Selective oxidative modification and affinity cleavage of pigeon liver malic enzyme by the Cu(2+)-ascorbate system.

Pigeon liver malic enzyme was rapidly inactivated by micromolar concentration of Fe2+ in the presence of ascorbate at neutral pH. The inactivated enzyme was subsequently cleaved by the Fe2+-ascorbate system at the chemical bond between Asp258 and Ile259 (Wei, C. H., Chou, W. Y., Huang, S. M., Lin, C. C., and Chang, G. G.(1994) Biochemistry, 33, 7931-7936), which was confirmed by site-specific mutagenesis (Wei, C. H., Chou, W. Y., and Chang, G. G.(1995) Biochemistry 34, 7949-7954). In the present study, at neutral pH, Cu2+ was found to be more reactive in the oxidative modification of malic enzyme and the enzyme was cleaved in a similar manner as Fe2+ did. At acidic pH, however, Fe2+ was found to be ineffective in oxidative modification of the enzyme. Nevertheless, Cu2+ still caused enzyme inactivation and cleaved the enzyme at Asp141-Gly142, Asp194-Pro195, or Asp464-Asp465. Mn2+ and L-malate synergistically protect the enzyme from Cu2+ inactivation at acidic pH. Cu2+ is also a competitive inhibitor versus Mn2+ in the malic enzyme-catalyzed reaction with Ki value 70.3 ± 5.8 μM. The above results indicated that, in addition to the previously determined Asp258 at neutral pH, Asp141, Asp194, and Asp464 are also the coordination sites for the metal binding of malic enzyme. We suggest that the mechanism of affinity modification and cleavage of malic enzyme by the Cu2+-ascorbate system proceed in the following sequence. First, Cu2+ binds with the enzyme at the Mn2+ binding site and reduces to Cu+ by ascorbate. Next, the local oxygen molecules are reduced by Cu+, thereby generating superoxide or other reactive free radicals. These radicals interact with the susceptible essential amino acid residues at the metal-binding site, ultimately causing enzyme inactivation. Finally, the modified enzyme is cleaved into several peptide fragments, allowing the identification of metal site of the enzyme. The pH-dependent different specificities of metal-catalyzed oxidation system may be generally applicable for other enzymes or proteins.

The role of metal ion in the enzymatic reaction entails providing a bridge between malate and the enzyme and functions as a second sphere complex with the substrate (Hsu et al., 1976). Therefore, identification of the ligands for metal binding is ultimately important in understanding the structure-function relationship of this enzyme. In most metalloproteins, the amino acid residues involved in metal binding are dispersed along the complete sequence (Regan, 1993). Without a three-dimensional crystal structure available, affinity cleavage at the putative metal-binding site by the metal-catalyzed oxidation system (MCO) 1 may be the optimal approach of reaching the above goal. Using this technique with the Fe 2ϩ -ascorbate system, Asp 258 is successfully identified in our previous study as one of the metal-binding sites (Wei et al., 1994), as confirmed by site-directed mutagenesis (Wei et al., 1995). In that study, some divalent metal ions were found to be capable of providing protection of the enzyme against Fe 2ϩ -induced inactivation. Among the divalent metal ions, only Cu 2ϩ was found to accelerate the Fe 2ϩ -ascorbate-induced enzyme inactivation rate.
In this work, the inactivation of malic enzyme by Cu 2ϩascorbate system is investigated. We demonstrate, for the first time, that at different pH values Cu 2ϩ -ascorbate system shows different specificities in protein modification and peptide bond cleavage. Taking the advantage of this selectivity, three more metal binding ligands of pigeon liver malic enzyme are successfully identified. Novel sequence motifs for the metal-binding site of malic enzyme are also deduced.

EXPERIMENTAL PROCEDURES
Enzyme Purification and Assay-Malic enzyme from pigeon liver was purified to apparent homogeneity according to published procedure (Chang and Chang, 1982). The enzyme activity was assayed by monitoring the formation of NADPH at 30°C as described previously (Chang et al., 1992).
