Reaction between reduced diphosphopyridine nucleotide and glutamate oxalacetate transaminases.

Abstract DPNH can be bound to pig heart cytoplasmic and bovine liver mitochondrial glutamate oxalacetate transaminases. The physiological function of this binding is not known, but DPNH does protect the mitochondrial enzyme from inactivation. Not only can DPNH be bound, but in the presence of ammonium ions it can be oxidized to DPN. This reaction is inhibited by phenylhydrazine. Consequently, it is believed that in this reaction a keto group which is tightly bound to the transaminase is reductively aminated. The product of the reaction other than DPN has not yet been identified. If DPNH with tritium on the B-side is incubated, then tritium is transferred from DPNH to water. While some pyridoxamine phosphate can be produced, it is produced in considerably lesser amounts than DPN. These reactions are apparently not the result of an artifact induced during the purification of the enzymes; they are quite slow and apparently take place independent of reactions between glutamate dehydrogenase and transaminase-bound pyridoxal or pyridoxamine phosphate. Additional evidence is presented which suggests that glutamate dehydrogenase can form an enzyme-enzyme complex with the transaminase, and in the presence of DPNH (or TPNH) plus ammonium ions, glutamate dehydrogenase reacts with transaminase-bound pyridoxal phosphate to produce equal amounts of DPN (or TPN) and pyridoxamine phosphate. In the presence of DPN (or TPN), glutamate dehydrogenase reacts with transaminase-bound pyridoxamine phosphate to produce equal amounts of DPNH (or TPNH) plus pyridoxal phosphate. In these reactions glutamate dehydrogenase is apparently reacting directly with pyridoxal or pyridoxamine phosphate. That is, these reactions are apparently not mediated by some keto or amino acid. Glutamate dehydrogenase must have direct physical contact with the transaminase for there to be interaction between the two enzymes. In the enzyme-enzyme complex, glutamate dehydrogenase is the actual catalyst and transaminase-bound pyridoxal or pyridoxamine phosphate is the substrate.

SUMMARY DPNH can be bound to pig heart cytoplasmic and bovine liver mitochondrial glutamate oxalacetate transaminases. The physiological function of this binding is not known, but DPNH does protect the mitochondrial enzyme from inactivation.
Not only can DPNH be bound, but in the presence of ammonium ions it can be oxidized to DPN. This reaction is inhibited by phenylhydrazine. Consequently, it is believed that in this reaction a keto group which is tightly bound to the transaminase is reductively aminated.
The product of the reaction other than DPN has not yet been identified.
If DPNH with tritium on the B-side is incubated, then tritium is transferred from DPNH to water. While some pyridoxamine phosphate can be produced, it is produced in considerably lesser amounts than DPN. These reactions are apparently not the result of an artifact induced during the purification of the enzymes ; they are quite slow and apparently take place independent of reactions between glutamate dehydrogenase and transaminase-bound pyridoxal or pyridoxamine phosphate.
Additional evidence is presented which suggests that glutamate dehydrogenase can form an enzyme-enzyme complex with the transaminase, and in the presence of DPNH (or TPNH) plus ammonium ions, glutamate dehydrogenase reacts with transaminase-bound pyridoxal phosphate to produce equal amounts of DPN (or TPN) and pyridoxamine phosphate.
In the presence of DPN (or TPN), glutamate dehydrogenase reacts with transaminase-bound pyridoxamine phosphate to produce equal amounts of DPNH (or TPNH) plus pyridoxal phosphate.
In these reactions glutamate dehydrogenase is apparently reacting directly with pyridoxal or pyridoxamine phosphate. That is, these reactions are apparently not mediated by some keto or amino acid. Glutamate dehydrogenase must have direct physical contact with the transaminase for there to be interaction between the two enzymes.
In the enzyme-enzyme complex, glutamate dehydrogenase is the actual catalyst and transaminase-bound pyridoxal or pyridoxamine phosphate is the substrate.
It has been previously demonstrated that in the presence of TPNH, NH4C1, tmnsaminases and gIutamste dehydrogenasc (L-glutamate NAD(P) oxido-reductase EC 1.4.1.3), TPNH is oxidized to TPN and the transaminase is converted from the pyridoxal to pyridoxamine form. If some keto acids are added to this enzyme-enzyme system, then the rate of TPNH oxidation is increased, pyridoxamine phosphate does not accumulate and the corresponding amino acid is produced. These results suggested t'hat in this enzyme-enzyme system the added keto acid is bound to, and reacts with, the pyridoxamine phosphate form of the transaminase (1, 2). This communication describes a reaction between I)PNH, ammonium ions and glutamate oxalacetate transaminase (L-aspartate-2 oxoglutarate amino transferase EC 2.6.11). 111 this reaction DPNH is converted to DPN even in the absence of glutamate dehydrogenase. TPNH cannot replace DPNH and even t,he transaminase apoenzyme reacts. Added keto acids have no effect on the initial rate of this reaction.
In addition reactions with DPN, glutamate dehydrogenase, and the pyridosamine phosphate form of the transaminase are described.

MATERIALS AND METHODS
Enqmes and Reagents-The bovine liver mitochondrial glutamate dehydrogenase and glutamate oxalacetate transaminase used in these experiments were prepared and crystallized with previously described methods (3-7).
The pig heart glutamate oxalacetate and glutamate pyruvate transaminases were obtained from Boehringer Mannheim Corp. The pig heart malate soluble during dialysis. Therefore, this solution was clarified by centrifugation, and experiments were performed with the supernatsnt solution.
The pyridosamine form of t,he transaminase was prepared by adding glutamate (200 mtil) and dialyzing or chromatographing on Sephadex G-25.
The concentrations of glut,amate, malate, and lactate dehydrogenase were measured spectrophotometrically using previously determined values of the extinction coefficients of these enzymes (16)(17)(18).
