Ketoglutarate Dehydrogenase Complex - Semantic Scholar

2 downloads 0 Views 1MB Size Report
malate dehydrogenase (1). However, a feature of most hetero- enzyme complexes is that the enzymes in the complex have binding sites for a common ligand.
Vol. 264, No. 21, Issue of July 25, pp. 12303-12312,1989 Printed in U.S.A.

THEJOURNALOF BIOLOGICAL CHEMISTRY 01989 by The American Society for Biochemistry and Molecular Biology, Inc

Kinetic Advantagesof Hetero-Enzyme Complexes with Glutamate Dehydrogenase and the &-Ketoglutarate Dehydrogenase Complex* (Received for publication, October 7, 1988)

Leonard A. Fahien$, Michael J. MacDonaldS, J a n K. Tellern, Barbara Fibich, and Catherine M. Fahien From the Departments of Pharmacology and §Pediatrics, University of Wisconsin Medical School, Madison, Wisconsin 53706

We have found previously (Fahien, L. A., Kmiotek, complex have been described previously (3), but to date only E. H., MacDonald, M. J., Fibich, B., and Mandic, M. a few kinetic advantages have been found due to association (1988) J. Biol. Chem. 263, 10687-10697) that gluta- of glutamate dehydrogenase with the aminotransferase and mate-malate oxidation can be enhanced by cooperative malate dehydrogenase (1).However, a feature of most heterobinding of mitochondrial aspartate aminotransferase enzyme complexes is that the enzymes in the complex have and malate dehydrogenase to the a-ketoglutarate debinding sitesfor a common ligand.This suggests that a major hydrogenase complex. The present results demonstratefunction of hetero-enzyme complexes is to provide a means that glutamate dehydrogenase, whichforms binary for a direct transfer of the ligand from one enzyme to the complexes with these enzymes, adds to this ternary other. Advantages of direct transfer over diffusion of the complex and thereby increases binding of the other ligand from one enzyme to the other have beendiscussed enzymes. Kinetic evidence for direct transfer of a- previously (4). ketoglutarate andNADH, within these complexes, has In the work reported here, we have studied the possibility been obtained by measuring steady-state rates of Ez of a direct transfer of a-ketoglutarate from aspartate aminowhen most of the substrate coenzyme or is bound to the aminotransferase or glutamatedehydrogenase ( E l ) . transferase to glutamate dehydrogenase and of NADH beRates significantly greater than those which can be tween glutamate dehydrogenase and malate dehydrogenase. accounted for by the concentrationof free ligand, cal- Furthermore, since glutamate dehydrogenase shares the ligands GTP and a-ketoglutarate with GTP-linked succinate culated from the measuredvalues of the El-ligand dissociation constants, require that the El-ligand com- thiokinase and the a-ketoglutarate dehydrogenase complex, plex serve asa substrate forEZ(Srivastava, D. K., and respectively, we have sought evidence for the direct transfer Bernhard, S. A. (1986) Curr. Tops. Cell Regul. 28, 1- of these ligands within complexes of the relevantenzymes. In 68). By this criterion, NADH is transferred directly addition, since glutamate dehydrogenase and the a-ketoglufrom glutamate dehydrogenase to malate dehydrogen-tarate dehydrogenasecomplex both associate withmalate ase and a-ketoglutarate is channeled from the amino- dehydrogenase and the aminotransferase, we have determined transferase to both glutamate dehydrogenase and the whether there are competing complexes or if a complex can a-ketoglutarate dehydrogenase complex. Similar evi- be formed between all four enzymes. A quaternary complex dence indicatesthat GTPbound to an allosteric site on organized around the a-ketoglutaratedehydrogenase complex glutamate dehydrogenase functions as a substrate for could facilitate transfer of ligands between the constituent succinic thiokinase.The potentialphysiological advan- enzymes. of activators and inhibitors well as tages to channeling as substrates withinmultienzyme complexes organized EXPERIMENTALPROCEDURES aroundthea-ketoglutaratedehydrogenase complex are discussed. Enzymes and Reagents-We found greater binding of citrate syn-

Glutamate dehydrogenase (L-glutamate:NAD(P)+ oxidoreductase, EC 1.4.1.3) and the a-ketoglutarate dehydrogenase complex (EC 1.2.4.2) both associate with mitochondrial malate dehydrogenase (L-malate:NAD+ oxidoreductase EC 1.1.1.37) andaspartateaminotransferase (L-aspartate:oxalacetate aminotransferase EC 2.6.1.1) (1-3). Kinetic advantages resulting from association of malate dehydrogenase and the aminotransferase with the a-ketoglutarate dehydrogenase * This work was supported by National Institutesof Health Grants CA 40445 and AM 28348 and by a grant to the University of Wisconsin Medical School from the NIH Division of Research Facilities and Resources. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact. j T o whom correspondence should be addressed. ll Permanent address: Dept. of Animal Physiology, Poznan University, Poland.

