Physicochemical properties of isocitrate dehydrogenase from lactating bovine mammary gland: Effect of substrates and cofactors

ARCHIVESOFBIOCHEMISTRYAND BIOPHYSICS Vol. 274, No. 2, November 1, pp. 453-462,1989

Physicochemical Properties of lsocitrate Dehydrogenase from Lactating Bovine Mammary Gland: Effect of Substrates and Cofactors VIRGINIA

L. SEERY’ AND HAROLD

M. FARRELL,

JR?

US. Department of Agriculture, Eastern Regional Research Center, 600East Mermaid Lane, Philadelphia, Pennsylvania 19118 Received March 27.1989, and in revised form June 30,1989

The influence of substrates and cofactors on the oligomeric structure of the cytosolic form of NADP+-specific isocitrate dehydrogenase (IDH) from lactating bovine mammary gland was investigated using analytical ultracentrifugation and kinetic methods. In guanidine-HCl, the monomer molecular weight for reduced and carboxymethylated IDH was found to be 50,000 to 52,000. In nondenaturing solvents IDH behaves as a homogeneous solute with a molecular weight of 97,200. When added separately, manganous isocitrate, isocitrate, manganous citrate (substrate analog), and a mixture of the substrate analog and NADP+ do not significantly alter the sedimentation coefficient or the molecular weight of IDH as judged by direct observation of the enzyme at 0.1 to 3 PM using sedimentation velocity and equilibrium. Active enzyme sedimentation (AES) was used to assess the degree of dissociation of IDH at lower concentrations, and Kd for the dimer-monomer equilibrium was estimated to be 2 nM. In enzymatic studies, the specific activity at several levels of substrate does not vary as the subunit concentration of enzyme is reduced from 10 to 0.3 nM. Estimates for Kd by AES indicate the presence of a significant fraction of monomer at assay concentrations of 1 nM and below, where the weight fraction of monomer is predicted to be 0.6. If the monomer has a lower activity than the dimer, a drop in specific activity is expected below 1 nM. Significant decreases occur only when the IDH is not protected from denaturation. The concentration of cytoplasmic IDH in bovine mammary tissue is estimated to be 5.7 PM, at least loo-fold greater than our estimates of Kd. Since over 90% of the enzyme is present in the dimeric form, ligand-induced changes in aggregation state cannot play a significant role in the regulation of the cytosolic form of IDH in situ in this tissue. 8 1989 Academic Press, Inc.

The cytosolic isozyme of NADP+-specific isocitrate dehydrogenase [three-D,-isocitrate:NADP+ oxidoreductase (decarboxylating); EC 1.1.1.421 is a major source of reducing equivalents in the form of NADPH for biosynthetic reactions in several animal tissues including mammary gland, adrenal gland, and liver. In lactating ruminant mammary tissue isocitrate dehydro-

genase (IDH)3 provides NADPH for fatty acid and cholesterol synthesis (1, 2). The mechanisms which control the activity of the enzyme are unknown. Unlike the NAD+-requiring form which appears to be lacking in mammary gland (2), NADP-specific IDH is not allosterically regulated. One of the possible mechanisms for regulation of NADP-specific IDH is a ligand-mediated change in polymerization of the enzyme as has been suggested by Carlier and Pantaloni (3) and Kelly and Plaut (4) for

1Current address: Department of Biochemistry, Philadelphia College of Osteopathic Medicine, Philadelphia, PA 19131. aTo whom correspondence should be addressed.

3 Abbreviations used: IDH. isocitrate dehydrogenase; SDS, sodium dodecyl sulfate. 453

lw.s9861/~Y Copyright All rights

$3.00

0 1989 by Aeademie Press, Inc. of reproduction in any form reserved.

454

SEERY AND FARRELL

the cytosolic and mitochondrial isozymes, respectively. The fundamental subunits of all these enzymes have molecular weights in the range of 48,000-55,000 (3,5,6). Conflicting data have appeared in the literature regarding the aggregation state of these enzymes which have been reported to exist either primarily as monomers (7, 8) or as dimers (3, 9, 10) depending on the presence or absence of substrates and other effecters and upon the conditions chosen for physical measurements. Addition of the metal-isocitrate complex was shown either to have no effect on dimerization (6,10) or to favor this reaction (3, 8). In this report we use analytical ultracentrifugation and kinetic methods to investigate the relationship between quarternary structure and the binding of substrates and cofactors by bovine mammary gland IDH. The association behavior of the enzyme is compared to that of other isozymes of IDH. The implications of our results for the control of the enzyme in mammary gland are discussed. MATERIALS