Enzyme Modification and Cleavage-The inactivation experiments were performed at 0°C by adding freshly prepared solutions of cupric nitrate (6 M) and ascorbate (20 mM) into the enzyme solution (0.97 M) in sodium acetate buffer (66.5 mM, pH 5.0). The progress of enzyme inactivation was monitored by assaying the enzyme activity in small aliquots withdrawn at the designated time intervals. For modification of the enzyme at different pH values, Bis-Tris (pH 4.0 -7.0) or Bis-Trispropane (pH 6.3-9.0) was used as buffer. Other experimental conditions are provided in the figure legends.
For detecting the peptide bond cleavage, the samples withdrawn from the reaction mixture were added to EDTA solution (4 mM) to * This work was supported by National Science Council (Republic of China) Grants 84-2331-B016-063 (to W.-Y. C.) and 82-0412-B016-032 (to G.-G. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed: Dept. of Biochemistry, National Defense Medical Center, P. O. Box 90048, Taipei, Taiwan, 100, Republic of China. Fax: 886-2-365-5746. ** Portions of this work were submitted in partial fulfillment of a M. S. degree (Biochemistry), National Defense Medical Center, Taipei. prevent further reaction from occurring. Next, the protein samples were subjected to SDS-PAGE for separating the peptide fragments. The resulting gel pattern was quantified by densitometric analysis and the isolated peptide fragments were then transblotted to an Immobilon-P membrane and subjected to automatic amino acid sequence analysis as described previously (Wei et al., 1994).
Inhibition Study-Inhibition study was performed at several Cu 2ϩ concentrations, and the concentration of Mn 2ϩ was varied from 1 to 10 M. Concentrations of other components were held constant. The results were fitted to Equation 1 for a competitive inhibition by using the EZ-FIT computer program (Perrella, 1988).
where v is the observed enzyme activity, V m is the maximum enzyme activity, K mMn denotes the Michaelis constants for Mn 2ϩ , and K iCu denotes the inhibition constant for Cu 2ϩ .

Selective Inactivation of Pigeon Liver
Malic Enzyme-Pigeon liver malic enzyme was highly sensitive to metal-catalyzed oxidation (Wei et al., 1994). The inactivation rate of the enzyme by the MCO system was highly pH-dependent. At pH 7.0 and 0°C, 20 M Fe 2ϩ -20 mM ascorbate caused 98% enzyme activity loss in 1 h ( Fig. 1), i.e. in correlation with our previous observation (Wei et al., 1994). On the other hand, under otherwise identical conditions, the enzyme lost only 10% activity in 1 h at pH 5.0. Experimental results indicated that Cu 2ϩ caused a faster inactivation at a smaller concentration than Fe 2ϩ , especially at pH 5.0, in which Cu 2ϩ caused substantial inactivation (Fig. 1).
Inactivation of malic enzyme by Cu 2ϩ required ascorbate in the system; Cu 2ϩ or ascorbate alone did not cause any inactivation. Cu ϩ was much less effective; Cu ϩ (6 M)-ascorbate (20 mM) caused only 10% inactivation in 1 h under conditions that caused Ͼ95% inactivation by the Cu 2ϩ -ascorbate system.
Effect of pH on the Cu 2ϩ -catalyzed Inactivation of Malic Enzyme-The above results indicate that malic enzyme has a different sensitivity toward the metal-catalyzed oxidation system at two different pH values. Next, the inactivation of malic enzyme was investigated by the Cu 2ϩ -ascorbate system between pH 4.0 -9.0 in which the enzyme was stable. There are two optima for the inactivation rate: one at pH 6.0 -7.0, and the other at approximately pH 4.0 (Fig. 2). These results would suggest that different modification mechanisms involved in acidic or neutral pH. For obtaining manageable inactivation rates, the inactivation of malic enzyme by Cu 2ϩ at pH 5.0 is explored in the following experiments.