thase to the a-ketoglutarate dehydrogenase complex (3) than reported previously (5, 6). Preparing the a-ketoglutarate dehydrogenase complex with detergent could possibly alter hetero-enzyme interaction. Therefore, we isolated the a-ketoglutarate dehydrogenase complex by using a modification of the procedure of Stanley and Perham (7) in which detergent (0.5% w/v, Triton X-100) was omitted from all steps except theinitialextraction.The complex was storedin 50 mM potassium phosphate, pH 7.0, 0.1 mM EDTA, 1 mM dithiothreitol, and 10 pg/ml leupeptin a t -70 “C, and used without further treatment. As observed previously(8), thiamine pyrophosphate dissociated from the complex during its isolation and was rebound slowly in the presence of M e . Preincubation of the isolated complex with thiamine pyrophosphate/MF had noeffect on its binding to the enzymes discussed in this report. The specific activity of the purified, preincubated complex assayed under conditions (5 mM a-ketoglutarate, 5 mM Ca-EGTA,’ pH 7.4) found experimentally to providea saturating concentration of aketoglutarate was 25 pmol of NADH/min/mg at 30 “C. Lower values in the literature were determined a t a nonsaturating concentration of a-ketoglutarate (9). Analysis by polyacrylamide gel electrophoresis The abbreviations used are: EGTA, [ethylenebis(oxyethylenenitri1o)ltetraacetic acid; ANS, 8-anilino-1-naphthalene sulfonic acid; CoA, coenzyme A PMP, pyridoxamine-P; PLP, pyridoxal-P.

12303

Multienzyme Complexes

12304

in sodium dodecyl sulfate (10) indicated that the preparation was over 90% pure. Bovine and/or rat liver mitochondrial aspartate aminotransferase, malate dehydrogenase, and glutamate dehydrogenase were prepared with previously described methods (11-15). Pig heart mitochondrial malate dehydrogenase and citrate synthase were obtained from Boehringer Mannheim. Other enzymes were obtained from Sigma. Polyethylene glycol (8000) and 8-anilino-1-naphthalene sulfonic acid (ANS) were obtained from Fisher and Aldrich, respectively. Other coenzymes, substrates, andreagents were obtained from Sigma. Stock solution of all reagents used in assays were adjusted to the pH of the assays. Solutions of succinyl-CoA and coenzyme A were prepared fresh daily. The concentration of succinyl-CoA was determined by measuring its absorbance at 260 nm and correcting for the amount of coenzyme A by measuring the concentration of sulfhydryl groups as described previously (16). Methods used to dialyze and prepare enzymes for use in these experiments were described previously (3). Concentration of Enzymes-The concentration of the a-ketoglutarate dehydrogenase complex was measured using the method of Lowry et al. (17) with bovine serumalbumin as a standard. The concentrations of glutamate dehydrogenase, malate dehydrogenase, citrate synthase, glyceraldehyde-3-phosphate dehydrogenase, NADP isocitrate dehydrogenase, and succinate thiokinase were measured spectrophotometrically at 280 nm as described previously (3). The concentration of the aminotransferase is expressed in terms of the concentration of boundpyridoxal-P or pyridoxamine-P and was calculated from the absorption spectrum as described previously (3, 12). Analysis of Kinetic Data-Michaelis constants were obtained from the Michaelis-Menten relationship (Equation 1).

concentrations according to the reaction

Therefore, in these experiments, the level of free a-ketoglutarate was calculated from Equation4 (the equilibrium relationship for the aminotransferase half reaction) where KI is the equilibrium constant of the half reaction

Measurement of KI-The value of KI was determined in separate experiments as described previously (21). The pyridoxal-P form (22 p M ) was titrated with 0-200 p M glutamate and the pyridoxamine-P form (22 p M ) was titrated with 0-40 p~ a-ketoglutarate. Titrations were followed spectrophotometrically at 360 nm. In these titrations, the initial concentrations of aminotransferase (ET),glutamate (GT), and wketoglutarate (ST) were much smaller than the dissociation constant of a-ketoglutarate ( k , / k , ) and glutamate (k3/kl), which according to previous results (22), can be as high as 0.7 and 12 mM, respectively. Consequently, the level of E,X was negligible, and there were nospectralsignsresulting from the accumulation of E,X. Therefore, the absorbance at 360 nm (AOb.) equals: Aobs

=

c[ET]+ Ac[El-PLP] (5)

where ~ ( 2 . 5X lo3) is the measured molar absorbance of El-PMP and A45 X lo3) isthe molar absorbance of El-PLP minus 6 . When titrations were made with a-ketoglutarate: [E1-PLP] = [GI [El-PMP]

In Equation 1, u, V , and K, are the initial velocity, maximal velocity, and Michaelis constant, respectively, and S is the concentration of free substrate. The effect of an allosteric activator (A) on glutamate dehydrogenase activity in the presence of saturating levels of substrate can be expressed by Equation 2 (18,19).