AND METHODS

Materials. NADP+-specific IDH was isolated from frozen bovine mammary glands as previously described (5). All coenzymes, substrates, and biochemicals were purchased from Sigma Chemical CO.~Tris base (ultrapure) and guanidine hydrochloride (ultrapure) were products of Schwarz/Mann. All other chemicals were reagent grade. Preparation of samples for physical measurements. Samples of IDH were dialyzed for 12 to 16 h at 4°C vs a buffer and filtered through Nucleopore membrane filters (0.4 Mm) before use. Concentrated ligand solutions were added to a measured volume of the protein solution from a microsyringe; total dilution did not exceed 4%. Protein concentration was determined spectrophotometrically using the absorbance index (l%, 1 cm) reported in this study (vide infra). Molar concentrations are expressed in terms of the subunit molecular weight of 55,000 (5). IDH was reduced and alkylated by the method of Schechter et al. (11); amino acid analysis showed >96% conversion to Scarboxymethyl cysteine. Absorbance index. The refractive index increment of an IDH solution at ca. 20 fiM in a buffer containing

4 Reference to a firm or brand name does not constitute endorsement over similar products not mentioned.

0.02 M Tris-Cl and 0.08 M KC1 at pH 7.4 was measured in a Brice-Phoenix differential refractometer (Model BP-2000-V) at 436 nm and 25°C. The protein solution was diluted gravimetrically to ca. 10 pM and its absorbance at 278 nm determined in a Perkin-Elmer Lambda 7 uv/vis spectrophotometer. The absorbance index was calculated from An assuming a value of 0.19 ml/g at 436 nm for dn/o!c of a typical globular protein (12). lX?acentwe techn@es Ultracentrifugation was performed in a Spinco Model E analytical ultracentrifuge equipped with an electronic speed control and photoelectric scanning system. Double-sector centerpieces were used for both sedimentation velocity and sedimentation equilibrium. All cell components had orientation marks so that the cell could be reproducibly assembled. High-speed sedimentation equilibrium as described by Yphantis (13) was used for the determination of molecular weight. During the time required for the attainment of equilibrium (16 to 20 h), the activity of control samples of IDH stored at room temperature remained essentially constant. The output of the photoelectric scanner was interfaced with a Mod Comp III computer. The computerscanner interface and the methods used to handle data for both equilibrium and velocity experiments have been described by Kumosinski et al. (14). Data from two to three equilibrium scans were combined by an averaging program. The equilibrium concentration distribution was fitted directly by an iterative procedure based on gauss-Newton nonlinear regression analysis to obtain the weight average molecular weight. The partial specific volume, G, of mammary gland IDH was assumed to be equal to the apparent specific volume of 0.728 ml/g which was determined by Carlier and Pantaloni (3) for IDH from beef liver at 7.5 g/liter. A correction factor of di2dT = 0.0005 ml/g per degree (15) was applied to V, when the temperature was below 20°C. For the ultracentrifuge experiments in guanidine HCl at 24,000 rpm, reduced and S-carboxymethylated IDH was dissolved at 0.3 g/ liter in 6.3 M guanidine HCl containing 0.05 M TrisCl, and 0.05 M KC1 (pH 7.4) and dialyzed against this solvent for 2 days at room temperature. Other experimental details are given in the table footnotes. Active enzyme sedimentation (16) was carried out at 20°C and 56,000 rpm using a Vinograd-type band forming centerpiece. The standard assay mixture described below was modified by increasing the concentration of NADP* to 0.22 mM (20 K,) and by adding 0.1 M KC1 and 0.1 mM dithiothreitol. The activity of IDH is unaffected by the inclusion of 0.1 to 0.2 M KC1 in the assay mixture (5). Formation of NADPH was monitored at 340 nm. Sedimentation coefficients were calculated both from the inflection point of the boundary and from the maximum ordinate of the difference curves for successive scans (17). Observed sedimentation coefficients were corrected to standard

