3-Mercaptopicolinate. A reversible active site inhibitor of avian liver phosphoenolpyruvate carboxykinase.

The inhibition of chicken liver phosphoenolpyruvate carboxykinase by 3-mercaptopicolinic acid (3-MP) has been investigated. Kinetic studies show 3-MP to be a noncompetitive inhibitor relative to all substrates and to the activator, Mn2+. EPR studies demonstrate that Mn2+ binding to the enzyme is unaffected by 3-MP. Proton relaxation rate studies demonstrate that 3-MP binds to the binary E X Mn complex with a KD of 0.5 X 10(-6) M and gives a ternary enhancement of 8.0. Additional proton relaxation rate studies detected formation of the quaternary complexes E X Mn X IDP X 3-MP, E X Mn X ITP X 3-MP, and E X Mn X CO2 X 3-MP. High resolution 1H nuclear relaxation rate studies suggest that 3-MP binds in close proximity to the activator cation, Mn2+, but not in the first coordination sphere. Active site models suggest that the 3-MP-binding site may partially overlap the phosphoenolpyruvate-binding site. The NMR studies, which detected formation of the quaternary E X Mn X 3-MP X phosphoenolpyruvate complex, also demonstrated that the binding of one of these ligands affects the interactions of the other ligand with E X Mn. Calorimetric studies of the E X Mn complex demonstrated that 3-MP causes an increase in the transition temperature midpoint without an increase in enthalpy. These results indicate that 3-MP causes a conformational change in the enzyme but does not increase the thermostability of the ternary complex. The experiments reported herein suggest that inhibition by 3-MP is due to specific and reversible binding within the active site of phosphoenolpyruvate carboxykinase.

The inhibition of chicken liver phosphoenolpyruvate carboxykinase by 3-mercaptopicolinic acid (3-MP) has been investigated. Kinetic studies show 3-MP to be a noncompetitive inhibitor relative to all substrates and to the activator, Mn2+.
EPR studies demonstrate that Mn2+ binding to the enzyme is unaffected by 3-MP. Proton relaxation rate studies demonstrate that 3-MP binds to the binary E-Mn complex with a KD of 0.5 X M and gives a ternary enhancement of 8.0. Additional proton relaxation rate studies detected formation of the quaternary complexes E-Mn*IDP*3-MP, E*Mn-ITP*3-MP, and E*Mn-C02*3-MP.
High resolution 'H nuclear relaxation rate studies suggest that 3-MP binds in close proximity to the activator cation, Mn2+, but not in the first coordination sphere. Active site models suggest that the 3-MP-binding site may partially overlap the phosphoenolpyruvate-binding site. The NMR studies, which detected formation of the quaternary E*Mn*3-MP*phosphoenolpyruvate complex, also demonstrated that the binding of one of these ligands affects the interactions of the other ligand with E-Mn.
Calorimetric studies of the E*Mn complex demonstrated that 3-MP causes an increase in the transition temperature midpoint without an increase in enthalpy. These results indicate that 3-MP causes a conformational change in the enzyme but does not increase the thermostability of the ternary complex.
The experiments reported herein suggest that inhibition by 3-MP is due to specific and reversible binding within the active site of phosphoenolpyruvate carboxykinase.
In vertebrates, phosphoenolpyruvate carboxykinase (GTP:oxalacetate carboxylase (transphosphorylating), EC 4.1.1.32) catalyzes an essential reaction in the formation of glucose from three and four carbon precursors: the conversion of oxalacetate to P-enolpyruvate. Mitochondrial and cytosolic * This work was partially supported by Grant AM17049 from the National Institutes of Health and by a grant from the Research Corp. The support for the 300-MHz NMR spectrometer was provided by General Medical Sciences Division of the National Institutes of Health and the Department of Chemistry, University of Notre Dame. 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 a fellowship from The Indiana Affiliate, Inc. of The American Diabetes Association. forms of the enzyme are known, and the ceilular distribution is species-dependent (Utter and Kolenbrander, 1972). Studies have shown the mitochondrial and cytosolic forms to be different proteins in rat liver (Ballard and Hanson, 1969), in monkey liver (Utter and Chuang, 1978), and in guinea pigliver.' The two forms are thought to be subject to different forms of control.
