for longterm resistance to DFMO, this inactivated ODC does not accumulate to become a major cell protein. This is due to the fact that in all of these variants, except for one in which the structure of the protein was substantially altered (35), ODC retains its characteristically short halflife (36-38). Since rapid ODC degradation is independent of this large variation in ODC protein concentration, it appears that either there is no specialized component required for this degradation or any specific proteins required for rapid ODC degradation have been coamplified along with ODC expression. In our attempts to select for an overproducing variant from HTC cells, an established rat hepatoma cell line, we developed an unusual DFMO-resistant variant, DH23b. This variant exhibits ODC protein levels several thousand times that of the parental cells; however, the rate of ODC synthesis is not elevated proportionally. Unlike all other DFMO-selected variants, this one accumulates ODC pro-

tein because it does not selectively and rapidly degrade ODC. This variant is also quite distinct from the a-methylornithine-selected variant, HMOA, in that it is completely resistant to DFMO, and both its constitutive and polyamine-enhanced ODC-degradation pathways are inoperative. In this study we have characterized this peculiar cell line and the amplified ODC protein in an attempt to increase our understanding of the processes that are normally responsible for the rapid degradation of this enzyme. EXPERIMENTAL

PROCEDURES

Materials. L-[l-‘4C]Ornithine (51 Ci/mol) and [32P]orthophosphate DL-a-[&4(600 mCi/ml) were obtained from ICN Radiochemicals. 3H]DFM0 (20 Ci/mmol) and 35S protein-labeling mix were obtained from DuPont-New England Nuclear. S-Adenosyl-L-[carbozy-‘4C]methionine (62 Ci/mol) was purchased from Amersham Corp. Biochemicals were from Sigma and Pierce Chemicals. DFMO was a generous gift of Merrell Dow Research Institute (Cincinnati, OH). Monospecific rabbit antiserum prepared against purified mouse kidney ODC was kindly provided by Dr. Lo Persson. Cell culture. Rat hepatoma (HTC) cells and the subclone (HMOA) were grown in suspension culture in Swim’s 77 medium containing 10% calf serum. Horse serum was substituted where indicated to minimize alteration of added polyamines. The DFMO-resistant variant, DH23b, was obtained by growing HTC cells in stepwise-increasing concentrations of DFMO starting at 0.02 mM. Finally cells that could grow in 10 mM DFMO at the same rate as the parental cells in the absence of inhibitor were selected. After the initial selection period of about 6 months the cells were maintained on 10 mM DFMO for 25 passages and were designated DH23a. After 60 passages at this DFMO level the cells appeared to be quite homogeneous and were designated DH23b. This variant has not been successfully cloned. ODC activity assay and protein determination. ODC activity was assayed by measuring the release of 14C02 from L-[l-“Clornithine as described previously (21, 39). One unit is defined as the release of 1 nmol 14COz/h. In assaying ODC from DH23b cells grown on DFMO it was necessary to dialyze the homogenates against assay buffer or use sufficiently small samples to minimize the possibility of DFMO carrying over to the assay reaction. Total cell protein was measured by the method of Bradford (40). ODC protein was determined by radioimmunoassay using monospecific antibody prepared in rabbits against ODC purified from the DH23b cells. Samples of this purified ODC were used as standards in all radioimmunoassays. Purified DH23b ODC was labeled with “‘1 using the Iodo-Bead method of Pierce Chemicals.

RAT

HEPATOMA

ORNITHINE

DECARBOXYLASE 30

400

STABILITY

145

VARIANT

Preparation of ‘H- and 32P-labeled ODC. ODC derived from HTC cells was labeled by in vitro inactivation with [3H]DFM0 as previously described (39). 32P-labeling of DH23b cells was accomplished by placing 1.25 X lo7 cells in 25 ml of a phosphate-free medium to which 0.6 mCi of 32P, was added. After 3.5 h at 37°C the cells were washed extensively with phosphate-buffered saline. Extracts of these cells were precipitated with monospecific antiserum as described above. Slab isoelectric focusing gels. A 0.75-mm slab gel containing 4% acrylamide, 9.2 M urea, 0.02% Chaps detergent, 1.25% Pharmalyte (Sigma Chemicals), pH 3-10,2.5% Pharmalyte, pH 2.5-5, and 6% Pharmalyte, pH 5-6, was constructed in a Mighty Small II (Hoefer Scientific Instruments). Samples were suspended in a lysis buffer containing 5 mM DTT, 1 g/liter urea, 0.4% Brij 35, and 0.4% Pharmalyte, pH 3-10, for application to the gel. Gels were run for 30 min at 4 W followed by 3 h at 5 W. The gradient was determined by eluting l-cm strips with 1 ml degassed, distilled water and measuring pH with a microelectrode.

