A Proton Nuclear-Magnetic-Resonance Study of Self-stacking in Purine and Pyrimidine Nucleosides and Nucleotides

Size: px

Start display at page:

Download "A Proton Nuclear-Magnetic-Resonance Study of Self-stacking in Purine and Pyrimidine Nucleosides and Nucleotides"

Angelina Young
5 years ago
Views:

Eur. J. Biochem. 88, 149 154 (1978) A Proton NuclearMagneticResonance Study of Selfstacking in Purine and Pyrimidine Nucleosides and Nucleotides Paul R.

1 Eur. J. Biochem. 88, (1978) A Proton NuclearMagneticResonance Study of Selfstacking in Purine and Pyrimidine Nucleosides and Nucleotides Paul R. MITCHELL and Helmut SIGEL Institute of lnorganic Chemistry, University of Basel (Received November 28, 1977) The concentration dependence of the chemical shifts of the protons H2, and HI of A TP and of Mg(ATP), of all nonlabile protons of adenosine, of H5, H6 and HI of UTP, and of H5, H6, Hl, and H2 of uridine have been measured. The results for the purine derivatives are consistent with the isodesmic model of indefinite noncooperative stacking; for adenosine K = 15 k 2 M, for ATP K = 1.3 _+ 0.2 M and for Mg(ATP) K = 3.6 & 0.3 Ml. For the pyrimidines, uridine and UTP, stacking is much weaker and the stability constant could only be estimated; for uridine K Ml, and for UTP K z 0.3 M. Selfassociation of adenine [l] and purine [l], and of several purine nucleosides [2], including adenosine and inosine, has been observed by vapour pressure osmometry. It was concluded that indefinite noncooperative stacking occurred, and stability constants were calculated [2]; for adenosine K = 4.5 M. For pyrimidines, e.g. uridine, the interaction is very much weaker, and the calculated [l] stability constant is only 0.6 Ml. The NMR spectra of purine nucleosides are shifted in concentrated (0.1 M) solution, and this was attributed to the effect of the ring current in the stacked complexes ; however no significant shift was observed with the pyrimidine nucleosides uridine and thymidine [3] owing to the much smaller ring current in the pyrimidines [4]. In this early work external references were used; Bovey [5] repeated the experiment on purine using internal sodium trimethylsilylpropane sulphonate as reference, and concluded that, as no significant upfield shift of H2 or H6 is observed and as actually moves downfield, no stacking occurs. However sodium trimethylsilylpropane sulphonate is an unreliable reference in the presence of aromatic compounds [6,7], and thus, despite the use of an external reference, the earlier work is probably more reliable. An upfield shift due to stacking has also been observed [8 101 in NMR studies of adenosine 2: 3 and 5 monophosphates : the relaxation times of the protons H2 and, and the variation of the upfield chemical shift of the resonances H1, H2, and were said [9] to indicate head to tail stacking, with the fivemem Abbreviation. NMR, nuclear magnetic resonance. bered and sixmembered rings alternating in the stack. These H NMR studies [8 101, a H NMR study [ll], ultracentrifuge studies 112,131 and a kinetic study [14] show that adenosine 5 monophosphate [8 121 and other adenine derivatives [13,14] do not merely dimerize, but form indefinite stacks. Although difference ultraviolet spectroscopy [15] and difference circular dichroism [15,16] have been used to study the selfassociation of ATP, and the effect of ph, temperature, and of magnesium ions on this association, shifts in the H NMR spectrum of ATP, which would confirm that this association is due to stacking, have not been reported. In the course of a series of studies[l7 191 on the stacking in binary systems such as bipyridyl/atp4, phenanthroline/atp4, and bipyridyl/utp4, and in the ternary systems such as Zn2 /bipyridyl/atp4, Mg2 /phenanthroline/atp4, and Zn2 I /bipyridyl/ UTP4, it was necessary to study the selfassociation of the nucleotides themselves, and also of 2,2 bipyridyl and of 1,lophenanthroline, by NMR. For comparison we also studied adenosine and uridine. Although the selfstacking is better [I61 at lower ph, we carried out our measurements at ph 78, in the absence and presence of Mg, as this is then more relevant to the stacking of ATP4 in living systems. MATERIALS AND METHODS The disodium salt of ATP was from Serva Feinbiochemica GmbH (Heidelberg, F.R.G.). The tri

