Eur. J. Biochem. 258,239-243 (1986) 0FEBS 1986

Binding of nucleotides to nucleoside diphosphate kinase monitored by rose Bengal Ioan LASCU, Elena PRESECAN and loan PROINOV Department of Biochemistry, Medical and Pharmaceutical institute, Cluj-Napoca (Received January 15,1986) - EJB 860044

The binding of nucleotides to nucleoside-diphosphate kinase from pig heart was studied using the dye rose Bengal as an optical probe. By difference absorption spectroscopy and by equilibrium dialysis it was shown that one dye molecule strongly bound per enzyme subunit. By competition with nucleotides it was shown that two nucleotide-binding sites exist on each subunit of either unphosphorylated or phosphorylated enzyme: one of them binds ATP or ADP tightly, the other one binds rose Bengal tightly and ADP loosely. As detected by different affinities for rose Bengal the enzyme exists in two conformations: a ‘relaxed’ conformation induced by the binding of ADP, and a ‘tense’ conformation induced by the binding of ATP or by phosphorylation. Nucleoside diphosphate kinase (NDP kinase) is a key enzyme in the rephosphorylation pathway of the non-adenine nucleoside diphosphates to the corresponding triphosphates. It was purified to homogeneity from a variety of procariotic and eucariotic sources [I - 51. The enzyme, regardless of the origin, is built up of six polypeptide chains having a molecular mass of about 17 kDa and has a ping-pong mechanism, with the formation during catalysis of a covalent bond between the phosphate group being transfered and a histidine side-chain [l, 61. In contrast to earlier studies, where the maximum phosphate incorporation was about 3 -4/hexamer [7, 81, recent studies demonstrated the phosphorylation of each polypeptide chain [4, 91. It seems therefore, that subunits are equivalent. In this paper we studied the binding of the adenine nucleotides to the pig heart nucleoside diphosphate kinase, by using the optical probe rose Bengal (2,4,5,7-Tetraiodo3’,4,5’,6‘-tetrachlorofluorescein).Since it tightly binds to both the unphosphorylated and phosphorylated NDP kinase, with a marked change in its spectral properties, it is a useful probe for monitoring the binding of nucleotides to the enzyme. We report the results of these studies, which led for the first time to the conclusion that NDP kinase has two nucleotide-binding sites per subunit. EXPERIMENTAL PROCEDURE Chemicals

The nucleotides were obtained from Boehringer, Mannheim, except for 8-bromoinosine5‘-diphosphate,which was synthesized according to [lo]. All nucleotides were purified by ion-exchangechromatography on DEAE-Sephadex A-25 [111. Lactate dehydrogenase and pyruvate kinase were obtained as described [12]. Rose Bengal was purified by preparative thin-

layer chromatography on silica gel plates (Merck, Darmstadt), with 2-butanol as eluent. The concentration of the dye was calculated using an absorption coefficient of 95 000 Mcm-’ at 545 nm [13]. The buffer used throughout was 20 mM sodium phosphate (pH 6.9) containing 1.0 mM EDTA. Enzyme preparations

NDP kinase was purified from pig hearts as previously described [3]. Its concentration was calculated using an A:&nm of 12.6 [9]. The molar enzyme concentration was expressed as subunit concentration assuming a molecular mass of 17 kDa [3]. The phosphorylation was performed as described [9].The enzyme bound 1.O f 0.05 mol phosphate/mol subunit. Both the unphosphorylated and the phosphorylated NDP kinase had a ratio of the absorbances at 280 nm and 260 nm equal to 1.70 f 0.02, which indicates that tightly bound nucleotides were absent. Spectrophotometric measurements

Difference absorption spectra were recorded with a Specord UV-VIS spectrophotometer (Zeiss, Jena, GDR). Difference absorbances at fixed wavelength were measured with a VSU-2D single-beam spectrophotometer (Zeiss, Jena, GDR). This instrument allows sample illumination for brief periods of time during measurements only, thus avoiding the dye-mediated photooxidation. Titrations were made with Hamilton syringes using the repetitive dispenser PB-600. The differential absorption coefficient was measured by titration of the dye with a large molar excess of NDP kinase, plotting the reciprocal of the difference absorbance versus the reciprocal of the enzyme concentration and extrapolation at infinite enzyme concentration [14]. Equilibrium dialysis

Correspondence to I. Lascu, Catedra de Biochimie, Institutul de MedicinZ $i Farmacie, Strdda Pasteur 6, R-3400 Cluj-Napoca, Roumania Abbreviation. N D P kinase, nucleoside-diphosphate kinase. Enzyme. Nucleoside-diphosphate kinase (EC 2.7.4.6).

