Biochemistry 1993, 32, 68216827
6821
Articles Characterization of the Inhibition of Rabbit Muscle Adenylate Kinase by Fluoride and Beryllium Ionst JBrdme Garin'J and Pierre V. Vignaiss Laboratoire de Chimie des Protbines and Laboratoire de Biochimie (URA 1I30 CTVRS),Dbpartement de Biologie Molbculaire et Structurale, Groupement CEA- CNRS-INSERM- UJF, Centre d'Etudes Nuclbaires de Grenoble, 85X,38041 Grenoble Cedex, France Received November 13, 1992; Revised Manuscript Received March 22, 1993
ABSTRACT: In the presence of NaF, BeC12, ADP, and Mg2+, rabbit muscle adenylate kinase was strongly inhibited. Molecules responsible for the inhibition were identified after subjecting the inhibited enzyme to filtration-centrifugation through small Sephadex G-50 columns. Full inhibition was correlated with the entrapment of 2 mol of ADP, 1 mol of beryllium, and 1 mol of magnesium per mole of enzyme; when an excess of AMP was present together with ADP, only 1 mol of bound ADP was detected in the desalted inhibited enzyme, due to the release of a loosely bound AMP. The fluorometal species identified in the ADP-fluoroberyllate entrapped complex were BeF+, BeF2, and BeF3-. When inhibited adenylate kinase was diluted in a medium deprived of the inhibitory species, the enzyme activity was progressively recovered, and a k0R value of 0.23 min-l for the release of the inhibitor complex was calculated. The possibility that the inhibitory nucleotide-fluoroberyllate complex behaves as a transition-state analog is discussed.
Inhibition or activation of a number of nucleotide binding proteins by millimolar concentrations of fluoride has been reported for many years (Hewitt & Nicholas, 1963). Only recently has the role of contaminating aluminum or beryllium ions been demonstrated in this process (Sternweis & Gilman, 1982). Al3+ and Be2+ are known to bind up to six and four fluoride anions, respectively, generating stable fluorometal complexes (Goldstein, 1964; Mesmer & Baes, 1969). In solution these complexes combinewith nucleotides, as revealed by NMR (Issartel et al., 1991a). The fluorometal complexes behave as phosphate analogs. In the case of transducin, a G protein involved in visual excitation, they combine with GDP in the nucleotide binding site of the protein, and the resulting GDP-fluorometal complex mimicks GTP, thus promoting the activation of transducin (Bigay et al., 1985). Likewise, fluorometals have been found to affect the activities of mitochondrial ATPase (Lunardi et al., 1988),tubulin (Carlier et al., 1988), actin (Combeau & Carlier, 1988, 1989), and myosin (Maruta et al., 1991). The present study deals with the effect of the fluoroberyllate complex on adenylate kinase, a low molecular mass phosphotransferase that catalyzes the transfer of a phosphoryl group from ATP to AMP in the presence of Mgz+ according to the reaction MgATP + AMP e=r MgADP ADP (Noda, 1958;Kubyetal., 1962;Hamada & Kuby, 1978). Thekinetics of inhibition of adenylate kinase in the presence of fluoride and beryllium has been studied, and the correlation of inhibition with the binding of nucleotidefluorometal has been established.
+
This work was supported by grants from the "Direction des Sciences du Vivant" of CEA, from the "Centre National de la Recherche Scientifique" (URA 1130), and from the "Universitd Joseph Fourier, Facultd de Medecine de Grenoble". * Author to whom correspondence should be addressed. Laboratoire de Chimie des Prot6ines. 8 Laboratoire de Biochimie.
