Time-dependent inhibition of aspartate aminotransferase isozymes by DL-glyceraldehyde 3-phosphate.

Abstract dl-Glyceraldehyde 3-phosphate in low concentrations has been found to be a time-dependent inhibitor of aspartate aminotransferase (EC 2.6.1.1) isozymes. The cationic isozyme is more susceptible than the anionic isozyme to this inhibition. The l-isomer of glyceraldehyde 3-phosphate was as effective as the d-isomer for each isozyme. Study of various glycolytic intermediates indicated that the conjoint presence of the free aldehyde group and the phosphoryl residue of glyceraldehyde 3-phosphate was necessary for inhibition. The extent of inhibition was dependent on the duration of preliminary incubation, the concentration of glyceraldehyde-3-P, and the pH of the preliminary incubation mixture, but was independent of the concentration of enzyme. Half-maximum inhibition of each isozyme was obtained between pH 6.5 and 6.7. Maximum inhibition of the anionic isozyme was obtained at pH values of 8.4 and above, while the cationic isozyme was optimally inhibited at pH 7.4. The presence of α-ketoglutarate in the preliminary incubation mixture decreased inhibition while aspartate accentuated it. After 30 min of preliminary incubation with glyceraldehyde-3-P, dialysis of the anionic isozyme against buffer, aspartate, α-ketoglutarate, or pyridoxal 5-phosphate resulted in a substantial release of inhibition. In the case of the cationic isozyme, substantial release of inhibition was obtained only by dialysis against α-ketoglutarate. The addition of α-ketoglutarate to either the anionic or cationic isozyme-inhibitor preliminary incubation mixtures released inhibition, but only in the case of the anionic isozyme was the extent of release dependent on the duration of preliminary incubation. Prolonged incubation of this isozyme with glyceraldehyde-3-P resulted in the appearance of a second phase of inhibition. The mechanism of the inhibition of the isozymes by glyceraldehyde-3-P is discussed.

The cationic isozyme is more susceptible than the anionic isozyme to this inhibition.
The L-isomer of glyceraldehyde j-phosphate was as effective as the D-isomer for each isozyme.
Study of various glycolytic intermediates indicated that the conjoint presence of the free aldehyde group and the phosphoryl residue of glyceraldehyde 3-phosphate was necessary for inhibition.
The extent of inhibition was dependent on the duration of preliminary incubation, the concentration of glyceraldehyde-3-P, and the pH of the preliminary incubation mixture, but was independent of the concentration of enzyme.
Halfmaximum inhibition of each isozyme was obtained between pH 6.5 and 6.7. Maximum inhibition of the anionic isozyme was obtained at pH values of 8.4 and above, while the cationic isozyme was optimally inhibited at pH 7.4. The presence of a-ketoglutarate in the preliminary incubation mixture decreased inhibition while aspartate accentuated it. After 30 min of preliminary incubation with glyceraldehyde-3-P, dialysis of the anionic isozyme against buffer, aspartate, cr-ketoglutarate, or pyridoxal S-phosphate resulted in a substantial release of inhibition.
In the case of the cationic isozyme, substantial release of inhibition was obtained only by dialysis against a-ketoglutarate.
The addition of a-ketoglutarate to either the anionic or cationic isozyme-inhibitor preliminary incubation mixtures released inhibition, but only in the case of the anionic isozyme was the extent of release dependent on the duration of preliminary incubation.
Prolonged incubation of this isozyme with glyceraldehyde-3-P resulted in the appearance of a second phase of inhibition.
The mechanism of the inhibition of the isozymes by glyceraldehyde-3-P is discussed.
Asp&ate aminotransferase (n-aspartate :2-oxoglutarate aminotransferase, EC 2.6.1.1) has been characterized and purified * This work was supported in part by Grant CA-08748 from the National Cancer Institute, National Institutes of Health, and Grant T-431L from the American Cancer Society. from various mammalian tissues (l-5).
Among its several functions, liver aspartate aminotransferase has been implicated in the synthesis of glucose from noncarbohydrate precursors (8,9). Although both isozymes of aspartate aminotransferase are involved in the gluconeogenic process (8,9), it is the activity of the anionic isozyme which is regulated by dietary and hormonal treatments, while that of the cationic isozyme remains essentially unchanged (9)(10)(11).
