Internal Dynamics of tRNA Dynamic Light Scattering
Studied by Depolarized
A. PATKOWSKI,' W. EIMER,' and TH. DORFMULLER2 'Molrc L i l x Biophysics Laboratory, Institute of Physics, A. Mickiewicz University, Grunwaldzka 6, 60-780 Poznan, PoLInd, and 'Faculty of Chemistry, University of Bielefeld, D-4800 Bielefeld 1, Federal Republic of Germany
SYNOPSIS
The collective internal dynamics of transfer RNAPhefrom brewer's yeast in solution was studied by depolarized dynamic light scattering (DDLS). Within the melting region of tRNA the depolarized spectra consist of two Lorentzian, where the narrow (slow) component describes the overall rotation of the macromolecule. The broad component is attributed to the collective reorientation of the bases within the biopolymer. At high temperature only this relaxation process is observed in the spectrum. The viscosity dependence of the collective internal relaxation process is described by the Stokes-Einstein-Debye equation for rotational diffusion. Estimates of the internal orientational pair correlation factor from the integral depolarized intensities of tRNAPhesolutions indicates that the observed dynamics correspond to the collective reorientation of approximately 5 bases. A comparison of the results presented with DDLS studies on the aggregation of the mononucleotide guanosine5'-monophosphate confirms this result. For a further characterization of the relaxation process we studied the effect of hydrostatic pressure (1-1000 bar) on the depolarized spectra of tRNA. While other spectroscopic methods like nmr, fluorescence polarization anisotropy decay, or ESR give information about the very local motion of a single base within the DNA or RNA, this study shows that by DDLS one can characterize collective internal motions of macromolecules.
INTRODUCTION The x-ray crystallography provides information about the static structure of biopolymers that are averaged over the time of the experiment and the population of macromolecules in the crystal.' There is, however, growing experimental and theoretical evidence on the internal dynamics of proteins and nucleic The biological function of proteins and nucleic acids, and especially the variety of functions of transfer ribonucleic acids ( t R N A ) , cannot be understood on the basis of the static x-ray structures alone. Therefore, it is essential to have a detailed picture of the internal dynamics of the biopolymers. In this respect x-ray crystallography is not very helpful unless high-intensity x-ray sources
IC' 1990 John Wiley & Sons, Inc. CCC 0006-3525/90/9-10975-09 $04.00 Biopolymers, Vol. 30,975-983 (1990)
and new detection techniques with a time resolution in the nano- and picosecond range are developed. Since such techniques are not available, we have to study internal dynamics of biopolymers using spectroscopic methods such as ESR, fluorescence polarization anisotropy decay ( F P A ), nmr, and depolarized dynamic light scattering ( DDLS ) . In general, biological macromolecules exhibit a complex dynamic behavior t h a t covers a broad time range from lop1' to 10' s. Each of the various spectroscopic techniques has its specific sensitivity to different relaxation modes. This has to be kept in mind if one compares results from the different methods, but it also bears the chance to use complementary techniques for a more detailed analysis of the complicated internal dynamics of DNA and RNA. Besides that, each technique has its specific experimental limitations. FPA and ESR require the use of a label and, actually, the dynamics being measured is the dynamics of the probe molecule. Together with the various nmr relaxation methods, 975
976
PATKOWSKI, EIMER, AND DORFMULLER
ESR and FPA are similar in that the data analysis is complicated and relies heavily on the dynamic models being used. DDLS, as nmr, has the advantage that it is a nonperturbing method. For small macromolecules the analysis of the spectra is straightforward. T h e difference of DDLS compared to all the other methods is that it provides information about static and dynamic correlations between the molecules. This might be a n advantage-but also a disadvantage-depending on the problem focused on. The term "internal motion" is widely used in the literature for such different relaxation processes as the motion of a n individual subgroup within a macromolecule as well as for long-range collective motions. In recent years, nmr,'-12 ESR,13-17 and FPA 18-20 have been extensively used t o characterize the very local internal dynamics of the bases-e.g., the wobbling or librational motion of a single base. The results concerning the magnitude of the internal motion have been controversial.21f22 The experiments were mostly carried out on longer DNA and RNA fragments of various lengths. Besides the fact that the system studied was often polydisperse, the long oligonucleotides exhibit a rather complex relaxation mechanism. This complicates the interpretation of the experimental data. Neglecting the collective twisting and bending motions of the macromolecule, the amplitude of the librational motion was estimated to about 30" and in some cases even larger values. The relaxation times for the motion were calculated t o some nanoseconds. Early ESR studies on the dynamics of spin labels in tRNAVa',I5tRNAPhe,I6and tRNATy' li gave correlation times in the range of 1-2.4 ns and a n activation energy of E, = 41.8 kJ/molI7 for the relaxation process a t high temperature. Assa-Munt et a1.l' studied the internal dynamics of poly ( dA-dT) by means of proton nmr. They obtained a good fit t o the experimental data assuming a three-state jump model with angular displacement of the bases of k32" relative to the helix axes and a correlation time T = 0.7 ns. Mirau et a1.l' studied the dynamics of duplexes of the B- and Z-DNA form of poly(dGd C ) poly( dG-dC) by means of proton nmr. They were able to fit the experimental data with two models: ( a ) assuming out-of-plane wobbling of the bases with angular amplitudes of +30° and a correlation time of 1 ns, and ( b ) assuming decoupled longitudinal (out of plane) and azimuthal (in plane ) local motions of the bases. Within this model the amplitude for the longitudinal motion was estimated t o 25" for B-DNA and 32" for Z-DNA, respectively, with a relaxation time in the range of 1-100 ns. For
the azimuthal motion they assumed a jump amplitude of 30"-90" and a relaxation time in the range of 0.1-1.5 ns. Kao and BobstI4 investigated local base motions in spin-labeled polyribo- and polydeoxyribonucleotides by means of ESR. For singlestranded RNA and DNA they found a local relaxation time (for a three-state jump model) amounting to 1.2 ns. This relaxation time increased to 4.0 ns for double-helical RNA and DNA. In a previous paper we studied the dynamics of the bases in unfolded calf thymus DNA by DDLS23and obtained a n internal dynamics relaxation time of 1.5 ns a t 97". On the other hand, time resolved FPA experiments2' on DNA show a very rapid decay in the range of some picoseconds. Furthermore, a reanalysis of nmr results," based on a model that takes into account long-range fluctuations, has indicated that the amplitudes for the wobbling motion are much smaller ( N 10"-20" ) and probably much faster ( < 100 p s ) than originally believed. A recent combined DDLS and nmr relaxation study24 of very short oligonucleotides, where collective motions are less important, confirms these results. The long-range collective motions of DNA have been studied by dynamic light s ~ a t t e r i n g ' ~and .~~ transient electric b i r e f r i n g e n ~ e .Recent ~~ developments in DNA preparation techniques make monodisperse samples available. Therefore, DNA has become a very useful system t o test various dynamic models. But DNA and RNA are also interesting macromolecules by themselves, because information about the flexibility and collective internal dynamics is important in order to understand the role of conformational changes for biological processes. In a separate paper2' we reported results from DDLS on the overall reorientation of tRNAPhe. Above the melting temperature, we observed a second, much faster relaxation process. Here we will analyze this broad component in the DDLS spectrum in more detail, in order t o determine its origin. We will show that the measured relaxation times as well as the calculated hydrodynamic volume must be attributed to the collective reorientation of the bases and do not reflect the very local librational motion of a single base. Due to the coherent character of DDLS, the signal is strongly determined by static orientational pair correlations between the bases. Therefore, the measured dynamics reflect the correlated rather than the uncorrelated motion of the scattering units. Thus, DDLS is a complementary method to nmr, ESR, and FPA, and allows us to study the more collective internal dynamics of polymers consisting of optically anisotropic subunits.
