Biochem. J. (1981) 195, 735-743 735Printed in Great Britain

Evidence for an essential arginine residue at the active site ofATP citratelyase from rat liver

Seethala RAMAKRISHNA and William B. BENJAMINDiabetes Research Laboratory, Department ofPhysiology and Biophysics, School ofMedicine,

State University ofNew York, Stony Brook, Long Island, NY 11794, U.SA.

(Received 13 January 1981/Accepted 18 February 1981)

Rat liver ATP citrate lyase was inactivated by 2,3-butanedione and phenylglyoxal.Phenylglyoxal caused the most rapid and complete inactivation of enzyme activity in4-(2-hydroxyethyl)- 1-piperazine-ethanesulphonic acid buffer, pH 8. Inactivation by bothbutanedione and phenylglyoxal was concentration-dependent and followed pseudo-first-order kinetics. Phenylglyoxal also decreased autophosphorylation (catalyticphosphate) of ATP citrate lyase. Inactivation by phenylglyoxal and butanedione wasdue to the modification of enzyme arginine residues; the modified enzyme failed to bindto CoA-agarose. The V declined as a function of inactivation, but the Km values wereunaltered. The substrates, CoASH and CoASH plus citrate, protected the enzymesignificantly against inactivation, but ATP provided little protection. Inactivation withexcess reagent modified about eight arginine residues per monomer of enzyme. Citrate,CoASH and ATP protected two to three arginine residues from modification byphenylglyoxal. Analysis of the data by statistical methods suggested that theinactivation was due to modification of one essential arginine residue per monomer oflyase, which was modified 1.5 times more rapidly than were the other arginine residues.Our results suggest that this essential arginine residue is at the CoASH binding site.

ATP citrate lyase [ATP citrate (pro-3S)-lyase;EC 4.1.3.81 is widely distributed in animal tissuesand has been purified to homogeneity from rat liver(Inoue et al., 1966; Linn & Srere, 1979) and adiposetissue (Ramakrishna & Benjamin, 1978, 1979a).ATP citrate lyase has a molecular weight of440000-450000 and consists of four identicalsubunits (M, 110000) (Singh et al., 1976). Theenzyme is known to be a phosphoprotein with 2 molof structural phosphate/mol of tetramer. When theenzyme is incubated with ATP, 2mol of phosphateare covalently linked to a histidine residue/mol ofprotein (Cottam & Srere, 1969; Ramakrishna &Benjamin, 1979a). The phosphoenzyme can reactwith Mg-citrate, forming a citryl enzyme whichreacts with CoASH to form acetyl-CoA andoxaloacetate. ADP and P, are released at any of theintermediate steps.Though the mechanism of this enzyme reaction

has been studied in great detail, little is known aboutthe structure of the catalytic site. Enzyme arginineresidues are known to play an important role in

Abbreviation used: Hepes, 4-(2-hydroxyethyl)-l-piper-azine-ethanesulphonic acid.

Vol. 195

binding of anionic substrate(s), cofactor(s) or effec-tor(s) (Riordan et al., 1977). The role of arginineresidues in enzymatic activity can be studied byusing specific arginine modifiers such as phenyl-glyoxal and 2,3-butanedione. Enzymes may beprotected from inactivation by dicarbonyl reagentsby the presence of substrates that bind at arginineresidues. Ribulose bisphosphate carboxylase isprotected by ribulose bisphosphate (Schloss et al.,1978), NADPH-dependent aldehyde reductase isprotected by NADPH (Davidson & Flynn, 1979),and pyruvate kinase is protected by ATP andphosphoenolpyruvate (Berghauser, 1977). In pre-liminary studies (Ramakrishna & Benjamin, 1979b)we found that the phosphorylation of an adiposetissue phosphoprotein, which we later identified asATP citrate lyase (Ramakrishna & Benjamin,1979a), was markedly inhibited by phenylglyoxal.

Materials and methods

MaterialsDEAE-cellulose (DE-52) was purchased from

Whatman. Sephadex G-25 (fine) was obtained fromPharmacia. 1,2-Cyclohexanedione and 2,3-butane-

0306-3275/81/060735-09$01.50/1 (© 1981 The Biochemical Society

S. Ramakrishna and W. B. Benjamin

dione were obtained from Aldrich. Tris base,2-mercaptoethanol, dithiothreitol, CoASH, ATP,ADP, NADH, malic dehydrogenase, hydroxyl-amine hydrochloride, phenylglyoxal and bovineserum albumin were purchased from Sigma. [y-32PIATP (2-5 Ci/mmol) was prepared as describedby Glynn & Chappell (1964). ['4C]Phenylglyoxal(28mCi/mmol) was purchased from Research Pro-ducts International. Agarose-hexane-CoA (type 1)was a product of P-L Biochemicals. All otherchemicals used were of analytical reagent grade.