The adaptive regulation of the anionic isozyme occurs over a period of days (g-II), but other more rapidly acting control mechanisms may also operate to complement these processes. Preliminary observations indicated that when fasted rats, which have elevated levels of the liver anionic isozyme, were fed fructose, the activity of this isozyme returned within 6 hours toward the levels present in the fed animal.
This change did not occur when glucose was fed. The refeeding of either sugar did not change the level of activity of the cationic isozyme.
These observations led us to investigate the effects on purified aspartate aminotransferase isozymes of various metabolites which are either unique to fructose metabolism (12) or have been shown to change significantly in rat liver upon fructose feeding (13). Of the various compounds tested only glyceraldehyde-3-P had a substantial effect; it inhibited both isozymes.
The present work describes the time dependence of this inhibition, and the influence of inhibitor concentration, enzyme protein concentration, pH, and substrates. Possible mechanisms for the int.eraction of glyceraldehyde-3-P with the isozymes are discussed.

EXPERIMENTAL PROCEDURE
Purijication of Aspartate Arninotransferase Isozymes-The isozymes of aspartate aminotransferase were purified 150-to 200fold from rat liver essentially as described by Nisselbaum and Bodansky (5). The final preparation of each isozyme was dissolved in 5 mM Tris-acetate buffer, pH 7.4. The specific activity of the anionic isozyme was 150 units per mg and that of the cationic isozyme was 210 units per mg. Cross-contamination of the isozyme preparations was less than 0.1% as determined by starch gel electrophoresis (14). Enzyme Assay-The activity of aspartate aminotransferase was measured by a modification of the coupled reaction described by Karmen (15). One unit is equal to the utilization of 1 pmole of substrate per min at 37". Reaction mixtures were buffered with sodium barbital-HCl for the following reasons. Tris has 2 '.,_ 6 40 Samples of 0.2 ml were brought up to 2.5 ml with the appropriate buffer and incubated at 37" for the indicated times. Enzyme activity was determined after adding 0.5 ml of previously warmed substrate mixture.
The results are expressed as percentage of the activity at zero time for each isozyme in the presence of sodium barbital-HCl, pH 7.4. Anionic isozyme without chloride (0) and with 0.034 M chloride (m). Cationic isozyme without chloride (0) and with 0.034 M chloride (0). been shown to react with glyceraldehyde-3-P (16), phosphate has been shown to inhibit the cationic isozyme (17), and chloride affects the activity of aspartate aminotransferase isozymes (17,13). Therefore, 0.04 M sodium barbitaLHC1 buffer, pH 7.4, was prepared by titrating sodium barbital to pH 7.4 with HCI.
The chloride concentration in this buffer, 0.034 M, was found to yield maximal reaction velocities for both isozymes.
This concentration of chloride also maintained the activity of the diluted isozymes for at least 2 hours at 37" (Fig. 1). Dilutions of the stock isozyme preparations that had been kept in 5 112~ Trisacetate buffer were made in 0.04 M sodium barbitaLHC1 buffer containing 0.15% human serum albumin; the concentrations of Tris-acetate buffer in these dilutions were negligible. Each reaction component was dissolved in 0.04 M sodium barbital-HCI buffer and readjusted to pH 7.4. The reactions were started by the addition of substrate mixture at 37" to the initially incubated isozyme solution.
The final reaction volumes of 1.0 or 3.0 ml contained 16.7 mM aspartate, 6.7 mu cY-ketoglutarate, 0.067 mg per ml of NADH, 0.064 unit per ml of malic dehydrogenase, and 6.7 X 10m3 unit per ml or 7.0 x 10m3 unit per ml of the anionic and cationic isozymes, respectively. Initial velocities were determined by following the rate of decrease in absorbance at 340 rnp in a Beckman model DU spectrophotometer fitted with a Gilford 2000 automatic sample changer and recorder.