The reaction catalyzed by P-enolpyruvate carboxykinase is thought to be a key regulatory site for gluconeogenesis, yet little is known regarding the control of either form of the enzyme. The cytosolic enzyme from rat liver has been extensively studied, and several factors have been reported to play a role in the control of its activity. These include: diabetes, fasting and hormones (Shrago et at., 1963), a ferroactivator protein (Bentle et al., 1976), and metal ion levels (Colombo et al., 1981;Schramm et al., 1981). Divalent metal ions have also been suggested as playing a regulatory role in mitochondrial P-enolpyruvate carboxykinase from avian liver (Lee et al., 1981).
Tryptophan metabolites and structurally related derivatives of picolinic acid have been reported to alter the rate of gluconeogenesis both in rats and in uitro studies (Snoke et al., 1971). Particular interest has centered on 3-mercaptopicolinic acid since it was first shown to be a potent hypoglycemic agent in laboratory animals and that it inhibits glucose synthesis from three carbon precursors in uitro (DiTullio et al., 1974). Later studies demonstrated that 3-MP2 specifically inhibits gluconeogenesis at the level of P-enolpyruvate carboxykinase (Kostos et al., 1975).
Studies of rat liver cytosolic P-enolpyruvate carboxykinase show 3-MP to be a noncompetitive inhibitor of the P-enolpyruvate formation reaction with a KT -3 pM, and it was suggested that 3-MP removes a tightly bound inhibitory metal other than Mn2+ (Jomain-Baum et al., 1976). MacDonald (1978) has reported that an Fez+. 3-MP complex inhibits crude fractions of cytosolic or mitochondrial P-enolpyruvate carboxykinase from five species equally well. His studies suggested that 3-MP will not bind well to a Mn2+-containing Penolpyruvate carboxykinase. Reynolds (1980) has reported 3-MP to be a competitive inhibitor with respect to P-enolpyruvate using P-enolpyruvate carboxykinase from rat liver and from tapeworm.
Other investigators have reported that the cytosolic enzyme is more sensitive than the mitochondrial enzyme to inhibition by 3-MP from studies of enzymes from guinea pig liver (Robinson and Oei, 1975) and from kidney tubules of five

3-Mercaptopicolinate Inhibition 11655
species (Watford et al., 1980). Initial observations in our laboratory suggested that 3-MP was the only derivative of picolinic acid which specifically affected chicken liver mitochondrial P-enolpyruvate carboxykinase activity, causing inhibition at micromolar concentrations.
The inhibitor 3-MP could be a useful tool in understanding the mechanism of regulation of this enzyme. A detailed kinetic and spectroscopic investigation of the inhibition of chicken liver mitochondrial P-enolpyruvate carboxykinase by 3-MP was undertaken in an effort to elucidate the mechanism by which enzymic activity, and thereby the rate of glucose synthesis, is controlled in vertebrates.
The results presented here suggest that 3-MP inhibits avian liver P-enolpyruvate carboxykinase activity by reversibly and specifically binding within the active site. This binding causes a conformational change which is unfavorable for catalysis.

EXPERIMENTAL PROCEDURES
Materials P-enolpyruvate carboxykinase was purified from chicken liver as previously described by Hebda and Nowak (1982a) and modified by Duffy (1982) and Lee (1983). Lactate dehydrogenase, malate dehydrogenase, and pyruvate kinase were purchased from Boehringer Mannheim. P-enolpyruvate, oxalacetate, ITP, IDP, and NADH were purchased from Sigma. 3-Mercaptopicolinic acid and the other derivatives tested were generous gifts from Drs. Harry Saunders and Nicholas DiTullio, Smith Kline and French Laboratories, Philadelphia, PA and were prepared fresh daily by dissolving in HZ0 and neutralizing to pH 7 with KOH prior to final dilution. All other reagents were of the highest grade commercially available. Distilled, deionized H,O was used routinely to prepare all solutions.

Methods
Kinetic Assays-Initial velocity studies of the carboxylation of Penolpyruvate to form oxalacetate were done using the continuous assay described by Hebda and Nowak (1982a), which couples Penolpyruvate carboxykinase to malate dehydrogenase. The disappearance of NADH absorbance at 340 nm was monitored with either a Gilford 240 or 250 spectrophotometer. Initial velocity studies of the P-enolpyruvate formation reaction were done by monitoring the decrease in absorption of NADH at 340 nm using an assay which couples P-enolpyruvate carboxykinase to pyruvate kinase and lactate dehydrogenase (Lee et al., 1981). Enzyme activity is expressed as micromoles of NAD+ formed per ml/min/mg of protein (units/mg).