In Equation 2, bv is the change in velocity produced by the activator, V is the velocity in the absence of activator, V, is the velocity in the presence of saturating levels of activator and K,, is the dissociation constant of the activator from the coenzyme-substrate-enzyme complex. Equations 1 and 2areidentical empirically, and therefore, the kinetic constants of both equations can be evaluated in the same manner with the least square method as described previously (20). Studies of Direct Transfers-An enzyme ( E 2 )which catalyzes the reaction

E2 + S + E2S + E2 + product can be inhibited by a high level of a second enzyme (El)if the second enzyme hasa sufficiently high affinity for the substrate. If the substrate is not altered by El, then theexpected amount of inhibition produced by E, can be obtained by calculating the level of free S from Equation 3 and

[SI

=

[ET]- [El-PLP]

[ST] - [El-PLP]

(6) (7)

(8)

When titrations were made with glutamate: [SI = [El-PMP] =

[ET]- [El-PLP] (9)

[GI = [GT] - [El-PMP]

(10)

The value of K1 was calculated by substituting the concentration of El-PLP (obtained from Equation 5) into Equation 4 and Equations 6-8 or 9 and 10. The most reliable estimates of KI were obtained at concentrations of glutamate (100-200 KM) where the change in the concentrationof glutamate was within experimental error and the concentrations of El-PLP and El-PMP atequilibrium were nearly equal. The average value calculated for KI was 14 & 4. The maximum difference in the calculated concentration of a-ketoglutarate at equilibrium was 22% for this range of values of KI under the conditions of the experiment described under "Results." The values of the dissociation constants of aspartate and oxalacetate for mitochondrial aspartate aminotransferase are two low (22) for the equilibrium constant (KII) for the aminotransferase halfreaction with oxalacetate to be measured by titrating the pyridoxamine-P form of the aminotransferase with oxalacetate. However, the equilibrium constant (Keq)of the over-all aminotransferase reaction Keq =

[Aspartate][S] [Oxalacetate][G]

is between 6 and 7 under a wide variety of conditions (22-25). Since KI, = (Keq)(Kl),a value of 84 was used for KI,. Measurements of KD"The dissociation constants (KD) of NADH and NADPH for glutamate dehydrogenase were measured with an substituting the calculated value of S into the Michaelis-Menten relationship for E, (Equation 1). InEquation 3, KO, [E1] and [E,.SI Aminco Bowman spectrofluorometer at 25 "C, as described previously are the dissociation constant of S for El, concentration of free El, (26, 27). Small aliquots of reduced coenzyme (5-10 pl) were added to and the concentration of the E,. S complex, respectively. Failure of glutamate dehydrogenase (1.0 mg/ml; 18 &M) in 20 mM potassium E, to inhibit E2 to the extentpredicted by Equations 1 and 3 indicates phosphate, 0.1 mM EDTA, pH 7.0, plus the indicated additions. After that S can be directly transferred from El to E,, within a complex (4) each addition of reduced coenzyme to enzyme, mixing was accomplished with a plasticplunger. Excitation and emission were measured between the two enzymes (E1. S .Ed. Studies of a direct transfer of a-ketoglutarate (S) from the pyri- at 340 and 450 nm, respectively. The difference in fluorescence ( A F ) doxamine-P form of mitochondrial aspartate aminotransferase (E1- at each addition of reduced coenzyme was obtained by subtracting PMP) to glutamate dehydrogenase ( E 2 )are more complicated. When the fluorescence of the sameconcentration of reduced coenzyme n-ketoglutarateis added to E,-PMP, it reactsinstantaneously to added in the absence of glutamate dehydrogenase from the total produce the pyridoxal-P form (E1-PLP) and glutamate (G) in equal fluorescence of the mixture (reduced coenzyme plus glutamate dehy-

12305

Multienzyme Complexes drogenase). The fractional saturation of enzyme was calculated from the ratio of AF/AF,, where SF, is the maximal increase in fluorescence at saturating concentrations of reduced coenzyme (obtained from a double-reciprocal plot of A F versus concentration of reduced coenzyme). The number of binding sites (n) and KO were then determined from plots of [ S T ] / [ E . Suersus ] ( [ E T] [ E . S ] )according to Equation 12 (28),

enzyme which receives thesubstrate (Ed. Thus, failure of El to inhibit was not due to the presence of EPas a contaminant of E,. All assays were performed spectrophotometrically at 340 nm (3). Additional Methods-Coprecipitation of enzymes in polyethylene glycol, preparation of the aminotransferase in the pyridoxamine-P form, absorption spectroscopy, and measurements of the molar absorbance of the pyridoxamine-P and the pyridoxal-P form of the aminotransferase were performed as described previously (3, 33, 34). RESULTS

