Multinuclear NMR study of the interaction of vanadate with mononucleotides, ADP, and ATP

Multinuclear NMR Study of the Interaction of Vanadate with Mononucleotides, ADP, and ATP Carlos F. G. C. Geraldes and M. Margarida C. A. Castro Chemistry Department,

University of Coimbra, Coimbra, Portugal

ABSTRACT The interaction of vanadate with 5 ‘-mononucleotides, ADP, ATP, and various molecules containing some of their chemical moieties was studied in aqueous solution in the pH region of 5-9 using proton, 13C, 31P, and Q V nuclear magnetic resonance (NMR) spectroscopy. All the compounds studied formed noncyclic vanadate esters through interaction of monovanadate or divanadate with the hydroxyl groups of the ribose ring. Noncyclic anhydrides were also formed with the phosphate groups of ribose S-phosphate, the mononucleotides, ADP, ATP, phosphate, pyrophosphate, and tripolyphosphate. In particular, ADP and ATP analogs resulted from AMP (AMPV and AMPVr) and from ADP (ADPV) . Cyclic esters of trigonal bipyramidal geometry resulted from the interaction of vanadate with two ribose ring cis hydroxyl groups. AMP, CMP, and UMP formed two such complexes of 1: 1 and 1:2 stoichiometries, similar to what has been observed for uridine and other nucleosides. However, 2’-deoxy- AMP does not yield this type of complexes. ADP and ATP also form similar cyclic ester complexes with vanadate, which does not chelate their pyrophosphate and tripolyphosphate moieties. Nevertheless, the separate phyrophosphate (PP) and tripolyphosphate (PPP) ligands form cyclic anhydrides of octahedral geometry with vanadate. However, their binding to vanadate is weaker than that of the ribose ring of nucleotides. Competition experiments between ethylene glycol and phosphate (P), pyrophosphate (PP), or tripolyphosphate (PPP) show that the relative strength of the interaction of these ligands with vanadate is PP > ethylene glycol > PPP > P.

INTRODUCTION Vanadium is an essential element of widely recognized biological importance [l-5]. For example, it is part of the catalytical center of certain nitrogenases and halide peroxydases [6,7], and, acting as a phosphate analogue, it has a large influence on the activity of many enzymes [g-lo]. Many cases of this later effect are known, such as Address reprint requests to: Dr. C. F. G. C. Geraldes, Department of Chemistry, University of Coimbra, 3000 Coimbra, Portugal. Journal of Inorganic Biochemistry 37, 213-232 (1989) 0 1989 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas,

213 NY, NY 10010

0162-0134/89/$3.50

214

C. F. G. C. Geraldes and A4. h4. C. A. Castro

3

A

2

B

FIGURE 1. A) Diagrams showing the (Y and /3 anomers of ribose Sphosphate (R = CH20P032-, which exist in the furanose form; B) The general structure of 5 ‘-mononucleotides, where B is the base adenine, cytosine or uracil, and R is CH20P032- for monophosphates, CH2P03P04 ‘- for ADP and CH2P03P03P04*- for ATP.

the inhibition of ribonuclease [l l] and the modulation of the activity of many phosphatases and ATPases [lo, 12-151. In order to elucidate the biological function of vanadium, many studies have been carried out involving the interaction of vanadate with small biologically relevant molecules in aqueous solution [ 16-241. 5’ V nuclear magnetic resonance (NMR) spectroscopy is a particularly useful technique in this type of study, providing detailed information on the number of complexes formed, their structure, and metal coordination environment [25]. This type of information is particularly useful in the interpretation of the enzyme regulatory properties of vanadate, which depend on the vanadate esteritication of phosphate or hydroxyl groups of small biological molecules forming analogues of phosphorylated enzyme substrates, cofactors, or transition states

WI. In this work, we report an extensive multinuclear NMR study, using observation of 51V, *H, i3C, and 31P nuclei, of the interaction, in aqueous solution, of vanadate with several 5 ’ -mononucleotides, adenosine diphosphate (ADP), and adenosine triphosphate (ATP) (Fig. l), as simple models to understand the role of vanadate as an inhibitor of ATPases. To help a detailed interpretation of the results obtained, the interaction of ribose Sphosphate, phosphate (P), pyrophosphate (PP) [27], tripolyphosphate (PPP), and 2’-deoxy-adenosine 5 ‘-monophosphate (d-AMP) was also analyzed. These results were compared with previous investigations by 51V NMR of the interaction of monosaccharides, uridine, and adenosine 5 ‘-monophosphate (AMP) with vanadate [23, 241, as well as a previous multinuclear NMR investigation of the binding of vanadate to various monosaccharides and nucleosides [26].

