Dissociation of Aspartate Aminotransferase into Subunits EFFECT OF LIGANDS UPON THIS DISSOCIATION*

Frontal and zonal analysis of the chromatography of aspartate aminotransferase (EC 2.6.1.1), pig heart cytosolic enzyme, on Bio-Gel P150 shows that holo-and apoenzyme can dissociate at pH 8.3. Ultracentrifugation and fluorescence depolarization confirm this result. Kineticanalysis of the fluorescence depolarization experiments favors a biphasic phenomenon: a few minutes for the faster one and several hours for the slower one. The apparent dissociation constant is 0.8 pM for the apoenzyme and 0.18 fiM for the pyridoxal 5’-phosphate form of the holoenzyme. In the presence of sucrose or 0.1 M L-aspartate or a mixture of 70 mM L-glutamate and 2 mM Lu-ketoglutarate, the holoenzyme is dimeric at concentrations higher than 5 nM. The addition of a mixture of the substrates L-glutamate and oc-ketoglutarate

The same experimental conditions were used in frontal analysis, but the volume of the samples was only 0.5 ml instead of 10 ml. In this case, the observations are only qualitative (13 to their intersection with the y axis, and the value of l/P, in the formula is replaced by the value of l/P', corresponding to this point intersection for each isotherm. A more accurate calculation has been made by Gotlib and Wahl (18), but the Weber (17)  smaller with dilution, and the decrease was time-dependent. Since the Perrin plot of concentrated and diluted holoenzyme do not extrapolate to the same apparent UP, value for TIT equal to zero (that means that the dye is a little less rigidly bound to the diluted enzyme), there is no linear relation between l/P and rotational relaxation time so that the kinetics was difficult to study quantitatively.
A study of the variations of the fluorescence polarization of a 0.48 FM solution with time shows that the polarization decrease is slow (Fig. 4). Fluorescence depolarization measurements also allow the reversibility of the phenomenon to show. The addition of concentrated dye-free holoenzyme to diluted labeled holoenzyme causes a rapid increase of polarization, showing that the phenomenon observed by diluting the solution is reversible.
The only possible phenomenon that could be correlated to a variation of the concentration for diluted solutions of protein (less than 1 mg/ml) is a change in the aggregation state of the molecules, and since it was proved in our experiments that the dye is not released in the buffer, the observed decrease in fluorescence depolarization and rotational relaxation time can be correlated to a dimer-monomer equilibrium.
Hence two independent methods-molecular sieving and fluorescence depolarization-give the same result: the dimeric pyridoxal form of aspartate aminotransferase dissociates into monomers at pH 8.3 when the concentration is lowered from 1 PM to 10 nM. The dissociation is either a slow phenomenon, or it is rapid and the monomer is slowly transformed. sieving experiments, apoenzyme, at pH 8.3 in Tris buffer can dissociate much more easily than holoenzyme (Fig. 5). In a concentration which is 0.4 pM in dimer, one-half of the molecules are dissociated and a dissociation constant equal to 0.8 FM can be deduced, but the calculated curves do not fit well with the experimental values. At 0.2 PM, apoenzyme is almost entirely dissociated, and addition of one pyridoxal phosphate per monomer gives the elution volume of the dimer. The elution volume of the apoenzyme subunit is consistent with a molecular weight of 46,000.
Since apoenzyme can dissociate in the concentration range where the protein absorbance is 0.02 at 280 nm, ultracentrifugation can be used with a scanning method and double sector cells. Some experiments were made to improve the results obtained by the gel filtration method. With dilution, the diffusion coefficient increases as the sedimentation coefficient falls off as well as the s/D ratio (Fig. 6). These results indicate a drop in the average molecular weight, i.e. a dissociation of the apoenzyme is induced by dilution in the same concentration range where the elution volume increases. Seven hours after dilution, a sample of apoenzyme at 0.2 pM gave an s equal to 2.9 x lo-l3 s and a D equal to 5 + 0.5 cm2/s, but the activability was only TO%, instead of 95% obtained 1 hour after dilution. Effect of Substrates-When the aldiminic form of aspartate aminotransferase can react with the substrate L-aspartate, it is turned into the aminic form of the enzyme (330 nm absorption), and intermediate covalent complexes (360, 430, and 330 nm absorption) exist together (19). Using Bio-Gel chromatog- Under these conditions l/R, remains constant, 1.22, corresponding to a dimer in the concentration range from 5 nM to 2 FM. Another experiment was performed to show that substrates really induce a reassociation of the enzyme. Monomeric enzyme (without substrates) was obtained by diluting a holoenzyme solution to 4.6 nM protein concentration, then 70 mM L-glutamate and 2 mM oc-ketoglutarate were added, and the mixture was eluted on Bio-Gel P-150.
A l/R, equal to 1.21 was detected, meaning that substrates induce the reassociation of monomer into dimer.
Effect of Sucrose-When the molecular weight of aspartate aminotransferase was measured in the presence of sucrose, the absence of dissociation was always observed (5,16). In the presence of lo%, by weight, sucrose, we observed a constant R, between 5 nM and 1 WM on Bio-Gel; its value corresponds to the dimeric protein. This result confirms the lack of dissociation observed under these particular conditions (5,16). We did not try to find out why sucrose inhibits the subunit dissociation.

