- Email: [email protected]
Nuclear magnetic relaxation studies of cytidine and its complexing in dimethyl sulfoxide solution
349
Biochimica et Biophysica Acta, 425 (1976) 349--355 © Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands
BBA 98538
NUCLEAR MAGNETIC RELAXATION STUDIES OF CYTIDINE AND ITS COMPLEXING IN DIMETHYL SULFOXIDE SOLUTION
SHIGEZO S H I M O K A W A , T E T S U R O Y O K O N O
and JUNKICHI S O H M A
Faculty of Engineering, Hokkaido University, Sapporo 060 (Japan) (Received August 26th, 1975)
Summary ~sC spin-lattice relaxation times (T1 values) in cytidine were determined experimentally to investigate molecular motions of both metal-free and ion-complexed cytidines in dimethylsulfoxide solutions. It was found that the correlation times of the protonated carbons were equal within experimental error, and this equality of correlation times of different sites of the molecule suggests strongly isotropic random motion of the molecule. Correlation times for interhal motion of the amino group obtained from the observed T~ of the amino protons are 4.6 • 10 -~ s, 2.0 • 10 -9 s and 1.1 • 10 -~° s for the metal-free cytidine and the cytidine complexed with either CaC12 or ZnC12, respectively. An experimental value of T~ of the H e proton of the cytidine base agrees very well with the value estimated from a conformation determined by the nuclear "Overhauser" effect. Spin-lattice relaxation time measurements of the 7Li nucleus in the LiCl/cytidine system strongly suggested that the 7Li cation is directly coordinated with cytidine.
Introduction
It :is known that metal ions play a very important role in the living system [ 1]. We have studied the interaction between nucleosides and metal ions by the use of NMR (PMR and lsC NMR) spectroscopy [2--5]. In a previous paper [ 5], changes in the chemical shifts induced by addition of metal salts to dimethylsulfoxide (Me2SO) solutions of cytidine and guanosine can not be explained as the single effect of either metal ion or anion but has some complexity. Recently, several researchers [6,7] reported the spin-lattice relaxation times (T1 values) of the lsC and ~H in order to investigate molecular motion and the molecular structure of the metal-free and ion-complexed molecules. In the present report, the T~ values observed for both lsC and ~H of metal-free as Abbreviation: Me2SO, dimethylsulfoxide.
350 well as ion-complexed cytidines are presented and molecular motions will be discussed based on the experimental results of these T~ values. Materials and Methods Cytidine was obtained from Kohjin Co. and dried b y pumping in a vacuum desicator over silica gel for several days. Me2SO was purified twice by vacuum distillation after drying over calcium hydride. Commercial grade CaCI2 was dried in an oven before use. Anhydrous ZnC12 was obtained by heating in a quartz vessel filled with dry HC1 gas. LiC1 of reagent grade was used without further purification. All the solutions were degassed b y a repetition of the freeze and pump-thaw m e t h o d until dissolved gasses were sufficiently removed. The solutions were then re-frozen in a sample and the tube sealed. Nuclear spinlattice relaxation times of both ~H and ~3C were determined from the partially relaxed Fourier transform method, using the 180-r-90 ° pulse sequence at either 90 MHz for 1H or 22.63 MHz for ~aC. For T~ measurements of the 7Li nucleus, a conventional 180-r-90 ° pulse sequence at 34.98 MHz was performed. A Bruker SXP 4-100 pulse and FT spectrometer was used for all T~ measurements in the experiment. Viscosity of the solutions was measured by an Ostwald viscosimeter, at r o o m temperature. Results and Discussion
1. ~3Cspin-lattice relaxation Although there are three possible mechanisms [8,9], for ~3C spin-lattice relaxation, namely chemical-shift anisotropy, spin rotation and ~3C-~H dipoledipole interaction, the last is reasonably assumed to be predominant for a prot o n a t e d 13C. Furthermore, a molecular reorientation is assumed to be isotropic and described by a single rotational correlation time, TR. An approximation of the extreme narrowing can be made for a system such as an Me2SO solution of nucleosides. Thus, a spin-lattice relaxation time, T~, is related to a rotational correlation time, TR, b y the equation: _11 = Ta2~,c 2"),H~ rCHi-6TRN
