of Rat Liver Acetyl-coA Car

THE J O ~ A L OF BIOUX~ICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 20, Issue of May 20, pp. 14438-14445, 1994 Printed in U.S.A.

Unique Structural Features and Differential Phosphorylation of the 280-kDa Component (Isozyme) of Rat Liver Acetyl-coA Carboxylase*

(Received for publication, February 21, 1994, and in revised form, March 16, 1994)

Robert Win& Daniel HessB, Ruedi AebersoldS&l, and Roger W. BrownseySII From the Department of $Biochemistry and Molecular Biology and $The Biomedical Research Centre, The University of British Columbia, Vancouver, British Columbia V6T 123, Canada

Rat liver acetyl-coA carboxylase (ACC, EC 6.4.1.2) exhibits major and minor subunits (M, of 265,000 and 280,000 respectively), the structure and function of which are compared in this study. The two subunits co- purified and each contained biotin as demonstrated by avidin reactivity and direct determination of biocytin. In agreement with previous studies, the ACC subunits could be distinguished with specific monoclonal antibodies and differential tissue expression. We now report extensive differences in primary structure revealed by peptide mapping, mass spectrometric analysis of peptides following reverse phase high performance liquid chromatography, and microsequencing of selected peptides. Four peptides derived from the 265-kDa subunit were sequenced and matched sequences within the predicted structure of rat 265-kDa ACC. Although one identical peptide sequence was detected within both subunits (residues 2009-2024 of the 265-kDa subunit), 12 peptides derived from the 280-kDa subunit exhibited entirely novel sequences or matched partially (average 70% identity) with sequences within the 265-kDa subunit. The 280-kDa subunit may also exhibit distinct func- tional properties, since the initial rate of phosphorylation was at least 10-fold greater than that of the 265-kDa subunit in the presence of CAMP-dependent protein kinase. Two-dimensional mapping demonstrated that the tryptic phosphopeptides released from the two ACC subunits are distinct. These structural studies suggest that the 265- and 280-kDa components (isozymes) of ACC are so distinct they may be encoded by separate genes, while the differential phosphorylation observed in vitro suggests a key role for the 280-kDa subunit in regulating enzyme activity within intact cells.

Acetyl-coA carboxylase (ACC, EC 6.4.1.2.)’ catalyzes the for- mation of malonyl-CoA in a two-step reaction, which involves initial ATP-dependent coupling of a carboxyl group (donated by bicarbonate ion) to the enzyme biotinyl prosthetic group and subsequent transfer of the carboxyl moiety to acetyl-coA.

the Medical Research Council of Canada (to R. W. B.). The costs of * This work was supported in part by operating Grant MA8676 from

publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ”advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

n Recipient of a Research Scholarship from the Medical Research Council of Canada.

I/ To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology, The University of British Columbia, Medical Block “A,” 2146 Health Sciences Mall, Vancouver, British Columbia V6T 123, Canada. Tel.: 604-822-3810; Fax: 604-822-5227. ’ The abbreviations used are: ACC, acetyl-coA carboxylase; PAGE, polyacrylamide gel electrophoresis; Mops, 3-(N-morpholino)propanesul- fonic acid; BCA, bicinchoninic acid; Caps, 3-(cyclohexylamino)-l-pro- panesulfonic acid; HPLC, high-performance liquid chromatography; MS, mass spectrometry or mass spectrometer.

Three critical catalytic functions, including the site for attach- ment of biotin and two active sites (biotin carboxylase and carboxyl transferase) are encoded by separate polypeptides in Escherichia coli but within a single multifunctional polypeptide in eukaryotes (for reviews, see Numa and Yamashita (19741, Volpe and Vagelos (19761, Lane et al. (19741, Numa and Tanabe (19841, Brownsey and Denton (19871, and Hardie, 1989). In mammals, regulation of the production of malonyl- CoA is brought about in large part through the control of ACC activity and appears to profoundly affect rates of de novo fatty acid biosynthesis in fat, liver, and lactating mammary gland (Lane et al., 1974; Volpe and Vagelos, 1976; Brownsey and Denton, 1987). In addition, in these tissues, as well as other cell types which exhibit very low rates of lipogenesis (such as heart and skeletal muscle), the provision of malonyl-CoA is likely to play a significant role in regulating fatty acid oxidation at the level of carnitine acyltransferase-I (McGarry et al . , 1978; Awan and Saggerson, 1993; Saddik et al . , 1993). The regulation of ACC is exerted through complex and concerted mechanisms including regulation of protein synthesis and degradation, as well as more acute mechanisms involving the binding of allo- steric ligands and the phosphorylation of as many as 8 serine residues (Brownsey and Denton, 1987; Hardie, 1989; Kim et a l . , 1989; Quayle et al . , 1993). The latter two mechanisms most likely lead to a shift in the equilibrium which exists between ACC dimers with very low specific activity and highly active polymers comprised of up to 10-20 dimers (Gregolin et al . , 1966; Lane et al., 1974). These properties of purified ACC appear also to be reflected in polymerization within intact cells (Clarke et al., 1979; Borthwick et al., 1987).

Determination of the size and composition of the subunits which link pairwise to form the ACC dimers has been difficult because of the susceptibility of the enzyme to limited proteoly- sis (Song and Kim, 1981). Molecular cloning of cDNAs encoding rat mammary (Bai et al . , 1986) and chicken liver (Takai et al., 1988) forms of ACC has provided a definitive estimate of subunit size of 265 kDa for the major form ofACC present in liver, fat, and mammary gland. The related but distinct sequence of yeast ACC has also been recently determined (Al-Feel et al., 1992). It has also been recognized that an additional isoform of ACC exists in a number of tissues, including heart and skeletal muscle where it is the predominant form (Thampy, 19891, and other tissues such as liver and lactating mammary gland where it is a minor component (Bianchi et al. 1990). The newly dis- covered isoform exhibits a subunit with molecular mass of approximately 280,000 and can be distinguished from the 265- kDa isoform immunologically and on the basis of kinetic properties (Thampy, 1989; Bianchi et al., 1990). Although the 265-kDa isoform ofACC is apparently encoded by a single gene (Bai et al . , 1986; Takai et al., 19881, several distinct species of mRNA have been described. A number of mRNA species have been identified which differ in the 5”untranslated region (Luo et al., 1989; Lopez-Casillas et al . , 1991) and an additional

14438

Isozymes of Acetyl-CoA Carboxylase 14439

mRNA encodes a form of ACC from which amino acids 1198- 1206 have been deleted (Kong et al., 1990). None of these mRNA variants would appear to encode a form of ACC which differs sufficiently from the 265-kDa isoform to be able to account for the observed difference in subunit M, observed with the larger isoform. An important caveat is that even quite subtle changes in structure may lead to unexpectedly large changes in apparent subunit size observed upon gel electrophoresis. For example, changes in phosphorylation may alter apparent mobility of some proteins during SDS-PAGE, although no effects of phosphorylation or dephosphorylation of ACC (even at multiple sites) has ever been noted to induce a shift in apparent subunit Mr.

We first became interested in the 280-kDa component of purified preparations of ACC from rat liver when we noted distinct immunoreactivity not shared by the major 265-kDa component (Winz and Brownsey, 1989). Here we provide detailed analysis which compares the structural features of the two high molecular weight components of ACC purified from rat liver. The results obtained underline the extensive differences in structure of the two ACC subunits which suggest that they may be encoded by separate genes. The 280-kDa subunit ofACC is also shown to undergo differential phosphorylation in vitro, raising the possibility that it may play a distinctive func- tional and regulatory role within intact cells.

