A multilocus system for studying tissue and subcellular specialization. The pH and temperature dependence of the two major NADP-dependent isocitrate dehydrogenase isozymes of the fish Fundulus heteroclitus.

In the teleost fish Fundulus heteroclitus, there are three NADP-dependent isocitrate dehydrogenase isozymes. IDH-B2 is the only cytoplasmic isozyme, and IDH-C2 dominates the mitochondria of all tissues other than liver, where IDH-A2 is expressed. Since fish are ectotherms, their intracellular temperature and pH change directly with environmental temperature. In order to evaluate the influence of these environmental parameters on a model fish NADP-isocitrate dehydrogenase system, the major cytoplasmic (IDH-B2) and mitochondrial (IDH-C2) isozymes were kinetically evaluated as a function of pH and temperature. Whereas Vfmax and KmISOCm (where ISOC is isocitrate) were pH-independent, the Km for NADP was pH-dependent for both isozymes. The cytoplasmic isozyme (IDH-B2) had smaller KmNADP values between pH 7.0 and pH 8.0 than the mitochondrial form (IDH-C2). Vfmax and Km for substrate and coenzyme were temperature-dependent. Energy of activation for IDH-B2 and IDH-C2 was 10.6 and 12.8 kcal/mol, respectively. Both proteins had delta G not equal to values of about 15.8 kcal/mol, with significantly different distributions between delta H not equal to and delta S not equal to. The cytoplasmic isozyme (IDH-B2) appears to have a greater rate of catalysis than the mitochondrial enzyme (IDH-C2) at temperatures less than 30 degrees C. Moreover, the IDH-B2 isozyme had lower KmNADP values than the IDH-C2 isozyme at all temperatures, whereas the KmISOC values for the two isozymes were indistinguishable. Our data suggest that the two major NADP-dependent isocitrate dehydrogenase isozymes have unique physiological and metabolic functions that are adapted to the tissues and cellular compartments in which they are expressed.


A Multilocus System for Studying Tissue and Subcellular Specialization* T H E p H AND TEMPERATURE DEPENDENCE OF THE TWO MAJOR NADP-DEPENDENT ISOCITRATE DEHYDROGENASE ISOZYMES OF THE FISH FUNDULUS HETEROCLITUS
(Received for publication, April 17, 1986) Lucia Irene Gonzalez-Villaseiior$ and Dennis A. Powers5 From the McCollum-Pratt Znstitute and the Department of Biology, The Johns Hopkins Uniuersity, Baltimore, Maryland 21218 In the teleost fish Fundulucl heteroclitus, there are three NADP-dependent isocitrate dehydrogenase isozymes. IDH-B, is the only cytoplasmic isozyme, and IDH-C, dominates the mitochondria of all tissues other than liver, where IDH-A, is expressed. Since fish are ectotherms, their intracellular temperature and pH change directly with environmental temperature. In order to evaluate the influence of these environmental parameters on a model fish NADP-isocitrate dehydrogenase system, the major cytoplasmic (IDH-B2) and mitochondrial (IDH-C,) isozymes were kinetically evaluated as a function of pH and temperature.
Whereas VL, and K$socm (where ISOC is isocitrate) were pH-independent, the K , for NADP was pH-dependent for both isozymes. The cytoplasmic isozyme (IDH-B2) had smaller KZmP values between pH 7.0 and pH 8.0 than the mitochondrial form (IDH-CZ). Via, and K,,, for substrate and coenzyme were temperaturedependent. Energy of activation for IDH-B2 and IDH-C, was 10.6 and 12.8 kcal/mol, respectively. Both proteins had AG+ values of about 15.8 kcal/mol, with significantly different distributions between A H + and AS+. The cytoplasmic isozyme (IDH-B,) appears to have a greater rate of catalysis than the mitochondrial enzyme (IDH-C,) at temperatures less than 30 OC.
Moreover, the IDH-B, isozyme had lower KZmP values than the IDH-C, isozyme at all temperatures, whereas the KAsoc values for the two isozymes were indistinguishable. Our data suggest that the two major NADPdependent isocitrate dehydrogenase isozymes have unique physiological and metabolic functions that are adapted to the tissues and cellular compartments in which they are expressed.