EXPERIMENTAL The mononucleotides adenosine 5 ’ -phosphate (5 ’ -AMP), cytidine 5 ’ -phosphate (5 ’ CMP) , uridine 5 ’ -phosphate (5 ’ -UMP) , and 2 ’ -deoxy-adenosine 5 ’ -phosphate (dAMP), as well as ribose 5-phosphate, adenosine diphosphate (ADP), and adenosine triphosphate (ATP), were obtained from Sigma Chemical Co. Analytical grade ammonium vanadate and the salts sodium phosphate, sodium pyrophosphate, and sodium tripolyphosphate were purchased from Merck. DzO, DCl, and NaOD were obtained from Stohler Isotope Chemicals. These biological substances and inorganic salts were, respectively, lyophilized from DzO and dried at adequate temperatures in order to minimize their HZ0 content and dissolved in D20 for proton NMR measurements, or in HZ0 enriched with 10% D20

VANADATE

WITH NUCLEOTIDES

215

for heteronuclear measurements. The concentrations of the stock solutions containing vanadate and the ligands were established by weight and were mixed in appropriate volumes. The pH was adjusted cautiously (avoiding formation of the orange-colored decavanadate [26]) by addition of dilute solutions of DC1 and NaOD. The pH values quoted are direct pH meter readings obtained on a Metrohm E520 apparatus (room temperature) after standardization with aqueous (H*O) buffers. Proton, i3C 31P, and 51V NMR spectra were obtained on a Varian XL-200 and a General Electhc GE-500 NMR spectrometer at probe temperature (2 1 f 1 “C). 13C and 3’P spectra were broadband proton decoupled. Some proton spectra required the solvent HOD peak suppression using a presaturation pulse sequence. TSS and tbutanol were used as internal references for ‘H and i3C nuclei, but those shifts were recalculated relative to TMS. Phosphoric acid and trimethylphosphate were used, respectively, as an external and internal reference for 31P shifts, whereas VOC13 was the external reference for 5’V shifts.

RESULTS

AND DISCUSSION

Because of the complexity of the chemistry of vanadium (V) in aqueous solution including polymerization and protonation of vanadate species [ 11, and because of its coordination properties [ 16-241, the complexation of vanadate may depend on various factors, such as pH, metal-to-ligand ratio and total metal and ligand concentration. These factors were studied by comparison in similar conditions, of ‘H, 13C, and 31P spectra of ligand solutions or “V spectra of metal ion solution with those of mixtures of metal and ligand. Slow-exchange conditions, observed in most cases, allowed information about the number of complexes present, their relative concentrations, stoichiometries, and stability to be obtained from signal areas, whereas chemical shifts provided, in some cases, information about their structure [26].

Mononucleotides The potentially complex coordination behavior of the 5 ‘-mononucleotides as ambidentate ligands [28] towards vanadate required an investigation of its interaction with each of their structural units and their possible combinations. The interaction with ribose [26] and other sugars [23,26], nucleosides [24,26], and phosphate [27,29] has already been studied. Therefore, we concentrated on the bases, ribose 5-phosphate, and finally the 5 ‘-mononucleotides themselves. No complexation was found between vanadate and the bases adenine, cytosine, and uracil in aqueous solution, irrespective of the pH and metal-to-ligand ratio studied. It is known that phosphate interacts weakly with vanadate at pH near neutrality, forming a mixed anhydride (PV), analogous to pyrophosphate [27]. At acid pH and high V:P ratios, a 1: 14 vanadophosphate heteropolyanion, HnPV14042(9-n)- has been detected [29], but these results are not relevant for the present study. Figure 2 shows representative proton, 31P, and 51V NMR spectra of solutions containing vanadate and ribose 5-phosphate (RP) in 1: 1 mole ratio and at pH 5.6-7.2. The 5’V spectrum (Fig. 2C) displays the usual resonances between - 550 and - 580 ppm assigned to the vanadate monomer (Ti), dimer (T2), tetramer (T4), and pentamer (T5). The signals from the monomer (Ti) and dimer (T2) are considerably broadened, reflecting formation of unidentate monoesters or anhydrides of vanadate (RPV) and (RPVJ with tetrahedral geometry about the V(V) atom. These species are formed by

216

C. F. G. C. Geraldes and M. M. C. A. Castro

4a

A

w 5a

30 1

50 MP

2a

1.!