BOVINE MAMMARY

GLAND ISOCITRATE

455

DEHYDROGENASE

Using the value of 0.19 ml g-l for dn/dc of a typical globular protein, an absorbance SEDIMENTATIONCOEFFICIENTOFIDH ASA FUNCTION index (O.l%, 1 cm) of 1.37 + 0.03 is calcuOFINITIAL PROTEINCONCENTRATIONS (c,,)* lated. A value of 1.29 for the absorbance index (O.l%, 1 cm) was reported for beef Co(FM) %o,w6% liver IDH (3). 2.8-3.2 5.66(0.07)* Sedimentation velocity studies. The sedi1.2 5.53 mentation coefficient of IDH was found to 5.67 0.7 be -5.7 S and to remain essentially constant over the range of initial protein cona Samples were centrifuged at 48,000 rpm at 15°C centrations from 1 to 3 PM (Table I). Homoin solvent containing 0.02 M Tris-Cl, 0.08 M KCl, and geneity of the enzyme preparations was in0.1 mM EDTA (pH 7.4). dicated by two criteria. Values of szO,w *The number in parenthesis represents the stancalculated from the inflection point dard deviation of the average of three measurements. method were either identical to those of Table I (equivalent boundary method) or differed by a maximum of 2%. The shape conditions using the viscosity and density of the sol- of the sedimenting boundary was fitted vent. with the integral of a single Gaussian Sedimentation velocity and sedimentation equilibfunction (RMS error < 0.005). The molecurium experiments were interpreted by assuming that lar weight calculated from a combination the aggregation properties of IDH can be described by a rapid interconversion between a monomer and a of the sedimentation coefficient and the Stokes radius of 4.1 nm (5) is 102,000.Table dimer. Data at concentrations c 3 pM and the II presents the sedimentation coefficient of equations of Gilbert and Gilbert (18) were used to the enzyme in the presence of various lipredict sedimentation coefficients for different values gands. The additions include the substrate of the monomer-dimer association constant. Steady-state activity meo.surements. IDH activity (manganous isocitrate) at saturating conwas measured at 25°C by following the production of centration, the substrate analog (mangaTABLE I

NADPH at 340 nm in a Gilford Model 252 recording spectrophotometer equipped with a thermostated cell block (5). Absorbance scales of 0 to 0.05,O to 0.1, and 0 to 0.2 were used. Assay mixtures contained 0.1 M Tris-Cl buffer at pH 7.4,0.06 to 0.11 mM NADP, and other components as described in the legends to the figures. The association constants for the manganous ion complexes were taken as 1.15 X lo3 Mm’for isocitrate, 3.7 M-i for chloride (19), 5.2 X 103M-’ for citrate (20), and 467 M-’ for NADP (21). The enzyme was diluted with buffer (pH 7.4) containing 0.05 to 0.1 M Tris-Cl, 0.1 mM dithiothreitol, and 0 to 15 pM albumin using vials and micropipet tips made of polypropylene. Specific activity is expressed as millimoles of NADPH per minute per gram of protein. RESULTS

Absorbance index of IDH. To convert absorbance data obtained from ultracentrifuge experiments to concentration in grams per liter, an accurate value of the absorbance index of IDH is required. IDH exhibits an absorption maximum at 278 nm. A protein solution with an absorbance of 1.276 at 278 nm had a refractive index increment of 1.77 + 0.05 X low4 at 436 nm.

TABLE II SEDIMENTATIONCOEFFICIENTOFIDH IN THE PRESENCEOFLIGANDS Addition

Concn (mM)

Sedimentation velocity” None Mn’+-D-isocitrate Mn’+-citrate Mn’+-citrate + NADPH NADPH Active enzyme* Complete assay mixture

0.27 1.0 1.0 + 0.002 0.002 -

S2O.W

w 5.72 5.62 5.52 5.45 5.5 5.3

a The buffer contained 0.02 M Tris-Cl and 0.08 M KC1 at pH 7.4. The protein concentration was 3.2 pM. The samples were centrifuged simultaneously at 36,000 rpm and 15°C. * The experimental conditions are described under Materials and Methods. The loading concentration of IDH ranged from 13 to 34 nM for four experiments at 56,000rpm.