Enzyme Preparation for Physical Studies-Prior to all physical studies, enzyme was prepared by desalting on a Bio-Rad P-6-DG column (1.1 X 25 cm), having 2 cm of Chelex 100 on the top, which had been equilibrated in 65 mM Tris-HCI buffer, pH 7.4. The enzyme was concentrated using either an Amicon Model 8MC ultrafiltration system with microvolume accessory or a Minicon B-15 concentrator. All solutions used in the physical studies, with the exception of MnC12, were passed through a Chelex 100 column.
3-MP Concentration-When an exact determination of 3-MP concentration was required, it was determined spectrophotometrically by reaction with 5,5'-dithiobis(2-nitrobenzoic acid). Aliquots of 3-MP were added to an excess of 5,5'-dithiobis(2-nitrobenzoic acid) (0.4-ml final volume), and the change in absorption a t 412 nm, due to release of the nitromercaptobenzoate anion, was monitored. Using the reported extinction coefficient of 13,600 M" cm" (Habeeb, 1972), the moles of anion released could be calculated. The amount of anion released is stoichiometric with the amount of 3-MP.

Mn2+
Biding Studies-Mn2+ binding to P-enolpyruvate carboxykinase in the presence of 3-MP was investigated using EPR techniques (Cohn and Townsend, 1954;Mildvan and Cohn, 1963). A Varian model 4500 spectrometer was used, and the probe temperature was maintained at 22 f 1 "C by NO flow. Individual samples (0.05-ml final volume) contained varying Mn2+ Concentrations and fixed concentrations of enzyme and 3-MP in 65 mM Tris-HC1, 100 mM KCl, pH 7.4. The control samples lacked enzyme. The samples were drawn into 1-mm (inner diameter) quartz capillary tubes and the spectra were taken. The peak height of the Mn2+ spectrum was used as a measure of free MnZ+, and it was assumed that bound Mn2+ makes no contribution to the observed signal height. Bound MnZ+ was calculated as the difference in concentrations of total Mn2+ and free Mn2+. Data were treated by the Scatchard (1949) method. The same procedures were used to measure the binding of Mn2+ to 3-MP.
Proton Relaxation Rate Studies-Formation of the ternary E. Mn.
3-MP complex and quaternary E . Mn. 3-MP. S complexes was investigated using PRR techniques (Mildvan and Cohn, 1970;. Proton relaxation rates were measured at 22 f 1 "C with a Seimco pulsed NMR spectrometer operating at 24.3 MHz, using the Carr-Purcell (1954) 180"-~-90" sequence. Sample tubes contained, in a final volume of 0.05 ml, a fixed concentration of enzyme and Mn2+ in 65 mM tris-HC1, 100 mM KCl, pH 7.4. A sample prepared identically but which also contained 3-MP was titrated into the first tube, and the proton relaxation rate was measured at each addition. No dilution of enzyme or Mn2+ thus occurred during the titration. Formation of the quaternary E.Mn.3-MP.S complexes was investigated in a similar manner. The sample tubes contained a fixed concentration of enzyme, Mn2+, and the substrate. A second sample contained identical amounts of enzyme, Mn2+, and the substrate and 3-MP was also present. Aliquots of this sample were titrated into the first tube, and the PRR was measured with each addition. The reverse titrations were also performed. In this case, sample tubes which contained fixed concentrations of enzyme, Mn2+, and 3-MP were titrated with an identical sample which also contained the substrate to be titrated. The PRR was measured at each addition.
The dissociation constant (KD) of ligand from the enzyme. Mn or the enzyme. Mn. ligand complex was obtained by generating a "best fit" to the data. Titration data were fit using the program of Reed et al. (1970) which has been adapted for the Wang calculator and which varies KD until a minimum per cent standard deviation is obtained.
High Resolution 'H NMR-The 'H NMR spectra were obtained at 300 MHz using a Nicolet NTC-300 spectrometer equipped with a 239A pulse system and a Nicolet 1180E computer. The spectra taken at 100 MHz were obtained with a Varian XL-100-15 spectrometer interfaced to a Nicolet 1080 computer. The spectrometer was locked is the concentration of ligand 3-MP being measured. The magnitude of this effect can be used to calculate the distance, r, between the nucleus measured and the paramagnetic metal ion by using the Solomon-Bloembergen equation (Solomon, 1955;Bloembergen, 1957). For Mn2'-'H dipolar interactions, the equation simplifies to r (A) = 812 (TIM.f(Te))l" where the correlation function is as follows.
In the above equation, T'M is the relaxation time of the nucleus measured in the Mn. ligand complex. If the values l/pTl, are in the fast exchange region, then l/pT, = 1/TlM . The values w1 and w, are the nuclear and the electron precession frequencies, respectively, and T , is the correlation time of the electron-nuclear interaction. The value for T, can be estimated by several methods  including measuring l/pTl, as a function of w,. This method was utilized in the studies reported herein.