2 150 Selfstacking in Purine and Pyrimidine Nucleosides and Nucleotides sodium salt of UTP was from Sigma (St Louis, U.S.A.). type 111, > 98 % Na3UTP. HzO; it was found to contain traces of methanol (z 0.25 %> and ethanol (< 0.25%). All other chemicals were of the best purity available from Merck AG (Darmstadt, F.R.G.). The 'H NMR spectra were recorded using a Varian Anaspect EM360 spectrometer (60 MHz) at 35 "C, or a Bruker WH90FT spectrometer ( MHz) at 27"C, using the centre peak of the tetramethylammonium ion resonance as internal reference. The p2h of the solutions was obtained by adding 0.40 to the ph meter reading [20]. The experimental results were analysed using a HewlettPackard 9821 A calculator connected to a 9862A calculator plotter. RESULTS AND DISCUSSION The NMR spectrum of adenosine changes considerably as the concentration is increased from M to 0.05 M. The resonances of H2 and of the adenine part and of Hl', H2', H3', and H4' of the ribose part shift upfield by 0.150, 0.085, 0.067, 0.057, 0.016, and ppm respectively, whereas the resonances of the H5' protons do not shift significantly; i.e. the resonances of the protons of the adenine part, which is actually involved in the stacking, are shifted significantly more than those of the ribose. The shift of the resonances of H2,, and H1 ' was already known [2], although the use of an external reference in the earlier study is unsuitable for a more detailed study, and we have therefore used (CH3)4N+ as internal reference throughout this work. Similar, but somewhat larger, shifts are observed for ATP4; H2, and HI' shift by 0.337,0.163 and ppm respectively as [ATP4] is increased from 0.01 M to 0.4 M. The shifts for Mg(ATP)2 are similar to those for ATP4; 0.33, 0.19, and 0.16 ppm respectively. The variation of the upfield shifts of H2,, and H1' of adenosine, of ATP4, and of Mg(ATP)2 as a function of the concentration is shown in Fig The theory developed by Heyn and Bretz [15] for indefinite noncooperative association, in which the equilibrium constants for the equilibria (ATP4), + ATP4 G (ATP4), + I K= [(ATP), + 1],"(ATP4),] [ATP4] are all assumed to be equal, was adapted for NMR, giving the relationship Sobs = SO + (6, do) 3 K[A] K2[A12 H2 H1' H 2' + (hi + 1 / 1 2 K' [A]' Px x x x x x ~ ~ 380 xxxx X H4' [Adenosine] (mm) I I x H5 Fig. 1. The variation of the chemical shift of the protons of adenosine with varying concentration. Spectra were measured on a Bruker FT 90 at MHz (27 C pzh = 7, I = 0.1 M, NaNO, in 'HzO), relative to internal (CH3)4N+ and converted to values relative to sodium trimethylsilylpropane sulphonate by adding The curves shown in this and the other two figures are the computercalculated best fit of the experimental data using the indefinite noncooperative stacking model; the shifts of the different protons all give the same stability constant within experimental error between the observed chemical shift (Sob,) in a solution of total concentration [A] and the chemical shifts of a free molecule (i.e. the shift at infinite dilution) (SO), of a molecule at the end of a stack (S& of a molecule within a stack (hi), and the stability constant. As the distance between stacked molecules is typically of the order of 0.35 nm [21], and as the upfield shifts due to a ring current falls off very rapidly with increasing separation of the rings 1221, only the ring current in adjacent molecules in the stack is expected to have a significant' effect on the upfield shift. More l The contribution due to nextnearest neighbours, and yet more remote molecules, may be estimated by extrapolation of published results [22] as < 6% in a tetramer, 4 9% in a pentamer, 4 11 in a hexamer, etc., and 4 20% in a infinitely long stack. Moreover the concentration of molecules larger than pentamer is small in our experiments.