B

a3

Pf

CC

~

4.0

6 (pm)

I

I

+4

+2

I

I

0

-2

14

Wppm)

T5 L I

-500

I

I

I

I

- 600

- 550 6

(ppmhs

VOClj

FIGURE 2. NMR spectra of vanadatekibose 5phosphate, 50 mM/50 mM: a) 200 MHz proton spectrum, pH = 7.2; B) 81 MHz 31P spectrum, pH = 5.6; C) 52.6 MHz ‘IV spectrum, pH = 7.2. direct interaction of the vanadate species with each of the hydroxyl groups of the ribose ring and the phosphate group [16, 17, 19,271. A rather broad assymetric signal C(t) at - 521 ppm, containing at least two unresolved components, is observed and assigned to pentacoordinate products of possible trigonal bipyramidal coordination geometry. Similar 51V signals have been observed for vanadate in presence of ethylene

218

C. F. G. C. Geraldes

and M. M. C. A. Castro

I

I

I

I

- 600

- 550

-500

6 (ppm)vsVOCIJ FIGURE 3. 52.6 MHz 51V NMR spectra of aqueous solutions of A) vanadateld-AMP, 5OmM15OmMat pH = 7.9; B) vanadate/AMP, 100 mMllO0 mM at pH = 7.2. “1

Ti+l

P BP1

[BP~l=Kl[Til[ll

(2)

WV= KzW’I 1

(3)

K2

2BP, * BP2 K3

2Ti+21

* BP*

K3 = K1 2K2

(4)

Kq

BP;

P BP;

[BP; I = &[BP;

I

(5)

where & = 9.7 x lo9 Mm3. BP1 and BP2 are trigonal bipyramidal 1:l and 2:2 complexes with equilibrium constants Kt and K2. The overall stability constant for BP2 is Ks = Kt2K2, whereas the species BP; and BP; are two geometric isomers of the 2:2 complex, [BP21 = [BP;] + [BP;‘]. Based on a plot of ([BP, + 2 [BP2])/ (KC- 1’4[T41“4[1]) versus fr,] ‘“[l], where ([BP, + 2 @3P2])was given by the total area of the C(t) signal (see Fig. 4D), a reasonable least squares linear fit of the experimental

220

C. F. G. C. Geraldes and M. A4. C. A. Castro

data gave Ki = 0 and K3 = K12 K2 = 3.4 x lo6 Mm3. Therefore, based on this linear fit of the concentration dependence of the C(t) signal area, these authors [24] concluded that the 1: 1 complexes are not significantly formed and that C1 and C2 are two geometrical isomers of 2:2 stoichiometries with overall formation constants almost five orders of magnitude larger than the analogous value obtained for ethylene glycol [27]. An even larger value of K3 = 2.8 x 10’ Me3 has been obtained for the complexation of vanadate by uridine [23]. Alternatively, in the case of the vanadate/uridine system, we have presented evidence [26] that the C, and C2 complexes are 1: 1 and 1:2 complexes (see structures 14 and 15, Fig. 8C). We then considered the equilibria: Ti+l

Ki * BP,

[BPII=K;

[Til[ll

(6)

W21=

W11[11

(7)

Ki BP,+1

ti BP2

K;

where BP1 and BP2 are now the 1: 1 and 1:2 complexes C, and C2 with stability constants K,’ and K; . A linear plot of ([BP,] + [BP2])/([Ti][l]) versus [l], where ([BP,] + [BP2]) was obtained from the total area of the C(t) signal, gave K,’ = 130 M- 1 and K; = 123 M-l [26]. These binding constants for chelate formation are in better agreement with those reported for ethylene glycol than those obtained using the first chemical model [23]. We also obtained Job plots for the vanadate/uridine complexes, based on their H6 ligand signal intensities, which gave maxima at 1: 1 and 1:2 metal-to-ligand ratios, therefore supporting the stoichiometries considered in our chemical model. In an attempt to get further independent evidence for one or the other chemical model of the C, and C2 complexes of the vanadate/UMP system, we compared the relative areas of their proton (Fig. 4A) and 51V (Fig. 4D) signals. Using either the Hs or the H,’ resonances of the ligand, their relative areas gave populations P, = 0.34 and P2 = 0.66, assuming 1:l and 1:2 stoichiometries, and P, = 0.20, P2 = 0.80, assuming two 2:2 isomers. We fitted the composite C(t) 51V signal (Fig. 4D) to the sum of two lorentzian lines (Fig. 5) using a computer program and, from their relative areas, obtained populations PI = 0.30 and P2 = 0.70, therefore supporting the view that Ci and C2 are, respectively, 1:l and 1:2 complexes. The line widths of these two signals are 830 Hz and 1150 Hz, respectively, again indicating that C2 is the larger complex, causing more efficient quadrupolar relaxation of its 51V nucleus. The complexation shifts induced by vanadate on the proton, 13C and 31P signals of the 5 ‘-mononucleotides are shown in Table 1. The base 13C shifts are quite small, but the ribose shifts are significantly larger, with positive values in the relative order C; - c; * c; > c,l > c;, indicating binding of vanadate to the C; and C; hydroxyl groups of the ribofuranose ring [26]. The proton complexation shifts show similar trends, but the two complexes C, and C2 have quite .lifferent Hh(Hs) and H ,’ shifts, indicative of different structures. These shift differences between C, and C2 are larger in the 5 ‘-mononucleotides than for the nucleosides [26]. ADP and ATP The complexation of vanadate by ADP and ATP should, in principle, be somewhat similar to that by the 5 ‘-mononucleotides AMP, CMP, and UMP, resulting from the direct interaction of vanadate with the C; and C; hydroxyl groups of the ribose ring

VANADATE

WITH NUCLEOTIDES

-5io FIGURE 5. lines.

