Biochem. J. (1969) 111, 225 225 Printed in Great Britain

Conditions for the Self-Catalysed Inactivation of Acetyltransferase A NOVEL FORM OF ENZYME INHIBITION

By J. F. A. CHASE AND P. K. TUBBS Department of Biochemi8try, University of (Jambridge (Received 2 Augw8t 1968)

1. Carnitine acetyltransferase is very rapidly inhibited in the presence of bromoacetyl-(-)-carnitine plus CoA or of bromoacetyl-CoA plus (-)-carnitine. 2. Under appropriate conditions, the enzyme may be titrated with either bromo- acetyl substrate analogue; in each case about lmole of inhibitor is required to inactivate completely 1 mole of enzyme of molecular weight 58 000 + 3000. 3. Inhibition by bromoacetyl-CoA plus (-)-carnitine results in the formation of an inactive enzyme species, containing stoicheiometric amounts of bound adenine and (-)-carnitine in a form that is not removed by gel filtration. This is shown to be S-carboxymethyl-CoA (-)-carnitine . 4. The inhibited enzyme recovers activity slowly on prolonged standing at 4°. 5. Incubation with S-car- boxymethyl-CoA (-)-carnitine ester causes a slow inhibition of carnitine acetyl- . 6. The formation of bound S-carboxymethyl-CoA (-)-carnitine ester by the enzyme is discussed. Presumably the resulting inhibition reflects binding of the ester to both the CoA- and carnitine-binding sites on the enzyme and its consequent very slow dissociation. These observations confirm that carnitine acetyltransferase can form ternary enzyme-substrate complexes; this also appears to be the case with carnitine palmitoyltransferase and choline acetyltransferase.