yl
(1)
where N is the number of protons attached directly to the carbon, rCH i is the distance between laC and the i-proton, and ~'Hi and 7c are the gyromagnetic ratios of 1H i and ~3C. The T~ values observed for IaC at the various sites of the molecule are listed in Table I for b o t h the metal-free and the complexed cytidines. Assignment of each peak of the 13C spectrum in cytidine is taken from the literature [10]. Longer relaxation times, of 13C at the carbon positions 2 and 4 (see Table I), correspond to the u n p r o t o n a t e d form. This apparent correspondence supports the assumption that the dipole-dipole interaction between 13C and IH predominates in the ~ac relaxation process. The carbons other than C2 and C4 may be classified into three groups from the view point of the observed spin-lattice relaxation; C~' having a longer T~ of 0.23 s; Cs' having a shorter T~ of 0.10 s; and the others, C6, Cs, C4', C2' and Ca', having nearly same T~ value of 0.15--0.20 s, which is within experimental error +15%. The relaxation time of Cs' having the t w o protons is nearly one half of
351 TABLE I 13 C S P I N - L A T T I C E R E L A X A T I O N IN M e 2 S O
TIMES (T 1 VALUES) FOR CYTIDINE AND ITS CaCl 2 COMPLEX
Carbon atom
Tl.cytidine a (s)
Tl.cytidine. CaC12 b (s)
C4 C2 C6 C5
2.05 1.02 0.19 0.16
0.58 1.05 0.20 0.15
C1' C 4' C 2' C 3' C5'
0.28 0.15 0.20 0.20 0.10
0.13 0.16 0.20 0.17 0.09
a 0.5 mol/1 cytidine. b 0 . 5 m o l / l c y t i d i n e , 0 . 5 m o l f l CaC12.
those of carbons having one proton in the third group; this shorter relaxation time can be explained by Eqn. 1, in which the number o f protons directly attached to the carbon is t w o for Cs' and one for the others. This quantitative agreement reconfirms that the main relaxation path of 13C is the dipolar interaction. Furthermore, the fact that the observed T1 is inversly proportional to the number of the attached protons involves a similarity of correlation times for the different carbons of the cytidine molecule. On the basis of this fact and the considerations presented above, one can conclude that random motion of the molecule can be described in terms of single correlation time, rR, at least as a good approximation. This conclusion is consistent with an assumption of the isotropic rotation of the bulk molecule. It was also derived, from the fact that the different carbons of the base and the ribose in the third group have the same T1 values, that there is no intermolecular base stacking. This is because the stacking provides an additional relaxation path and the T1 values of the base carbons would therefore be different from those of the sugar [6], in the case o f stacking between cytidine bases. Using Eqn. 1, and the observed value of 0.17 s for the TI of a base carbon (the average TI values for the carbons at positions 5 and 6 in free cytidine), one obtains rR = 1.5 • 10 -1° (S). This value agrees well with the value reported for the TR of a similar molecule like 2',3'isopropylideneadenosine [11]. Inserting the above value for 7R, a measured viscosity, ~ = 0.039 g/cm per s, of the Me2SO solution and the temperature into the well-known formula [7] : rR = 4 ~ a 3 / 3 k T
(2)
the radius of a cytidine molecule, which is approximated as a sphere, can be estimated as 3.34 A. This approximated radius agrees well with the value, 3.4 A, obtained from a molecular model. It was found that addition o f CaC12 salt to an Me2SO solution of cytidine reduces the relaxation time TI of ~3C at carbon position 4, which is attached to the amino group, to nearly a quarter. Further, it was found that observed changes of TI values of the other carbons of
352 the base remain within experimental error and are actually unvaried (Table I). This fact indicates t h a t metal salts interact mainly with the amino group and/or the nitrogen atom at position 3. The above results are compatible with the 1HNMR chemical shift variation induced by addition of the metal salts [4].