DAYS OFF OF DFMO FIG. 1. ODC activity and protein content in DH23b cells that have been removed from DFMO. A suspension culture of DH23b cells grown on 10 mM DFMO was washed and resuspended in fresh medium without DFMO. Every 24 h the culture was again washed and resuspended in fresh medium. Samples withdrawn at the indicated times were analyzed for ODC protein (w) using radioimmunoassay and for ODC activity (0).

S-Adenoxyl-L-methionine decarboxyiuse activity assay. The assay for S-adenosylmethionine decarboxylase follows the exact same procedures as that for ODC except the assay buffer contains 2.5 mM putrescine and 0.05 @i (0.015 pmol) S-[“‘Cladenosylmethione is added instead of the L-ornithine. ODC purification and antibody preparation. Pellets (approx 1 X lo9 cells) of DH23b cells grown O-3 days without DFMO were disrupted by sonication in 20 ml of ice-cold 0.02 M Tris-HCI (pH 7.2) containing 0.1 M NaCl and 5 mM DTT, and the homogenate was centrifuged in a Ti80 rotor for 1 h at 35,000 rpm. The supernatant was chromatographed on a Mono Q column (Pharmacia) using a NaCl gradient from 0.1 to 0.5 M. The peak of ODC activity was then applied to a 20-ml pyridoxamine 5’-phosphate affinity column that was constructed using Affigel-10 (BioRad) using the manufacturer’s directions. This column was washed with approx 10 column vol of 0.02 M phosphate buffer (pH 7.0) containing 0.25 M NaCl, 1.0 mM DTT, and 0.02% Brij 35. The enzyme was then eluted by adding 25 pM PLP to this buffer and concentrated by ultrafiltration. Protein purity was checked by 10% polyacrylamide gel electrophoresis (PAGE) in the presence of the protein denaturant sodium dodecyl sulfate (SDS). Protein content was determined by the Lowry method as adapted by Peterson (41). Monospecific ODC antibody was generated by subcutaneous injections of the purified enzyme, along with Freund’s complete adjuvant for the first injection and incomplete adjuvant for each booster, into New Zealand white rabbits. Immunoprecipitation and preparation for SDS-PAGE. Specific precipitation of labeled ODC from cell homogenates was achieved by diluting samples up to 0.2 ml with 0.02 M phosphate buffer (pH 7.0) containing 0.25 M NaCl, 1.0 mM DTT, 0.02% SDS, and 0.1% bovine serum albumin. Antiserum (2-4 ~1) was added and incubated 2 h at room temperature and overnight at 4’C. The next day 10 ~1 of 10% insoluble protein A was added and the mixture incubated 30 min at 4°C. The precipitate was centrifuged using an Eppendorf centrifuge and the pellet was washed three times with 0.02 M Tris buffer (pH 7.2) containing 1.0 mM DTT, 0.1% SDS, 0.1% Brij 35, and 0.1% bovine serum albumin. This pellet was finally extracted with 50 pl of 1% SDS gel sample buffer, heated for 90 s in a boiling water bath, and centrifuged for 5 min using an Eppendorf centrifuge. Small aliquots of the supernatant were subsequently chromatographed on 10% polyacrylamide SDS gels. Following Coomassie blue staining and destaining, gels were treated with Amplify (Amersham), dried, and visualized using Kodak XAR-5 film.

RESULTS

Characterization of an HTC Cell Variant That Accumulates Large Amounts of ODC Protein Rat hepatoma HTC cells were exposed to selective pressure by increasing concentrations of the irreversible ODC inhibitor, DFMO, eventually establishing a variant

HTC

DH 238

c35s -0DC

FIG. 2. Comparison of 3sS incorporation into HTC and DH23b cells during a 20.min labeling. A culture of DH23b cells was resuspended in fresh medium without DFMO 6 h before the start of this experiment. At the same time a culture of HTC cells was induced for ODC activity by resuspension in fresh medium containing 0.25 fiM MGBG and 0.125 mM cyclohexylamine. These inhibitors of spermidine synthesis extend the half-life of ODC in HTC cells to 3 h or more (11). The HTC and DH23b cells were subsequently washed, resuspended at 5.0 X lo6 cells/ ml in methionine-free medium without inhibitors, and labeled for 20 min with 250 &i/ml [?S]methionine. After multiple washes with icecold, isotonic phosphate-buffered saline the cells were homogenized by sonication and samples containing 5 pg total protein were immunoprecipitated, separated on 10% SDS-PAGE gels, and Auorographed.