EXPERIMENTAL PROCEDURES Materials-Most laboratory chemicals and solvents were obtained

from BDH Chemicals Canada Ltd. (Vancouver, British Columbia). Bio- chemicals were mostly obtained from Sigma, including proteinase in- hibitors, the catalytic subunit of MP-dependent protein kinase (and the corresponding inhibitor peptide), and serum albumin, which was the defatted bovine form. Lys-C protease (Achromobacter Protease 1) was from Wako (Dallas, TX). Radioisotopes were from Amersham Int. (Oakville, Ontario, Canada) and reagents for electrophoresis were from Bio-Rad Ltd. (Mississauga, Ontario), except membranes for electrophoretic transfer which were from Millipore Corp., and Quickstain for protein visualization on Immobilon CD membranes from Zoion Re- search Inc. (Allston, MA). Supplies for chromatography were from Phar- macia-LKB (Canada) Inc. (Baie d'Urfe, Quebec), except Whatman DEAE-cellulose (DE-52) from Whatman Int. (Maidstone, Kent, United Kingdom). Avidin was kindly donated by The Canadian Lysozyme Co. (Abbotsford, British Columbia). Thin layer chromatography plates (without fluorescent indicator) were from Kodak (number 13255). Fal- con disposables for tissue culture were from Becton Dickinson Canada Inc. (Mississauga, Ontario).

Purification ofACC from Rat Essues-Wistar rats were maintained on a 12-h lighudark cycle (dark 20.00 to 8.00 h) and allowed free access to laboratory chow (Purina) up to the time of killing (usually 9.00-10.00 h) by asphyxiation in carbon dioxide or by cervical dislocation. Livers were taken from rats larger than 280 g, epididymal fat pads from male rats 140-180 g, and mammary glands from females within 24 h after weaning of the offspring. Upon removal, livers and mammary tissue were placed in ice-cold phosphate-buffered saline while fat pads were placed in ice-cold Krebs-Henseleit buffer (Krebs and Henseleit, 1932), which had been pre-gassed (oxygenxarbon dioxide in ratio 955%. v/v) and which contained glucose (11 m), reduced CaC1, (1.25 mM), and defatted bovine serum albumin (5 mg/ml). After rinsing, all tissues were homogenized in ice-cold extraction buffer (pH 7.4) which contained Mops (50 mM), sucrose (250 mM), EGTA (2 mM), EDTA (5 mM), dithiothreitol (2 mM), benzamidine (10 mM), soybean trypsin inhibitor (100 pg/ml), sodium azide (3 mM), serum albumin (5 mg/ml), pepstatin (14 phf), leupeptin (23 p), and phenylmethylsulfonyl fluoride (1.0 mM). Homogenization employed a glass-Teflon Potter-Elvhjem for liver, or a Polytron (setting number 6 for 10 or 30 s, respectively, for adipose or mammary tissue). Extracts were then subjected to a sequence of centrifugation steps (all at 4 "C): 15 min at 2,300 x g, 30 min at 37,000 x g, and 60 min at 215,000 x g. The final supernatant fraction was loaded onto a DE-52 column (2.6 x 26 cm) which had been pre-equilibrated with "DEAE buffer" (pH 7.5), containing Mops (50 mM), sucrose (250 mM), EDTA (2 mM), EGTA (2 mM), sodium azide (3 mM), pepstatin (3 p ~ ) , leupeptin (5 PM), glycerol (5% w/v), and dithiothreitol (2 m). After removal of unbound protein and further washing with DEAE buffer,

ACC and other proteins were eluted with "avidin buffer" (pH 7.51, which resembled the DEAE buffer except for the addition of Hepes (50 mM, to replace Mops) and NaCl (0.5 M) and the omission of sucrose. Protein eluted from the DEAE column was applied to an affinity column in which avidin had been coupled to Ani-Gel-10, monomerized and pre- loaded with biotin as described (Kohanski and Lane, 1990). The column was washed with the same buffer until protein in the wash fractions was essentially undetectable. ACC was then eluted with the same buffer, which was supplemented with biotin (1 mM) and sodium citrate (10 m ~ ) . Fractions containing the protein peak were concentrated with a Centriprep 30 concentrator (Amicon, Beverly, M A ) or by dialysis against Aquacide I11 (Calbiochem) and then stored in aliquots a t -80 "C.

Assay of ACC Actiuity-The procedures are as initially described (Halestrap and Denton, 1973) and subsequently modified (Brownsey et al., 1979). Samples containing ACC were preincubated for 30 min at 37 "C in the presence of citrate (10 mM), NaCl (150 m ~ ) , and serum albumin (1 mg/ml) and assays were initiated with the addition of 50 1.11 of preincubated enzyme to 450 pl of assay medium (pH 7.2) containing Hepes (50 mM), MgSO, (10 m ~ ) , EDTA (0.5 mM), ATP (5 mM), P-mercap- toethanol (5 m), serum albumin (10 mg/ml), acetyl-coA (0.6 mM), and sodium [14C]bicarbonate (15 m~ and -500-1000 dpdnmol). Reactions were stopped after 2 min at 37 "C with the addition of 200 pl of HCl(5 N). After vortexing and centrifugation (5 min, Eppendorf Microfuge), 500 pl of the acid supernatant was transferred to scintillation bottles and dried prior to liquid scintillation counting (ACS mixture, from Amersham, Oakville, Ont.). 1 milliunit ofACC activity is defined as the amount of enzyme required to catalyze the conversion of 1 nmol of acetyl-CoNmin.

Protein Determination-Protein concentration of samples was rou- tinely determined using the dye-binding assay (Bradford, 1976), using bovine immunoglobulin G as standard. When a detergent such as SDS was present, as in gel-eluted samples, the commercial BCA method (Pierce) was used, with serum albumin as standard.

Biotin Content-The biocytin content of the ACC subunits was assayed in a noncompetitive fashion using avidin and c3H1biotin (Ebrahim and Dakshinamurti, 1987). The procedure was modified by precipitat- ing the [3Hlbiotin-avidin complex with ZnSO, and NaOH, instead of Bentonite (Hood, 1979). ACC subunits were electroeluted from gel chips (Isco electrophoretic concentrator, model 1750, Lincoln, NE) in buffer (pH 11) containing Caps (10 mM) and SDS (0.01%, w/v). ACC protein was precipitated by incubation overnight (-20 "C) following the addition of acetone (4 volumes) and collected by centrifugation (5 min, Ep- pendorf Microfuge). Pellets were suspended in buffer (pH 7.5) containing Hepes (40 mM), NaCl(150 mM), CaCl, (10 mM), heated for 10 min at 100 "C, and then cooled to 25 "C before the addition of Pronase (1 pg/lO pg of ACC). Incubation was continued at 37 "C for 48 h with a second, equal, addition of Pronase after 24 h. The reaction was stopped by heating at 100 "C for 10 min, with subsequent cooling to room tempera- ture and the addition of EDTA (5 mM) and phenylmethylsulfonyl fluoride (1 mM). Biocytin was then assayed as described (Ebrahim and Dakshinamurti, 1987).