3 P. R. Mitchell and H. Sigel over the upfield shifts caused by the two adjacent molecules on a molecule within a stack are expected to be additive, i.e. (Si SO) = 2 (6, SO). Expression (1) then simplifies to Sobs = 60 + (hi 80) '\ (1 + [1 V KGGmK[AI 1. (2) The shift in an infinitely long stack (6,) is then the same as the shift of a molecule within a shorter stack. If species larger than dimers are ignored then analogous derivation of the relationship between the observed upfield shift and the total concentration gives : I I I I [ATPI (mm) Fig. 2. The variation of the chemical shift of H2, and HI' of ATP4 with varying concentration. Spectra were measured on a BrukerFT90 at MHz (27"C, p2h = 8.4, I= 0.1 zz 2M, NaN03 in 2Hz0). Owing to the more complicated nature of the 2', 3', 4: and 5' resonances of ribose in ATP4 and in Mg(ATP)', and to the smaller shifts of these protons, exact measurements of the upfield shifts are more difficult; see legend to Fig t I 1 I, I [Mg(ATP)'] (mm) Fig. 3. The variation of the chemical shift of H2,, and HI' of Mg(ATP)' with varying concentration. Spectra were measured on a Varian Anaspect EM360 spectrometer at 60 MHz (35 "C, ph = 8.5, I = 0.1 zz 2 M, NaN03 in H20). Other details as in Fig. 1 and 2.., Expressions (2) and (3) are identical except that Si is replaced by &, the upfield shift in a dimer, and K is replaced by 2 KD, twice the equilibrium constant for the dimerization : ATP4 + ATP4 $ (ATP)i KD = [(ATP):]/ [ATP4I2. Computercalculated leastsquares fits of the variation of the upfield shifts of each of the protons with increasing concentration were performed using all three models. The concentration range which could be studied was limited by the solubility of adenosine, and by problems in preparing solutions of ATP4 stronger than z 0.4 M; this prevented determination of the ratio (Si So)/(S, SO); however, for all the resonances the computer fit of the results was not improved by taking (Si 60) < 2 (8, SO), and a very slight improvement by taking (Si SO) > 2 (6, So) was not statistically significant. This tends to confirm our expectation that the ring currents are additive and that the influence of nextnearest neighbours is not significant. The concentration dependence of the chemical shifts of H2,, and H1 ' of ATP4, of Mg(ATP)', and of H2,, Hl', and H2' of adenosine fit expression (2) well ; the equilibrium constants obtained are listed in Table 1. The resonances of H3' and H4' of adenosine also shift upfield slightly: although the shifts are too small to permit calculation of the stability constant, they are also consistent with K = 15 M '. The resonances of the inequivalent CHz protons H5' do not shift significantly. The dimer model naturally fits the results just as well, giving stability constants half of those given by the indefinite noncooperative stacking model (Table 1).