INTERNAL DYNAMICS OF tRNAPhe
THEORY The spectrum IVH ( K , w ) measured in a DDLS experiment from a dilute solution of a polymer that consists of optically anisotropic monomer units has been theoretically calculated by Pecora:
977
If static correlations between monomers have to be taken into account, Bauer e t al. have shown that the collective internal relaxation time rFntis given by31
29330
I w ( K , 01
=-
A 27rp
s
exp(-iwt)S,,(K,
t ) dt
(1)
where A is a constant, p the number density of the polymer, S,, ( K , t ) the correlation function, and K the scattering vector K = [ ( 4 x n )/ A ] sin ( 8 / 2 ) . Here n is the refractive index of the solvent, X the wavelength of the incident light, and 8 is the scattering angle. For a polymer chain of n monomer units, the correlation function is given by3's3'
where F is the static internal correlation function. If the collective reorientation of the bases is diffusional, one can assume that the dependence of rotational relaxation time 7 Fnt on temperature T and solvent viscosity 77 can be described by the StokesEinstein-Debye ( SED ) equation,32 (7) where V, is the hydrodynamic volume of the relaxing unit, kB the Boltzmann constant, and r,, a constant.
MATERIALS A N D METHODS where of,( t ) is the y z component of the laboratoryfixed polarizability a t time t and r 1is the position of the i t h monomer unit. T h e monomer units i and j belong to the same chain. The K-dependent term can be ignored since its relaxation is much slower than the relaxation of a;, ( t ). For a rigid polymer, i.e., when the relaxation of monomer units within the polymer is much slower than the overall rotation of the macromolecule or if the internal motion does not change oYzsignificantly,
S,,(t)
N
exp(-6DRt),
(3)
where L)' is the rotational diffusion coefficient of the macromolecule. If the monomer units can reorient independently on a time scale much faster than the overall rotation of the macromolecule, then
where r f n tis the internal dynamics relaxation time of the monomer units within the polymer. In this case the spectrum consists of a single Lorentzian,
Transfer RNA from brewer's yeast ( BoehringerMannheim) was used without further purification. A tRNA solution with a concentration of 50 mg/ mL in 10 m M Tris HC1, 0.2 AT NaC1, pH 7.2, buffer was prepared. Before measurements the tRNA solutions were filtered into dust-free light scattering cells through Millipore filters (pore size 0.22 p m ) and centrifuged for a t least half a n hour a t 2000 X g in order to remove air bubbles. Two kinds of light scattering cells were used: a t atmospheric pressure a 4 X 4 X 40 mm fluorescence cell (Hellma) and a t high pressure a round cell sealed with a teflon piston and O-rings. A single mode argon ion laser (Spectra Physics, Model 165) with a n output power between 200 and 400 mW was used a s a light source. The VH component of scattered light was selected using polarizers with extinction coefficients of and for the incident and scattered light. The spectrum of the depolarized component of scattered light was measured by a planar Fabry-Perot interferometer with a free spectral range of FSR = 3.8 GHz and a finesse of about 80, stabilized by a DAS-10 system (Burleigh Instruments). The fluorescence light from a dye present in the tRNAPhepreparation was removed by a narrow band interference filter. Due to a very low depolarized intensity, it was necessary to accumulate up to 1024 sweeps of 5 s for each spectrum. The spectrum was measured in the temperature range from 65 to 93°C while the temperature was stabilized within +.l"C.
978
PATKOWSKI, EIMER, AND DORFMULLER
High-pressure measurements were performed using a specially designed high-pressure cell33that was modified for DDLS measurements in the following way: The cell was equipped with five windows-four of them a t a n angle of 90” and the fifth one a t 45”. In order t o avoid pressure-induced birefringence of the windows, a specially selected lead glass was used. T h e windows were coated with antireflective layers.
previously by a confocal Fabry-Perot interferometer of FSR = 750 MHz. On the other hand, a t high temperatures (about 99°C) no contribution due to overall rotation in the depolarized spectrum was observed, which indicates that the fluctuations of the anisotropy of “melted” tRNA predominantly follows a n intramolecular rather than a n overall tumbling mechanism. In Figure 1 the internal relaxation time 7Fnt is plotted vs q / T , where q is the viscosity of the solvent. From the slope the hydrodynamic volume is calculated to V, = (4.7 -t 0.5) X lo3 A3. Within the observed temperature range the experimental relaxation times scale with q / T according t o the StokesEinstein-Debye equation. Therefore, it is evident that T fnt corresponds to the same relaxation process a t different temperatures. Furthermore, this motion is diffusive, determined by the viscosity of the surrounding medium. The question is, what kind of relaxation process is monitored by the fast component in the DDLS spectrum. Is it the very local librational motion of a single base or a more collective reorientation of the bases. A comparison with results from a study on the aggregation of the mononucleotide guanosine-5’m o n ~ p h o s p h a t egives ~ ~ some interesting clues to the problem. From the dynamic behavior of guanosine5‘-monophosphate it was possible to estimate the size of the single helical stacks that are formed in solution. T h e dimensions of the aggregates were calculated from the rotational and translational diffusion coefficients based on a hydrodynamic theory for cylindrically symmetric molecules. The absolute correlation times as well as the hydrodynamic volume of tRNA a t high temperature, as obtained from the fast component in the DDLS spectrum, indicate that the dynamics of “melted” tRNA is similar t o the collective reorientation comprising approximately seven to eight stacked mononucleotides. Another way to estimate the number of bases that, on the average, participate in the collective reorientation process of tRNA is to compare the integral intensities of the depolarized spectra under conditions where only the overall rotation or the internal motion is observed. As discussed in detail later, this yields a value of about 5 for the static pair correlation
RESULTS AND DISCUSSION Temperature Dependence of
7Fnt
The values of the collective internal relaxation time Tfnt for tRNAPhein the temperature range of 6593°C are listed in Table I. These relaxation times were measured above the “melting temperature” ( T, = 60°C) of tRNAPhe(see Fig. 6 in Ref. 2 8 ) . At temperatures below T, the depolarized spectrum consists of a single Lorentzian of high intensity t h a t corresponds t o the overall rotation of tRNAPhe.28 At T < T,, there was no indication for a second, broader component in the spectrum which would correspond to the internal reorientation of the bases. This may be due to the fact that either the internal dynamics of tRNAPh”below T, is much slower, or the local dynamics relax only a very small part of the optical anisotropy (very restricted rotation). Thus, the intensity connected with internal dynamics is much lower than the intensity corresponding to the overall rotation of tRNA, and therefore, the fast component does not significantly contribute to the depolarized spectra. Within the temperature range of the melting transition ( 50-70°C) both, overall rotation and internal dynamics, could be measured. The collective internal correlation times were obtained from a double Lorentzian fit to the spectrum deconvoluted with the instrumental function.34 As mentioned above, the second, much slower component is due to the overall rotation of tRNA. The spectral resolution of the planar Fabry-Perot interferometer was adjusted in such a way that the narrow line (slow relaxation process) was not resolved. However, the rotational dynamics of tRNA has been studied
Table I
Temperature Dependence of the Collective Internal Relaxation Time 7fntfor tRNAPhe
Temperature (“C) (PS)”
7:”t
a
65.5 467
The experimental error amounts to *5%.