Enzyme velocities were expressed as change in absorbance per 3 min.
ikfaterials-The chemicals used were of the highest purity commercially available. All compounds were obtained as, or converted to, the sodium or potassium salts. Dipotassium Dglucose l-phosphate, disodium D-frUCtOSe l-phosphate, disodium DL-glycerol l-phosphate, dihydroxyacetone phosphate dimethylketal, barium nn-glyceraldehyde 3-phosphate diethylacetal,' sodium n-3-phosphoglycerate, sodium 2-phosphoglycerate, trisodium phosphoenolpyruvate, and n-glyceraldehyde were pur-1 Solutions of nn-glyceraldehyde-%P were stable for at least 6 hours in sodium barbital-HCl buffer at pH 7.4 when kept on ice. At 37", however, this compound was destroyed at a rate of approximately 28'% per hour over a 3-hour period. chased from Sigma.
Barium n-fructose B-phosphate and rabbit muscle n-glyceraldehyde 3-phosphate dehydrogenase were purchased from Boehringer Mannheim.
Aspartic acid was obtained from Cycle Chemical Corporation or Calbiochem. n-Glyceraldehyde-3-P was prepared from fructose B-phosphate by periodate oxidation according to the procedure of Szewczuk et al. (19). The product obtained was 65% n-glyceraldehyde-3-P and was further purified.
A solution, 14 ml, that contained approximately 500 pmoles of organic phosphorus of which 316 pmoles were n-glyceraldehyde-3-P, as determined by enzymic analysis (20), was placed on a column, 1.3 x 10.0 cm, of Dowex AG-X8 that had been equilibrated with 1 mM HCl. The column was eluted with a linear gradient between 35 ml of 1 mM HCl and 35 ml of 1 mM HCI-0.1 M NaCl.
This was followed by an additional 35 ml of 1 mM HCl-0.1 M NaCl.
Fractions of 3.7 ml were collected, and those tubes in which the total phosphorus and the alkali labile phosphorus (10 min at room temperature in 1 N NaOH) were the same were pooled. Both the n-glyceraldehyde-3-P and the impurities were eluted close to the upper limit of the gradient.
Because of overlapping of the peaks only 161 pmoles (about 50%) of the n-glyceraldehyde-3-P could be recovered free of impurities.
The compound was precipitated as the calcium salt and dried at room temperature, in a vacuum, over silica gel. The calcium salt was transformed to the hydrogen form with Dowex 50, and the stock solution was kept frozen at pH 2.5. Aliquots were adjusted to pH 7.4 immediately before use. A comparison of the phosphorus analysis with enzymic analysis by means of glyceraldehyde-3-P dehydrogenase showed 83 kmoles of organic phosphorus, 76 pmoles of alkali labile phosphorus, and 78 pmoles of enzymically active compound. Thus, the final preparation was at least 94% pure.

E$ect of Glycolytic Intermediates on Activities of Aspartate Aminotransferase
Isozymes- Table  I shows the effect of preliminary incubation of aspartate aminotransferase isozymes with various glycolytic intermediates for 30 min at 37". Of the compounds tested, only nn-glyceraldehyde-3-P substantially inhibited both isozymes.
The cationic isozyme was more sensitive than the anionic isozyme to this inhibition.
The auxiliary enzyme in the assay of aspartate aminotransferase, malic dehydrogenase, was not affected by nn-glyceraldehyde-3-P.
The inhibition of the cationic isozyme by dihydroxyacetone-P was low but reproducible.
Enzymic analysis of the preparation of dihydroxyacetone-P used in these studies showed that it contained no more than 0.24% n-glyceraldehyde-3-P.
This amount of contamination would not have affected the activity of either isozyme when dihydroxyacetone-P was tested at 2 mM ( Table  I). None of the other compounds had an effect at the concentrations used. With respect to each isozyme, D-and nn-glyceraldehyde-3-P were equally effective inhibitors (Fig. 2). Thus, the n-isomer was as inhibitory as the n-isomer of glyceraldehyde-3-P.
Time Dependence of Inhibition of Aspartate Aminolransjerase Isozymes by DL-Glyceraldehyde-S-P- Fig.  3A shows that the inhibition of the anionic isozyme by nn-glyceraldehyde-3-P was time-dependent and attained maximal values within 30 min and remained constant for up to 1 hour. In the absence of either substrate in the preliminary incubation mixture, the maximal inhibitions were 18 and 47% at concentrations of 0.2 and 2.0 mM nn-glyceraldehyde-3-P, respectively. When 16.7 mM aspartate was included in the preliminary incubation mixture, the maximal values of inhibition increased to 51 and 70% respectively.