Enzyme assaysWhen phenylglyoxal was used for modification of

ATP citrate lyase, activity was assayed by thehydroxamate method (Takeda et al., 1969), becauseof the possibility of inhibition of coupling enzyme inthe malic dehydrogenase-coupled assay by phenyl-glyoxal carried over with the sample. The assaymixture contained 200,umol of Tris/HCl buffer,pH 8.7, lO,umol of dithiothreitol, 20,umol of potas-sium citrate, lO,umol of MgSO4, 0.33,umol ofCoASH, lOO,umol of hydroxylamine and 5,umol ofATP in a final volume of 1 ml. Assays were carriedout at 370 C. Reactions were terminated by theaddition of 1.5 ml of 20% (w/v) trichloroaceticacid/0.4 M-FeCl3. The acetylhydroxamate formedwas measured by its absorbance at 520nm. One unitof enzyme activity is defined as the amount ofprotein necessary to catalyse the formation of 1,umolof acetylhydroxyamate/min or the amount neces-sary to catalyze the oxidation of 1 pmol of NADH/min. Phosphate incorporation was assayed by thephenol-extraction method as described previously(Ramakrishna & Benjamin, 1979a).

Purification ofA TP citrate lyaseATP citrate lyase was purified from rat liver by a

modification of the procedure of Linn & Srere(1979). In the DE-52 chromatography step, proteinwas eluted with a linear gradient of 0.02-0.20M-potassium phosphate, pH 7.5. Fractions with mostATP citrate lyase activity were pooled, precipitatedwith (NH4)2SO4 (35g/100ml) and the precipitatewas dissolved in buffer A [0.02 M-Tris/HCl, pH 7.2,containing 3 mM-dithiothreitol, 0.1 mM-EDTA,1 mM-MgSO4 and 10% (v/v) glyceroll and dialysedovernight against 50 vol. of the same buffer. Thedialysed DE-52 fraction was chromatographed onCoA-agarose (column 1 cm x 20cm) equilibratedwith the same buffer and washed until protein wasno longer eluted. ATP citrate lyase was eluted with0.1 M-citrate in buffer A. The enzyme could also beeluted from the affinity gel by 1 mM-CoASH or0.5M-NaCl. Fractions containing most of the en-zyme were pooled and concentrated by precipi-tation with (NH4)2SO4 (35g/100ml). The precipi-tate was dissolved in a minimal volume (protein

concn. 6-10mg/ml) of buffer A and dialysedovernight against 100vol. of this buffer. The purifiedenzyme was homogeneous by the criterium of gelelectrophoresis, without any observable proteolysis.The specific activity of the pure enzyme by the malicdehydrogenase-coupled assay method was about4.95 units/mg of protein at 20°C and 9.1 units/mg ofprotein at 370C, which is comparable with that ofthe preparation of Linn & Srere (1979).

Details of the purification of the rat adipose tissueenzyme will be published elsewhere. Protein wasestimated by the method of Lowry et al. (1951), bythe dye-binding method (Bradford, 1976) or by itsabsorbance at 279 nm (A 1lom = 11.4).

Reaction with arginine-modifying reagentsUnless otherwise stated, ATP citrate lyase (150-

300,ug/ml) was incubated at 300C with either2,3-butanedione in 50mM-borate buffer, pH8.0, orphenylglyoxal in 50mM-Hepes buffer, pH8.0. En-zyme incubated with buffer alone served as theuntreated control. At the time points indicated,samples (100,ul) were removed and assayed forenzyme activity. In experiments to determine theoptimum pH and buffer conditions to producemaximal effects, borate buffer was replaced byTris/HCl or potassium phosphate buffers at variouspH values. Chemical modifiers were prepared freshdaily.

Protection studiesThe ability of various ligands to protect ATP

citrate lyase against inactivation by butanedione orphenylglyoxal was tested by preincubation of theenzyme (300C for 5 min) with the ligand followed byincubation with the modifier. Samples were removedat periodic intervals and assayed for enzymicactivity.

Kinetic studies ofthe modified enzymeProtein incubated with buffer alone (control), with

3mM-phenylglyoxal for 10min or 30min, or pre-incubated with 5mM-citrate plus 0.25mM-CoASHfor 5min followed by 30min incubation with 3mM-phenylglyoxal in 50mM-Hepes buffer, pH8.0, waschromatographed on a Sephadex G-25 (fine) column(1 cm x 20 cm) to remove the unreacted reagent.Protein was eluted from the column with buffer Acontaining 100mM-NaCl. Fractions (0.75ml) con-taining enzyme were pooled and the protein con-centrations were made similar by appropriate dilu-tion. The time required from the application of thesample on the column to recovery of the protein wasapprox. 12 min during which no additional enzymeinactivation was detected. Kinetic parameters ofthese gel-filtered protein samples were determined byassaying ATP citrate lyase activity by the coupledmalic dehydrogenase method.