70.2 436
75.4 416
80.6 382
85.3 345
90.0 310
92.9 307
INTERNAL DYNAMICS OF tRNAPhe
060
-
979
5
where Iy' is attributed to the overall reorientation of the molecule, while I@Aarises from the relaxation of the optical anisotropy due to local motion. Pecora and Wang3' have calculated the time correlation function for the diffusion of a rigid rod in a cone. As long as the cone angle H,, 5 50", the following approximation is valid:
t
/ 3.7 5 750 11.3 15.0
OO
q / T w lo6 [ P / K Figure 1. The collective internal relaxation time as a function of q / T .
+
1 T&
factor ( 1 F,ntra), which indicates that the collective internal reorientation comprises approximately 5 bases. Both estimates are rather crude, but nevertheless they clearly show that the relaxation process observed in the DDLS experiments is different from the very local internal motion of a single base as it is observed by other spectroscopic methods. A t low temperatures (ordered tRNA form) the internal relaxation is strongly limited by hydrogen bonding and stacking interactions between the heterocyclic bases. In such a case the internal dynamics will result in very small changes of the polarizability anisotropy and most of the optical anisotropy will relax by the overall rotation of the macromolecule. That means, most of the depolarized intensity is attributed t o the rotation of the whole molecule. For this reason we were not able to measure the internal dynamics of tRNAPhea t low temperatures. A t high temperatures (unfolded structure) the macromolecule becomes more flexible and allows collective rotat ional motion of the heterocyclic bases. This decreases the relaxation times and also leads to a loss in the static pair correlations, i.e., a lower scattering intensity. As mentioned in the introduction, it is well established that a t room temperature the single bases perform a librational motion with a rms amplitude of about 20". In the depolarized spectra of tRNAPh", however, there is no evidence for such a relaxation process below the melting temperature. If we assume that the local internal motion of the bases can be described by the restricted rotational diffusion in a cone it is possible to estimate the fraction of the component of the depolarized scattered light Zvh for this relaxation process. For a single scatterer, the total intensity is given by
where C?, C:, and v: are dependent on the cone angle, and 6 ( w ) is the Dirac 6 function. Dk is the rotational diffusion coefficient characteristic for the internal relaxation process. Within the dynamic model the orientation of the cone is fixed. In our system the whole molecule rotates, but nevertheless in a first approximation we can consider C: proportional to IYH, due to overall reorientation, and C proportional to the integrated intensity I$',!, of the local motion of a n individual base. For a cone angle of 20" the amplitude of the librational motion amounts t o approximately 20% and for Bo = 30" C:/C: is N 0.5 ( a figure of Ci/C: as a function of the cone angle do is given in Ref. 38). Thus, for a monomer unit the restricted rotational diffusion would give a significant contribution to the total integrated intensity a t a cone angle that is typical for the bases within DNA. The amplitude of the librational motion in tRNAPheshould not be very different. This estimate is correct for a single base or nucleotide, but now we have to account for the situation in a macromolecule. The integrated intensity IYH, due to the overall rotation is proportional to ( y ')
I&
x
(y 2 )
(10)
where (7') is the mean square optical anisotropy of the macromolecule
with p the optical anisotropy of the monomer units and ( 1 Fin,,,)a measure of the internal orientational pair correlation between the subunits. Therefore, the intensity ratio for tRNA is given by
+
980
PATKOWSKI, EIMER, AND DORFMULLER
For a flexible chain molecule Fintra is about 0. But in tRNA the orientation of the bases is strongly correlated and the static pair correlation function (1 Fintra) is expected to be significantly larger than zero and may amount to about 75 ( = number of bases per tRNA molecule). Therefore, the contribution of the librational motion of the individual bases to the depolarized spectra is much smaller than estimated from Eq. ( 9 ) . It has to be emphasized that the relaxation time for the local internal motion is expected to be < 250 ps. Due to the FSR of the planar Fabry-Perot interferometer used in these experiments this fast component might disappear in the baseline. On the other hand, we did not observe an unusual high background in the spectra which would have supported this possibility. An Arrhenius plot of the internal relaxation time rfntis shown in Figure 2. The best fit to the experimental data was obtained for E, = (17 2 ) k J / mol. The activation energy for the collective internal relaxation process is smaller than E, for the overall reorientation ( E , = 20.3 -t 0.4 kJ/mol) as observed by the same experimental method.28 Nevertheless, the difference is not very pronounced, and therefore it is reasonable to assume that both motions are characterized by similar relaxation mechanisms. On the contrary, ESR measurements of spin-labeled tRNAVa',17which reflect the mobility of the aminoacetyl end of tRNA molecules, show an abrupt transition in the nature of the motion at low ionic strength. The transition temperature coincides with the melting temperature measured by the optical density. The activation energy for the relaxation process at high temperature is almost twice as high as E, for T < T,. A comparison with our DDLS
+
*
OSO'lo
100 I
90 I
80
70
60
I
I
Table I1 Pressure Dependence of the Correlation Time T;,,~for tRNAPheat 90°C.
P (bar) 71t
(PSI" a
50
250
500
1000
303
379
406
560
The experimental error amounts to +5%.
results clearly indicates that both experimental methods monitor different aspects of the rather complex conformational fluctuations of tRNA. Pressure Dependence of the Internal Relaxation Time
In order to characterize the collective reorientational relaxation process in more detail, we have also investigated the effect of pressure on the internal dynamics of tRNAPhe.Measuring the internal dynamics relaxation time r Fnt provides the activation volume AV# of the relaxation process
where R is the gas constant, T the absolute temperature, and P the pressure. The internal relaxation time rFnt has been measured at pressures from 1to 1000bar at temperatures of 70 and 90°C. At 70°C no pressure effect was observed. The pressure dependence of r;ntat 90°C is given in Table I1 and plotted in Figure 3. According to Eq. ( 1 3 ) , the activation volume for the relaxation process was calculated to AV# = (30 4 ) A3. This is a very small volume, even if compared with the hydrodynamic volume of a single base. It is com-
*
50 II
0.50: -
-040 ul
-C
50.30
0.20
-
0.20 ; 250 500 750 1000 0 2.70 280 290 3.00 330 1;
2.60
Pressure [bar ]
l/T*1000 [ l / K I Figure 2. Arrhenius plot of the collective internal correlation time 7fnt.
Figure 3. 90°C.
Plot of 1n(7Fn,) vs pressure for tRNAPheat
981
INTERNAL DYNAMICS OF tRNAPhe
parable t o the activation volume AV' = 15 A3 obtained in a similar experiment for guanosine-5'm o n ~ p h o s p h a t e Since . ~ ~ the activation volume reflects the extra volume required by the rotating moiety to move from the equilibrium to a transition state, it means that this transition state requires only a small extra volume. In order to interpret the experimental activation volume one should also recall that this quantity contains a usually negative contribution from the rearrangement of the environment. We have also studied the influence of hydrostatic pressure on the overall rotation relaxation time of tRNAPh' at temperatures below T,. No effect of pressure on the measured dynamics has been observed.