Measurements of Producfs-The concentration of DPN produced in t,hc presence of DPNH and enzymes was calculated from the change in absorbance at 340 and 360 nm (1, 2). The validity of these calculations was confirmed by measuring the amount of DPN fluorometrically after deproteinization and destruction of DPNH (19). Also, in some cases the incubation mixture was chromatographed on a column of DEAE-Sephadex X-25 of sufficient size to separate DPN from DPNH (20) (see below).
The fractions from t'his column were assayed for DPN fluorometrically and by measuring the increase in optical density at 340 nm, following the addition of aliquots from this column to solutions of 0.028 mg per ml of glutamate dehydrogenase, 40 mM glutamate and 0.06 M Tris acetate, pH 8.8. It was found that all three assays (fluorescence, spectroscopy, and enzymatic) agreed with respect to the amount of DPN produced.
Standard Enzyme Assays-Glutamate dehydrogenase activity was measured in the presence of 2 mM cr-ketoglutarate, 50 mM NH&Y, and 100 PM DPNH (3). Glutamate oxalacetate transaminase activity was measured in the presence of 1 mM oxalacetate and 40 mrvr glutamate (4). Transaminase dehydrogenase activity (conversion of the transaminase from the pyridoxal to pyridoxamine form) was measured in the presence of 100 PM TPNH, 50 mM NH&l and 0.02 to 0.08 mg per ml of glutamate dehydrogenase (I, 2). The effect of transaminase on the reaction between glutamate dehydrogenase and pyruvate was measured in the presence of 100 PM TPNH, 50 mM NH&l, 0.9 mM pyruvate and 0.5 mg per ml of glutamate dehydrogenase (1, 2). The assay used to measure the rate of oxidation of DPNH in the presence of transaminase and absence of glutamate dehydrogenase or keto acids (unless specified to the contrary) was 100 PM DPNH and 50 111~ NH&l.
All of these assays were performed in 0.025 M sodium arsenate, 0.1 mrvr EDTA, pH 7.8 at 25". In some cases (when radioactive compounds were used) 0.02 M Tris chloride was present in the assays for DPNH oxidizing activity of the transaminase.
The addition of Tris chloride to this assay did not alter results, and similar results were obtained if only Tris buffer was used. All assays were performed with a Gilford model 2000 recorder and a Beckman DU monochromator as previously described (l-3).
Electrophoresis-Electrophoresis on cellulose polyacetate strips wit,h a Gelman electrophoresis chamber was performed according to previously described methods (3). Alanine with tritium on the a-carbon was prepared by incubating the purified tritiated B-form of DPNH (100 PM) with pyruvate (1.0 m&r), NH&l (50 mM), and glutamate dehydrogenase (0.5 mg per ml) in 0.02 M Tris chloride, pH 7.5 for 6 hours at 24".
Glutamate with tritium on the a-carbon was prepared as described above for alanine with oc-ketoglutarate (10 InM) being used instead of pgruvatc.
Acid mashed Norit (0.1 g per ml) was added to the supernatant solution (to remove coenzymes) and the mixture was then filtered.
Glutamate made wit.11 either method gave identical experimental results.
Column Chromatography-The tritiated DPNH, glutamat,c, and alanine synthesized in the above reactions were purified by chromatographing on a column of DEAE-Sephadex A25 (1.5 X 13 cm) which had been equilibrated with 0.02 M Tris chloride, pH 7.5. The column was developed with a linear KC1 gradient consisting of 200 ml of 0.02 M Tris chloride, pH 7.5, in the mixing vessel and reservoir.
The reservoir also contained 0.8 M KCl. Fractions with a volume of 2.8 ml were collected from the column. The entire procedure was performed at 24". Only fractions of the highest purity and concentration (peak fractions) were used in subsequent experiments.
Reaction mixtures of DPNH, NH&I, transaminase plus other additions, as well as samples of transaminase or glutamate dehydrogenase alone were chromatographed on an identical type of DEAE column. Chromatography of tritiated water on Dowex columns was performed as described previously (21). Samples of the tritiated glutamate and alanine which were synthesized were applied to amino acid analyzer.
It was found that these compounds were eluted as the standard amino acids. Radioactivity and ninhydrin peaks from the analyzer were identical Only freshly prepared tritiated amino acids were used in these experiments.
Amino Acid Analysis-Amino acid analyses were performed with a Beckman model 120C amino acid analyzer.
Solutions containing enzymes were deproteinized (by heating at 100" for 1 to 3 min), clarified by centrifugation, and brought to pH 2.0 (with HCl) before applying to the analyzer.
Standard solutions of amino acids were added to constituents of reaction mixtures and treated in a similar manner.
It was found that these constituents (Tris, coenzyme, etc.) and this treatment had no sig-nificant effect on the elution time of the amino acids studied. In some experiments the eluate from the analyzer (with or without ninhydriu) was collected at 2-min intervals and assayed for radioactivity. Assay joo7. Radioactivity-These assays were performed with a Packard Tri-Carb (model 3310) liquid scintillation spectrophotometer at 4". Aliquots (100 ~1) of samples were added to 10 ml of either Bray's solution (22) or another previously described standard scintillation solution (23). Both solutions gave comparable results with respect to the distribution of tritium in various compounds.
Solutions to be assayed were incubated for 15 min at 4' prior to assaying.

Soluble Glutamate
Oxalacetate Transaminase There is no oxidation of TPNH in the absence of glutamate dehydrogenase and presence of TPNH, NH4+, and transaminase (1, 2). However, if DPNH, NH.,+, and transaminase are incubated, DPNH is oxidized to DPN in the absence of glutamate dehydrogenase.
The init'ial velocity was the same if 0.02 M Tris chloride, 0.1 mM EDTA, pH 7.8, was used instead of arsenate buffer.