The enzyme carnitine acetyltransferase [acetyl- Preliminary reports of part of this work have CoA-(-)-carnitine O-acetyltransferase, EC 2.3.1.7] been published (Chase & Tubbs, 1966b, 1967). is widespread and active in animal tissues (Beenakkers & Klingenberg, 1964), but as yet no MATERIALS precise physiological role can be ascribed to it (Fritz, 1965). It might be expected, however, that Enzymes. Pigeon breast muscle carnitine acetyltransferase a study of the metabolic disturbances caused by was obtained from Boehringer Corp. (London) Ltd. (Lon- specific inhibitors of the enzyme would yield in- don, W. 5) and recrystallized twice to a specific activity of formation relevant to this problem. 120-130 units/mg. (Chase, Pearson & Tubbs, 1965). The effects of bromoacetyl derivatives of CoA Citrate synthase (EC 4.1.3.7) was prepared from pig and carnitine on isolated preparations of carnitine heart by the method of Srere & Kosicki (1961). Chemical&. CoA was bought from Boehringer Corp. acetyltransferase have been investigated. It was (London) Ltd.; desulpho-CoA was prepared from it by the hoped that such substrate analogues, after being method of Chase, Middleton & Tubbs (1966). Chloroacetyl specifically bound to the enzyme, would alkylate chloride, bromoacetyl bromide, bromoacetic acid, thio- susceptible amino acid residues at or near the active phenol and ,B-alanine were the reagent grade of British centre. Similar active-site-directed inhibitors have Drug Houses Ltd. (Poole, Dorset). S-Carboxymethyl-2- been designed for a number of enzymes (Baker, mercaptoethylamine was prepared from 2-mercaptoethyl- 1967). Although it appears that bromoacetyl-(-)- amine hydrochloride (Koch-Light Laboratories Ltd., carnitine at least may be used to inhibit the enzyme Colnbrook, Bucks.) by the method of De Marco, Riva & and irreversibly and probably in the manner envisaged Dupre (1964) and recrystallized twice. (+)-, (-)- it (±)-Carnitine hydrochlorides were obtained from Koch- above, has also become clear that under certain Light Laboratories Ltd.; samples of the two optically active can cause a more conditions either analogue much isomers from British Drug Houses Ltd. were also used. rapid inhibition, which is slowly reversible. It is (+±)-[Me-14C]-Carnitine hydrochloride, nominally 11-65/Lc/ the somewhat novel nature of this rapid inhibition ,umole, was a gift from Dr P. B. Garland. Before use it that is reported here. was chromatographed on a column of Zeo-Karb 225 8 Bioch. 1969, 111 226 J. F. A. CHASE AND P. K. TUBBS 1969 (SRC 16) cation-exchange resin (British Drug Houses Ltd.) Preparation and assay of bromoacetylcarnitine. The by the procedure of Friedman, McFarlane, Bhattacharyya bromoacetyl derivatives of (+)-, (-) and (±)-carnitine & Fraenkel (1960). An emergent peak with 95% of the were prepared by treating the appropriate carnitine hydro- radioactivity of the starting material proved to contain all chloride, dissolved in bromoacetic acid, with bromoacetyl the carnitine. This material was pooled and stored frozen. bromide. The procedure was otherwise identical with that 5,-5'-Dithiobis-(2-nitrobenzoic acid) was obtained from described by Fraenkel & Friedman (1957) for acetyl- the Aldrich Chemical Co. Inc. (Milwaukee, Wis., U.S.A.). carnitine. Bromoacetylcarnitine was recrystallized twice Tris (Trizma base; Sigma Chemical Co., St Louis, Mo., from acetic acid-acetone (Chase & Tubbs, 1966a) and stored U.S.A.) was neutralized with HCI. buffers were in a desiccator away from light. prepared from KH2PO4 and KOH. No specific assay for bromoacetylcarnitine is available, Other reagents were of analytical grade and glass- as this compound strongly inhibits carnitine acetyltrans- distilled water was used throughout. ferase. Two indirect methods have therefore been used. In the first, bromoacetylcarnitine was hydrolysed by incu- bation at room temperature for 10min. in the presence of a METHODS 0-1 M concentration excess of KOH. After neutralization Preparation of chloroacetyl-CoA and bromoacetyl-CoA. In the (-)-carnitine released was determined enzymically the early stages ofthis work these compounds were prepared (Pearson & Tubbs, 1964). This procedure can only be used by treating CoA solutions with the appropriate haloacetyl with derivatives of (-)- or (±)-carnitine. Bromoacetyl halide (Seubert, 1960). However, as this procedure gives of all carnitine enantiomers were also estimated rather poor yield (20-30%) of impure haloacetyl-CoA, a chemically by a hydroxamate method (Friedman & milder technique has been developed based on transhalo- Fraenkel, 1955). The first method measures both free and acetylation between thiophenol and CoA. alkali-labile carnitine in the sample, whereas the second (i) Preparation of S-bromoacetylthiophenol. A 2-Oml. measures all esters of carnitine, so that the reported con- portion (19-2m-moles) of thiophenol was added to 2-4ml. centrations of bromoacetylcarnitine are maximum values. (26-4m-moles) of freshly redistilled bromoacetyl bromide Sub8trate8 of carnitine acetyltransferase. Solutions of in a Thunberg tube. HBr was removed as formed by means CoA, acetyl-CoA, (-)-carnitine and acetyl-(-)-carnitine of a water pump. After about 60min. at room temperature were prepared, stored and assayed as described by Chase no further gas was evolved and saturated KHCO3 solution & Tubbs (1966a). was added to bring the pH of the mixture to about 8. This Measurement8 of carnitine acetyltran8ferase activity. precipitated bromoacetylthiophenol as colourless crystals. Except where specific conditions are mentioned, carnitine The mixture was extracted twice with ether, which dissolved acetyltransferase activities were compared by adding the crystals, and the aqueous supernatant was washed enzyme samples to a system containing tris-HCl buffer, twice with this solvent. The ethereal solution was washed pH7-8 (100mm), acetyl-CoA (0-1mM) and (-)-carnitine three times with dilute aqueous KHCO3 to remove traces hydrochloride (1-25mM) in a final volume of 2-Oml. The of bromoacetic acid and dried with Na2SO4. Removal of initial rate of decrease of E232, due to the deacetylation of the ether in a stream of N2 left a pale-yellow solid, which CoA, was observed by using a Beckman DK-2 recording was recrystallized from aqueous ethanol. A yield of 2-7g. spectrophotometer with the housing maintained at (57% of theory) of colourless bromoacetylthiophenol (m.p. 30+0-30. 