146

MITCHELL,

HOFF,

AND

BAREYAL-LEYSER

25 TIME IN CYCLOHEXIMIOE

? 25 I2 n k P

(HRS)

100 60 60

0

2

4

6

TIME IN CYCLOHEXIMIOE

6 (HRS)

FIG. 3. Stability of ODC activity and protein in DH23b cells. In A, DH23b cells grown on 10 mM DFMO were washed and resuspended at 5 X 10’ cells/ml in fresh medium without DFMO. After 18 h, 0.2 mM cycloheximide was added and samples were withdrawn for ODC activity assay (0) and ODC protein determination by radioimmunoassay (0). For comparison, a culture of HTC cells was resuspended in fresh medium for 4 h before cycloheximide was added and samples were withdrawn for ODC activity (0). Initial values were 3.3 fig ODC protein and 46.2 U ODC activity per milligram total protein of the DH23b cells, and 3.4 U/mg protein for the HTC cells. In B, a culture of DH23b cells was prepared and treated with cycloheximide as above, and the cell samples were assayed for S-adenosylmethionine decarboxylase activity (w, initial value, 0.37 nmol ‘%/h/mg protein) as well as for ODC activity (0; initial value, 68 U/mg protein).

(DH23b) that grew in medium containing 10 mM of this inhibitor. When grown at this level of inhibitor, the variant has the same doubling time (24 h) as the parental line in the absence of this compound; however, the variant cells contain approximately 2000 times as much ODC protein (Table I). Inhibitors of spermidine synthesis had been used previously to block the polyamine feedback response and enhance ODC activity in HTC cells (1 l), yet even that induction was only a fraction (0.2%) of the ODC protein levels observed in the variant cells. Consistent with observations made on other ODCoverproducing lines (S), during growth in the presence of DFMO almost all (>99%) of the enzyme is catalytically inactive due to covalently bound inhibitor. The level of ODC protein is maximal in these cells during the first day off of DFMO (Fig. 1) and decreases with subsequent culturing in the absence of this compound. ODC activity,

however, increases during the first 5 days after removal from DFMO, and the portion of the total ODC protein bound to DFMO decreases such that by 10 days the specific activity approaches that of the parental HTC cells (1.2 U/rig of ODC protein). Unlike other DFMO-resistant cell lines that have been described, this variant is actually dependent upon DFMO. Deterioration, with substantial cell lysis, begins 4-5 days after they are removed from the inhibitor (data not shown). All DFMO-resistant mammalian cells previously studied had been shown to compensate for the DFMO inactivation of ODC by gene amplification and increased ODC-mRNA production (30-34), which result in greatly increased biosynthesis of ODC protein. Increases of several hundred to several thousand fold, up to 15% or more of the total protein synthesis, have been observed in such cell lines (8,30,42). Pulse labeling of the DH23b cells for

RAT

HEPATOMA

ORNITHINE

DECARBOXYLASE

Ii k-2 L 100 f;

60

k5

60

e

40

z Fr u Q

20

i

10 0

2

4

6

6

10

0

2

TIME AFTER ADDITION

4

6

8

10

(HRS)

FIG. 4. Effect of spermidine on ODC activity in HTC and DH23b cells. A culture of HTC cells (left) was induced for ODC by resuspension in fresh medium containing 10% horse serum, 0.25 pM MGBG, and 0.125 mM cyclohexylamine. After 19 h this was split into three cultures to which 1.0 mM spermidine (A), 0.2 mM cycloheximide (Cl), or both spermidine and cycloheximide (0) were added. Samples were subsequently drawn and assayed for ODC activity (initial activity, 4.1 U/ mg). A culture of DH23b cells (right) was removed from 10 mM DFMO and resuspended in fresh medium containing 10% horse serum. After 19 h this culture was divided, additions were made, and samples were withdrawn and assayed exactly as with the HTC culture. The initial DH23b ODC activity was 68 U/mg protein.