In these reactions pyridoxamine phosphate is not produced and the specific activity (initial rate of DPNH oxidation per transaminase peptide chain) of the pyridoxamine phosphate form of the transaminase (prepared by adding glutamate a.nd dialyzing (2)) was about 1.5.fold faster than that of the pyridoxal phosphate form. The specific activity of the resolved transatninase was essentially the same as that of the holoenzyme. The phenylhydrazine-treated enzyme did not oxidize DPNH at a measurable initial rate. The above results are summarized in Table I. The transaminase preparation was chromatographed on a column of Sephadex G-200 (1). The fractions from this column containing the peak t'ransaminase activity were concentrated by dialyzing versus a solution 80% saturated with respect to ammonium sulfate, pH 7.0. The ammonium sulfate suspension was centrifuged, the precipitate solubilized in 0.025 M sodium arsenate, 0.1 mM EDTA, pH 7.8, and dialyzed two times versus 2 liters of arsenate buffer.
It was found that the above treatmcnts of the transaminase preparation (chromatography, concentrating with ammonium sulfate and dialysis) did not alter the specific activity of DPNH oxidation by the transaminase. This transaminase was then chromatographed on DEAE-Sephadex. Three widely separated peaks of transaminase activity were eluted.
On the basis of the absorption spectrum and elution profile of these three peaks it was concluded that the two most anionic peaks were the previously described fi-and ysubforms of the transaminase (24). The specific activity of DPNH oxidation with the @subform and the least anionic fraction (which was mainly the oc-subform) were, respectively, 3fold lower and 1.7-fold higher than that of the original prepara- tion. The y-subform did not oxidize DPNH. These results are summarized in Table I. 9 series of kinetic experiments were performed with the least anionic (predominantly a-subform) fraction from the DEAE-Sephadex column.
A plot of the rate of DPNH oxidation as a function of transaminase concentration is shown in Fig. 1. It can be seen that this is essentially a linear relationship. These experiments were performed with either the pyridoxal or pyridoxamine phosphate forms of the transaminase (prepared by adding glutamate and dialyzing).
It can be seen that DPN is much more inhibitory t.han TPN.
Also, DPN inhibits slightly more when the transaminasc is in the pyridoxamine form.
Time Course of DPN Production--As shown in Fig. 4, about 2 moles of DPN are produced per mole of incubated transaminase peptide chain in 40 min when transamiuase (22 PM with respect t,o peptide chains), DPNH (100 PM), and NH&l (50 mM) are incubated.
An additional 2 moles of DPN per mole of peptide chain are produced slowly over the next 2 hours.
In similar experiments it was found that the addition of the previously mentioned (see section on initial rate) keto acids, amino acids, or purine nucleotides did not alter the time course of DPN production shown in Fig. 4. Similarly, dialyzing the transaminase versus adding dithioerythritol (0.8 mM) had no effect on these results.
Also, substituting 0.02 M Tris chloride, 0.1 mM EDTA, pH 7.8, for arsenate buffer did not alter the time course of DPN production.
The time course of DPN production was similar if either the pyridoxal or pyridoxamine form of the transaminase was incubated (Fig. 4). The addition of DPN (but not TPN) decreased the amount of DPN produced (Fig. 4). As shown in Table I, less DPN is produced in 4 hours if the /3-subform of the transaminase is iucubated and the y-subform does not oxidize DPNH.
While the initial velocity is faster with the predominantly c-r-subform than with the mixture of subforms, the amount of DPN produced after 4 hours incubation is the same in both cases (Table I). If a low concentration of transaminase is incubated with 100 PM DPNH and 50 mM NH&l for 3 hours, then as many as 15 moles of DPN rather than 3 to 4 are produced per peptide chain (Fig. 5). Thus, the amount of DPN produced during 3 hours incubation is apparently related to the initial velocity of DPNH oxidation (which is directly proportional to the concentration of transaminase incubated (Fig. l)), the amount of DPN produced (which inhibits DPNH oxidation (Figs. 3 and 4)), t.he initial ratio of DPNH per transaminase peptide chain and the amount of DPNH remaining at any given time. That is, DPNH oxidation does not stop when a certain amount of DPN per incubated peptide chain has been produced but apparently proceeds until the ratio of DPN to DPNH concentration is quite high. This is also illustrated by the results shown in Table II. If a high concentration of transaminase (40 pM with respect to peptide chains) is incubated with a high concentration of DPNH (1 mw) and 50 mM NH&l for 20 hours, then at the end of this time about 20 moles of DPN are produced per P-7  I  I  I  I  I  I  I  I  I  I  I  I  I  I   mole of transaminase peptide chain. Therefore, again in the presence of a high initial ratio of DPNH to transaminase the reaction slowly proceeds until the ratio of DPN to DPNH con centration is quite high or most of the incubated DPNH has been converted to DPN.
This is not the case if the apoenzyme or the pyridoxamine phosphate form of the transaminase is incubated (Table II) in the presence of a considerable excess of DPNH.
In these latter cases only 6 or 8 moles, respectively, of DPN are produced per mole of incubated peptide chain even after 20 hours incubation.
Similarly, if the apoenzyme is incubated with 0.1 or 0.2 mM DPNH, then after 4 hours 2 to 3, rather than 4 to 5, moles of DPN are produced per incubated peptide chain (Table II) or malate dehydrogenase to DPSH (100 PM), NH&l (50 mnf) and transaminase (20 pM with respect to peptide chains) had no effect on the initial rate of DPNH oxidation or the time course of DPN production in 3 to 4 hours. The levels of these dehydrogenases used in t'hese experiments were sufficienbly high to react with low levels (20 to 100 pM) of their keto acid substrates. When the level of glutamate dehydrogenase is higher (0.01 mg per ml), then this enzyme increases the initial rate of DPNH oxidation until a plateau is reached when the level of this enzyme is about 0.1 mg per ml. At this concentration glutamate dehydrogenase increases the initial rate of DPNH oxidation about 1.4.fold.
The addition of glutamate dehydrogenase not only increases DPN production but now pyridoxamine phosphate is rapidly produced. Fig. 6 shows a comparison between the reactivity of DPNH versus TPNH in the presence of low levels of glutamate dehydrogenase. It can be seen that in the presence of low levels of glutamate dehydrogenase, DPN is produced in excess of pyridoxamine phosphat,e.
In the presence of any level of glutamate dehydrogenase, pyridoxamine phosphate and TPN are produced in equal amounts.