1981

736

Modification of arginines in ATP citrate lyase

Affinity chromatography ofmodified enzymeEnzyme (1 mg/ml) was incubated with 5 mM-

phenylglyoxal for approx. 40min. This treatmentdecreased enzyme activity by more than 95%. Thissample was first chromatographed on SephadexG-25 and then loaded at room temperature ontoCoA-agarose (1 ml column volume). The columnwas eluted with buffer A, collecting 1 ml fractions.After 12 fractions, the column was eluted with0.1 M-citrate in buffer A. An untreated proteinsample, processed identically, served as control.

Incorporation of(14ClphenylglyoxalRadiolabelling of the enzyme by phenylglyoxal

was studied by incubating the protein (1.1 mg/ml)for various times with [14C]phenylglyoxal (1.5mM,61 16 c.p.m./nmol) in 50mM-Hepes buffer, pH 8.0, at300C. At different times, aliquots (100,ul) werepipetted onto 2.5 cm2 Whatman 3MM filter diskswhich were immediately dipped in cold 10% (w/v)trichloroacetic acid. The filter disks were washedfour times with 10% (w/v) trichloroacetic acidfollowed by washes with ethanol and diethyl ether.Disks were air-dried and placed in a toluene-basedscintillation mixture [0.5% 2,5-diphenyloxazole/0.03% 1,4-bis(5-phenyloxazol-2-yl)benzenel andcounted in a LKB 1210 Ultrabeta scintillationspectrometer.

Amino acid analysisNative and phenylglyoxal-modified enzyme sam-

ples (200,ul; 1.1 mg of protein/ml) were diluted with40vol. of 6M-HCI containing 0.1 M-2-mercapto-ethanol and hydrolysed at 110°C for 24h. Since theneutral and acidic amino acids were not modified byphenylglyoxal treatment, the hydrolysates werechromatographed on the short column of a Beck-man amino acid analyzer.

Results

Inactivation of A TP citrate lyase by phenylglyoxaland 2,3-butanedione

Rat liver ATP citrate lyase was incubated for10min with 10mM each of phenylglyoxal, 2,3-butanedione, and 1,2-cyclohexanedione in 50mM-borate buffer at pH values 7, 7.5, 8.0, and 8.5. Since1,2-cyclohexanedione decreased the enzyme activityminimally this reagent was not studied further.Phenylglyoxal and butanedione showed maximalinhibition (85 and 35% respectively) of ATP citratelyase activity at pH 8.0. Phenylglyoxal was used formost of the studies since it inactivated the enzymemost rapidly.The logarithm of enzyme activity remaining

plotted as a function of time at various phenyl-glyoxal and butanedione concentrations was linearto less than 10% of the initial enzyme activity (Figs.

Vol. 195

1 and 2). Prolonged incubation with phenylglyoxalcompletely inactivated the enzyme. In controls,enzyme activity remained unchanged when incu-bated without the inhibitor. Inactivation by thesechemical modifiers appeared to follow first-orderkinetics at any fixed concentration of phenylglyoxalor butanedione. The reaction order with respect tophenylglyoxal and butanedione concentrations forthe inactivation of the lyase was determined from theequation:

kapp. = k(M)nwhere kapp is the apparent first-order rate constantfor inactivation, k is the second-order rate constant,and n is the reaction order of phenylglyoxal (M)reacting with the enzyme residues to yield aninactive complex. By replotting the data, logkapp.versus log (M) (inset of Figs. 1 and 2), a value of 1.0was obtained for the reaction order (n). Similarinactivation data were observed for ATP citrate

2.0

1.6

0C1

-a

boO J .-)

0.80 1 5 30 45 60

Time (min)Fig. 1. Inactivation of A TP citrate lyase by phenyl-

glyoxalATP citrate lyase (0.3 mg/ml) was incubated in5OmM-Hepes buffer, pH8.0, at 300C with 1.5mM(0), 3.0mm (A), 6.0mM (O) and 10.0mM (0)phenylglyoxal. At specified times, aliquots (100,ul)were assayed for lyase activity by the hydroxamatemethod. Protein concentrations were in the linearrange of the assay method. Inset: determination ofthe order of reaction with respect to phenylglyoxal.-log (kapp.) was plotted against -log([phenylgly-oxall) (M). The slope of the line gives the reactionorder (n).

737

S. Ramakrishna and W. B. Benjamin

lyase from adipose tissue treated with phenyl-glyoxal. The apparent second-order rate constant (k)for the inactivation of ATP citrate lyase wascalculated from the slope of the plot to be

2.0

1.6

C)

= 1.2

0.

0.4

20 40Time (min)

Fig. 2. Inactivation ofATP citrate lyase by butanedioneATP citrate lyase (0.3-mg/ml) was incubated in50mM-borate buffer, pH 8.0, at 30°C with 5mM(0), 10mM (A), 20mM (O), and 30mM (0)butanedione. The enzyme was incubated with20mM-butanedione in 50mM-Tris/HCI buffer,pH8.0 (A), 50mM-Hepes buffer, pH8.0 (x), and50mM-potassium phosphate buffer, pH 8.0 (v).Inset: determination of the order of reaction withrespect to butanedione in borate buffer. See legendto Fig. 1 for details.