The quaternary E . Mn. P-enolpyruvate .3-MP complex was detected using 'H nuclear relaxation techniques. In this experiment, microliter quantities of 3-MP were titrated into a buffered solution which contained fixed concentrations of enzyme, P-enolpyruvate, and Mn2+. The 2 ' ' values of the P-enolpyruvate protons were measured at each addition using the same technique as described above. Differential Scanning Calorimetry-Calorimetric measurements were done using an MC-1 Scanning Calorimeter (Microcal, Amherst, MA). Calorimetric scans were performed at 45 "C/h under constant pressure (-7.0 cmHg) using the procedures described by Ploplis et al. (1981). The buffer system for these studies was 65 mM HEPES,

RESULTS
Kinetic Studies-Several derivatives of picolinic acid were screened for their effect on chicken liver P-enolpyruvate carboxykinase activity. The compounds tested were 3-MP, 3aminopicolinic acid, quinolinic acid, 3-S-methylpicolinic acid, 3-S-phenylpicolinic acid, 3-hydroxypicolinic acid, 3-bromopicolinic acid, and 2-hydroxymethyl-3-mercaptopicolinic acid. All compounds tested inhibited P-enolpyruvate carboxykinase activity, but only 3-MP caused inhibition at concentrations of 0.1 mM or less. None of the other compounds showed effects at concentrations less than 0.25 mM. At higher concentrations, these compounds caused abrupt losses in enzyme activity (data not shown). This phenomenon is believed to be due to some nonspecific effect, such as removal of the activator cation, Mn2+, rather than specific inhibition of enzymatic activity.
Detailed kinetic studies of 3-MP inhibition against P-enolpyruvate, oxalacetate, and the nucleotide substrates exhibit mixed noncompetitive inhibition. The results of one such study are shown in Fig. 1A. 3-MP affects both the KM and the V,,,,,. If the experiments are performed against HCOF or Mn2+, only a V, , effect is observed, a pattern typical of "classical" noncompetitive inhibition. The results of one such experiment are shown in Fig, 1B. The slope and the intercept replots are linear in all the experiments performed (ie. insets of Fig. 1, A and B ) . Table I    @-mercaptoethanol as those obtained using an assay system without reducing agent. These results suggest inhibition by 3-MP is not due to mixed disulfide bond formation. The inhibition of enzyme activity by 3-MP is reversible. Enzyme was incubated with excess 3-MP for 5 min and enzymatic activity was lost. The sample was then passed over a Bio-Rad P-6-DG desalting column. The enzyme was separated from free 3-MP under these conditions and full enzymatic activity was recovered. Experiments which involved high concentrations of 3-MP and P-enolpyruvate carboxykinase could occasionally become hazy after several hours at room temperature. Enzyme activity could not be recovered in these cases. Whenever this phenomenon was observed, the data were discarded.
Mn2+ Binding-Mn2+ binding to chicken liver P-enolpyruvate carboxykinase in the presence of 3-MP was investi-3-Mercaptopicolinate Inhibition 1 O P M enzyme, and the concentration of Mn2+ was varied from 2.3 X gated by EPR techniques. The data are shown in Fig. 2 in the form of a Scatchard plot. Values for KO, the dissociation constant for E . Mn, and n, the number of Mn2+-binding sites on the enzyme, were determined to be 3.4 X M and 0.99, respectively. These values are in good agreement with those previously reported for native enzyme (Hebda and Nowak, 1982b). A binding study was also done using PRR techniques (data not shown) and similar results were obtained. These results confirm the kinetic data which indicated that the presence of 3-MP had no effect on the affinity of the enzyme for Mn2+. Mn2+ binding to 3-MP was also investigated by EPR techniques (data not shown). Although the data was somewhat scattered, a best fit to the data gives a value of approximately 6.1-9.1 X M for KD and a value of 1.3-1.7 for n. These values are in good agreement with those reported by Jomain-Baum et al. (1976) and MacDonald and Lardy (1978). Ternary E . Mn. 3-MP Complex Formation-The binding of 3-MP to enzyme .Mn2+ was investigated by PRR methods (Fig. 3). The data were fit as described under "Methods." The results demonstrate that 3-MP interacts with the enzyme. Mn complex with a KD range of approximately 0.1-1.0 X M and gives a ternary enhancement, CT, of 8.0. With this enzyme the data for tight binding ligands (KO < 1O"j M) can be fit equally well when KO varies by a factor of 5-10. A partial frequency dependence study of the PRR was done (data not ~h o w n ) .~ T h e data is nearly identical with results obtained with the ternary E . Mn. P-enolpyruvate complex (Duffy, 1982). The observed frequency dependence of PRR and lack of a substantial temperature effect suggests that the water is in rapid exchange with the bound Mn'+. This is in contrast to the positive temperature dependence of PRR measured with the enzyme from pig liver which suggests that A. L. Makinen and T. Nowak, unpublished observations. in that case the water is in slow exchange (Miller et al., 1968).