In these experiments the initial rate of oxidized pyridine nucleotide production is about 5-fold faster when DPNH rather than TPNH is the coenzyme.
However, the initial rate and amount of pyridoxamine phosphate produced over the entire incubation period is essentially the same with either coenzyme.
Results obtained in the presence of DPNH, NH&l, transaminase, and a higher "optimal" concentration of glutamate dehydrogenase (0.11 mg per ml) are shown in Fig. 7 Even when the transaminase is completely in the pyridoxamine phosphate form, DPN is still slowly produced.
The amount of product produced after 1 hour incubation in the presence of various concentrations of glutamate dehydrogenase is shown in Fig. 8. In the presence of high levels of glutamate dehydrogenase, 40 /.LM (with respect to peptide chains) and 20 PM transaminase, DPX and pyridoxamine phosphate are produced in essentially equal amounts.
The above results raise the possibility that the tra.nsa.minase has tightly bound keto acids. Conceivably these keto acids in the presence of reduced pyridine nucleotides, NH+4 and glut'a-7561 mate dehydrogenase could leave the transaminase, bind to glutamate dehydrogenase, and be converted to amino acids. These amino acids would then leave glutamate dehydrogenase, react with transaminase-bound pyridoxal phosphate to produce the pyridoxamine phosphate form of the transaminase. The rejuvenated keto acid would be too tightly bound to this latter form to leave the transaminase and react again with glutamate dehydrogenase.
Thus, there would not be an enzyme-enzyme complex but merely a transfer of keto acids from the transaminase to glutamate dehydrogenase.
To test this possibility, a l-ml solution of transaminase (48 pM with respect to peptide chains), NH&l (50 mM) and TPNH (100 PM) was placed into a dialysis sac and dialyzed versus a j-ml solution of TPNH (100 PM), NH&l (50 mM) and glutamate dehydrogenase (0.6 mg per ml). Both the inside and outside compartments of the system contained 0.025 M sodium arsenate, 0.1 InM EDTA, pH 7.8. The sample was then dialyzed for 6 hours at 25". At this time the dialysis sac was removed and the sample dialyzed versus 2 liters of the arsenate buffer. The dialyzed transaminase was in the pyridoxal phosphate form. Adding glutamate converted the enzyme to the pyridoxamine phosphate form. In a control experiment the same amount of t'ransaminase not enclosed in a dialysis sac was incubated within an identical 5-ml solution of TPNH, NH&l, and glutamate dehydrogenase.
In these control experiments the transaminase is rapidly converted to the pyridoxamine phosphate form and remains in this form during subsequent dialysis.
Therefore, it seems likely that the two enzymes must have contact with each other if the transaminase is to be converted from the pyridoxal to the pyridoxamine phosphate form.
E$ect of Ammonium Ions-The endogenous concentration of ammonium ions in a solution was decreased by adding glutamate dehydrogenase (0.17 mg per ml) and cY-ketoglutarate (1 mM) to transaminase (36 PM with respect to peptide chains) and DPNH (500 PC(M). If this solution was incubated for as long as 20 hours, then at this time the concentration of DPN was only 114 (IM. However. 50 PM DPN was produced almost instantaneously and this was found to be the level of endogenous ammonium ions in the incubated solution (as measured by assaying the solutions of glutamate dehydrogenase and transaminase used independently for ammonium ions (25)). Therefore, in the presence of ac-ketoglutarate, glutamate dehydrogenase transaminase, and no added ammonium ions, very little DPNH is oxidized (Table II). If corrections are made for the endogenous level of ammonium ions, then only 64 PM DPN or about 2 moles per mole of incubated transaminase peptide chain was produced in 20 hours in the absence of ammonium ions. Neither a-ketoglutarate nor these low levels of glutamate dehydrogenase alone can prevent large amounts of DPNH from being oxidized in the presence of transaminase and 50 mM NH&l.
Properties of Incubated Enzyme-Transaminase (24 PM with respect to peptide chains) was incubated with DPNH (500 pM) plus NH&I (50 ITIM) for 24 hours. At t,his time about 15 moles of DPN were produced per incubated peptide chain. The incubated enzyme solution was then dialyzed.
A quantitative ninhydrin (26) performed on a O.l-ml aliquot of the incubated dialyzed enzyme revealed that there were only two to three more amine groups per peptide chain on the incubated enzyme than a control which was not incubated with DPNH.
If the incubated dialyzed enzyme was deproteinized and applied to the amino acid analyzer, several different amino acids were found. Horn-ever, the total amount of amino acids found was 2 to 3 per incubated peptide chain. Of these t'he most abundant were serine and by guest on March 17, 2020 http://www.jbc.org/ While there is no evidence that pyridoxamine phosphate is produced when low levels of DPNH are incubated with transaminase for 3 to 4 hours, the absorption spectrum of the above described incubated dialyzed enzyme (incubated with high levels of DPNH for a long period of time) suggests that now the transaminase is at least partially in the pyridoxamine phosphate form. This conclusion is supported by the fact that if a-ketoglutarate is added to this incubated dialyzed enzyme, the absorption spectrum slowly shifts to that of the pyridoxal phosphate form of the transaminase (Fig. 9).
The above results reveal that after incubation and dialysis, some amines or amino acids are found.
However, the total amount of amines or amino acids found is considerably less t,han the amount of DPNH oxidized.
If the enzyme is incubated for 24 hours (as described above), not dialyzed, but deproteinized, and applied to the amino acid analyzer, then even fewer amino acids are found than in the dialyzed sample (Table III).
Since serine and glycine were the most abundant amino acids found, the transaminase (30 PM) was incubated with DPNH (10 to 100 PM) plus glyoxalate reductase (0.05 mg per ml) for as long as 48 hours in the absence of NH*+.
Incubating with glyoxalate reductase did not result in any more DPN production than the small amount observed in the absence of NH4+. At the end of the 48.hour incubation NH&l (50 mM) was added to the transaminase glyoxalate solution and a control solution which did not contain this latter enzyme. In both cases DPNH was oxidized in the typical manner observed in the presence of NH&l.