2. IH spin-lattice relaxation It has been reasonably assumed that a spin-lattice relaxation time of a proton is determined by the dipolar interaction with neighboring protons which belong to either of the two molecules. Moreover, the spin-lattice relaxation of a proton in our experiments can be approximated between extremely narrow limits and T1 is represented by the following equation [12]
p:~h:74Hrc~j
rpj-6
(3)
where rpj is the inter-proton distance between the protons j and p, re is the correlation time o f the molecular motion involved in this interaction, and 7H is the gyromagnetic ratio of 1H. Since the amino group has an additional freedom of rotation about the C-N axis, plus an overall reorientation of a molecule, TI of the amino protons is determined with an effective correlation time r¢ expressed by the following equation [6]. r; 1 " r~ 1 + r['
(4)
where rR is a correlation time for overall reorientation of a molecule and r, is that of an internal m o t i o n of a group attached to a skelton undergoing random reorientation. Therefore, the T1 of the amino protons is affected by the internal motion of the group as well as the overall motion of the molecule. The effective correlation time rc was determined to be 0.35 • 10 -1° (s) by Eqn. 3, in which the observed value of T, and the separation between the amino protons had been inserted. On the other hand, the correlation time of the overall reorientation of a molecule in Eqn. 4 is reasonably assumed to be identical with the correlation time determined from the measured T, values of the carbons of the cytidine base, so far as the C-H bond is rigid. Then, using 1.5 • 10 -1° (s) and 0 . 3 5 . 1 0 - 1 ° (s) for rR and rc in Eqn. 4, respectively, one obtained 4.6 • 10 -1' (s) for ri, in the case of metal-free cytidine. The value obtained for ri is much shorter than rR. This shorter correlation time for the amino group suggests that the amino group undergoes a kind of internal rotation in the Me2SO solution of metal-free cytidine. It was f o u n d t h a t the relaxation times in the amino protons were reduced by addition of the salts (CaC12, ZnC12) (See Table II). Two plausible explanations will be put forward for the observed reduction of T1 in complexing of cytidine. (1) No additional path for the relaxation is assumed and the dipolar interaction still governs the relaxation even in the case of complexed cytidine. Based on this assumption, the correlation times of the internal rotation of the amino group, ri, were determined from the observed TI values of the complexed cytidine by the m e t h o d similar to t h a t given above. The correlation time obtained in such a way is 2.0 • 10 -9 (s) and 1.1 • 10 -1° (s) for cytidine • CaC12 and cyti-
353 T A B L E II 1 H S P I N - L A T T I C E R E L A X A T I O N T I M E S ( T 1 V A L U E S ) F O R C Y T I D I N E A N D I T S CaCI 2 A N D Z n C l 2 C O M P L E X E S IN [ 2 H 5 ] D I M E T H Y L S U L F O X I D E
Hydrogen atom
Tl.cytidine a
Tl.Cytidine. CaCI 2 b
Tl.cytidine. ZnCl 2 c
NH 2 H6
0.40 0.45
0.10 0.45
0.22 0.35
a 0 . 1 m o l / l cytidine. b 0.1 m o l ] l c y t i d i n e , 0 . 3 m o l / l CaC12. c 0.1 m o l / l c y t i d i n e , 0 . 3 tool/1 Z n C I 2.
dine. ZnCl2, respectively. The values obtained for ri of the complexes are larger than that of the metal-free cytidine, and ri in these cases are either nearly same as rR or longer. This means that the internal rotation of the amino group is retarded and probably stops in the molecular frame by addition of the metal salts. This hinderance of the internal motion of the amino group strongly suggests that a kind of complexing, presumably a metal-ion complex, occurs at the amino group. This plausible conclusion is consistent with the result derived from the 13C relaxation mentioned in the last section, as well as with the observation of separate peaks for the amino protons in the cytidine complex. [4] Recently, Raszka et al. [13] reported that the amino group for CMP in water at neutral pH exhibited a restricted rotation at room temperature. They observed that the NMR peak from the amino group appeared to be split into two peaks at lower temperatures, such as --5°C. We also found that similar separation occurs for cytidine in water at neutral pH while no separation was observed in Me2SO solution. (2) The additives may provide an additional path for the proton relaxation, for example, an electric-quadripolar interaction of an anion, CI-, coordinated to the amino group. In the previous paper [5] we reported that the anion interacts mainly with the amino protons i n cytidine and, therefore, an extra relaxation path for the amino protons through a rapid quadripolar relaxation of coordinated C1- cannot be denied. Through either one of the two mechanisms mentioned above a restriction of the amino group rotation may be caused. The relaxation time of the proton at the H 6 position is considered to reflect the conformation around the glycosidic bond. The relaxation time of the H 6 proton is assumed to be expressed by the following equation: ( ~1 ) H ~ = ( -~11)H~'Hs + ~ ( ~1 ) H~-(ribose protons)
(5)
The distance between the H 6 and H s protons is 2.44 A, obtained from X-ray crystallographic data [14] in cytidine and the distance between H e and the ribose proton (rA) can be estimated from the nuclear Overhauser effect (NOE) measurement by using the following relation [15] where the coefficent A is 1.8 1
NOE
.-- 2 + A r 6
(6)
354 • 10 -2 and the value of NOE * is obtained from the results of the nuclear Overhauser effect [4,16]. The distance between H 6 and one of the ribose protons is estimated to be 2.65 A through Eqn. 6 and the observed NOE value. From Eqn. 5 and the distances mentioned above, the additional relaxation rates due to the couplings with the ribose protons were evaluated, and therefore the total relaxation rate determined by Eqn. 5 was estimated as 0.89 (s) for H 6. This value agrees fairly well with the experimental value 0.45 (s). Consequently, the results of both the nuclear Overhauser effect and T~ of the H 6 indicate that the cytidine in the metal-free system takes the intermediate conformation, which was proposed by the authors in a former report [4]. By examing Table I, one can claim that addition of CaCI2 reduces appreciably the relaxation time of C1' while the Tt values of the other carbons are virtually unchanged, except for C4 which was discussed in the last section. It was reported that the nuclear Overhauser effect was observed between H 6 and H l' only in cytidine • CaC12 and n o t in metal-free cytidine, and the observed changes in the nuclear Overhauser effect were ascribed to the conformational change induced by complexing of cytidine [4]. This nuclear Overhauser effect indicates an additional relaxation path for the 1' proton through this dipole-dipole coupling with H 6. This additional relaxation path, which appears only in the cytidine- CaC12 system, may give rise to an appreciable increase of the relaxation of ~3C at the 1' site in the case of cytidine • CaC12. Based on this argument, one can safely conclude that the observed reduction in T1 of C~' in the case o f cytidine • CaC12 supports the conformational change around the glycosidic bond induced by this addition of the metal salt.