456

SEERY AND FARRELL

in good agreement with the estimate from s~,~ SUMMARYOFMOLECCJLARWEIGHTSFORIDHINTHE and the Stokes radius. The agreement between the observed data and the concenPRESENCEOFSUBSTRATE,SUBSTRATEANALOG,OR tration distribution calculated for a single MIXTURESOFNADP+ANDTHESUBSTRATEANALOG~ species with a molecular weight of 97,200 No.of Concn is illustrated in Fig. 1 for a typical experiexperiments Addition (mM) ikf, (10m3) ment on the native enzyme. RMS error associated with the fit was 0.002, indicating 8 Noneb 97+2 that the concentration distribution of IDH 1 Isocitrate 1.0 88 was close to that predicted for a homoge1 Mn’+-isocitrate 0.3 91 neous solute. A plot of In c versus r2 was 2 Mn2+-citrate 1.0 91f3 constructed from the data in Fig. 1; the 1 Above + NADP 0.01 92 slope of 2.340 f 0.008 corresponds to a mo1 Mn2+-citrate 5.0 91 lecular weight of 94,300 f 3000. The In c ’ Experiments were carried out at 15”C, 20,009 rpm versus r2 data were divided into three rein an analytical ultracentrifuge equipped with a scan- gions containing 38-54 points. Molecular ning uv-visible absorption system. The solvent con- weights of 93,000 f 3000 and 98,200 f 400 tained 0.02 M Tris-HCI, 0.08 M KCl, pH 7.4. were calculated from the meniscus and ’ One of these samples also contained 0.1 mM dithiobase regions of the ultracentrifuge cell, rethreitol; the result was not significantly different. spectively. These values differ by ~4% from that obtained from all the data. In another series of four experiments on IDH nous citrate), the reduced coenzyme, and a in which the centrifugation speed was mixture of the reduced coenzyme and sub- 17,000 rpm, a molecular weight of 94,000 strate analog. Values of .s~,~ in the pres- f 1000 was obtained (data not shown). All ence and absence of Mn2+-isocitrate as of the ligands listed in Table III including well as the values for all ligands overlap isocitrate, manganous isocitrate, mangawithin the standard error of the individual nous citrate at two levels, and a mixture measurements. The sedimentation coeffiof the substrate analog and the coenzyme cient of the native enzyme was not changed decreased the observed molecular weight by more than 0.26 S by any of these additions. At concentrations of l-3 PM the en- by less than 10%. The decreases in molecuzyme exists in a highly associated state lar weight are probably not significant. In with and without ligands. This result does any case dissociation of the dimer into monomer units is far from complete under not imply that the ligands are without effect. Factors including the magnitude of all conditions tested. The linearity of our the monomer-dimer association constant plots of In c vs r2 in the meniscus region of the ultracentrifuge cell at 0.1 to 0.4 I.LM limit the ability to detect ligand-induced changes in association properties (22). suggests that concentrations of ~0.1 PM Changing the angular velocity from 48,000 are required in order to observe appreciarpm (Table I) to 36,000 rpm (Table II) did ble dissociation of the dimer. The apparent homogeneity of mammary not alter the sedimentation coefficient, suggesting that IDH undergoes rapid self- gland IDH in sedimentation equilibrium indicates that the fundamental subunit association (23). has a molecular weight of about 48,000. Sedimentation equilibrium studies. The molecular weight of IDH was analyzed in Sedimentation equilibrium of reduced and S-carboxymethylated IDH in 6.3 M guanigreater detail over the concentration range from 0.1 to 8 PM by the technique of dine HCl yielded a value of 8270 + 40 for sedimentation equilibrium. The results ob- the quantity M (1 - Up). If the binding of tained at 20,000 rpm under various condi- guanidine HCl to IDH is negligible (V = 0.728), 1M,of the subunit is calculated to tions are shown in Table III. The molecular weight of the enzyme in the absence of spe- be 51,600 +-200. Assuming that Vis lowered cific ligands was found to be 97,000 + 2000 by 0.006, the average reduction in Gfound TABLE III

BOVINE MAMMARY

GLAND ISOCITRATE

DEHYDROGENASE

457

0.8

0.6 r 5

0.4

a 0.2

7.02 r(cm) FIG. 1. Sedimentation equilibrium of IDH. The enzyme at an initial concentration of 5 pM was centrifuged at 20,000 rpm in a buffer containing 0.02 M Tris-HCl, 0.08 M KCl, pH 7.4. The concentration distribution as a function of the radius is shown: observed data (++++) and simulated data for a molecular weight of 97,200kDa (-).