20 min with [“5S]methionine, followed by ODC-specific antibody precipitation and SDS gel chromatography, revealed that ODC was produced more rapidly in this variant cell than in the parental HTC line (Fig. 2). By cutting and counting the ODC bands from such separations this difference was quantitated, revealing that ODC synthesis in the DH23b cells was approximately 9.5 times that detected in the HTC cells. The ratio of ODC to total protein synthesis in DH23b cells may vary somewhat with induction conditions. In six labelings of cells 0 to 6 h off of DFMO we have observed values from 0.39 to 1.18% of total protein synthesis. This is only a fraction of that observed in other DFMO-resistant cells and certainly not enough to account for the 2900-fold increase in the ODC protein of this variant. As shown in Fig. 3A, the very high levels of ODC protein in DH23b cells can be attributed, at least in part, to a substantial increase in the stability of this enzyme. In comparison to the half-life of 20 min for ODC in the parental HTC line, DH23b ODC activity declined, with a half-life of greater than 12 h. Even this modest decrease in activity may be due largely to continued inactivation by residual DFMO, as ODC protein itself showed less than a 20% decrease in this 24-h period. The DH23b cells do not appear to have a general deficiency in rapid protein degradation as another unstable enzyme, S-adenosylmethionine decarboxylase, is inactivated normally (Fig. 3B). Previous studies have shown that the half-life of normal mammalian cell ODC can be greatly extended by the use

STABILITY

147

VARIANT

of inhibitors that depress cellular spermidine levels (11, 43). Conversely, ODC instability recurs when cellular spermidine levels are increased by addition of this polyamine to the medium (Fig. 4, left). As noted in this figure and previously reported (11,12,44), the induction of rapid ODC degradation by spermidine or spermine does require new protein synthesis and can be substantially inhibited by the simultaneous addition of cycloheximide. Contrary to this response by the HTC cells, the addition of 1.0 mM spermidine to the DH23b cells does not stimulate a rapid inactivation of existing ODC (Fig. 4, right). The stability of ODC protein in the variant cells in the absence and presence of exogenous spermidine was confirmed by pulse-chase studies. DH23b cells labeled with [35S]methionine were placed in unlabeled medium with and without 0.2 mM spermidine, and the loss of ODC protein was evaluated by comparison of antibody-precipitated material eluting at the proper location on SDS gels (Figs. 5A-5C). Quantitative evaluation of these gel separations (Fig. 5C) confirmed that there was essentially no loss in [35S]ODC during the 8-h study period in the absence of added spermidine, and even in the presence of 0.2 mM spermidine the half-life of the ODC protein was at least 8 h. In similar experiments where 0.5 and 1.0 mM spermidine were added, the half-life of [35S]ODC was calculated at between 15 and 20 h. This insensitivity of DH23b ODC to exogenous polyamines was not due to a failure of these cells to incorporate spermidine. Conversely, studies on the uptake of [14C]spermidine into these cells revealed that the polyamine is transported into DH23b cells somewhat more rapidly than into the parental HTC line (data not shown). Furthermore, the DH23b cells, but not the parental HTC cells, appear to be adversely affected by exogenous spermidine and some cell lysis and growth inhibition are noted after 12-24 h. Purification and Characterization Protein from DH236 Cells

of the Stable

ODC

ODC protein was purified from the DH23b cells and extensively characterized to determine whether the stability of ODC in the DH23b cells was associated with a structural modification in the protein itself. Because of the high concentration of ODC in these cells, a three-step procedure was sufficient to produce 1 mg of electrophoretically pure ODC from batches of about 10’ cells. Since purifications started with cells only l-3 days off of DFMO, the specific activity of the resultant purified enzyme was naturally quite low, about 0.04 U/mg. Anti-ODC antibody prepared in rabbits, using a portion of this purified protein, was found to precipitate only one protein on SDSPAGE (Figs. 5 and 6). This band was indistinguishable from [3H]DFMO-labeled ODC prepared from HTC cells, as seen both on SDS (Fig. 6) and on urea isoelectric focusing gels (Fig. 7). The catalytic activities of enzyme from DH23b and HTC cells were also compared. The K, for ornithine was