Electrophoresis-For optimal separation of the 280- and 265-kDa subunits ofACC, SDS-PAGE was carried out with the discontinuous pH system (Laemmli, 1970) employing a separating gel in which the concentrations of acrylamide and cross-linker were, respectively, 3% (w/v) and 0.66% (w/v). Riboflavin (0.6 pg/ml) and ammonium persulfate (10 mg/ml) were used as catalysts and gels were run at a current of 5 W g e l overnight. Protein bands were futed and stained with Coomassie Blue R-250 or silver (Rabilloud et al., 1988). For amino acid composition or sequence analysis and for subsequent Cleveland digestion or electroelution, gels were stained (20 rnin), without prior furing, in aqueous Coomassie Blue (0.5%, w/v) or Quickstain (Harlow and Lane, 1988).

Western Blotting-Electroblotting onto nitrocellulose or polyvinyli-

of 0.2A in 20 m Tris acetate buffer (pH 7.4) containing EDTA (2 mM) dene fluoride membranes was carried out for 16 h at a constant current

and SDS (O.Ol%, w/v). Membranes were placed in phosphate-buffered sal ineheen and protein bands lightly stained by adding a few drops of Amido Black (0.025%, w/v) in ethano1:acetic acid:water (20:7:73, viviv). Destaining and blocking was performed overnight in phosphate-buffered salinemeen containing serum albumin (2%, w/v) and powdered milk (5%, w/v). Final immunoreactions employed anti-ACC antibodies and goat anti-mouse immunoglobulin conjugated to biotinylated acid phosphatase (OConnor and Ashman, 1982) or streptavidin conjugated with alkaline phosphatase (as per kit from ENZO Biochem, New York).

Amino Acid Composition Analysis-Analysis of the two ACC subunits was performed with an Applied Biosystems (Foster City, CA) model 420 derivatizer/analyzer. For this analysis, ACC subunits were separated by SDS-PAGE and then immobilized onto polyvinylidene

14440 Isozymes of Acetyl-coA Carboxylase fluoride membranes as described above. The same techniques were used when determination of biotin content ofACC subunits was carried out, except that SDS was omitted from the electrophoretic transfer buffer and methanol (lo%, w/v) was added.

Reverse-phase HPLC and Mass Spectrometry of Peptides-Purified rat liver ACC (0.5-1 mg) was first reduced (20 min, 95 "C) in sample buffer (pH 6.8) containing Tris (125 mM), SDS (5%, w/v), dithiothreitol (10 mM), EDTA (6 mM), sucrose (0.3 M), and bromphenol blue (0.002%. w/v); then after cooling to 25 "C, the subunits were carboxyamidom- ethylated (30 min, 25 "C in the dark) in the presence of iodoacetamide (22 mM). The two subunits of ACC were separated by SDS-PAGE, re- covered by electroelution in buffer (pH 8.6) containing Tris acetate (40 mM), EDTA (2 mM), and SDS (0.01%, w/v) and subjected to a second round of SDS-PAGE prior to electrophoretic transfer to Immobilon CD membranes (Patterson et al., 1992). Protein bands visualized with Quickstain were cut into small pieces and incubated at 37 "C in 10 pl of Tris buffer (0.1 M) (pH 8.5) containing NaCl(1 M), urea (0.2 M), and Lys-C protease from Achromobacter (0.1 pg). After 15 h a second addition of protease (0.1 pg) was made and the incubation continued for a further 6 h a t 37 "C. Peptide mixtures obtained by proteolytic cleavage on Immobilon CD were acidified by the addition of 1 pl of trifluoroacetic acid (lo%, v/v). After removal of the primary supernatant, membrane pieces were extracted with a further 10 pl of trifluoroacetic acid (0.1%, v/v) in MeCN (lo%, v/v) and the combined supernatants injected into the HPLC system. Peptides were separated on a 300 A Reliasil C18 column (1 x 150 mm) using an Ultrafast Microprotein Analyzer (Michrom BioResources Inc., Pleasanton, CA). Solvents A and B for the reverse-phase separation contained, respectively, trifluoroacetic acid (0.05 or 0.045%. v/v) plus acetonitrile (2 or 80%, v/v) in water. The applied gradient was from 5 to 50% solvent B in 60 min using a flow rate of 50 pl/min. W absorbance was monitored at 214 nm in a flow cell with a path length of 2 mm and volume of 200 nl. Solution emerging from the flow cell was directed to a low dead-volume flow splitter which directed 90% of the flow to a fraction collector and the remaining solution to the ion source of an API I11 triple quadrupole MS (Sciex, Thornhill, Ont.) equipped with an ion-spray ion source (Bruins et al., 1987). Peptides were ionized in the HPLC effluent a t a flow rate of 5 pl/min subjected to nebulizer gas at 40 p.s.i. and an ion-spray voltage of 5000 V. Analyses were carried out with the MS in single quadrupole mode, using quadrupole number 3, set to scan the mass range 300-2400 Da at a step size of 0.5 Da with a dwell time of 1 mdDa. Full details of the configuration of the MS together with its characterization have been described previously (Hess et al., 1993).

Amino Acid Sequencing-Internal amino acid sequencing proved es- sential since attempts at direct N-terminal sequencing were unsuccessful, suggesting both ACC subunits possessed blocked N termini. Inter- nal sequencing employed the technique described previously (Aebersold et al., 1987; Aebersold and Leavitt, 1990) using trypsin or Lys-C digestion after transfer of subunits to nitrocellulose or to Immobilon CD membranes (Patterson et al., 1992). Fractions from the HPLC column were collected manually according to the W absorbance signal, taking account of the delay (typically 5 s) between the W flow cell and the fraction collector. Peptide fractions were applied to precycled Polybrene- treated glass fiber discs and automated sequencing was carried out using standard protocols with a pulsed-liquid phase sequencer (Applied Biosystems model 477A, Foster City, CA) fitted with a model 120A analyzer.

ACC Phosphorylation and Phosphopeptide Analysis-For phosphorylation by traces of endogenous protein kinases within the ACC preparation, enzyme was diluted into buffer (pH 7.2) containing Hepes (50 m ~ ) , sucrose (0.25 M), EDTA(2 mM), glycerol (lo%, w/v), sodium azide (3 m ~ ) , dithiothreitol(5 mM), and pepstatin plus leupeptin (both 3 pglml). It is important to note that when highly purified, ACC is very suscep- tible to irreversible aggregation, especially a t low salt concentrations. Manipulation of the purified enzyme by direct dilution rather than by dialysis or membrane ultrafiltration was critical for reliable protein recovery. Phosphorylation was carried out by incubation for up to 2 h a t 30 "C in the presence of 0.2 mM [y32PlATP (500-1000 dpdpmol) and MgCl, (10 mM). Reactions were stopped by heating (95 "C for 5 min) with an equal volume of SDS-PAGE sample buffer. For phosphorylation ofACC with CAMP-dependent protein kinase, conditions were as above except for the addition of kinase catalytic subunit (0.5 unitdpg ACC) and microcystin-LR (100 m). After SDS-PAGE, gels were fixed, stained, dried, and subjected to autoradiography, by exposure (-80 "C) to Kodak XAR-5 film in cassettes containing DuPont Hi-Plus Cronex intensifjmg screens. Subunit bands were then excised and either solubilized for quantitation of [32Plphosphate incorporation by scintillation counting (Hames, 1990) or digested with trypsin in preparation for two-dimen-

A OR

280 ,

266"

TD

Li At Ma I

FIG. 1. The major subunits of ACC purified from rat tissues. Left panel (A) illustrates Coomassie Blue-stained proteins following SDS-PAGE analysis of 5 pg of ACC purified from rat liver. Symbols indicate the origin of the separating gel (OR ), the major ACC subunits (280- and 265-kDa, upper and lowerpaired markers), and the tracking dye (TD). The most prominent contaminant (subunit M, -150,000) is most likely pyruvate carboxylase (removable by subsequent size-exclu- sion chromatography). The right panel ( B ) illustrates the relative abundance of the 280- and 265-kDa subunits (upper and lower arrowheads) of ACC purified from rat liver (Li), white adipose tissue (At ), or lactating mammary gland (Ma). Approximately 5 pg of each preparation was subjected to SDS-PAGE and protein bands visualized by silver staining. The concentration of acrylamide in the separating gels differed forpanel A (7%, w/v) and panel B (4%, w/v).

sional peptide mapping as described (Brownsey and Denton, 1982; Mitchell et al., 1989). In other experiments, the first dimension of SDS- PAGE was followed by staining without fEing, protease digestion, and peptide mapping by a second round of SDS-PAGE (Cleveland et al., 1977).