for eight globular enzymes in 6 M guanidine HC1(24), M, would be 49,500 + 200. Active enzyme sedimentation. The sedimentation coefficient of an enzyme can be determined at nanomolar concentrations by following the appearance of product using the technique of active enzyme sedimentation. Values for s~,,~ of 5.2 +_0.1 and 5.3 st 0.2 S were calculated for mammary gland IDH from the inflection point and difference curve methods for four experiments at 56,000 rpm in which the initial concentration of layered enzyme was varied from 13 to 34 nM. The difference curves exhibited Gaussian distribution with no skewness or indication of a second component. The trend toward a lower s,,, calculated by this technique is consistent with some dissociation of the dimer, as the enzyme concentration is reduced from the micromolar (Tables I and II) to the nanomolar range. Efect of storage on the aggregation prop erties of IDH. A species with a sedimentation coefficient of less than 5.7 S was observed after prolonged storage at -20°C in Tris-Cl buffer (50 mM) at pH 7.4 containing 0.15 M NaCl and 10% glycerol. Although the specific activity of a preparation stored

for 1 year was essentially unchanged, heterogeneity with respect to molecular weight was indicated by the lack of agreement between the sedimentation coefficients calculated by the inflection point (5.4 S) and the equivalent boundary method (4.2 S). RMS error associated with the fit of the sedimenting boundary to the integral of a single Gaussian function increased with time and was 0.01 when the boundary had passed the midpoint of the solution column. Weight average sedimentation coefficients are obtained from the position of the equivalent boundary and reflect the sedimentation of all species, while the migration of the inflection point of the boundary measures the sedimentation of the major component. Aged preparations of IDH clearly contain species smaller than the dimer and may also contain larger aggregates. The results of active enzyme sedimentation of aged IDH were somewhat variable with s~,~ of 4.0, 5.3, and 4.8 S obtained at 24,34, and 50 nM respectively. Specific activity versus IDH concentration. Ligand-mediated, self-associating enzyme systems can be studied at low concentrations (nM) by the method of Kurganov

458

SEERY AND FARRELL

0.41 0

I 0.2

I I I I I 0.4 0.6 0.8 1.0 1.2 ENZYME CONCENTRATION (nM)

I 1.4

FIG. 2. Albumin protects IDH from nonspecific denaturation. An enzyme solution (4.6 PM) in 0.1 M Tris-Cl, 0.1 mM dithiotheitol at pH 7.4 was diluted in the same buffer with (0) and without (A) the addition of 15 FM albumin. Aliquots (10 ~1) of the diluted enzyme at 6 to 368 nM with respect to the subunit were added to 2.7 ml of the reaction mixture containing 1.5 mM DL-isocitrate, 1.7 mM MnS04, and 0.06 mM NADP. The experimental points are averages of at least two assays.

(25) which has been applied to cytoplasmic (26) and mitochondrial (4) forms of IDH. For a monomer-dimer equilibrium where binding of substrate to the monomer promotes formation of the more active dimer, specific activity decreases as the enzyme and substrate are diluted to concentrations which favor dissociation. Before applying the procedure of Kurganov to mammary gland IDH, conditions which would stabilize the enzyme during dilution were established. Many enzymes including IDH can be denatured by dilution to low concentration (0.1 PM). Inactivation can be prevented if a carrier protein such as albumin is added to the dilution buffer. Figure 2 presents a comparison of the specific activity at saturating levels of manganous isocitrate as a function of enzyme concentration for dilutions of IDH in the presence (0) and absence (A) of 15 PM albumin. The dilution buffer had an ionic strength of 0.05 M and contained 0.1 InM dithiothreitol. Following dilution the activities of the stock enzyme solutions (46 nM) were stable for 2 to 3 h at 0°C during the course of the experiments. Solutions of lower protein concen-

tration were prepared at the time of assay. Final dilution of IDH and albumin in the substrate mixture was 270-fold for all assays. In the presence of albumin (56 nM) the specific activity of IDH is constant from 0.2 to 1.3 nM of enzyme subunits. Activity is a function of enzyme concentration and falls rapidly below 1 nM when dilutions are made without albumin. Activity at the lowest enzyme concentration (0.05 nM) falls to 50% of the maximum value with an inflection point for the activity loss occurring at ca. 0.7 nM. Using manganous DL-iSOCitrate (0.69 mM) instead of albumin in the dilution buffer prevented some of the activity loss; relative activity dropped from 0.86 to 0.60 over the range of enzyme concentration from 0.34 to 0.02 nM (data not shown). Albumin has no direct influence on the activity of IDH. This was demonstrated by preparing two samples of IDH (92 nM) from the same stock solution (11 PM), one with and the other without 7.5 j&M albumin. After dilution to 0.34 nM of IDH in the assay, the sample without albumin was only 60% as active as the sample containing a final concentration of 28 nM