where ET and ST are the total concentrations of glutamate dehydroBinding of Reduced Coenzyme-Plots of the fluorescence of genase and coenzyme, respectively, and E.S is the concentration of the complex ([ET]AF/AFm). the reduced coenzyme-glutamate dehydrogenasecomplex verThe dissociation constant of GTP for glutamate dehydrogenase sus the concentration of NADH or NADPH areshown in Fig. was determined by a similar spectrofluorometric technique with ANS 1 (upper). The results are consistent with onereduced coenas the fluorescent probe (29). The ANS-glutamate dehydrogenase complex was titrated with GTP in the presence of a constant level of zyme-binding site with a KO of 7 and 5 p M for NADH and NADH at excitation and emission wavelengths of 390 and 480 nm, NADPH, respectively. Citratesynthaseandoxalacetate, respectively. It was determined that glutamate dehydrogenase (18 which are constituentsof the malatedehydrogenase assay (see p ~ was ) saturated with both ANS (100 p M ) and NADH (100 FM), and thus changes in fluorescence were the direct result of binding of “Experimental Procedures”), did not alter bindingof NADH GTP toANS-NADH-glutamate dehydrogenase. At the excitation and or NADPH to glutamate dehydrogenase (Fig. 1, upper). emission wavelengths employed, fluorescence of the NADH-glutaThere can be problems in these titrations when the level of mate dehydrogenase complex was minimal. Also, changes in fluoresreduced coenzyme is high. NADH can associate with both cence resulting from binding of NADH to glutamate dehydrogenase active and allosteric sites on glutamate dehydrogenase ( 3 5 ) . were negligible because the concentration of NADH was greater than five times that of the enzyme. Previous experiments (30) demon- However, binding of reduced coenzyme at allosteric sites strated that ANS and GTP do not compete for glutamate dehydrowould have been negligible at the low (140p M ) concentrations genase. (13, 35) andisnot a of NADH used inthesetitrations Compatibility between Measurements of K,, KI, and KD-Cofactors and additional substrates of the reaction catalyzed by E2 can possibly consideration with NADPH which only binds to the active alter binding of S to E,. Therefore, application of Equations 1,3, and site (35). While an innerfilter effect may have caused AF to 4 to studies of direct transfer requires measuring the KO or KI of E, under conditions similar to those used to measure the K, of Ez. KD, be artifactually low at >40 p~ reduced coenzyme, it should Kl, and K , were all measured in the presence of 20 mM potassium not invalidate Ku estimated from the data at the lower conphosphate, 0.1 mM EDTA, pH 7.0, at 25 “C. The constituents (100 centrations with AF, obtained by extrapolation on the basis p~ NADPH and 20 mM NH4C1)of the assay for glutamate dehydroof the known, single, noninteracting, high affinity, reduced genase were found previously (31, 32) not to alter the half-reaction coenzyme site/polypeptide chain. Ourvalues of KOare in fact of the aminotransferase with a-ketoglutarate, andwe found that this was also the case for the constituents (2 mM MgC12,200 pM thiamine slightly higher than those obtained previously with different pyrophosphate, 100 p~ CoA, and 1 mM NAD) of the a-ketoglutarate methods (36, 37). This is because our experiments were perdehydrogenase assay. Since NADPH, unlike NADH, is not bound to formed in 20 mM phosphate, which would produce the slight the aminotransferase (31) NADPH (100 p M ) was also used as the coenzyme in studies of adirect transfer of oxalacetate from the increase in KD (37). aminotransferase to malate dehydrogenase. Furthermore, thespecific Direct Transfer of Reduced Coenzyme-The expected effect activity of malate dehydrogenase was 100-fold less with NADPH than of glutamate dehydrogenase on malatedehydrogenase activity with NADH, and thus with NADPH as coenzyme, malate dehydrogenase activity was not detected in our preparationof the aminotrans- in the absence of hetero-enzyme interaction is shown in Fig. ferase (22 p ~ which ) had been further purified as described in Ref. 1 (center,curves C and D ) and Fig. 2 (curve C). These 33. theoretical curves were calculated fromthe Michaelis-Menten Double-reciprocal plots of the velocity of the a-ketoglutarate derelationship (Equation 1) for the malate dehydrogenase rehydrogenase complex versus the concentration of a-ketoglutarate were linear when the level of a-ketoglutarate was less than 100 p ~ . action with NADH (Fig. 1, center, curue C; and Fig. 2, curve Higher levels of a-ketoglutarate produced substrate activation. There- C) or NADPH (Fig. 1, center, curue D).The values of K, fore, the K,,, used in Equation1for the a-ketoglutaratedehydrogenase used in the calculations were 45 and 77 ~ L Mfor NADH and reaction was obtained by linear extrapolation throughconcentrations NADPH, respectively (Fig. 1, lower). The concentration of of a-ketoglutarate lower than 100 p ~ . free reduced coenzyme substituted into Equation 1 was calIn studies of a directtransfer of NADH or NADPH from glutamate dehydrogenase to malate dehydrogenase, the KO of NADH and culated with Equation 3, using values of l and 5 PM for the NADPH for glutamate dehydrogenase and the K,,, of reduced coen- KD of NADH and NADPH, respectively (Fig. 1, upper). The zyme for malate dehydrogenase were both measured in the presence observed effect of glutamate dehydrogenase on malate dehyof oxalacetate (100 p ~ and ) citrate synthase(50 pglml). A variety of proteins protected malate dehydrogenase from inactivation; citrate drogenase is shown in Fig. 1 (center, curves A and B; and Fig. synthase was used so that the results could be correlated with sub- 2, curue A ) . It can be seen that glutamate dehydrogenase sequent investigations of malate oxidation at pH7.0, where coupling produces considerably less inhibition than expected, a result with citrate synthase facilitates measurements of the kinetic conwhich indicates a direct transfer of reduced coenzyme from stants of malate dehydrogenase (3). In studies of a direct transfer of GTP from glutamate dehydrogenglutamate dehydrogenase ( E 1 )to malate dehydrogenase ( E 2 ) ase to GTP-linkedsuccinate thiokinase, the KOof GTP for glutamate within a complex ( E , .S .E2) between the two enzymes. It has dehydrogenase was determined in the presence of all of the constituents required for the assay of succinate thiokinase except lactate been demonstrated previously that in the absence of subdehydrogenase and pyruvate kinase. The level of these auxillary strates, glutamate dehydrogenase can associate with malate enzymes (20 pg/ml) was much less than that of glutamate dehydro- dehydrogenase (1, 38). The present results indicate that in genase (1 mg/ml), and thus the auxillary enzymes would not have the presence of saturating levels of oxalacetate, NADH-glualtered binding of GTP toglutamate dehydrogenase. In all studies of a directtransfer, itwas determined that theenzyme tamate dehydrogenase also associates with malate dehydrowhich donates the substrate (E,) was not contaminated with the genase as shown in Mechanism 1:

Multienzyme Complexes

12306

MECHANISM 3

Since the malate dehydrogenase reaction is ordered with NADH adding to the enzymebefore oxalacetate(39),the general steady-state rate equationfor Mechanism 1 would be quite complex. However, the mechanism was simplified to essentially a one-substrate reactionby using saturating levels of oxalacetate (40). Furthermore, when the level of free glutamate dehydrogenase was low, glutamate dehydrogenase did not alter either the rate of a nonsaturating concentration of NADH (Fig. 1, center, curue A ) or the rate as a function of the total, asopposed to thefree concentration of NADH (Fig. 2, curue A uersus B ) . Therefore, in Mechanism 1, k k’ and the K, of NADH and NADH-glutamate dehydrogenase are also essentially the same, i.e. NADH-glutamate dehydrogenaseand free NADHare essentially equivalentsubstrates. Frieden (40) has derived Equation 13, which is a steady-state rate equationfor a mechanism similar to Mechanism1 where the modifier (in this case glutamate dehydrogenase) does not alter the maximal velocity (V).

-“

(GLUTAMATE DEH

I

-0.025

0

0025

0 05

IAREDUCED COENZYME)~M-’

FIG. 1. Transfer of reduced coenzyme between glutamate dehydrogenase and malate dehydrogenase. The upper panel shows the difference in fluorescence (arbitrary units)between reduced coenzyme plus bovine liver glutamate dehydrogenase and reduced coenzyme alone (excitation, 340 nm; emission, 450 nm) uersw the concentration of reduced coenzyme added. Additions were 0, no additions; 0, 50 pg/ml citrate synthase; and A, 100 p~ oxalacetate plus 50Fg/ml citrate synthase. The data points of curues A for NADPH and B for NADH are experimental. The solid curues were calculated on the basis of one reduced coenzyme-binding site/glutamate dehydrogenase polypeptide chain with a KO of 5 and 7 ptM for In Equation 13, [“S”]is the sum of the concentration of the NADPH andNADH, respectively. The center panel shows the velocequivalent substrates S and E,. S and K, and K3 are dissocia- ity of pig heart malate dehydrogenase as a function of glutamate . and El. Ez dehydrogenase concentration. These assays were performed in the tion constants of the stepsEt.S .EP E , E P S presence of 30 @ NADH (curue A ) or 20 p~ NADPH (curue B ) plus El E2,respectively. According to Equation 13,the slight 100 pM oxalacetate and 50 pg/ml citrate synthase. The concentration inhibition produced by high levels of free glutamate dehydro- of malate dehydrogenase was 0.01 pg/ml and 1pg/ml in curues A and genase (Fig. 1, center, curue A ) results from K5 exceeding K3. B, respectively. Curves C and D are theoretical curves for the reaction and The solid curves (Fig. 1, center, curue A and Fig. 2, curue B ) in the absence of hetero-enzymeinteractionwithNADH NADPH, respectively. They were calculated with Equations 1 and 3 have been calculated from Equation 13 for K5 = 80 pM and (see “Experimental Procedures”) using the values of KO and K, for K3 = 44 pM with the measuredvalue of K,,, (Fig. 1,lower) and reduced coenzyme obtained from the results shown in the upper and [El] obtained from Equation 3. Consistent with the conclusion lowerpanels, respectively. The solid line (curue A ) was calculated with Equation 13 using the measured value of K, (lowerpanel), thevalues that K5 > K3, we havepreviouslyfound in direct binding of E , and S calculated from Equation 3 and the data shown in the experiments that NADH enhancesdissociation of the gluta- upper panel, letting K3 = 44 pM, and K5 = 80 pM. The experimental mate dehydrogenase-malate dehydrogenase complex (38). A values represent mean & standard error of three experiments. Doublecorollary of NADH enhancing dissociation of the heteroreciprocal plots of the velocity of the malate dehydrogenase reaction enzyme complex, and the fundamental reason for inhibition uersus the concentration of either NADH and NADPH are shown in Curves A and B, respectively, of the lower panel. These assays were by free glutamate dehydrogenase, is that its binding to malateperformed in the presence of 100 p~ oxalacetate and 50 pg/ml citrate dehydrogenase decreases the rate constantof association (k,) synthase. The data points are experimental and the solid lines are of NADH with malate dehydrogenase (40). Since the K, of least square fits to Equation 1. Velocity is in units of nanomoles of oxidized coenzvme Droduced/milliliter/minute. All experiments were NADH equals V/k, (where V is determined by therate performed in h e presence of 20 mM’potassium phosphate and 0.1 constants of dissociation of malate and NAD from malate mM EDTA, pH 7.0, at 25 “C.