36.3-37.3°, uncorr.) was obtained. On storage at 40 in Molecular-weight determination8. A single determination vacuo in the dark, bromoacetylthiophenol develops a yellow of the molecular weight of carnitine acetyltransferase was colour, and it was freshly recrystallized before use. made by using the Archibald technique as modified by (ii) Bromoacetylation of CoA. A 0-5g. portion (2-Om- Ehrenberg (1957). Observations on the meniscus at the moles) of bromoacetylthiophenol was dissolved in 3ml. of air/solvent interface were carried out on a solution con- acetone-0-1M-KHCO3 (5:2, v/v) and mixed with 20mg. taining 9-85mg. of protein/ml. in 0- M-phosphate buffer, (approx. 20,umoles) of CoA dissolved in the same solvent. pH7-5, by using a Spinco model E analytical ultracentrifuge This system was usually monophasic; if it was not, drops at 7950 rev./min. and 170. The subsequent artificial- ofacetone or KHCO3 were added as required. After 45min. boundary experiment was conducted at 16000 rev./min. at room temperature, the solution was brought to pH 1-2 Calculated values for the molecular weight after 48, 64 and with conc. HCI and the acetone removed with a stream of 80min. were 59300, 58900 and 57500 respectively. A value N2. The aqueous slurry that remained was extracted ten of 0-74 was taken for the partial specific volume of the times with ether to remove thiophenol and acylthiophenol, enzyme. This was obtained from the amino acid com- and the last traces of ether were evaporated with N2 as position (J. F. A. Chase, unpublished work) by the proce- before. dure of McMeekin & Marshall (1952). The pH of the solution was adjusted to 3 5-4-5 with Two experiments were performed by using the 'short- KHCO3. About 75% of the CoA originally present was column' sedimentation-equilibrium technique of Yphantis recovered as bromoacetyl-CoA, which may be stored frozen (1960). With 4-5mg. of protein/ml. dissolved in 0-03M- with only a low rate 'of decomposition (less than 5%/week). phosphate buffer, pH7-8, containing NaCl (0-1 m), these A8aay of haloacetyl-CoA derivative8. Both chloroacetyl- gave values for the molecular weight of 61120 and 60580. CoA and bromoacetyl-CoA are substrates for citrate A previous estimate of the molecular weight of carnitine synthase. Solutions of these derivatives may therefore be acetyltransferase, obtained by gel filtration (Andrews, assayed in the system described for acetyl-CoA (Chase, 1964), was 55000 (Chase et al. 1965). Taking this result 1967a), although the time taken for the reaction to with those reported above, the molecular weight of the reach equilibrium must be considerably extended and enzyme has here been assumed to be 58000+3000. larger amounts (100-200,ug.) of citrate synthase used. Protein concentration. Carnitine acetyltransferase solu- Vol. 111 INHIBITION OF CARNITINE ACETYLTRANSFERASE 227 tions were estimated spectrophotometrically, by taking for rapid inhibition of the enzyme to occur, it must El%. to be 8-25 at 280m,u. This value was obtained by be exposed simultaneously to CoA and bromo- submitting enzyme samples to micro-Kjeldahl digestion acetylcarnitine. Neither acetyl-CoA nor desulpho- and determining their nitrogen content (as ammonia) with CoA will substitute for CoA in promoting inhibition. ninhydrin. It was then assumed that the protein contained 16% of nitrogen (Bailey, 1962, p. 300). In experiments similar to those of Fig. 1, bromo- Amino acid analyses. Acid hydrolysates were analysed acetyl-(+)-carnitine was found to be very much by using a Technicon AutoAnalyzer. Known samples of less inhibitory than the (-)-derivative. Thus, even ,B-alanine andS-carboxymethyl-2-mercaptoethylamine were at a concentration of 2mM, only a very slow inhibi- used to determine the elution positions and ninhydrin tion was observed and this may well have been due colour yields of these compounds. Under the usual condi- to trace (about 1 %) contamination with the natural tions suggested by the manufacturers of the AutoAnalyzer isomer of the (+ )-carnitine used in the synthesis of for the estimation of amino acids, /-alanine was eluted bromoacetyl-( + )-carnitine. 23-26min. before S-carboxymethyl-2-mercaptoethylamine, Establishment of experimental conditions for the which was itself eluted 40-44min. before ammonia. Radioactivitymeasurements. Samples(lOOuil.)ofsolutions rapid inhibition of carnitine acetyltransferase by containing ['4C]carnitine dissolved in 01 M-phosphate haloacetyl-CoA. In Fig. 2(a) it is shown that the buffer, pH7-0, were plated on to planchets, dried in vacuo addition of 0 4,uM-bromoacetyl-CoA to the reaction and counted with a Geiger-Muller tube fitted with a Panax mixture used above caused virtually instant type D558 scaler. abolition of carnitine acetyltransferase activity. Even 0 02,tM-inhibitor stopped the reaction in RESULTS about 1 min. (Fig. 2b). Fig. 2(c)-2(e) establish that for rapid inhibition to occur the enzyme must be Establishment of experimental conditions for the simultaneously exposed to bromoacetyl-CoA and rapid inhibition of carnitine acetyltransferase by (-)-carnitine. (+ )-Carnitine, the unnatural iso- bromoacetylcarnitine. Fig. l(a) shows the effect of mer, did not promote inhibition by bromoacetyl- adding 100/,m-bromoacetyl-( ± )-carnitine to a CoA (Fig. 2f). reaction system containing carnitine acetyltrans- Enhancement of chloroacetyl-CoA inhibition by ferase and its substrates, acetyl-CoA and (-)- (-)-carnitine. Under the conditions of Fig. 2, carnitine. Complete inhibition of the enzyme is chloroacetyl-CoA inhibited in the same way as observed within 1 min.; a second sample of enzyme bromoacetyl-CoA. Loss of enzyme activity was, added to the same system, which now contains however, less rapid and this made it easier to investi- some free CoA, is inhibited even more rapidly. The gate the quantitative effects of (-)-carnitine on experiments shown in Figs. l(b)-1(d) establish that, the rate of inhibition. Fig. 3(a) shows that no