BOVINE MAMMARY

GLAND ISOCITRATE

459

DEHYDROGENASE

55 0

T

0

45

- 1 A

: 5 I= 35 :: " ii 3 itm 25

0

~

T I A 0 0

0

1

0 Oo

1 6

A

T 1

0

15L 0

0

I

I

I

I

I

2

4

6

8

IO

ENZYME

CONCENTRATION

12

(nM1

FIG. 3. Specific activity as a function of isocitrate dehydrogenase concentration at several levels of manganous isocitrate. A stock enzyme solution (2.8 pM) in 0.1 M Tris-Cl, 0.1 mM dithiothreitol at pH 7.4 was further diluted with the same buffer containing 7.5 j.~tMalbumin. Aliquots of the enzyme (10 pl) were mixed with 2.7 ml of the reaction mixture. Enzyme concentrations are expressed in terms of the subunit molecular weight. The concentrations of free Mn’+, calculated as described under Materials and Methods, were 0.75 PM at the highest substrate level (0) and 0.08 pM for all other levels; concentrations of manganous D-isocitrate were 0.35 mM (0), 15 pM (A), 1.5 pM (O), and 0.75 pM (0). Experimental points are averages of at least two determinations. The lines represents the average of all points at the same substrate concentration; the standard deviations are indicated.

albumin. The inclusion of albumin (28 nM) in the substrate mixture failed to restore the activity of diluted samples which did not contain albumin prior to the assay. In the next series of experiments activity was determined as a function of IDH concentration at several levels of manganous D-isocitrate (350 to 0.75 PM). The selected levels represent the following multiples of Km: 115,5,0.5, and 0.25 (5). The final concentration of albumin in the assay mixtures was 28 nM for these assays. As illustrated in Fig. 3, specific activity is independent of protein concentration at each level of substrate. DISCUSSION

In the range of protein concentration which we examined by direct ultracentrifuge methods (0.1 to 8 PM), mammary gland IDH essentially behaves as a homogeneous solute with a molecular weight of

97,200. The range of 49,500 to 51,600 for the size of the fundamental subunit in guanidine HCl is compatible with a dimeric structure for the native enzyme. Due to the uncertainties regarding the binding of guanidine HCl, the most accurate value for the molecular weight of the monomer is one-half that of the dimer or 48,500. This value is slightly lower than estimates of 55,000 given by SDS-gel electrophoresis in both the presence and the absence of reducing agents (5). Since no change in szO,w was observed as the concentration of IDH was reduced from 3 to 0.7 PM (Table I), the upper limit for the dimer-monomer dissociation constant (&) predicted from the equations of Gilbert and Gilbert (18) is 20 nM. The weight average molecular weight calculated from the meniscus region of the ultracentrifuge cell in sedimentation equilibrium experiments at 0.1 to 0.4 PM (93,000 k 3,000) was close to that of the whole cell average value (94,300 + 300) indicating

460

SEERY

AND

that dissociation of the dimer was negligible even at 0.1 pM. Velocity and equilibrium techniques agree in establishing a maximum value for Kd of 20 nM. The results of active enzyme sedimentation experiments which were conducted at enzyme concentrations in the nanomolar range provide a means for refining the value of Kd calculated above. The single reaction boundary observed in these experiments at loading concentrations of 13 to 34 nM is consistent with a rapid, reversible interconversion between a monomer and a dimer (17). The sedimentation coefficient of 5.3 f 0.2 S at 25 nM is lower than that obtained from direct observation of the enzyme at 3 pM in moving boundary velocity experiments (5.62 + 0.07; Table II, line 2). Although the decreased sedimentation coefficient could be the result of an alteration in shape, we were unable to detect any change in szo,wdetermined from moving boundary experiments as a consequence of ligand binding. An alternate explanation is partial dissociation of the dimer at the lower concentration of enzyme. A Kd of 2 nM is predicted from the initial concentration of 25 nM and the observed .a20 w of 5.3 S. Since the reacting enzyme band’is diluted during sedimentation, the value of Kd is somewhat overestimated. Even if Kd is 2 nM, the weight fraction of monomer at 25 nM is still only 0.2. Thus, the native form of mammary gland IDH is a highly associated dimer. The only direct indication of a species with a molecular weight lower than the dimer comes from sedimentation velocity of aged preparations where some denaturation may have occurred. The major difference in the physical properties of the cytosolic isozymes of IDH from mammary gland and from beef liver is their aggregation state in the absence of metal-substrate. The complex of the liver enzyme and its substrate had a molecular weight of 89,600 f 500, whereas molecular weights of 40,000 and 96,000 were calculated from the meniscus and base regions of the ultracentrifuge cell, respectively, without substrate (3). In our experiments on the unliganded enzyme, the plot of In c vs r2 was linear; the molecular weight of 94,300 f 300 obtained from the slope of

