Alanine Dehydrogenase from Soybean Nodule Bacteroids KINETIC MECHANISM AND pH STUDIES*

The kinetic mechanism of alanine dehydrogenase from soybean nodule bacteroids was studied by initial velocity experiments with or without product inhibi-tors, dead-end inhibitors, or alternate substrates. Without inhibitors, double-reciprocal plots of initial velocity experiments showed intersecting lines, indicating a sequential mechanism. These initial velocity experiments also revealed rapid-equilibrium ordered binding of NH: prior to pyruvate. nonlinear, concave down double-reciprocal was obtained. by pyruvate L-ala-nine with cosubstrates inhibition and pyruvate L-alanine L-alanine NH+

The kinetic mechanism of alanine dehydrogenase from soybean nodule bacteroids was studied by initial velocity experiments with or without product inhibitors, dead-end inhibitors, or alternate substrates. Without inhibitors, double-reciprocal plots of initial velocity experiments showed intersecting lines, indicating a sequential mechanism. These initial velocity experiments also revealed rapid-equilibrium ordered binding of NH: prior to pyruvate. When NAD was varied at changing-fixed concentrations of L-alanine, a nonlinear, concave down double-reciprocal plot was obtained. Substrate inhibition by pyruvate or L-alanine with cosubstrates varied was uncompetitive giving further support to an ordered mechanism. Product inhibition studies showed that both NAD and NADH and pyruvate and L-alanine were competitive. This suggested a Theorell-Chance mechanism. When product inhibition by L-alanine was studied with NH+ varied in a series of experiments at increasing concentrations of pyruvate, the inhibition was eliminated, as expected for a Theorell-Chance mechanism. Furthermore, when NADH, NH:, and pyruvate were varied simultaneously, maintaining their concentrations at a constant ratio to each other, an infinite V,,, was obtained. pH studies of the kinetic parameters indicated that NH:, rather than NHs, was the true substrate that binds to a residue on the enzyme with a pK of 8.1. In conclusion, the kinetic mechanism at pH 8.5 was determined to be a Ter-Bi Theorell-Chance. In the amination direction, the substrates add in the order: NADH, NH:, pyruvate, with NH: binding in rapidequilibrium. In the reverse direction, NAD adds first, followed by L-alanine.
Alanine dehydrogenase (EC 1.4.1.1) catalyzes a reversible reaction converting pyruvate to alanine using NADH as a n oxidation/reduction cofactor. NADH + NH: + pyruvate +-+ L-alanine + NAD + H 2 0 (Eq. 1) The reaction has a pH optimum of 8.5 in the amination direction and 10.0 in the direction of deamination. The enzyme exhibits substrate inhibition with pyruvate or alanine * This work was supported by United States Department of Agriculture Competitive Grant 91-37305-3301. This paper is Contribution 11,855 of the Agricultural Experiment Station of the University of Missouri. 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.
$ To whom correspondence should be addressed Dept. of Biochemistry, 117 Schweitzer Hall, University of Missouri, Columbia, MO 65211; Fax: 314-882-5635. and apparent substrate inhibition at high concentrations of NH:. A study of partially purified alanine dehydrogenase from soybean bacteroids shared some of these characteristics (1,2). The enzyme exists in several microorganisms and usually has a high K,,, for NH:, in the range of 20-300 mM. In Bradyrhizobium japonicum bacteroids, however, the appare n t K,,, for NH: is 10-100 times less (4-7 mM). Since the enzyme potentially plays an important role in nodule nitrogen metabolism, an attempt was made to determine the kinetic mechanism in order to better understand how the enzyme functions.