+

+ +

dehydrogenase, Ref. 39) and glutamate dehydrogenase does notalter V, high levels of free glutamate dehydrogenase increase the K, of NADH for malate dehydrogenase. Equation 13 is also consistent with the results obtained with NADPH(Fig. 1,center, curueB ) . The primarydifference between the experiments with NADH and NADPH is that

K,/[“S”] (see Equation 13) was larger with NADPH so that inhibition by free glutamate dehydrogenase was more pronounced. In contrast, we found, in agreement with (41), that glutamate dehydrogenase inhibited oxidation of NADPH by

Multienzyme Complexes I

' 75

I/

I

O

O

12307 I

I

I

U

5.0

(NADH) pM FIG.2. Effect of NADH concentration on malate dehydrogenase activity in the presence of glutamate dehydrogenase. Assay conditions were the same as those used to study the effect of glutamate dehydrogenase concentration on the rate of NADH oxidation (Fig. 1, center, curue A ) except glutamate dehydrogenase was either absent (curue A ) or constant at 36 p M (curue B ) . The points shown with curves A and B are experimental. The solid curue B is theoretical for the effect of NADH in the presence of hetero-enzyme interaction. Curve B was calculated as described for calculating curve A of Fig. 1 (center).The dashed Curve C is theoretical for the reaction in the absence of hetero-enzyme interaction and was calculated as described for calculating curves C and D for Fig. 1 (center).

0

IO

20

30

[AMINOTRANSFERASE] pM

NADP-isocitrate dehydrogenase essentially as predicted from Equations 1 and 3. Direct Transfer of Keto Acid-The inhibitory effect of the pyridoxamine-P form of mitochondrial aspartate aminotransferase on glutamate dehydrogenase, the a-ketoglutarate dehydrogenase complex, and malatedehydrogenase in the presence of 20 p~ of their keto acid substrates is shown in Fig. 3 (upper, curves B-D, respectively). The expected inhibition in FIG. 3. Effect of the pyridoxamine-P form of mitochondrial the absence of hetero-enzyme interaction is shown in Fig. 4 aspartate aminotransferase on dehydrogenases. The effects of (curve B, upper and lower). These theoretical curves were the pyridoxamine-P form of bovine liver mitochondrial aspartate calculated by substituting the concentration of the free keto aminotransferase on the activity of bovine liver glutamate dehydrogenase, the bovine heart a-ketoglutaratedehydrogenase complex, and acid substrate and the measured value of the K , (Fig. 3, lower) bovine liver malate dehydrogenase are shown in the upper panel into the Michaelis-Menten relationship (Equation 1) for the curves B-D, respectively. The effect of the pyridoxal-P form on all dehydrogenase reaction. The values of K , used were 56 p~ three dehydrogenases is shown in curve A (upper).The keto acid was for a-ketoglutarate in the a-ketoglutarate dehydrogenase re- 20 p~ a-ketoglutarate in all assays of glutamate dehydrogenase and action, 100 p~ for a-ketoglutarate in the glutamate dehydro- the a-ketoglutaratedehydrogenase complex and 20 p~ oxalacetate in assays of malate dehydrogenase. Additional experimental conditions genase reaction,and 50 p~ for oxalacetate in themalate are described below (lower panel). In these assays the dehydrogenase dehydrogenase reaction (Fig. 3, lower). The concentration of was added last. The experimental values represent mean t standard free keto acid was calculated withEquation 4 using 14 and84 error of three experiments. The lower panel shows double-reciprocal for the equilibrium constant of the aminotransferase halfplots of the velocity uersus the concentration of keto acid. Assay conditions: A , malate dehydrogenase (0.6 pg/ml), oxalacetate varied, , reaction with a-ketoglutarate ( K I )and oxalacetate ( K I z ) respectively (see "Experimental Procedures"). Additional theo- 100 p~ NADPH; B, a-ketoglutarate dehydrogenase complex (1.2 pg/ ml),a-ketoglutarate varied, 2 mM MgC12, 100 p M CoA,200 p~ retical curves (Fig. 4,curves A-C, upper and lower) are shown thiamine pyrophosphate, 1 mM NAD, 1 mM dithioerythritol; C, gluto demonstrate that thesecurves are similar over a range of tamate dehydrogenase (0.75 pg/ml), a-ketoglutarate varied, 100 p~ values of K,, K I , and KI1. The theoretical curves are similar NADPH, 20 mM NH,Cl. All assays were done in 20 mM potassium because, over the indicated rangeof values of K z and KII,the phosphate, 0.1 mM EDTA, pH 7.0, at 25 "C. Velocity is in units of equilibrium level of keto acid substrate is about the same andchange in absorbance at 340 nm/min. The data points are experidecreases almost linearly with increase in the concentration mental and the solid lines are least square fits to Equation1. of the pyridoxamine-P form. Since the level of keto acid substrate is considerably lower than the K,,, of the dehydro- inhibition of glutamate dehydrogenase than expected in the genases, the theoreticalvelocity also decreasesalmost linearly. absence of hetero-enzyme interaction (Fig. 4, upper, curvesB The pyridoxamine-P form of the aminotransferase inhib- versus E ) . For example, at initial concentrations of 20 p M CYited malate dehydrogenase by decreasing the level of oxal- ketoglutarate and 22 p~ pyridoxamine-P-aminotransferase, acetate, almost as predicted from Equations 1 and 4 (Fig. 4, the apparent concentration (15 p M ) of free a-ketoglutarate lower, curue B versus D ) . The presence of an inhibitor in the estimated (Equation 1) from the velocity of the glutamate (3.6 f 0.4 p ~ ) aminotransferase preparation was ruled out because the pyr- dehydrogenase reactionisfourtimesthat idoxal-P form had noeffect on malatedehydrogenase activity calculated (Equation 4)from the average value of K z (14 f 4) (Fig. 3, upper, curve A) and because increasing the level of for the aminotransferasehalf-reaction. The higher concentraoxalacetate from 20 to 50 p ~ in, the presenceof the pyridox- tion cannot be correct because it would give acalculated (Equation 5 ) increase in the absorbance at 360 nm that is amine-P form, eliminated inhibition (data not shown). In contrast with results obtained with malate dehydrogen- only 1/4 the observed increase, whichwent from 0.055 to 0.138, ase,thepyridoxamine-Pform producedconsiderablyless and a value of K I 140 to 180-fold lower thanthe value