Since several different amino acids were eluted when the incubated transaminase was applied to the amino acid analyzer, the transaminase (40 PM) was assayed for peptidase activity with denatured casein (27) 11'1~. 9. Absorption spectrum of the cytoplasmic transaminase. Curve A shows the absorption of a 36 NM (with respect to peptide chains) solution of the pyridoxamine phosphate form of the transaminase. Curve B shows the absorption spectrum of the enzyme after it had been incubated in 9 hours with 760 PM DPNH and 50 rnM NH&l and dialyzed. Curves C and D show the absorption spectrum of the enzyme 1 and 20 min after 2 mnx cr-kctoglut,arat.e has been added to the incubated dialyzed enzyme (these absorption spectra have been corrected for the absorbance of cu-ketoglutarate).
The absorption spectrum shown in Curre D is ident,ical with that of a 36 PM (with respect to peptide chains) solution of the transaminase in the pyridoxal phosphate form. These experiments were performed in the presence of 0.025 M sodium arsenate, 0.1 mM EDTA, pH 7.8 at 25".
Transaminase-bound pyridoxamine phosphate was prepared by incubating the pyridoxal phosphate form with glutamate; DPNH plus NHd+, with or without glutamate dehydrogenase, and TPNH, NH4+ plus glutamate dehydrogenase (Table IV). These incubated solutions were then dialyzed or chromatographed on Sephadex G-25. In the experiments with glutamate sufficient amounts of '*C-labeled glutamate were added so that even trace amounts of glutamate or a-ketoglutarate associated with the transaminase could be detected (12). No radioactivity was detected in the transaminase peaks from the Sephadex G-25 column.
After dialysis or chromatography the transaminase was predominantly in the pyridoxamine phosphate form if this form was prepared by incubating with glutamate or TPNH, NH4+ plus glutamate dehydrogenase.
If this form was prepared by incubating with DPNH plus NHd+, with or without glutamate dehydrogenase, then only about 0.3 to 0.4 mole of pyridoxamine phosphate per transaminase peptide chain was present after dialysis (Table IV). This is in spite of the fact that incubating with DPNH, NH,+ plus glutamate dehydrogenase converts the transaminase completely into the pyridoxamine phosphate form. In the presence of DPN and glutamate dehydrogenase the dialyzed transaminase-bound pyridoxamine phosphate (prepared by any of the above described methods) is rapidly converted to pyridoxal phosphate plus DPNH.
Pyridoxal phosphate and DPNH were produced in equal amounts throughout the incubation.
These reactions were complete in at least 8 min and additional incubation did not alter the amount of product formed. If DPN alone was added to transaminase-bound pyridoxamine phosphate, there was no reaction even if the pH of the solution Other conditions of the incubations were NH.&1 (50 mM), Tris chloride (0.01 M), sodium arsenate (0.025 M), EDTA (0.1 m&f), pH 7.5 at 25" in a volume of 2 ml. At the end of the incubation (various times over 1 hour were used to give nearly complete conversion of DPNH to DPN), the solution was chromatographed on DEAE-Sephadex A-25 as described in the legend to Fig. 10 was increased to 8.8 (by adding Tris). When the enzyme is completely in the pyridoxal phosphate form, it does not react with DPN, with or without glutamate dehydrogenase (Table  IV).
Therefore, these results demonstrate that pyridoxamine phosphate is the only amine on the transaminase which can react with DPN and glutamate dehydrogenase is required for this reaction.
Stereospecificity of Reaction-If the A-form of tritiated DPNH was incubated with transaminase plus NH&l and this solution was then chromatographed on DEAE-Sephadex, it was found that tritiated DPN was produced (Table V). If the &form of tritiated DPNH was incubated and this solution was chromatographed on DEAE-Sephadex, then no radioactivity was found in the fractions containing DPN.
In bhese experiments radioactivity was found in a peak which was eluted immediately after the major transaminase peak (Fig. 10). No radioactivity was found associated with the transaminase.
If the radioactive compound eluted from the DEAE-Sephadex column was applied to the amino acid analyzer, it was eluted as a peak after 28 min. This compound was volatile, it did not react with ninhydrin, it would pass through a dialysis membrane, it did not absorb light, it was not absorbed to Norit, and it did not react as a keto aeid. It was eluted from a Dowex 50 column and distilled like tritiated water (21). The same tritiated compound was produced (as judged by volatility, chromatography on Dowex 50, DEAE-Sephadex, and the amino acid analyzer) if either glutamate or alanine (with tritium on the a-carbon) were incubated with transaminase (Table VI).
It is known that the tritiated product in these latter reactions is tritiated water (28, 29). Therefore, for this reason and since the tritiated compound had the properties described above, it was concluded that when the &form of tritiated DPNH is incubated with transaminase and XH&I, DPN is produced and in this reaction tritium is transferred from
The reaction mixture consisted of the B-form of tritiated DPNH (235 nmoles with a specific activity of 430 cpm per nmole), NH&l (100 pmoles), transaminase (4.5 mg), and glutamate dehydrogenase (0.2 mg) in a volume of 2 ml. The incubation was in 0.02 M Tris chloride, 0.025 M sodium arsenate, 0.1 mM EDTA, pH 7.5 for 1 hour at 24'. At the end of this time the mixture was chromatographed.
The column (1.5 X 13 cm) was equilibrated with 0.02 M Tris chloride, pH 7.5, and elution was performed with an increasing linear gradient of KCl. The mixing chamber cont,ained 200 ml of 0.02 M Tris chloride, pH 7.5, and the reservoir contained an equal volume of this buffer plus 0.8 M KCl. The column was developed at 24". Fractions of 2.5 ml were collected. The shaded area indicates radioactivity.
the B-side of DPNH to water.
The results of several of these experiments are summarized in Table V. It can be seen that similar results were obtained if cy-ketoglutarate or glutamate dehydrogenase was added to the incubation mixture. Also, tritiat'ed water was eluted if the incubated solution was deprot'einized (by heating) before it was added to the DEAE-Sephadex column.