221

-550 PPM Computer tit of the “V C(t) signal for vanadate/UMP as a sum of

two

lorentziar

and with the phosphate groups. Therefore, we started by investigating the interaction of vanadate with pyrophosphate (PP) and tripolyphosphate (PPP) by 51V and 31P NMR. The 51V spectra of vanadate/PP (Fig. 6A, B) and of vanadate/PPP (Fig. 6C-E) show broadening of the Ti and Tz resonances reflecting the formation of linear monoand dianhydrides with pyrophosphate [27] and tripolyphosphate (Vl, Viz and Vzl, 1= PP or PPP) of tetrahedral geometry about the V(V) atom, which cause only a slight broadening of the 31P signals of these ligands (Fig. 7). Another broad 5’V resonance, C(t), appears at - 540 ppm, for some conditions of pH and concentration (Fig. 6AC), and is assigned to octahedral cyclic dianhydrides of PP [27] or PPP. A decrease of pH causes increased broadening of Ti and favors formation .of octahedral versus tetrahedral complexes [24]. Formation of octahedral complexes is also favored by increase of the ligand-to-metal ratio (see Fig. 6). The octahedral complexes for the vanadate/PP system give more than one 3’P signal (Fig. 7A) reflecting formation of ,more than one isomer of the 1:2 stoichiometry (see Fig. 8A, structures 2 and 3). In the case of vanadate/PPP, the octahedral complexes yield 31P signals, which show identical shifts for the P, and PY atoms of PPP (Fig. 7B), reflecting tridentate binding of PPP to vanadate. If octahedral 1: 1, 1:2, or 2:2 complexes are formed, various geometrical isomers are possible, (Fig. 8B, structures 7-13), which may account for the somewhat broader 51V resonance at - 540 ppm for V/PPP versus V/PP. The octahedral complexes for V/PPP are formed at somewhat lower pH that for V/PP, in the same ligand-to-metal ratio conditions (Fig. 6). At pH 5.8, vanadate/PPP gives rise to another broad “V signal at - 530 ppm, possibly reflecting formation of trigonal bipyramidal complexes. A binuclear M& complex with tridentate ligand binding has been previously detected for Tris and Tris-eth complexes [3 11. As a model to investigate the competition between the two types of possible vanadate coordination sites in AMP, ADP, and ATP, i.e., the C; and C; hydroxyl groups of the ribose moiety and the phosphate groups, the effect of adding increasing amounts of phosphate (P), pyrophosphate (PP), or tripolyphosphate (PPP) on the 51V spectrum of vanadatejethylene glycol mixtures was investigated. Figure 9A shows the

H6

2.65 12.06 12.65 1.76 1.47; - 0.65;

0.10’ 0.00 0.00 0.55 0.29

0.14;

-0.27; 0.00

-

1.03 - 0.80

0.50 C C

-

- 0.06

y Positive shifts are to high frequency. ’ Not observed. c Specific assignments not carried out.

PC. PB P,

HS HE H2 H,, Hz, H31 Hd, Hs1.n G G G G G G, C21 Cl, C4/ CS,

V(V)/UMP

0.12

0.10;

-

b b b b b b

b b b b

c C

0.45

-

- 0.02

V(V)/CMP

0.10

-

b b b b b b b

b

C e

0.42

+0.35 -0.50

0.08;

-

b b b b b b

b h b b b b

0.42 c C C C b

- 0.65 - 0.50 - 0.60

0.16;

b

0.43 C C C c

c C

c c

0.00

-

-0.27; 0.00 0.21; 0.60

-

V(V)/ADP

V(V)/AMP V(V)/ATP

Shifts. A 6 (ppm), u Resulting from the Interaction of Various Nucleotides

- 0.28; 0.00

1. Proton, 13C, and 3’P Complexation Solution (pH = 6.8) _

Nucleus

TABLE

- 0.44 -

b b b b b

C e C c C

-

-

V(V)/rib-SP

with Vanadate in Aqueous

I

I

-525

(ppm)vsVOCl3

6

I

-57s

-550

1

4

*5

1

-525

C

I

-550

c(t)

I

6

Ti

1

I

(ppmhsVOCl3

-575

1

-600

FICXJRE 6. 52.6 MHz ” V NMR spectra of aqueous solutions containing: A) vanadate/PP, SOnM/2OOmM, pH = 7.6; B) vanadate/PP, 1 mM/ 2 mM, pH = 7.6; C) vanadate/PPP, 50 mM/200 mM, pH = 7.6; D) vanadate/PPP, 1 mM/6 mM, pH = 7.1; E) vanadate/PPP, 1 mM/6 mM, pH = 5.18.