(a) (b) (c) (d) A I Ct 4 A' c~1 B

-1I

min. Fig. 1. Inhibition of carnitine acetyltransferase by bromoacetyl-( ± )-carnitine. Progress curves are shown for the reaction between acetyl-CoA and (-)-carnitine as followed at 232mp. (a) Acetyl-CoA, 55,UM; (-)-carnitine, 1 25mM; tris-HCl buffer, pH77, 0-1M. The reaction was started at A by the addition of enzyme. At B bromo- acetyl-( ± )-carnitine was added (final concn. 100iM). A second addition of enzyme was made at A'. The broken line shows the progress of the uninhibited reaction. (b) Enzyme preincubated with 100tuM-bromoacetyl-(±)- carnitine for 2min. in buffer at 300. The reaction was started at C by adding acetyl-CoA and (-)-carnitine (final concentrations as before). A second addition of enzyme was made at A'. (c) Enzyme preincubated as in (b) with 100liax-bromoacetyl-(± )-carnitine, but plus 50tkM-CoA. Additions at C and A' were as in (b). (d) Enzyme pre- incubated as in (c), but with 55,uM-desulpho-CoA in place of CoA. Acetyl-CoA and (-)-carnitine were added at C as in (b) and (c). 228 J. F. A. CHASE AND P. K. TUBBS 1969 (aj (b) (c) (d) (e) (f)

F v A I I , i B ASu iXlAI i~~~~~1 To

.1

min. Fig. 2. Inhibition of carnitine acetyltransferase by bromoacetyl-CoA. Progress curves are shown as in Fig. 1. (a) Acetyl-CoA, 55pm; (-)-carnitine, 1 4mM; tris-HCl buffer, pH7-7, 0-1M. The reaction was started at A by the addition ofenzyme (0-58p,g.). At B bromoacetyl.CoA was added (final concn. 0 35,uM). The broken line shows the progress of the uninhibited reaction. (b) Reaction started at A as in (a). At C bromoacetyl-CoA was added (final conen. 17nM). A second addition of enzyme was made at A'. (c) Enzyme (0.58,ug.) preincubated for 2min. in buffer at 30° with 17nm-bromoacetyl-CoA. The reaction was started at D by adding acetyl-CoA and (-)- carnitine (final concentrations as before). (d) Enzyme preincubated as in (c) with 17nM-bromoacetyl-CoA plus 55px-acetyl-CoA. The reaction was started at E by adding (-)-carnitine. (e) Enzyme preincubated as in (c) with 17nm-bromoacetyl-CoA plus 1.4mM-(-)-carnitine. The reaction was started at F with acetyl-CoA. (f) Enzyme preincubated as in (e) with 1-25mm-(+ )-carnitine in place of (-)-carnitine. The reaction was started at D with acetyl-CoA and (-)-carnitine.

(b) .11 bq0 a ElI Ca Ca 0 P-3 40 ._- ._ o o cS 40 a

P- 0 co 0

20 30 -I *0 1.0 20 3 0 Time (sec.) 10-4/[(-)-Carnitine] (M-1) Fig. 3. Effect of (-)-carnitine concentration on inhibition by chloroacetyl-CoA. (a) In a series of experiments carnitine acetyltransferase was mixed with (-)-carnitine, at the indicated concentration, in lOOmM-tris-HCl buffer, pH7-8, at 300. Chloroacetyl-CoA (final conen. 0-38,UM) was added at zero time and, after various intervals, acetyl-CoA and (-)-carnitine were added simultaneously (final concns. 100,umL and 900/-LM respectively) to start the transferase reaction. Rates were linear for 30-60sec., showing effective protection by acetyl-CoA against further inhibition. Initial (-)-carnitine concentrations: *, zero; A, 36,um; EO, 72,UM; A, 1801,M; 0, 900,UM. (b) Reciprocal ofthe pseudo-first-order inactivation rates (in arbitrary units) observed at the various (-)-carnitine concentrations in (a) plotted against the reciprocal of the corresponding (-)-carnitine concentrations.

