Inactivation of Glucose-6-phosphate Dehydrogenase by 4-Hydroxy-2-nonenal SELECTIVE MODIFICATION OF AN ACTIVE-SITE LYSINE*

Incubation of glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides with 4-hydroxy-2-nonenal (HNE) results in a pseudo first-order loss of enzyme activity. The pH dependence of the inactiva- tion rate exhibits an inflection around pH 10, and the enzyme is protected from inactivation by glucose 6-phosphate. Loss of enzyme activity corresponds with the formation of one carbonyl function per enzyme subunit and the appearance of a lysine-HNE adduct. The data presented in this paper are consistent with the view that the e-amino group of a lysine residue in the glucose 6-phosphate-binding site reacts with the double bond (C,) of HNE, resulting in the formation of a stable secondary amine derivative and loss of enzyme activity. We have described a mechanism by which HNE may, in part, mediate free radical damage. In addition, a method for the detection of the lysine-HNE adduct is introduced.

Reactive oxygen species readily interact with membrane lipids, often resulting in the formation of unsaturated aldehydes such as hydroxyalkenals (1). These aldehydes are more stable than free radical species and may more readily diffuse into cellular media, where they are available for facile reaction with various biomolecules. Modification of protein and other biomolecules by lipid peroxidation products is believed to play a central role in many pathophysiological conditions often associated with free radical damage (2)(3)(4)(5). The mechanisms and relative contributions of several potential reactions, however, are not well understood.
4-Hydroxy-2-nonenal (HNE)' (Fig. l), one of the major products of membrane peroxidation, shows many biological effects such as high toxicity to Ehrlich ascites tumor cells (6), the lysis of erythrocytes (7), and inhibition of the synthesis of DNA and protein (3,8). HNE is likely to exert these effects because of its reactivity with biological materials, particularly with protein. It has been demonstrated that much of the conjugation of protein with HNE involves the sulfhydryl group of cysteine (2, 9). Recently, histidine-HNE adducts * 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.
$ Present address: Dept. of Food Science and Technology, School of Agriculture, Nagoya University, Nagoya 464-01, Japan. were also identified as major products in various proteins treated with HNE (10). Further characterization of the modification of protein by HNE may lead to a better understanding of the mechanism of a large number of biological effects induced by HNE and other lipid peroxidation products.
We report here results of an i n uitro study on the HNE modification of glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides. Glu-6-P dehydrogenase from L. mesenteroides was selected for this study because it lacks cysteine residues (11); therefore, any effect of HNE on catalytic activity cannot be due to modification of sulfhydryl groups. We found that incubation of Glu-6-P dehydrogenase with HNE results in a pseudo first-order loss of enzyme activity. We provide evidence that the loss of activity is due to the reaction of a lysine residue in the glucose 6-phosphatebinding site with the double bond (C-3) of HNE. A method for the detection of this kind of lysine-HNE adduct on proteins was developed. These results provide the basis for future i n uiuo studies to assess the role of protein modification by HNE and other lipid peroxidation products in certain diseases, pathophysiological conditions, and aging.

MATERIALS AND METHODS
Preparation of Glucose-6-phosphate Dehydrogenase-Glu-6-P dehydrogenase purified from L. mesenteroides was purchased from Sigma a t a concentration of 8.8 mg/ml in 3.2 M (NH4)2S04. Prior to exposure of the enzyme to HNE, (NH4)2S04 was removed by passage through a Sephadex G-25M PD-10 column (Pharmacia LKB Biotechnology Inc.). The enzyme was eluted with 10 mM MOPS, 100 mM KC1 at pH 7.4. This process was then repeated to ensure complete removal of (NH&S04. The enzyme, a t a final concentration of -2.0 mg/ml, was stable a t room temperature for up to 12 h. 4-Hydroxy-2-nonenal Preparation-4-Hydroxy-2-nonenal dimethylacetal was synthesized as previously described (12). Prior to use, HNE was generated by acid treatment (1 mM HCl) of 4-hydroxy-2nonenal dimethylacetal. The 'H NMR spectra of 4-hydroxy-2-110nenal dimethylacetal and 4-hydroxy-2-nonenal were in agreement with published results (12). The concentration of HNE was determined by measurement of UV absorbance a t 224 nm with a molar absorptivity of 13,750 M" for HNE.