-500

c(t)

T4

224

C. F. G. C. Geraides

-10

and hf. M. C. A. Castro

-15

-20

-25 5 bpm)

FIGURE 7. 81 .O MHz 31P NMR spectra of aqueous solutions of 50 mM vanadate, pH = 7.6, in the presence of A) 200 mM PP; B) 200 mM PPP (A ’ is an expansion of A).

51V spectrum of 1 mM vanadate in the presence of 3M ethylene glycol, which consists of the signals of the tetrahedral Tl and Tlz complexes (at - 557 ppm and - 553 ppm, respectively) and of the 2:2 trigonal biyramidal complex BP2, at - 523 ppm [19]. Addition of 0.1 M phosphate causes the disappearence of the Tl and Tlz signals, but not of BP*, which disappears only at a larger phosphate concentration of 1 M. It is well known [ 171 that phosphate catalyzes the hydrolysis of tetrahedral vanadate complexes, Tl and Tlz formed by hydroxyl groups. The signal C(t) (at - 561 ppm or at - 579 ppm, Fig. 9B, C) results from coalescence of the 51V NMR signal from Tl and Tla with that of the tetrahedral VP species. Addition of 0.1 M PP to vanadate/ethylene glycol (Fig. 9D) completely destroys the

VANADATE WITH NUCLEOTIDES

a

225

226

C. F. G. C. Geraldes

I

-500

and M. M. C. A. Castro

I

-525

I

-550

I

-575 6 (ppm)

vs VOCI3

FIGURE 9. 52.6 MHz *‘V NMR spectra of aqueous solutions of vanadatejethylene glycol, 1 n&l/ 3 M, pH = 7.6; A) with no added ligand; B) with 0.1 M phosphate; C) with 1 M phosphate; D) with 0.1 m PP; E) with 0.1 M PPP. Tl, Tlz and BP2 complexes.

The broad resonance C (at - 540 ppm) corresponds to formation of octahedral cyclic dianhydrides of PP. A similar addition of 0.1 M PPP (Fig. 9E) does not affect the BP2 complex but causes coalescence of the signals from Tl and Tlz into a broad resonance C(t) (at - 551 ppm), which possibly results from formation of octahedral vanadate complexes with mixed ligands PPP and ethylene glycol. These competition experiments clearly illustrate the relative stability of the vanadate chelates, which is PP > ethylene glycol + PPP. The decreased stability of the PPP chelates could result from steric repulsion of the coordinated groups. All these chelate binding modes are more stable than the monodentate binding of phosphate and ethylene glycol, which is P > ethylene glycol, as could be inferred from published stability constants [19, 271. However, this model system may not be completely accurate for vanadate binding to AMP, ADP, and ATP, as it has been proposed that