RESULTS

Presence of lltvo Distinct High M, Components in Purified Preparations of Acetyl-coA Carboxylase from Rat Liver-ACC was purified from high-speed (150,000 x g) supernatant fractions from rat liver, white adipose tissue, and lactating mammary gland through DEAE-cellulose and avidin-Sepharose affinity chromatography as described under "Experimental Procedures." In all cases, yields of ACC activity were in the range 3650% based on activity in high-speed supernatant fractions measured after activation by preincubation in the presence of citrate and defatted serum albumin. The final preparations exhibited specific activities of 2 4 unitslmg of protein and subunit profiles as illustrated (Fig. 1) following SDS- PAGE under reducing conditions. In agreement with previous studies, ACC purified from white adipose tissue exhibited a single subunit with molecular mass approximately 265,000 (265 m a ) , whereas an additional band with subunit molecular mass of 280,000 (280 kDa) was detected in preparations from liver and mammary tissue. The relative abundance of the 280- kDa subunit, assessed by densitometric scanning of stained gels, differed in preparations from liver and mammary gland, representing approximately 30 and 5-lo%, respectively, of the total silver-stained ACC protein. We have also observed (data not shown) the 280-kDa subunit of ACC in preparations from heart muscle as described previously (Thampy, 1989).

Although the two high molecular mass subunits of ACC differed by approximately 15 kDa in apparent mass as estimated by electrophoretic mobility during SDS-PAGE, it is possible that such a difference may be explained by rather subtle differences in structure, perhaps brought about by post-transla-

Isozymes of Acetyl-coA Carboxylase 14441

tional modification. Distinct immunological properties of the 280- and 265-kDa subunits of acetyl-coA carboxylase have been noted (Thampy, 1989; Bianchi et al., 1990). We also obtained evidence from development of subunit-specific monoclonal antibodies that the structure of the 280-kDa subunit of rat liver ACC may differ more substantially from that of the major 265-kDa subunit (Winz and Brownsey, 1989). As judged by Western blotting, antibodies designated 5D8 and 1D4 recognized only the 280-kDa subunit of rat liver ACC despite the greater abundance of the 265-kDa subunit (data not shown). Using these antibodies, the 280-kDa subunit of ACC was con- vincingly detected in supernatant fractions from rat liver, heart, and lactating mammary gland. In contrast, very little 280-kDa subunit could be detected in supernatant fractions from brown adipose tissue and none at all in extracts from white adipose tissue, even though equivalent amounts of cell protein were applied to the SDS gels initially and the 265-kDa subunit of ACC could be readily detected.

The 280- and 265-kDa Subunits of Acetyl-coA Carboxylase from Rat Liver Both Contain Biotin-An enzymatically active 280-kDa isoform ofACC has been reported in heart muscle, and although highly related, it was suggested to differ somewhat in immunoreactivity from the equivalent subunit in preparations of ACC from rat liver (Thampy, 1989). It is therefore important to obtain evidence for the catalytic activity of the 280-kDa subunit present in preparations of ACC from rat liver. We have so far been unable to separate the 280- and 265-kDa subunits of rat liver ACC except by PAGE under denaturing conditions, so that direct assessment of the enzyme activity of the isolated 280-kDa subunit has not so far been possible. Since the presence of covalently coupled biotin is necessary for carboxylase activity, evidence was sought for the presence of this prosthetic group. The presence of biotin in each of the separate ACC subunits has been confirmed with three independent techniques. In the first approach, the presence of biotin was demonstrated in a 280-kDa protein present in purified ACC from rat liver and in high-speed supernatant fractions from rat liver and mammary gland by using Western blotting based upon the recognition of biotinylated proteins with streptavidin (data not shown). Surpris- ingly, the more abundant 265-kDa subunit of ACC was not clearly detected by this technique, perhaps because of differential inhibition of the biotin-streptavidin interaction by milk proteins during immunodetection (Hoffman and Jump, 1989).

In the second approach, use was made of the remarkably slow dissociation of avidin from biotin or biotinylated proteins, even during denaturing SDS-PAGE. The two major subunits of ACC both undergo a quantitative shift in apparent molecular mass judged by SDS-PAGE following exposure to avidin, the alteration in subunit size of the protein bands being equivalent to the addition of one avidin subunit (approximately 16 kDa) per ACC subunit (data not shown).

In the third approach, protein-bound biotin was determined directly following isolation of the subunits of ACC by SDS- PAGE (described under “Experimental Procedures”). Two separate preparations of ACC from rat liver were analyzed and the mean biotin contents found to be 0.65 and 0.39 mol of biotidmol of ACC for the 265- and 280-kDa subunits, respectively. The individual values for picomole of protein analyzed (determined by amino acid composition) and picomole of biotin detected were as follows: for the 265-kDa subunit, 209 and 134 pmol of ACC yielded 128 and 96 pmol of biotin; for the 280-kDa subunit, 129 and 114 pmol ofACC yielded 41 and 52 pmol of biotin, respectively.

Structural Relationships between the 280- and 265-kDa Com- ponents of Rat Liver Acetyl-CoA Carboxylase-In one group of studies, ACC affinity-purified from rat liver was subjected to Cleveland mapping, which employed SDS-PAGE to resolve the

two subunits followed by digestion with Staphylococcus aureus V8 protease prior to initiation of the second stage of electrophoresis. A number of differences in the pattern of fragments released from the 280- and 265-kDa ACC subunits, were detected by inspection of silver-stained gels or by autoradiography following digestion of 32P-labeled ACC. Most notably, with the highest concentration of protease employed, the most abundant 32P-labeled fragment derived from the 265-kDa subunit of ACC migrated with an apparent molecular mass of -12,000. In contrast, the digestion of the 280-kDa subunit of ACC yielded significant intermediate fragments with molecular mass values in the 20-30,000 range and the most abundant fragment following most extensive digestion had an apparent mass of -10,000.

On the basis of the immunological properties and the results of peptide analysis described above we set out to further define the differences in structure of the two major subunits of purified rat liver ACC by analyzing the amino acid composition of the intact subunits and by detailed analysis of proteolytic fragments by HPLC/MS and microsequencing of selected peptides. ACC was purified from rat livers through the avidin-Sepharose affinity chromatography step and approximately 1.5 mg of the purified protein subjected to preparative SDS-PAGE to separate the major subunits. After brief staining in aqueous Coo- massie Blue, the protein bands were electrophoretically transferred to polyvinylidene fluoride membranes and the blotted proteins subjected to amino acid analysis. The average values (expressed as mol % of the individual amino acids detected) were obtained from two separate analyses and compared with values reported previously in the literature on the basis of direct analysis as well as the values predicted for the rat mammary 265-kDa subunit based on sequencing of cDNA. Accurate values for tryptophan, methionine, and cysteine were not obtained in our own analysis. Values for abundance of 12 amino acids in the 280-kDa subunit (Glx, Ser, Gly, His, Val, Leu, Phe, Lys, k g , Thr, Ala, and Ileu) differed by less than 5% from the values previously analyzed or deduced for the 265-kDa subunit. The most marked differences in abundance of amino acids (as mol %) in the 280- and 265-kDa subunits, respectively, were: proline (5.2 uersus 4.6), tyrosine (2.2 versus 2.71, and aspartate plus asparagine (9.4 versus 10.7).