Dependence of Cu 2ϩ -catalyzed Inactivation of Malic Enzyme on Cu 2ϩ Concentration-At pH 5.0, the inactivation of malic enzyme activity does not follow a pseudo-first-order kinetics as the natural logarithmic of residual activity versus time does not result in a straight line. The inactivation rate is clearly dependent on Cu 2ϩ concentration (Fig. 3), i.e. much smaller in a concentration than Fe 2ϩ requiring to cause the same extent inactivation.
Similar to Fe 2ϩ -induced inactivation, Cu 2ϩ -induced inactivation could be stopped by EDTA (4 mM), which, however, did not reverse the already inactivated enzyme activity.
Protection of Malic Enzyme against Cu 2ϩ -induced Inactivation by Substrates-For demonstrating that the inactivation of enzyme activity was due to modification of essential amino acid residues in or near the active site, the inactivation process was examined in the presence of various combinations of substrates. In contrast to Fe 2ϩ -ascorbate inactivation of malic enzyme, which was completely protected by some divalent cations, Mn 2ϩ (4 mM) alone only protected 56% enzyme activity against the Cu 2ϩ -ascorbate induced inactivation at pH 5.0 ( Fig.  4). L-Malate, which did not give any protective effect in the Fe 2ϩ system, provided 15% protection in the inactivation induced by the Cu 2ϩ system. L-Malate plus Mn 2ϩ yielded synergistic protection (ϳ90%). Nucleotide NADP ϩ , on the other hand, did not provide any protection by itself or in combination with Mn 2ϩ and L-malate in both Fe 2ϩ and Cu 2ϩ systems.
Inhibition of Pigeon Liver Malic Enzyme by Cu 2ϩ -The above results indicate that at pH 5.0 the binding mode between divalent cation and the enzyme is different from that at pH 7.0. Direct kinetic evidence for the binding of Cu 2ϩ at the Mn 2ϩ binding site of malic enzyme was provided by inhibition studies shown in Fig. 5, where Cu 2ϩ was demonstrated to be a competitive inhibitor with respect to Mn 2ϩ with K i value of 70.3 Ϯ 5.8 M. This result indicates that Cu 2ϩ and Mn 2ϩ compete for the same binding site. The fitted value of K mMn was 2.7 Ϯ 0.23 M, which is within the range (1.8 -9 M) as determined previously (Hsu et al., 1976).
For comparison, Cu ϩ was also tested for inhibition of malic enzyme. Cu ϩ as high as 160 M showed no inhibition on the enzyme. At 169 M, the inhibition was only 7%. These results strongly suggest that only divalent Cu 2ϩ has a high affinity with malic enzyme.
Peptide Bond Cleavage Pattern of the Cu 2ϩ -inactivated Malic Enzyme-We have demonstrated previously that, at pH 7.0, the Fe 2ϩ -ascorbate system cleaved malic enzyme at Asp 258 -Ile 259 (Wei et al., 1994). In this study, the cleavage pattern of the enzyme inactivated by Cu 2ϩ -ascorbate system is also ex- amined. The results shown in Fig. 6 clearly indicate that Cu 2ϩ induced the same cleavage as Fe 2ϩ at pH 7.0. However, different specificities were observed at pH 5.0. Fe 2ϩ showed minimum cleavage at pH 5.0, which is in correlation with the minimum enzyme inactivation observed (Fig. 1). Cu 2ϩ produced entirely different cleavage patterns at pH 5.0 (Fig. 6,  lane 6). The two fragments IV and V, which are those seen at pH 7.0, were relatively small in quantity at pH 5.0. Instead, six major fragments are detected (EЈ I , EЈ II , EЈ III , EЈ VI , EЈ VII , and EЈ VIII ). On the basis of M r estimation, these fragments seem to be due to three cleavages. Site A cleavage produces EЈ II (48,000) and EЈ VII (M r 17,000), site B cleavage produces EЈ III (M r 38,000) and EЈ VI (M r 27,000), and cleavage at site C produces EЈ I (M r 55,000) and EЈ VIII (M r 10,000). Probably because of a low molecular mass, the diffusion problem renders EЈ VII and EЈ VIII less defined visually. However, we successfully obtained a discrete amino acid sequence that is the fourth coordination site of Mn 2ϩ of malic enzyme as described in the following sections.