VANADATE

WITH NUCLEOTIDES

227

the hydroxyl groups of uridine of AMP [24], with their favorable spatial orientation, form much more stable chelates with vanadate than with ethylene glycol. Fig. 10 shows the 5’V NMR spectra of vanadate/ADP and vanadate/ATP at pH = 6.8 and pH = 9.4. Their relevant features are a broadening of the Tr resonance and a new resonance T; at - 575 to - 578 ppm, corresponding, respectively, to formation of tetrahedral ADPV (or ATPV) and ADPVz (or ATPV2) complexes. These weak interactions cause broadening of the 31P resonances of free ADP and ATP (see Fig. 1 l), because of intermediate exchange conditions. Therefore, the tetrahedral species result not only from esterification of hydroxyl groups of ADP and ATP but also from formation of anhydrides with their terminal phosphate groups. A broad composite signal, C(t), is also observed in the - 530 ppm region of the 5’V spectrum, corresponding to, at least, two components Cr and C2 at pH 6.8. The broad C2 component predominates at pH 9.4. This composite signal C(t) corresponds to trigonal bipyramidal complexes resulting from binding of vanadate to the C2 I and C3’ hydroxyl groups of the ribose ring of ADP and ATP. These stronger chelate trigonal bipyramidal complexes yield 31P resonances in slow exchange with those of the free nucleotides and are generally shifted to low frequency relative to free ADP and ATP (Fig. 11). An exception is the Pa signal of the ADP complexes, which is shifted to high frequency. The magnitudes of these shift are much smaller, and their signal is generally opposite to those observed for their molybdate chelates [32]. This indicates that, contrary to Mo(Vl), which chelates directly the ADP and ATP phosphate groups, V(V) does not form chelates with these moieties. The small 31P complexation shifts are due to weak electric field and geometry effects in the complexes containing vanadate bound to the ribose rings. The proton spectra of vanadate/ADP (Fig. 12A) and vanadate/ATP (Fig. 12B) indicate the presence of at least four major trigonal bypyramidal complexes, Ci (i = 1, - - - 4). Their base proton resonances are not well resolved and are difficult to assign. However, based on the relative areas of the H ,’ resonances, it can be concluded that in the case of V(V)/ADP, like in V(V)/UMP, the populations of the polymeric C3 and C4 complexes are much smaller than those of C, and CZ. However, in the case of V(V)/ ATP, the polymeric complexes have comparable populations to those of C1 and C2. Therefore, the analysis of the 5’V C(t) signal in terms of two lorentzian components, using the procedure described above for vanadate/UMP, is still acceptable for V(V)/ ADP but not for V(V)/ATP. The composite “V C(t) signal obtained for V(V)/ADP was decomposed into two lorentzians, with relative areas that gave populations Pi = 0.38 and P2 = 0.62 for the Ci and CZ complexes, with line widths, respectively, of 680 Hz and 1660 Hz. We did not attempt to obtain the stoichiometries of these complexes using the procedure described for V(V)/UMP, as the extensive overlap of the proton resonances did not allow a correct comparison of proton and 51V signal areas. However, the presence of a sharper (Cl) and a broader (C,) component, like in the case of V(V)/UMP, leads us to propose that Ci and CZ are 1: 1 and 1:2 complexes, and C3 and C4 are polymeric (e.g., 2:2 or higher) complexes. the 5*V C(t) signal of V(V)/ATP (Fig. 1OC) also contains sharper and broader components components, and, although no quantitative conclusions are possible, a similar complexation picture is also proposed for this system. CONCLUSIONS Vanadate reacts spontaneously with the hydroxyl groups of 5 ‘-mononucleotides, ADP, and ATP to form noncyclic vanadate esters. It also forms cylic esters through chelation of the cis Cz, and Cs, hydroxyl groups of the ribose ring. It has been

8

7

A

,

- 520

,,,,

a

Cl

-540

-540

C

-5;o

c2

-560

T2

I

Ti

TL

-580

PPM

‘5

PPM

D

-520

c0) Cl

-540

-540

J

- 580

A) V(V)/ADP,

pH = 6.8; B) V(V)IADP,

72

Ti

i

2

-580

-580

1

pH = 9.4; C) V(VYATP,

-560

-560

J

:,..:y,)-

L

-560

Ti

Ti

u,I’:*

Tt5

FIGURE 10. 131.5 MHz 5’V NMR spectra of vanadatehucleotides 6.8; D) V(V)/ATP, pH = 9.4.

,‘_slo

,

h

n

PH =

PPM

J

5

VANADATE

WITH NUCLEOTIDES

229

B

A

1

0

I

I

-4

I -8

I

I

I

-12

-16

-20

I

PPM

-24

FIGURE 11. 202.5 MHz 31P NMR spectra of A) V(V)/ADP, 50 mM/SO mM, pH = 6.8; B) V(V)/ATP, 50 mM/50 mM, pH = 6.8.

proposed [23, 241 that these cyclic esters are binuclear complexes containing two ligands, which can yield stereoisomers resulting from different relative orientations of the two nucleotide molecules, each coordinated to a pentacoord+te V(V) ion in the dimeric species. Following previous studies with nucleosides [26], we now presented further evidence that, though oligomeric or polymeric species are formed, the two major cyclic esters of vanadate formed by 5’-mononucleotides, ADP and ATP are mononuclear trigonal bipyramidal complexes of 1: 1 and 1:2 stoichiometries. Vanadate and divanadate also form linear anhydrides with phosphate, pyrophosphate, tripolyphosphate, 5 ‘-mononucleotides, ADP, and ATP. Therefore, vanadatecontaining analogues of ADP and ATP can be formed in various ways: AMPV is an analogue of ADP, whereas AMP!!* and ADPV are analogues of ATP. Pyrophosphate and tripolyphosphate also form cyclic mono- and dianhydrides of octahedral geometry. However, ADP and ATP do not form these types of cyclic anhydrides, becuase of preferential chelation of vanadate by their ribose ring. The different modes of binding of vanadate to sugars, nucleosides, and nucleotides, illustrated in this study as well as in previous work [23, 24, 261, are the basis for the interpretation of the regulatory action of vanadate on the activity of many enzymes, by acting as a phosphate analogue. Formation of stable cyclic esters with vicinal hydroxyl groups of ribose rings, yielding complexes of trigonal bipyramidal geometry that

230

C. F. G. C. Geraldes and A4. M. C. A. Castro

li

8.6

0.4

0.2

6.5

6.0

6.0

5.5

5.0 PPM

21

I

B

8b

-..--J,

/

F

4

1

0 FIGURE

12.