EXPERIMENTAL PROCEDURES
Enzyme Assays-Assays were done as described previously.' Alanine dehydrogenase purified from soybean nodule bacteroids was assayed by following the oxidation of NADH or reduction of NAD at 340 nm in a 1-cm quartz cuvette. Assays were done in 100 mM TAPS' buffer, pH 8.5, or CAPS buffer, pH 10.0. Buffers and NH&l stock solutions were kept at 25 "C, as were the assays. The rest of the substrates and enzyme were kept on ice until use. The ionic strength was kept constant with NaCl. Assays were done on a Gilford 250 spectrophotometer, using a deuterium lamp. The spectrophotometer was equipped with a Gilford 6051 chart recorder and a Gilford 2451-A automatic cuvette positioner. Full scale sensitivity of A340 and a chart speed of 1 cm/min were used.
pH Studies-A mixture of BisTris, HEPES, and Tricine buffers or TAPS, CHES, and CAPS buffers was used at 50 mM concentration each. Assays at overlapping pH values between the two mixtures were used. The pH of the reaction mixture without added enzyme was measured, and the final pH at the end of the assay was checked. The enzyme was stable throughout the pH range as determined previ-Data Processing-The nomenclature used is that of Cleland (12). Reciprocal initial velocities were plotted versus reciprocal substrate concentrations, and the experimental data were fitted to Equations 2-10, ously.' by the least squares method (13), using an IBM XT computer and the compiled Fortran programs of Cleland (14), except Equations 3 and 10, which were fitted using the least squares Minsq program by MicroMath (Salt Lake City, UT), kindly supplied by Dr. Peter Tipton. Except where noted, the points in the reciprocal plots are the experimentally determined values, while the lines are calculated from the fits of these data to the appropriate rate equation. Data were usually the average of four replicates, and the error analysis was expressed as the standard deviation. Linear double-reciprocal plots were fitted to Equation 2, while sequential initial velocity data were fitted to Equation 4. Nonlinear concave up intercept replots, indicating uncompetitive substrate inhibition, were fitted to Equation 3. Data conforming to linear competitive, linear noncompetitive, and linear uncompetitive inhibitions were fitted to Equations 5-7, respectively. The pH dependence of log V / K for pyruvate was fitted to Equation 8, describing a half-bell-shaped curve with a drop in activity at high pH, with a slope of 1. The pH dependence of log V / K for NH: was fitted to Equation 9, which describes a bell-shaped curve with slopes of 1. Equation 10 describes a rapid-equilibrium Theorell-Chance mechanism and was derived (15) from an equation used to describe the mechanism for malic enzyme (16). RESULTS Initial Velocity Patterns-In the forward direction, intersecting lines were obtained with all combinations of substrates, confirming a sequential mechanism (Table I). In the reverse direction, when L-alanine was varied a t different levels of NAD, intersecting lines were also obtained, indicating a sequential mechanism in this direction as well (Fig. 1). However, statistical analysis of the data gave kinetic constants with a relatively large error, and when the same data was plotted with NAD as the varied substrate, a curued line concave down at the l / u axis was obtained (Fig. 2). A similar pattern has been obtained, when NAD was varied, with the enzyme from A. cylindrica (7). This apparent negative cooperativity (17) has been demonstrated in several dehydrogenases. This property was not investigated further in this study and should be considered tentative until confirmed by binding studies. In the amination direction, with pyruvate varied and NH: as the changing-fixed substrate, lines that intersect on the l / u axis were obtained (Fig. 3). This type of pattern in an initial velocity plot occurs when the changing fixed substrate (NH:) adds in rapid-equilibrium fashion to the enzyme. The same data plotted with ammonia as the varied substrate gave lines intersecting to the left of the l / u axis ( Table 1). The slope in this plot became nearly horizontal at high concentra-tions of pyruvate, making the velocity appear independent of NH: concentration. T o confirm the rapid-equilibrium binding of NH:, the slope replot was examined (Fig. 4), which extrapolated through the origin, as predicted. This type of pattern can only be obtained in a rapid-equilibrium ordered mechanism with pyruvate adding after NH: (18).