inhibition occurred in the absence of (-)-carnitine; a manner that suggests that it is an enzyme- in its presence the progressive inhibition obeyed carnitine complex with Ks 1*13 x 10-4M, and not psuedo-first-order kinetics. The apparent first- the free enzyme itself, that is susceptible to order rate constant for the development of inhibi- attack by chloroacetyl-CoA (Fig. 3b). A value of tion depends on the (-)-carnitine concentration in 1.20 x 10-4M has been reported for the dissociation Vol. Ill INHIBITION OF CARNITINE ACETYLTRANSFERASE 229 Table 1. Protection by deoulpho-CoA again8t bromoacetyl-CoA inhibition of carnitine acetyltransferase Two incubation mixtures were set up containing: tris-HCl buffer, pH 77, 100mM; (-)-carnitine, 1 tuM; desulpho- CoA, 0 or 112pM; carnitine acetyltransferase, 0 81 UM; water to final volume 0-98ml. At zero time 0-02ml. of bromoacetyl-CoA solution was added (final concn. 125IgM) to each mixture. Samples (0-05ml.) were removed for assay after 0, 20, 120 and 600sec. The percentages of the original enzyme activity remaining in each sample are shown. Activity remaining (%) Concn. of desulpho-CoA (,tM) Time ... 20sec. 120sec. 600sec. 0 17-5 8 4 112 46-5 34.5 21

2-0r

0 cq *5

- 0o 4.;CO *0

-4 25 0 5

I._

11 0 0-5 I-O 1-5 2-0 2 5 A; 0 0.01 0 02 0-03 0.04 0 05 Bromoacetyl-(-)-carnitine added (nmoles) Carnitine acetyltransferase added (nmole) Fig. 4. Titration with bromoacetyl-(-)-carnitine. To each Fig. 5. Titration with bromoacetyl-CoA. Samples of car- of two tubes kept at room temperature was added 0-2ml. nitine acetyltransferase (0-0052nmole) were incubated at of carnitine acetyltransferase (1-25nmoles) dissolved in 300 in tris-HCl buffer, pH7-8 (100mM), EDTA (0.25mm), 01 m-phosphate buffer, pH7-0, and lO,ul. (lOOnmoles) of (-)-carnitine (0-25mM) and bromoacetyl-CoA in a final CoA. Samples (2,ul.) were removed for enzymic assay and volume of 19ml. After lOmin. residual enzyme activity then 1,ul. (0.129nmole) of bromoacetyl-(-)-carnitine was was measured by the addition of 0-05ml. of lOOmM-acetyl- added to one tube and 1'pu. of water to the control. After (±)-carnitine and 0-05ml. of LOmM-CoA. Bromoacetyl- 5min. the enzyme activity of each tube was determined as CoA added: *, none; A, 0-0148nmole; 0, 0-0296nmole. before. Further additions of inhibitor and water were made in the same way. 0, Titration with inhibitors; o, control. Enzymic rates are expressed in arbitrary units. enzyme's activity is lost, the degree of inhibition observed is proportional to the amount of inhibitor added. By extrapolation it may be calculated that constant of the enzyme-carnitine complex formed lmole of enzyme is inactivated by 0-97mole of during the catalytic activity of carnitine acetyl- inhibitor. transferase (Chase & Tubbs, 1966a). The close Titration of carnitine acetyltran8fera8e with bromo- similarity between these values strongly suggests acetyl-CoA. Fig. 5 summarizes the results of an that the same enzyme-carnitine complex is in- experiment in which the residual activities of car- volved in catalysis and in the promotion of halo- nitine acetyltransferase solutions were measured acetyl-CoA inhibition. after incubation with bromoacetyl-CoA in the Protection by de8ulpho-CoA again8t inhibition by presence of excess of (-)-carnitine. The molar bromoacetyl-CoA. As shown in Table 1, the CoA proportion of enzyme to inhibitor was varied from analogue desulpho-CoA, which is bound to carnitine about 0-25:1 to 2:1, and lmole of bromoacetyl- acetyltransferase similarly to other CoA derivatives CoA in this system completely inactivated up to (Chase et al. 1966), affords considerable protection 1- 17 moles of enzyme. Further additions of enzyme against inhibition by bromoacetyl-CoA. above the amount required to titrate the inhibitor Titration of carnitine acetyltran8ferase with bromo- had undiminished activity. acetyl-(-)-carnitine. Fig. 4 shows the effect of The results shown in Figs. 4 and 5 suggest that successive additions, each of 0*129nmole, of the enzyme has only one active centre/molecule. bromoacetyl-(-)-carnitine on the activity of car- Recovery of activity by inhibited enzyme. Carnitine nitine acetyltransferase (1.25mnoles) in the pres- acetyltransferase, after inhibition with bromo- ence of excess of CoA. Until at least 85% of the acetyl-CoA plus (-)-carnitine and the subsequent 230 J. F. A. CHASE AND P. K. TUBBS 1969 (bli Enz *BrAcCoA +Cn EnzB*rAcCn*CoASH