The dialysis membrane was boiled in a solution containing 10 mM EDTA and 50 mM sodium bicarbonate. The homemade dialysis cells had unequal compartments (0.2 ml for the protein and 4.0 for the free ligand compartment). This was

240 necessary for a precise estimate of the free dye concentration. The samples were rotated for 20 h at 4°C. It was shown in control experiments that the equilibrium was 95% complete in about 3 h. The concentration of the dye was measured at 545 nm after dilution in 0.2 M NaOH for protein denaturation.

2 0.83

Activity measurements

2 0.6-

1.0 c

1

0

al

v)

The activity of NDP kinase was measured spectro- 0 d photometrically by a couple assay [lo, 151. Rose Bengal, at z 0 4 the concentrations used, was non-inhibitory for the coupling ' enzymes. No significant loss of enzyme activity was detected & c under the conditions described in this paper. All the experiments reported were performed at 25°C unless otherwise stated.

=

0.4 0.2 a2 0.4 0.6 0.8 1.0 I3

U

1.0

RESULTS Effect of rose Bengal on the catalytic activity of NDP kinase Rose Bengal was found to inhibit the NDP kinase activity. The inhibition was competitive with respect to ATP, the donor nucleotide. The Ki was 1.15 f 0.3 pM. Binding of rose Bengal to the unphosphorylated NDP kinase The difference between the spectrum of the enzyme-bound dye and the spectrum of the dye in solution had a positive maximum at 562 nm and a negative maximum at 540 nm. The presence of an isosbestic point (at 552 nm) is consistent with the existence of only one mode of binding and allows the quantitative treatment of the data. The value of A E , ~ ,50 f 2 mM- cm- at 560 nm was not affected by the phosphorylation or by the presence of 207 pM ADP or 177 pM ATP. By spectrophotometric titration it was found that one dye molecule bound per enzyme subunit (range 0.85 - 1.1 using three enzyme preparations), without apparent cooperativity (Fig. 1). The KD (dye dissociation constant) was somewhat dependent on the enzyme concentration. In a series of experiments we found KD values of 0.4 pM, 0.65 pM and 0.9 pM at the enzyme concentrations of 1.2 pM, 2.4 pM and 4.66 FM respectively, but the number of binding sites was not significantly different from one. The decrease of K D with the NDP kinase concentration decrease was probably due to enzyme dissociation, a phenomenon which is known to affect NDP kinases of various origins [3, 5, 161. In all the experiments reported the concentration of the NDP kinase was about 5 pM. Equilibrium dialysis (Fig. 1) confirmed the figures obtained in the spectrophotometric experiments. About 0.92 dye molecule bound/NDP kinase subunit, with a KD of 1.35 and 1.52 pM (two experiments). In addition, the value of Ki measured kinetically agreed fairly well with the KD. EfSect of ADP and ATP on the binding of rose Bengal to the unphosphorylated NDP kinase All the nucleotides tested decreased the affinity of the enzyme for the dye. In order to obtain valuable functional information about the enzyme it is essential to know whether or not the nucleotide and the dye share the same binding site. If the dye (D) and the nucleotide (A) bind to the same binding site of the enzyme (E), the average number ( B ) of dye molecules bound per NDP kinase subunit, assuming a one-

4.0 5.0 [Rose Bengal Ifree( p M 1

2.0

3.0

6.0

Fig. 1. Binding of rose Bengal to the unphosphorylated NDP kinase. ( 0 ) Data from spectrophotometric titration. The sample cell contained 4.5 pM NDP kinase in 20 m M phosphate buffer (pH 6.9), EDTA 1 mM, whereas the referencecell contained buffer alone (final volume 1.0 ml, 0.5-cm optical path). Successive portions of 1 p1 35 pM dye solution were added to each cell and the difference absorbance at 560 nm was measured. After a correction for dilution, B was calculated (mol dye bound/mol NDP kinase subunit), using AE,,,M, obtained as described in Experimental Procedure, and F,the concenData from equilibrium dialysis experitration of the free dye. (0) ments. The concentration of NDP kinase was 80pM. The inset contains the Scatchard plots from the same data

to-one stoichiometry and a molar excess of nucleotide over the enzyme, is

where [E,] is the total enzyme concentration and

KA is the dissociation constant of the enzyme-nucleotide complex and K"Jp is the apparent dissociation constant of the dye in the presence of nucleotide. If, however, the two ligands bind to separate, but interacting sites, from the law of the mass action, by simple substitutions, it follows: B=