148 A

MITCHELL, Hours in unlabeled

HOFF,

AND

media

BAREYAL-LEYSER Hours in 0.2mM SPD

B 0

1

2

3

4

5

6

II

-%-ODC

300 0

I

I

I

I

2

4

6

8

TIME IN UNLABELED MEDIA (HRS) FIG. 5. Effect of spermidine on ODC degradation in DH23b cells. A culture of DH23b cells grown on 10 mM DFMO was washed and resuspended in DFMO-free medium for 4 h, then lo7 cells were resuspended in 5 ml methionine-free medium to which 0.5 mCi [?S]methionine labeling mix was added. After 1 h the labeled cells were washed twice and resuspended at 3 X lo5 cells/ml with medium containing 1 mM each of unlabeled cysteine and methionine with (A) or without (B) 0.2 mM spermidine. Samples were withdrawn at the indicated times and homogenized by sonication, and 6.5pg protein aliquots were precipitated with excess ODC antibody. The resulting pellets were washed, run on 10% SDS-PAGE gels, and fluorographed. The appropriate ODC bands for control samples (0) and those exposed to spermidine (a) were cut from the gels, extracted, and counted by liquid scintillation counting (C).

determined to be 0.055 and 0.058 mM for the enzyme from the variant and wild-type cells, respectively. Sensitivity to the selective inhibitor, DFMO, was also determined (Fig. 8) and the calculated K1 of 33 PM agrees with previously published values for this cell line (45). Since phosphorylation has been suspected as a possible regulator of ODC stability (21, 46, 47), DH23b ODC was examined for this particular post-translational modification. [32P]Phosphate presented in a phosphate-limited culture medium was readily incorporated into immunoprecipitable protein coeluting with ODC on SDS-PAGE, as shown in Fig. 6. As reported in several cell systems (47, 48), the phosphate was found to associate only with the more negatively charged ODC species, ODC-II (Fig.

7). As shown in these studies, immunoprecipitates of crude DH23b as well as the highly purified enzyme both contain ODC-I and ODC-II. As we and others have reported (4951), phosphorylation of ODC within an intact mammalian cell appears to be predominantly, if not exclusively, on a serine residue. Paper electrophoresis of hydrolyzed DH23b ODC immunoprecipitated from 32P-labeled cells reveals that this enzyme also appears to be labeled only on a serine residue (Fig. 9). DISCUSSION

The DFMO-resistant variant described in this work is unique in that it is unable to rapidly and selectively de-

RAT

HEPATOMA

DH238

HTC P

ORNITHINE

DECARBOXYLASE

3H

BSA-

I-ODC ov-

FIG. 6. Comparison of ODC from DH23b and HTC cells on SDS PAGE. Immunoprecipitates of ODC from DH23b cells labeled with ?S were prepared as in Fig. 5. DH23b cells were labeled with inorganic 32P as described under Experimental Procedures and a portion was immunoprecipitated. ODC prepared from HTC cells was labeled by inactivation with [3H]DFM0 as described previously. All samples were chromatographed on a 10% SDS-PAGE gel and visualized by fluorography.

grade ODC. In all other DFMO-resistant cell lines the characteristic instability of this enzyme (half-life of 2050 min) has been maintained even though the synthetic rates were increased several hundred fold (31,32,36,37). This most unusual variant provides an unprecedented opportunity to enhance our currently limited understanding of the mechanisms and control of ODC degradation. It is conceivable that the DH23b cells contain an altered ODC gene that produces enzyme lacking a particular sequence required for rapid proteolytic degradation. For example, PEST regions on this protein have been identified as possible proteolysis determinants whose modification could retard any preferential protein degradation (16). Specifically, Ghoda et al. (17,52) have demonstrated that removal of only a short segment of the C-terminal end of this protein greatly promotes the stability of this enzyme. Such an alteration in the primary sequence of the enzyme would likely involve a detectable difference in size, charge, or catalytic kinetics, yet we were unable to detect any of these differences in comparisons with ODC isolated from the parental cell line. Of course, silent amino acid substitutions or minor deletions could still exist and be undetectable without direct sequence comparison of the variant and parental genes. It is also possible that the DH23b cell is defective in some post-translational modification of this enzyme such