3. VLi relaxation Concerning the relaxation mechanism [17] of the 7Li nucleus in an aqueous solution, one must consider t w o mechanisms, the magnetic dipolar and the electric quadripolar. To elucidate a direct effect of the metal ion on cytidine, 7Li relaxation experiments in LiC1 and LiCI • cytidine aqueous solutions were performed. The relaxation times obtained for LiC1 (1 M) and LiC1 (1 M)cytidine (0.5 M) are 19 (s) and 12.5 (s), respectively. The appreciable change of the 7Li relaxation time reveals that there exists a direct interaction between the metal ion (Li ÷) and cytidine. References 1 Eichhorn, G.L. (1973) Inorganic Biochemistry (Eichhorn, G.L., ed.), Chapts 33 and 34, Elsevier, Amsterdam 2 Sohma, J., Shimokawa, S. and Hotta, K (1968) R e c e n t D e v e l o p m e n t of Magnetic Resonance in Biological Systems (Fujiwara, S. and PIette, L.H., eds.), pp. 57--69, Hirokawa Publishing Co., T o k y o 3 Shimokawa, S., Fukui, H., Sohma, J. and Hotta, K. (1973) J. Am. Chem. Soc. 95, 1 7 7 7 - - 1 7 8 2 4 Y okono, T., Shlmokawa, S., Fukui, H. and Sohma0 J. (1973) J. Chem. Soc. Japan, Chem. I n d u c t . Chem. 2, 201--206 5 Y o k o n o , T,, Shimokawa, S. and Sohma, J. (1975) J. Am. Chem. Soc. 97, 3 8 2 7 - - 3 8 2 9 6 Anerhand, A.+ Doddrell, D. and Komoroski~ R. (1971) J. C h e m . Phys. 55, 189--197 7 Fedaxko, M.C. (1 973) J. Mag. Res. 12, 30--35 8 Abragam, (1961) The Principles of Nuclear Magnetic Resonance, pp. 64 Oxford, L o n d o n * The nuclear Overhauser effect in Eqn. 6 is d e f i n e d as (I--Io)/Io, w h e r e I 0 is the NMR i nt e ns i t y w i t h o u t r-f irradiation and I is t h e same q u a n t i t y with r-f irradiation.