[ D - E ] + [ D . E - A ][Dl [Etl [D] K"D"

+

(3)

where (4) (&A, the dissociation constant of the dye from the ternary enzyme-dye-nucleotide complex, may be larger or smaller than KD, the dissociation constant of the enzyme-dye complex). The two situations can be distinguished by plotting Ka8p, measured at different fixed concentrations of nucleotide, versus the nucleotide concentration. If Eqn (2) holds, there is

24 1

0.i IATPHmM)

0.2

1.0 20 IADPl (mM)

Fig.2. Binding of rose Bengal to the unphosphorylated N D P kinase in the presence of ATP or ADP. The binding isotherms were obtained as described in the legend to Fig. 1, in the absence and in the presence of fixed concentrations of nucleotide. The KD was calculated by nonlinear least-square computer fitting. The bar represents the standard error. (A) Effect of ATP. The curve was traced by computer according to Eqn (4) with K D = 0.85 pM, K A = 3.0 pM, and KDA = 14 pM. (B) Effect of ADP. The line was traced according to Eqn (2). With KD = 1.3 pM and KA = 310 pM.The two sets of data were obtained separately, which explains the different value for KO. The nucleotides were in large molar excess over the enzyme, therefore the concentration of the free nucleotide was approximated by the total nucleotide concentration

0.2

1 1.o

2 .o

[ADPI (mM)

a linear dependence. On the other hand, Eqn (4) predicts a curve, with the limiting value equal to K D A , at very high concentrations of nucleotide. When NDP kinase was titrated with rose Bengal in the presence of fixed concentrations of ADP or ATP, the Scatchard plots were linear and had a common intercept on the x axis. The replot of PJP versus the concentration of ATP was a curve (Fig.2A). By computer fitting according to Eqn (4) one obtained K A T p = 1 - 5 pM and K D , ATP = 820 pM. The quantitative estimate of the constants is rather poor because of the inaccuracy of the values of K"8P greater than about 5 pM. The replot of PJp versus the fixed concentration of ADP was a straight line (Fig.2B). The data were fitted according to Eqn (2) and one obtained K A D p = 300 pM. We concluded that rose Bengal and ADP compete for the same binding site, whereas ATP bound to a different site and induced an enzyme conformation with reduced affinity for the dye. Effect of A D P on the binding of ATP Fig. 3 A shows the displacement of the bound dye from its complex with the unphosphorylated NDP kinase by ATP in the presence of fixed concentrations of ADP. One notes the marked decrease in the efficiency of ATP as a competitor in the presence of ADP. When ADP concentration was varied in the presence of fixed concentrations of ATP (Fig. 3 B) there was an initial increase in the bound dye Concentration. The concentration of ADP for half-maximum effect depended on the ATP concentration: it was 60 pM and 300 pM when the concentration of ATP was 120 pM and 1200 pM respectively. Apparently ADP bound to the ATP site but was unable to induce the conformational change to the state with low affinity for the dye, as ATP did. At larger ADP concentrations the dye was displaced by competition. This experiment proves the

Fig.3. Effect of ATP and ADP on the binding of rose Bengal to the unphosphorylated NDP kinase. The sample cell contained 6.8 pM NDP kinase and 11.4 pM rose Bengal, whereas the reference cell contained the same concentration of rose Bengal in buffer (final volume 1.0 ml, 0.5 cm optical path). B represents mol rose Bengal bound/mol NDP kinase subunit. (A) Effect of ATP. In addition to enzyme and dye the cells contained ADP at the final concentration of 0 pM ( O ) , 45 pM (A) and 225 pM (0),and ATP as indicated. Since the difference absorbance decreased slowly, the points represent the final stable readings after 1- 5 min. (B) Effect of ADP. The cells contained enzyme, dye, the fixed concentration of 0 pM ( O ) ,125 pM (0),1200 pM ( W ) of ATP, and ADP as indicated

existence of two kinds of ADP-binding sites, distinguishable by the competition with rose Bengal and ATP. Binding of rose Bengal to the phosphorylated NDP kinase Effect of ADP and ATP. Rose Bengal bound to the phosphorylated NDP kinase with reduced affinity ( K D = 6 - 8 pM as compared to the value of 0.7 - 1.4 pM corresponding to the unphosphorylated enzyme) (Fig. 4). The stoichiometry of the binding remained the same. It is worth noting that the A E , ~was exactly the same for both the unphosphorylated and the phosphorylated NDP kinase. This indicates a similar environment of the bound dye in both cases. ATP increased the K D of the dye, but the high errors in measuring KD values larger than 8 - 10 pM precluded a thorough analysis. ADP, at low concentrations, increased the affinity of the dye for the phosphorylated NDP kinase, but at higher concentrations decreased it, a feature closely reminiscent of the effect of ADP on the binding of rose Bengal to the unphosphorylated NDP kinase in the presence of ATP. In Fig. 5 it is apparent that the two opposite effects occur at ADP concentrations different by two orders of magnitude. 8-Bromoinosine 5'-diphosphate,