For further comparison of the structures of the 265- and 280-kDa subunits of ACC, subunits were isolated by SDS- PAGE, transferred to Immobilon CD membranes, subjected to proteolytic cleavage (endoproteinase Lys-C), and the peptides separated by HPLC prior to MS analysis. The profiles of peptides eluted from the HPLC column were complex as anticipated for such large polypeptides and differed substantially for the two different ACC subunits, on the basis of U V absorption (not shown) and by mass analysis (Fig. 2). For example, the differences between the subunits is readily observed with the profile of the total ion current. Thus major peaks derived from the 280-kDa subunit of ACC eluted at 28.9, 29.8, 30.3, 34.7, 42.9, and 46.2 min into the gradient but were only weakly represented in the profile of the 265-kDa subunit. Conversely, major peaks eluting a t 33.0, 42.9, and 46.2 min were far more abundant in the analysis of the 265-kDa subunit.

Analysis of the peptide masses within each of the peaks following HPLC has demonstrated substantial divergence in the structures of the 265- and 280-kDa subunits of ACC. The data presented (Table I) includes only the more abundant peptide masses for which at least two charge states were apparent, with a minimum ion count of 10,000 for the more abundant ion. According to these criteria 44 peptide masses were derived from the 265-kDa subunit of ACC and 36 of these matched exactly (within 1.0 Da) with peptide sequences predicted on the basis of Lys-C digestion of the deduced sequence of the 265-kDa


14 A

‘ 1 B

30.0 40.0 50.0

*

0 ’ 30.0 40.0 50.0

Time (min)

FIG. 2. HPLCNS analysis of peptides derived from the 265- and 280-kDa subunits of rat liver ACC. Individual subunits ofACC were isolated by SDS-PAGE, transferred to Immobilon-CD membranes, and subjected to digestion with Lys-C protease. Proteolytic digests were subjected to reverse-phase HPLC and 10% of the column effluent analyzed by on-line mass spectrometry. The figure illustrates the total ion current of the HPLC chromatogram obtained following digestion of the 280-kDa subunit (A) or 265-kDa subunit ( B ) . Vertical axis represents the ion current of all ions between 500 and 2400 Da in relation to time of elution from the HPLC column (min). The peaks which contained the

highlighted (*). one peptide matching exactly in mass and amino acid sequence are

subunit of rat ACC. Together, these peptide masses potentially span 420 amino acids of the full-length ACC (2345 residues) deduced from cDNA sequencing and suggest confirmation of 18% of the deduced sequence of the complete polypeptide. Most peptide masses were apparent in only one HPLC fraction, although in five cases, essentially identical masses were detected in consecutive HPLC fractions and were therefore assigned to the fractions in which they were most abundant (masses of 1181.6, 1767.8, 1958.3, 2120.8, and 2541.0 Da). One species (mass of 758.2) could not be assigned unambiguously since two possible peptide sequences (residues 86-91 or 1781-1787) within the ACC 265-kDa subunit would account for this mass. Of the 43 other major peptides detected, 35 could be assigned to unique sequences identifiable within the deduced sequence of 265-kDa subunit of ACC (Table I). Eight peptides detected in the Lys-C digest of the 265-kDa subunit of ACC have masses which fail to match with any of the deduced peptide masses. These differences may be accounted for by nonspecific cleavage, peptide modification (either post-translational or during peptide isolation), or by true differences in amino acid sequence.

Analysis of peptides derived from the 280-kDa subunit of ACC revealed 41 major peptide masses using the same criteria as for the 265-kDa subunit peptides. Only one of these peptides could be matched with a corresponding peptide from the 265- kDa subunit on the basis of both identity of mass (+1 Da) and also a close correspondence in time of elution during HPLC

TABLE I Analysis of the masses of peptides derived from the 265- and

280-kDa subunits of rat liver ACC Individual subunits of ACC were isolated by SDS-PAGE, transferred

to Immobilon-CD membranes and subjected to digestion with Lys-C protease. Proteolytic digests were subjected to RP-HPLC and the column effluent analyzed by electrospray mass spectrometry. The table illustrates the mass (in Da) and the elution time (min) from HPLC, of the major peptides identified according to the criteria described in de- tail in the text. The first two columns, to the left, describe values for peptides derived from the 265-kDa subunit. The two central columns indicate the position and mass of peptides within the sequence, deduced from cDNA studies, for rat ACC (265 kDa), which match the observed values (within 1 Da). Crosses (x) indicate a lack of detectable match between observed and deduced sequences within the 265-kDa subunit of ACC (using the MacBioSpec program). The values for peptides from the 280-kDa subunit (right two columns) are presented so as to achieve optimum alignment with corresponding peptide masses from the 265- kDa subunit. The one peptide from both ACC subunits found to be identical in mass, elution time, and sequence is highlighted in bold, with underlining. Marginal arrowheads indicate peptides subjected to confirmatory amino acid sequence analysis. One peptide (mass 758.2, from the 265-kDa subunit) may correspond to one of two possible deduced sequences but all other assignments were unique.

265-kDa peptide Observed

265-kDa peptide Deduced

280-kDa peptide Observed

~

Mass Time Sequence Mass

477.2 564.8 660.8 676.1 686.6 701.9 707.3 711.3 722.4 745.9 758.2

760.4 782.7 806.9 824.4 930.7 947.2

1008.2 1009.6 1069.0 1155.6 1181.6 1187.6 1265.4 1293.4 1302.4 1305.9 1320.6 1321.6 1423.5 1474.0 1541.5 1591.1 1712.1

F1834.2 1767.8

30.7 22.4 40.9 28.0 33.9 33.0 44.5 45.3 31.0 33.3 35.4

30.7 33.0 31.5 45.3 32.2 32.2 35.7 37.6 39.3 36.9 38.7 43.2 37.6 37.6 41.7 40.9 38.1 40.9 37.6 32.0 49.4 36.9 47.6 44.8 48.5

1663-1666 1334-1337 311-318

X X

1580-1585 X X

2025-2030 3 9 4 4

1781-1787 86-91

292-298 741-746 163-169 217-223 817-824

X 2260-2267 2054-2061 1505-1513 2293-2301

1354-1362 120-131

1008-1017 1904-1915 1749-1758 299-310

X X

803-816 747-760

1916-1928

1279-1293 728-740

1564-1579 2009-2024

Mass Time

477.3 564.3 661.3

702.4

722.4 745.4 758.4 758.5 760.4 782.5 807.4 824.6 931.4

1008.5 1010.5 1068.7 1155.6 1181.7 1186.6 1264.7 1293.7 1302.7 1305.7

1423.8 1473.7 1541.8 1591.8 1713.0

1834.0 1768.0

716.1 758.4 782.1 827.9 871.4 871.4 903.4 915.2

918.5 919.5

1002.5

1054.0 1256.4 1352.7 1372.9 1374.7 1387.9 1425.2 1470.0 1486.7 1522.8 1600.3 1699.0 1757.6 1774.2 1774.4 1833.7

25.4 38.2 29.0 36.04 36.0 38.2 47.14 30.3

32.4 43.5 37.2

34.7 30.3 39.34 30.3 39.3 29.0 34.74 29.84 29.0 35.44 44.2 36.0 41.24 37.2

47.14 41.2

1958.3 1991.0

2059.0 2120.8

2489.0 2541.0

2964.2

3271.6

42.9 52.1

42.3 44.8

40.3 49.8

44.5

49.8

2276-2292 224-242

1546-1563 1929-1946

825-847 93-115

X

10-38

1958.0 1833.9 1990.0 1855.4

2024.7 2059.0 2041.0 2120.1 2057.3

2087.3 2490.3 2102.8 2541.3 2553.0

2808.1 2830.4 2966.0

3272.6 2980.0 2987.7

48.7 47.1 38.6 38.6 43.5 46.2 46.2 43.5 46.2 46.2 42.9 42.9 42.9


[M+2H12+

A loo I

918.0

75 .