Even if the apoenzyme is incubated with the B-form of trit,iated DPNH plus NH&l, then tritium is transferred from DPT\'H to water.
The elution profiles of the above incubations with the B-form of DPNH on DE,LE-Sephadex were essentially all identical to that shown in Fig. 10. Tritium was not eluted in the transaminase peak in any of these experiments.
E$ect of PuriJication on Reaction-The above experiments were performed with highly purified pig heart cytoplasmic glutamate oxalacetate transaminase.
In one of the purification steps employed, the transaminase is heated in the presence of dicarboxylic acids (12). To test the possibility that this heat step in some way induces an alteration in the t,ransaminase so that it can now react with DPNH plus NH4+, the enzyme was prepared without a heat step. The heart tissue was homogenized in 0.1 M sodium succinate, pH 6.0. The cytoplasmic extract was fract.ionated with ammonium sulfate, the fraction containing transaminase activity was t.hen chromatographed on DEAE-Sephadex and carboxymethyl-Sephadex (12, 24). After chromat,ography on DEAE-Sephadex, the enzyme was sufficiently pure to assay for DPPJH oxidizing activity.
It was found that the t,ransaminase at this stage of purification oxidized DPNH in After incubation, the solutions (1 to 2 ml) were deproteinized and applied to the amino acid analyzer.
Fractions from the analyzer (with or without ninhydrin) were collected at 2-min intervals and radioactivity was measured in these fractions. In all of these experiments less than 7y0 of the amino acid was converted to a keto acid even though there was complete transfer of tritium to water.
Incubations were in 0.02 M Tris chloride, pH 7.5 at 24". The specific activity of the glutamate incubated was 58.5 cpm per pmole and 829 cpm per pmole, respectively, when the glutamate concentration was 1.0 mM and 76 PM. The specific activity of alanine incubated was 8.25 cpm per gmole. the presence of DPNH plus NH4+. Also, in the presence of TPNH, NH,+, glutamate dehydrogenase, and transaminase eluted from the DEAE-Sephadex column, TPN and pyridoxamine phosphate were produced in equal amounts.
These activities paralleled transaminase activity during additional purification procedures.
Therefore, a heat step with dicarboxylic acids is not necessary to induce these activities.
Furthermore, a heat step performed with cY-ketoglutarate (12) on this purified enzyme did not alter these activities.

Xitochondrial
Glutamate Oxalacetate Transaminase The initial velocity of DPNH oxidation with the mitochondrial transaminase is quite slow compared with the cytoplasmic and essentially too slow to measure accurately. Therefore, measurements were made of the amount of DPN produced during prolonged incubations (Fig. 11). This reaction, like the cytoplasmic, is dependent upon ammonium ions, and TPNH is much less reactive than DPNH.
With the mitochondrial enzyme, only 0.6 to 0.7 mole of DPN is produced per incubated peptide chain in 100 min. However, now a small amount (0.2 to 0.3 mole per peptide chain) of pyridoxamine phosphate is produced during this time (Fig. 12). Even if these solutions are incubated for 17 hours, then only 0.9 mole of DPN and about 0.3 mole of pyridoxamine phosphate per peptide chain is produced.
If at this time this solution is dialyzed, then it is found that after dialysis t,he enzyme is in the pyridoxamine phosphate form. If a-ketoglutarate (2 mM) is added to this incubated dialyzed transaminase, then the pyridoxal phosphate form is rapidly produced. If DPNH or TPNH (100 PM (1)) of the incubated dialyzed enzyme is the same as a nonincubated control.
Like the cytoplasmic transamiuase, the addition of cr-ketoglutarate (1 InM) to an incubation of DPNH (100 FM), NHdCl (50 mbf), and mitochondrial transaminase (20 FM with respect to peptide chains) has no effect on DPN production However, in the presence of cu.ketoglutarabe, pyridoxamine ph0sphat.e is not produced.
The pyridoxamine phosphate form of the mitochondrial, unlike the cytoplasmic (prepared by adding glut.amate and dialyzing (2)), does not oxidize DPNH in the presence of NH&l.
If glutamate dehydrogenase is added to DPNH, NH&l, and mitochondrial transaminase, then the initial rate of DPN and pyridosamine phosphate production is increased considerably (Figs. 11 and 12). Initially DPN and pyridoxamine phosphate are produced in equal amounts.
After about 10 min, when about half of the transaminase is in the pyridoxamine phosphate form, DPN exceeds pyridoxamine phosphate production. After prolonged incubation (100 min), the transaminase is completely in the pyridoxamine phosphate form and about I.&fold more DPN than pyridoxamine phosphate has been produced. Unlike experiments performed with the cytoplasmic transaminase, little additional DPN is produced after the mitocbondrial transaminase is completely in the pyridoxamine phosphate form. The initial velocity of reduced pyridine nucleotide oxidation is 3.5-fold greater when DPNH, rather than TPNH, is the coenzyme in the incubation of the mitochondrial transaminase and glutamate dehydrogenase (condit,ions identical with those described in the legends to Figs. 11 and 12). When TPNH (100 PM) is the coenzyme, TPN and pyridoxamine phosphate are produced in equal amounts throughout the course of the incubation (100 min).
A series of experiments were performed with the mitochondrial transaminase, tritiated DPNH, and NH&l.
These experiments were performed essentially the same as those with the cytoplasmic transaminase.
After incubating, the reaction mixture was assayed for DPN and pyridoxamine phosphate, chromatographed on DEAE-Sephadex, and the eluted products assayed. If either the A-or B-forms of tritiated DPNH were incubated with the mitochondrial transaminase in either the presence or absence of glutamate dehydrogenase, then both tritiated DPN and water were eluted from the DEAE-Sephadex columns.
No tritium was found in the transaminase fractions even when the transaminase was eluted predominantly in the pyridoxamine phosphate form, i.e. in the presence of glutamate dehydrogenase. When the B-form of tritiated DPNH is incubated in the absence of glutamate dehydrogenase, then 2-fold more tritium is recovered in the DPN than tritiated water fractions from the column (Table VII).
