Aspartate aminotransferase. Determination of the active site occupancy pattern indicates independent transamination of the two subunits.

Aspartate aminotransferase (LY subform from pig heart cytosoI~, a dimeric enzyme composed of two identical subunits, was tested for subunit interactions imposed on the transamination reaction. The two functional variants of the enzyme arising from its double displacement mechanism, i.e the pyridoxal-5’-P form (LL dimer) and the pyridoxamined’-P form (MM dimer), as well as the enzyme whose aldimine bonds to pyridoxal-5’-P had been reduced with sodium borohydride (RR dimer), were found to be separable by isoelectric focusing. Separation of a half-reduced pyridoxal-5’-P enzyme preparation yielded a 1:2:1 distribution of LL homomer, LR hybrid, and RR homomer (isoelectric points 5.68, 5.87, and 6.03, respectively). The enzymatic activity of the isolated LR hybrid is exactly half of that of the native LL homomer. Labeling of the L subunits by reduction with sodium borohydride was employed to determine the distribution of L and M subunits among aspartate aminotransferase dimers at transamination equilibrium. The substrate pair 2-ketoglutarate and glutamate in a concentration ratio appropriate to result in the formation of equimolar concentrations of L and M subunits was added to a solution of the enzyme. After removal of the substrates to substoichiometric concentrations, the LLILMIMM dimer mixture was reduced with sodium borohydride to a RRIRMIMM mixture, which was converted to a readily separable RR/RL/LL mixture by the addition of excess 2-ketoglutarate. Isoelectric focusing yielded a 1:2:1 distribution of LL homomer (originally MM), LR hybrid (originally ML), and RR homomer (originally LL). The binomial active site occupancy pattern demonstrates random interconversion of the individual L and M subunits of the enzyme dimers during transamination, i.e. independent catalytic function of the two active sites of the aspartate aminotransferase dimer.

Aspartate Aminotransferase DETERMINATION OF THE ACTIVE SITE OCCUPANCY PATTERN INDICATES INDEPENDENT  TRANSAMINATION  OF THE TWO SUBUNITS* (Received for publication, January 21, 1977, andin revised form, May 19, 1977) HAN~J~RG SCHLEGEL,$ PETER E. ZAORALEK, § AND PHILIPP CHRISTEN From the Biochemisches Institut d-er Universitit Ziirich, CH-8028 Ziirich, Switzerland Aspartate aminotransferase (LY subform from pig heart cytosoI~, a dimeric enzyme composed of two identical subunits, was tested for subunit interactions imposed on the transamination reaction. The two functional variants of the enzyme arising from its double displacement mechanism, i.e the pyridoxal-5'-P form (LL dimer) and the pyridoxamined'-P form (MM dimer), as well as the enzyme whose aldimine bonds to pyridoxal-5'-P had been reduced with sodium borohydride (RR dimer), were found to be separable by isoelectric focusing. Separation of a half-reduced pyridoxal-5'-P enzyme preparation yielded a 1:2:1 distribution of LL homomer, LR hybrid, and RR homomer (isoelectric points 5.68, 5.87, and 6.03, respectively). The enzymatic activity of the isolated LR hybrid is exactly half of that of the native LL homomer.
Labeling of the L subunits by reduction with sodium borohydride was employed to determine the distribution of L and M subunits among aspartate aminotransferase dimers at transamination equilibrium.
The substrate pair 2-ketoglutarate and glutamate in a concentration ratio appropriate to result in the formation of equimolar concentrations of L and M subunits was added to a solution of the enzyme. After removal of the substrates to substoichiometric concentrations, the LLILMIMM dimer mixture was reduced with sodium borohydride to a RRIRMIMM mixture, which was converted to a readily separable RR/RL/LL mixture by the addition of excess 2-ketoglutarate. Isoelectric focusing yielded a 1:2:1 distribution of LL homomer (originally MM), LR hybrid (originally ML), and RR homomer (originally LL). The binomial active site occupancy pattern demonstrates random interconversion of the individual L and M subunits of the enzyme dimers during transamination, i.e. independent catalytic function of the two active sites of the aspartate aminotransferase dimer.