50 '

77500

" 500 1000 1500 2000

d Z

38500

wr 1835.4

I " 500 1000 1500 2000

d z

FIG. 3. Mass spectra of matching peptides present in both 280- and 265-kDa subunits of rat liver ACC. Peptides were generated

HPLC/MS analysis as described in the legend to Fig. 2. Illustrated are from the individual subunits of purified rat liver ACC and subjected to

the individual mass spectra of peptide peaks derived from the 280-kDa subunit of ACC (panel A, peak with elution time of 47.1 min) and the 265-kDa subunit (panel B, peak with elution time of 48.5 min). Note the double (918) and single (1835) charge state of the peptide species which match exactly with residues 2009-2024 of the deduced sequence of the 265-kDa subunit ofACC as confirmed by direct sequencing. The values indicated at the top right of each spectrum indicate the total ion count of the major (double charge) species.

(Table I, bold and underlined values). The individual mass spectra of the HPLC peaks which contained these matching peptides (masses of 1834.2 and 1833.7 for the 265- and 280-kDa subunits, respectively) are illustrated (see Fig. 3) and the identity of these peptides was further confirmed by direct sequencing (see below). Although other matching masses may be dis- cerned, these represent peptides which eluted from the HPLC column with substantially different elution times (differing by 2 min or more) and therefore most likely represent different species. Most apparent in this regard were peptide masses of 758.2 (or 758.4) and 782.7 (or 782.1) derived from the 265-kDa (or 280 kDa) subunit ofACC. In eight cases, sequence analysis confirmed values observed by MS analysis (Table I, values indicated with marginal arrows). This analysis therefore con- firms that a large proportion of the major proteolytic peptides derived from the 280-kDa subunit of ACC are structurally distinct from fragments derived from the 265-kDa subunit.

Peptide Sequence Analysis-Attempts to determine N-terminal sequences were unsuccessful for both the 280- and 265-kDa subunits suggesting that the N termini were blocked in both cases, as previously observed for the 265-kDa subunit of rat and chicken ACC (Bai et al., 1986; Takai et al., 1988). Further studies were therefore initiated in order to obtain sequences of internal peptides. For the determination of internal sequences, individual subunits of ACC were isolated, electroblotted, and digested with trypsin. Peptides were isolated by reversed- phase HPLC and selected major peaks subjected to microsequencing. Following digestion of the isolated 265-kDa subunit, four major peptides which were isolated and sequenced

matched exactly with deduced sequences within the 265-kDa subunit of rat mammary ACC. The actual sequences determined were as follows (in single letter code, where Xrepresents a residue which could not be unambiguously ascribed): 1) XXXIPVGWAWT, corresponding to the tryptic peptide LGGI- PVGWAWTR at positions 1977-1990; 2)XNNADDFPNL, corresponding to VNNADDFPNLFR a t 323-334; 3) FTVAS- PAEFVT, corresponding to DFTVASPAEFVTR at 98-110; 4) XLEPALAF, corresponding to HLEPALAFQLELNR at 1371- 1384.

So far 12 of the more abundant peptides derived from the 280-kDa subunit of ACC have been sequenced. Of these peptides five represent sequences spanning a total of 78 amino acids which show no obvious homology to sequences within the 265-kDa ACC isoform. The remaining seven peptides, spanning a total of 91 amino acids, contain sequences which are distinct from but related to sequences within the 265-kDa subunit, possessing amino acid identities a t 64 of the 91 residues (70% identity overall). The number of identities in individual peptides were as follows (with the full-length of the peptides in parentheses) 10(11), 8(12), 7(12), 12(13), 13(15), 9(14), and 5(8).

Only one of the peptides derived from the 280-kDa subunit of ACC has revealed a perfect match with a corresponding sequence from the 265-kDa form of rat mammary ACC. The matching peptide was the one peptide identified as exactly matching in the Lys-C digests from both 265- and 280-kDa subunits on the basis of mass and HPLC co-elution. The co- eluting peptides displayed masses of 1834.2 and 1833.7 and identical sequences: IIQQAGQVWFPDSAFK, representing position 2009-2024 of the 265-kDa ACC. Although a limited number of the ACC peptides have so far been subjected to sequence analysis, those analyzed have confirmed structures which match exactly with masses determined following HPLCMS. Taken together with the other evidence published previously and new evidence described above, this analysis indicates that the 280- and 265-kDa polypeptides present in purified preparations of ACC from rat liver are related but distinct proteins.

Differential Phosphorylation in Vitro of the 280- and 265-kDa Components of Purified Rat Liver Acetyl-coA Carboxylase-In view of the established significance of reversible phosphorylation in the regulation of ACC, it was of interest to investigate the phosphorylation of the individual 280- and 265-kDa polypeptides. In initial experiments, ACC was phosphorylated by incubation in the presence of [y3"?P]ATP (see "Experimental Procedures"). Phosphorylation by endogenous protein kinases present in the purified preparations of ACC led to apparently equivalent incorporation of [32Plphosphate into each of the major subunits, so that autoradiograms or direct scintillation counting of gel slices gave a ratio of incorporation which reflected the relative abundance of the polypeptides determined by scanning stained gels. In contrast, a markedly different distribution of [32Plphosphate between the two major subunits of ACC was observed after incubation in the presence of the catalytic subunit of cyclic AMP-dependent protein kinase (Fig. 4). In this case, the phosphorylation of the 280-kDa subunit of ACC proceeded more rapidly and to a higher stoichiometry than did the phosphorylation of the corresponding 265-kDa subunit. Thus the phosphorylation of the 280-kDa subunit reached an apparent plateau of 1.8 mol of [P]/subunit within 30-60 min, rising slowly to 2.0 mol of [P]/subunit at 2 h. The 280-kDa subunit reached half-maximal phosphorylation within 2 min. In contrast, the phosphorylation of the 265-kDa subunit reached 0.5 mol of [P]/subunit within 30 min and continued to rise through 1.0 mol of [P]/subunit through 2 h. For comparison, parallel studies were carried out using ACC purified from rat white adipose tissue, which contains exclusively the 265- kDa subunit. The phosphorylation of the white adipose ACC