Measurement of Glucose-6-phosphate Dehydrogenase Activity-Glu-6-P dehydrogenase was diluted in 10 mM MOPS, 100 mM KC1 at pH 7.4 to give a final protein concentration of -25 ng/ml. The enzyme-catalyzed reaction, performed a t 25 'C, was initiated by the addition of a 50-111 aliquot of 10 mM NADP+ and 100 mM glucose-6 phosphate to 950 pl of enzyme solution. Glu-6-P dehydrogenase activity was then determined by following the appearance of NADPH spectrophotometrically at 340 nm using a Hewlett-Packard Model 8452.4 diode array spectrophotometer. Enzyme activity was linear under all conditions tested over the time assayed.
Incubation of Glucose-6-phosphate Dehydrogenase with 4-Hydroxy-2-nonenal-Glu-6-P dehydrogenase (-0.5 mg/ml) was incubated a t 37 "C in 10 mM MOPS, 100 mM KC1 at pH 7.4 (except where stated) with various concentrations of HNE for the indicated periods of time. Inactivation was arrested by diluting the inactivation mixture 20,000fold in 10 mM MOPS, 100 mM KC1 at pH 7.4. Enzyme activity was then determined as described above. Prior to amino acid analysis or treatment of the modified enzyme with 2,4-dinitrophenylhydrazine or NaB3H4, unreacted HNE was removed by passage over a Sephadex G-25M PD-10 column.
Treatment of Modified Glu-6-P Dehydrogenase with 2,4-Dinitrophenylhydrazine-The carbonyl content of the protein was determined using a variation of a previously described technique (13). Enzyme (-0.25 mg), taken to dryness using a rotary evaporator, was resuspended in 500 p1 of 10 mM 2,4-dinitrophenylhydrazine, 6.0 M wanidine HC1, and 0.5 M KH2P04, pH 2.5. The enzyme was incubated with 2,4-dinitrophenylhydrazine a t 25 "C for 90 min. Derivatized protein was then separated from excess reagent by filtration of the reaction mixture through a Zorbax GF250 gel column (Du Pont-New England Nuclear) using a Hewlett-Packard Model 1090 high performance liquid chromatograph equipped with a diode array UV detector. Derivatized enzyme was eluted with a solution containing 6.0 M guanidine HC1 and 0.5 M KHzPO,, pH 2.5. Elution profiles were detected a t 276 and 370 nm. Molar ratios of carbonyl content to enzyme subunit were then calculated using molar absorptivities of 9500 and 22,000 M" for the hydrazone a t 276 and 370 nm, respectively. The molar absorptivity for Glu-6-P dehydrogenase in 6.0 M guanidine HCl and 0.5 M KHZPO,, pH 2.5, is 60,000 M" at 276 nm.
Treatment of Modified Glu-6-P Dehydrogenase with NaBH4-A 400-p1 aliquot of HNE-modified Glu-6-P dehydrogenase (0.25 mg/ ml) was mixed with 0.1 M EDTA (40 pl) and 1 N NaOH (40 pl) in a 1.5-ml Sarstedt tube fitted with an O-ring and cap. A 40-p1 aliquot of 0.1 M NaB3H4 in 0.1 N NaOH was then added, and the mixture was incubated a t 37 "C for 1 h. The reaction was terminated by the addition of 100 pl of 1.0 N HC1. T o separate 3H-labeled protein from other radioactive contaminants, the reaction mixture was chromatographed over a Sephadex G-25M PD-10 column using 6.0 M guanidine HCl as eluant. Recovery of protein was determined by the measurement of UV absorbance at 278 nm with a molar absorptivity of 60,000 M" for Glu-6-P dehydrogenase. The protein carbonyl content was calculated from radioactive measurement of the suitable aliquots. The

FIG. 2. Fractional loss of glucose 6-phosphate dehydrogenase activity as function of time in presence
of 4-hydroxy-2-nonenal. Enzyme a t a concentration of 0.5 mg/ml was incubated with 1.5 mM 4-hydroxy-2-nonenal at 37 "C for the times indicated on the abscissa. At the indicated times, enzyme was separated from unreacted 4-hydroxy-2-nonenal by gel filtration chromatography, and remaining activity was determined as described under "Materials and Methods." The inset presents the same data expressed as the log of fractional loss of glucose 6-phosphate dehydrogenase activity versus time. 100 80 60 40 20 specific radioactivity of NaB3H4 was determined by reaction with 4.0 M acetone in 100 mM NaOH (14) and the subsequent measurement of incorporated 3H.