For the E. Mn (Duffy, 1982). These results suggest that 3-MP binding causes the same type of change in the environment of enzyme-bound Mn2+ as P-enolpyruvate binding does: displacement or immobilization of one-half of the water molecules bound in the first coordination sphere of Mn2+.
Formation of Quaternary E . Mn .3-MP. S Complexes-PRR titrations were performed by titration of 3-MP into a solution which contained enzyme, Mn2+, and substrate to detect the formation of an enzyme. Mn2+. substrate -3-MP complex. The reverse titration sequence was also performed where substrate was titrated into the E .Mn. 3-MP complex. Table I1 summarizes the results of the PRR titrations. The results were fit by computer simulation of the data.
For P-enolpyruvate, there was no evidence for quaternary E . Mn . P-enolpyruvate .3-MP complex formation by either type of titration. No change in observed enhancement (e*) was detected upon the addition of inhibitor to the E. Mn. Penolpyruvate complex (e* = 5.5) or upon the addition of Penolpyruvate t o the E . Mn.3-MP complex (c* = 4.9). High resolution 'H NMR experiments ( d e infra) demonstrate that the quaternary E . Mn. P-enolpyruvate. 3-MP complex does form, however. Since P-enolpyruvate binding and 3-MP binding to enzyme.Mn cause the same type of change in the environment of bound Mn", addition of the second ligand does not cause any further perturbations in the environment of bound Mn2+; therefore, the quaternary complex cannot be detected by PRR techniques.
When 3-MP was titrated into E. Mn. IDP, a titration curve was obtained (Fig. 3). Computer fits to the data give an eT of 3.5 and a KD of approximately 5 X M. Titration of IDP into E.Mn.3-MP causes no change in e* (e* = 4.3).
Addition of 3-MP to E .Mn.ITP resulted in a titration curve which was fit with a KD of 0.1-1.0 x M and an cT of 1.4. Titration of ITP into E. Mn. 3-MP showed no change in the observed enhancement (e* = 4.2).
Formation of the quaternary E. Mn. CO,. 3-MP complex was detected by PRR techniques. The 3-MP binds to the ternary complex and causes a decrease in the observed enhancement. The titration curve was fit with a KO of 0.1 x M and an cT of 9.6 was obtained. HCO, was titrated into the E .Mn.3-MP complex and a small increase in enhancement was observed ( € * = 4.5). A graphical estimation of the dissociation constant for HCO; from E .Mn.3-MP gives a value of 2.0 X lo-' M. This value is in good agreement with the KM of HCO; and the K3 and K, values for HCO, binding to E . Mn previously reported (Hebda and Nowak, 1982a and 198213). This result is consistent with the kinetic experiments which suggest that HCO; binding to P-enolpyruvate carboxykinase is unaffected by 3-MP. As previously stated, with tight binding ligands, good fits to the data can be obtained even when KD values vary by a factor of 5-10. For example, the per .0 X lU5 M; K,, the binding of ligand to E, varied from 1 X to 1 X 10" M with little effect; K1, the binding constant for Mn2+. titrant was 0.7 X M for 3-MP and the other values were those previously used (Hebda and Nowak, 1982b).

Complex
Titrant

ITP.
High Resolution ' H NMR-A proton NMR spectrum of 3-MP in D20 (pD 7.4) is shown in Fig. 4. Resonance peak assignments were made on the basis of chemical shifts, coupling constants, and homonuclear decoupling experiments.
The effect of added Mn2+ on the l/Tl values of the 3-MP protons in the presence and absence of enzyme was measured as described under "Methods," and the l/pTl, values were determined. The relaxation rates of the three ring protons of 3-MP were differentially affected by the paramagnetic Mn" ion (Fig. 5 , A and B).