II SrCH2 11 C=O BrCH2-C=O S OH SH O

0 , '-- HS, (Cf (d) C02- CH2-C-O CH2 S 0 0 S OH

.0

0 10 20 30 40 (e) 11 Enz-Cn (rn CH2-c=o Time (days) + S -carboxymethyl -CoA, s 0 C S / \ |~~~~~Enz Fig. 6. Recovery of activity by inhibited enzyme. A 40mg. sample (0.69,umole) of carnitine acetyltransferase Scheme 1. Proposed scheme for the inhibition of carnitine in 0 1 M-phosphate buffer, pH 7 15, was treated with 1 Ou-05 acetyltransferase by the bromoacetyl derivatives of CoA moles of (-)-carnitine and 0-93 Lmole of bromoacetyl-CoA and (-)-carnitine. Abbreviations: Cn, (-)-carnitine; in a final volume of3-0ml. After 5min. at room temperature BrAcCn, bromoacetyl-(-)-carnitine; BrAcCoA, bromo- residual enzyme activity was less than 1%. Excess of acetyl-CoA; CoA*S * CH2 *CO. O . Cn, S-carboxymethyl-CoA inhibitor was then removed from the enzyme by passage (-)-carnitine ester. through a column of Sephadex G-25 as described for the experiment shown in Fig. 7. The eluted protein in a volume of 24ml. was stored at 40, and 1,I. samples were assayed for carnitine acetyltransferase activity over a period of same site as other CoA derivatives (Chase & Tubbs, about 6 weeks. Control samples of enzyme showed less than 1966a). A hypothesis developed from these obser- 5% loss of activity after a corresponding time in the vations is illustrated in Scheme 1. It is assumed refrigerator. that, after independent binding of bromoacetyl- CoA and (-)-carnitine to the usual enzyme sites for CoA and carnitine derivatives (Scheme la), removal of excess of inhibitor by gel filtration, transfer of the bromoacetyl group from CoA to showed a slow progressive recovery of activity on carnitine takes place by the normal catalytic standing at 4° (Fig. 6). The rate of this process mechanism (Scheme lb). This is not unreasonable followed pseudo-first-order kinetics with a half- as the enzyme has a broad specificity for the transfer time of about 14-8 days. After 37 days 85% of the of short-chain acyl groups (Chase, 1967b). At this original activity was present; thereafter reactiva- stage, the newly formed enzyme-bound bromo- tion took place more slowly and never exceeded acetyl-(-)-carnitine is involved in an alkylation 89% (51 days). reaction, not of a group on the enzyme, but rather Enzyme that had been inhibited with bromo- of the terminal thiol group of the other substrate, acetylcarnitine plus CoA (Fig. 1) also regained CoA (Scheme lc). The resulting adduct, S-carboxy- activity (5-10% of the uninhibited value) on methyl-CoA (-)-carnitine ester, having affinity for dialysis overnight against 0 1 M-phosphate buffer, both the CoA- and carnitine-binding sites on the pH 7 0, at 40. enzyme, might be expected to dissociate extremely The finding that the inhibition of carnitine slowly. Inhibition of carnitine acetyltransferase acetyltransferase by bromoacetyl-CoA plus car- by bromoacetylcarnitine plus CoA would presum- nitine, or by bromoacetylcarnitine plus CoA, was ably follow an abbreviated scheme without a reversible on prolonged incubation was unexpected, preliminary transbromoacetylation (Scheme lb-c). as these inhibitors had been designed to act irre- Such a sequence of events is compatible with the versibly by alkylating groups on the protein. experiments described in Figs. 4 and 5 on the Consideration had therefore to be given to possible stoicheiometry of inhibition, and the further work alternative mechanisms for their action. The described here was designed to establish that the striking quantitative similarity in the enzyme's above hypothesis is correct. requirement for (-)-carnitine, and not (+)-car- Large-8cale preparation of inhibited enzyme. nitine, both for the inhibition by haloacetyl-CoA Carnitine acetyltransferase (086,umole) in 3-5ml. derivatives (Figs. 2 and 3) and for its normal of OlM-phosphate buffer, pH7-0 (14.2mg. of catalytic function, suggested that the processes of protein/ml.), was mixed with 250,u1. (2.08,umoles) inhibition and catalysis have features in common. of 8.3mM-( + )-[14C]carnitine hydrochloride (specific Further, it appears likely from the results of Table radioactivity 3-6 x 105 counts/min./,umole). Then 1 that bromoacetyl-CoA binds to the enzyme at the 200 pI. (1.12,umoles) of 5 6mM-bromoacetyl-CoA Vol. III INHIBITION OF CARNITINE ACETYLTRANSFERASE 231 to a column (2cm. x 35cm.) of Sephadex G-25 equilibrated with 0- M-phosphate buffer, pH7-0, and was eluted with the same buffer; 26-drop fractions were collected automatically. Fig. 7 shows the elution pattern of protein and unchanged bromoacetyl-CoA as determined from the E260 value of each fraction. Samples (100ul.) were also taken for radioactivity determinations to show the fate of the carnitine label. Ab8orption 8pectrum of the inhibited enzyme. A difference spectrum (curve C in Fig. 8) may be calculated by comparing the spectrum of inhibited 0 5 10 15 20 25 30 carnitine acetyltransferase (curve A in Fig. 8) with Fraction no. that of the native enzyme (curve B in Fig. 8). It has a peak at 260m,u, which could be caused by an Fig. 7. Elution profile from Sephadex G-25 of carnitine adenine nucleotide residue bound to the inhibited acetyltransferase treated with bromoacetyl-CoA plus enzyme. (+)-[14C]carnitine. *, E260 (1Omm. light-path); A, radio- Radioactivity a8sociated with the inhibited enzyme. activity. Fractions 11-15 of the eluate from Sephadex G-25 (Fig. 7) were pooled and found to contain 41-3mg. of protein (0-71 ,umole). Assay of this enzyme, 0 immediately after gel filtration, showed it to have 35j 18% of the specific activity of the original carnitine 0 30k acetyltransferase. The radioactivity associated with this protein was equivalent to that from of the 0-25 - 0'43,umole (-)-carnitine used to promote inhibition. Thus, if it is assumed that no radio- activity was bound to reactivated enzyme, each 0-20 - E mole of inhibited enzyme contained 0-74mole of a (-)-carnitine derivative. 0-l51 Relea8e of radioactivity from the inhibited enzyme. Preliminary experiments showed that neither 0-101 precipitation of the enzyme with ammonium c \ sulphate nor its adsorption on calcium phosphate 0-05 gel and subsequent elution promoted release of radioactivity from the protein or the return of its 0 --o- - - - - b. catalytic function. However, irreversible denatura- 240 250 260 270 280 290 300 tion of the enzyme with trichloroacetic acid Wavelength (m,u) followed by passage through a Millipore filter left more than 90% of the radioactivity in the filtrate. Fig. 8. Altered absorption spectrum of the inhibited To 9-5 ml. of inhibited enzyme (39mg., 0-67,umole) enzyme. Curve A (0), spectrum of 7-2,UM inhibited car- from the experiment shown in Fig. 7 was added nitine acetyltransferase obtained by diluting fraction 12 of an the Sephadex G-25 eluate shown in Fig. 7 with O-1M- equal volume of cold 5% (w/v) trichloro- phosphate buffer, pH7-1. Extinction measurements over acetic acid. Denatured protein was removed by a 10mm. light-path were made by using a Unicam SP. 500 centrifugation and washed once with cold 2-5% spectrophotometer. The protein concentration of the trichloroacetic acid. The volume of the combined inhibited enzyme solution was obtained by the micro- supernatant and washings was reduced to 4-7ml. biuret method (Bailey, 1962, p. 294), with untreated car- by rotary evaporation (bath temperature 300) and nitine acetyltransferase as a standard. Curve B (A), trichloroacetic acid was removed by extracting six corresponding spectrum of 7-2,uM native enzyme. Curve times with ether. C (o), difference spectrum obtained by subtracting curve The sample was diluted to 50ml. with water and B from curve A. applied to a column (1-S5cm. x 21cm.) of DEAE- cellulose equilibrated with 3mM-hydrochloric acid. Elution was performed with a salt gradient, the was added and, after 5min. at room temperature, mixing vessel originally holding 100ml. of 3mM- less than 0-3% of the original enzyme activity was hydrochloric acid, and the reservoir 100ml. of detectable. 3mM-hydrochloric acid containing 0-25M-potassium The entire reaction mixture (3 95ml.) was applied chloride (Moffatt & Khorana, 1961); 26-drop 232 J. F. A. CHASE AND P. K. TUBBS 1969 fractions were collected automatically. The E260 neutral solution the maximum was at 260mp, the value of each fraction was measured and 25,1 . minimum at 228m,u and the E2so/E260 ratio was samples were taken for radioactivity determinations 0-136. These values agree well with published data (Fig. 9). A single sharp peak of 260m,-absorbing for adenine (Beaven, Holiday & material emerged coincidentally with all the radio- Johnson, 1955). Assuming the molar extinction activity of the sample. Fractions 45-48, which coefficient for adenine nucleotides in acid solution comprised this peak, were pooled for analysis. to be 14-2 x 103cm.-l at 260m,u (Beaven et al. 1955), Analy8i8 of the pooled material: identification of the total yield of nucleotide was calculated to be S-carboxymethyl-CoA (-)-carnitine e8ter. (i) Ultra- 0-33,umole. violet-absorption spectrum. This was measured (ii) Alkali-labile (-)-carnitine content of the both at pH2.7 and at pH 7.0. In acid there was an nucleotide. The radioactivity associated with the extinction maximum at 257 m,, a minimum at nucleotide was that present in 0-366,umole of the 230-5m,u and the E280/E260 ratio was 0-202. In [14C]carnitine used to promote inhibition of the original enzyme. Three samples of nucleotide were adjusted to 0-1 M concentration excess of potassium hydroxide 2Or and left at room temperature for 20min. They were then neutralized with hydrochloric acid and assayed for (-)-carnitine (Chase & Tubbs, 1966a), ft .111 together with another nucleotide sample that had 1.5'- not been exposed to alkali. It was found (Table 2) that each mole of nucleotide contained about 1 mole of (-)-carnitine and that at least 95% of this carnitine was released only on alkaline . to e~I0W 1 0 (iii) Acid hydrolysis of the nucleotide. A sample containing 0-205B,mole of nucleotide was evapor- ated to dryness in a rotary evaporator, dissolved in 1 ml. of 6M-hydrochloric acid and hydrolysed in a 0 sealed evacuated tube for 18hr. at 1050. The 5[ hydrolysate, after evaporation to dryness, was dis- solved in 1ml. of 0-1M-hydrochloric acid and applied to a Technicon AutoAnalyzer. It proved 7 : DAME LEM ) to contain stoicheiometric amounts of ,-alanine 0 l0 20 30 40 50 60 70 and S - carboxymethyl - 2 - mercaptoethylamine Fraction no. (Table 2). Fig. 9. Chromatography on DEAE-cellulose of the radio- This analysis strongly suggests that the nucleo- active nucleotide released from inhibited carnitine acetyl- tide formed during the inhibition ofcarnitine acetyl- transferase by acid denaturation. *, E260 (1Omm. light- transferase by bromoacetyl-CoA plus (-)-carni- path); A, radioactivity. tine is indeed S-carboxymethyl-CoA (-)-carnitine