14444 Isozymes of Acetyl-coA Carboxylase

A t 2 3 4 5 6 7 8 9

Incubation Time (mid

FIG. 4. Different rates of phosphorylation of the 280- and 265- kDa subunits ofACC with catalytic subunit of cAMF"dependent protein kinase. ACC (10 pg) purified from liver or white (epididymal) adipose tissue was incubated with [3zP]ATP in the presence of free catalytic subunit of cAMP-dependent protein kinase (5 units) for the indicated times a t 30 "C. Following SDS-PAGE, incorporation of [3'Plphosphate into each subunit was measured by excision of bands, digestion with hydrogen peroxide, and liquid scintillation counting. The upper panel (A) illustrates an autoradiogram obtained following phosphorylation of purified rat liver ACC for 1, 2, 5, 10, 30, 60, 120, or 180

ACC incubated (180 min) in the absence of added kinase. The lower min (lanes 1-8, respectively). Lane 9, at the extreme right, represents

panel ( B ) illustrates the time course for incorporation of ["Plphosphate into the 280-kDa (W) and 265-kDa (0) subunits of rat liver ACC or into the single subunit of adipose tissue ACC (A), following incubation for the indicated times in the presence of the catalytic subunit of cyclic AMP-dependent protein kinase. These results were obtained following SDS-PAGE, band excision, and counting and are typical of at least three independent experiments with different preparations of ACC.

proceeded with kinetics even slower than either of the two liver ACC subunits. In view of the difference in the concentrations of the two subunits in preparations of rat liver ACC, the 280-kDa subunit appears to be a highly preferred substrate for the cyclic AMP-dependent protein kinase.

In addition to the estimation of overall phosphorylation of intact ACC subunits, the distribution of [32Plphosphate among the different potential phosphorylation sites was determined by two-dimensional thin layer mapping of tryptic peptides derived from the separate polypeptides. Digestion of the 265-kDa subunit released one major and several minor phosphopeptides, giving a phosphopeptide pattern closely resembling that observed previously with isolated adipose tissue ACC phosphorylated in vitro (with cyclic AMP-dependent protein kinase) or within intact fat cells stimulated with adrenaline (Brownsey and Denton, 1982, 1987). The phosphopeptide map observed following digestion of the 280-kDa subunit with trypsin differed markedly from that of the 265-kDa subunit. The major phosphopeptide migrated differently in both electrophoretic and chromatographic dimensions and the minor phosphopeptides also showed clear differences (data not shown). The two major subunits of rat liver ACC are therefore distinguished kinetically by cyclic AMP-dependent protein kinase and the major phosphorylation sites on the two subunits are structurally distinct.

DISCUSSION

The studies reported above have investigated the structural relationships between the two major biotinylated components

of rat liver ACC. Although direct demonstration of the catalytic activity of the 280-kDa subunit of ACC from rat liver has not proved possible (the 265- and 280-kDa subunits have been separated only under denaturing conditions), several lines of evidence confirm its identification as an enzyme isoform. First, the 265- and 280-kDa subunits both contain covalently bound biotin. Second, in confirmation of previous reports (Thampy, 1989; Bianchi et al., 1990), the rat liver 280-kDa subunit is immunologically distinct from the 265-kDa subunit while antibodies directed against the 280-kDa subunit do cross-react with the cardiac ACC subunit of 280 kDa which has demonstrated enzyme activity. Third, the inability to separate the 265- and 280-kDa subunits except under denaturing SDS- PAGE, suggests they associate very tightly, perhaps by forming heterodimers. We have obtained further evidence for the for- mation of 265:280 subunit heterodimers by co-immunoprecipi- tation of approximately 1:l complexes with anti-280 monoclonal antibodies 1D4 and 5D8 as well as co-migration during nondenaturing gel electrophoresis (data not shown). Finally, microsequence analysis of peptides derived from the 280-kDa subunit demonstrates at least one peptide fragment (16 residues in length) which is identical and a further seven peptides which are related to sequences within the published sequence of the 265-kDa subunit. Taken together, these studies confirm that the 280-kDa subunit of rat liver ACC may be considered a true enzyme isoform.

A striking conclusion from the studies reported here is the extent of divergence between the structures of the two subunits of rat liver ACC. At the outset, it was considered that the two high M, subunits may be derived by alternate splicing of mRNA. Several discrete species of ACC mRNA as well as one splice variant of the 265-kDa ACC protein have already been demonstrated (Luo et al., 1989; Lopez-Casillas and Kim, 1991; Kong et al., 1990). The detailed analysis described here has, however, indicated that the differences in primary sequences between the two subunits are greater than anticipated and indeed far greater than those which might be explained by mRNA splicing variants described so far. Thus whereas four major peptides derived from the 265-kDa subunit of rat liver ACC yielded sequences readily identified within the published sequence of mammalian ACC, only one peptide identity has been so far detected among the proteolytic digests of the 280- kDa subunit. A further 12 peptides from the 280-kDa subunit are either related but distinct structures or are apparently entirely distinct from sequences within the 265-kDa subunit. We do not yet have any means of locating the possible positions of the unique sequences within the overall structure of the ACC 280-kDa subunits and so it must remain possible that alternative splicing of mRNA may, in part, yet explain the differences detected here. However, we feel this is very unlikely when also considering the extent of divergences detected by MS analysis of a large proportion of all the peptides derived from the two separate subunits and the possibility that the 280-kDa subunit of ACC is encoded within a separate gene must also be considered. To this end, further peptide sequencing as well as molecular cloning is under way to attempt to identify longer sequences of the 280-kDa subunit.

The use of HPLC coupled with ion-spray mass spectrometry for the analysis of the peptides derived from the subunits of ACC has provided a powerful approach for the comparative analysis of the 280- and 265-kDa isoforms. The application of this technique has provided convincing evidence that the 265- kDa subunit of the preparations of rat liver ACC does indeed correspond to the polypeptide for which the sequence has been deduced from studies of cDNA, since the masses of more than 40 detected peptides match exactly with masses which are predicted for the proteolytic fragments. Together, the fragments so


far analyzed may account for approximately 20% of the intact polypeptides and this could be extended with similar ap- proaches by using different proteinases. The analysis of the masses of peptides is therefore feasible even though they may be generated from very large polypeptides such as the ACC subunits studied here and the validity of the analysis has been supported by parallel sequence analysis. These studies therefore support the application of MS analysis to aid in confirmation of protein structure (here, the ACC 265-kDa subunit), as well as comparative analysis where an unknown structure (the 280-kDa subunit) may be matched with that of an appropriate reference sample of defined structure.

At present, the precise function of the 280-kDa subunit of ACC can only be speculated. For example, its abundance in rat heart muscle suggests it might play a special role in providing malonyl-CoA for the regulation of fatty acid oxidation and strong supporting evidence has been presented recently (Awan and Saggerson, 1993; Saddik et al., 1993). On the contrary, there is no apparent increase in relative abundance of the 280- kDa compared with the 265-kDa subunit in the regions of liver reportedly adapted to fatty acid oxidation by “zonation” (Evans et al., 1990). I t is conceivable, perhaps, that the subunits may distribute selectively among different compartments within individual cells. Mitochondrial forms of ACC have been reported (Roman-Lopez et al., 19891, with apparent enrichment of the 280-kDa subunit (Allred et al., 1989). However, using the extraction conditions described in the studies reported here, we find little evidence for membrane association of the 280-kDa subunit of rat liver ACC. An alternative role for the 280-kDa subunit of ACC, on the basis of the evidence presented above, may be to exert an important regulatory influence over polymer dimer equilibrium. For example, the 280-kDa subunit may be positioned strategically in rat liver ACC (approximately every fourth subunit, or one subunit in each alternate dimer) to be able to facilitate regulation of polymerization and depolymer- izatrion. In this regard, the rapid rate of phosphorylation of the 280-kDa subunit may offer a means for rapid initiation of de- polymerization upon activation of CAMP-dependent protein kinase (which phosphorylates the 265-kDa subunit relatively slowly in vitro and perhaps not at all within intact cells). The differences in rates of subunit phosphorylation, as well as the distinct phosphopeptide patterns may suggest that entirely different sites exist on the 280-kDa subunit or perhaps that the peptides containing Ser-77 and Ser-1200 of the 265-kDa are modified so as to conform more exactly with ideal concensus sequences for phosphorylation by cyclic AMP-dependent protein kinase. Whatever the role in liver and cardiac muscle, it is intriguing that the 280-kDa subunit is not apparently expressed in adipose tissue, suggesting that the important sites of lipogenesis might exhibit distinctive control mechanisms. Clearly, many critical details of the structure and function of the subunits of this important regulatory protein remain to be defined.