Product Inhibition-Product inhibition by NADH with NAD varied (in the region of linear response) gave competitive inhibition, demonstrated by lines intersecting on the l / u axis (Table 11). This is consistent with NADH and NAD adding first and coming off last, respectively (Scheme l), as occurs in many other dehydrogenases with ordered mechanisms. The reverse experiment, with NADH varied and NAD the changing-fixed product inhibitor, gave a similar pattern (Table 11).
When pyruvate was varied at changing-fixed levels of Lalanine, a competitive inhibition pattern was also obtained (Table 11). This is consistent with either a Theorell-Chance or random mechanism with dead-end complexes (18). However, a random mechanism would not be consistent with the rapid-equilibrium ordered pattern observed with NH: and pyruvate. It is also inconsistent with an ordered mechanism, which would show only one competitive product inhibition pattern (18).
If product inhibition by L-alanine with NH: as the variable substrate were examined at saturating and nonsaturating concentrations of pyruvate, a distinction between mechanisms could be made. For an ordered mechanism with pyruvate adding last, saturation with pyruvate would convert noncompetitive inhibition to uncompetitive inhibition. In an ordered mechanism with pyruvate adding second, saturation with pyruvate would have no effect, and the inhibition by alanine would remain noncompetitive. In a random mechanism, the inhibition would also remain noncompetitive, as NH: can add last and remain reversibly connected to alanine. A Theorell-Chance mechanism would give a unique pattern upon saturation with pyruvate. That is, the noncompetitive inhibition by alanine would be eliminated by saturation with pyruvate (18), causing the intersecting lines to converge to a single line.
However, substrate inhibition by pyruvate prevents saturation with this substrate.' Nevertheless, a series of experiments with NH: varied uersus changing-fixed concentrations of L-alanine at increasing concentrations of pyruvate (below substrate inhibition levels) should indicate a trend toward elimination of inhibition (18). The patterns obtained with 0.5 and 2.3 mM pyruvate are shown in Figs. 5 and 6, respectively. The plots of these experiments, which were otherwise carried out under the same conditions, show a compression of the lines at the higher pyruvate concentration. This implies that at saturating pyruvate, the lines would compress into a single line, indicating no inhibition. When a tertiary plot of the slopes of these plots, including data from an experiment at 1.25 mM pyruvate was made, the trend at infinite pyruvate concentration approached zero (no change in slope). This is consistent with an elimination of inhibition at saturating pyruvate. These experiments provide further evidence for the existence of a Theorell-Chance mechanism.
Determination of the True Maximum Velocity-A prediction of the Theorell-Chance mechanism is that at infinite substrate concentration, the Vmax should also become infinite (16). The true V, , in a ter-reactant mechanism can be determined by varying all the substrates simultaneously, maintaining their concentrations a t a constant ratio to each other (18). Extrapolation to infinite substrate concentration in a double-reciprocal plot should give the true V, , , . This was done for alanine dehydrogenase in the amination direction,     in Fig. 7 describe a curve that is concave up at low pyruvate concentrations, as expected for an ordered mechanism (18). The dashed line that is tangent to this curve passes through the origin, indicating that the velocity extrapolates to infinity a t infinite substrate concentration.
Substrate Inhibition-L-Alanine and pyruvate both give substrate inhibition a t higher levels of substrate.' Ammonium also shows apparent substrate inhibition at concentrations greater than 50 mM. When NH: was varied at different fixed levels of pyruvate sufficient to give substrate inhibition, the pattern in Fig. 8 was obtained. At noninhibitory levels of Product inhibition by L-alanine with ammonium varied. Pyruvate concentration was 2.3 mM. Data were fit to Equation 6 (see "Experimental Procedures"). The u value for the fit to Equation 6 was 0.020. pyruvate (dashed line), the pattern was intersecting to the left of the l / u axis as expected. As pyruvate was increased to substrate inhibition levels (solid lines), the lines eventually became parallel, indicating uncompetitive substrate inhibition. The uncompetitive substrate inhibition was confirmed by the concave up intercept replot. This is consistent with an ordered mechanism (18) with pyruvate adding to the E-NAD complex, the enzyme form to which L-alanine normally binds (Scheme 1). The combination of pyruvate as a dead-end inhibitor with any other enzyme form would give noncompetitive or competitive inhibition. It is also inconsistent with a random mechanism, which would give a noncompetitive pattern. When substrate inhibition by L-alanine was analyzed at variable concentrations of NAD (64-170 PM), which were within the range that results in linear response on doublereciprocal plots, a similar pattern was obtained (Fig. 9). At higher concentrations of alanine, the lines again became parallel, indicating uncompetitive substrate inhibition. The replot of the intercepts confirmed the uncompetitive substrate inhibition since it was concave up at the l / u axis. This is consistent with an ordered mechanism, with alanine binding to the form of the enzyme that normally binds pyruvate. Thus, the substrate inhibition patterns imply that the substrates bind in a compulsory ordered manner in both directions.