Table 2. Analysi8 of S-carboxymethyl-CoA (-)-carnitine e8ter The values for the amount of ester taken for analysis are based on E260 (see the text). Amount of component found Amount of ester taken for analysis (moles/mole (nmoles) Component estimated (nmoles) of ester taken) 18-3* (-)-Carnitine 0-7 0-04 18-3t (-)-Carnitine 15-9 0-87 21-It (-)-Carnitine 22-4 1-06 36-6t (-)-Carnitine 32-5 0-89 0-921 (-)-Carnitine 1-02 1-11 211 P-Alanine§ 181 0-86 211 S-Carboxvmethyl-2-mercaptoethylamine§ 191 0-91 * (-)-Carnitine estimated without alkaline hydrolysis of the ester. t (-)-Carnitine estimated after alkaline hydrolysis of the ester. I Radioactivity measurement. § Estimated after acid hydrolysis of the ester by using the Technicon AutoAnalyzer. Vol. III INHIBITION OF CARNITINE ACETYLTRANSFERASE 233 Aoo1-A4 A may be represented by Scheme 1. Certain aspects A A ofthe inhibition, however, require further comment. First, the alkylation of CoA by bromoacetyl-( -)- .-1 75 carnitine, which takes place within the enzyme- substrate ternary complex, is an extraordinarily 4-- 50- fast reaction. Bromoacetylcarnitine is not in itself an unusually powerful alkylating agent. Thus, in ..! an attempt to prepare S-carboxymethyl-CoA 25 (-)-carnitine ester synthetically, a mixture con- taining 15mM-CoA and l40mM-bromoacetyl-(-)- --4-- carnitine was kept at room temperature at pH 8-9; 0 1 2 1 8 20 22 24 26 complete reaction, as judged by the mixture Incubation time (hr.) becoming nitroprusside-negative, was achieved only Fig. 10. Time-course (0) of the inhibition of carnitine after 60-90min. In contrast, in Fig. 2(a), at a acetyltransferase (6-8,UM) by S-carboxymethyl-CoA (-)- maximum concentration of enzyme-bound bromo- carnitine ester (15/.4M). A, Control from which the ester acetyl-CoA and (-)-carnitine of 5nm, complete was omitted. inhibition takes less than 5sec. This short time is sufficient for both the alkylation and for the essential preliminary transbromoacetylation (Scheme la-b); in fact the latter must determine the ester, as predicted in Scheme 1. Its spectrum is rate at which inhibition develops, as the stoicheio- that of an adenine nucleotide and it contains metry of the reaction of carnitine acetyltransferase (-)-carnitine in a form with the alkali-lability with bromoacetyl-CoA (Fig. 5) requires that vir- expected ofacid-soluble carnitine O-esters (Pearson, tually none of the bromoacetylcarnitine formed as Chase & Tubbs, 1969). The finding of ,B-alanine and an intermediate (Scheme lb) is released from the S-carboxymethyl-2-mercaptoethylamine is crucial enzyme before taking part in the alkylation of in that it shows the CoA moiety of the nucleotide CoA. Presumably the enzyme's ability to accelerate to be intact and also requires that the process of the alkylation reaction results solely from its inhibition involves an enzyme-catalysed trans- capacity to bind the reactants in close proximity bromoacetylation (Scheme 1c), since alkylation of and in a favourable orientation (Koshland, 1960), CoA must have occurred. Although (± )-[14C]- rather than from any specific chemical catalysis. carnitine had to be used to prepare the inhibited An explanation must also be sought as to why enzyme, the close agreement between the values of the enzyme is so powerfully inhibited by S-carboxy- the carnitine content of the nucleotide estimated methyl-CoA (-)-carnitine ester formed in 8itU. from its radioactivity and by the stereospecific This adduct of the two substrates is imagined as enzyme assay after alkaline hydrolysis (Table 2) forming a 'bridge' between the independent CoA- indicate that only the laevorotatory isomer can and carnitine-binding sites, previously proposed to have been incorporated into the nucleotide, again exist on this enzyme on kinetic grounds (Chase & providing support for Scheme 1. Tubbs, 1966b; Chase, 1967a,b). On a naive inter- Inhibition of carnitine acetyltranaferase by isolated pretation, it might be expected that this inhibitor S-carboxymethyl-CoA (-)-carnitine e8ter. An incu- would be bound to the enzyme with a free energy bation mixture containing 6 8 uM-carnitine acetyl- of binding roughly equal to the sum of the binding transferase, 15,nm-S-carboxymethyl-CoA (-)-car- energies of CoA and carnitine derivatives. Thus, nitine ester (isolated from inhibited enzyme as by taking the product of the dissociation constants described above) and 25mM-phosphate buffer, for short-chain acyl-CoA derivatives and for acetyl- pH 7-0, was kept at room temperature. Samples (-)-carnitine [34tLM (Chase, 1967b) and 350pM were removed at intervals and assayed for enzyme (Chase & Tubbs, 1966a), respectively], a value of activity. A very slow loss of activity was observed 1-2 x 10-8M might be expected for the K, of the (Fig. 10), the degree of inhibition never exceeding adduct. However, the equilibrium formation of an 80%, even after 26hr. A control from which the enzyme-inhibitor complex with such a dissociation nucleotide was omitted showed no loss of activity constant cannot explain the observed results. For over this period. example, in the experiment shown in Fig. 4, in which the enzyme was assayed at a concentration DISCUSSION of 6-25nM, virtually no inhibition would have been observed when less than stoicheiometric amounts The results described here demonstrate that the of bromoacetylcarnitine had been added. It is, of rapid inhibition of carnitine acetyltransferase by course, possible that the dissociation constant bromoacetyl derivatives of CoA and (-)-carnitine deduced above for the adduct might be much too 234 J. F. A. CHASE AND P. K. TUBBS 1969 high, for simple multiplication of the component stant, but rather of the very low rate of adduct dissociation constants may not be appropriate. release, i.e. it is a kinetic and not an equilibrium For metal complexes it is well known that the free effect. Possible explanations of this slow dissocia- energy of binding of a multidentate ligand (here tion have been discussed above. It is probable the adduct) is greater than the sum of the free that the recovery from inhibition (Fig. 6) is largely energies of binding of its individual parts; this is governed by the hydrolysis of S-carboxymethyl- due to slow dissociation of the complex rather than CoA (-)-carnitine ester either in 8itu or after to its accelerated formation (Martell & Calvin, release (Scheme lc-f). The equilibrium constant 1952). Such a chelate effect could presumably also for the reversible binding of the adduct by the contribute to enzyme-inhibitor (or, indeed, to enzyme has not been measured. Indeed, for reasons enzyme-substrate) interactions. Another possi- already discussed above, it would not be directly bility is that adduct formation is accompanied by relevant to the observed inhibition effects. a conformational change in the enzyme-inhibitor The ability of the enzyme to promote the forma- complex, which enhances binding. tion of S-carboxymethyl-CoA (-)-carnitine ester A number of cases are known of protein-small directly demonstrates that it can bind CoA and molecule complexes with very low dissociation carnitine simultaneously, confirming the existence constants. Although in principle such constants of a ternary complex previously inferred from give no indication of the absolute rates offormation kinetic evidence (Chase & Tubbs, 1966a). Indeed, and breakdown of complexes, in practice low values the isolated inhibited enzyme may be regarded as are usually caused by slow breakdown. For a 'frozen' ternary enzyme-substrate complex. example, the pseudo-first-order release of FAD Since carnitine palmitoyltransferase (EC 2.3.1.