0.33 M-1 s- for phenylglyoxal in Hepes buffer,pH 8.0, and 0.042M-1 s- for butanedione in boratebuffer, pH 8.0.Among the various buffers tested at pH 8.0,

enzyme inactivation by butanedione (20mM) (Fig.2) was, highest in borate buffer, followed bypotassium phosphate and Hepes buffers, and least inTris/HCl buffer. Inactivation by phenylglyoxal wasmaximal in Hepes buffer (k = 9.33 M-1 s-) followedby potassium phosphate buffer (k = 0.25 M-1 s-1)and then borate buffer (k = 0.078 M-1 - s-1) (data notshown). Hence, in all further experiments, inacti-vation by phenylglyoxal was studied in 50 mM-Hepesbuffer, pH 8.0, and the effects of butanedione werestudied in 50 mM-borate buffer, pH 8.0.The first step in the catalytic reaction of ATP

citrate lyase is the formation of phosphoenzymefrom Mg-ATP (Srere, 1972). Treatment of ATPcitrate lyase with phenylglyoxal inhibited enzymeautophosphorylation, which is most likely at thecatalytic site (Ramakrishna & Benjamin, 1979b).This inactivation appeared to follow a pseudo-first-order reaction process.

Properties ofmodified enzymeThe kinetic properties of thQ enzyme modified by

treatment with 3mM-phenylglyoxal for 10min and30min, and by treatment with 3mM-phenylglyoxalfor 30min in the presence of 0.25 mM-CoASH and5mM-potassium citrate, were investigated (Table 1).The apparent Km values for ATP, citrate andCoASH were found to be unaltered during thecourse of inactivation by phenylglyoxal. However,the V of the enzyme decreased in proportion to theinactivation of the enzyme. The pseudo-first-orderrate constant for the decrease in V was 0.053 min-at 3mM-phenylglyoxal. This value is similar to thekapp. obtained for inactivation of ATP citrate lyaseby 3 mM-phenylglyoxal (Fig. 1). Protection of theenzyme with CoASH plus citrate limited thedecrease in V. Inactivation by butanedione also didnot alter the Km values though the rate of decrease in

Table 1. Effect ofinactivation by phenylglyoxal on kinetic parameters ofA TP citrate lyaseThe units for V are millunits (mU) per 4,pg protein at 200C.

Kinetic parameters for

ATP Citrate* CoASH, A 5 _ 5~~~~

Treatment Km (gM) V (mU) Km (gM) Km (mM) V (mU) Km (pM)Control 144 19.2 117 1.5 18.8 213mM-Phenylglyoxal (1Omin) 161 13.7 122 1.5 13.5 213mM-Phenylglyoxal (30min) 155 3.8 108 1.5 3.8 220.25 mM-CoASH + 5 mM-citrate + 156 8.5 127 1.5 9.1 23

3 mM-phenylglyoxal (30 min)* Under normal assay conditions a low and a high Km values are obtained for citrate concentrations.

V (mU)18.913.34.29.1

1981

738

Modification of arginines in ATP citrate lyase

V values with time was similar to the rate ofinactivation (results not shown).

In order to study whether arginine residues wererequired for substrate binding, enzyme which had

0.3

0.2

0.1

0Fraction no.

Fig. 3. Chromatography of phenylglyoxal-inactivatedA TP citrate lyase on CoA-agarose

Enzyme (1 mg/ml) was inactivated by incubationwith 5 mM-phenylglyoxal at 30°C for 40min in50mM-Hepes buffer, pH8.0, and chromatographedon CoA-agarose as described in the Materials andmethods section. Inactivated enzyme (0) anduntreated control (0). At the arrow, elution wasstarted with 0.1 M-citrate in buffer.

first been inactivated with 5 mM-phenylglyoxal toless than 5% of original activity was chromato-graphed on Sephadex G-25 to remove unreactedphenylglyoxal. The inactivated enzyme was thenchromatographed on CoA-agarose. The modifiedprotein was not retained (Fig. 3), but untreatedenzyme was retained on the column and was elutedfrom the column with 0.1 M-citrate in buffer A. Thesefindings suggest that ATP citrate lyase inactivationby phenylglyoxal may be due to modification of theCoASH binding site.