Amino Acid Analysis-A 100-p1 aliquot of modified Glu-6-P dehydrogenase (0.25 mg/ml) was treated with 10 pl of 10 mM EDTA, 10 p1 of 1 N NaOH, and 10 p1 of 100 mM NaBH4 in 0.1 N NaOH. After 1 h a t 37 "C, the reaction was terminated by the addition of 40 pl of 1.0 N HC1. The solution was brought to dryness using a rotary evaporator, and the protein was subsequently hydrolyzed with 6 N HCl (200 pl) for 20 h a t 110 "C under nitrogen atmosphere. The hydrolyzed sample was evaporated to dryness and resuspended in 2.0 ml of 50 mM sodium phosphate buffer, pH 8.0, containing 1 mM EDTA. A lO-pl aliquot of the hydrolyzed protein solution was labeled with o-phthaldehyde (15). Derivatized amino acids were analyzed by reverse-phase HPLC using a CIS column (Jones Chromatography) fitted to a Hewlett-Packard Model 1090 chromatograph equipped with a Hewlett-Packard Model 1046A programmable fluorescence detector.
Lysine-4-Hydroxy-2-nonenal Adduct-The standard sample of the HNE-lysine adduct was prepared by reaction of HNE with N-acetyllysine. N-Acetyllysine (50 mM) was treated with 5 mM HNE in 50 mM sodium phosphate buffer, pH 7.2, for 20 h at 37 "C. Formation of products was determined by HPLC. A linear gradient of 0.05% trifluoroacetic acid in water (solvent A)/acetonitrile (solvent B) (time = 0, 100% A 20 min, 0% A) at a flow rate of 1 ml/min was used with a TSK-GEL ODs-80TM column (TOSOHAAS). The major component was collected and further characterized by amino acid and mass spectral analyses. Amino acid analysis was performed as outlined above. A JEOL JMS-SXlO2 mass spectrometer was used for fast atom bombardment-mass spectrometry.

RESULTS AND DISCUSSION
Inactivation of Glu-6-P Dehydrogenase by 4-Hydroxy-2-nonenal-Incubation of Glu-6-P dehydrogenase with HNE resulted in a loss of enzyme activity. Approximately 70% of the initial activity was lost when the enzyme was exposed to 1.5 mM HNE at pH 7.4 and 37 "C for 90 min (Fig. 2). A semilogarithmic plot of remaining activity versus incubation time (Fig. 2, inset) is linear over the experimental interval examined, demonstrating that inactivation occurred by a pseudo first-order process. The fractional loss of Glu-6-P dehydrogenase activity was independent of enzyme concentration over a range pertinent to these experiments. Inactivation was quenched by dilution. There was no regain in activity following dilution of the inactivation mixture or separation of the enzyme from HNE by gel filtration.
