Glucose Phosphorylation INTERACTION OF A 50-AMINO ACID PEPTIDE OF YEAST HEXOKINASE WITH TRINITROPHENYL ATP *

A 50-amino acid peptide predicted three-dimensional structure

A 50-amino acid peptide predicted by chemical modification studies of yeast hexokinase to contain an ATP-binding site has been synthesized and purified. The peptide, which includes residues from glutamate '78 at the NH&erminal end to leucine 127 at the COOH-terminal, resides within the smaller of the two lobes found in the three-dimensional structure of yeast hexokinase.
It is this region which has been reported recently to exhibit significant sequence homology with hexokinase types I and IV of higher eukaryotic cells and sequence homology with the active site of protein kinases.
A 5-fold enhancement is observed when S pM peptide interacts with 20 pM TNP-ATP. The stoichiometry of binding is very close to 1 mol of TNP-ATP/mol peptide. Also, similar to native yeast hexokinase, the fluorescent enhancement observed upon TNP-ATP binding to the synthetic peptide is greater than that observed upon TNP-ADP binding. Finally, TNP-AMP exhibits a much lower fluorescent enhancement in the presence of hexokinase or the synthetic peptide.
The additional findings that ATP can readily prevent TNP-ATP binding and that TNP-ATP can substitute for ATP as a weak substrate for hexokinase in the phosphorylation of glucose indicate that the synthetic peptide described here comprises part of the catalytic site.

Hexokinase
(ATP: D-hexose 6-phosphotransferase, EC 2.7.1.1) commits glucose to catabolism by catalyzing the phosphorylation of glucose using ATP in the presence of Mg*+. Only the three-dimensional structure of the yeast enzyme has been elucidated (l-6). As shown in Fig. lA, the enzyme consists of a single polypeptide chain of 50 kDa which folds into a three-dimensional structure in which there is a deep central cleft that divides two lobes (l-4). Upon binding glucose, the smaller lobe rotates relative to the larger lobe partially closing the cleft around the glucose molecule (4). The region of the hexokinase molecule involved in ATP binding remains controversial.
Although the x-ray studies indicate that the ATP-binding site lies within the larger of the two lobes, it is important to note that 8-BrAMP rather than ATP was used (6). Recent work from other laboratories now indicates that the major region of ATP binding to yeast hexokinase may lie within the smaller lobe. Thus, Tamura et ul. (7) demonstrated that the ATP affinity label PLP-AMP' (pyridoxyl5 '-diphospho-5'-adenosine) binds covalently to lysine 111 within the smaller lobe. Also, amino acid sequence analysis of several eukaryotic hexokinases, including types I and IV, and yeast types A and B have revealed considerable sequence homology within the smaller lobe (8-12). Part of this region matches closely the ATP binding region found in protein kinases (13)(14)(15)(16).
In order to test more directly the role of the smaller lobe of yeast hexokinase in ATP binding, we have synthesized a 50amino acid peptide which includes residues from glutamate 78 at the NH*-terminal to leucine 127 at the COOH-terminal (Fig. 1B). This peptide includes both the region of homology found in various hexokinase forms and lysine 111 which binds the ATP affinity label (7). Data presented below demonstrate that this synthetic peptide retains significant structure and binds the fluorescent ATP analog TNP-ATP.