STABILITY

149

VARIANT

as the N-terminal arginylation (15) or phosphorylation (21, 47), as these have also been associated with ODC stability. In this study we observed that phosphorylation of ODC does proceed in the variant ceil in a manner that is at least qualitatively similar to that of the parental line. If DH23b ODC protein is indeed normal then it is quite probable that the underlying defect responsible for the uncharacteristic stability of ODC in this cell variant will be found in the cellular mechanism for protein degradation. Since these cells demonstrate the short half-life characteristic of the enzyme S-adenosylmethionine decarboxylase, any defect would probably be limited to the ODC-specific pathway. A defect in a process required for rapid ODC degradation is also thought to exist in the LYmethylornithine-resistant line, HMOA, which had been derived from the same parental line, HTC (22-24). In both variants the half-life of ODC is greatly extended; however, distinct mechanisms of action seem to be involved, because the cells are strikingly different in their physiology of ODC regulation. Most noticeable is the observation that DH23b cells must be routinely maintained in 10 mM DFMO, while the HMOA line cannot grow for more than 1 week in 0.1 mM of this irreversible inhibitor. In HMOA cells the level of ODC activity is 2- to s-fold greater than that in the HTC cells (22), while ODC protein in DH23b cells is increased approximately 2000-fold. The stability of ODC noted in the HMOA variant (half-life of 5 to 10 h) is only slightly greater than that observed in the parental (HTC) cell when grown in the presence of inhibitors of spermidine synthesis (11, 46). By contrast, the half-life of ODC in the DH23b is in excess of 24 h.

-pH

5.7

-pH

5.2

ODC-I ODC-II

FIG. 7. Charge forms of ODC isolated from HTC and DH23b cells. Samples of ODC from several sources were denatured in 8 M urea and compared on a slab isoelectric focusing gel. Lane 1, HTC ODC labeled with [3H]DFM0, visualized by fluorography. Lane 2, Coomassie blue stain of an antibody-precipitated pellet of ODC prepared from DH23b cells labeled with inorganic “P Lane 3, Auorograph of lane 2. Lane 4, Coomassie blue stain of purified ODC prepared from DH23b cells.

150

MITCHELL,

HOFF,

AND

BAREYAL-LEYSER

MINUTES

-0.1

0

0.2

0.1

11 DFMOI

0.3

0.4

(FM)

FIG. 8. Sensitivity of ODC isolated from DH23b to inactivation by DFMO. A sample of DH23b cells that had been growing for 24 h without DFMO was homogenized and dialyzed extensively against 0.02 M PO, (pH 7.2) buffer containing 2 mM DTT, 50 pM PLP, and 0.1% bovine serum albumin. Aliquots were then incubated at 37’C in 0 (U), 3 (m), 5 (O), 10 (O), 20 (a), or 40 (A) FM DFMO. At the time points indicated replicate 5-~1 samples were removed and assayed for remaining ODC activity. Half-lives were determined from the initial slopes in A and used to create the double-reciprocal plot in B.

Another unique aspect of the DH23b physiology is suggested by the observation that the addition of spermidine did not stimulate rapid ODC proteolysis. Mammalian cells, including the mutated HMOA line, generally respond to exogenous polyamines by shortening the halflife of ODC to less than 20 min (11, 25-28). Conversely, as shown in Fig. 4 and previous studies (11,43), the halflife of ODC can be greatly extended by lowering the cellular levels of spermidine. While this suggests that the normal mammalian ODC degradation pathway contains a polyamine-dependent step, some investigators think that there are two paths for rapid ODC degradation, only one of which is stimulated by polyamines (17, 52). The stability of DH23b ODC in both the presence and the absence of added spermidine suggests that there is only one ODC degradation pathway, or if there are two, they contain a common step that is defective in this cell. A rather puzzling observation made in this variant is that dialyzed homogenates from cells grown on DFMO

exhibit ODC activity at least 10 times as great as that of HTC cells induced by fresh serum. Why has all of this enzyme not been inactivated with the ubiquitous DFMO? This is not just newly synthesized enzyme as inhibition of protein synthesis in cells still on DFMO did not lead to a more rapid loss in enzyme activity than that described in Fig. 3A (unpublished observations). DFMO is thought to enter mammalian cells by passive diffusion (53). Homogenates made from DH23b cells do indeed contain DFMO and it must be dialyzed away before ODC activity can be assayed. Similarly, the DH23b ODC is not insensitive to DFMO since enzyme purified from these cells was found to have the exact same sensitivity to DFMO as enzyme from the parental cells. In order for the ODC within the cell to be inactivated by DFMO the irreversible inhibitor must be within the same cell compartment and the enzyme must be enzymatically active. Enzymatic activity does not appear restricted by insufficient pyridoxal 5’-phosphate levels as greater than 85% of the activity in

RAT

HEPATOMA

ORNITHINE

DECARBOXYLASE

STABILITY

151

VARIANT

2. Pegg, A. E. (1986) Biochem.

J. 234,

249-262.