Increases in HNE concentration resulted in increased rates of inactivation. The linear relationship between HNE concentration and kob. (Fig. 3) indicates that the rate of inactivation is first-order with respect to HNE concentration. The fact that the intersect at the ordinate is -0 indicates that the inactivation reaction is essentially irreversible. As shown in Observed rate constant (kobs) for inactivation of glucose-6-phosphate dehydrogenase by 4-hydroxy-2-nonenal as function of pH. Enzyme at a concentration of 0.5 mg/ml was incubated with 1.0 mM 4-hydroxy-2-nonenal a t 37 "C for 0-2 min a t the pH values indicated on the abscissa. Inactivation was arrested by dilution, and remaining activity was determined as described under "Materials and Methods." Fig. 3 (inset), the double-reciprocal plot of the initial rate versus HNE concentration is linear with an intercept near the origin. These data suggest that the Kd for the Glu-6-P dehydrogenase-HNE complex is much larger than 5.0 mM, if such a complex indeed forms for the inactivation reaction. It is likely that inactivation is the result of a bimolecular collision between an exposed nucleophilic residue on the protein and HNE, with a second-order rate constant of 11.0 min" mM". inactivation mixture was tested to identify sites of modification. We found that glucose 6-phosphate protected Glu-6-P dehydrogenase from HNE inactivation. A double-reciprocal plot of the difference between the rate of inactivation in the absence and presence of glucose 6-phosphate versus glucose 6-phosphate concentration (Fig. 4) reveals that glucose 6phosphate inhibits inactivation of the enzyme by HNE in a noncompetitive fashion with a Ki of -1.5 mM. Glucose 1phosphate, glucose, phosphate, and NADP+ had no effect on inactivation of the enzyme by HNE, indicating that protection afforded by glucose 6-phosphate is highly specific. As previously noted, HNE does not appear to form a stable complex with Glu-6-P dehydrogenase. Therefore, the noncompetitive pattern obtained with glucose 6-phosphate does not exclude the possibility that HNE reacts with amino acid residue(s) in or near the glucose 6-phosphate-binding site. The observed Ki (1.5 mM) for glucose 6-phosphate inhibition of HNE inactivation is 10 times greater than the K, (150 PM) of Glu-6-P dehydrogenase for glucose 6-phosphate. This difference between K; and K , is not surprising since the K, of Glu-6-P dehydrogenase for glucose 6-phosphate is not a measure of binding affinity (16).
In an attempt to provide information on the amino acid residue(s) involved in the inactivation of Glu-6-P dehydrogenase by HNE, we explored the effect of pH on this process. As shown in Fig. 5, the rate of inactivation increased with increasing pH. The inflection point was a t a pH of 10, a value comparable to the pK, values of lysine and arginine. Ionic strength had little or no effect on the rate of inactivation (data not shown). In the absence of HNE, no loss of enzyme activity was observed over the pH range used in this study throughout the 2-min incubation period. At pH values above 11.5, the enzyme was highly unstable. Furthermore, glucose 6-phosphate protected the enzyme from HNE inactivation at all pH values tested (data not shown). This fact and the demonstration that Glu-6-P dehydrogenase from L. mesenteroides contains a lysine residue required for the binding of the phosphate moiety of glucose 6-phosphate (17)(18)(19) are consistent with the view that loss of activity results from reaction of HNE with the lysine residue involved in the binding of glucose 6-phosphate. Like other a,&unsaturated aldehydes, 4-hydroxy-2-nonenal should be susceptible to nucleophilic addition at both the double bond (C-3) and the carbonyl moiety (CJ. Accordingly, the eamino group of lysine residues on a protein may react with HNE to form a secondary amine derivative with Frc. 8. Molar ratio of carbonyl to enzyme subunit as function of fractional loss of glucose-6-phosphate dehydrogenase activity. Following treatment of the enzyme as described for Fig. 2, 2,4-  base with retention of the double bond (Fig. 7). The relative contributions of these two kinds of enzyme modification may be deduced by studying the reactions of HNE-modified protein with 2,4-dinitrophenylhydrazine and with sodium borotritide. If the t-amino group of lysine reacts with HNE at the double bond of HNE (Fig. 6), the carbonyl function of the protein adduct will react with 2,4-dinitrophenylhydrazine to form a nondissociable hydrazone derivative. However, the reaction of 2,4-dinitrophenylhydrazine with an enzyme-HNE Schiff base adduct will lead to cleavage of the protein-HNE adduct at the Schiff base bond and to the formation of the free HNE hydrazone derivative, which is readily separated from protein (Fig. 7). By contrast, sodium borotritide will react with both kinds of protein-HNE conjugates to yield stable 3H-labeled protein derivatives. It follows that the fraction of total HNE adduct formed by reaction of the t-amino group of lysine with the double bond of HNE is given by the molar ratio of the protein-bound product formed in the reaction with 2,4-dinitrophenylhydrazine and with sodium borotritide, i.e. moles of protein-bound hydrazone/moles of pro- tein-bound 3H. Reaction of 2,4-dinitrophenylhydrazine with Glu-6-P dehydrogenase that had been inactivated from 0 to 70% indicated a one-to-one relationship between the fraction of activity lost and the moles of hydrazone derivative incorporated per mole of enzyme subunit (Fig. 8). Treatment of the enzyme with HNE for extended periods of time resulted in a decrease in this ratio. Reaction of partially inactivated enzyme with sodium borotritide gave similar results: a oneto-one relationship between the fraction of activity lost and the moles of 3H incorporated per mole of enzyme subunit. Protection of Glu-6-P dehydrogenase from HNE inactivation by the addition of glucose 6-phosphate resulted in a level of 100 .4 2,4-dinitrophenylhydrazine-detectable carbonyl which correlates with the degree of inactivation rather than the time of incubation. These data indicate that inactivation is the result of selective modification of Glu-6-P dehydrogenase by HNE: 1 mol of HNE is bound per mol of enzyme subunit and appears t o result in inactivation. The molar ratio of the product formed in the reaction of 2,4-dinitrophenylhydrazine and sodium borotritide is also 1, indicating that there is no detectable reaction between protein and the carbonyl group of HNE. Instead, a nucleophilic residue on the enzyme reacts with the double bond ((2-3) of HNE, resulting in the incorporation of a stable carbonyl group on the protein and loss of enzyme activity.