242 a good substrate for NDP kinase despite its syn conformation [lo, 151, exerts the same activatory effect as ADP, at a concentration about four times larger. ATP greatly decreased the activatory effect of ADP, presumably by direct competition. DISCUSSION Binding of nucleotides to NDP kinase The results obtained in studying the effect of ADP and ATP on the binding of rose Bengal to NDP kinase strongly suggest that the enzyme has two nucleotide-binding sites per subunit. Table 1 summarizes the estimates of the dissociation constants. There are two alternative possibilities: either NDP kinase has a two-site ping-pong mechanism [17], such as that hypothesized for the adenosine kinase from L1210 cells [18], or one

of the nucleotide-binding sites is regulatory and the other is the catalytic site. The latter possibility was suggested by Colomb et al. [7] for beef heart NDP kinase, based on steadystate kinetic experiments. There is a discrepancy between the competitive character of the inhibition of NDP kinase activity with respect to ATP by rose Bengal and the evidence that ATP and the dye bind to separate sites. We believe that the explanation is the presence of Mg2+ions in the kinetic experiments and their absence in the binding experiments. Probably for the same reason Robinson et al. [4], using tetraiodofluorescein as an optical probe, detected competitiveness between the dye and ATP. The simultaneous presence of Mg2+ions and ATP caused the phosphorylation of NDP kinase and the formation of ADP. Further evidence that ATP and MgATP bind differently, and possibility at different sites on the enzyme, is the finding that ATP in the absence of Mg2 ions binds much more slowly than the turnover of the enzyme (unpublished experiments). +

Conformational changes The existence of conformational changes in NDP kinase has previously been noted [l, 191. Roisin and Kepes [20]

I

I

I

t

W

m

a2

0.6

Ok

a8

I

5

1D

B Fig.4. Scatchard plots of the binding data of rose Bengal to the phosphorylated NDP kinase obtained by specfrophotometric titration. For details see the legend to Fig.1. The additions were none (O), 21.8 pM ADP (a),48 pM ADP (+), 960 pM ADP (0),580 p M ATP (Q), and 580 pM ATP plus 21.8 pM ADP ( x )

10

15

20

I

i' 1000

I 2000

[ A D P 1 (pM1 Fig. 5. Effect of ADP on the bonding of rose Bengal to the phosphoryluted NDP kinase. The sample cell contained 3.6 pM phosphorylated NDP kinase and 7.5 pM rose Bengal, whereas the reference cell contained 7.5 pM rose Bengal in buffer (final volume 1.0 ml, 0.5-cm

optical path). Increasing amounts of ADP were added to both cells

Table 1. Binding constants The ligands are placed left or right of E to indicate the existence of two binding sites on each subunit Equilibrium

+

E RBeRB.E E + ATPeE.ATP E + ADPeADP.E RB.E + A D P e R B . E - A D P E.ATP + R B + R B . E - A T P E P + RB+RB-E P E P + ADP+ADP.E P RB.E P + ADPeRB.E P.ADP E - P + A T P e E P.ATP E - P.ADP + RB=RB.E P.ADP

--

--

- -

5

Dissociation constant

Data source

PM 0.85" 1-5 310b binds" 8-20 6.0" 230 10 - 25 bindsd 1.5"

Fig. 1, slope of the Scatchard plot Fig.2A, computer fitting of Eqn (4) Fig.2B, computer fitting of Eqn (2) Fig. 3 Fig.2A, computer fitting of Eqn (4) Fig.4, slope of the Scatchard plot Fig. 5, computer fitting using Eqn (4) from [4] Fig.4, fitting of Eqn (4) Fig. 4 Fig.4, computer fitting of Eqn (4)

Standard error of the estimate is & 15% of the value reported. Standard error of the estimate is f 30% of the value reported. The binding was detected, but the dissociation constant was not reliable because of the indirect way of measuring it. * The binding was detected, but a study of the competiveness between ATP and rose Bengal was not performed, because of the weak binding of the dye, which associates with large errors in the estimated KO. a

