J. Phys. Chem. 1980, 84, 911-913
911
Ultrasonic: Relaxation and 'H NMR Evidence Demonstrating That Urea Destacks 6-Methylpurine Aggregates Paul Hemmes, Leslie Oppenheimer, Robert Rhlnesmith, Gloria Anderle, David Saar, and Frank Jordan Department of Chemistry, Rutgers University, The State University of New Jersey, Newark, New Jersey 07102 (Received July 18, 1979)
The ultrasonic relaxation behavior of 0.075 M 6-methylpurinewas investigated in detail in water-urea mixtures ranging from 0 to 7 M urea. The concentration-dependent ultrasonic relaxation, presumably due to the intermolecular stacking association process, decreases in magnitude but occurs at nearly the same frequency in the entire range of solvent mixtures employed. Since it was previously shown that the glycosyl isomerization and stacking processes are coupled in nucleosides and nucleotides, it is proposed that 7 M urea be employed in demonstration of intrinsic glycosyl isomerization kinetics and thermodynamics. 'H NMR studies support the ultrasonic relaxation results.
Introduction While the importance of both stacking interactions and syn-anti glycosyl conformational preferences to the maintenance of nucleic acid aggregates has been well recognized, it is only recently that clear evidence for the interdependence of these two phenomena has been provided by nuclear magnetic re~onancel-~ and ultrasonic relaxation spectroscopic4~5measurements. Base stackiing in water has been detected by ultrasonic relaxation spectroscopy for NG,Ng-dimethyladenine,6 NGfl-dimethyladenosine,6and 6-methylp~rine.~~~ In order to uncouple th,e syn-anti conformational equilibrium from the stacking equilibrium to study the former process alone, we have found that urea effectively destacks nucleosides and nucleotides. To establish a well-defined protocol for the quantitative aspects of this destacking, we have studied the ultrasonic relaxation behavior of 6-methylpurine (a compound only capable of stacking not of hydrogen bonding and of course not of glycosyl equilibration) in water-urea mixtures. Our results clearly point to the fact that urea effectively destacks 6-methylpurine and we suggest possiblle mechanisms whereby this destacking is accomplished. We have also performed 'H NMR experiments in 2Hz0-[2H4]ureamixtures. Both methods lead to similar conclusions. Materials and Methods 6-Methylpurine was purchased from Aldrich Chemical Co. Purine solutions (0.075 M) were made up in distilled, deionized, degassed (by sonication to remove the bubbles which would interfere with the ultrasonic measurements) water to which high-purity urea was added to make up the appropriate molar concentration solutions. A swept frequency interferometer (resonator) was employed in the ultrasonic measurements at 1-40 MHz. The instrumentation resembled the one described by Eggers and FunckagAccurate frequency measurement was made by Tektronix IlC503 frequency counter. Data collection and analysis were accomplished by interfacing a Tektronix 31 programmable calculator to the PSM-5 via a Tektronix 153 instrument interface. Ultrasonic relaxations were evaluated by a template technique followed by numerical curve fitting. The curves are single relaxations within the experimental error. The lH NMR measurements were performed at 28 "C a t 100 MHz on a JOEL PS FT-100 instrument on approximately 0.0'75 M 6-methylpurine dissolved in mixtures 0022-3654/80/2084-09 11$01.OO/O
TABLE I: Ultrasonic Parameters for 0.075 M 6-Methylpurine in Water-Urea Mixtures [urea], M 0.0 0.5 1.0 2.0 3.0 4.0 6.7
104pmax relaxation freq, MHz
2.0 1.68 1.55 1.35 0.97 0.78 0.38
-
35i 2 35k 2 35 ?: 2 35+ 2 35+ 2 35 f. 2 35+ 2
of 2H20-[2H4]urea(both purchased from Stohler Isotlope Chemicals, Rutherford, NJ). In general 10 transients gave noise-free spectra. The chemical shifts were recorded against sodium 3-trimethyl~ilyl[2,2,3,3-~H~]propana1ate (TSP) and are reproducible to izO.003 ppm.