-) is from D-amino acid oxidase has a half-time of powerfully inhibited by 2-bromolauroyl-CoA (or 133sec., K, for coenzyme binding being 2 5 x 10-7M 2-bromostearoyl-CoA) in a (-)-carnitine-depen- (Dixon & Kleppe, 1965); aminopterin, which in- dent manner (Tubbs & Chase, 1967), its action too hibits folate reductase with KJ about 10-10M, dis- presumably involves the simultaneous binding of sociates from the enzyme-inhibitor complex with both substrates. Such also appears to be the case a half-time of 5-4 x 105sec. (Werkheiser, 1961); and with choline acetyltransferase (EC 2.3.1.6), for the in the extreme case of the combination of biotin inhibition of this enzyme by bromoacetyl-CoA is with avidin, Ks 10-15M is accompanied by a half- choline-dependent (Chase & Tubbs, 1966b), al- time for dissociation of 1.1 x 107sec. (Green, 1963). though such a conclusion is at variance with the Fig. 6 shows that carnitine acetyltransferase kinetic studies of Schuberth (1966). recovers from inhibition by bound adduct with a Finally, it may be noted that carnitine acetyl. half-time of 14-8 days. Since this virtually com- and palmitoyl- can be inhibited in plete recovery occurs in comparatively concentrated intact mitochondria by the mechanism described solution (28tM) it cannot be due to the slow dis- in this paper. This fact has been used to demon- sociation of an inhibitor of high affinity. Fig. 10 strate (Tubbs & Chase, 1966, 1967) that each of shows that, as expected, added S-carboxymethyl- these enzymes exists in two mitochondrial pools, CoA (-)-carnitine ester inhibits the enzyme. only one ofwhich is inactivated by externally added However, for a non-covalent interaction, the bromoacyl-CoA plus (-)-carnitine. The other pool process is unusually slow and contrasts strongly is, however, inhibited by bromoacylcarnitine plus with the virtually instant inhibition observed when intramitochondrial CoA. the adduct is formed in sttu. Other examples of com- We thank the Medical Research Council for an expenses sluggish combination between enzymes and grant and Dr A. Feinstein and Mr N. Buttress for perform- plex small molecules are known, e.g. the slow ing the ultracentrifuge runs. reactivation of yeast apo-pyruvate decarboxylase by (Morey & Juni, 1968). REFERENCES Possibly the favoured configuration of the adduct in free solution is unsuitable for binding; it is well Andrews, P. (1964). Biochem. J. 91, 222. known that NAD and FAD, which have structural Bailey, J. L. (1962). Techniques in Protein . similarities to CoA derivatives, have preferred con- Amsterdam, London and New York: Elsevier Publishing formations (Kaplan, 1960; Beinert, 1960). Equally Co. the structure of the protein may be such as to Baker, B. R. (1967). Design of Active-Site-Directed Irre- and of versible Enzyme Inhibitors. New York: John Wiley and hinder sterically both binding desorption Sons Inc. the adduct. Beaven, G. H., Holiday, E. R. & Johnson, E. A. (1955). The exceedingly potent inhibition of carnitine In The Nucleic Acids, vol. 1, p. 493. Ed. by Chargaff, E. acetyltransferase by S-carboxymethyl-CoA (-)- & Davidson, J. N. New York: Academic Press Inc. carnitine ester, when formed in 8itU, is evidently a Beenakkers, A. M. Th. & Klingenberg, M. (1964). Biochim. consequence, not of a very small dissociation con- biophys. Acta, 84, 205. Vol. 111 INHIBITION OF CARNITINE ACETYLTRANSFERASE 235 Beinert, H. (1960). In The Enzymes, vol. 2, p. 339. Ed. by Green, N. M. (1963). Biochem. J. 89, 585. Boyer, P. D., Lardy, H. & Myrback, K. New York: Kaplan, N. 0. (1960). In The Enzymes, vol. 3, p. 105. Ed. Academic Press Inc. by Boyer, P. D., Lardy, H. & Myrback, K. New York: Chase, J. F. A. (1967a). Biochem. J. 104, 503. Academic Press Inc. Chase, J. F. A. (1967b). Biochem. J. 104, 510. Koshland, D. E. (1960). Advanc. Enzymol. 22, 45. Chase, J. F. A., Middleton, B. & Tubbs, P. K. (1966). McMeekin, T. L. & Marshall, K. ( 1952). Science, 116, Biochem. biophy8. Res. Commun. 23, 208. 142. Chase, J. F. A., Pearson, D. J. & Tubbs, P. K. (1965). Martell, A. E. & Calvin, M. (1952). Chemistry of the Metal Biochim. biophys. Acta, 96, 162. Chelate Compounds, chapter 4. New York: Prentice- Chase, J. F. A. & Tubbs, P. K. (1966a). Biochem. J. 99, 32. Hall Inc. Chase, J. F. A. & Tubbs, P. K. (1966b). Biochem. J. 100, Moffatt, J. G. & Khorana, H. G. (1961). J. Amer. chem. 47P. Soc. 83, 663. Chase, J. F. A. & Tubbs, P. K. (1967). Ab8tr. 4th Meet. Morey, A. V. & Juni, E. (1968). J. biol. Chem. 243, 3009. Fed. Europ. biochem. Soc., 08o, p. 135. Pearson, D. J., Chase, J. F. A. & Tubbs, P. K. (1969). In De Marco, C., Riva, F. & Dupre, S. (1964). Analyt. Biochem. Methods in Enzymology. Ed. by Colowick, S. P. & 8, 269. Kaplan, N. 0. New York: Academic Press (in the Press). Dixon, M. & Kleppe, K. (1965). Biochim. biophys. Acta, Pearson, D. J. & Tubbs, P. K. (1964). Biochem. J. 91, 23c. 96, 357. Schuberth, J. (1966). Biochim. biophys. Acta, 122, 470. Ehrenberg, A. (1957). Acta chem. scand. 11, 1257. Seubert, W. (1960). Biochem. Prep. 7, 80. Fraenkel, G. & Friedman, S. (1957). Vitam. & Horm. 15,73. Srere, P. A. & Kosicki, G. W. (1961). J. biol. Chem. 236, Friedman, S. & Fraenkel, G. (1955). Arch. Biochem. 2557. Biophys. 59, 491. Tubbs, P. K. & Chase, J. F. A. (1966). Biochem. J. 100, Friedman, S., McFarlane, J. E., Bhattacharyya, P. K. & 48P. Fraenkel, G. (1960). Biochem. Prep. 7, 26. Tubbs, P. K. & Chase, J. F. A. (1967). Abstr. 4th Meet. Fed. Fritz, I. B. (1965). In Recent Research on Carnitine, p. 199. Europ. biochem. Soc., Oslo, p. 135. Ed. by Wolf, G. Cambridge, Mass.: Massachusetts Werkheiser, W. C. (1961). J. biol. Chem. 236, 888. Institute of Technology Press. Yphantis, D. A. (1960). Ann. N. Y. Acad. Sci. 88, 586.