In the teleost fish Fundulus heteroclitus, there are three NADP-dependent isocitrate dehydrogenase isozymes (isocit-* This research was supported by Grants DEB-791221 and DEB 82-0-7006 from the National Science Foundation. This is contribution 1317 from The Department of Biology. 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" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Supported by the Institute of Marine Science and Limnology at the University of Mexico (Universidad Nacional Autonoma de Mexico) and partially by the National Council of Science and Technology (Consejo Nacional de Cienciay Tecndogia), Mexico. Presented the material in this paper as part of the thesis submitted to the School of Sciences, National University of Mexico, Mexico City, D. F., Mexico, in partial fulfillment of the requirements for the Degree of Doctorate of Sciences.
To whom correspondence should be addressed.
rate:NADP+ oxidoreductase (decarboxylating), EC 1.1.1.42) that are coded for by three independent loci (1). The three isozymes, IDH-A,, IDH-B, and IDH-C,,' are thought to have been derived from gene duplication and subsequent evolutionary divergence and specialization of gene function (1). The isocitrate dehydrogenase isozymes exhibit differences in tissue and subcellular distribution (l), suggesting that they have significantly diverged and adapted through evolutionary processes toward the specific metabolic needs of the tissues in which they are expressed. Recently, these enzymes were purified to homogeneity and characterized in order to assess their potential physiological roles in different tissues and cell compartments (2). The results of those studies have shown that these multilocus enzymes are structurally and functionally nonequivalent. Whereas kinetic studies need to be done in the reverse direction, the differences observed between the kinetic parameters of the liver cytoplasmic isozyme (IDH-B,) and mitochondrial isozyme (IDH-A,) for the forward reaction were consistent with the shuttling of reducing equivalents (NADPH) into the liver cytoplasm, as suggested by Kaplan (3). However, the role of the mitochondrial isozyme found in all other tissues (IDH-C2) is less well defined. Since the IDH-B2 cytoplasmic isozyme is present in most tissues and the dominant mitochondrial form in all tissues other than liver is IDH-C, (4), they were selected as models for further studies to examine the physiological significance of tissue specificity and subcellular specialization.
Fish are cold-blooded organisms (i.e. ectotherms) subjected to external environmental temperature fluctuations that directly affect their internal temperature and pH (5-7). These parameters have a profound effect on protein structure and function (5)(6)(7). Since the fish F. heteroclitus is subjected to large environmental temperature fluctuations (8) and the kinetics of many enzymes are affected by changes in both temperature and pH (9), the effects on these variables on the rate of isocitrate oxidative decarboxylation (forward reaction) were studied as a model to probe the role of these environmentally important parameters for the two predominant multilocus NADP-isocitrate dehydrogenase isozymes (IDH-B, and IDH-C,). Whereas the liver mitochondrial isozyme (IDH-A,) may be involved in the shuttling of reducing equivalents into the cytoplasm (see Ref. and a previous report on the NADP-dependent isocitrate dehydrogenase isozymes (l), they are designated IDH-A2, IDH-B2, and IDH-C2, respectively. reducing equivalents for energy production. Thus, this study provides insight concerning the biochemical basis of the subcellular localization and tissue distribution of the IDH-B, and IDH-C, isozymes.

EXPERIMENTAL PROCEDURES~
The purification, physical characterization, and general kinetic procedures for the forward reaction at 25 "C and pH 7.4 for the IDH-B, and IDH-C, isozymes are described in a previous paper (2).
Kinetic Procedures-The following buffers were used for pH studies at 25 "C and constant ionic strength: 33 mM imidazole chloride was used for kinetic studies at pH 6.5 and 7.0; 30 mM triethanolamine chloride (concentration in the anion) was used for kinetic studies at pH 7.4-8.0; and 30 mM Tris chloride was used for kinetic studies at pH 8.5-9.0. Temperature studies were carried out with 30 mM triethanolamine chloride at pH 8.0.