355 9 Hertz, H.G. (1967) Progress in Nuclear Magnetic Resonance Spectroscopy (Ems/ey, J.W., Feeney, J. and Sutcliffe, L.H., eds.), VoL 3, Chapt. 5, Pergamon Press, London 10 Jones, A.J., Grant, D.M., Winkley, M.W. and Robbins, R.K. (1970) J. Amer. Chem. Soc. 92, 4079-4087 I I Akasaka, K., Imoto, T. and Hatano, H. (1973) Chem. Phys. Lett. 2 1 , 3 9 8 - - 4 0 0 12 Solomon, I. (1955) Phys. Rev. 99, 126--131 13 Raarka, M. and Kaplan, N.O. (1972) Proc. Natl. Acad. Scl. U.S.A. 69, 2075 14 Furberg, S., Petersen, C.S. and Romming, C. (1965) Acta. Crystalogr. 18, 313--320 15 Bell, R.A. and Saunders, J.K. (1970) Can. J. Chem. 48, 1114--1122 16 Hart, P.A. and Davis, J.P. (1969) Biochem. Biophys. Res. Commun. 34, 733--739 17 Woessner, D.E., Snowden, Jr., B.S. and Ostroff, A.G. (1968) J. Chem. Phys. 4 9 , 3 7 1 - - 3 7 5
Biochimica et Biophysica Acta, 425 (1976) 349--355 © Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands
BBA 98538
NUCLEAR MAGNETIC RELAXATION STUDIES OF CYTIDINE AND ITS COMPLEXING IN DIMETHYL SULFOXIDE SOLUTION
SHIGEZO S H I M O K A W A , T E T S U R O Y O K O N O
and JUNKICHI S O H M A
Faculty of Engineering, Hokkaido University, Sapporo 060 (Japan) (Received August 26th, 1975)
Summary ~sC spin-lattice relaxation times (T1 values) in cytidine were determined experimentally to investigate molecular motions of both metal-free and ion-complexed cytidines in dimethylsulfoxide solutions. It was found that the correlation times of the protonated carbons were equal within experimental error, and this equality of correlation times of different sites of the molecule suggests strongly isotropic random motion of the molecule. Correlation times for interhal motion of the amino group obtained from the observed T~ of the amino protons are 4.6 • 10 -~ s, 2.0 • 10 -9 s and 1.1 • 10 -~° s for the metal-free cytidine and the cytidine complexed with either CaC12 or ZnC12, respectively. An experimental value of T~ of the H e proton of the cytidine base agrees very well with the value estimated from a conformation determined by the nuclear "Overhauser" effect. Spin-lattice relaxation time measurements of the 7Li nucleus in the LiCl/cytidine system strongly suggested that the 7Li cation is directly coordinated with cytidine.
Introduction
It :is known that metal ions play a very important role in the living system [ 1]. We have studied the interaction between nucleosides and metal ions by the use of NMR (PMR and lsC NMR) spectroscopy [2--5]. In a previous paper [ 5], changes in the chemical shifts induced by addition of metal salts to dimethylsulfoxide (Me2SO) solutions of cytidine and guanosine can not be explained as the single effect of either metal ion or anion but has some complexity. Recently, several researchers [6,7] reported the spin-lattice relaxation times (T1 values) of the lsC and ~H in order to investigate molecular motion and the molecular structure of the metal-free and ion-complexed molecules. In the present report, the T~ values observed for both lsC and ~H of metal-free as Abbreviation: Me2SO, dimethylsulfoxide.
350 well as ion-complexed cytidines are presented and molecular motions will be discussed based on the experimental results of these T~ values. Materials and Methods Cytidine was obtained from Kohjin Co. and dried b y pumping in a vacuum desicator over silica gel for several days. Me2SO was purified twice by vacuum distillation after drying over calcium hydride. Commercial grade CaCI2 was dried in an oven before use. Anhydrous ZnC12 was obtained by heating in a quartz vessel filled with dry HC1 gas. LiC1 of reagent grade was used without further purification. All the solutions were degassed b y a repetition of the freeze and pump-thaw m e t h o d until dissolved gasses were sufficiently removed. The solutions were then re-frozen in a sample and the tube sealed. Nuclear spinlattice relaxation times of both ~H and ~3C were determined from the partially relaxed Fourier transform method, using the 180-r-90 ° pulse sequence at either 90 MHz for 1H or 22.63 MHz for ~aC. For T~ measurements of the 7Li nucleus, a conventional 180-r-90 ° pulse sequence at 34.98 MHz was performed. A Bruker SXP 4-100 pulse and FT spectrometer was used for all T~ measurements in the experiment. Viscosity of the solutions was measured by an Ostwald viscosimeter, at r o o m temperature. Results and Discussion
1. ~3Cspin-lattice relaxation Although there are three possible mechanisms [8,9], for ~3C spin-lattice relaxation, namely chemical-shift anisotropy, spin rotation and ~3C-~H dipoledipole interaction, the last is reasonably assumed to be predominant for a prot o n a t e d 13C. Furthermore, a molecular reorientation is assumed to be isotropic and described by a single rotational correlation time, TR. An approximation of the extreme narrowing can be made for a system such as an Me2SO solution of nucleosides. Thus, a spin-lattice relaxation time, T~, is related to a rotational correlation time, TR, b y the equation: _11 = Ta2~,c 2"),H~ rCHi-6TRN
yl
(1)
where N is the number of protons attached directly to the carbon, rCH i is the distance between laC and the i-proton, and ~'Hi and 7c are the gyromagnetic ratios of 1H i and ~3C. The T~ values observed for IaC at the various sites of the molecule are listed in Table I for b o t h the metal-free and the complexed cytidines. Assignment of each peak of the 13C spectrum in cytidine is taken from the literature [10]. Longer relaxation times, of 13C at the carbon positions 2 and 4 (see Table I), correspond to the u n p r o t o n a t e d form. This apparent correspondence supports the assumption that the dipole-dipole interaction between 13C and IH predominates in the ~ac relaxation process. The carbons other than C2 and C4 may be classified into three groups from the view point of the observed spin-lattice relaxation; C~' having a longer T~ of 0.23 s; Cs' having a shorter T~ of 0.10 s; and the others, C6, Cs, C4', C2' and Ca', having nearly same T~ value of 0.15--0.20 s, which is within experimental error +15%. The relaxation time of Cs' having the t w o protons is nearly one half of
351 TABLE I 13 C S P I N - L A T T I C E R E L A X A T I O N IN M e 2 S O
TIMES (T 1 VALUES) FOR CYTIDINE AND ITS CaCl 2 COMPLEX
Carbon atom
Tl.cytidine a (s)
Tl.cytidine. CaC12 b (s)
C4 C2 C6 C5
2.05 1.02 0.19 0.16
0.58 1.05 0.20 0.15
C1' C 4' C 2' C 3' C5'
0.28 0.15 0.20 0.20 0.10
0.13 0.16 0.20 0.17 0.09
a 0.5 mol/1 cytidine. b 0 . 5 m o l / l c y t i d i n e , 0 . 5 m o l f l CaC12.