where n is the number of monomer units per macromolecule and p is the optical anisotropy of the monomer. We have measured the depolarized intensities a t 2 and 9O"C, and obtained
Comparison with Other Studies For a rigidly parallel stacked polymer the pair correlation factor (1 Fin,,,) can amount to n , the The results from various experimental methods on number of monomer units in the macromolecule. the magnitude and the time constant of the local Since for tRNAPhe(1 Fin,,,)2 76, with Eq. (15) internal motion of the bases within DNA and RNA we can estimate (1 Fintra) t o I5 for the collective are still controversial. In Table I11 we compare nmr reorientational process observed at high temperature and ESR data on the internal correlation time with ( T2 90°C). our results. Comparing our numerical values of Assuming that the high temperature data can be 7Fnt with the results from other studies, one has to keep in mind that DDLS measures the collective extrapolated t o room temperature, we obtain 7Fn,(20"C) = 1.2 ns [see Eq. ( 7 ) ] , which gives relaxation time [ see Eq. ( 6 ) 1, while the other methods yield the single particle correlation time 7 % , . In 7ynt(20°C) N 0.25 ns. Our estimate for the correlaprinciple, the static pair correlation factor ( 1 Fintra) tion time of the internal motion lies a t the lower can be calculated from the integral depolarized inend of comparable data shown in Table 111. More recent studies 12,22,24 lead t o the conclusion that the tensities I,, of the monomer unit and the polymer
+
+ +
+
Table I11 Internal Dynamics of Nucleic Acids: Comparison with Other Studies
Reference
Sample
Method
Present study
tRNAPh' (yeast) tRNA""' (E. coli) tRNAPhe(E. coli) tRNATYr(yeast) DNA (calf thymus) poly(dA-dT) Single-stranded RNA Single-stranded DNA Double-helical RNA Double-helical DNA poly(dG-dC)* poly( dG-dC) (B, Z DNA) tRNAPhe
DDLS ESR ESR ESR DDLS Proton nmr ESR
15 16 17 23 10 14
11
12
Tint
v,
(20", w), ns
-
0.2 0.1-1 0.36-2.40 1.75 3 0.7 1 4
Proton nmr
0.1-1.5
13C-nmr
-
0.02
A3
(4.7
0.1-lo3)
982
PATKOWSKI, EIMER, AND DORFMULLER
time constant of the librational motion of the bases is in the range of picoseconds. The main argument against longer times, found in other studies, is that long-range twisting and bending motions of the relatively long fragment studied have been neglected in the model calculations. Moreover, nmr does not provide the time resolution for the very fast processes, but the information about the internal correlation time was extracted from relaxation processes that decay on a much slower time scale. It was shownz4that the amplitude of the librational motion can be estimated quite accurately from the nmr relaxation parameters, while this is not possible for the time constant. From all this it becomes clear that so far the dynamics of the local internal reorientation of the bases within DNA and RNA have not yet been fully characterized quantitatively by any one of the experimental methods used.
CONCLUSIONS We have studied the collective internal reorientation of the bases in tRNAPhe.Within the melting region the depolarized spectra show two clearly distinct relaxation processes, a slow component, which is attributed to the overall rotation of the macromolecule, and a broad line, which describes the collective reorientation of the bases within tRNAPhe.Well above the melting temperature only the fast relaxation motion is observed. The absolute correlation time and the experimental hydrodynamic volume indicate that the segmental motion extends over I 5 bases. Therefore, the fast relaxation process cannot be attributed to the librational motion of a single base, as observed by other spectroscopic methods. This very local motion, which is also present below the melting temperature, does not significantly contribute to the depolarized spectra. Due to the correlated static orientation of the bases within the macromolecule, it is rather the overall reorientation of the molecule that dominates Ivh below T,. We have shown that DDLS provides valuable additional information about the collective motion of the bases within tRNAPhe,which is not available by other methods. I n fact, we are now able t o describe the dynamics of tRNA on three different levels. The local one-base reorientation monitored by nmr, FPA and ESR, the collective reorientation as monitored by DDLS above the melting temperature, involving a small number of cooperatively rotating bases, and finally, the overall reorientation of the whole macromolecule.
This work was supported by the Complex Liquids Project of the ZIF (Center for Interdisciplinary Research), the Minister fur Wissenschaft und Forschung des Landes NRW, the Fonds der chemischen Industrie, and in part ( t o A P ) by the research project RP.II.13.1.10.
REFERENCES 1. Saenger, W. ( 1984) Principles of Nucleic Acid Struc-
ture, Springer-Verlag, New York. 2. Clementi, E. et al., Eds. ( 1985) Structure and Motion,
Nucleic Acids and Protein, Academic Press, New York. 3. Rigler, R. & Wintermeyer, W. (1983) Ann. Rev. Biophys. Bioeng. 12, 475-505. 4. McCammon, J. A., Gelin, B. R. & Karplus, M. ( 1977) Nature 2 6 7 , 585-590. 5. Harvey, S. C., Prabhakaran, M., Mao, B. & McCammon, J. A. (1984) Science 2 2 3 , 1189-1191. 6. Harvey, S. C., Prabhakaran, M. & McCammon, J. A. (1985) Biopolymers 2 4 , 1169-1188. 7. Prabhakaran, M., Harvey, S. C. & McCammon, J. A. (1985) Biopolymers 24, 1189-1204. 8. Early, T. A. & Kearns, D. R. ( 1979) Proc. Natl. Acad. Sci. U S A 76,4165-4169. 9. Mirau, P. A. and Kearns, D. R. (1984) Biochemistry 23, 5439-5446. 10. Assa-Munt, N., Granot, J., Behling, R. W. & Kearns, D. R. (1984) Biochemistry 23,944-955. 11. Mirau, P. A., Behling, R. W. & Kearns, D. R. (1985) Biochemistry 2 4 , 6200-6211. 12. Schmidt, P. G., Sierzputowska-Gracz, H. & Agris, P. F. (1987) Biochemistry 2 6 , 8529-8534. 13. Bobst, A. M., Kao, S.-C., Toppin, R. C., Ireland, T. C. & Thomas, I. E. (1984) J . Mol. Biol. 173, 6374. 14. Kao, S.-C. & Bobst, A. M. (1985) Biochemistry 24, 5465-5469. 15. Hoffman, B. M., Shofield, P. & Rich, A. (1969) Proc. Natl. Acad. Sci. USA 62, 1195-1202. 16. Caron, M., Brisson, N. & Dugas, H. (1976) J . Biol. Chem. 251, 1529-1530. 17. Weygand-Durasevic, I., Kruse, T. A. & Clark, B. F. C. (1981) Eur. J. Biochem. 116,59-65. 18. Barkley, M. D. & Zimm, B. H. ( 1979) J . Chem. Phys. 70, 2991-3007. 19. Millar, D. P., Robbins, R. J. & Zewail, A. H. ( 1981) J. Chem. Phys. 76, 2080-2094. 20. Magde, D., Zappala, M., Knox, W. H. & Nordlund, T. H. (1983) J . Chem. Phys. 87,3286-3288. 21. HBrd, T. (1987) Biopolymers 26, 613-618. 22. Schurr, J. M. & Fujimoto, B. S. (1988) Biopolymers 27, 1543-1569. 23. Patkowski, A., Fytas, G. & Dorfmuller, Th. (1982) Biopolymers 21, 1473-1477. 24. Eimer, W., Williamson, J. R., Boxer, S. G. & Pecora, R. ( 1990) Biochemistry 29, 799-811.
INTERNAL DYNAMICS OF tRNAPhe
25. Bloomfield, V. A. (1985) in Dynamic Light Scattering: Applications of Photon Correlation Spectroscopy, PeCora, R., Ed., Plenum, New York. 26. Sorlie, S. S. & Pecora, R. ( 1988) Macromolecules 2 1 , 1437-1449. 27. Lewis, R. J., Pecora, R. & Eden, D. ( 1986) Macromolecules 16, 134-139. 28. Patkowski, A,, Eimer, W. & Dorfmuller, Th., (1990) Riopolymers 30,93-105. 29. Pecora, R. (1968) J . Chem. Phys. 49, 1036-1043. 30. Berne, B.