Protection by substrates against inactivation byphenylglyoxal and butanedione

Inactivation rates produced by these reagents inthe presence of various concentrations of enzymesubstrates were studied to determine if substrateswould protect the enzyme from inactivation byphenylglyoxal and butanedione. Protection studieswere performed by preincubating the enzyme withthe protector followed by incubation with themodifier. Pseudo-first-order kinetic data were ob-served for the inactivation of ATP citrate lyase byphenylglyoxal and butanedione in the presence ofadded ligands (Fig. 4a). The ability of variousligands to protect the lyase from reacting with themodifier is shown in Table 2. Mg-ATP at con-centrations lower than 0.4mM slightly reduced theinactivation produced by a-dicarbonyl reagents.

0 10 20Time (min)

30 40

0 3 6

Phenylglyoxall (mM)9

Fig. 4. Protection ofATP citrate lyasefrom inactivation by phenylglyoxal(a) Enzyme (0.3 mg/ml) was preincubated for 5 min with 50mM-Hepes buffer, pH 8.0, (0) or with buffer containing5 mM-citrate (E), 0.25 mM-CoASH (A\) and 5 mM-citrate plus 0.25 mM-CoASH (0). Note that the concentrations ofprotecting agents given are final concentrations after the addition of modifier. During pretreatment the volumes were3/4 of the final volume after addition of phenylglyoxal. The pretreated enzyme samples were then incubated with1.5 mM-phenylglyoxal and at specified time points aliquots (lOO,ul) were assayed for enzymic activity as described inthe legend to Fig. 1 and the Materials and methods section. (b) Kinetic analysis of protection by ligands. Values forkapp. were obtained from a series of experiments in which the concentration of phenylglyoxal was varied in thepresence of fixed concentrations of various ligands. kapp was plotted as a function of phenylglyoxal concentration(mM). No added ligand (0), 5 mM-citrate (A), 0.25 mM-CoASH (O) and 5 mM-citrate plus 0.25 mM-CoASH (5).

Vol. 195

739

000

S. Ramakrishna and W. B. Benjamin

Table 2. Effect ofvarious ligands on the rate of inactivation ofA TP citrate lyase by phenylglyoxal and 2,3-butanedioneValues of kapp. were obtained from the plots of 2.3 x log (percentage residual activity) against time of incubationwith reagent.

ReagentPhenylglyoxal (5 mM)

Butanedione (10mM)

Ligand

Citrate (2mM)Citrate (5 mM)Citrate (15 mM)Citrate (40mM)Mg-ATP (0.4mM)Mg-ATP (0.8mM)CoASH (0.25 mM)CoASH (0.5 mM)Citrate (2.0mM) + Mg-ATP (0.8mM)

Citrate (5 mM)Citrate (20mM)Mg-ATP (0.25 mM)CoASH (0.25 mM)CoASH (0.25 mM) + citrate (5 mM)

103 x kapp. (min-')93.052.645.240.837.067.8108.724.424.447.122.210.08.3

14.85.94.9

However, Mg-ATP at concentrations of 0.5 mm andabove did not protect but augmented the inacti-vation by phenylglyoxal. Citrate (5mM) decreasedthe rate of inactivation by 5 mM-phenylglyoxal byapprox. 50% and that by l0mM-butanedione by65%. With increasing concentrations of citrate (to40mM), protection against inactivation by phenyl-glyoxal increased, but in a non-linear fashion.Citrate plus Mg-ATP did not produce additionalprotection compared with citrate alone. CoASH(0.25 mM) protected the enzyme from inactivation byapprox. 75%. Protection by CoASH plus citrate washigher than when either ligand was added alone (Fig.4a). ADP, a competitive inhibitor of citrate lyase,did not protect against inactivation by phenyl-glyoxal. Of the several ligands tested, CoASHappeared to be the most effective in protecting ATPcitrate lyase against inactivation by phenylglyoxal orbutanedione. Among the substrate combinationstested, citrate plus CoASH gave maximal pro-tection. Inactivation of ATP citrate lyase in thepresence of added ligands was also examined atdifferent phenylglyoxal concentrations. Plots of kappas a function of phenylglyoxal concentrations withdifferent ligands were linear (Fig. 4b). The apparentsecond-order rate constant (calculated from theslope) decreased with protection and the lowestvalue (k = 0.13 M-1 s-) was observed with citrateplus CoA.

Determination of the number of arginine residuesmodified

Fig. 5 illustrates the correlation between mol of[14C]phenylglyoxal incorporated into each mol ofsubunit and the extent of inactivation. The degree of

inactivation during the initial 80-90% loss ofenzyme activity was directly proportional to [14C]-phenylglyoxal incorporation. About 17 mol ofphenylglyoxal were calculated to be incorporated permonomer of enzyme at 100% inactivation. Studiesof the amino-acid composition of the modifiedenzyme suggested that only arginine residues weremodified by phenylglyoxal. A linear relation wasalso observed between the arginine residues de-stroyed by phenylglyoxal and loss of enzyme activity(Fig. 5). From these data it was calculated that forevery arginine residue modified, approximately 2 molof phenylglyoxal were incorporated per mol ofenzyme.