those of carbons having one proton in the third group; this shorter relaxation time can be explained by Eqn. 1, in which the number o f protons directly attached to the carbon is t w o for Cs' and one for the others. This quantitative agreement reconfirms that the main relaxation path of 13C is the dipolar interaction. Furthermore, the fact that the observed T1 is inversly proportional to the number of the attached protons involves a similarity of correlation times for the different carbons of the cytidine molecule. On the basis of this fact and the considerations presented above, one can conclude that random motion of the molecule can be described in terms of single correlation time, rR, at least as a good approximation. This conclusion is consistent with an assumption of the isotropic rotation of the bulk molecule. It was also derived, from the fact that the different carbons of the base and the ribose in the third group have the same T1 values, that there is no intermolecular base stacking. This is because the stacking provides an additional relaxation path and the T1 values of the base carbons would therefore be different from those of the sugar [6], in the case o f stacking between cytidine bases. Using Eqn. 1, and the observed value of 0.17 s for the TI of a base carbon (the average TI values for the carbons at positions 5 and 6 in free cytidine), one obtains rR = 1.5 • 10 -1° (S). This value agrees well with the value reported for the TR of a similar molecule like 2',3'isopropylideneadenosine [11]. Inserting the above value for 7R, a measured viscosity, ~ = 0.039 g/cm per s, of the Me2SO solution and the temperature into the well-known formula [7] : rR = 4 ~ a 3 / 3 k T
(2)
the radius of a cytidine molecule, which is approximated as a sphere, can be estimated as 3.34 A. This approximated radius agrees well with the value, 3.4 A, obtained from a molecular model. It was found that addition o f CaC12 salt to an Me2SO solution of cytidine reduces the relaxation time TI of ~3C at carbon position 4, which is attached to the amino group, to nearly a quarter. Further, it was found that observed changes of TI values of the other carbons of
352 the base remain within experimental error and are actually unvaried (Table I). This fact indicates t h a t metal salts interact mainly with the amino group and/or the nitrogen atom at position 3. The above results are compatible with the 1HNMR chemical shift variation induced by addition of the metal salts [4].