When the apparent substrate inhibition by NH+ was examined, the pattern in Fig. 10  The iminopyruvate may then undergo catalysis by the enzyme to yield the normal products. Their analysis thus identifies this pattern as being due to an alternate reaction rather than true substrate inhibition.

Dead-end Inhibition-To provide
further support for the addition of pyruvate following the addition of NH:, dead-end inhibition experiments were carried out with the pyruvate analogs: oxamate, propionate, and methylglyoxal. When pyruvate was varied in the presence of each of these analogs, however, lines intersecting to the left of the l / u axis were obtained, rather than the expected competitive inhibition pattern (Table 111). There are two ways such a pattern could be generated. First, if the pyruvate analog were adding to the enzyme before the addition of pyruvate and reversibly connected to it, a noncompetitive inhibition pattern would result. Second, these patterns could be generated, if the pyruvate analogs are adding to two different enzyme forms. The pyruvate analogs could add to the same enzyme form as pyruvate, giving a slope effect, as well as adding to the E-NAD complex, the form which normally binds L-alanine, giving an intercept effect (Scheme 1). The combined effect of binding to both enzyme forms would give a slope and intercept effect, produc-

-6.5
since the pyruvate analog either binds to the same enzyme form as NH: or upstream from it. In the second case, the pyruvate analog would be expected to show uncompetitive from NH: and is irreversibly connected to it. When this experiment was done, using propionate as the pyruvate analog (Table 111), the pattern was uncompetitive. This concurs with -8.5 0 the substrate inhibition data in which high levels of pyruvate were uncompetitive with respect to NH:. Thus, it is not too -9 surprising that the pyruvate analogs behave in the same way.  (Table 111). P-Hydroxypyruvate has curved line in the log Vplot was not fit but represents the best curve about 15% of the activity of pyruvate.' Thus it would be through the data points. expected to bind to the same form of the enzyme as pyruvate. When pyruvate was varied at changing-fixed concentrations of P-hydroxypyruvate (Table 111), a noncompetitive pattern was obtained, consistent with the results of the dead-end inhibition experiments.
These results, together with the dead-end and substrate inhibition experiments, and the initial velocity data, confirm the ordered addition of NH: and pyruvate. 0 insight into the mechanism of alanine dehydrogenase, the variation of kinetic parameters with p H was studied. The log V/K and log V plots for NH: are shown in Fig. 11. The log 5 -5.35 -V / K profile fit Equation 9, which describes a bell-shaped curve with slopes equal to 1. The profile gave a pK of 9.0 on 8 the basic side of the curve, which is very close to the pK of -5.n5 9.2 for NH:. The acidic side of the curve gave a pK of 8.1, due, presumably, to a group on the enzyme which is responsible for binding NH:. Since the log V/K rises with an increase in pH, this group must be unprotonated in order to bind - 6.2 . , , . I . , , , , , . , , I , , . , I . , , , , , , , , I , , ,~ NH:. The decline in log V/K after the pK of NH: implies 7 7.5 8 8.5 9 9.5 10 1( that NH: rather than NH2 is the sDecies that binds to the pH " -4.5 p H Profiles of Kinetic Parameters-In order to gain further -4*925 -log V/K profile fit Equation 8, which describes a single group that must be protonated in order to bind the substrate. The apparent pK of this group was 8.8. The results of the pH studies are listed in Table IV.