Results and Discussion Table I summarizes the ultrasonic relaxation data. Figures 1and 2 show the dependence of pmax(ultrasonic) and chemical shift changes in 6-methylpurine solutions, as a function of urea concentration. Both sets of data indicate asymptotic behavior, Le., at approximately 6-7 M urea the ultrasonic results approach p = 0.2 X 104-0.3 X lo4 (about 10% of the value found in H,O); the chemical shifts also approach a limiting value. Under these conditions the only associative mechanism available to 6-methylpurine is stacking. Therefore, the ultrasonic data indicating a gradual disappearance of a relaxation must pertain to the stacking phenomenon. The upfield shift accompanyiing stacking of nucleic bases has been known for many years. Therefore, the downfield shifts accompanying urea addition can be attributed to destacking. The question of the origin of the observed destacking by urea is not settled. It is not impossible, however, that urea specifically associates with the purine base and thereby causes destacking. In support of this hypothesis, one can analyze the chemical shift data as representing formation of a 1:l complex between urea and 6-methylpurine. When the variable concentration of urea is at letst 10 times that of 6-methylpurine, a Benesi-Hildebrand type plot can be applied ( l / A vs. l/[urea]) as indicated for the various proton resonances in Figure 3. Clearly the plots are all linear, implying a K association -2 M-' (28 "(2). Another observation suggesting that urea destacks by specific interactions with the nucleic base is that 8 M LiCl, an often employed denaturant, had little effect on the pa 0 1980 American Chemical Society
912
The Journal of Physical Chemistry, Vol. 84, No. 8, 1980
Hemmes et al.
1.5
1.0
Figure 3. Benesi-Hildebrand double reciprocal plots. Xintercept leads to an association constant K = 2.0.
0.5
I 6
4
2
0
Figure 1. Plot of excess sound absorption per wavelength, pmx,vs. urea concentration for 0.075 M 6-methylpurine. C h e m i c a l Shifts o f b-Methyl-Purlne
ln
DZO 11'
m
1
6-Yethyl-Purine 1.0
0.20. C8H
0.18
Cli,
0.14.
-
c211
0.10 , 0.08
..
0.06. 0.04
0.1
C T (M)
0.L
Figure 4. Plot of T-' vs. molar concentration of 6-methylpurine in 3 M urea (squares) and water (circles).
0.16.
0.12
I
,
0.02 0.0
1.0
2.0
3
n
4.0
5.0
6.0
7.0
Figure 2. Chemical shifts of 6-methylpurine protons vs. urea concentration.
lower the rate of stack formation. This follows from the fact that the extent of stacking decreases (that is the stack formation constant is lower) but the reverse rate constant is the same, therefore the forward constant is necessarily smaller. There does not seem to be a viscosity effect since 8 M LiCl showed no similar effect on f,. Furthermore, the lack of dependence off, upon high concentrations of LiCl shows that no free water molecules are produced when 6-methylpurine stacks. If free water molecules were released upon stacking, the reaction could be represented as M, + M1
y kf
HzO
%+I+
observed in the ultrasonic spectrum. In addition to the above measurements we carried out a limited rate study on the stacking process in 3 M urea. Concentrations of 6-methylpurine ranged from 0.075 to 0.020 M. The data are shown in Figure 4 along with aqueous data from ref 7. Squares on our data represent an uncertainty of A2 MHz in the relaxation frequency. No error limits were given for the aqueous data. The interpretation of the slope and intercept of these plots is dependent on the model employed. For isodesmic or modified isodesmic models it was shown by Garland and Pate17 that the intercept of the plot in Figure 4 is
where M, and M1 represent n-mer and monomer, respectively. Then h, would be a pseudo-first-order rate constant and would be equal to k , k,'aH where k,' is a bimolecular rate constant and uHzO is h e activity of water. In LiCl and urea solutions the activity of water is markedly different from unity and thus a lowering of the rate of dissociation would be expected. No such effect is observed within experimental error. The independence of k, on urea concentration can be explained if the transition state for stacking resembles the product and we assume that addition of urea lowers the free energy of the monomers. This is in agreement with our previous contention that specific complex formation occurs between urea and the base.
with the kD the dissociation rate constant for the stack. Since kD is solvent independent within the error limits we can conclude that the major kinetic effect of urea is to
Acknowledgment. We are especially thankful to the General Medical Science Institute of the National Institute of Health (GM 19338) for financial support of this work.
+
J. Phys. Chem. 1980, 84, 913-916
References and Notes (1) Zens, A. P.;Bryson, T. A.; Dunlap, R. B.; Fisher, R. R.; Ellis, P. D. J. Am. Chem. Soc. 1976, 98, 7559. (2) Lee, C-H.; Ezra, F. S.;Kondo, N. S.;Sarma, R. H.; Danyluk, S. S. Biochemistrv 1976. 15.3627. (3) Chachav, C.; Yokono, T.; Son, T-D.; Guschlbauer,W.Biophys. Chem. 1977, 6 , 161.
913
(4) Hemmes, P. R.; Oppenheimer. L.; Jordan, F. J. Chem. Soc., Chem. Commun. 1976.,929-930. (5) Hemmes, P. R.; Oppenheimer, L.; Jordan, F., unpublished results. (6) Porschke, D.;Eggers, F. fur. J . Biochem. 1972, 26, 490. (7) Garland, F.; Patel, R. C. J . Phys. Chem. 1974, 78, 848. (8) Garland, F.; Christian, S. D. J . Phys. Chem. 1975, 79, 1247'. (9) Eggers, F.;Funck, Th. Rev. Sci. Instrum. 1973, 44, 969.