4 ~ 152 Selfstacking in Purine and Pyrimidine Nucleosides and Nucleotides Table 1, Equilibrium constants,fbr stacking, and chemical shifts of monomeric and stacked ATF, Mg( ATPI' and ndenosine K for ATP was measured at 27 "C, 1 = M (NaN03), for Mg(ATP)2 at 35 'C, I = M (NaN03) and for adenosine at 27 "C, I = 0.1 M (NaN03). Values of KD for ATP and Mg(ATP)' are given for comparison only, as association proceeds beyond the dimer stage. Chemical shifts are measured relative to internal (CH3)4N+ and converted to values downfield from sodium trimethylsilylpropane sulphonate by adding ppm ~ ~ Compound Proton K KD 60 6, Upfield shift ATP H2 H1' Average M~(ATP)~ H2 HI ' Average. Adenosine H2 HI ' H2' H3' H4' H5'" H5'" Average ~ M' PPm ~ & 0 30 (1 28 f 0 12)" (064 & 008)" f 0 05 (0 54 i 0 1)" (1 0 & 02)" ~ ~ & f f _ ~ b b b b 15 2 h b h b 7.5 k ~ f & f i: f f a Values in parentheses were determined in a different experiment using a different batch of ATP, at 35 C in HzO on a 60MHz spectrometer K could not be determined as these shifts were too small. The two H5' protons are inequivalent. However the upfield shifts (Table I), especially for H2 and of ATP4 and of H2 of Mg(ATP)', are much higher than would be expected for the shift due to a single adjacent molecule, as in the dimer, either from calculations of the size of the ring current in the adenine moiety [4] or from the upfield shift caused by ATP on the protons of phenanthroline in the ternary stacked complexes Mg. (phenanthroline). (ATP)' or Zn. (phenanthroline). (ATP)' [18]. The shifts are about twice what would be expected, suggesting that the indefinite noncooperative stacking model, in which the shift is caused by two adjacent molecules, is preferable. Indeed there is abundant evidence that association of AMP proceeds beyond the dimer stage [8 121 and this is presumably also true for ATP [15] and possibly also for adenosine. Using the above stability constants the upfield shifts of an infinitely long stack in the table can be calculated. It is noticeable that for H2,, and HI ' the upfield shifts decrease in the series ATP4 =. Mg(ATP)' > adenosine. This difference, which cannot be explained by assuming that the calculated stability constants are in error for some unknown reason, may be due to variations in the relative orientation of the molecules in the stack, which result from changes in the repulsion between the phosphate groups on formation of the Mg(ATP)' complex; in adenosine there is no such repulsion. A head to tail geometry was said to be more consistent with the ultraviolet absorption and circular dichroism results on ATP [15], and with NMR results on AMP [9]; the orientation of the molecules in stacked adenosine is not known. The upfield shift of ATP4 is enhanced by increasing the concentration of Na', particularly at lower [ATP4]. The stability constant of Na(ATP)3 is known[23,24], K{t(ATP) = 15 M', and so it appears that Na(ATP)3 selfstacks slightly better than ATP4. Therefore the equilibrium constant given above for ATP4 must be considered an upper limit, as in the more concentrated ATP4 solutions considerable Na(ATP)3 was present, thus affecting the result. Comparison of the equilibrium constants measured at 35 "C for selfstacking of ATP4 (K = 1.0 M') and of Mg(ATP)' (K = 3.6 M') shows that Mg(ATP)' selfstacks much better than ATP. This increased stacking for Mg(ATP)' is of considerable relevance to the properties of ATP in living systems. For example at the high concentrations (sz 0.1 M) of ATP present in the adrenal chromaffin granules [25], considerable selfstacking probably occurs.