Radioactive phenylglyoxal incorporation intoboth the protected and unprotected enzyme wasmeasured to determine the number of arginineresidues protected by the ligands. The differences inthe extent of incorporation between unprotected andprotected enzyme at each time point was plottedagainst the enzyme activity lost from the un-protected enzyme, according to the method ofSchloss et al. (1978). A difference of about 4-6molof phenylglyoxal incorporated per mol of protectedenzyme subunit compared to the unprotected wascalculated by extrapolation to 100% inactivation.Thus the ligands (citrate, CoASH, or CoASH pluscitrate) protected 2-3 arginine residues per subunitfrom the total of 8.5 residues modified by phenyl-glyoxal.To determine the number of essential arginine

residues, the data were analysed by the statisticalmethod of Tsou (1962) as described by Horiike etal. (1979). This method requires the determinationof both the number of residues modified and theresidual enzyme activity in partially modified pro-

1981

740

Modification of arginines in ATP citrate lyase

0 25 50 75 100

Enzyme activity lost from unprotected enzyme(% of initial)

-

0

0

E

0

0

a_

E

0

._

Ec0._

x15 °

a>..o

r_E

C o

10 0 Eo_

o* 0

v- o0.~

i-.0Q).5 .I

C) %-

0'd v

Fig. 5. Phenylglyoxal incorporation per monomer ofATP citrate lyase as afunction ofenzyme activity lostThe enzyme (1.1 mg/ml) was incubated with 1.5 mM-[14C]phenylglyoxal. Arginines residues modified weredetermined by amino acid analysis (-); (0), ['4C]phenylglyoxal incorporated into monomer of enzyme. Thedifference in incorporation between unprotected enzyme and enzyme protected by either 5 mM-citrate (v),0.3mM-ATP (0), 0.25 mM-CoASH (A) or 0.25 mM-CoASH plus 0.3 mM-ATP (U) at corresponding time points isgiven in the lower curves.

tein samples. The data given in Fig. 5 allows such ananalysis from the equation:

m-n(l-x) = n-pal"i-(n-p)aa/i (1)

where n is the total number of modifiable arginineresidues in the enzyme, among which p of theresidues, of which i are essential, react withphenylglyoxal at a pseudo-first order rate constantk, and (n-p) residues that are not essential react ata pseudo-first-order rate constant k2 (=ak,), a is thefraction of activity remaining, x is the fraction ofunmodified arginine residues and m is the number ofarginine residues modified per molecule. Eqn. (1) canbe written in the following form (Tsou, 1962;Horiike et al., 1979):

nx 1 a-1log [a' -pj =log (n-p)+ ( .1loga (2)

In eqn. (2), the left-hand component was calculatedwith n = 8.5, p = i = 1 and plotted against a (Fig. 6)which gave a satisfactory linear plot. This suggeststhat one essential arginine residue per monomer isinvolved in the catalytic function of ATP citratelyase. A value of 0.68 was obtained for a = (kl/k)from the slope (a- 1/i) of the curve, which suggeststhat the rate of modification of the essential arginineresidue is about 1.5-fold faster than for othermodifiable arginyl residues. Fig. 6, inset, shows therelation between the enzyme activity remaining, a,and the number of arginine residues modified, m,that was calculated from eqn. (1) with n = 8.5,p = i = 1 and a = 0.68. The experimental valuesobtained were also plotted. The experimental curveagreed closely with the calculated values.

Vol. 195

1.5 0.2

0 2 4 6 8

)-t 1.0

bo0

0.5-2.0 -1.5 - 1.0 -0.5 0

log (a)

Fig. 6. Analysis of A TP citrate lyase-phenylglyoxalreaction data by the method of Tsou (1962)

Experimental conditions are the same as in Fig. 5.The left-hand component in eqn. (2) was calculatedassuming n = 8.5, p = i = 1 and plotted againstlog (a). From the slope of the curve the value of acan be determined. Inset, the relationship betweennumber of arginine residues modified (m) andfraction of enzyme activity (a). Using eqn. (1) withn = 8.5, p = i =1 and a = 0.68 the number ofarginine residues modified for each value of a werecalculated and plotted (0) and for comparison theexperimentally obtained values (A) were alsoplotted.

Discussion

This paper describes the effects of arginine-specific reagents, phenylglyoxal and butanedione, on

741

S. Ramakrishna and W. B. Benjamin

ATP citrate lyase. Phenylglyoxal caused a rapidinactivation of ATP citrate lyase, the reaction beingpseudo-first-order. Butanedione inactivated enzymeactivity but to a lesser degree, and cyclohexane-dione was least effective. Butanedione has beenshown by others (Riordan, 1973; Chollet, 1978;Cipollo & Dunlap, 1978; Jordan & Wu, 1978;McTigue & Vanetten, 1978; Rogers et al., 1978) andin this report to have maximal effects in boratebuffer, possibly due to stabilization of the reactionproduct of the butane-dione-guanidino group byesterification with borate (Riordan, 1973). Inactiva-tion by phenylglyoxal was lowest in borate buffer,and the most rapid inactivation was observed inHepes buffer, pH 8.0. Inactivation by the a-dicarbonyl reagents appeared to be a direct con-sequence of the modification of enzyme arginineresidues.