The effect of Mn2+ on the relaxation rates of the protons of 3-MP in the binary Mn.3-MP and ternary E.Mn.3-MP complexes is summarized in Table 111. The relaxation rates for the ternary complex were measured at both 100 and 300 MHz. For the binary complex a value of 3.48 X lo-'' s for T, was used to calculate r. In these calculations it was assumed that for small complexes re equals T,, the rotational correlation time. The 7, for Mn(H20)i+ (3 X lo-'* s) was estimated to increase by approximately 15% upon formation of the Mn2+.   Results obtained at 100 MHz.
ligand complex, proportional to a change in molecular weight of the complex. A T~ of 3.9 X lo-' s was used to calculate r in the ternary complex. This value was obtained from the frequency dependence of the relaxation rates of the 3-MP protons in the ternary complex (Table 111) and assuming no frequency dependence of T=. Assuming a maximum frequency dependence of 7, yields a value for T~ which is smaller by a factor of 8 (0.5 X s). This value is reflected in a 15% smaller value of f ( 7 J at 100 MHz and a factor of 4 smaller value at 300 MHz. Resultant variations in the distances calculated are reflected in the error levels given in Table 111. Studies of several ternary P-enolpyruvate carboxykinase complexes in our laboratory have shown that T= varies little in the ternary complexes investigated, regardless of the method used to determine T~ (Duffy, 1982;Lee, 1983).
In calculating the Mn nuclei distances, it was assumed that This assumption is based on the observation that the three protons of 3-MP experience a different paramagnetic effect (Fig. 5, A and B ) and a frequency dispersion of l/Tlp is measured (Table 111). If the ligand was in slow exchange, 1/ T M would dominate l/pT, and identical l/pTlP values would be expected for all three protons. Furthermore, if the larger value of l/pTlp measured (H6 proton at 100 MHz) is partially limited by chemical exchange, l/pT, < l / T l~, the calculation of 7, would then lead to an apparent shorter value for T~. This is not the case as a value of T , = 3.9 ns is calculated from this data. The distances were also calculated based on the assumption that q, the number of 3-MP molecules associated with E . Mn, is 1. Two lines of evidence support this assumption. First, the PRR titrations were fit using a model where q = 1, and good fits (<lo% S.D.) were obtained for all the titrations. Futhermore, in the PRR experiments where well defined titration curves were obtained (ie. Fig. 3), the maximum effect is observed when the concentration of 3-MP is approximately the same as the concentration of E .Mn, suggesting that 3-MP forms a 1:l complex with the E . Mn complexes. The other assumption used is that all of the Mn2+ added to the 'H NMR experiment forms an E . Mn .3-MP complex. This assumption is consistent with the frequency dispersion of l/pTlp observed (Table 111). The binary Mn.3-MP complex shows no frequency dispersion in this frequency range. Also, a calculation of the Mn2+ distribution indicates that >97% of the Mn2+ is in the enzyme complex.
The Mn 'H distances calculated for the ternary complex (Table 111) are larger than those measured for the binary complex. This indicates that 3-MP binds close to the activator cation Mn2+ in chicken liver P-enolpyruvate carboxykinase, although it does not bind with the same coordination scheme as it does with free Mn2+.
Models of E . Mn. 3-MP (Fig. 6) were constructed based on the interatomic distances reported herein. Similar models of the E . Mn . P-enolpyruvate complex were constructed based on data reported elsewhere (Duffy, 1982). These models suggested that the 3-MP-binding site could possibly overlap the P-enolpyruvate-binding site. Since the quaternary E . Mn. Penolpyruvate. 3-MP complex was not detected by PRR techniques, studies were done to determine if the quaternary complex does form, or whether P-enolpyruvate and 3-MP binding to E . Mn is mutually exclusive.
Attempts to do direct P-enolpyruvate binding studies, using radiolabeled P-enolpyruvate, were unsuccessful. The enzyme is unstable over the time course of an equilibrium dialysis experiment and would frequently precipitate. This occurred in the presence or absence of 3-MP. P-enolpyruvate carboxykinase elution from gel filtration columns is retarded (Hebda and Nowak, 1982a), and this phenomenon is thought to interfere with Hummel-Dreyer-type binding studies. Skewed peaks and unequal peak and trough sizes were consistently obtained when binding studies were attempted using this method. In all cases, the enzyme concentration was 1.34 mg/ml (18.6 pM) and the buffer system was 0.065 M HEPES, pH 7.4, which contained 0.1 M KC1 and 0.01 M dithiothreitol. The heating rate was 45 "C/h. A high resolution 'H nuclear relaxation rate study of the quaternary complex was performed. If 3-MP was competitive with P-enolpyruvate for a binding site on the enzyme, increasing concentrations of 3-MP would displace P-enolpyruvate from the enzyme, and the relaxation rates of the P-enolpyruvate protons would approach the values for free P-enolpyruvate.