Buffers were chosen with a pK. that would allow for any ApK. that might occur due to temperature changes and were prepared at constant ionic strength (10) (see Table I). Appropriate standardizations of buffers, substrate, and coenzyme were carried out (see Miniprint).
The isozymes were diluted in 30 mM triethanolamine chloride buffer, pH 7.4, containing 1% (w/v) bovine serum albumin. No loss in activity was observed for dilutions kept on ice for 8-10 h.
Velocity measurements (ABlO/min) for the forward reaction were made by steady-state kinetics using a Beckman Acta cIII UV-visible, dual-beam spectrophotometer with an automatic sampling system which monitors four reactions sequentially. The change in absorbance at 340 nm for each 3-ml cuvette was recorded every 15 s and converted to digital output by a Beckman 52700 teletype intercoupler interfaced to the spectrophotometer. The enzyme reactions were kept at constant temperature (k0.5 "C) with a Haake T52-T33 circulating constant temperature bath connected to the water jacket of the spectrophotometer sample compartment. Velocities were obtained by an unweighted least square analysis. Quadruplicate assays were averaged for each velocity. The kinetic parameters, K , and V,,,, were calculated using the parametric weighted regression method of Cleland (11). Statistical tests and analysis of the kinetic constants were performed by standard procedures (12,13).  where v is the initial rate, V, , , is the maximum velocity, [A] is the concentration of isocitrate or NADP in the forward direction, and K, is the Michaelis constant for either NADP or isocitrate.
Temperature Studies-For kinetic studies at constant pH 8.0 and variable temperature, K,,, and V, , values for isocitrate and NADP were obtained at saturating concentrations of either coenzyme or substrate while changing the concentration of the other variable at 10,15, 20, 25, and 30 "C. Reaction mixtures contained 30 mM triethanolamine chloride buffer, 1.5 mM MnS04, and 2 mM isocitrate when NADP was varied and 100 p~ NADP when isocitrate was varied. In all cases, the reactions were initiated by addition of enzyme.
Data obtained from temperature studies were fitted in Equation 2 above.
The energy of activation (E.) was determined from an Arrhenius plot, i.e. log V, , The apparent E. value was calculated from the slope of the plot (i.e. slope = -Ea/2.303R) by linear regression analysis (15).
The thermodynamic quantities, AH* (enthalpy) and AS* (entropy), of the activated state were calculated from a modification of the linear Arrhenius equation (15), where k b is the Boltzman's constant and h is the Plank's constant.
The data from the plot V,.,/T (K-') versus 1/T (K-') were analyzed by the least square method. The free energy of activation (AG' ) was estimated from Equation 4.
In a rapid equilibrium system, like the NADP-dependent isocitrate dehydrogenases (17)(18)(19)(20)(21)(22), K, values may approximate the dissociation constants between enzyme and substrate (i.e. K,,, G K.) (15). Under such conditions, the apparent dissociation constants (K,) for isocitrate or NADP can be used to calculate apparent free energies, enthalpies, and entropies of substrate binding (9). If the Arrhenius plots are linear, the following equation can be employed, where K " , is the appropriate K, for either isocitrate or NADP when the other substrate is saturating but not inhibiting (4), AHH. is the enthalpy of substrate binding, AS, is the entropy of substrate binding, R is the gas constant, and T is absolute temperature. The free energy of substrate binding (AG,) was estimated from the following equation.

RESULTS AND DISCUSSION
pH Studies-pH can dramatically affect enzyme function (9,14). A shift in pH may change the ionization of substrates and/or the ionization of amino acid side chains that directly or indirectly affect catalysis. pH is particularly important for cold-blooded organisms since their intra-and extracellular pH is directly affected by the external environment. The study of pH in relation to enzyme catalysis provides insight concerning these possibilities.