Studied by Depolarized
A. PATKOWSKI,' W. EIMER,' and TH. DORFMULLER2 'Molrc L i l x Biophysics Laboratory, Institute of Physics, A. Mickiewicz University, Grunwaldzka 6, 60-780 Poznan, PoLInd, and 'Faculty of Chemistry, University of Bielefeld, D-4800 Bielefeld 1, Federal Republic of Germany
SYNOPSIS
The collective internal dynamics of transfer RNAPhefrom brewer's yeast in solution was studied by depolarized dynamic light scattering (DDLS). Within the melting region of tRNA the depolarized spectra consist of two Lorentzian, where the narrow (slow) component describes the overall rotation of the macromolecule. The broad component is attributed to the collective reorientation of the bases within the biopolymer. At high temperature only this relaxation process is observed in the spectrum. The viscosity dependence of the collective internal relaxation process is described by the Stokes-Einstein-Debye equation for rotational diffusion. Estimates of the internal orientational pair correlation factor from the integral depolarized intensities of tRNAPhesolutions indicates that the observed dynamics correspond to the collective reorientation of approximately 5 bases. A comparison of the results presented with DDLS studies on the aggregation of the mononucleotide guanosine5'-monophosphate confirms this result. For a further characterization of the relaxation process we studied the effect of hydrostatic pressure (1-1000 bar) on the depolarized spectra of tRNA. While other spectroscopic methods like nmr, fluorescence polarization anisotropy decay, or ESR give information about the very local motion of a single base within the DNA or RNA, this study shows that by DDLS one can characterize collective internal motions of macromolecules.
INTRODUCTION The x-ray crystallography provides information about the static structure of biopolymers that are averaged over the time of the experiment and the population of macromolecules in the crystal.' There is, however, growing experimental and theoretical evidence on the internal dynamics of proteins and nucleic The biological function of proteins and nucleic acids, and especially the variety of functions of transfer ribonucleic acids ( t R N A ) , cannot be understood on the basis of the static x-ray structures alone. Therefore, it is essential to have a detailed picture of the internal dynamics of the biopolymers. In this respect x-ray crystallography is not very helpful unless high-intensity x-ray sources
IC' 1990 John Wiley & Sons, Inc. CCC 0006-3525/90/9-10975-09 $04.00 Biopolymers, Vol. 30,975-983 (1990)
and new detection techniques with a time resolution in the nano- and picosecond range are developed. Since such techniques are not available, we have to study internal dynamics of biopolymers using spectroscopic methods such as ESR, fluorescence polarization anisotropy decay ( F P A ), nmr, and depolarized dynamic light scattering ( DDLS ) . In general, biological macromolecules exhibit a complex dynamic behavior t h a t covers a broad time range from lop1' to 10' s. Each of the various spectroscopic techniques has its specific sensitivity to different relaxation modes. This has to be kept in mind if one compares results from the different methods, but it also bears the chance to use complementary techniques for a more detailed analysis of the complicated internal dynamics of DNA and RNA. Besides that, each technique has its specific experimental limitations. FPA and ESR require the use of a label and, actually, the dynamics being measured is the dynamics of the probe molecule. Together with the various nmr relaxation methods, 975
976
PATKOWSKI, EIMER, AND DORFMULLER
ESR and FPA are similar in that the data analysis is complicated and relies heavily on the dynamic models being used. DDLS, as nmr, has the advantage that it is a nonperturbing method. For small macromolecules the analysis of the spectra is straightforward. T h e difference of DDLS compared to all the other methods is that it provides information about static and dynamic correlations between the molecules. This might be a n advantage-but also a disadvantage-depending on the problem focused on. The term "internal motion" is widely used in the literature for such different relaxation processes as the motion of a n individual subgroup within a macromolecule as well as for long-range collective motions. In recent years, nmr,'-12 ESR,13-17 and FPA 18-20 have been extensively used t o characterize the very local internal dynamics of the bases-e.g., the wobbling or librational motion of a single base. The results concerning the magnitude of the internal motion have been controversial.21f22 The experiments were mostly carried out on longer DNA and RNA fragments of various lengths. Besides the fact that the system studied was often polydisperse, the long oligonucleotides exhibit a rather complex relaxation mechanism. This complicates the interpretation of the experimental data. Neglecting the collective twisting and bending motions of the macromolecule, the amplitude of the librational motion was estimated to about 30" and in some cases even larger values. The relaxation times for the motion were calculated t o some nanoseconds. Early ESR studies on the dynamics of spin labels in tRNAVa',I5tRNAPhe,I6and tRNATy' li gave correlation times in the range of 1-2.4 ns and a n activation energy of E, = 41.8 kJ/molI7 for the relaxation process a t high temperature. Assa-Munt et a1.l' studied the internal dynamics of poly ( dA-dT) by means of proton nmr. They obtained a good fit t o the experimental data assuming a three-state jump model with angular displacement of the bases of k32" relative to the helix axes and a correlation time T = 0.7 ns. Mirau et a1.l' studied the dynamics of duplexes of the B- and Z-DNA form of poly(dGd C ) poly( dG-dC) by means of proton nmr. They were able to fit the experimental data with two models: ( a ) assuming out-of-plane wobbling of the bases with angular amplitudes of +30° and a correlation time of 1 ns, and ( b ) assuming decoupled longitudinal (out of plane) and azimuthal (in plane ) local motions of the bases. Within this model the amplitude for the longitudinal motion was estimated t o 25" for B-DNA and 32" for Z-DNA, respectively, with a relaxation time in the range of 1-100 ns. For
the azimuthal motion they assumed a jump amplitude of 30"-90" and a relaxation time in the range of 0.1-1.5 ns. Kao and BobstI4 investigated local base motions in spin-labeled polyribo- and polydeoxyribonucleotides by means of ESR. For singlestranded RNA and DNA they found a local relaxation time (for a three-state jump model) amounting to 1.2 ns. This relaxation time increased to 4.0 ns for double-helical RNA and DNA. In a previous paper we studied the dynamics of the bases in unfolded calf thymus DNA by DDLS23and obtained a n internal dynamics relaxation time of 1.5 ns a t 97". On the other hand, time resolved FPA experiments2' on DNA show a very rapid decay in the range of some picoseconds. Furthermore, a reanalysis of nmr results," based on a model that takes into account long-range fluctuations, has indicated that the amplitudes for the wobbling motion are much smaller ( N 10"-20" ) and probably much faster ( < 100 p s ) than originally believed. A recent combined DDLS and nmr relaxation study24 of very short oligonucleotides, where collective motions are less important, confirms these results. The long-range collective motions of DNA have been studied by dynamic light s ~ a t t e r i n g ' ~and .~~ transient electric b i r e f r i n g e n ~ e .Recent ~~ developments in DNA preparation techniques make monodisperse samples available. Therefore, DNA has become a very useful system t o test various dynamic models. But DNA and RNA are also interesting macromolecules by themselves, because information about the flexibility and collective internal dynamics is important in order to understand the role of conformational changes for biological processes. In a separate paper2' we reported results from DDLS on the overall reorientation of tRNAPhe. Above the melting temperature, we observed a second, much faster relaxation process. Here we will analyze this broad component in the DDLS spectrum in more detail, in order t o determine its origin. We will show that the measured relaxation times as well as the calculated hydrodynamic volume must be attributed to the collective reorientation of the bases and do not reflect the very local librational motion of a single base. Due to the coherent character of DDLS, the signal is strongly determined by static orientational pair correlations between the bases. Therefore, the measured dynamics reflect the correlated rather than the uncorrelated motion of the scattering units. Thus, DDLS is a complementary method to nmr, ESR, and FPA, and allows us to study the more collective internal dynamics of polymers consisting of optically anisotropic subunits.