3-MP was titrated into a buffered solution of E . Mn. Penolpyruvate, and the effect on the relaxation rates of Penolpyruvate protons was monitored a t each addition. The results are shown in Fig. 7. In the ternary E . P-enolpyruvate complex, the l/Tl value for the cis proton is approximately 1.2 s-', and this value remains constant upon addition of 3-MP. The l/Tl value for the trans proton of P-enolpyruvate decreases upon the addition of 3-MP. These data indicate that although 3-MP perturbs the environment of enzymebound P-enolpyruvate, it does not displace P-enolpyruvate from its binding site. Both the substrate and the inhibitor bind, forming a quaternary E . Mn. P-enolpyruvate. 3-MP complex. The pdotons of the substrate are estimated to be greater than 9 A from the Mn2+ in the quaternary complex.
In the ternary E . Mn . P-enolpyruvate complex, the distances from Mn2+ to the cis and the trans protons are -5.7 and 6.6 A, respectively (Duffy, 1982).
The relaxation rates of the 3-MP protons were determined from the same experiment. In the quaternary complex these protons all experience a smaller paramagnetic effect than in the ternary complex (data not shown). If one uses these values to calculate Mn2+ 'H distances using the assumptions made previously and the sape value of re, the values obtained are all approximately 2 A larger than those calculated for the ternary complex. This again suggests that while both 3-MP and P-enolpyruvate bind to form the quaternary complex, their interactions with E . Mn are affected by the presence of the other ligand.
During the time course of the 'H NMR experiments, 3-MP appears to exist in solution as a monomer, as evidenced by a single set of resonance peaks. Over extended periods of time, however, a second set of resonance peaks appears, shifted downfield by -0.3 ppm. This phenomenon occurs both in the absence and presence of enzyme and is possibly due to formation of 3-MP dimers. The relative concentrations of these are minor, however.
Differential Scanning Calorimetry-Transition endotherms for enzyme. Mn, enzyme. Mn. 3-MP, and enzyme. Mn. P-enolpyruvate are shown in Fig. 8. In each case only one thermal denaturing transition is observed. The presence of saturating 3-MP causes an increase in the TM from 54.4 to 58.3 "C without an increase in the measured calorimetric enthalpy. Saturating P-enolpyruvate, however, causes a large increase in the TM to 63.8 "C which is accompanied by an increase in the calorimetric enthalpy, indicative of an increase in thermostability of the ternary complex over the binary E .Mn complex. The binding of 3-MP does not elicit the same effects on the protein as does the substrate P-enolpyruvate.

DISCUSSION
Mixed type kinetic inhibition may be thought of as partial competitive inhibition (KM effects) and pure noncompetitive inhibition (V,,, effects), although no simple unified model for such inhibition exists. The competitive portion of this effect suggests at least partial overlap of the inhibitor and the substrate for the active site. With P-enolpyruvate, oxalacetate, and the nucleotide substrates, KI.1 # KI,s, which indicates that 3-MP has a different affinity for E' than for E ' . S, where S is the variable substrate and E' represents enzyme in the presence of fixed substrates and cofactors. Since K,,S < KIJ, this suggests that these substrates affect formation of the E . 3-MP complex. The E ' . S complex has a lower affinity for 3-MP than does free E. The inhibitor constants were corrected for the presence of the fixed substrate to obtain a KI value ( Table I) Kinetic constants, where comparable, are lower when measured in the direction of Penolpyruvate formation apparently due to differences in experimental conditions for the two assays. Analogous behavior has been reported previously with rat liver P-enolpyruvate carboxykinase (Jomain-Baum et al., 1976) and with the en-zyme from rat, guinea pig, dog, rabbit, and man (Watford et al., 1980). The enzyme from chicken liver has been reported to have different requirements for metal ion and sulfhydryl reagent in the forward and reverse reactions (Lee et al., 1981;Hebda and Nowak, 1982a). The differences in magnitude of the Kl values for 3-MP in the forward and reverse reaction indicate that the interactions of 3-MP with the enzyme are influenced by assay conditions.
The value of Kl, measured for 3-MP in the P-enolpyruvate formation direction, is approximately the value of KO for 3-M P (Table 11). The presence of IDP weakens the binding of 3-MP, and HCO, causes no effect as suggested by inhibition kinetics. The fit to the binding data in the presence of ITP is not sufficiently precise to detect a factor of 4 increase in K D for 3-MP.