In contrast, when 6.7 mM cY-ketoglutarate instead of aspartate was included in the preliminary incubation mixture the maximal inhibitions were decreased to zero at 0.2 InM and 11% at 2.0 InM nn-glyceraldehyde-3-P. Fig. 3B shows that essentially the same patterns were obtained for the cationic isozyme.
In the absence of substrates in the preliminary incubation mixtures, maximal inhibitions of 14 and 67% were attained at 30 min with 0.06 and 0.6 m&r nn-glyceraldehyde-3-P, respectively.
Inthe presence of 16.7 mM aspartate, levels of inhibition were increased to 33 and 86% and were attained after about 15 min. On the other hand, the presence of 6.7 mM cr-ketoglutarate during preliminary incubation led to a marked decrease of the inhibition of the cationic isozyme. Fig. 3 also shows that the time course of inhibition for each isozyme, whether or not substrate was present, was essentially the same at two concentrations of nn-glyceraldehyde-3-P that differed by IO-fold.
Dependence of Extent of Inhibition on Concentration of DL-Glyceraldehyde-3-P-k the absence of either substrate the extent of inhibition of the anionic isozyme after 30-min preliminary incubation increased with increasing concentration of nL-glyceraldehyde-3-P until it reached a maximum of 60% at 3 mM  4B shows that, for the cationic isozyme, the presence of aspartate during preliminary incubation did not affect the maximal extent of inhibition, 90%, but decreased the concentration of nn-glyceraldehyde-3-P required to attain this degree of inhibition
Anionic isozyme was incubated in 0.83 ml of buffer that contained n-glyceraldehyde-3-P (0) or nn-glyceraldehyde-3-P (0 mM. In agreement with the results obtained with the anionic isozyme, 6.7 mu cr-ketoglutarate substantially protected the cationic isozyme against inhibition.
Inhibition of Aspartate Aminotransferase Isoxymes by DL-Glyceraldehyde-S-P as Function of Isozyme Protein Concentration-It has been shown that the isozymes of aspartate aminotransferase may dissociate into subunits in dilute solutions (21,22). The possibility that such subunits might have different sensitivity to inhibition by glyceraldehyde-3-P was explored. Fig. 5A shows that the activity of the anionic isozyme was a linear function of its concentration between 2.6 and 260 ng of protein per ml after preliminary incubation for 30 min both in the absence and in the presence of 0.4 or 4.0 mM nL-glyceraldehyde-3-P. Similar results were obtained when the cationic isozyme at concentrations between 1.16 and 116 ng of protein per ml was initially incubated with 0, 0.2, or 1.0 mM nn-glyceraldehyde-3-P (Fig.  5B). The presence of either 16.7 mM aspartate or 6.7 mM (Yketoglutarate with or without DL-glyceraldehyde-3-P in the preliminary incubation mixture did not alter the linear relationship between reaction velocity and protein concentration for either isozyme.
Effect of pH on Inhibition of Aspartate Aminotransferase Isozymes by DL-Glyceraldehyde-S-P-The isozymes were diluted in 0.08 M NaCl-0.004 M sodium barbital-HC1-0.15~0 human serum albumin, adjusted to the desired pH. Diluted isozyme, 0.2 ml, was mixed with either 0.3 ml of nn-glyceraldehyde-3-P solution or 0.3 ml of 0.004 M sodium barbitaLHC1 in controls, adjusted to the same pH. The chloride concentration was approximately 0.034 M in the preliminary incubation mixture. The pH of each mixture was measured before incubation.
After 30 min at 37", 2.5 ml of complete substrate mixture in 0.04 M sodium barbital-HCI, pH 7.4, was added to each sample and enzyme activity was determined.
This concentration of buffer was sufficient to bring the pH in the reaction to 7.4 f 0.05. Control experiments in which DL-glyceraldehyde-3-P was omitted provided a measure  of the stability of the isozymes over the pH range studied (top curves, Fig. 6). The percentage inhibition of each isozyme was calculated at each pH value from the activity of the corresponding control. Fig. 6A shows that the anionic isozyme was maximally inhibited between pH 8.4 and 10.3 in the presence of either 0.4 or 4.0 mM nL-glyceraldehyde-3-P.