FARRELL

2.340 f 0.008 is comparable to that of 97,200 + 200 obtained by direct fitting of the concentration distribution data. Our preparations of cytosolic IDH clearly differ from those of Carlier and Pantaloni (3). For the mitochondrial enzyme, Reynolds et al. (10) and Bailey and Colman (6) reported that the dimer is the predominant species in 0.05 to 0.1 M triethanolamine hydrochloride at pH 7 with and without ligands. The association behavior of IDH observed by the latter workers is very different from that of the mitochondrial isozyme used by Kelly and Plaut (8), perhaps indicating differences in enzyme preparation. In the absence of specific ligands the enzyme from porcine heart, as prepared by Kelly and Plaut, had a sedimentation coefficient of 4.8 S and a molecular weight of 58,000; addition of magnesium DL-iSO&rate resulted in the formation of a dimer with s~,~ of 6.7 S. In contrast, denaturating agents are required to effect complete dissociation of mammary gland cytoplasmic IDH. Under all other conditions tested, including the presence of substrate, coenzyme, or substrate/coenzyme analogs, the dimer is the predominant form of mammary cytosolic IDH which is in agreement with the results of Reynolds et al. (10) and Bailey and Colman (6) for mitochondrial IDH. Experiments of other investigators (4, 26) using the Kurganov technique were interpreted as indicating dissociation of IDH at low concentrations of enzyme to yield a less active monomer. Carlier and Pantaloni (26) observed a drop in the activity of beef liver cytosotic IDH below 1 nM. This loss of activity was attributed to enzyme dissociation which was favored by both low protein and low substrate concentration. In our experiments no loss of enzyme activity occurred when albumin was added to the dilution buffer. Significant losses of activity occurred without albumin, and denaturation is a more likely explanation than dissociation. Kelly and Plaut (4) also reported conditions under which the activity of mitochondrial IDH from porcine heart is not decreased upon dilution. They concluded that agents such as bovine serum

BOVINE

MAMMARY

GLAND

ISOCITRATE

albumin, NADPH, or EDTA prevent inactivation by favoring the dimeric form of the enzyme. In our experiments the loss in enzyme activity which was observed at 0.34 nM (Fig. 2) was not reversed by the addition of albumin to the substrate mixture prior to assay, indicating that an irreversible change had occurred upon dilution without albumin. The constancy of the specific activity at various levels of substrate (Fig. 3) indicates the absence of inactive forms of the enzyme as the concentrations of both enzyme and substrate are reduced. Using a Kd of 2 nM, which was estimated from active enzyme sedimentation at saturating levels of substrate, the weight fraction of monomer is predicted to be 0.6 under assay conditions. If the monomer has lower activity than the dimer, a drop in specific activity is expected at concentrations < 1 nM. This was not observed from mammary gland IDH. The discrepancies in the physical and kinetic properties of the various isozymes of IDH, cited above, may have their origin in the methods of isolation. Our procedure for mammary gland IDH (5) and a recent method for the isolation of porcine heart IDH (6) use affinity chromatography as the major purification step without any heat treatment. In contrast, IDH is isolated from beef liver with heating to 60°C for 5 min in the presence of magnesium citrate (3) and from porcine heart with heating to 50°C for 1 h under Nz in the absence of a substrate analog (8). Since both of these preparations show ligand-induced associations, it is possible that heat treatment alters the aggregation behavior of IDH. Documentation of such changes has been reported by Luther et aZ. (23) who directly compared samples of rabbit muscle phosphofructokinase purified with and without heat treatment. The association properties of mammary gland IDH can be changed by external conditions as indicated by the appearance of a modified species with a sedimentation coefficient lower than that of the dimer after prolonged storage at -20°C. Finally, the role of ligand-induced association phenomena in the control of mammary gland can be assessed. We can esti-