Correlation between Enzyme Activity Inactivation and Peptide Bond Cleavage-To correlate the Cu 2ϩ -induced enzyme inactivation and peptide bond cleavage, the modification was performed in various stages, the reaction was stopped with EDTA, and the protein samples were subjected to SDS-PAGE to examine the peptide bond cleavage. Results shown in Table  I clearly indicate that with the increasing of incubation time, a rapid loss in enzyme activity occurred and the peptide bond cleavage increased. However, the enzyme activity lost proceeded much faster than the peptide bond cleavage. When the enzyme activity was down to 27%, 91% of the enzyme molecules were still intact. We can conclude that peptide bond cleavage follows the enzyme inactivation, as for other MCO systeminduced enzyme inactivation. Site C is a minor cleavage site as compared to sites A and B. Since EЈ II ϩ EЈ VII and EЈ III ϩ EЈ VI appear to have an equal amount (Table I), site A and site B seem to have a similar probability of cleavage. However, these sites do not cleave simultaneously in the same enzyme molecule. Similar mutually exclusive cleavages were also observed for pig heart isocitrate dehydrogenase with Fe 2ϩ -ascorbate system (Soundar and Colman, 1993).
The switch of specificity was examined by monitoring the protein cleavage pattern at various pH values. As the pH decreased from neutral to acidic, fragments I, II, III, VI, VII, and VIII gradually increased, while fragments IV and V decreased (data not shown). Fig. 6, we see that, at pH 5.0, Cu 2ϩ cleaved malic enzyme at three sites that are different from the Asp 258 -Ile 259 observed at pH 7.0. The SDS-PAGE-separated peptide fragments were electrophoretically transblotted onto an Immobilon-P membrane and each peptide was analyzed via an automatic protein sequencer. The N terminus of fragment II was found to have the following sequence: Gly-Glu-Arg-Ile-Leu-Gly-Leu-Gly-Asp-Leu-X-X-X-Gly-Met-Gly-Ile-X-X-Gly, which is identified as the known sequence between Gly 142 -Gly 161 of the cDNA sequence of pigeon liver malic enzyme . Peptide EЈ III has the N-terminal sequence Pro-Leu-Tyr-Ile-Gly-Leu-Arg-His-Lys-Arg-Ile-Arg-Gly-Gln-Ala-Tyr-Asp-Asp, which is identified as Pro 195 -Asp 212 of the cDNA sequence. Peptide EЈ VIII has the N-terminal sequence Asp-Val-Phe-Leu-Thr-Thr-Ala-Glu-Val-Ile-Ala-Gln-Glu-Val-Ser-Glu-Glu-Asn-Leu-Gln, which is identified as Asp 465 -Asn 484 of the cDNA sequence.

Identification of the Metal-binding Site of Pigeon Liver Malic Enzyme-From
Corresponding to peptide EЈ II and EЈ III are peptides EЈ VII and EЈ VI , which contain the N-terminal peptides Met 1 -Asp 141 and Met 1 -Asp 194 , respectively, and should not detect any sequence since the N terminus of malic enzyme is blocked (Wei et al., 1994). When peptides EЈ VI and EЈ VII were subjected to amino acid sequence analysis, multiple small peaks were detected and identifying any reliable sequence was relatively difficult. When consulting the gel pattern shown in Fig. 6, we can conclude that while peptides EЈ II -EЈ VII represent the major cleavage products, minor cleavage at other sites also occurred. Peptide EЈ VI contains sequence Met 1 -Asp 194 and some other minor peptides. This is also true for peptide EЈ VII . The failure in identifying some amino acid residues in EЈ II may be due to the interference of other minor peptides. A sufficient quantity of fragment I for sequence analysis was not collected in this study to confirm that it is blocked at the N terminus.