Kinetic analysis of the phenylglyoxal-modifiedATP citrate lyase showed that the Km values for thesubstrates were not significantly altered. The Vdecreased with inactivation in a pseudo-first-orderfashion. This suggests that modification of theenzyme by phenylglyoxal or butanedione com-pletely abolishes enzyme activity and that anyresidual activity in treated samples is the activityderived from unmodified molecules with normalactive sites.

There are approx. 40 arginine residues per mol ofenzyme monomer (Srere, 1972). Binding datashowed that approx. 17mol of phenylglyoxal werebound per subunit. Our data is consistent with thenotion that for every two mol of phenylglyoxalbound to protein, one mol of arginine residue ismodified (Takahashi, 1968; Daemen & Riordan,1974; Lobb et al., 1975; Schloss et al., 1978).Anionic substrates (Mg-ATP, citrate and CoASH)protected the enzyme against inactivation, though todiffering degrees. At concentrations less than0.5 mM, Mg-ATP reduced the inactivation byphenylglyoxal and butanedione but at concentra-tions higher than 0.5 mM, Mg-ATP increased therate of inactivation. The ambiguity in protection byMg-ATP cannot be explained by our data. Higherconcentrations of Mg-ATP may bring conforma-tional changes at the active site of the enzyme, thusrendering essential arginine more reactive towardsa-carbonyl reagents. CoASH and CoASH pluscitrate gave excellent protection from inactivation byphenylglyoxal and butanedione. The binding datashowed that of the 8.5 arginine residues of theenzyme susceptible to phenylglyoxal modification,2-3 residues were protected by substrates. Thecombination of substrate-induced conformationalchanges and direct binding of the anionic groups ofthe substrates to arginine residues may be re-sponsible for this protection.

The number of essential amino acid residues

required for catalytic activity has been determinedfor pyridoxamine phosphate oxidase (Horiike et al.,1979), pepsin (Paterson & Knowles, 1972) andtransferrin (Rogers et al., 1978) by analysing thedata produced by chemical modification withspecific amino acid modifiers by the statisticalmethod of Tsou (1962). By a similar analysis, wedetermined that each ATP citrate lyase subunit hasan essential arginine residue required for catalyticfunction. As discussed by Horiike et al. (1979), themethod for the determination of the number ofessential amino acid residues from a plot of a as afunction of m by extrapolation of the initial linearportion of the curve to 100% inactivation, can beused only when a= 1 and n(=p) = i= 1 or a< 1 andp = i = 1. Since in the present case a was calculatedto be 0.68, this extrapolation method was not used.The order of reaction obtained with phenylglyoxaland butanedione was 1, indicating that one arginineresidue is modified at the active site to produceinactivation.The reaction catalyzed by ATP citrate lyase

involves a complex series of transformations at theactive site, possibly involving phosphoryl-enzyme(covalent) enzyme-citryl phosphate (noncovalent),citryl-enzyme (covalent) and enzyme-citryl-CoA(noncovalent) intermediates. The last-named undergoes Claisen cleavage, giving rise to the productsoxaloacetate and acetyl-CoA carbanion, whichfinally is converted to acetyl-CoA (Walsh & Spector,1969; Walsh, 1979). The glutamate y-carboxylgroup was suggested as the nucleophile involved inthe formation of both phosphoryl-enzyme andcitryl-enzyme (Suzuki et al., 1969; Suzuki, 1971). Incontrast, the phosphate residue in the phospho-enzyme was identified as histidyl N-phosphate(Cottam & Srere, 1969; Ramakrishna & Benjamin,1979a). From the studies on modification of arginineresidues, enzyme inactivation cannot be explaineddirectly in terms of the Walsh (1979) reactionmechanism. However, modification of arginineresidues in the close vicinity of the active site mightmake the nucleophilic group less approachable to theelectrophilic groups of the substrate and thuscause inhibition. Alternatively, arginine could be thebase involved in Claisen cleavage of citryl-CoA thatis modified by a-dicarbonyl reagents; hence leadingto inhibition of enzyme activity. However, moreexperimental data is needed to support any of theseassumptions.The bulky nature of substitution (about 8 arginine

residues modified per subunit) could bring aboutconformational changes in the enzyme structurewhich could at the most cause partial inhibition butnot complete inhibition. In majority of the casesstudied, enzyme inactivation by phenylglyoxal andbutanedione was shown to be due to modification ofan essential arginine residue which prevents binding