5 134. P. R. Mitchell and H. Sigel 153 Although one earlier study [I61 by circular dichroism reported that at ph 2.7 the addition of Mg2+ decreases the stacking, and the equilibrium constant falls from 158 M' to 22 M', this is certainly not due to Mg(ATP)2 as under these conditions this species will not be formed. However in another similar study [I51 at ph 7.7 to 8.7, it was observed that the equilibrium constant for indefinite stacking (at 25 "C) increased from 3.1 M' for ATP4 to 5.6 M' for Na(ATP)3 and to 8.2 M' for Mg(ATP)2. This pattern is similar to that which we observe although our stability constants, measured at 35 "C, are somewhat lower. The NMR spectrum of uridine was also observed to change slightly as the concentration is increased from 0.01 M to 0.2M; however the change is very small; the upfield shifts of H5, H6, H1 ' and H2' are only 0.010, 0.005, 0.010, and ppm respectively. In the concentration range 0.01 to 0.1 M the resonances of H5 and H1' of uridine 5'triphosphate (UTP) shift by about 0.01 ppm. The resonance of H6 is not concentrationdependent, but does shift upfield by z 0.02 ppm as "a'] is increased from 0.02 M to 0.6 M. Under the conditions used, (I = 0.6 M, NaN03) the UTP is predominantly Na(UTP)", assuming the stability constant K E (UTP) to be similar to that of KEt (ATPI [23,24]. Although these shifts show that stacking can occur, they are far too small to allow a stability constant to be calculated. The ring current in the pyrimidines is much lower than in the purines [4] and the relative size of the upfield shifts caused by UTP compared with that caused by ATP can be estimated by comparing the shifts of bipyridyl in the ternary complexes Zn. bipyridyl. (UTP)2 [I91 and Zn. bipyridyl. (ATP)2 (Mitchell, P. R. and Sigel, H., unpublished results). The shift caused by UTP is only about one fifth of that of ATP, and moreover the relative shift differs for the four bipyridyl protons; indeed the upfield shift in these stacked pyrimidines probably arises at least as much from the deshielding by the C = C double bond [26] as from the very small ring current. It can thus be estimated that the stability constant, K, for selfassociation of uridine is certainly less than 1 M ' and is probably /< 0.5 M'; osmotic pressure measurements by Ts'o [l] suggested K = 0.6 M'. For UTP the stability constant is probably about 0.3 M'; thus there is much less difference between the selfstacking tendencies of uridine and Na(UTP)3, than between those of adenosine and N~(ATP)~. Since this work was completed, Lam and Kotowycz 1271 have reported that Bovey's failure [S] to observe selfstacking in ATP is due to the unreliability of the sodium trimethylsilylpropane sulphonate that Bovey used as reference. They also reported concentrationdependent shifts in ATP, using fertbutyl alcohol as internal reference, which is only slightly affected by hydrophobic interactions [28] ; they did not, however, calculate any stability constants. We have calculated stability constants from their published results [27], and find that the value obtained, K = 0.8 k 0.4 M' (27 "C), agrees well with our value. Their results[27] also indicate that at [ATP] > 0.3 M significant hydrophobic interaction occurs between ATP (either monomer or selfstacked) and tertbutyl alcohol, the resonance of the latter also being shifted upfield by z 0.06 ppm at [ATP] = 0.56 M. The stability constant for this hydrophobic interaction must be very small, K < 0.3 MI. In the light of our recent results on hydrophobic interactions between the trimethylsilyl moiety and 2,2'bipyridyl or 1,lophenanthroline [7], and our studies on stacking interactions between the bases of nucleotides and aromatic coinpounds [17 191, this observation [27] seems very interesting as it suggests that hydrophobic interactions between purine moieties of nucleotides and aliphatic groups, e.g. isopropyl side chains of amino acids, may occur in nature. We thank Mr K. Aegerter for recording the 90MHz NMR spectra, the CIBA Foundation for a grant towards the costs of these spectra, and the Swiss National Science Foundation for a research grant. REFERENCES 1. Ts'o, P. 0. P., Melvin, 1. S. & Olson, A. C. (1963) J. Am. Chem. SOC. 85, Broom, A. D., Schweizer, M. P. & Ts'o, P. 0. P. (1967) 1. Am. Chem. Soc. 89, Schweizer, M. P., Chan, S. I. & Ts'o, P. 0. P. (1965) J. Am. Chem. Soc. 87, GiessnerPrettre, C. & Pullman, B. (1965) C. R. Hehd. S&ancc,s Acad. Sci. (Paris) 261, Bovey, F. A. (1972) HighResolution NMR of Macromolecules, p. 399, Academic Press, New York. 6. Hand, E. S. & Cohen, T. (1965) J. Am. Chem. Soc. 87,133 ~ 7. Mitchell, P. R. & Sigel, H. (1976) Angeiz. Chem. Int. Ed. n,ol. 15,548, and Angew. Chem. 88, Schweizer,M.P., Broom, A.D., Ts'0,P.O. P. & Hollis, D. P. (1968) J. Am. Chem. Soc. 90, Son, T.D. & Chachaty, C. (1973) Biochim. Biophys. Acta, 335, Evans, F. E. & Sarma, R. H. (1974) Biupolymers, 13, Egan, W. (1976) J. Am. Chem. Soc. 98, Rossetti, G. P. & Van Holde, K. E. (1967) Biochem. Biophys. Res. Commun. 26, Bretz, R., Lustig, A. & Schwarz, G. (1974) Biophys. Chem. 1, Heyn, M. P., Nicola, C. U. & Schwarz, G. (1977) J. Phys. Chem. 81, Heyn, M. P. & Bretz, R. (1975) Biophys. Chem. 3, Gilligan, T..I.& Schwarz, G. (1976) Biophys. Chem. 4, Chaudhuri, P. & Sigel, H. (1977) J. Am. Chem. Soc. 99, Mitchell, P. R. & Sigel, H. (1978) J. Am. Chem. Soc. 100, Fukuda, Y., Mitchell, P. R. & Sigel, H. (1978) Helv. Chim. Acta, 61,