12308

Multienzyme Complexes higher than expected velocity apparently resulted from glutamate dehydrogenase reductively aminating the a-ketoglutarate associated with the hetero-enzyme complex as shown below: E , I'LP

I 1 I

+ ~ ~ l u l a m a t e=

B,.l'RII"-~l,.~et~,y~,l,~,~'~,~

I I

NH~---B.--"N.AL)I'H

-

i-iH:"-).',"__

'~

YAUI'H

b:( r w + E + Y.WP + ( ; I ~ ~ ~ ~ ~ Equilibrium of the aminotransferase half-reaction would be restored by the reaction of a-ketoglutarate with the free E,PMPpresent inseveralfoldmolar excess over glutamate dehydrogenase.

E,. P M P

~

~

+ a-ketoglutarate + El. P M P . a-ketoglutarate + E , . PLP

+ glutamate

Presumably the rateof the glutamatedehydrogenase reaction is slightly slower in the presence than in the absence of the aminotransferase because the a-ketoglutarate-aminotransferase complex is a less effective substrate than free a-ketoglutarate. An alternativeexplanation, namely that binding to the aminotransferase markedly decreased the K,,, of glutamate (AMINOTRANSFERASE) pM dehydrogenase for a-ketoglutarate,is unlikely since the pyriFIG. 4. Theoretical versus observed plots of percent inhibition versus the concentration of the pyridoxamine-P form doxal-P form, to which it also binds, had no effect (Fig. 3, of the aminotransferase. The upper panel shows the theoretical curve A ) . Theaminotransferasealsobindstothea-ketoglutarate effect of the pyridoxamine-P form in the absence of hetero-enzyme interaction on glutamate dehydrogenase (curues A-C) and the a- dehydrogenase complex (3). As with glutamate dehydrogenketoglutarate dehydrogenase complex (also curue B ) . The theoretical ase, the pyridoxamine-P form caused less inhibition of the curves were calculated from Equations 1 and 4, with the measured dehydrogenasecomplex thanpredictedintheabsence of value of K,,, (Fig. 3, lowerpanel) and with Kr (the equilibrium constant of the aminotransferase half-reaction with a-ketoglutarate)equal to hetero-enzyme interaction (Fig. 4, upper, curve B versus D ) , 18,14, and 10 for curves A-C, respectively. The actual results obtained whereas the pyridoxal-P form had no effect (Fig. 3, upper, with the a-ketoglutarate dehydrogenase complex and glutamate de- curve A ) . The pyridoxamine-P formdid, however, inhibit the hydrogenase are shown in curves D and E, respectively. The lower oxidative decarboxylation of a-ketoglutarate by the dehydropanel shows the theoretical effect of the pyridoxamine-P form in the genase complex more than it did the reductive amination of absence of hetero-enzyme interaction on malate dehydrogenase. The a-ketoglutarate by glutamate dehydrogenase (Fig. 4, upper, theoretical curves were calculated as described above with KII (the curveD versus E ) . Therefore,thea-ketoglutarate-aminoequilibrium constant of the aminotransferase half-reaction with oxalacetate) equal to 100, 84, and 60, respectively, for curves A-C. The transferase complex is apparently less reactive with the aketoglutarate dehydrogenasecomplex than with glutamate actual results are shown in curve D. Experimental conditions are described in the legend to Fig. 3. dehydrogenase. Quaternary Complex Containing the a-Ketoglutarate Dehydrogenase Complex and Glutamate Dehydrogenase-The a TABLEI ketoglutarate dehydrogenase complexcan forma ternary comBinding of glutamate dehydrogenase to mitochondrial aspartate plex with m-aspartate aminotransferase and m-malate dehyaminotransferase in the presenceof substrates In these experiments,0.02 mg of glutamate dehydrogenase (GDH) drogenase (3). As shown in Table I1 (line 21, glutamate dewas incubated either alone orwith an excess (0.4 mg) of the pyridoxhydrogenase also associates with the a-ketoglutarate dehydroamine-P form of mitochondrial aspartate aminotransferase (AspAT) genase complex. However, under a wide variety of conditions, for 20 min in 1.0 ml of 14% (w/v) polyethylene glycol containing 20 including phosphate buffer which stabilizes glutamate dehyPM a-ketoglutarate, 20 mM NH,Cl, 100 p~ NADPH, 20 mM potassium phosphate, and 0.1 mM EDTA, pH 7.0, at 25 "C. At the end of drogenase (13), glutamate dehydrogenase binds most tightly a much lower affinthe incubation, the mixturewas centrifuged for 20 min at 25 "C, and to m-aspartate aminotransferase and has ity for the a-ketoglutaratedehydrogenase complexand malate the precipitate and supernatant were assayed for enzyme activity. The values represent mean & standard error of three experiments. dehydrogenase (Table I1 and Reference 38). For example, when the concentration of each enzyme is less than or equal Enzymes precipitated Enzymes incubated to 0.1 mg/ml, glutamate dehydrogenase essentially only asGDH AspAT 11, lines 3-5). sociates withtheaminotransferase(Table % Higher levels of malate dehydrogenase and the a-ketoglutarGDH, AspAT 90 ? 2.8 4 1.2 ate dehydrogenase complex are required for significant bindGDH 5 ? 1.2 ing of glutamate dehydrogenase (Table 11, line 2, and Ref. 1). A variety of other proteins, even in high levels, do not assomeasured by us and others (22). The measured value of KI ciate with glutamate dehydrogenase (1, 38). The a-ketoglutaratedehydrogenase complex bound slightly and the data in Table I indicate that in this experiment, 75% of the aminotransferase would have been in the pyridoxal-P more aminotransferase and malate dehydrogenase than did form, 82% of the a-ketoglutaratewould have been converted glutamate dehydrogenase (Table 11, line 6 versus 7). Yet an to glutamate, and most of the 13 nM glutamate dehydrogenase excess of the dehydrogenase complex (0.2 mg/ml) added to glutamate dehydrogenase (0.1 mg/ml), malate dehydrogenase would have been bound to the aminotransferase. Thus, the