At pH values below 8.4 there was a progressive decrease of inhibition.
At pH 5.4, the inhibition was less than zero, indicating a slight protection of the anionic isozyme against the inactivation observed in the control. Fig. 6B shows that, in the presence of 0.16 or 0.6 mM DL-glyceraldehyde-3-P, the inhibition of the cationic isozyme increased from zero at pH 5.4 to a maximum at pH 7.4 and decreased at higher pH values. The lowered levels of inhibition at pH values other than the optima were not the result of changes in the rates of interaction between the isozymes and the inhibitor, since it was found that maximal inhibition was attained within 30 min at pH 6.4 and 9.2 as well as at pH 7.4. There was no relationship between the effects of pH on inhibition by DLglyceraldehyde-3-P and the effects of pH on the stability of the two isozymes.
Reversal of Inhibition-As has been noted (Figs. 3 and 4) aspartate potentiated and oc-ketoglutarate prevented, to varying extents, the inhibition of both isozymes by nn-glyceraldehyde-3-P. In order to study the effect of adding either substrate after preliminary incubation of the isozymes with on-glyceraldehyde-3-P, the anionic isozyme was initially incubated with 3.6 mM nn-glyceraldehyde-3-P and the cationic isozyme with 0.36 mM DL-glyceraldehyde-3-P for varying periods. Fig. 7A shows that, in the case of the anionic isozyme, addition at 15 min of 01-ketoglutarate to yield a final concentration of 6.7 mM resulted in a time-dependent recovery of enzyme activity during the following 30 min of incubation.
In contrast, addition of aspartate to another portion of the preliminary incubation mixture after 15 min resulted in a further time-dependent loss of activity. When buffer instead of substrate was added to the preliminary incubation mixture, the extent of inhibition remained constant during the following 30 min. Thereafter a progressive loss of activity, which may be regarded as a second phase, was observed in all three types of incubation mixtures, albeit at differing rates. With regard to the cationic isozyme (Fig. 7B), the directions of changes in activity during the 30 min after addition of substrate or buffer were similar to those observed with the anionic isozyme.
However, on further incubation and in contrast to the anionic isozyme, all three types of mixture showed progressive increases in activity.
This release of inhibition of the cationic isozyme was shown to be the result of the instability' of DL-glyceraldehyde-3-P. Cationic isozyme was incubated for 30 min with aliquots of DLglyceraldehyde-3-P that had been kept at 37" in buffer, pH 7.4, for varying periods of time. The longer the inhibitor was kept at 37", the less the inhibition.
In spite of the instability of glyceraldehyde-3-l', a second phase of inhibition of the anionic isozyme was observed during prolonged incubation with this inhibitor.
As has been noted by Webb (23) 6. Effect of pH on the inhibition of aspartate aminotransferase isozymes by nL-glyceraldehyde-3-P.
Each isozyme was incubated with DL-glyceraldehyde-3-P at 37" for 30 min in 0. At each pH value, percentage inhibition was calculated from the controls. 3-P, or to reaction of this compound with other, more slowly reacting groups.
In order to determine whether the release of inhibition, observed after the addition of a-ketoglutarate, and the potentiation of inhibition obtained after aspartate addition were dependent on the time of preliminary incubation of the isozyme with DL-glyceraldehyde-3-P, a mixture of these in buffer, pH 7.4, were initially incubated for 3 to 4 hours. At intervals during this period, aliquots were added to solutions of buffer, cr-ketoglutarate, or aspartate to yield the concentrations described in the preceding sections.
The mixtures were then incubated at 37" for an additional 30 min in the case of the anionic isozyme or 60 min for the cationic isozyme.
The activity of the inhibited anionic isozyme initially incubated without substrate was constant between 30 and 75 min, then decreased (Fig. 8A). In contrast, the activity of the inhibited cationic isozyme showed a moderate increase after 30 min (Fig. 8B). Fig. 8A shows that the extent of reactivation of the anionic isozyme by cr-ketoglutarate decreased markedly with increasing time of preliminary incubation with DL-glyceraldehyde-3-P.