Initial studies at low temperatures (10 "C) proved difficult employing the matrix approach (see "Experimental Procedures'') because low substrate and coenzyme concentrations yielded initial velocities at 10 "C that approached background variation of the spectrophotometer. Whereas increasing the path length of the reaction cuvette improved our measurements, this approach did not eliminate the problem. Varying only one substrate at a time allowed accurate measurements at 10 "C; thus, we adopted the single variable approach for our pH and temperature studies as described under "Experimental Procedures." Since we had used the matrix approach at p H 7.4 and 25 "C in an earlier study (2), we were curious when initial K, values obtained by the single variable approach were slightly higher than those obtained by the matrix approach while the V,,, values were identical. In order to gain insight concerning these differences and to ensure that any K,,, and V,,, trends in relation to pH would be independent of the method of calculation, we studied the effect of pH on the various kinetic parameters, Vi,,, KEADP, and KEoc (where ISOC is isocitrate), by both the matrix and single variable approach at 25 "C. Those results and their statistical significance, or lack thereof, are tabulated in Tables I1 and 111. Whereas the absolute K, values appeared to be sensitive to the method of calculation, the trends in relation to experimental variables like p H were similar, no matter which way they were calculated (Figs. 3 and 4). The differences between the methods probably arise because: (a) the single variable method only requires the extraction of two constants, whereas the matrix approach must extract four constants; ( b ) the single variable method employs weights and thus minimizes variation due to error at low substrate and coenzyme concentration, whereas the matrix approach most commonly employed (2, 16) does not normally allow weighting of data; (c) the matrix approach appears to minimize K, differences by absorbing variation in the data set in the K, term (see well as other equations to determine the most parsimonious approach to analyzing such kinetic data. However, it should be emphasized that the kinetic trends in relation to variables like pH and temperature appear to be independent of the method of calculation. On the other hand, differences in the method of calculation may account for numerical differences in K, estimates between various studies reported in the literature. Also, it should be clear that data collected using a number of different variables like pH, ionic strength, temperature, etc., should be collected and analyzed by the same method (ie. either matrix or single variable) in order to generate a self-consistent data set that will allow appropriate comparisons across variables. Fig. 3A and Fig. 4A illustrate the KEoc in relation to pH for the IDH-B2 and IDH-C, isozymes as determined by both the single variable and matrix approaches. Of the six pH values, there were three KEoc values that were significantly different between the isozymes. The three values that were different using the matrix approach were pH 6.5, 7.4, and 9.0, whereas pH 7.0, 8.0, and 8.5 were different employing the single variable method (Tables I1 and 111). Since the differences were not consistent between calculation methods, we shall be conservative and conclude that no differences in Kfoc exist between the isocitrate dehydrogenase isozymes. The matrix and single variable methods both indicated no significant pH dependence of KEoc. This point is obvious upon perusal of the KEoc values in Tables I1 and 111. Since Kfoc is apparently pH-independent, the ionizable groups on the free enzyme and/or isocitrate do not affect enzyme-substrate affinity or the effect is complementary and opposite, i.e. no net pH effect. This phenomenon has been observed for arginase (23).
Whereas KZoc does not vary with pH, Fig. 3B illustrates that KEADP does change in relation to pH. Inspection of the Kf;IADP estimates (Tables I1 and 111) indicates a pH dependence for both isozymes, with apparent optima around pH 8.0. Moreover, the differences are consistent between calculation methods, i.e. matrix and single variable. The KZADP values appear to converge at pH 6.5 and 9.0 for both isocitrate dehydrogenase isozymes (Fig. 3B). Since the ionization of NADP is the same for both isocitrate dehydrogenase isozymes, differences between IDH-B, and IDH-C2 in the Kf;IADP pH dependence suggest different ionization of groups on the free enzymes which control the binding of the coenzyme. The pK values of the major ionizable groups on NADP are pK, = 3.7 and pK2 = 6.1 (24). Of these, only one phosphate titrates over the experimental pH range. The increase in Kj;iADP (Fig. 3B) at low pH may be partially the result of the protonation of NADP. The increase in KEADP at high pH probably results from the deprotonation of a basic amino acid side chain on the free enzyme, which, in turn, affects the cofactor binding. The divergence of the two KZADP values (Fig. 3B) over the middle pH range suggests different ionization states of amino acid residues on the isozymes, whereas convergence a t high and low pH suggests the deprotonization of similar residues.