Aspartate aminotransferase is a dimeric enzyme composed of two identical subunits. In the course of the transamination reaction, the subunits shuttle between the pyridoxal-5'-P form and the pyridoxamine-5'-P form (1, 2 and functional properties of the singIe active site of both the apo/holo hybrid (6-S) and a hybrid containing one subunit whose internal aldimine had been reduced with sodium borohydride (9, 10) were indistinguishable from those of the active sites in the native holo homomer. These experiments with semiactive enzyme hybrids, however, do not conclusively rule out interactions between the two catalytically active subunits of the native enzyme. Therefore, the present study was designed to test whether the individual subunits of the aspartate aminotransferase dimer undergo transamination independently from each other. The experimental objective of this investigation was to measure the relative concentrations of LL homomer, LM hybrid, and MM homomel3 in a population of aspartate aminotransferase dimers after attainment of the transamination equilibrium. In an enzyme solution that, after removal of the substrates, contains equal concentrations of L and M subunits, a 1:2:1 distribution of the three dimer species would be expected if the two subunits of the enzyme underwent transamination independently from each other. Any deviation from the binomial distribution would indicate the existence of functionally important subunit interactions.
Three variants of the enzyme were employed in the present study, the LL, the MM, and the enzymically inactive RR3 dimer prepared from the LL homomer by reduction with sodium borohydride.
On isoelectric focusing the three homomers, analyzed separately or combined, reproducibly accumulate in distinct bands at defined positions in the pH gradient, the LL homomer at pH 5.68 and the RR and MM homomers in close proximity at pH 6.03 and 6.07, respectively (S.D., 20.05 pH unit). These pH values correspond approximately to the isoionic pH values of the enzyme The LR hybrid could readily be isolated from a LL homomer preparation that had been half-reduced with sodium borohydride. Isoelectric focusing separated the half-reduced dimer mixture into three components with a distribution of protein of 1:2:1 (Fig.  1). The position in the pH gradient, the absorption spectrum, and the specific activity showed Peak I to be the LL homomer (350 units/mg) and Peak III to be the RR homomer (26 uuits/mgl. The center peak II was identified as the LR hybrid; its pyridoxal-5'-P content was 1.0 mol/mol of dimer, its absorption spectrum (in 50 mM sodium phosphate (pH 7.5)) corresponded exactly to the arithmetical mean of those of the LL and RR homomers, and its specific activity was 180 units/ mg (i.e. half of that of the LL homomer). On addition of glutamate (10 mM) or cysteine sullinate (2 rnM)? the LR hybrid was converted to the MR hybrid, whose absorption spectrum again was indistinguishable from that of an equimolar mixture of MM and RR homomer. Refocusing of the isolated LR hybrid after storage for 6 days at 4" in 50 mM sodium phosphate (pH 7.5), yielded one symmetrical peak at the same position in the pH gradient with unchanged specific activity and spectral properties.
Similarly, isoelectric focusing of an equimolar mixture of LL and RR homomers (28 pM dimer each in 50 mM sodium phosphate (pH 7.5)) that had been kept for 10 days at room temperature yielded unaltered LL and RR homomers without hybrid having been formed. Analogous experiments showed that the MR hybrid and the MM and RR homomers also do not interchange their subunits. Attempts to isolate the LM hybrid by a procedure analogous to that used for the LR hybrid were unsuccessful. The substrate pair 2-ketoglutarate plus glutamate was added to an enzyme solution in saturating concentrations at such a ratio that the enzyme, after removal of the substrates by gel filtration, contained equimolar concentrations of L and M subunits. Isoelectric focusing of this transamination equilibrium mixture yielded LL and MM homomers, each representing approximately half of the total protein. A small intermediary peak 3 Capitals are used as symbols for the subunits in the pyridoxal-5'-P (L), pyridoxamine-5'-P (M), and reduced (R) form (internal aldimine of pyridoxal-5'-P and lysyl residue 258 of the enzyme reduced with sodium borohydride (13, 15)). proved to be subject to a time-dependent decrease during the focusing run (see Fig. 2  focusing does not sufficiently separate the RR/RM/MM mixture, it was converted to the readily separable RR/RL/LL mixture by the addition of 2 mM 2-ketoglutarate. On isoelectric focusing a 1:2:1 distribution was obtained, similar to the one observed on focusing the half-reduced LL preparation (see Fig. 1). The specific activities (350, 190, and 35 unitstmg) and the absorption spectra of the enzyme in the three different peaks confirmed their identification as the LL homomer, the LR hybrid, and the RR homomer. Prolonged storage of the dimer mixture either before or after removal of the substrate pair up to 15 days prior to reduction and focusing did not alter the 1:2:1 distribution.