243 obtained evidence that the heat stability of NDP kinase of Escherichia coli was increased by ATP and decreased by ADP. In this study we obtained additional lines of evidence that the binding of nucleotides induced conformational changes. The values reported in Table 1 suggest that NDP kinase has two conformations (as detected by different affinities for the dye): a ‘relaxed’ conformation, in the unphosphorylated NDP kinase and induced by the binding of ADP to the ‘ATP site’; and a ‘tense’ conformation, induced by the binding of ATP and by phosphorylation. A common conformation in the two later cases is to be expected if hydrogen bonds and salt bridges are formed between the phosphate(s) group(s) and amino acid side-chains in the active site. Recent studies showed that several enzymes have the nucleotide-binding site made of amino acids from neighbouring subunits [21, 221. This could be the case for NDP kinase as well, since the subunits are small, and much of their surface be in contact with other subunits. Possibly the binding of ATP in the intersubunit area exerts an ‘allosteric-like’ effect, although the conformation of the polypeptide chains remains unchanged. We thank Dr. loan Petrescu for generously donating rose Bengal, Dr. Horea Demian for help in purifying the dye, Dr. Horea Porumb for useful discussions and Eppendorf Gerltebau (Hamburg) for making available to us the photometer used in the kinetic studies.

REFERENCES 1. Parks, R. E. & Aganval, R. P. (1973) in The enzymes (Boyer, P. D., ed.) vol. 8 A, pp. 307- 333, Academic Press, New York. 2. Colomb, M. G., Chkruy, A. & Vignais, P. V. (1972) Biochemistry 11, 3370-3378. 3. Lascu, I., Duc, M. & Cristea, A. (1981) Anal. Biochem. 113,207211.

4. Robinson, J. B., Brems, D. N. & Stellwagen, E. (1981) J. Biol. Chem. 256,10169-10773. 5. Islam, K. & Burns, R. G. (1984) Anal. Biochem. 137,8-14. 6. Colomb, M. G., ChBruy, A. & Vignais, P. V. (1972) Biochemistry 11,3378-3386. 7. Colomb, M. G., ChBruy, A. & Vignais, P. V. (1974) Biochemistry 13,2269 - 2277. 8. Garces, E. & Cleland, W. W. (1969) Biochemistry 8, 633-640. 9. Lascu, I., Pop, R. D., Porumb, H., Presecan, E. & Proinov, I. (1983) Eur. J. Biochem. 135,497 - 503. 10. Kezdi, M., Kiss, L., Bojan, O., Pavel, T. & BPrzu, 0.(1976) Anal. Biochem. 70,361 -364. 11. Petrescu, I., Lascu, I., Goia, I., Markert, M., Schmidt, F. H., Deaciuc, I., Kezdi, M. & Bbrzu, 0. (1982) Biochemistry 21. 886 - 893. 12. Lascu, I., Porumb, H., Porumb, T., Abrudan, I., Tgrmure, C., Petrescu, I., Presecan, E., Proinov, I. & Telia, M. (1984) J . Chromatogr. 283, 199 - 210. 13. Wu, C.-W. & Wu, F. Y.-H. (1973) Biochemistry 12,4349-4355. 14. Dorgan, L. J. & Schuster, S. M. (1981) Arch. Biochem. Biophys. 207, 165-174. 15. Lascu, I., Kezdi, M., Goia, I., Jebeleanu, G., Btrzu, O., Pansini, A., Papa, S. & Mantsch, H. H. (1979) Biochemistry 18,48184826. 16. Palmieri. R.. Yue, R. H., Jacobs, H. K., Maland, L. M., Wu, L. & Kuby, S. A. (1973) J. Biol. Chem. 248.4486-4499. 17. Cleland, W. W. (1973) J . Biol. Chem. 248, 8353-8355. 18. Chang, C.-H., Cha, S., Brockman, R. W. & Bennett, L. L. (1983) Biochemistry 22,600 -611. 19. Katz, M. & Westley, J. (1980) Arch. Biochem. Biophys. 204,464470. 20. Roisin, M. P. & Kepes, A. (1978) Biochim. Biophys. Acta 526, 418 -428. 21. Muirhead, H. (1983) Trends Biochem. Sci. 8 , 326 - 329. 22. Kantrowitz, E. R., Pastra-Landis, S. C. & Lipscomb, W. N. (1980) Trends Biochem. Sci. 5 , 150- 153.