Acknowledgments-We thank Sandy Kielland and the staff of the Regional Facility at the University of Victoria, British Columbia for amino acid analysis, and Edward Bures of the Biomedical Research

Centre of the University of British Columbia for protein sequence analysis. We are also most grateful to The Canadian Lysozyme Co. for the generous donation of avidin.

REFERENCES

Aebersold, R. & Leavitt, J . (1990) Electrophoresis 11, 517-527 Aebersold, R., Leavitt, J., Saavedra, R. A., Hood, L. E. & Kent, S. B. 8. (1987) Proc.

Al-Feel, W., Chirala, S. S. & Wakil, S. J. (1992) Proc. Natl. Acad. Sci. U. S. A. 89,

Allred, J. B., Roman-Lopez, C. R., Jurin, R. R. & McCune, S. A. (1989) J . Nutr 119,

Awan, M. M. & Saggerson, E. D. (1993) Biochem. J. 295, 61-66 Bai, D. H., Pape, M. E., Lopez-Casillas, F., Luo, X. C., Dixon, J. E. & Kim, K-H.

Bianchi, A,, Evans, J . L., Iverson, A. J., Nordlund, A-C., Watts, T. D. & Witters, L.

Natl. Acad. Sci. U. S. A. 84, 6970-6974

4534-4538

478-483

(1986) J. Bid. Chem. 261,12395-12399

Borthwick, A. C., Edgell, N. J. & Denton, R. M. (1987) Biochem. J . 241,773-782 A. (1990) J. Biol. Chem. 265, 1502-1509

Bradford, M. M. (1976) Anal. Biochem. 72,248-254 Brownsey, R. W. & Denton, R. M. (1982) Biochem. J . 202,7746 Brownsey, R. W. & Denton, R. M. (1987) in The Enzymes: VolumeXVlZZ, Control by

Academic Press, Orlando, FL Phosphorylation, Part B (Boyer, P. D., and Krebs, E. G., eds) pp. 123-146,

Brownsey, R. W., Hughes, W. A. & Denton, R. M. (1979) Biochem. J. 184,23-32 Bruins, A. P., Covey, T. R. & Henion, J. D. (1987)AnaL Chem. 59,2642-2647 Clarke, S. D., Watkins, P. A. & Lane, M. D. (1979) J. Lipid Res. 20,974-985 Cleveland, D. W., Fischer, S. G., Kirschner, M. W. & Laemmli, U. K. (1977) J. Biol.

Ebrahim, H. & Dakshinamurti, K. (1987) Anal. Biochem. 162,319-324 Evans, J. L., Quistorf€, B. & Witters, L. A. (1990) Biochem. J . 270, 665-672 Gregolin, C., Ryder, E., Warner, R. C. , Kleinschmidt, A. K. & Lane, M. D. (1966)

Halestrap, A. P. & Denton, R. M. (1973) Biochem. J . 132, 509-517

Hardie, D. G. (1989) Prog. Lipid Res. 28, 117-146 Hames, B. D. (1990) Gel Electrophoresis of Proteins, IRL Press Ltd., London

Harlow, E. & Lane, M. D. (1988) in Antibodies: A Laboratory Manual (Harlow, E., and Lane, M. D., eds) p. 63, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY

Hess, D., Covey, T. R., Winz, R., Brownsey, R. W. & Aebersold, R. (1993) Protein Sci. 2,1342-1351

Hoffman, W. L. & Jump, A. A. (1989) Anal. Biochem. 181, 318-320 Hood, R. L. (1979) Methods Enzymol. 62, 279-283 Kim, K.-H., Lopez-Casillas, F., Bai, D. H., Luo, X. & Pape, M. E. (1989) FASEB J.

Kohanski, R. A. & Lane, M. D. (1990) Methods Enzymol. 184, 194-200

Krebs, H. A. & Henseleit, K. (1932) Hoppe-Seyler’s Z. Physiol. Chem. 210,33-36 Kong, I-S., Lopez-Casillas, F. & Kim, K-H. (1990) J. Biol. Chem. 266,13695-13701

Laemmli, U. K. (1970) Nature 227, 680-685 Lane, M. D., Moss, J. & Polakis, S. E. (1974) Curr Top. Cell. Regul. 8, 139-195 Lopez-Casillas, F. & Kim, K.-H. (1991) Eur. J. Biochem. 201, 119-127 Luo, X., Park, K., Lopez-Casillas, F. & Kim, K-H. (1989) Proc. Natl. Acad. Sci.

McGarry, J. D., Takabayashi, Y. & Foster, D. W. (1978) J. Bid. Chem. 253, 8294-

Mitchell, F. E., Marais, R. M. & Parker, P. J. (1989) Biochem. J. 261, 131-136 Numa, S. & Tanabe, T. (1984) in Fatty Acid Metabolism and Its Regulation (Numa,

Numa, S. & Yamashita, S. (1974) Curr Top. Cell. Regul. 8, 197-246

Patterson, S. D., Hess, D., Yungwirth, T. & Aebersold, R. (1992) Anal. Biochem. OConnor, C. G. & Ashman, C. K. (1982) J. Zmmunol. Methods 54, 267-271

Quayle, K. A., Denton, R. M. & Brownsey, R. W. (1993) Biochem. J . 292,7544 Rabilloud, T., Carpentier, G. & Tarroux, P. (1988) Electrophoresis 9, 286-289 Roman-Lopez, C. R., Shriver, B. J., Joseph, C. R. & Allred, J. B. (1989) Biochem. J.

Saddik, M., Gamble, J., Witters, L. A. & Lopaschuk, G. D. (1993) J. Biol. Chem.

Song, C. S. & Kim, K-H. (1981) J. Biol. Chem. 256,778G7788 Takai, T., Yokoyama, C., Wada, K. & Tanabe, T. (1988) J. Biol. Chem. 263,2651-

Thampy, K G. (1989) J. Biol. Chen. 264,17631-17634 Volpe, J . J. & Vagelos, P. R. (1976) Physiol. Reu. 56, 339-417 Winz, R. & Brownsey, R.W. (1989) Proceedings ofthe 32nd Annual Meeting of the

Canadian Federation of Biological Societies, Calgary, Alberta, Canada, June

Abstr. 362 1417, 1989, Canadian Federation of Biological Societies, Ottawa, Ontario,

Chem. 252,1102-1106

Proc. Natl. Acad. Sci. U. S . A. 56, 1751-1758

3,2250-2256

U. S. A. 86, 4042-4046

8300

S., ed) pp. 1-27, Elsevier Science, Amsterdam

202,193-203

260,927-930

268,25836-25845

2657

of Rat Liver Acetyl-coA Car

Documents