The same distribution was obtained when the transamination equilibrium was established by adding the substrates at much lower, substoichiometric concentrations (5 nM glutamate and 5 nM 2-keteglutarate) to an equimolar mixture of LL and MM homomers. In this case gel filtration prior to the reduction was omitted. Cuntrol experiments described in the miniprint supplement showed that hybrid formation strictly depends on the occurrence of transamination and is not due to interchange of enzyme-bound pyridoxal-Y-P and pyridoxamine+S-P or to interchange of holo subunits.

DISCUSSION
The present approach to the detection of functionally important subunit interactions in oligomeric enzymes is based on the determination of the active site occupancy pattern in a catalytically engaged enzyme population. In principle, the approach is feasible if two requirements are met: firstly, there must be means to label the catalytically engaged subunits, e.g. by stabilization of covalent enzyme substrate intermediates or by the use of substrate analogs that react irreversibly with the enzyme. Secondly, the labeled enzyme must be physically separable from the unoccupied enzyme. Aspartate aminotransferase, in view of its double displacement mechanism, would prima facie appear to be directly amenable to the determination of the active site occupancy pattern; its two functional states, the L and the M forms, should be stabilizable by the mere removal of the transaminating substrates. However, the slow disappearance of the LM hybrid during isoelectric focusing necessitated stabilization of the transamination equilibrium mixture prior to isoelectric focusing by reduction with sodium borohydride. in saturating concentrations and subsequent gel filtration, and (b) addition of the substrates in low substoichiometric concentrations to an equi-molar mixture of LL and MM homomers. In both procedures, the reduction of the transamination equilibrium mixture occurs in the presence of low, subsaturating concentrations of the substrates. According to the above-mentioned considerations, the observed binomial active site occupancy pattern excludes the existence of subunit interactions on the binding level, whereas the existence of weak interactions involving the covalent phase of catalysis is not completely ruled out.
The structural and functional properties of the isolated LR hybrid are fully compatible with functional independence of the two subunits of aspartate aminotransferase with respect to both the binding and the covalent transformation of the substrates. Cytosolic aspartate aminotransferase from pig heart has been reported to retain its dimeric structure also in very dilute solution (1 x lo-'" M) (16). Thus, the LR hybrid may be assumed to have remained a dimer at the concentrations used in this study to measure enzymic activity or to record the absorption spectra. In both the LR and MR hybrids, the absorption spectra of the coenzyme chromophores serving as intrinsic optical probes of the active site region correspond with the arithmetic mean of those of the two respective homomers. Further, the unchanged catalytic activity of the nonreduced subunit, as well as the random occurrence of the reduction, testify to the absence of functionally significant subunit interactions. Similar observations have been made in previous experiments with an LR hybrid obtained by hybridization of LL and chemically modified RR homomers under denaturing conditions (9, 10).
In conclusion, two different experimental approaches have demonstrated the mutual functional independence of the two active sites of cytosolic aspartate aminotransferase from pig heart: (a) studies with isolated hybrid dimers containing one active and one inactive subunit (apo (see the introduction) or with reduced internal aldiminef; and (b) determination of the active site occupancy pattern in the fully active enzyme dimer at transamination equilibrium.