DISCUSSION
Preliminary experiments on diluted holoaspartate aminotransferase, studied by zonal analysis of Bio-Gel P-150 chromatography, have shown that this enzyme can be dissociated by dilution. The phenomenon is complex and includes a dissociation and a transconformation; one of these steps is slow. More quantitative results were obtained by frontal analysis. By this method, the dissociated monomeric holoenzyme appears to be eluted as a 60,000-dalton molecule and the dimer as a 90,000 one. An apparent dissociation constant of 0.18 pM is detected. Nevertheless, the results do not fit well the Ackers-Winzor (13,14) equation. The best fitting is obtained by assuming a l/R, equal to 1.18 for the dimer and 1.34 for the.monomer, but the fitting is not good. A possible explanation is that the equation applies only to quickly equilibrating systems. The results obtained with the frontal analysis are confirmed by the experiments of fluorescence depolarization. This last method gives two other informations: the global phenomenon is slow and its reversibility is confirmed. Frontal analysis also shows that the apoenzyme can undergo dissociation, and the results are confirmed by an independant method-ultracentrifugation.
Our results are consistent with all published observations and may explain some apparent contradictions between them. Actually, others (5, 16) who could not detect enzyme dissociation at 3 nM used sucrose in their experiments, and we observed that a reassociation of the monomer is induced by sucrose. There is some discrepancy between our results of zonal chromatography and that of Feliss and Martinez-Carrion (5). These authors affirm they did not observe any dissociation by Sephadex G-200 chromatography, but they did not give their experimental conditions and they did not explain why their results are different from that of Banks et al. or Melander (3), so it is difficult to explain this discrepancy.
Our results on the dissociation of the apoenzyme are consistent with that of Banks et al. (4). These authors failed to obtain a stable apoenzyme and could not reactivate the apoenzyme after passage through Sephadex G-100 columns. They found indirect evidence that the apoenzyme dissociates into monomer by studying the kinetics of its recombination with coenzyme. They showed that the apoenzyme undergoes a concentration-dependent change from a more to a less reactive species as the concentration is raised. In our experiments apoenzyme prepared without phosphate does not inactivate on Bio-Gel P-150, so that the chromatographic method can be used and gives a direct measure of the apoenzyme dissociation into monomer. The concentration range is the same for the dissociation into subunits observed by us and for the change in reactivity toward the coenzyme. It means that the hypothesis of Banks et al. is justified: the monomeric apoenzyme is more reactive uersus pyridoxal or pyridoxamine phosphate than the dimeric form.
The most important result of the present work is that the stability of the dimeric form is in the order, holoenzyme + substrates > holoenzyme > apoenzyme, with about 1 order of magnitude between the dissociation constants. With the succinylated enzyme, Polyanovsky and Pikhelgas (20) have shown that the dissociation is easier for the apoenzyme than for the holoenzyme. The order of stability of the quaternary structure and that of the tertiary structure does not seem to be the same, since the complexes with inhibitors are less stable than the pyridoxal enzyme form and more stable than the pyridoxamine form (21).
An important problem is the activity of the dissociated enzyme. Is the monomer really active since substrates induce the dimerization?
It seems that in very diluted solutions the activity of the enzyme increases (2,3,7), and this increase has been attributed to a lowering of K, values for aminoacids in enzyme monomeric form. If these differences of K, values are not due to the methods used to determine the activity at high enzyme concentration, it means that the substrate affinity increases when the enzyme is partially dissociated and that the monomer is active. This result is hardly compatible with the substrate-induced association of the monomer into dimer. The only interpretation compatible with all the observations should be an anticooperativity between the two sites of dimer.
Rene Cohen for the program used to simulate the ultracentrifugation experiments. We would like to thank Professor J. Yon and G. Nemethy for helpful comments and careful reading of the manuscript. Excellent technical assistance of Mme A. Larousse (ultracentrifugation) and Mme J. Carrette (preparation of the enzyme) is very gratefully acknowledged.