0006-296019310432-6821$04.0010
MATERIALS AND METHODS Chemicals. All reagents used were of the purest grade commercially available. Adenylate kinase, pyruvate kinase, lactate dehydrogenase, AMP, ADP, ATP, GDP, and IDP were from Boehringer (Mannheim). ApsA and d-ADP1were from Sigma. Beryllium chloride was obtained from Fluka, and ultrapure sodium fluoride was from Riedel de Haen. NaF solutions were made and stored in plastic containers. All reactions were performed in plastic tubes. Adenylate Kinase Preparation. Before use, rabbit muscle adenylate kinase (380 units/mg) was further purified to apparent homogeneity. Fifteen milligrams of the enzyme (ammonium sulfate suspension) was resuspended in 3 mL of buffer A (20 mM Tris-HC1, 100 mM NaC1, 5 mM DDT, 1 mM EDTA, and 55% ammonium sulfate, final pH 7.5). The turbid solution was centrifuged for 15 min at l5OOOg (+4 "C). The supernatant was loaded on a Sephacryl S-100 HR (Pharmacia) column (100 X 2.5 cm) equilibrated with buffer A, except the ammonium sulfate was omitted. The fractions corresponding to the main peak of absorbancy at 280 nm were pooled, and ammonium sulfate was added to reach 85% saturation. As shown by analysis of the enzyme by HPLC (Cq column, 7 pm, 1000 A), the Sephacryl chromatography efficiently removed five minor contaminants representing about 25% of the total protein content. The specific activity of the purified protein was about 500 units/mg, and its molecular mass analyzed by ion spray mass spectrometry was 21 680 f 2 Da. This mass value agrees with that predicted from the enzyme amino acid sequence (Kuby et al., 1984), but does not suggest that the N-terminal acetyl-methionyl residue is oxidized as reported by the same authors. Abbreviations: Ap5A, bis(5'-adenosyl) pentaphosphate;d-ADP, 2'deoxyadenosine 5'-diphosphate; EDC, l-ethyl-3-[3-(dimethylamino)propyllcarbodiimide; LDH, lactate dehydrogenase; PEP, phosphoenolpyruvate;pF, d o g of the free fluoride concentration; PK, pyruvate kinase.
0 1993 American Chemical Societv
6822 Biochemistry, Vol. 32, No. 27, 1993 We have tested the effectiveness of our adenylate kinase preparation by its ability to bind the high-affinity ligand [3H]Ap5A. [3H]ApsA was synthesized by a procedure adapted from the procedure of Ng and Orgel (1987). l-Ethyl-3-[3(dimethylamino)propyl]carbodiimide(EDC) was used as the coupling agent between [3H]ADP and ATP. [3H]ApsAwas purified by a two-step HPLC procedure, and its purity was checked by thin-layer chromatography on silica gel plates in dioxane/%-propanol/20%NH40H/H20 (4020:5030). Commercial ApzA, ApsA, Ap4A, ApsA, and Ap6A were used as references. From the [3H]ADPradioactive specific activity and the [3H]ApsA radioactivity and UV absorbancy was calculated a molar extinction coefficient of 27 300 M-l cm-1 at 256 nm for [3H]ApsA,a value in good agreement with that reportedin theliterature(Hol1eret al., 1983). For anaccurate measurement of binding of [3H]ApsA to adenylate kinase, the equilibrium dialysis method was found to give more reproducible results than centrifugation-filtration on short columns of Sephadex G-50. The equilibrium binding measurements were conducted at +4 OC in a Dianorm apparatus equipped with microcells (2 X 250 pL) and Spectrapor 2 (Spectrum Medical Industries) membranes (M,cutoff 12000-14000). Onehalf-cellwas filled with 180 pL of the desalted enzyme solution. The other halfcell was filled with 180 pL of 100 mM KC1,50 mM Tris C1, and 2 mM MgClz (pH 7.5) (KTMg medium) containing [3H]ApsA. The same operation was conducted with increasing concentrations of [3H]Ap~A.Dialysis lasted for 20 h at +4 OC under continuous rotation at 12 rpm. Preliminary assays showed that dialysis equilibration was totally achieved after this period of time and that less than 5% of the enzyme activity was lost. Furthermore, it was verified that no protein escaped through the dialysis membrane. Using our adenylate kinase was found to be 32 nM and preparation, the Kd for [3H]Ap~A the amount of bound [3H]Ap~A(mol/mol of enzyme) was 0.75 f 0.05 mol (three assays). As adenylate kinase binds ApsA in a 1:l stoichiometry(Feldhaus et al., 1975; Reinstein et al., 1990), it was concluded that only 75% of our adenylate kinase preparation was functional. Therefore, the experimental binding data obtained for the various ligands tested in the course of this work were corrected appropriately. Kinetic Analysis. For kinetic analysis, a coupled colorimetric assay of the forward reaction catalyzed by adenylate kinase was used. The assay medium consisted of 2 mM MgC12, 100 mM Tris-HC1 (pH 7.5),80 mM KCl, 200 pM NADH, 400 pM PEP, 80 mM KCl, and 10 units of lactate dehydrogenase (LDH) and pyruvate kinase (PK) (Berghauser, 1975). Where not specified, AMP and ATP concentrationswere 150 and 200 pM, respectively. The final volume was 1 mL. It was checked to ensure