DISCUSSION
Kinetic Mechanism-The intersecting patterns obtained with the initial velocity experiments clearly indicated a sequential as opposed to a ping-pong mechanism (Table I). Thus, all of the substrates must be bound to the enzyme before any products are released.
The mechanism appears to be ordered as opposed to random on the basis of several lines of evidence. Substrate inhibition by both L-alanine (Fig. 9) and pyruvate (Fig. 8) was uncompetitive, which would be predicted for an ordered rather than a random mechanism (18). The initial velocity experiment with pyruvate and NH: gave a unique pattern expected for a rapid-equilibrium ordered mechanism (Fig. 3). This pattern would not be obtained with a random mechanism. Product inhibition by alanine with NH: varied at increasing concentrations of pyruvate (Figs. 5 and 6) also supported an ordered addition, as discussed previously, rather than random addition. This ordered addition of NH: followed by pyruvate was also supported by dead-end inhibition studies, which showed that pyruvate analogs added after NH: and NADH (Table  111). This was further supported by experiments with the alternate substrate, P-hydroxypyruvate (Table 111).
Although the substrates add in a compulsory order to the enzyme, the mechanism is not a simple ordered one. The product inhibition studies gave two competitive inhibition patterns: one with the substrate-inhibitor pair NAD-NADH, and the other with pyruvate and alanine (Table 11). A simple ordered mechanism, however, would only give one competitive product inhibition pattern, due to the two reactants binding to the free enzyme. That is, the first substrate to add and the last product to come off the enzyme would give the competitive pattern. The two competitive inhibition patterns are consistent with either a Theorell-Chance or a random mechanism with dead-end complexes forming (18). In the Theorell-Chance mechanism, the additional competitive inhibition pattern is obtained with the last substrate to add and the first product to be released. This is due to the rapid breakdown of the central complexes to undetectable levels, so that the enzyme forms to which these reactants bind appear to interconvert directly. A rapid-equilibrium random mechanism would give all competitive product inhibition patterns unless substrates are able to bind to enzyme forms as dead-end inhibitors. In that case, they would produce apparently noncompetitive inhibition patterns due to the combination of competitive product inhibition and uncompetitive dead-end inhibition. However, as demonstrated under "Results," a random mechanism does not hold for this enzyme. Therefore, the Theorell-Chance fits best with the data. The product inhibition pattern with NH: varied at changing-fixed concentrations of pyruvate gives strong support to this conclusion. As pyruvate concentration was extrapolated to infinity, product inhibition by alanine was eliminated. As demonstrated under "Results," this would only hold in a Theorell-Chance mechanism (18). This is because alanine cannot convert E-NAD back to E-NADH-NH: when pyruvate is saturating, which forces the flux of enzyme forms back to E-NAD. The Theorell-Chance mechanism was further confirmed by the fact that the velocity extrapolates to infinity a t infinite substrate concentrations (Fig. 7). A Theorell-Chance mechanism was suggested for alanine dehydrogenase from Mycobacterium tuberculosis (19). Departure from a simple ordered mechanism was also indicated by the rapid-equilibrium rather than steadystate addition of NH:. This was demonstrated by the initial velocity patterns of NH: and pyruvate (Figs. 3 and 4, Table  I). The Theorell-Chance mechanism has been demonstrated for only a few enzymes ( Ref. 22 and references therein).
Since NADH and NAD are the first to add and the last to be released (Table 11), and pyruvate adds after NH: (Figs. 3  and 4, Table 11), the order of addition for the forward reaction is: NADH, NH:, and pyruvate (Scheme 1). For the reverse reaction, since NAD was the first to add and alanine is reversibly connected to pyruvate (Table 11), the order is: NAD, followed by alanine (Scheme 1).