The maximal rate of catalysis ( V i a x ) was pH-and calculation method-independent (Fig. 3C). The respective Viax values were significantly different between the two isozymes (see Tables I1 and 111). These Vi,, differences may reflect either different kcat values or a difference in active enzyme concentration ([EO]) because Vmax = kc,, [E,,]. We shall return to this point below.
Temperature Studies-Temperature is particularly important for cold-blooded organisms because it has a direct effect on protein structure and function and induces changes in enzyme kinetics. The K,,, values for isocitrate and NADP in this study are temperature-dependent (Figs. 5 and 6). The K F values increased with temperature (Fig. 5), but there were no significant differences between the isocitrate dehydrogenase isozymes at 25 "C (Table IV). On the other hand, there were significant differences in the K2ADP values between the isozymes at all temperatures (Table IV), with IDH-B2 having the lower KN,ADP (Fig. 6). Figs. 5 and 6 are linear which supports the use of K,,, values as apparent dissociation constants ( K a ) for enzyme-substrate complexes. In a rapid equilibrium system, like that of the NADP-dependent isocitrate dehydrogenase (17-22), the dissociation constants are usually in the range of to IO-' M so that AG, is in the range of -7.7 to -2.6 kcal/mol at 25 "C. The apparent dissociation constants for the NADP-isocitrate dehydrogenase isozymes (KzADP and ICEoc) are within this range (Tables 11-IV), and free energies (AG,) calculated from these apparent dissociation constants also fall within the expected range (e.g. -7.3 to -7.7 kcal/mol at 25 "C).
The maximal rate of isocitrate oxidative decarboxylation (V' ,,,) for the NADP-isocitrate dehydrogenase isozymes resulted in linear Arrhenius plots (Fig. 7), suggesting that there has been no change in mechanism over the temperature range studied and that a single rate constant is driving the catalytic process for both isozymes. Moreover, the slopes (Fig. 7) suggest that the Vi,, differences between isozymes result from    15.8 f 0.9
unique kc,, values, otherwise the slopes would be parallel. Table 1V indicates that the Vi,, values between 10 Table V shows the energy of activation values (E,) calculated from the slope of Fig. 7, as well as the thermodynamic activation parameters, A H + , ASf, AG+, and Qlo for both isozymes. Whereas differences in the E,,, Qlo, A H f , and AS+ parameters were observed between the two isocitrate dehydrogenase isozymes, the free energies of activation (AG+) were not statistically different. This finding appears, at first, to be contradictory to the above statement that kcat values may vary between the isocitrate dehydrogenase isozymes. This arises from the fact that the calculation of AGf includes a very large constant such that any significant differences in kcat can be statistically eclipsed in the propagation of errors and the logarithmic nature of the equations for the calculation of AGf. The maximum differences between the isocitrate dehydrogenase isozyme kcat values are on the order of 40%. However, a 2-fold difference would only result in a AG+ change of about 400 cal which would be buried by the propagation of errors when estimating AG+. It would take almost an order of magnitude difference in kc,, in order to detect a significant difference in free energies ( AG+) above a background error of 5%. Since Vi,, values differ significantly between the isocitrate dehydrogenase isozymes, it follows that the respective AGf values must also differ, but, due to the propagation of errors, the logarithmic nature of the equations, and the loss of degrees of freedom, these differences in Viax are eclipsed during the free energy calculations.