that the coupling enzymes, PK and LDH, were devoid of contaminating adenylate kinase activity and that beryllium and fluoride ions, per se, had no effect on the regenerating system under our experimental conditions. The amount of adenylate kinase added to initiate the reaction was between 50 and 100 ng per reaction. The decrease in the absorbance at 340 nm due to the consumption of NADH was recorded using a Uvikon 930 (Kontron) spectrophotometer. The rates of ATP hydrolysis were calculated by extrapolating the slope to t = 0. To test the effect of fluoroberyllate on adenylate kinase, the enzyme was preincubated with nucleotide(s) and fluoride and beryllium ions, and then the inhibited preparation was diluted 100-fold prior to the assay of the remaining activity. Binding Measurements. The reversible binding of radiolabeled nucleotides to purified adenylatekinase was performed
Garin and Vignais as follows. A sample of the stock solution of adenylate kinase in ammonium sulfate was pelleted by centrifugation for 15 min at 15000gat 4 OC. The supernatant was discarded. The pellet was dissolved in KTMg medium and then subjected to desalting by centrifugation-filtration,using 1-mL tuberculin syringes filled with Sephadex G-50 (fine), and inserted in conical centriguetubes as described by Penefsky (1977). These small G-50 Sephadex columns were equilibrated in KTMg medium. Samples of desalted adenylate kinase adjusted to a final concentration of about 20 pM were incubated for 5 min at 25 OC with radiolabeled nucleotides in the presence or absence of NaF and BeCl2 and then subjected to centrifugation-filtration through G-50 Sephadex columns. About 60% of the protein was recovered by centrifugation at 450g for 2 min. Part of this material was used for the determinations of protein content by the method of Bradford (1976) and of bound radioactivity by scintillationcounting. For the analysis of bound fluorometal and nucleotides, another fraction of the eluate was subjected to total digestion by trypsin (1:lO (w/ w)) for 30 min at 37 OC. Trypsin digestion was preferred to protein precipitation by perchloric acid or heat treatment at 95 OC. These two methods of extraction of protein-bound nucleotides were tested in preliminary experiments. It was found that adenylate kinase was resistant to heat denaturation and that the KOH-neutralized perchloric extract, even after removal of the precipitate in the cold, often interfered with the fluorescence assay of beryllium (see below). Magnesium in the tryptic digest was analyzed by flame atomic absorption spectrometryusing a Perkin-Elmer Model 2380 spectrometer (Paschen & Fuchs, 1971), and beryllium was measured by fluorescence (Petidier et al., 1985). The nature (mono-, di-, or triphosphate) of the nucleotide bound to the inhibited adenylate kinase was determined by HPLC of the tryptic digest. It was checked that the 30-min incubation period at 37 OC used for adenylate kinase digestion by trypsin did not modify the ATP content of the samples. The ODS-2 column (15 X 0.46 cm, 3 pm; Soci6t6 Franpaisede Chromatographie, Eragny, France) used for HPLC was perfused at a flow rate of 500-pL/min with a buffer containing 50 mM KH2PO4 and 2 mM tetrabutylammonium (pH 6.0). The nucleotides were eluted by a 040% linear gradient of 60% acetonitrile in 2 mM tetrabutylammonium and 50 mM KH2P04 (pH 6.0) in the following order: AMP at 8.8 min, ADP at 12.7 min, and ATP at 16.5 min. With an amount of injected nucleotide ranging between 100 and 500 pmol, which corresponded to routine conditions, the average nucleotide recovery calculated with reference to external standards was 85%.
RESULTS The Simultaneous Presence of ADP, Mg2+,NaF, and Bee12 with Adenylate Kinase Results in Inhibition of the Enzyme. The activity of rabbit muscle adenylate kinase was inhibited by more than 90% after a 5-min preincubation with ADP, MgCl2, BeC12, and NaF present together at appropriate concentrations (Table I, row 1). Under similar conditions, Escherichia coli adenylate kinase was also inhibited, but far less effectively than rabbit muscle adenylate kinase (less than 20%; not shown). This might be attributed to the differences which exist in the amino acid sequences of the E. coli and the rabbit muscle enzymes, despite the global homology which exists between the two proteins (Schulz et al., 1986). All of the experiments described here were therefore carried out with the rabbit muscle enzyme. AlC13 was found to be as efficient as BeC12 for eliciting inhibition, but no inhibition was observed when either BeClz,
Inhibition of Adenylate Kinase by Fluorometals
Biochemistry, Vol. 32, No. 27, 1993 6823
Table I: Effect of BeClz, NaF, Mez+, and ADP on the Activity of Rabbit Muscle Adenylate Kinas@ experiment 1 2 3 4 5
NaF
+ + + + -
BeC12
+ + + -
+
Mg2+
ADP
+ + + +
+ + + +
inhibition (W) 92 f 5