DISCUSSION
Pigeon liver malic enzyme is among the most sensitive enzymes toward MCO system. We have demonstrated previously that this enzyme was rapidly inactivated by the Fe 2ϩ -ascorbate system at neutral pH. The inactivation was rapid that incubation at 0°C was necessary to slow down the reaction rate. In this study, the Cu 2ϩ -ascorbate system was found to be more effective than the Fe 2ϩ system. Experimental results indicated that the Cu 2ϩ system produced an entirely different cleavage pattern at pH 5.0 as compared to the pattern at pH 7.0. The following criteria indicate that Cu 2ϩ -ascorbate caused affinity modification and cleavage of pigeon liver malic enzyme at the putative metal-binding site. (a) Cu 2ϩ is essential for the process; Cu ϩ is much less effective, suggesting that Cu 2ϩ must bind with the enzyme at the divalent metal ion binding site before inactivation takes place. (b) Cu 2ϩ is a competitive inhibitor versus Mn 2ϩ for the enzyme, indicating that Cu 2ϩ and Mn 2ϩ compete for the same binding site. (c) Inactivation of the enzyme is prevented by Mn 2ϩ plus L-malate, indicating that modification is at the active site.
Based on the above discussion, the reaction sequence of oxidative modification and peptide bond cleavage of pigeon liver malic enzyme by the Cu 2ϩ -ascorbate system can be summarized in Scheme I. First, Cu 2ϩ binds with the enzyme at the Mn 2ϩ binding site (step 1). Second, ascorbate reduces Cu 2ϩ to Cu ϩ (step 2), which, in the presence of dissolved O 2 , generates reactive free radicals (e.g. O 2 . , OH ⅐ ) that in turn modify the essential amino acid residue(s) nearby and forming the inactivated enzyme (EЈ) (step 3). Finally, depending on pH of the solution, the enzyme molecule is cleaved at four possible sites giving peptide fragments EЈ I ϩ EЈ VIII , EЈ II ϩ EЈ VII , or EЈ III ϩ EЈ VI at pH 5.0, or EЈ IV ϩ EЈ V at pH 7.0 (step 4). In this manner, different specificities are achieved by manipulating the reaction conditions. We suggest that this strategy can be generally applied to other enzymes or proteins in elucidating the metalbinding sites. We propose that the catalytically essential carboxyl group of Asp 258 has a pK a value of 6.7, as determined by chemical modification experiments (Chang et al., 1985). At pH 5.0, this carboxyl group is protonated and loss its metal-binding ability and is not reactive toward oxidative modification. Under this circumstance, Fe 2ϩ is inactive but other metal ligands Asp 141 , Asp 194 , or Asp 464 , which might have pK a values near 4.7, are modified by the more reactive Cu 2ϩ . Interestingly, the metal ligands Asp 141 , Asp 194 , Asp 258 , and Asp 464 of malic enzyme were all aspartate residue, which has been indicated to be the major metal-binding ligand for many metal-proteins (Higaki et al., 1992;Vallee and Auld, 1993;Traut, 1994). Furthermore, isocitrate dehydrogenase, which catalyzes a similar oxidative decarboxylation reaction as malic

TABLE I
Correlation of Cu 2ϩ -catalyzed inactivation of pigeon liver malic enzyme activity and cleavage of the metal binding site Experimental conditions were the same as in Fig. 6. The enzyme samples inactivated to various degrees were subjected to SDS-PAGE separation of the cleaved and uncleaved molecule and quantified with a densitometer. Original enzyme amount was taken as 100%. The minor fragments I and VIII were not included in the calculation.  (Hurley et al., 1990); this enzyme from pig heart was also sensitive to MCO system (Soundar and Coleman, 1993). Examination of the sequences shown in Figs. 7-9 reveals that Asp 141 and Asp 194 , the major cleavage sites by the Cu 2ϩ system, are strictly conserved in all malic enzyme with known amino acid sequences. Asp 464 is also highly conserved; however, this region has higher variations among malic enzyme of different origins. This site is a minor cleavage site. Maximum alignment of other malic enzyme sequences with Asp 464 of pigeon enzyme reveals that Glu 464 of duck, Asn 512 of ascaris, and Glu 557 of maize enzymes may be the metal coordinates. Although Asn and Glu are found as metal ligand in many metal-proteins (Villafranca and Nowak, 1992;Higaki et al., 1992;Vallee and Auld, 1993;Traut, 1994), an observation of the nearby amino acid residues reveals that Asp 465 of duck, Asp 511 or Asp 513 of ascaris, and Asp 558 of maize enzyme may be the authentic metal ligands. These results enforce the critical value of aspartate residue as the metal coordinate in proteins.