INTERNAL DYNAMICS OF tRNAPhe
THEORY The spectrum IVH ( K , w ) measured in a DDLS experiment from a dilute solution of a polymer that consists of optically anisotropic monomer units has been theoretically calculated by Pecora:
977
If static correlations between monomers have to be taken into account, Bauer e t al. have shown that the collective internal relaxation time rFntis given by31
29330
I w ( K , 01
=-
A 27rp
s
exp(-iwt)S,,(K,
t ) dt
(1)
where A is a constant, p the number density of the polymer, S,, ( K , t ) the correlation function, and K the scattering vector K = [ ( 4 x n )/ A ] sin ( 8 / 2 ) . Here n is the refractive index of the solvent, X the wavelength of the incident light, and 8 is the scattering angle. For a polymer chain of n monomer units, the correlation function is given by3's3'
where F is the static internal correlation function. If the collective reorientation of the bases is diffusional, one can assume that the dependence of rotational relaxation time 7 Fnt on temperature T and solvent viscosity 77 can be described by the StokesEinstein-Debye ( SED ) equation,32 (7) where V, is the hydrodynamic volume of the relaxing unit, kB the Boltzmann constant, and r,, a constant.
MATERIALS A N D METHODS where of,( t ) is the y z component of the laboratoryfixed polarizability a t time t and r 1is the position of the i t h monomer unit. T h e monomer units i and j belong to the same chain. The K-dependent term can be ignored since its relaxation is much slower than the relaxation of a;, ( t ). For a rigid polymer, i.e., when the relaxation of monomer units within the polymer is much slower than the overall rotation of the macromolecule or if the internal motion does not change oYzsignificantly,
S,,(t)
N
exp(-6DRt),
(3)
where L)' is the rotational diffusion coefficient of the macromolecule. If the monomer units can reorient independently on a time scale much faster than the overall rotation of the macromolecule, then
where r f n tis the internal dynamics relaxation time of the monomer units within the polymer. In this case the spectrum consists of a single Lorentzian,
Transfer RNA from brewer's yeast ( BoehringerMannheim) was used without further purification. A tRNA solution with a concentration of 50 mg/ mL in 10 m M Tris HC1, 0.2 AT NaC1, pH 7.2, buffer was prepared. Before measurements the tRNA solutions were filtered into dust-free light scattering cells through Millipore filters (pore size 0.22 p m ) and centrifuged for a t least half a n hour a t 2000 X g in order to remove air bubbles. Two kinds of light scattering cells were used: a t atmospheric pressure a 4 X 4 X 40 mm fluorescence cell (Hellma) and a t high pressure a round cell sealed with a teflon piston and O-rings. A single mode argon ion laser (Spectra Physics, Model 165) with a n output power between 200 and 400 mW was used a s a light source. The VH component of scattered light was selected using polarizers with extinction coefficients of and for the incident and scattered light. The spectrum of the depolarized component of scattered light was measured by a planar Fabry-Perot interferometer with a free spectral range of FSR = 3.8 GHz and a finesse of about 80, stabilized by a DAS-10 system (Burleigh Instruments). The fluorescence light from a dye present in the tRNAPhepreparation was removed by a narrow band interference filter. Due to a very low depolarized intensity, it was necessary to accumulate up to 1024 sweeps of 5 s for each spectrum. The spectrum was measured in the temperature range from 65 to 93°C while the temperature was stabilized within +.l"C.
978
PATKOWSKI, EIMER, AND DORFMULLER
High-pressure measurements were performed using a specially designed high-pressure cell33that was modified for DDLS measurements in the following way: The cell was equipped with five windows-four of them a t a n angle of 90” and the fifth one a t 45”. In order t o avoid pressure-induced birefringence of the windows, a specially selected lead glass was used. T h e windows were coated with antireflective layers.
previously by a confocal Fabry-Perot interferometer of FSR = 750 MHz. On the other hand, a t high temperatures (about 99°C) no contribution due to overall rotation in the depolarized spectrum was observed, which indicates that the fluctuations of the anisotropy of “melted” tRNA predominantly follows a n intramolecular rather than a n overall tumbling mechanism. In Figure 1 the internal relaxation time 7Fnt is plotted vs q / T , where q is the viscosity of the solvent. From the slope the hydrodynamic volume is calculated to V, = (4.7 -t 0.5) X lo3 A3. Within the observed temperature range the experimental relaxation times scale with q / T according t o the StokesEinstein-Debye equation. Therefore, it is evident that T fnt corresponds to the same relaxation process a t different temperatures. Furthermore, this motion is diffusive, determined by the viscosity of the surrounding medium. The question is, what kind of relaxation process is monitored by the fast component in the DDLS spectrum. Is it the very local librational motion of a single base or a more collective reorientation of the bases. A comparison with results from a study on the aggregation of the mononucleotide guanosine-5’m o n ~ p h o s p h a t egives ~ ~ some interesting clues to the problem. From the dynamic behavior of guanosine5‘-monophosphate it was possible to estimate the size of the single helical stacks that are formed in solution. T h e dimensions of the aggregates were calculated from the rotational and translational diffusion coefficients based on a hydrodynamic theory for cylindrically symmetric molecules. The absolute correlation times as well as the hydrodynamic volume of tRNA a t high temperature, as obtained from the fast component in the DDLS spectrum, indicate that the dynamics of “melted” tRNA is similar t o the collective reorientation comprising approximately seven to eight stacked mononucleotides. Another way to estimate the number of bases that, on the average, participate in the collective reorientation process of tRNA is to compare the integral intensities of the depolarized spectra under conditions where only the overall rotation or the internal motion is observed. As discussed in detail later, this yields a value of about 5 for the static pair correlation
RESULTS AND DISCUSSION Temperature Dependence of
7Fnt
The values of the collective internal relaxation time Tfnt for tRNAPhein the temperature range of 6593°C are listed in Table I. These relaxation times were measured above the “melting temperature” ( T, = 60°C) of tRNAPhe(see Fig. 6 in Ref. 2 8 ) . At temperatures below T, the depolarized spectrum consists of a single Lorentzian of high intensity t h a t corresponds t o the overall rotation of tRNAPhe.28 At T < T,, there was no indication for a second, broader component in the spectrum which would correspond to the internal reorientation of the bases. This may be due to the fact that either the internal dynamics of tRNAPh”below T, is much slower, or the local dynamics relax only a very small part of the optical anisotropy (very restricted rotation). Thus, the intensity connected with internal dynamics is much lower than the intensity corresponding to the overall rotation of tRNA, and therefore, the fast component does not significantly contribute to the depolarized spectra. Within the temperature range of the melting transition ( 50-70°C) both, overall rotation and internal dynamics, could be measured. The collective internal correlation times were obtained from a double Lorentzian fit to the spectrum deconvoluted with the instrumental function.34 As mentioned above, the second, much slower component is due to the overall rotation of tRNA. The spectral resolution of the planar Fabry-Perot interferometer was adjusted in such a way that the narrow line (slow relaxation process) was not resolved. However, the rotational dynamics of tRNA has been studied
Table I
Temperature Dependence of the Collective Internal Relaxation Time 7fntfor tRNAPhe
Temperature (“C) (PS)”
7:”t
a
65.5 467
The experimental error amounts to *5%.