500 MHz

7

’

I

I

6

proton NMR spectra of A) V(V)/ADP, V(V)/ATP, 50 mM/50 mM, pH = 6.8.

5

50

PPM

mML50 mM, pH = 6.8; B)

constitute transition state analogues, is the basis for the inhibition of enzymes such as ribonuclease [ll]. Spontaneous formation of vanadate esters by single hydroxyl groups of sugars (e.g., forming glucose 6-vanadate) leads to activation of enzymes that accept those esters as analogues of phosphorylated substrates, e.g., glucuse 6phosphate dehydrogenase [33]. Formation of vanadate esters by hydroxyl groups of enzyme residues may lead to their inhibition through blocking the formation of phophorylated enzyme intermediates, such as in the cases of acid phophatases [ 1, 34, 351 or the sodium-potassium pump [36]. The other mode of binding of vanadate, illustrated in this work for 5 ‘mononucleotides, ADP, and ATP, is the rapid, nonenzymatic formation of simple mixed anhydrides of vanadate with the nucle%ide terminal phosphate groups. These linear anhydrides, although not the major compounds formed by vanadate with

VANADATE

WITH NUCLEOTIDES

231

nucleotides, are biologically the most important species, as they can act as ADP or ATP analogues. AMPV, acting as an analogue of the substrate ADP, activates pyruvate kinase [27], but ADPV, an analogue of ATP, competitively inhibits myosin ATPase [ 121. Arsenate also spontaneously forms sugar or nucleoside esters, such as glucose 6arsenate [36] or adenosine 5 ‘-arsenate [37]. However, these compounds are kinetically and thermodynamically less stable than their vanadate analogs [36]. Therefore, arsenate, like vanadate, controls the activity of various enzymes, as illustrated by the observed arsenate activation of glucose 6-phosphate dehydrogenase [36] or of myokinase [37]. Because of ther extremely slow nonenzymatic formation [37], phosphorylated compounds present in metabolic pathways are formed by phosphate group transfer reactions under enzymatic control. The biological effects of vanadate, arsenate, and possibly of other oxyanions [ 1, 341 as phosphate analogues derive from spontaneous formation of analogues of these phosphorylated compounds. The lack of enzymatic control of their formation, although a useful characteristic in some cases, might be one of the reasons why those anions have no general natural role in metabolic processes. We thank Prof. A. Dean Sherry, from the University of Texas at Dallas, U.S.A., for the use of a General Electric GE-500 NMR Spectrometer and Dr. D. Buster for the help in running the 500 MHz spectra. We also thank Prof. Ferrer Correia, from the University of Aveiro, Portugal, for help in setting up the curve-fitting computer program. This is a contribution from the Centro de Investigaqio em Q&mica, supported by INIC of the Portuguese Ministry of Education.

REFERENCES 1. a) N. D. Chasteen, Struct. Bonding (Berlin) 53, 105 (1983); b) K. Kustin, G. C. Mcleod, T. A. Gilbert, and L. R. Briggs, 4th, Struct. Bonding (Berlin,’ 53, 239 (1983); c) D. W. Boyd and K. Kustin, Adv. Znorg. Biochem. 6, 311 (1984). 2. a) N. D. Chasteen, J. K. Grady, and C. E. Holloway, Inorg. Chem. 25, 2754 (1986); A. Butler and N. J. Danzitz, J. Am. Chem. Sot. 109, 1864 (1987). 3. P. Krauss, E. Bayer, and Z. Kneifel, 2. Naturforsch, B: Anorg. Chem., Org. Chem. 39B, 829 (1984). 4. M. Assi Nawi and T. L. Riechel, Znorg. Chim. Acta. 136, 33 (1987). 5. a) P. F. Frank, R. M. K. Carlson, and K. 0. Hodgson, Znorg. Chem 25, 470 (1986); b) S. G. Brand, C. J. Hawkins, and D. L. Perry, Znorg. Chem. 26, 627 (1987). 6. a) B. J. Hales, E. E. Case, J. E. Morningstar, M. F. Dzeda, and L. A. Mauterer, Biochemistry 25, 725 1 (1986); b) J. M. Arber, B. R. Dobson, R. R. Eady, P. Stevens, S. S. Hasnain, C. D. Garner, and B. E. Smith, Nature (London) 325, 372 (1987). a) H. Vilter, Bot. Mar. 26, 451 (1983); b) E. de Boer, Y. van Kooyk, M. G. M. Tromp, H. Flat, and R. Wever, Biochem. Biophys. Acta 48, 869 (1986); c) E. de Boer, M. G. M. Tromp, H. Plat, G. E. Krenn, and R. Wever, Biochem. Biophys. Acta 872, 104 (1986); d) H. Vilter and D. Rehder, Znorg. Chim. Acta 136, L7

(1987). a) A. Green, Biochem. .Z. 238, 663 (1986); b) T. Ramasanna

and F. L. Crane, Curr.

