A new method to monitor the rate of conformational transitions in RNA
A new method to monitor the rate of conformational transitions in RNAEmily J. Maglott and Gary D. Glick*
Department of Chemistry, University of Michigan, Ann Arbor, MI 48109-1055, USA
Received April 15, 1997;Revised and Accepted July 3, 1997
ABSTRACT
Many RNAs need Mg2+ to produce stable tertiary structures. Here we describe a simple method to measure the rate and activation parameters of tertiary structure unfolding that exploits this Mg2+ dependence. Our approach is based on mixing an RNA solution with excess EDTA in a stopped-flow instrument equipped with an absorbance detector, under conditions of temperature and ionic strength where, after chelation of Mg2+, tertiary structure unfolds. We have demonstrated the utility of this method by studying phenylalanine-specific transfer RNA from yeast (tRNAPhe) because the unfolding rates and the corresponding activation parameters have been determined previously and provide a benchmark for our technique. We find that within error, our stopped-flow method reproduces both the rate and activation enthalpy for tertiary unfolding of yeast tRNAPhe measured previously by temperature-jump relaxation kinetics. Since many different RNAs require divalent magnesium for tertiary structure stabilization, this technique should be applicable to study the folding of other RNAs.
INTRODUCTION
While significant advances have been made in studying native RNA structure (1 -7 ), relatively little is known about the rates of tertiary folding and unfolding, particularly on timescales faster than seconds. Such data are essential for developing mechanistic models of RNA folding and are particularly important in the context of present efforts to design RNA molecules that adopt stable structures possessing specific binding and catalytic properties (8 ). Recent approaches to investigate tertiary folding have used the intrinsic properties of native structure to report on folding transitions, such as self-splicing for ribozymes (9 -11 ), as well as indirect techniques like oligonucleotide hybridization (12 ,13 ), UV cross-linking (14 ) and chemical modification (10 ,15 ,16 ). With the exception of recent stopped-flow fluorescence experiments that monitor conformational changes that can occur upon substrate binding to ribozymes (1 7 ,18 ), these methods cannot observe transitions that occur faster than about several events per minute.
At present, temperature-jump (T-jump) relaxation is the only general method capable of directly measuring folding transitions in RNA that occur faster than seconds. Indeed, much of our understanding of RNA folding comes from the elegant work of Crothers (19 -22 ), Maass (23 -27 ) and Biltonen (28 ) who used T-jump to study the folding and unfolding kinetics of transfer RNAs (tRNA). These studies established that denaturation of tRNA tertiary structure to a `cloverleaf' is the first step in the unfolding pathway of tRNAs, and that this conformational change happens on millisecond timescales. It was also found that under the appropriate conditions, the various elements comprising a cloverleaf melt in a sequence-dependent fashion after tertiary structure has been disrupted (19 -28 ). While T-jump is an extremely powerful technique, it does have some limitations. For example, T-jump equipment is expensive and not widely available, current commercial instrumentation uses relatively large sample volumes, and photobleaching can lead to sample degradation. These latter points are particularly important when using RNAs that are only available in small quantities such as synthetic constructs possessing site-specific modifications (29 ).
Here, we describe a simple non-invasive method to determine the rate and activation parameters associated with unfolding of native RNA tertiary structure. This method exploits the fact that formation of native RNA structure often requires Mg2+ or other divalent metals (30 -40 ), and it entails mixing an RNA solution containing Mg2+ with excess EDTA in a stopped-flow instrument under conditions of temperature and ionic strength where tertiary structure unfolds after chelation of Mg2+. Low millisecond resolution can be achieved, and the unfolding transition can be observed by standard absorbance measurements. We chose to develop this method using phenylalanine specific transfer RNA from yeast (tRNAPhe) as a model system because the unfolding rates and the corresponding activation parameters for this molecule are already known and provide a standard for comparison (23 -25 ,41 ). Our results show that the stopped-flow technique is robust and versatile, which in principle can be used to examine a range of systems.