In contrast, the extent to which the cationic isozyme was reactivated by cu-ketoglutarate decreased only slightly with increasing time of preliminary incubation with the inhibitor (Fig. 8B). The degree of potentiation of inhibition in the presence of aspartate was largely independent of the time of preliminary incubation with DL-glyCeraldehyde-3-P for both isozymes. Fig. 9A shows that it was also possible to reverse the inhibition of the anionic isozyme substantially by dialysis for 3 hours at 37" after it had been initially incubated for 30 min with 2.0 The anionic isozyme (A) was incubated with 3.6 InM and the cationic isozyme (B) with 0.36 mM DL-glyceraldehyde-3-P for 15 min in a volume of 10 ml. At the end of this period, 0.66 ml of 0.04 M sodium barbital-HCI, pH 7.4, or substrate was added to 3.3-ml aliquots to give final concentrations of 16.7 mM aspartate (Asp) or 6.7 InM a-ketoglutarate (a-Kg) and 3.0 and 0.3 mM DL-glyceraldehyde-3-P in A and B, respectively. At intervals, 0.6-ml aliquots were assayed for enzyme activity after the addition of 2.4 ml of previously warmed substrate mixture.
The final substrate concentrations during assay were 16.7 mM aspartate and 6.7 mM tu-ketoglutarate.
Enzyme controls without aL-glyceralde-hyde3-P or substrate were incubated simultaneously.
Control enzyme activity (X), activity with m-glyceraldehyde-3-P and buffer (0 ), 16.7 mM aspartate (A), or 6.7 mM a-ketoglutarate (m). All activities were expressed as percentage of the control activity at zero time.  mM nn-glyceraldehyde-3-P. Dialysis against sodium barbital-HCI buffer or buffer that contained 8 mM a-ketoglutarate, or 0.4 mM pyridoxal-5-P restored the activity from 35 to 41y0 to 80 to 85% of that of a control enzyme solution without DLglyceraldehyde-3-P dialyzed under the same conditions. In the presence of 20 mM aspartate the extent of recovery was essentially the same, from 35% of the control activity before dialysis to 75%.
In contrast, cationic isozyme that had been initially incubated for 30 min with 0.6 mM nn-glyceraldehyde-3-P when dialyzed against buffer, or buffer that contained 20 mM aspartate or 0.4 mM pyridoxal-5-P increased in activity from 22 to 29% to only 51 to 55% of the control levels (Fig. 9B). Dialysis against Qketoglutarate gave almost complete recovery of enzyme activity, to 93% of the control value.
These results suggest that, in the absence of a-ketoglutarate, nn-glyceraldehyde-3-P was bound more tightly to the cationic isozyme than to the anionic isozyme. It is unlikely that pyridoxal-5-P was dissociated from either isozyme as a result of inhibition by nn-glyceraldehyde-3-P since dialysis against pyridoxal-5-P was no more effective in restoring activity than was dialysis against buffer alone.
The second phase of inhibition of the anionic isozyme (Figs. 7A and 8A) was apparently not reversible by dialysis. This isozyme was initially incubated at 37" with 2.0 mM nn-glyceraldehyde-3-P for the usual period of 30 min as well as for a longer period of 2 hours, then dialyzed against 0.04 M sodium barbital-HCI, pH 7.4, for 3 hours at 37". Controls without DLglyceraldehyde-3-P were treated in the same manner. As in the earlier experiment (Fig. 9A), enzyme activity rose substantially, from 46% of that in the control after 30 min of preliminary incubation to 87% after dialysis, whereas after 2 hours of pre- liminary incubation the activity was 38% of that in the controls and rose to only 50% after dialysis. Spectral An&&--The pyridoxal form of aspartate aminotransferase has a characteristic absorption spectrum with a peak at 360 mp because of a Schiff base linkage of pyridoxal-5-P and an e-amino group of lysine at the active center (1). Rupture of this linkage by glyceraldehyde-3-P might have been expected to result in a change in the spectrum similar to that observed upon addition of aspartate (24). We were unable to demonstrate any change in the absorption spectra between 310 and 435 rnp at pH 7.4 or 8.4 of either isozyme after the addition of DLglyceraldehyde-3-P at concentrations that resulted in approximately 70% inhibition. DISCUSSION Each isozyme of aspartate aminotransferase purified from rat liver was inhibited by nn-glyceraldehyde-3-P.