70.2 436
75.4 416
80.6 382
85.3 345
90.0 310
92.9 307
INTERNAL DYNAMICS OF tRNAPhe
060
-
979
5
where Iy' is attributed to the overall reorientation of the molecule, while I@Aarises from the relaxation of the optical anisotropy due to local motion. Pecora and Wang3' have calculated the time correlation function for the diffusion of a rigid rod in a cone. As long as the cone angle H,, 5 50", the following approximation is valid:
t
/ 3.7 5 750 11.3 15.0
OO
q / T w lo6 [ P / K Figure 1. The collective internal relaxation time as a function of q / T .
+
1 T&
factor ( 1 F,ntra), which indicates that the collective internal reorientation comprises approximately 5 bases. Both estimates are rather crude, but nevertheless they clearly show that the relaxation process observed in the DDLS experiments is different from the very local internal motion of a single base as it is observed by other spectroscopic methods. A t low temperatures (ordered tRNA form) the internal relaxation is strongly limited by hydrogen bonding and stacking interactions between the heterocyclic bases. In such a case the internal dynamics will result in very small changes of the polarizability anisotropy and most of the optical anisotropy will relax by the overall rotation of the macromolecule. That means, most of the depolarized intensity is attributed t o the rotation of the whole molecule. For this reason we were not able to measure the internal dynamics of tRNAPhea t low temperatures. A t high temperatures (unfolded structure) the macromolecule becomes more flexible and allows collective rotat ional motion of the heterocyclic bases. This decreases the relaxation times and also leads to a loss in the static pair correlations, i.e., a lower scattering intensity. As mentioned in the introduction, it is well established that a t room temperature the single bases perform a librational motion with a rms amplitude of about 20". In the depolarized spectra of tRNAPh", however, there is no evidence for such a relaxation process below the melting temperature. If we assume that the local internal motion of the bases can be described by the restricted rotational diffusion in a cone it is possible to estimate the fraction of the component of the depolarized scattered light Zvh for this relaxation process. For a single scatterer, the total intensity is given by
where C?, C:, and v: are dependent on the cone angle, and 6 ( w ) is the Dirac 6 function. Dk is the rotational diffusion coefficient characteristic for the internal relaxation process. Within the dynamic model the orientation of the cone is fixed. In our system the whole molecule rotates, but nevertheless in a first approximation we can consider C: proportional to IYH, due to overall reorientation, and C proportional to the integrated intensity I$',!, of the local motion of a n individual base. For a cone angle of 20" the amplitude of the librational motion amounts t o approximately 20% and for Bo = 30" C:/C: is N 0.5 ( a figure of Ci/C: as a function of the cone angle do is given in Ref. 38). Thus, for a monomer unit the restricted rotational diffusion would give a significant contribution to the total integrated intensity a t a cone angle that is typical for the bases within DNA. The amplitude of the librational motion in tRNAPheshould not be very different. This estimate is correct for a single base or nucleotide, but now we have to account for the situation in a macromolecule. The integrated intensity IYH, due to the overall rotation is proportional to ( y ')
I&
x
(y 2 )
(10)
where (7') is the mean square optical anisotropy of the macromolecule
with p the optical anisotropy of the monomer units and ( 1 Fin,,,)a measure of the internal orientational pair correlation between the subunits. Therefore, the intensity ratio for tRNA is given by
+
980
PATKOWSKI, EIMER, AND DORFMULLER
For a flexible chain molecule Fintra is about 0. But in tRNA the orientation of the bases is strongly correlated and the static pair correlation function (1 Fintra) is expected to be significantly larger than zero and may amount to about 75 ( = number of bases per tRNA molecule). Therefore, the contribution of the librational motion of the individual bases to the depolarized spectra is much smaller than estimated from Eq. ( 9 ) . It has to be emphasized that the relaxation time for the local internal motion is expected to be < 250 ps. Due to the FSR of the planar Fabry-Perot interferometer used in these experiments this fast component might disappear in the baseline. On the other hand, we did not observe an unusual high background in the spectra which would have supported this possibility. An Arrhenius plot of the internal relaxation time rfntis shown in Figure 2. The best fit to the experimental data was obtained for E, = (17 2 ) k J / mol. The activation energy for the collective internal relaxation process is smaller than E, for the overall reorientation ( E , = 20.3 -t 0.4 kJ/mol) as observed by the same experimental method.28 Nevertheless, the difference is not very pronounced, and therefore it is reasonable to assume that both motions are characterized by similar relaxation mechanisms. On the contrary, ESR measurements of spin-labeled tRNAVa',17which reflect the mobility of the aminoacetyl end of tRNA molecules, show an abrupt transition in the nature of the motion at low ionic strength. The transition temperature coincides with the melting temperature measured by the optical density. The activation energy for the relaxation process at high temperature is almost twice as high as E, for T < T,. A comparison with our DDLS
+
*
OSO'lo
100 I
90 I
80
70
60
I
I
Table I1 Pressure Dependence of the Correlation Time T;,,~for tRNAPheat 90°C.
P (bar) 71t
(PSI" a
50
250
500
1000
303
379
406
560
The experimental error amounts to +5%.
results clearly indicates that both experimental methods monitor different aspects of the rather complex conformational fluctuations of tRNA. Pressure Dependence of the Internal Relaxation Time
In order to characterize the collective reorientational relaxation process in more detail, we have also investigated the effect of pressure on the internal dynamics of tRNAPhe.Measuring the internal dynamics relaxation time r Fnt provides the activation volume AV# of the relaxation process
where R is the gas constant, T the absolute temperature, and P the pressure. The internal relaxation time rFnt has been measured at pressures from 1to 1000bar at temperatures of 70 and 90°C. At 70°C no pressure effect was observed. The pressure dependence of r;ntat 90°C is given in Table I1 and plotted in Figure 3. According to Eq. ( 1 3 ) , the activation volume for the relaxation process was calculated to AV# = (30 4 ) A3. This is a very small volume, even if compared with the hydrodynamic volume of a single base. It is com-
*
50 II
0.50: -
-040 ul
-C
50.30
0.20
-
0.20 ; 250 500 750 1000 0 2.70 280 290 3.00 330 1;
2.60
Pressure [bar ]
l/T*1000 [ l / K I Figure 2. Arrhenius plot of the collective internal correlation time 7fnt.
Figure 3. 90°C.
Plot of 1n(7Fn,) vs pressure for tRNAPheat
981
INTERNAL DYNAMICS OF tRNAPhe
parable t o the activation volume AV' = 15 A3 obtained in a similar experiment for guanosine-5'm o n ~ p h o s p h a t e Since . ~ ~ the activation volume reflects the extra volume required by the rotating moiety to move from the equilibrium to a transition state, it means that this transition state requires only a small extra volume. In order to interpret the experimental activation volume one should also recall that this quantity contains a usually negative contribution from the rearrangement of the environment. We have also studied the influence of hydrostatic pressure on the overall rotation relaxation time of tRNAPh' at temperatures below T,. No effect of pressure on the measured dynamics has been observed.