If the B-form of trjtiated DPNH is incubated in the presence of glutamate dehydrogenase, then 2-fold more tritium is found in the water than DPN fractions eluted from the column.
If the A-form of tritiated DPNH is incubated in the presence of glutamate dehydrogenase, then the reverse is the case, i.e. a-fold more tritium is found in the DPN than water fractions (Table VII).
The above results suggest that there are two different groups on the mitochondrial transaminase (A and B) which in the presence of ammonium ions can react with DPNH.
Group A removes hydrogen from the A-side and Group B from the B-side of DPNH.
The tritiated product produced in the reaction with either Group A or B can then exchange tritium with water. In the above experiments the amount of DPN eluted from the column was equal to the amount of DPN produced in the incubations and loaded on the column. The total counts eluted from the column in the DPN plus water peaks divided by the amount of DPN produced in the incubation was equal to the specific activity of the starting DPNH incubated.
Therefore, the total amount of DPN produced in the incubation and eluted from the column is apparently equal to the sum of the reaction with Group A plus Group B.
If the B-form of tritiated DPNH is incubated, the ratio of tritium recovered in DPN to water is equal to the ratio of the reaction with Group Tritiated DPNH (200 PM and 100 to 300 cpm per nmole), NH&I (50 mM) and mitochondrial glutamate oxalacetate transaminase (20 PM with respect to peptide chains) were incubated for 1 hour in 0.025 M sodium arsenate, 0.1 mM EDTA, pH 7.8 at 25". At the end of t,his time, the solution was assayed for l)PN and pyridoxamine phosphate and chromatographed on DEAE-Sephadex A-25 as described (see "Materials and Methods" and legend to Fig. 10). The amount of tritium in the water and DPN peaks eluted from the column was measured.
The values for the extent. of the reaction with A and B were calculated as described in the text. The result of these calculations are summarized in Table VII.
These calculations suggest that glutamate dehydrogenase increases the reaction with Group B since the addition of this enzyme doubles the reaction with this group and has no effect on the reaction with Group A. Also, in either the presence or absence of glutamate dehydrogenase the amount of Group 13 which reacts is essentially equal to the amount of pyridoxamine phosphate produced. In the absence of glutamate dehydrogenase, ADP has no effect on the reaction of the cytoplasmic transaminase with DPNH.
This nucleotide, however, increases the amount of DPN and pyridoxamine phosphate produced with the mitochondrial transaminase (Figs. 11 and 12). In the presence of DPNH, NH&l, mitochondrial transaminase, and glutamate dehydrogenase, ADP increases the initial rate of both DPNH oxidation and pyridoxamine phosphate production 2-fold. Initially, in both the presence and absence of ADP, Dl'N, and pyridoxamine phosphate production are equal. However, after 5 min incubation, about 1%fold more DPN than pyridoxamine phosphate is produced in the presence of RDP.
At the end of the incubation (100 min) the system is apparently close to equilibrium, the transaminase is completely in the pyridoxamine phosphate form, and the same amount of DPN is produced in either the presence or absence of ADP (1.8 versus 1.6 moles per mole of transaminase peptide chain).
The above experiments with mitochondrial transaminase were performed with this enzyme after it had been crystallized in the presence of cr-ketoglutarate (4-6). Before use in these experiments, the crystalline enzyme is extensively dialyzed, and it has been shown that interactions between this enzyme and glutamate dehydrogenase cannot be explained on the basis of trace amounts of free a-ketoglutarate (1, 2). Experiments similar t'o those shown in Figs. 11 and 12 were performed with the enzyme preparation before incubation in a-ketoglutarate and crystallization. It was found that this preparation oxidized DPNH in the absence of glutamate dehydrogenase at a 2-fold lower specific activity than the crystalline enzyme. This decrease in specific activity is apparently not related to impurities in the precrystalline preparation since the specific act,ivity of the transaminase reaction does not increase markedly with crystallization, and the precrystalline preparation behaves as the crystalline when subjected to electrophoresis on cellulose polyacctate (3). It was, however, found that the precrystalline enzyme \vas predominantly in the pyridoxamine phosphate form. If t,he precrystalline enzyme was incubated with ol-ketoglutarate (3 IIIM) for 10 min and then dialyzed, then it oxidized DPNH in the absence of glutamate dehydrogenase like t,he crystalline enzyme. Protection of Mitochondrial Transaminase by DPNH--To further confirm that DPNH can be bound to the transaminase, low concentrations of the mitochondrial enzyme (0.014 PM with respect to peptide chains) were incubated with DPNH. &it various time intervals O.l-ml aliquots of the incubated solutions were withdrawn and assayed for transaminase activity in a l-ml solution of a coupled assay system with ol-ketoglutarate, aspartate, DPNH, and malate dehydrogenase (4). It was ascertained that the small amounts of DPNH added to the assays from the incubated solution (the concentration of DPNII in the assays ranged only from 200 to 250 ELM) had no effect on t#he initial velocity of the transaminase reaction.
That is, the initial velocity of this reaction was the same after the transaminase had been incubated for 1 min, regardless of the concentration of DPNH in the incubation mixture. However, after longer periods of time, transaminase incubated in the absence of DPNH slowly lost activity while the addition of 400 PM DPNH cornpletely protected the enzyme against inactivation.
A plot of the log of the velocity of transaminase activity versus time was linear for 50 min. A plot of the slope of thcsc plots (k) (the first order rate constant of inactivation) versus the concentration of DPNH incubated is shown in Fig. 13. Protection was specific for DPNH since glutamate (40 mM), cr-ketoglutamtc (1 mM), DPN (1.2 m&f), TPNH (0.3 mM), ADP (5 mM), and ATP (7 mM) did not protect the transaminase against inactivation.
In these experiments no NH4+ is added to the incubation and the concentration of transaminase is quite low. Therefore, there is no DPNH oxidation during the incubation.