As noted previously, there was no stereospecific requirement with regard to the second carbon of glyceraldehyde-3-P.
A survey of the compoundstestedshowed that the conjoint presence of a free aldehyde and a phosphate group was necessary for inhibition.
The pKz for glyceraldehyde-3-P is 6 (Fig. 6). Half-maximal inhibition of each isozyme occurred between pH 6.5 and 6.7, suggesting that the divalent ion of glyceraldehyde-3-P was the inhibitory species. As the pH was increased above 7.4 where glyceraldehyde-3-P exists completely as the divalent anion, maximal inhibition of the anionic isozyme was attained at approximately pH 8.4 and remained constant up to pH 10.3. The inhibition of the cationic isozyme was optimal at pH 7.4 and decreased substantially at higher pH values. Above pH 7.4 there is decreased ionization of the positively charged groups that are responsible for the electrophoretic behavior of the cationic isozyme, as indicated by the finding that this isozyme which is a cation near neutrality has no net charge at pH 8.8 (14). It would appear therefore that the change in net charge at high pH values renders the cationic isozyme less sensitive to inhibition by nn-glyceraldehyde-3-P.
The active center of aspartate aminotransferase contains pyridoxal-5-P linked to an e-aminolysyl residue through an internal Schiff base (1,26). Hughes, Jenkins, and Fischer (26) have isolated from the enzyme, after reduction with sodium borohydride and subsequent digestion, a tetradecapeptide which contains one E-pyridoxyllysine as well as a 2nd lysyl residue. Transamination is accomplished by transfer of the amino group from an a,mino acid to a keto acid through a series of Schiff base reactions (27,28). Ivanov and Karpeisky (29) have reviewed the geometry of this reaction, and have postulated that the interaction of the enzyme with the substrates involves rotations of the pyridoxal-5-P molecule during transamination. Since we found no change in the characteristic absorption spectrum of either isozyme upon incubation with glyceraldehyde-3-P, this compound did not interact with the e-amino group of the lysine bound in the internal aldimine linkage of the pyridoxal enzyme and the possibility existed that it combined with the e-amino group of the 2nd lysyl residue.
It may be postulated, therefore, that such an interaction and the binding of the divalent phosphor-y1 end of the molecule to a positively charged locus in the active site of the enzyme imposed restrictions on the orientation of the pyridoxal-5-P and thus inhibited transamination. Aspartate may potentiate the inhibition by opening the internal aldimine linkage between pyridoxal-5-P and the e-amino group of the 1st lysyl residue, thus freeing an additional site for interaction with the aldehyde group of glyceraldehyde-3-P. cr-Ketoglutarate may protect or reactivate the isozymes by competing for the same site to which the phosphoryl group of glyceraldehyde-3-P is bound.
Turano, Giartosio, and Fasella (31) have noted that the extent of interaction of sulfhydryl groups in aspartate aminotransferase with p-chloromercuribenzoate parallels the degree of inhibition.
The possibility exists that glyceraldehyde-3-P may react with sulfhydryl groups of the enzyme and that this reaction may also play a role in inhibition.
It is unlikely, however, that such a reaction would have a major role in the inhibition since it was reversed substantially by ru-ketoglutarate during the early stages of preliminary incubation.
We propose to examine further the mechanism of inhibition by kinetic studies of the interaction between the isozymes of aspartate aminotransferase and glyceraldehyde-3-P in the presence and absence of substrates as well as by attempting to isolate a glyceraldehyde-3-P derivative of the inhibited isozymes in a manner similar to that employed by Hughes et al. (26) for the isolation from holoaspartate aminotransferase of a peptide containing e-pyridoxyllysine. Our results raise the possibility that glyceraldehyde-3-P may be implicated in the regulation of gluconeogenesis in vivo by affecting the activity of the isozymes of aspartate aminotransferase. Such regulation would depend not only upon the intracellular concentration of glyceraldehyde-3-P and substrates but also upon their distribution between the mitochondria and the cytosol.