where n is the number of monomer units per macromolecule and p is the optical anisotropy of the monomer. We have measured the depolarized intensities a t 2 and 9O"C, and obtained
Comparison with Other Studies For a rigidly parallel stacked polymer the pair correlation factor (1 Fin,,,) can amount to n , the The results from various experimental methods on number of monomer units in the macromolecule. the magnitude and the time constant of the local Since for tRNAPhe(1 Fin,,,)2 76, with Eq. (15) internal motion of the bases within DNA and RNA we can estimate (1 Fintra) t o I5 for the collective are still controversial. In Table I11 we compare nmr reorientational process observed at high temperature and ESR data on the internal correlation time with ( T2 90°C). our results. Comparing our numerical values of Assuming that the high temperature data can be 7Fnt with the results from other studies, one has to keep in mind that DDLS measures the collective extrapolated t o room temperature, we obtain 7Fn,(20"C) = 1.2 ns [see Eq. ( 7 ) ] , which gives relaxation time [ see Eq. ( 6 ) 1, while the other methods yield the single particle correlation time 7 % , . In 7ynt(20°C) N 0.25 ns. Our estimate for the correlaprinciple, the static pair correlation factor ( 1 Fintra) tion time of the internal motion lies a t the lower can be calculated from the integral depolarized inend of comparable data shown in Table 111. More recent studies 12,22,24 lead t o the conclusion that the tensities I,, of the monomer unit and the polymer
+
+ +
+
Table I11 Internal Dynamics of Nucleic Acids: Comparison with Other Studies
Reference
Sample
Method
Present study
tRNAPh' (yeast) tRNA""' (E. coli) tRNAPhe(E. coli) tRNATYr(yeast) DNA (calf thymus) poly(dA-dT) Single-stranded RNA Single-stranded DNA Double-helical RNA Double-helical DNA poly(dG-dC)* poly( dG-dC) (B, Z DNA) tRNAPhe
DDLS ESR ESR ESR DDLS Proton nmr ESR
15 16 17 23 10 14
11
12
Tint
v,
(20", w), ns
-
0.2 0.1-1 0.36-2.40 1.75 3 0.7 1 4
Proton nmr
0.1-1.5
13C-nmr
-
0.02
A3
(4.7
0.1-lo3)
982
PATKOWSKI, EIMER, AND DORFMULLER
time constant of the librational motion of the bases is in the range of picoseconds. The main argument against longer times, found in other studies, is that long-range twisting and bending motions of the relatively long fragment studied have been neglected in the model calculations. Moreover, nmr does not provide the time resolution for the very fast processes, but the information about the internal correlation time was extracted from relaxation processes that decay on a much slower time scale. It was shownz4that the amplitude of the librational motion can be estimated quite accurately from the nmr relaxation parameters, while this is not possible for the time constant. From all this it becomes clear that so far the dynamics of the local internal reorientation of the bases within DNA and RNA have not yet been fully characterized quantitatively by any one of the experimental methods used.
CONCLUSIONS We have studied the collective internal reorientation of the bases in tRNAPhe.Within the melting region the depolarized spectra show two clearly distinct relaxation processes, a slow component, which is attributed to the overall rotation of the macromolecule, and a broad line, which describes the collective reorientation of the bases within tRNAPhe.Well above the melting temperature only the fast relaxation motion is observed. The absolute correlation time and the experimental hydrodynamic volume indicate that the segmental motion extends over I 5 bases. Therefore, the fast relaxation process cannot be attributed to the librational motion of a single base, as observed by other spectroscopic methods. This very local motion, which is also present below the melting temperature, does not significantly contribute to the depolarized spectra. Due to the correlated static orientation of the bases within the macromolecule, it is rather the overall reorientation of the molecule that dominates Ivh below T,. We have shown that DDLS provides valuable additional information about the collective motion of the bases within tRNAPhe,which is not available by other methods. I n fact, we are now able t o describe the dynamics of tRNA on three different levels. The local one-base reorientation monitored by nmr, FPA and ESR, the collective reorientation as monitored by DDLS above the melting temperature, involving a small number of cooperatively rotating bases, and finally, the overall reorientation of the whole macromolecule.
This work was supported by the Complex Liquids Project of the ZIF (Center for Interdisciplinary Research), the Minister fur Wissenschaft und Forschung des Landes NRW, the Fonds der chemischen Industrie, and in part ( t o A P ) by the research project RP.II.13.1.10.
REFERENCES 1. Saenger, W. ( 1984) Principles of Nucleic Acid Struc-
ture, Springer-Verlag, New York. 2. Clementi, E. et al., Eds. ( 1985) Structure and Motion,
Nucleic Acids and Protein, Academic Press, New York. 3. Rigler, R. & Wintermeyer, W. (1983) Ann. Rev. Biophys. Bioeng. 12, 475-505. 4. McCammon, J. A., Gelin, B. R. & Karplus, M. ( 1977) Nature 2 6 7 , 585-590. 5. Harvey, S. C., Prabhakaran, M., Mao, B. & McCammon, J. A. (1984) Science 2 2 3 , 1189-1191. 6. Harvey, S. C., Prabhakaran, M. & McCammon, J. A. (1985) Biopolymers 2 4 , 1169-1188. 7. Prabhakaran, M., Harvey, S. C. & McCammon, J. A. (1985) Biopolymers 24, 1189-1204. 8. Early, T. A. & Kearns, D. R. ( 1979) Proc. Natl. Acad. Sci. U S A 76,4165-4169. 9. Mirau, P. A. and Kearns, D. R. (1984) Biochemistry 23, 5439-5446. 10. Assa-Munt, N., Granot, J., Behling, R. W. & Kearns, D. R. (1984) Biochemistry 23,944-955. 11. Mirau, P. A., Behling, R. W. & Kearns, D. R. (1985) Biochemistry 2 4 , 6200-6211. 12. Schmidt, P. G., Sierzputowska-Gracz, H. & Agris, P. F. (1987) Biochemistry 2 6 , 8529-8534. 13. Bobst, A. M., Kao, S.-C., Toppin, R. C., Ireland, T. C. & Thomas, I. E. (1984) J . Mol. Biol. 173, 6374. 14. Kao, S.-C. & Bobst, A. M. (1985) Biochemistry 24, 5465-5469. 15. Hoffman, B. M., Shofield, P. & Rich, A. (1969) Proc. Natl. Acad. Sci. USA 62, 1195-1202. 16. Caron, M., Brisson, N. & Dugas, H. (1976) J . Biol. Chem. 251, 1529-1530. 17. Weygand-Durasevic, I., Kruse, T. A. & Clark, B. F. C. (1981) Eur. J. Biochem. 116,59-65. 18. Barkley, M. D. & Zimm, B. H. ( 1979) J . Chem. Phys. 70, 2991-3007. 19. Millar, D. P., Robbins, R. J. & Zewail, A. H. ( 1981) J. Chem. Phys. 76, 2080-2094. 20. Magde, D., Zappala, M., Knox, W. H. & Nordlund, T. H. (1983) J . Chem. Phys. 87,3286-3288. 21. HBrd, T. (1987) Biopolymers 26, 613-618. 22. Schurr, J. M. & Fujimoto, B. S. (1988) Biopolymers 27, 1543-1569. 23. Patkowski, A., Fytas, G. & Dorfmuller, Th. (1982) Biopolymers 21, 1473-1477. 24. Eimer, W., Williamson, J. R., Boxer, S. G. & Pecora, R. ( 1990) Biochemistry 29, 799-811.
INTERNAL DYNAMICS OF tRNAPhe
25. Bloomfield, V. A. (1985) in Dynamic Light Scattering: Applications of Photon Correlation Spectroscopy, PeCora, R., Ed., Plenum, New York. 26. Sorlie, S. S. & Pecora, R. ( 1988) Macromolecules 2 1 , 1437-1449. 27. Lewis, R. J., Pecora, R. & Eden, D. ( 1986) Macromolecules 16, 134-139. 28. Patkowski, A,, Eimer, W. & Dorfmuller, Th., (1990) Riopolymers 30,93-105. 29. Pecora, R. (1968) J . Chem. Phys. 49, 1036-1043. 30. Berne, B.