AND DISCUSSION
Purity and Characterization of the 50-Amino Acid Residue Hexokinase Peptide (HPP-50)-Procedures employed in the synthesis and purification of HPP-50 are described in detail under "Experimental Procedures." As shown in Fig. 2 and Table I, four different methods were used to examine the purity of HPP-50 prior to investigating its capacity to interact with ATP. Fig. 2A shows that upon HPLC chromatography on a PBondapak Cl8 column, HPP-50 elutes as a single peak, nearly gaussian in shape. The inset in Fig. 2A shows that HPP-50 migrates as a single band upon SDS-PAGE in a 10% polyacrylamide gel system. As predicted from its molecular weight of 6,210, HPP-50 migrates faster than cytochrome c (Mr = 11,700). HPP-50 appears by this criteria apparently homogeneous with neither larger nor smaller contaminants being apparent. In data not presented here, HPP-50 also showed a single band upon native PAGE in a 20% polyacrylamide gel. Results presented in Fig. 2I3   acid composition of HPP-50 compares within experimental error with the predicted amino acid composition of the peptide. Finally, the amino acid sequence (38 residues) of HPP-50 from the NH*-terminal glutamate 78 to proline 115 has been contirmed by direct amino acid sequence analysis as detailed under "Experimental Procedures." The results of the first 15 cycles of this sequence analysis, together with recoveries, are presented in Table I. Fig. 2C summarizes the circular dichroism spectrum of HPP-50. This measurement was performed because circular dichroism spectroscopy at wavelengths near the absorption band for amide bonds represents one of the most sensitive physical methods for detecting secondary structure in polypeptide chains. HPP-50 exhibits significant structure as revealed by the circular dichroism spectrum. The extent of various secondary structures were calculated using the PROSEC program as follows: @ sheet = 57.9%, a-helix = 5.9%, turns = 6.5%, and random = 29.7%. Significantly, a @strand stretch is located within the x-ray structure of hexokinase in the region represented by HPP-50 (1).
Interaction of HPP-50 with TNP-ATP-In order to determine whether HPP-50 exhibits the capacity to interact with adenine nucleotides, we used the fluorescent probe TNP-ATP. This nucleotide analog represents a sensitive probe for ATP-binding sites and has been used successfilly in probing the active sites of both the mitochondrial (23) and Na+,K+ ATPase (24). Results presented in Fig, 3A show that HPP-50 induces TNP-ATP to exhibit a strong fluorescent enhancement throughout the concentration range of TNP-ATP tested (520 PM). HPP-50 induces a similar response upon interacting with TNP-ADP but a much lower response in the presence of TNP-AMP.
Overall, the relative degree of interaction is in the order TNP-ATP = TNP-ADP >> TNP-AMP. Stoichiometric ratios (mean values) for TNP-ATP and TNP-ADP were found to be, respectively, 0.80 and 0.85 mol/mol HPP-50 with apparent dissociation constants (& values) of 4.4 and 5.2 FM.
Data presented in Fig. 3A show also that neither M$+ nor glucose are required for TNP-ATP binding to HPP-50. The latter observation is consistent with the x-ray structure that predicts that residues involved in glucose binding lie outside  Fig. 3. In all panels the concentration of TNP-nucleotides is indicated. In all cases the concentration of yeast hexokinase based on a molecular weight of 50,000 was 10 pM. Where indicated in A 1.0 mM glucose and 0.5 mM MgClg were present.
In C, 0.5 mM ATP was added prior to TNP-ATP.
the region of HPP-50 (1). Therefore, the binding of TNP-ATP to HPP-50 cannot experience a dependence on glucose binding as is the case in the intact enzyme (Fig. 4A). Finally, it is important to note in Fig. 3B that the addition of ATP prior to TNP-ATP to HPP-50 almost completely prevents a fluorescent response. In data not presented here, ATP addition after TNP-ATP addition also reversed the fluorescence response. Thus, the site on HPP-50 which interacts with TNP-ATP also appears to interact with ATP. Interaction of TNP-ATP with Yeast Hexokinase-As experiments on HPP-50 strongly implied that this peptide comprises at least part of an ATP binding domain within yeast hexokinase, it became important to determine whether the native enzyme also interacts with TNP-ATP, and if so, the characteristics of this interaction. Experiments in Fig. 4.4 show that, just as in the case of HPP-50, native yeast hexokinase induces TNP-ATP to exhibit a strong fluorescent response with or without added Mp.
In this case, and as expected, glucose markedly alters the fluorescent enhancement, as glucose binding is known to affect ATP binding in the native enzyme (6).
Similar also to studies obtained with the synthetic peptide (HPP-50), it can be seen in Fig. 4B that the relative degree of interaction of native yeast hexokinase with TNP-nucleotides is in the order TNP-ATP>TNP-ADP>>>TNP-AMP. Moreover, the addition of ATP to yeast hexokinase prior to TNP-ATP markedly reduces, in an apparently competitive manner, the fluorescent enhancement of the latter, a result also obtained with HPP-50 (Fig. 3B). In data not presented here, the addition of ATP after the addition of TNP-ATP was shown also to markedly reverse the fluorescent enhancement. Finally, stoichiometry and apparent KD values (mean ATP-binding Peptide of Yeast Hexokinme values) for bindin g of TNP-ATP and TNP-ADP to the native nical assistance and to Dr. Xavier Ysern for his help in obtaining the enzyme were in the same range as those for the synthetic computer graphics representation of hexokinase presented in Fig. L4 TNP-ATP binding to hexokinase appears much tighter than that of ATP, a finding consistent with the binding of TNP-ATP to the catalytic sites of the mitochondrial Fi-ATPase (23) and the Na+/K+ ATPase (24).) Overall, the data summarized in this report indicate that binding of TNP-ATP to both HPP-50 and to hexokinase is not due to nonspecific hydrophobic interactions, but reflects specific binding. Thus, binding is prevented or reversed by addition of the substrate ATP. Moreover, TNP-AMP binds poorly to both HPP-50 and to hexokinase.
In experiments not presented here, it should be noted also that TNP-ATP did not bind to several other proteins tested including lactoglobulin, chymotrypsinogen, and ovalalbumin. Also, TNP-ATP failed to bind to a synthetic, 50 amino acid, yeast hexokinase peptide (tyrosine 322 through isoleucine 371), which showed very little secondary structure.
Significantly, two different methods not presented here revealed that TNP-ATP can substitute for ATP in the hexokinase reaction as a weak substrate. Using the thin layer method normally used to assess the purity of TNP-nucleotides (see "Experimental Procedures"), the presence of hexokinase was observed visually to induce the formation of TNP-ADP. In a second assay (see "Experimental Procedures") in which the formation of [%]glucose 6-phosphate from [i%]glucose and TNP-ATP was monitored, a very low but significant rate (420 nmol of glu-6-P/h/mg protein) of product formation was observed. (This low rate may reflect tight binding of the product TNP-ADP to hexokinase, its slow release rate impairing further catalysis.) In summary, the studies presented here add support to the conclusion of Tamura et ul. (7) that the smaller lobe of hexokinase is involved in ATP binding. They also support the recent suggestion of Andreone et ul. (10) based on homology arguments that the region of yeast hexokinase (and other hexokinases), comprised here by the synthetic peptide HPP-50, includes an ATP-binding site similar to that found also in protein kinases. However, unlike putative ATP binding domains reported for adenylate kinase (27), the mitochondrial ATP synthase (28), and many other nucleotide-binding proteins (29), yeast hexokinase does not contain the consensus sequence GXdGKTXeI(V), either in the region comprised by HPP-50 or in other regions of the enzyme.