MATERIALS AND METHODS
Sample preparation
Yeast tRNAPhe was obtained from Sigma and was purified by denaturing polyacrylamide gel electrophoresis (8% polyacrylamide, 29:1 acrylamide:bis-acrylamide, 8 M urea, 31.0 cm * 38.5 cm * 0.8 mm). Product bands were excised from the gels and soaked in 4* TAE (450 [mu]l, 160 mM Tris, 80 mM acetic acid, 8 mM EDTA, pH 7.5) for 12 h. The tRNA samples were ethanol precipitated and stored at -20oC. Prior to use, samples were dialyzed against buffer 1 (600 ml, 2 h, 5 mM Na2HPO4, 22 mM NaCl, 3 mM Na2EDTA, pH 6.8) to remove any residual Mg2+ and then dialyzed against buffer 2 (600 ml, 2 h, 5 mM Na2HPO4, 22 mM NaCl, pH 6.8) to remove the EDTA. Lastly, the tRNA was dialyzed into the folding buffer (1.6 l, 6.5 h, 5 mM Na2HPO4, 22 mM NaCl, 1 mM MgCl2, pH 6.8). The dialyzed tRNA was diluted with folding buffer to a final concentration of 0.5 [mu]M. No more than 6 h prior to use, aliquots of this solution (1 ml) were heated to 72oC for 5 min and then cooled to room temperature over 30 min to properly fold the tRNA. Mg2+-free yeast tRNAPhe was prepared following the method of Levy et al. (28 ). Briefly, samples were dialyzed against buffer 1 (600 ml, 2 h) and buffer 2 (1 l, 4 h) and contained <0.5 molecules of Mg2+ per tRNA as determined by atomic absorption analysis with ICP detection.
Kinetic measurements
Kinetic experiments were performed using an Applied Photophysics Ltd model SX18MV stopped-flow spectrophotometer. In a typical experiment, a solution of folded tRNA (0.5 [mu]M) and an EDTA solution (ranging from 0.5 to 9 mM) in the same folding buffer were each equilibrated in the stopped-flow instrument at the desired temperature for 15 min prior to mixing. Mixing was commenced (deadtime of 1.4 ms) using equal volumes (50 [mu]l) of the two solutions and the unfolding transition was monitored at 268 nm (rather than 260 nm to reduce the background absorbance of EDTA). A minimum of seven kinetic traces were obtained at each temperature and each concentration of EDTA. All kinetic data obtained under a given set of conditions were averaged, and the average trace was analyzed using the Kinetic Spectrometer Workstation Software provided with the instrument. Errors to these fits were <6% of the rate. Standard deviations for the rates were calculated from exponential fits to each of the traces that were included in the average rate. To determine activation parameters, Arrhenius plots were constructed for each concentration of EDTA as described by Cole et al. (21 ).
UV thermal denaturation experiments
UV spectra were measured on a Cary 3 spectrophotometer equipped with a Varian Peltier. tRNA samples were prepared by dialysis against the requisite buffer. Aliquots of dialyzed tRNA were diluted with the same degassed buffer to a final volume of 1 ml in 10 mm long self-masking cuvettes and a final tRNA concentration of 0.5 [mu]M. Samples were heat-denatured at 85oC for 1 min and then cooled to 5oC at a rate of 3oC/min. The tRNA samples were equilibrated at 5oC for 30 min and then heated at a rate of 0.5oC/min while monitoring the absorbance.
RESULTS AND DISCUSSION
Experimental design and selection of buffer conditions
This stopped-flow unfolding experiment was first proposed nearly 30 years ago (28 ), yet the instrumentation of that era had relatively long dead times (e.g., 10 ms) and low sensitivity which made the approach impractical. However, hardware advances have circumvented these limitations and the basic outline of our experiment is shown in Figure 1 . While it should be possible to study folding by rapidly adding Mg2+ to an RNA solution void of divalent metals, we have concentrated on measuring the rates and energetics of unfolding because unfolding measurements start from the most defined state (a folded structure) and alternate/unproductive folding pathways due to misfolded intermediates are eliminated. Indeed, the advantages of using unfolding kinetics to examine folding pathways and transition-state structure have been addressed by Fersht and others in studies of protein folding (42 ,43 ).
Kinetics of yeast tRNAPhe unfolding
In preliminary stopped-flow experiments, we observed an increase in hyperchromicity when folded yeast tRNAPhe is mixed with a large excess of EDTA (e.g., 6-fold). The rates obtained from these measurements range from 51 (+- 18) s-1 at 28.5oC to 127 (+- 8)s-1 at 34.8oC (Fig. 2 ). The relative absorbance changes observed in the stopped-flow experiments match the UV melting data described above to within 5%. In addition, the maximum hyperchromicity increase observed in the stopped-flow experiments is ~6% and this determination agrees with the value measured previously by Leroy et al. under similar solution conditions (26 ,44 ).
ACKNOWLEDGEMENTS
This work was supported by NIH grant GM 53861 and an NIH Molecular Biophysics predoctoral fellowship to E.J.M. G.D.G. is the recipient of a National Science Foundation Young Investigator Award, a Camille Dreyfus Teacher-Scholar Award and a research fellowship from the Alfred P.Sloan Foundation. We thank J.Goodwin and M.Carey for helpful discussions.