3. Pegg, A. E. (1988) Cnncer Res. 48, 759-774. 4. Morris, D. R., and White, M. W. (1989) A&. Ezp. Med. Biol. 250, 241-252. 5. Hirvonen, A. (1989) Biochim. Biophys. Actu 1007, 120-123. 6. Katz, A., and Kahana, C. (1987) Mol. Cell. Biol. 7, 2641-2643. 7. Kameji, T., and Pegg, A. E. (1987) J. Biol. Chem. 262, 2427-2430. 8. Persson,

L., Holm, I., and Heby, 0. (1988) J. Biol. Chem. 263, 3528-3533. 9. Holm, I., Persson, L., Stjernborg, L., Thorsson, L., and Heby, 0. (1989) Biochem. J. 258, 343-350.

10. Russell, D. H., and Snyder, S. H. (1969) Mol. Pharmacol. 5, 253262. 11. Mitchell, J. L. A., Mahan, D. W., McCann, P. P., and Qasba, P. (1985) Biochim. Biophys. Acta 840, 309-316. 12. Glass, J. R., and Gerner, E. W. (1987) J. Cell. Physiol. 130, 133141. 13. Bercovich, Z., Rosenberg-Hasson. Y., Ciechanover, A., and Kahana, C. (1989) J. Biol. Chem. 264, l&949-15,952. 14. Rosenberg-Hasson, Y., Bercovich, Z., Ciechanover, A., and Kahana, C. (1989) Eur. J. Biochem. 185, 469-474.

.

15. Kopitz, J., Rist, B., and Bohley, P. (1990) Biochem. J. 267, 343348. 16. Rogers, S., Wells, R., and Rechsteiner, M. (1986) Science 234,364367.

Origin

FIG. 9. Identification of serine as the major phosphorylated amino acid in ODC from DH23b cells. DH23b cells were labeled with inorganic 32P as described previously. ODC-specific polyclonal antibody was used to precipitate the enzyme from 2 mg of total cell protein. These pellets were resuspended in 0.02 M Tris-HCI (pH 7.2) containing 1% SDS and chromatographed on a Superose 12 gel filtration column using a Pharmacia FPLC system. The [32P]ODC peak was then precipitated with 7% TCA, washed, lyophilized, and hydrolyzed in 6 N HCl for 1 h at 110°C. The hydrolyzed material was electrophoresed on a cellulose plate in a 5% acetic acid, 0.5% pyridine buffer (pH 3.5) for 40 min at 750 V. The plate was stained with ninhydrin, dried, and autoradiographed. Ninhydrin-stained areas and the inorganic phosphate spot are indicated by dashed outlines on the autoradiograph. Identifications were made by direct comparison with phosphoamino acid standards prepared and chromatographed at the same time.

17. Ghoda, L., Phillips, M. A., Bass, K. E., Wang, C. C., and Coflino, P. (1990) J. Biol. Chem. 265, 11,823-11,826. 18. Fong, W. F., Heller, J. S., and Canellakis, E. S. (1976) Biochim. Biophys.

Acta

428,

456-465.

19. Heller, J. S., and Canellakis, 217.

107,209-

E. S. (1981) J. Cell. Physiol.

20. Murakami, Y., and Hayashi, S. (1985) Biochem. J. 226,893-896. 21. Mitchell, J. L. A., and Chen, H. J. (1990) Biochim. Biophys. Acta 1037, 115-121.

P. S., Duchesne, M., Grove, J., and Tardif, C. (1978) 115,387-393.

22. Mamont, CellRes.

23. McCann, P. P., Tardif, C., Hornsperger, J. Cell. Physiol. 99, 183-190. 24. Pritchard, Chem.

J.-M., and Bohlen, P. (1979)

M. L., Pegg, A. E., and Jefferson,

257,

Enp.

L. S. (1982) J. Biol.

5892-5899.

25. Holm, I., Persson, L., Heby, O., and Seiler, N. (1988) Biochim. phys. Acta

crude homogenates can be detected without adding this coenzyme to the assay mixture. Yet, since we can detect a large amount of active ODC in DH23b cells grown on 10 mM DFMO, it may be that either the enzyme is sequestered from the DFMO or much of it is maintained in an inactive configuration when it is in the intact cell. The latter appears more likely in view of the low levels of putrescine and polyamines observed in these cells on DFMO (Mitchell et al., unpublished). Any such mechanism for reversibly modifying the enzymatic activity of ODC within an intact cell is of innate interest in the regulation of polyamine biosynthesis and certainly warrants further investigation. REFERENCES 1. Tabor, C. W., and Tabor, H. (1984) Annu. 790.