Some Physiological Roles of the NADP-Isocitrate Dehydrogenase Isozymes-Mitochondria contain two or more separate isocitrate dehydrogenases. One is NAD-linked and one or two are NADP-linked enzymes. However, the cytosol only contains and NADP-linked isocitrate dehydrogenase (see Refs. 25 and 26 for reviews). NAD-isocitrate dehydrogenase is a nonequilibrium pathway, whereas the NADP-dependent isocitrate dehydrogenase pathway is in equilibrium (27). Flux through the NAD-linked pathway is activated only under special conditions relating to the needs of the Krebs cycle. On the other hand, the NADP-linked isocitrate dehydrogenase performs the majority of the isocitrate oxidation and provides the mibochondrion with both NADPH and NADH when coupled to the transhydrogenase system (28). Weigl and Sies (29) have shown that mitochondrial NADPH reducing equivalents, in the form of isocitrate or citrate, may be provided to the cytoplasmic space for biosynthetic reactions and during metabolic limitations on the pentose phosphate pathway. In fact, Thurman and Scholz (30) and Weigl and Sies (29) have suggested that the major supply line for NADPH utilization at the endoplasmic reticulum is the export of citrate from mitochondria as well as NADPH generation by cytosolic NADP-linked isocitrate dehydrogenase.
Kaplan (3) suggested that the cytoplasmic NADP-isocitrate dehydrogenase is kinetically favored over the mitochondrial form for isocitrate oxidation and thus favors NADPH shuttling into the cytoplasm. However, there are several important factors that must be considered, such as: (a) the kinetic parameters for the reverse direction (i.e. K:Oz, Kf;IADPH, FAKG, and VL,., where aKG indicates a ketogluterate), ( b ) the substrate and cofactor inhibition constants, (c) the concentration of substrates and products in the cytoplasm and mitochondria of various tissues, ( d ) the role of NAD-isocitrate dehydrogenase in various tissues, (e) the metabolic linkages to other enzymes involved in the shuttling of reducing equivalents across the mitochondrial membrane (e.g. the malate dehydrogenase isozymes, the aspartate aminotransferase isozymes, the malic enzyme, etc.), (f) the differential transport of substrates across the mitochondrial membrane and the temperature and pH sensitivity of that transport, and (g) the relative concentration of the various NADP-isocitrate dehydrogenase isozymes in different tissues.
We have presented kinetic data in the forward direction previously (2) for the IDH-A, and IDH-B2 isozymes that are consistent with Kaplan's (3) concept for liver tissue. The fractional distributions of these isozymes (Table VI) are also consistent with the greater catalytic efficiency for the IDH-B, isozyme in liver. However, Table VI illustrates that the IDH-C, isozyme is the major NADP-isocitrate dehydrogenase in heart and brain and the only detectable NADP-isocitrate dehydrogenase form in white muscle. Whereas the VLaX/ K:ADP values for IDH-B2 and IDH-C, differ by a factor of 2, in red muscle and brain, these catalytic differences would be offset by the greater concentration of the IDH-C2 isozyme. Moreover, in white muscle, the only detectable NADP-isocitrate dehydrogenase is the IDH-C2 isozyme. Since IDH-C, is not present in liver, its role in the alleged NADPH shuttle must be very limited. These data are consistent with those of the literature. For example, Hogeboom and Schneider (31) and others (32, 33) have found that tissue such as liver, with high cytosolic demand, has 80% of its NADP-isocitrate dehydrogenase in the cytoplasmic form; whereas red muscle and heart, with low cytosolic NADPH demand, have 90% mitochondrial NADP-isocitrate dehydrogenase. Thus, the major role of IDH-C, must be to provide reducing equivalents in the form of NADPH or NADH when coupled into the transhydrogenase system to the electron transport chain. This point is emphasized by the results illustrated in Table  VII, wherein it is clear that NAD-isocitrate dehydrogenase makes up only 6-7% of the total isocitrate dehydrogenase activity in any given tissue. Since white muscle isocitrate dehydrogenase activity is 93% IDH-C2 (see Tables VI and  VII), it must supply the major oxidation of isocitrate and thus reducing equivalents for energy production. These results are consistent with those of Alp et al. (34), who surveyed tissues of numerous vertebrates and found that the vast majority of isocitrate dehydrogenase activity resided in the NADP-dependent isocitrate dehydrogenase enzyme.
These data, together with studies presented here and elsewhere (2, 3), lead us to conclude that the IDH-C2 isozyme is significantly involved in energy production, whereas IDH-B2 is involved in supplying cytosolic NADPH reducing equivalents for biosynthetic reactions.