Top. Cell Regul. 20, 247 (1981). a) C. E. Heylinger, A. G. Tahiliani, and J. H. McNeill, Science (Washington, D.C.) 227, (1985); b) J. Meyerovitch, Z. Farfel, J. Sack, and Y. Schechter, J. Biol. Chem.

232

C. F. G. C. Geraldes and M. M. C. A. Castro

83, 609 (1987); c) A. S. Tracey, and M. J. Gresser,

Proc. Natl. Acad. Sci. USA 83,

609 (1986). 10. a) J. J. Grantham. Am. J. Physiol. 97, 239 (1980); b) 0. Pick, J. Biof. Chem. 257, 6111 (1982). 11 .B. Borah, C. Chen, W. Egan, M. Miller, A. Wlodawer, and J. S. Cohen, Biochemis-

try 24, 2058 (1985). 12. C. C. Goodno, Proc. Natl. Acad. Sci. USA 76, 2620 (1979); b) C. C. Goodno, Proc. Natl. Acad. Sci, USA 79, 21 (1982); c) K. B. Munson, M. J. Smerdon, and R. G. Young, Biochemistry 25, 7640 (1986). 13. Y. Emoto, T. Kawamura, and K. Tawada, J. Biochem. 91, 1293 (1982). 14. P. Csermely, A Martonisi, G. Levy, and A. Ejchart, Biochem. J. 230, 807 (1985). 15. S. J. Youmans and W. A. B&sky, Biochem. Biophys. Acta 900, 88 (1987). 16. M. J. Gresser and A. S. Tracey, J. Am. Chem. Sot. 107, 4215 (1985). 17. A. S. Tracey, M. J. Gresser, and B. GalefE, Znorg. Chem. 27, 157 (1988). 18. A. S. Tracey and M. J. Gresser, Proc. Natl. Acad. Sci. USA 83, 609 (1986). 19. M. J. Gresser and A. S. Tracey, J. Am. Chem. Sot. 108, 1935 (1986). 20. A. S. Tracey, M. J. Gresser, and K. M. Parkinson, Znorg. Chem. 26, 62 (1987). 21. M. M. Caldeira, M. L. Ramos, N. C. Oliveira, and V. M. S. Gil, Can. J. Chem. 65, 2434 (1987). 22. M. M. Caldeira, M. L. Ramos, A. M. Cavaleiro, and V. M. S. Gil, J. Mol. Strut. 174, 461 (1988). 23. A. S. Tracey and M. J. Gresser, Znorg. Chem. 27, 2695 (1988). 24. A. S. Tracey, M. J. Gresser, and S. Liu, J. Am. Chem. Sot. 110, 5869 (1988). 25. D. Rehder, C. Weideman, A. Duch, and W. Priebsch, Znorg. Chem. 27, 584 (1988). 26. C. F. G. C. Geraldes and M. M. C. A. Castro, J. Znorg. Biochem. 35, 79 (1989). 27. M. J. Gresser, A. S. Tracey, and K. M. Parkinson, J. Am. Chem. Sot. 108, 6229 (1986). 28. R. B. Martin and Y. H. Mariam, in Metal Ions in Biofogicaf Systems, H. Sigel, Ed., Marcel Dekker, New York, 1979, vol. 8, chap. 2. 29. R. Kato, A. Kobayashi, and Y. Sasaki, Znorg. Chem. 21, 240 (1982). 30. C. F. G. C. Geraldes, M. M. C. A. Castro, M. E. Saraiva, M. Aureliano, and B. A. Dias, J. Coord. Chem. 17, 205 (1988). 31. A. S. Tracey and M. J. Gresser, Znorg. Chem. 27, 1269 (1988). 32. C. F. G. C. Geraldes and M. M. C. A. Castro, J. Znorg. Biochem. 28, 319 (1986). 33. A. F. Nour-Eldeen, M. M. Craig, and M. J. Gresser, J. Biol. Chem. 260, 6836

(1985). 34. R. L. VanEtten, P. P. Waymack, and D. M. Rehkop, J. Am. Chem. Sot. %, 6782 (1974). 35. M. J. Gresser, A. S. Tracey, and P. J. Stankiewicz, Adv. Protein Phosphat. 4, 35 (1987). 36. R. Lagunas, Arch. Biochem. Biophys. 20, 67 (1980). 37. R. Lagunas, D. Pestana, and J. C. Diez-Masa, Biochemistry 23, 955 (1984). Received January 3, 1989; accepted May 18, 1989