2. IH spin-lattice relaxation It has been reasonably assumed that a spin-lattice relaxation time of a proton is determined by the dipolar interaction with neighboring protons which belong to either of the two molecules. Moreover, the spin-lattice relaxation of a proton in our experiments can be approximated between extremely narrow limits and T1 is represented by the following equation [12]
p:~h:74Hrc~j
rpj-6
(3)
where rpj is the inter-proton distance between the protons j and p, re is the correlation time o f the molecular motion involved in this interaction, and 7H is the gyromagnetic ratio of 1H. Since the amino group has an additional freedom of rotation about the C-N axis, plus an overall reorientation of a molecule, TI of the amino protons is determined with an effective correlation time r¢ expressed by the following equation [6]. r; 1 " r~ 1 + r['
(4)
where rR is a correlation time for overall reorientation of a molecule and r, is that of an internal m o t i o n of a group attached to a skelton undergoing random reorientation. Therefore, the T1 of the amino protons is affected by the internal motion of the group as well as the overall motion of the molecule. The effective correlation time rc was determined to be 0.35 • 10 -1° (s) by Eqn. 3, in which the observed value of T, and the separation between the amino protons had been inserted. On the other hand, the correlation time of the overall reorientation of a molecule in Eqn. 4 is reasonably assumed to be identical with the correlation time determined from the measured T, values of the carbons of the cytidine base, so far as the C-H bond is rigid. Then, using 1.5 • 10 -1° (s) and 0 . 3 5 . 1 0 - 1 ° (s) for rR and rc in Eqn. 4, respectively, one obtained 4.6 • 10 -1' (s) for ri, in the case of metal-free cytidine. The value obtained for ri is much shorter than rR. This shorter correlation time for the amino group suggests that the amino group undergoes a kind of internal rotation in the Me2SO solution of metal-free cytidine. It was f o u n d t h a t the relaxation times in the amino protons were reduced by addition of the salts (CaC12, ZnC12) (See Table II). Two plausible explanations will be put forward for the observed reduction of T1 in complexing of cytidine. (1) No additional path for the relaxation is assumed and the dipolar interaction still governs the relaxation even in the case of complexed cytidine. Based on this assumption, the correlation times of the internal rotation of the amino group, ri, were determined from the observed TI values of the complexed cytidine by the m e t h o d similar to t h a t given above. The correlation time obtained in such a way is 2.0 • 10 -9 (s) and 1.1 • 10 -1° (s) for cytidine • CaC12 and cyti-
353 T A B L E II 1 H S P I N - L A T T I C E R E L A X A T I O N T I M E S ( T 1 V A L U E S ) F O R C Y T I D I N E A N D I T S CaCI 2 A N D Z n C l 2 C O M P L E X E S IN [ 2 H 5 ] D I M E T H Y L S U L F O X I D E
Hydrogen atom
Tl.cytidine a
Tl.Cytidine. CaCI 2 b
Tl.cytidine. ZnCl 2 c
NH 2 H6
0.40 0.45
0.10 0.45
0.22 0.35
a 0 . 1 m o l / l cytidine. b 0.1 m o l ] l c y t i d i n e , 0 . 3 m o l / l CaC12. c 0.1 m o l / l c y t i d i n e , 0 . 3 tool/1 Z n C I 2.
dine. ZnCl2, respectively. The values obtained for ri of the complexes are larger than that of the metal-free cytidine, and ri in these cases are either nearly same as rR or longer. This means that the internal rotation of the amino group is retarded and probably stops in the molecular frame by addition of the metal salts. This hinderance of the internal motion of the amino group strongly suggests that a kind of complexing, presumably a metal-ion complex, occurs at the amino group. This plausible conclusion is consistent with the result derived from the 13C relaxation mentioned in the last section, as well as with the observation of separate peaks for the amino protons in the cytidine complex. [4] Recently, Raszka et al. [13] reported that the amino group for CMP in water at neutral pH exhibited a restricted rotation at room temperature. They observed that the NMR peak from the amino group appeared to be split into two peaks at lower temperatures, such as --5°C. We also found that similar separation occurs for cytidine in water at neutral pH while no separation was observed in Me2SO solution. (2) The additives may provide an additional path for the proton relaxation, for example, an electric-quadripolar interaction of an anion, CI-, coordinated to the amino group. In the previous paper [5] we reported that the anion interacts mainly with the amino protons i n cytidine and, therefore, an extra relaxation path for the amino protons through a rapid quadripolar relaxation of coordinated C1- cannot be denied. Through either one of the two mechanisms mentioned above a restriction of the amino group rotation may be caused. The relaxation time of the proton at the H 6 position is considered to reflect the conformation around the glycosidic bond. The relaxation time of the H 6 proton is assumed to be expressed by the following equation: ( ~1 ) H ~ = ( -~11)H~'Hs + ~ ( ~1 ) H~-(ribose protons)
(5)
The distance between the H 6 and H s protons is 2.44 A, obtained from X-ray crystallographic data [14] in cytidine and the distance between H e and the ribose proton (rA) can be estimated from the nuclear Overhauser effect (NOE) measurement by using the following relation [15] where the coefficent A is 1.8 1
NOE
.-- 2 + A r 6
(6)
354 • 10 -2 and the value of NOE * is obtained from the results of the nuclear Overhauser effect [4,16]. The distance between H 6 and one of the ribose protons is estimated to be 2.65 A through Eqn. 6 and the observed NOE value. From Eqn. 5 and the distances mentioned above, the additional relaxation rates due to the couplings with the ribose protons were evaluated, and therefore the total relaxation rate determined by Eqn. 5 was estimated as 0.89 (s) for H 6. This value agrees fairly well with the experimental value 0.45 (s). Consequently, the results of both the nuclear Overhauser effect and T~ of the H 6 indicate that the cytidine in the metal-free system takes the intermediate conformation, which was proposed by the authors in a former report [4]. By examing Table I, one can claim that addition of CaCI2 reduces appreciably the relaxation time of C1' while the Tt values of the other carbons are virtually unchanged, except for C4 which was discussed in the last section. It was reported that the nuclear Overhauser effect was observed between H 6 and H l' only in cytidine • CaC12 and n o t in metal-free cytidine, and the observed changes in the nuclear Overhauser effect were ascribed to the conformational change induced by complexing of cytidine [4]. This nuclear Overhauser effect indicates an additional relaxation path for the 1' proton through this dipole-dipole coupling with H 6. This additional relaxation path, which appears only in the cytidine- CaC12 system, may give rise to an appreciable increase of the relaxation of ~3C at the 1' site in the case of cytidine • CaC12. Based on this argument, one can safely conclude that the observed reduction in T1 of C~' in the case o f cytidine • CaC12 supports the conformational change around the glycosidic bond induced by this addition of the metal salt.