DEHYDROGENASE

461

mate the concentration of IDH in mammary gland and predict its aggregation state on the basis of the range proposed for Kd of 2 to 20 nM. Using the specific activity of 52.5 U/mg for the purified enzyme (5) and the average specific activity of 17.3 U/ g of tissue in bovine mammary gland (2), the apparent concentration of IDH is estimated to be 5.7 PM. This latter figure is an underestimate, since the enzyme is cytoplasmic and the volumes occupied by nuclei, mitochondria, Golgi, secretory vesicles, and other organelles are ignored. A concentration of 5.7 /IM is still at least 100 times greater than our best estimate of Kd, indicating that cytoplasmic IDH essentially exists as a dimer in mammary tissue. Thus, a ligand-mediated decrease in the Kd for the dimer-monomer interconversion cannot play a significant role in the regulation and control of the enzyme in situ. Since bovine mammary gland appears to lack the NAD+ form of the enzyme which usually modulates Kreb’s cycle activity (2), an alternative form of regulation of this step may be necessary. REFERENCES 1. MOORE, J. H., AND CRISTIE, W. W. (1981) in Lipid Metabolism in Ruminant Animals (Cristie, W. W., Ed.), pp. 227-2’78, Pergamon, Oxford. 2. FARRELL, H. M., JR., DEENEY, J. T., TUBBS, K. A., AND WALSH, R. A. (1987) J. Dairy Soi. 70,781788. 3. CARLIER, M. F., AND PANTALONI, D. (1973) Eur. J. Biochem, 37,341-354. 4. KELLY, J. H., AND PLAUT, G. W. E. (1981) J. Bid Chmn. 256.335-342. 5. FARRELL, H. M., JR. (1980) Arch. Biochem Bitphys. 204,551-559. 6. BAILEY, J. M., AND COLMAN, R. F. (1985) &o&emistry 24,5367-5377. 7. COLMAN, R. F. (1972) J. Biol Chem 247, 67276729. 8. KELLY, J. H., AND PLAUT, G. W. E. (1981) J. Bid C%m 256,330-334. 9. SEELIG, G. F., AND COLMAN, R. F. (1978)AroA Bio them. Biophys. l&394-469. 10. REYNOLDS, C. H., KUCHEL, P. W., AND DALZIEL, K. (1978) Bia&m. J. 171,73.?-742. 11. SCHECHTER, Y., PATCHORNIK, A., AND BURSTEIN, Y. (1973) Biochemistry 12,3407-3413. 12. FASMAN, G. D. (1975) Handbook of Biochemistry and Molecular Biology: Proteins, Vol. 2, pp. 372-382, CRC Press, Cleveland.

462

SEERY AND FARRELL

13. YPHANTIS,D. A. (1964) Biochemistry 3,297-317. Publication No. 25 of the Chemical Society of 14. KUMOSINSKI,T. K., BROWN,E. M., AND GROVES, London, Alden Press, Oxford. M. L. (1981) J. Biol Chem 256,10,949-10,953. 21. KUCHEL, P. W., REYNOLDS,C. H., AND DALZIEL, K. 15. SVEDBERG,T., AND PEDERSEN,K. 0. (1940) in The (1980) Eur. J. Bioch.em. 110,465-473. Ultracentrifuge, Oxford Univ. Press, London. 22. NA, G. C., AND TIMASHEFF, S. N. (1985) in Meth16. COHEN,R., AND MIRE, M. (1971) Eur. J. B&hem. ods in Enzymology (Hirs, C. H. W., and Tima23,267-275. sheff, S. N., Eds.), Vol. 117, pp. 459-519, Aca17. COHEN,R., AND CLAVERIE,J. (1975) Biopolymers demic Press, San Diego. 14,1701-1716. 23. LUTHER, M. A., HESTERBERG,L. K., AND LEE, J. C. 18. GILBERT, L. M., AND GILBERT, G. A. (1973) in (1985) Biochemistry 24,2463-2470. Methods in Enzymology (Hirs, C. H. W., and 24. LEE, J. C., AND TIMASHEFF,S. N. (1974) BiochemTimasheff, S. N., Eds.), Vol. 27, pp. 273-296, Ac13,257-265. istry ademic Press, San Diego. 19. GRZYBOWSKI,A. K., TATE, S. S., AND DATTA, S. P. 25. KURGANOV,B. I. (1967) Md Biol (Moscow) 1,1727. (1970) J. Chem. Sot. A, 241-245. 26. CARLIER,M. F., AND PANTALONI,D. (1978) Eur. J. 20. SILEN, L. G., AND MARTELL, A. E. (1971) Stability B&hem. 89,511-516. Constants of Metal Ion Complexes, Special