Thus, the mechanism determined for alanine dehydrogenase from soybean bacteroids at pH 8.5 is an ordered Ter-Bi Theorell-Chance with rapid-equilibrium binding of NH: in the second position (Scheme 1). This mechanism was further supported by fitting the initial velocity data of pyruvate and NH: to Equation 10 (Fig. 3). This mechanism differs somewhat from the mechanism determined for the enzyme from B. subtilis (3), which has a Ter-Bi ordered mechanism with NH: adding after pyruvate. In this organism, as in most other organisms, the enzyme appears to favor the deamination reaction due to its high K,,, for NH3. In fact, NH3 does not actually bind to the enzyme but simply chemically reacts with the E-NADH-pyruvate complex (3). This would explain the binding order (NADH, pyruvate, NH3), because the NH3 cannot react until the other substrates are in place.
The kinetic mechanisms of alanine dehydrogenase from B. sphaericus ( 5 ) and P. freudenreichii (6) are also reported to be different from that of B. subtilis (3). The proposed mechanism of the enzyme from B. sphaericus is similar to that determined for soybean bacteroids. Ohashima and Soda ( 5 ) found a completely ordered mechanism with NADH, NH:, and pyruvate adding in that order. The deamination reaction was also the same, with NAD adding first, followed by alanine. Their data, however, showed linear NAD uersus alanine initial velocity plots and no Theorell-Chance mechanism at pH 9.0.
The mechanism for alanine dehydrogenase from P. freudenreichii was reported to be partial random with alanine and NAD adding in ordered fashion and NAD adding second (6). In the amination reaction, NH3 was reported to add first, followed by random addition of pyruvate and NADH.
pH Dependence of Kinetic Parameters-The log V/K profile for NH: on the basic side declined after the pK for NH3 (9.2), implying that the form that binds to the enzyme was NH:. Alanine dehydrogenase from B. subtilis, on the other hand, had a log V/K profile that rose with increasing pH and leveled off at pK 8.9, implying that NH3 was the true substrate in this organism (4). With the soybean bacteroid alanine dehydrogenase, the pK of 8.1 on the acidic side was presumably the group on the enzyme that binds NH: and must be unprotonated for activity. The log V profile for the enzyme from both organisms showed no variation with pH.
The pK of 8.8 from the log V/K plot for pyruvate with the B. japonicum alanine dehydrogenase may be due to the binding of NH:, since it must bind before pyruvate. The log V / K for pyruvate for the B. subtilis enzyme, on the other hand, had a pK of 7.9 (4).
In conclusion, there are several characteristics of the alanine dehydrogenase mechanism from B. japonicum bacteroids that would influence its role in the nodule. Rather than simply diffusing in and chemically reacting with the enzyme-substrate complex, NH: appears to bind directly to the enzyme. This may partially explain why the apparent K,,, for NH: was substantially lower in this organism than it is in other organisms. The fact that the enzyme binds NH: rather than NH3 would enable the enzyme to use the form of nitrogen that would be most abundant at physiological pH. Finally, the Theorell-Chance mechanism would help make most efficient use of the rapid-equilibrium binding of NH:. That is, since NH: binds and dissociates very rapidly, in order to make most efficient use of this substrate, the following steps should be as rapid as possible. In a Theorell-Chance mechanism, catalysis or release of the first product are theoretically capable of becoming infinite (16). Therefore, a greater percentage of the NH: that binds to the active site will be converted to products before it has time to dissociate than if catalysis were slower. All of these traits imply alanine dehydrogenase functions as an NH: assimilatory enzyme. The reaction is reversible, however, and with the proper pH, substrate concentrations, and reductant levels, the amination reaction could play an important role. In fact, it is possible that the enzyme could operate in both directions at different times during nodule development. Further studies of the enzyme mechanism coupled with physiological studies should help to further elucidate its role in nodule development and metabolism.