6 154 P.R. Mitchell and H. Sigel : Selfstacking in Purine and Pyrimidine Nucleosides and Nucleotides 20. Glasoe, P. S. & Long, F. A. (1960) J. Phys. Chem. 64, Naumann, C. F., Prijs, B. & Sigel, H. (1974) Eur. J. Biochem. 41, Johnson, C. E. & Bovey, F. A. (1958) J. Chem. Phys. 29, O'Sullivan,W. J. & Perrin,D.D. (1964) Biochemistry, 3, Botts, J., Chashin, A. & Young, H. L. (1965) Biochemistry, 4, Frausto da Silva, J. J. R. & Williams, R. J. P. (1976) Nature (Lond.) 263, Jackman, L. M. & Sternhill, S. (1969) Applications of Nuclear Magnetic Resonance in Organic Chemistry, 2nd edn, pp , Pergamon Press, Oxford. 27. Lam, Y.F. & Kotowycz, G. (1977) FEBS Lett. 78, Jones, R. A. Y., Katritzky, A. R., Murrell, J. N. & Sheppard, N. (1962) J. Chem. SOC Fan, S., Storer, A. C. & Hammes, G. G. (1977) J. Am. Chem. SOC. 99, Lam, Y.F. & Kotowycz, G. (1977) Can. J. Chem. 55, P. R. Mitchell and H. Sigel, Institut fur Anorganische Chemie der Universitat Basel, Spitalstrasse 51, CH4056 Basel, Switzerland Note Added in Proof. Two independent reports which mention the concentration dependence of the NMR of ATP have appeared since this paper was written [29,30]. One assumes that only dimers are formed 1291 and thus deduces KD = 0.45 M' ; the other, the definitive paper of the preliminary communication [27] mentioned above, assumes without evidence that the upfield shift is the same in dimers, trimers, tetramers, etc., and deduces that K = 0.9 M' for the monomer/dimer and dimer/trimer equilibria [30]. Both reports agree in general with ours, although, because of the different models used, rather lower stability constants are obtained. We believe that our results show that stacking proceeds beyond dimers, and that the shift in a dimer is less than that in a trimer. etc., owing to the increasing proportion of molecules with two nearest neighbours.

to all three phosphate groups P, P Q, P of ATP, Mg +, Ca +, p

to all three phosphate groups P, P Q, P of ATP, Mg +, Ca +, p Volume 4 Number 2 February 1977 Nucleic Acids Research A 31 P - NMR study of the interaction of M^2" 1 " ions with nucleoside diphosphates S.Tran-Dinh and J.M.Neumann Departement de Biologie - Centre d'etudes