Glutamate Pyruvate Transaminase
While this enzyme interacts with glutamate dehydrogenase (1, 2), it does not oxidize DPNH in the presence of NH4f and absence of glutamate dehydrogenase. DISCUSSIOX These results indicate that glutamate-oxalacetate transamnases have at least one tightly bound group other than pyridoxal phosphate which in the presence of NHd+ can react with I>PNH. It is believed that this group is a keto group since the reaction is inhibited by phenylhydrazine.
The group is apparently tightly bound to the transaminase because the enzymes can be extensively dialyzed and chromatographed without loss of this act,ivity. No known keto acids are so tightly bound to the pyridoxal phosphate form of the transaminase that they would remain with it through these procedures.
It is true that the y-subform of t,he cytoplasmic enzyme does not have this activity; however, this could mean that this subform does not have this group, or it is on this subform but does not react. This does not necessarily mean that this group is loosely bound and removed by chromatography.
Other arguments which support the concept that this group is tightly bound are that a plot of velocity versus 7.5r-----7 transaminase concentration is linear and the addition of several different keto acids does not increase activity.
It is believed that a group other than pyridoxal phosphate can react because the apoenzyme reacts, the pyridoxamine phosphate form of the cytoplasmic enzyme reacts, DPN, but not pyridoxamine phosphate, is produced initially when the cytoplasmic enzyme is incubated, and with both enzymes considerably more DPN than pyridoxamine phosphate is produced.
Several results suggest that oxidation by DPNH is catalyzed by the transaminase and llot some dehydrogenase impurity. Among these are that this activity remains with the transaminase through chromatography on DEAE-Sephadex and Sephadex G-200, and the addition of several enzymes which can react with keto acids does not increase activity.
A plausible possibility is that the reaction is catalyzed by a trace amount of glutamate dehydrogenase. This is because both transaminases have at least some B-specificity with respect to DPNH and ammonia is required for the reaction.
However, this enzyme is separated from the transaminnse on Sephadex G-200. Also, unlike reactions catalyzed by glutamate dehydrogenase, neither transaminase reacts with TPNH.
Furthermore, the addition of a-ketoglutarate to these transaminases in the presence of I )PNH plus ammonium ions does not increase the rate of DPNH oxidation.
The reaction with the mitochondrial transaminase has some A-specificit,y with respect to DPNH.
There is some malate dehydrogenase in these preparations (which has A-specificity) (30). However, we have found that neither mitochondrial nor cytoplasmic malate dehydrogenase can convert oc-ketoglutarate or oxalacetate to their corresponding amino acids in the presence of DPNII plus NH&l.
The products formed if the A-form of tritiated DPNH is incubated with these keto acids and malatc dehydrogenase are their respective tritiated hydroxyacids.
No tritiated water or tritiatcd DPN is found.
No amino acids are found when this incubate is chromatographed on the amino acid analyzer. While it is still possible that some dehydrogenase impurity with Aspecificity slowly reductively aminates a keto group on the mitochondrial transaminase, it is difficult to conceive of which dehydrogenase this might be. An impurity, however, would be only part of the reaction since the mitochondrial enzyme also has B-specificity with respect to DPNH, and as mentioned above there is no evidence of even a trace of glutamate dehydrogenase activity in these preparations. The fact that DPNH protects the mitochondrial transaminase against inactivation is a strong argument that DPNH is actually bound to this enzyme.
The mitochondrial enzyme seems to have one tightly bound keto group other than pyridoxal phosphate which slowly reacts with DPNH.
This reaction has A-specificity with respect to DPKH.
DPNH plus NIL+ also apparently reacts with pyridoxal phosphate to produce pyridoxamine phosphate. This latter reaction is slower and has B-specificity with respect to DPNH.
The number of groups on the cytoplasmic enzyme cannot yet be even estimated.
A logical interpretation of the reaction with both enzymes is Ohat an amine is produced by reductive amination. This amine can then exchange hydrogen with water perhaps by forming a Schiff base with a keto group or pyridoxal phosphate (28,29).
The results shown in Table VI reveal that hydrogen on the a-carbon of glutamate can completely and rapidly exchange with water even when the level of glutamate (1 mM) is considerably higher than that of the transaminase (0.1 mg per ml or 2.2 PM with respect to bound pyridoxal phosphate or peptide chains). Also, if a iow level of glutamate is present (76 PM), hydrogen on the a-carbon exchanges with water.
Therefore, these exchanges can be rapid and complete when either a high or low fraction of the transaminase is in the pyridoxamine phosphate form. An exchange between hydrogen on the product with water could also occur if the product was a sulfhydryl group or a reduced flavoprotein (31). However, if these were products of the reaction with DPNH and transaminase, it would be more difficult t,o explain why the reaction is not altered by adding dithioerythritol, the ammonia dependency, and the inhibition by phenylhydrazine.
Also, if a reduced flavoprotein was produced in amounts equivalent to the amount of DPN produced, then it would be expected that this flavoprotein could be detected spectrophotometrically.
This is not the case. At this time this complicated reaction cannot be completely described.
If amines are produced, their net production is considerably less than that of DPN.
A significant amount of amines arc apparently not produced which remain bound to the cytoplasmic trnnsaminase during dialysis and deproteinization. This is because a quantitative ninhydrin performed on the incubated dialyzed enzyme reveals only a few additional amines. Also, incubation plus dialysis does not alter the electrophoretic mobility of the transaminase.
It is also unlikely that amines are produced which are loosely bound and leave the enzyme during deproteinization and dialysis. This is because only a few amino acids are found when the incubated enzyme is deproteinized and applied to the amino acid analyzer without intervening dialysis.
Also, the incubated dialyzed cytoplasmic enzyme can again oxidize as many as 20 moles of DPNH per incubated peptide chain. One possibility is that an amine is produced which can be recycled back to the original keto group. This would be consistent with all present results. A logical possibility is that this proposed recycling results from amine oxidase activity.
It is known that pyridoxal can catalyze amine oxidase reactions (32). A reaction of this type facilitated by