*

,

~

12309

Multienzyme Complexes TABLEI1 Association of glutamate dehydrogenase with the ternary complex Inthese experiments, malate dehydrogenase (MDH),aspartate aminotransferase(AspAT),glutamate dehydrogenase (GDH),and the a-ketoglutarate dehydrogenase complex (KDH) were incubated as indicated in 1 ml of 14% (w/v) polyethylene glycol, 20 mM potassium phosphate, 0.1 mM EDTA, pH 7.0, a t 25 "C. After a 20-min incubation, the mixture was centrifuged and assayed for enzyme activity as described under "Experimental Procedures." Over 90% of the a-ketoglutaratedehydrogenase complex was precipitated in either the presence or absence of an additional enzyme. The values are averages of three experiments, and the standard errors areless than 10% of the mean and, therefore, are not shown. Enzymes precipitated

incubated

Enzymes

GDH AspAT MDH !4

Crg

7 " GDH (100) KDH (200) GDH (loo),21 GDH (loo), AspAT (100) 73 MDH(loo), GDH (100) 7 11 GDH (loo), KDH (100) KDH (loo), MDH (loo),AspAT (100) GDH (loo), MDH (loo), AspAT (100) 86 GDH (loo), MDH (50), AspAT (50) 43 KDH(200),GDH(loo),MDH (501, AspAT (50) 7532 MDH (50), AspAT (50), KDH (200) MDH ( 2 5 ) , AspAT (25), GDH (100) 15 AspAT (50), GDH (100) 43 MDH (50). GDH (100) 8

Pg

Pg

31 -

-

-

30 27 10 20 3 10 -

8

-

20 18 4

dehydrogenase from the ternary complex (3). It was found that citrate (2 mM) also dissociated (decreased the amount bound 2-fold) glutamate dehydrogenase from the quaternary complex. Actiuation by Succinyl-CoA-Succinyl-CoA is a ligand common toglutamate dehydrogenase andthea-ketoglutarate dehydrogenase complex (Fig. 5). Succinyl-CoA increases the rate of oxidation of glutamate by glutamate dehydrogenase to the same extent as theallosteric activators leucine and ADP (Table 111), but has no effect on the reverse reaction, the reductive amination of a-ketoglutarate (Fig. 5). The activation constant (100 p ~ in) the presence of Mg2+(Table 111) is lower than the mitochondrial level of succinyl-CoA (0.3-1.4 mM) and in the range of its K; (140 PM) for citrate synthase (46, 47). Thus, it seems likely that succinyl-CoA could be a physiologically significant activator of glutamate deamination. Succinyl-CoA, like leucine and ADP, essentially abolishes inhibition by the allosteric inhibitors GTP and palmitoylCoA (Fig. 5, curves C and D ;Refs. 48 and 49).Consequently, succinyl-CoA,leucine, andADP produceaconsiderably greater enhancement of velocity in the presence than in the absence of these allosteric inhibitors. Mg2+is a weak inhibitor

22 16 0.8 3

(0.05 mg/ml), and aminotransferase (0.05 mg/ml), instead of competing with glutamate dehydrogenase, increased precipitation of all three enzymes (Table 11, line 8 uersus 9). The following calculations set an upper limit to the amount of glutamate dehydrogenase that could have been bound to that portion of the individual enzymes not included in the ternary complex between malate dehydrogenase, theaminotransferase, and the a-ketoglutarate dehydrogenase complex. At least 2/3 of the 32 pg of bound aminotransferase (line9) would have been associated with the a-ketoglutaratedehydrogenase complex. The remaining28 pg of free aminotransferase could have bound only 15 pg of glutamate dehydrogenase (Table 11, line 11), and even if all of the a-ketoglutaratedehydrogenase complex and malatedehydrogenase had been free, they would have only bound 21 pg (Table 11, lines 2 and 11). Since the sum (