1981

742

Modification of arginines in ATP citrate lyase 743

of the substrate to the enzyme (Berghauser, 1977;Riordan et al., 1977; Jordan & Wu, 1978; Philipset al., 1978; Rogers et al., 1978; Vensel &Kantrowitz, 1980). It is not known whether ATPcitrate lyase has the same or different binding sitesfor different substrates. Complete loss of enzymeactivity by treatment with both phenylglyoxal andbutanedione suggests that an arginine residue(s)which interacts with substrates at the bindingsite (possibly CoASH binding site) was modified.Additional evidence for this hypothesis was thefinding that the modified enzyme failed to bind to aCoA-agarose column and the excellent protectionafforded by CoASH. Protection was minimal withcitrate, and Mg-ATP at higher concentrations failedto protect the enzyme. In view of the complexities ofthe protection patterns noted, we were not able todetermine if arginine residue(s) are involved in thebinding of ATP and citrate to ATP citrate lyase.However, our data strongly suggest that the in-activation ofATP citrate lyase by phenylglyoxal andbutanedione is at least in part due to the modi-fication of arginine residues at the CoASH bindingsite.

This study was supported by grant AM 18905 from theNational Institutes of Arthritis, Metabolism and Diges-tive Disease, National Institutes of Health.

References

Berghauser, J. (1977) Hoppe-Seyler's Z. Physiol. Chem.358, 1565-1572

Bradford, M. (1976) Anal. Biochem. 72, 248-254Chollet, R. (1978) Biochem. Biophys. Res. Commun. 83,

1267-1274Cipollo, K. L. & Dunlap, R. B. (1978) Biochem. Biophys.

Res. Commun. 81, 1139-1144Cottam, G. L. & Srere, P. A. (1969) Biochem. Biophys.

Res. Commun. 35, 895-900Daemen, F. J. M. & Riordan, J. F. (1974) Biochemistry

13, 2865-2871Davidson, W. S. & Flynn, T. G. (1979) J. Biol. Chem.

254,3724-3729

Glynn, I. M. & Chappell, J. B. (1964) Biochem. J. 90,147-149

Horiike, K., Tsuge, H. & McCormick, D. B. (1979) J.Biol. Chem. 254, 6638-6643

Inoue, H., Suzuki, F., Fukunishi, K., Adachi, K. &Takeda, Y. (1966) J. Biochem. (Tokyo) 60, 543-553

Jordan, F. & Wu, A. (1978) Arch. Biochem. Biophys.190,699-704

Linn, T. C. & Srere, P. A. (1979) J. Biol. Chem. 254,1691-1698

Lobb, R. R., Stokes, A. M., Hill, H. A. 0. & Riordan,J. F. (1975) FEBS Lett. 54, 70-72

Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall,R. J. (195 1) J. Biol. Chem. 193, 265-275

McTigue, J. J. & Vanetten, R. L. (1978) Biochim.Biophys. Acta 523, 422-429

Paterson, A. K. & Knowles, J. R. (1972) Eur. J. Biochem.31, 510-517

Philips, M., Roustan, C., Fattoum, A. & Pradel, L. (1978)Biochem. Biophys. Acta 523, 368-376

Ramakrishna, S. & Benjamin, W. B. (1978) Fed. Proc.Fed. Am. Soc. Exp. Biol. 37,477

Ramakrishna, S. & Benjamin, W. B. (1979a) J. Biol.Chem. 254,9232-9236

Ramakrishna, S. & Benjamin, W. B. (1979b) Fed. Proc.Fed. Am. Soc. Exp. Biol. 38, 1027

Riordan, J. F. (1973) Biochemistry 12, 3915-3923Riordan, J. F., McElvany, K. D. & Borders, C. L., Jr.

(1977) Science 195, 884-885Rogers, T. B., Borresen, T. & Feeney, R. E. (1978)

Biochemistry 17, 1105-1109Schloss, J. V., Norton, I. L., Stringer, C. D. & Hartman,

F. C. (1978) Biochemistry 17, 5626-5631Singh, M., Richards, E. G., Mukherjee, A. & Srere, P. A.

(1976) J. Biol. Chem. 251, 5242-5250Srere, P. A. (1972) Curr. Top. Cell. Regul. 5, 229-283Suzuki, F. (1971)Biochemistry 10, 2706-2710Suzuki, F., Fukunishi, K. & Takeda, Y. (1969) J.

Biochem. (Tokyo) 66, 767-774Takahashi, K. (1968) J. Biol. Chem. 243, 6171-6179Takeda, Y., Suzuki, F. & Inoue, H. (1969) Methods

Enzymol. 13, 153-160Tsou, C. L. (1962) Sci. Sinica 11, 1535-1558Vensel, L. A. & Kantrowitz, E. R. (1980) J. Biol. Chem.

255, 7306-7310Walsh, C. T. (1979) in Enzymatic Reactions Mechanisms,

pp. 765-768, Freeman, San FranciscoWalsh, C. T. & Spector, L. B. (1969) J. Biol. Chem. 244,4366-4374

Vol. 195