Reu. Biochem.

53,

749-

26. Murakami,

Y., Fujita,

Biochem.

J. 225,

27. Murakami,

K., Kameji,

T., and Hayashi,

S. (1985)

689-697.

Y., Nishiyama,

M., and Hayashi,

S.-I. (1989) Eur.

J.

180, 181-184.

Eiochem.

28. Kanamoto, Eur.

Bio-

972,239-248.

K., Utsunomiya,

K., Kameji, T., and Hayashi, S. (1986)

154,539-544.

J. Biochem.

29. Seely, J. E., Poso, H., and Pegg, A. E. (1982) Biochem.

J. 206,311-

318. 30. McConlogue, L., and Coffino, P. (1983) J. Biol. Chem. 258,12,08312,086. 31. Pegg, A. E., Secrist, J. A., III, and Madhubala, R. (1988) Cancer Res. 48,

32. Dircks,

2678-2682.

L., Grens, A., Slezynger, T. C., and Scheffler,

J. Cell. Physiol.

I. E. (1986)

126, 371-378.

33. Choi, J. H., and Scheffler, I. E. (1983) J. Biol. Chem. 258, 12,60112,608. 34. McConlogue, L., Dana, S. L., and Coffino, P. (1986) Mol. Cell. Biol. 6,2865-2871.

152

MITCHELL,

35. Poso, H., Karvonen,

E., Suomalainen,

HOFF,

H., and Andersson,

AND

L. C.

(1984) J. Biol. Chem. 259, 12,307-12,310. 36. Pegg, A. E., Madhubala,

R., Kameji, T., and Bergeron, R. J. (1988) J. Biol. Chem. 263, 11,008-11,014. 37. McConlogue, L., and Coffino, P. (1983) J. Biol. Chem. 258, 8384-

BAREYAL-LEYSER 46. Mitchell, J. L. A., Qasba, P., and Mahan, D. W. (1985) in Recent Progress in Polyamine Research (Selmeci, L., Brosnan, M. E., and Seiler, N., Eds.), pp. 55-64, Akademiai Kiado, Budapest. 47. Mitchell, J. L. A., Hicks, M. F., Chen, H. J., and Hoff, J. A. (1989) Adu. Exp. Med. Biol. 250, 55-70.

8388. 38. Choi, J., and Scheffler, I. E. (1981) Somatic Cell Genet. 7, 219-233.

48. Peng, T., and Richards, J. F. (1988) B&hem.

39. Mitchell, J. L. A., Rynning, M. D., Chen, H. J., and Hicks, M. F. (1988) Arch. Biochem. Biophys. 260, 585-594. 40. Bradford, M. (1976) Anal. Biochem. 72, 248-254. 41. Peterson, G. L. (1977) Anal. Biochem. 83, 346-356. 42. Poulin, R., and Pegg, A. E. (1990) J. Biol. Chem. 265, 4025-4032. 43. Karvonen, E., and Poso, H. (1984) Biochim. Biophys. Acta 791,

49. Meggio, F., Flamigni, F., Caldarera, C. M., Guarnieri, C., and Pinna, L. A. (1984) Biochem. Biophys. Rex Commun. 122,997-1004.

239-243. 44. Hayashi, S. I., Kameji, T., Fujita, K., Murakami,

Y., Kanamoto, R., Utsunomiya, K., Matsufuji, S., Takiguchi, M., Mori, M., and Tatibana, M. (1985) Adu. Enzyme Regul. 23,311-329. 45. Danzin, C., Bey, P., Shirlin, D., and Claverie, N. (1982) Biochem. Pharmacol. 3 1,3871-3878.

Biophys. Res. Commun.

153,135-141.

50. Pinna, L. A., Flamigni, F., Meggio, F., Guarnieri, C., and Caldarera, C. M. (1986) in Biomedical Studies of Natural Polyamines, pp. 3743, Cooperative Libraria Universitaria Editrice Bologna, Bologna, Italy. 51. Meggio, F., Flamigni, F., Guarnieri, C., and Pinna, A. L. (1987) Biochim. Biophys. Acta 929, 114-116. 52. Ghoda, L., van Daalen Wetters, T., Macrae, M., Ascherman, and Coffino, P. (1989) Science 243, 1493-1495.

D.,

53. Erwin, B. G., and Pegg, A. E. (1982) Biochem. Pharmacol. 31,2820-

2823.