Only bacillus malic enzyme was nonconservatively substituted this Asp with Val 456 . The actual metal ligand, however, may be Glu 457 , which is also a conservative substitution. Malic enzyme is a bifunctional enzyme. It catalyzes both the oxidoreduction and decarboxylation reactions. The reaction mechanism of the enzyme proceeds in two steps with hydride transfer preceding decarboxylation (Hermes et al., 1982). These two functions can be assessed separately by assaying the par- tial reactions with appropriate substrate (Hsu, 1982). One of the distinguishing features of the decarboxylase activity of malic enzyme is its pH optimum being at 4.5 (Salles and Ochoa, 1950). We propose that the proton released during dehydrogenase reaction provides a favorable local active site environment for the decarboxylation reaction, which involves metal ionstabilized enolate anion transition state (Hsu et al., 1976;O'Leary, 1992). During the catalytic cycle, the enzyme might undergo an isomerization that favors the decarboxylation reaction. Results presented in this study support the hypothesis that the enzyme exists as different conformational isoforms at neutral or acidic environment. However, another possibility that the observed pH effects were due to changes in the catalytic cleavage rate of different sites rather than to global conformational changes was not ruled out (Kufel and Kirsebom, 1994).
Metal-catalyzed oxidation of proteins has been suggested to be the marker for protein turnover in vivo (Stadtman and Oliver, 1991;Stadtman, 1992). According to this theory, the oxidized protein molecules are unstable and prone to degradation by multicatalytic proteases (proteasomes) in cells (Rivett, 1993). The normal copper concentration in serum is 16 -31 M (Murray et al., 1990); both iron and copper ions are normally presented in the cells, which also contain various reducing compounds. If malic enzyme is sensitive to only a few micromolar concentration of Cu 2ϩ in the cytoplasm, an intriguing question is that how could malic enzyme survive in vivo for a reasonable period of time to perform its metabolic roles? Protection of the enzyme by Mg 2ϩ or other divalent cations and substrate is one of the answers. Furthermore, cells contain other defense mechanisms against oxidative damage. The endogenous antioxidant enzymes (catalase, dismutase) are active in normal young tissues; a rapid physiological response of the translational or transcriptional control of the detoxification genes might play an important role when metal ion concentrations exceed a dangerous threshold (O'Halloran, 1993). These detoxification gene products may play important roles in cell protection, e.g. the high Fe 2ϩ content and oxygen carrier function of red blood cells indicate the high oxidation stress experience by red blood cells, which contains an abundant natural killer enhancing factor that protects red blood cells from oxidation injuries (Shau and Kim, 1994). These protecting proteins are highly homologous with yeast thiol-specific antioxidant, which protects yeast cells from oxidation insults (Chae et al., 1993). It is these natural protecting mechanisms that prevent cell proteins and other cellular components from experiencing oxidative damage. SCHEME I. Proposed reaction mechanism for the oxidative modification and affinity cleavage of pigeon liver malic enzyme by the Cu 2؉ -ascorbate system. The N and C under each peptide fragment denote the N-or C-terminal half before cleavage.