Concentration-Dependent Proton Magnetic Cross Relaxation in Aqueous Polyoxyethiylene Solutions B. Benko, V. Buljan, and S. Vuk-Pavlovi6" Macromolecular Biophysics Laboratory, Institute of Immunology, Rockefellerova 2, 4 1000 Zagreb, Yugoslavia (Received September 7, 1979)
Longitudinal proton magnetic relaxation rates were measured in aqueous solutions of polyoxyethylene (POE) 400 as a function of the polymer concentration. At all concentrations the plot of reduced magnetization vs. pulse separation in the 180°-t-900 sequence yields a nonlinear line which is easily separated into two relaxation processes. Both relaxation rates increase faster with the concentration above 50% POE than below. Above this concentration, the fractions of z magnetization ascribed to each of the two relaxation phases deviate from the values expected from the chemical composition of the solution. This demonstrates the transfer of magnetization between these phases, Le., the cross-relaxation phenomenon. The relaxation rates of POE and water protons in the absence of cross relaxation, the cross-relaxation rates, and the normalized intensities of the relaxation components for each of the relaxation phases were computed according to Edzes and Samulski after modifications required by the specificities of the POE-water system. According to the present data, along with those from literature, it is concluded that polymer ordering within the aqueous solution takes place above 50% POE, providing thus a structural basis for the apparent cross-relaxation phenomenon.
Early proton magnetic resonance studies of aqueous solutions of polyoxyethylenes (POE) revealed that the spin-lattice magnetic relaxation rate of water protons was significantly enhanced when POE was introduced as a so1ute.l Using the saturation recovery method, Liu and Ullmanl were able to measure separately the values of the relaxation times T1 of the polymer and water protons. They noted that polymer protons relaxed more efficiently than water protons in solutions of about 10% POE and elaborated the underlying physical mechanisms of proton magnetic relax'$1t'ion. In this paper we present the results of a systematic proton magnetic relaxation study of POE concentrated solutions in water. Using the pulsed nuclear magnetic resonance technique, we confirmed the existence of two proton relaxation rates in POE solutions. The analysis of the concentration dependence of the relaxation parameters in these solutions demonstrated that magnetic cross relaxation between water and polymer protons becomes prominent above 50% POE. We were able to apply the cross-relaxation modeY4 and to solve the unknown parameters without assumptions. One of the computed parameters could also be obtained experimentally and the close agreement between the computed and experimental data verifies the applicability of the model. The magnetic cross relaxation detected above 50% polymer indicates the concentration dependent ordering in the solution, a conclusion in agreement both with the NMR results and with those of other technique^.^
Experimental Section Materials and' Methods. Polyoxyethylene of molecular weight 400 (FishLer Scientific) was a gift of Dr. K. C. Ingham, Blood Research Laboratories, American National 0022-3654/80/2084-09 13$01.OO/O
Red Cross, Bethesda, Md. It was used without further purification. Polyoxyethylene was dissolved in 10 mM phosphate-1 mM ethylenediaminetetraacetic acid, pH 7.0, in lH20 or 2H20(99.9%; Prochem, B.O.C.) and the concentration was expressed as weight percents. POE concentrations in 2Hz0 were expressed as m p o ~[/ m P 0 + ~ m2HzOp'HzO/p2Hz0] (here m represents the mass and p the density of the respective component), in order to make data commensurate with those in lH,O. The proton longitudinal (spin-lattice) magnetic rehxation times, T1, were measured by the 180°-t-900 pulse sequence at 24 MHz, using a pulsed NMR apparatus with a digital readout (Xoief Stefan Institute, Ljubljana, Yugoslavia) equipped with a homonuclear pulsed field-frequency lock. The values of the amplitude of the free-induction decay were read 200 ps following the 90° pulse. A Bruker BE4OAll high-resolution magnet was used. Tlhe temperature of the samples was kept at 20 f 0.5 "C. For each T I measurement at least 30 data points (diifferent delays between 180 and 90' pulses) were taken. A graph of the logarithms of reduced magnetization m = -[M(t) - M , ] / 2 M , as a function of t was then constructed ( M , and M ( t ) denote the values of the projection of macroscopic magnetization on the z axis at equilibrium and at time t , respectively, measured as the amplitudes of the free-induction decay following the 90° pulse). Relaxation parameters were evaluated graphically and by a computer program written for a Hewlett-Packard H P 9820A computer.
Theory Due to the large difference between Tz*values of water protons and protons of the solute in some hydrated mac0 1980 American Chemical Society