3. VLi relaxation Concerning the relaxation mechanism [17] of the 7Li nucleus in an aqueous solution, one must consider t w o mechanisms, the magnetic dipolar and the electric quadripolar. To elucidate a direct effect of the metal ion on cytidine, 7Li relaxation experiments in LiC1 and LiCI • cytidine aqueous solutions were performed. The relaxation times obtained for LiC1 (1 M) and LiC1 (1 M)cytidine (0.5 M) are 19 (s) and 12.5 (s), respectively. The appreciable change of the 7Li relaxation time reveals that there exists a direct interaction between the metal ion (Li ÷) and cytidine. References 1 Eichhorn, G.L. (1973) Inorganic Biochemistry (Eichhorn, G.L., ed.), Chapts 33 and 34, Elsevier, Amsterdam 2 Sohma, J., Shimokawa, S. and Hotta, K (1968) R e c e n t D e v e l o p m e n t of Magnetic Resonance in Biological Systems (Fujiwara, S. and PIette, L.H., eds.), pp. 57--69, Hirokawa Publishing Co., T o k y o 3 Shimokawa, S., Fukui, H., Sohma, J. and Hotta, K. (1973) J. Am. Chem. Soc. 95, 1 7 7 7 - - 1 7 8 2 4 Y okono, T., Shlmokawa, S., Fukui, H. and Sohma0 J. (1973) J. Chem. Soc. Japan, Chem. I n d u c t . Chem. 2, 201--206 5 Y o k o n o , T,, Shimokawa, S. and Sohma, J. (1975) J. Am. Chem. Soc. 97, 3 8 2 7 - - 3 8 2 9 6 Anerhand, A.+ Doddrell, D. and Komoroski~ R. (1971) J. C h e m . Phys. 55, 189--197 7 Fedaxko, M.C. (1 973) J. Mag. Res. 12, 30--35 8 Abragam, (1961) The Principles of Nuclear Magnetic Resonance, pp. 64 Oxford, L o n d o n * The nuclear Overhauser effect in Eqn. 6 is d e f i n e d as (I--Io)/Io, w h e r e I 0 is the NMR i nt e ns i t y w i t h o u t r-f irradiation and I is t h e same q u a n t i t y with r-f irradiation.
355 9 Hertz, H.G. (1967) Progress in Nuclear Magnetic Resonance Spectroscopy (Ems/ey, J.W., Feeney, J. and Sutcliffe, L.H., eds.), VoL 3, Chapt. 5, Pergamon Press, London 10 Jones, A.J., Grant, D.M., Winkley, M.W. and Robbins, R.K. (1970) J. Amer. Chem. Soc. 92, 4079-4087 I I Akasaka, K., Imoto, T. and Hatano, H. (1973) Chem. Phys. Lett. 2 1 , 3 9 8 - - 4 0 0 12 Solomon, I. (1955) Phys. Rev. 99, 126--131 13 Raarka, M. and Kaplan, N.O. (1972) Proc. Natl. Acad. Scl. U.S.A. 69, 2075 14 Furberg, S., Petersen, C.S. and Romming, C. (1965) Acta. Crystalogr. 18, 313--320 15 Bell, R.A. and Saunders, J.K. (1970) Can. J. Chem. 48, 1114--1122 16 Hart, P.A. and Davis, J.P. (1969) Biochem. Biophys. Res. Commun. 34, 733--739 17 Woessner, D.E., Snowden, Jr., B.S. and Ostroff, A.G. (1968) J. Chem. Phys. 4 9 , 3 7 1 - - 3 7 5