Time (min)
It is difficult to determine which amino acid is lost upon HNE inactivation of Glu-6-P dehydrogenase by conventional amino acid analysis. This is particularly true for lysine, of which there are 36 residues/enzyme subunit (11,19). We therefore performed a product analysis of the reaction of Nacetyllysine with HNE. Fast atom bombardment-mass spectrometry analysis of the major product of this reaction revealed a quasi molecular ion of 345, consistent with the formation of the lysine-HNE adduct shown in Fig. 9. Following NaBH, reduction, amino acid analysis of this adduct revealed a peak distinct from that of unreacted lysine (Fig.  10). In addition, incubation of this compound with NaB3H4 resulted in the incorporation of 1 mol of 3H/mol of lysine-HNE adduct.
Amino acid analysis of Glu-6-P dehydrogenase inactivated t o varying degrees by reaction with HNE followed by stabilization with NaBH, revealed a peak on the HPLC chromatogram with the same elution profile as that of the lysine-HNE adduct formed by reaction of N-acetyllysine with HNE (Fig.  10). The area of this peak increased with a corresponding increase in the level of inactivation. Protection from inactivation by the addition of glucose 6-phosphate to the inactivation mixture prevented the formation of this compound. Upon near complete inactivation, -1 mol of lysine-HNE adduct was formed per mol of enzyme subunit. The levels of lysine-HNE adduct determined at various stages of inactivation were, however, difficult to quantify with much certainty due to the relatively low level of modification. As previously noted, HNE treatment of Glu-6-P dehydrogenase also resulted in a relatively small amount of histidyl modification. Since inactivation does not appear to involve strong binding of HNE to the enzyme (Fig. 3), it is likely that certain amino acid residues present in the glucose 6-phosphate-binding site are exposed to solvent, while other nucleophilic amino acid residues on the protein are not.
Structural and Functional Properties of 4-Hydroxy-2-nonenal-modified Glu-6-P Dehydrogenase-In an attempt to identify the properties of partially modified Glu-6-P dehydro-genase, we determined the molecular size, kinetic parameters, and thermal stability of enzyme inactivated to various degrees. As judged by gel filtration chromatography, modification did not result in cleavage of the peptide chain, dissociation of active dimer to inactive monomer, or any appreciable crosslinking under the conditions of our experiments. The K,,, values of the remaining active catalytic sites for glucose 6phosphate (-150 PM) and NADP' (-15 p~) remained virtually unchanged at various stages of inactivation, indicating that loss of enzyme activity is likely caused by reduction of active enzyme population. In addition, enzyme inactivated to various degrees by HNE exhibited little change in the thermal stability of remaining activity, as determined by incubation at 47 "C for varying periods of time.
Modification of protein and other biomolecules by lipid peroxidation products is believed to play a central role in many pathophysiological conditions often associated with free radical damage. We have described, using a model system, a mechanism by which HNE may, in part, mediate free radical damage. In addition, a method for the detection of the lysine-HNE adduct was introduced. Our results will therefore be useful for future in vivo studies that attempt to define the mechanisms of free radical impairment of cellular function.