The presence of 3-MP affects the formation of the E. S complexes, with 3-MP increasing the KM for these substrates.
These kinetic results are consistent with a model of 3-MP partially overlapping the P-enolpyruvate (oxa1acetate)-and the nucleotide-binding site. The K M value of HCO, is unaffected by 3-MP; Kl,s = Kl,s. This indicates that the interactions of 3-MP and CO, (the true substrate for this enzyme) with the enzyme are mutually independent. These results were confirmed by the PRR titration data and suggest that the CO, is located on the enzyme a t a site more remote from the 3-MP site. The kinetic and the binding data are consistent with a single mode of 3-MP binding, indicating that the same phenomenon is being investigated.
The reversibility of inhibition and lack of evidence for mixed disulfide bond formation support the hypothesis that inhibition of chicken liver P-enolpyruvate carboxykinase is due to a specific and reversible binding of the inhibitor to the enzyme. The relatively weak binding of 3-MP to Mn2+ further supports this idea, as inhibition by 3-MP at the concentrations reported herein could not be due to simple removal of the activator cation by 3-MP. The kinetic studies show that 3-MP and Mn2+ interact independently with the enzyme, while the EPR and PRR experiments demonstrate that both Mn2+ and 3-MP specifically bind to the enzyme. The lack of a positive synergistic effect of 3-MP on Mn2+ binding to Penolpyruvate carboxykinase is contrary to the effect observed by P-enolpyruvate (Hebda and Nowak, 1982b), suggesting different conformational effects by these two ligands although similar tT values are induced.
The PRR studies also display evidence for the formation of the quaternary complexes E. Mn . CO,. complex also forms. While the environment of the cis proton of P-enolpyruvate is unchanged in the quaternary complex from the ternary complex, the trans proton undergoes a substantial change in its environment upon 3-MP binding such that it experiences a negligible paramagnetic effect from the activator, Mn2+. Thus, although 3-MP and P-enolpyruvate bind to E . Mn, it is suggested that their binding sites are in very close proximity, possibly overlapping each other. Partial overlap of binding sites would be consistent with the KM effects observed and would also be consistent with the 'H relaxation rate studies which demonstrated that the interactions of both P-enolpyruvate and 3-MP with E. Mn are affected by the presence of the other ligand. Direct binding studies of P-enolpyruvate to E . Mn, in the presence and absence of 3-MP, would be helpful in clarifying whether the interactions of P-enolpyruvate with the enzyme are weaker in the presence of 3-MP. As discussed above, however, this type of experiment has not been successful to date.
While an overlap of binding sites may explain the KM effects, it is not sufficient to explain the Vmax effects. A classical noncompetitive inhibitor is generally thought to bind at a remote allosteric site where it causes a conformational change which is detrimental to catalysis. Both the E .S and the E. S . I (where I is inhibitor) complexes form, but the E.
S . I complex is nonproductive. The inhibition of chicken liver P-enolpyruvate carboxykinase by 3-MP is believed to be due to 3-MP binding within the active site, not at a remote site. Not only is there an overlap of substrate-and inhibitorbinding sites, but 3-MP binding also causes a conformational change which is unfavorable for catalysis. That 3-MP binding clearly causes a conformational change is shown by the increase in TM in the differential scanning calorimetry studies. Although tight binding of 3-MP to E. Mn is demonstrated by the PRR studies, it is unaccompanied by an increase in enthalpy. This suggests that no additional stabilizing interactions, such as hydrogen bonding, hydrophobic interactions, or electrostatic interactions, occur when 3-M P binds to E-Mn. When the substrate P-enolpyruvate binds to E. Mn, there is an increase in the enthalpy, indicating the formation of such additional interactions. These interactions increase the thermostability of the ternary complex.
The mixed type kinetic behavior of 3-MP inhibition of chicken liver P-enolpyruvate carboxykinase can thus be described in terms of specific enzyme-ligand interactions occurring within the active site of the enzyme: overlapping of the inhibitor and substrate sites ( K M effects) and an unfavorable conformational change (Vmax effects). The specific and reversible nature of 3-MP inhibition distinguishes it from other reported types of inhibitors, such as EDTA and sulfhydryl reagents, which elicit kinetic effects which appear noncompetitive.
The experiments reported here do not rigorously rule out the possibility that 3-MP binds at two independent sites, one affecting substrate binding and the second causing the unfavorable conformational change. A two-site model can only be ruled out by direct binding studies of 3-MP to the enzyme. The one-site model is the simplest model consistent with the data presented here.