A Single-Molecule Study of RNA Catalysis and Folding Xiaowei Zhuang, Laura E. Bartley, Hazen P. Babcock, Rick Russell, Taekjip Ha, Daniel Herschlag, and Steven Chu

For cleavage: 50 mM NaMOPS (pH 7.0), 10 mM MgCl₂, and various concentration of guanosine (G) at 22°C. For P1 docking: 50 mM NaMOPS (pH 7.0) and 10 mM MgCl₂ at 22°C. For unfolding: 50 mM NaMOPS (pH 7.0) at 25 °C. For refolding: 50 mM NaMOPS (pH 7.0) and 5 mM MgCl₂ at 25°C. All buffers contained an oxygen scavenger system (0.1 mg/ml glucose oxidase, 0.02 mg/ml catalase, 1% b-mercaptoethanol and 10% w/w glucose) to reduce the photobleaching rate.

Ribozyme Preparation

PCR and conventional cloning methods were used to modify the plasmid encoding the L-21 ribozyme, pT7L-21 (1), to encode the L-21ScaI ribozyme extended at its 3´-end by 26 nucleotides (ACCAAAAUCAACCUAAAACUUACACA). The resulting plasmid was prepared for in vitro run-off transcription by digestion with EarI, a recognition site for which had been introduced downstream of the sequence for the 3´-extended ribozyme. Ribozyme was transcribed for 1 to 3 hours at 37°C under conditions with limiting Mg²⁺ [100 mM TrisCl (pH 8.0), 5 mM MgCl₂, 1 mM each NTP, 1 mM spermidine, and 0.1 % triton X-100]. The Low Mg²⁺ concentration gave a larger fraction of fully extended ribozyme compared with standard conditions, presumably by minimizing ribozyme self-processing. Ribozyme was purified on Qiagen RNeasy columns and stored in 10 mM tris-Cl (pH 7.5), and 0.5 mM EDTA.

Ribozyme was annealed to a DNA tether complementary to the last 24 residues of the 3´-extension by incubation at 95°C for 30 s and then at 60°C for 5 min in 50 mM NaMES (pH 5.0-5.3), and 250 mM NaCl. Ribozyme was prefolded in 10 mM MgCl₂, 50 mM NaMES (pH 5.0), and 125 mM NaCl at 50°C for 30 min. To bind oligonucleotide substrate S, the folded ribozyme (>50 nM) was incubated with a stoichiometric concentration of S for 60 s, sufficient time to allow essentially complete binding.

To immobilize ribozyme to a surface, the surface of a quartz slide or a glass coverslip was incubated sequentially with biotinylated-BSA (1 mg/mL, Sigma), streptavidin (0.2 mg/mL, Molecular Probes, Inc.), and folded ribozyme bound to S (~50 pM).

Preparation of oligonucleotides

Substrates and the biotinylated DNA tethers (Dharmacon Research, Inc.) were labeled with Cy3 or Cy5 (Amersham/Pharmacia) post-synthetically, via a primary amine at the end of a three-carbon linker attached to the 3´-most phosphate. In detail, oligonucleotides were deprotected in 200 mM Na-Acetate (pH 3) for 15 min at 60°C. Then, the deprotected oligonucleotides were incubated with a fivefold excess of dye at 50°C, Na-Carbonate (pH 8.5) for 10 hours. Full-length, dye-labeled oligonucleotides were obtained by HPLC purification on a Dionex 100-PA ion exchange column in 10% acetonitrile, with an ammonium acetate gradient (0.01 to 1 M), followed by desalting on Waters Sep-Pak columns (Millipore).

Total internal reflection microscope

A circularly polarized 514 nm laser beam from an argon laser was internally reflected off the quartz surface. The resulting evanescent wave excites the fluorescent molecules immobilized on the quartz surface. A 60 water immersion objective with 1.2 numerical aperture (Nikon) mounted on a inverted microscope (TE300, Nikon) was used to image the fluorescence emission onto a 12-bit cooled, intensified CCD camera (Pentamax, Roper Scientific). A 540-nm long-pass filter was used to block laser light scattered from defects in the quartz. For FRET experiments, the fluorescence emission was split with a dichroic beamsplitter (635 nm long-pass). The short- and long-wavelength emissions were imaged onto the two halves of the camera. A home-built automated buffer exchange system was used to initiate reactions such as substrate cleavage, unfolding and refolding.

Scanning confocal microscope

A circularly polarized 514 nm laser beam from an argon laser was focused on the surface of a microscope coverslip by a 100 oil-immersion objective (Zeiss) with numerical aperture of 1.25, mounted on a inverted microscope (Axiovert 100, Zeiss). The fluorescence emission from molecules immobilized on the surface was imaged by the same objective onto silicon avalanche photodiodes (EG&G). In the FRET experiments, the fluorescence emission was split by a dichroic beamsplitter (625-nm long-pass). The short- and long-wavelength emission was detected after passing through a 590-nm band-pass filter and a 640-nm long-pass filter, respectively.

Rate limiting step for single-molecule cleavage reactions

Previously, the chemical step has been shown to be rate limiting based on the effects of phosphorothioate substitutions at the cleavage site and 2´-substitutions at the -1 position (see Fig. 1) on the cleavage reaction, and a log-linear pH dependence of the reaction rate constant with a slope of one (2-6). In this study, single-molecule and ensemble measurements gave the same rate constants and pH dependencies of cleavage, suggesting that the chemical step, not GA-Cy3 release, is rate limiting.

Measurements of the rate constant for substrate release

In the single-molecule measurements, release of the Cy3-labeled S results in the disappearance of fluorescence signal from individual molecules, allowing the determination of the rate constant for substrate release, k^s_off. A substrate with a 2´-deoxyribose at the -1 position (see Fig. 1) (-1dS) was used in order to reduce the rate constant for cleavage in the absence of G (7). We found that k^s_off = 3 ± 1 min^-1 at 40°C. For ribozyme molecules free in solution, k^s_off was obtained using the following method: k^s_off for all ribose S was first determined using k^s_off = K^E_d · k^s_on, and the previously determined values of the equilibrium constant K^E_d and association rate constant k^s_on for binding of S (8, 9). The value was then corrected to account for the threefold destabilizing effect of the 2´-deoxyribose at position -1, giving k^s_off = 2 ± 1 min^-1 at 40°C for -1dS (10).

Effects of dye-labeling and 3´-extension on cleavage activity

Cleavage of the Cy3-labeled S by the L-21 ScaI ribozyme gave the same rate constants (within 20%) as those shown in Fig. 2, indicating that the 3´-extension and base-pairing with the DNA tether had no effect on the cleavage activity. The Cy3 dye attached to the 3´-end of S gave only a 2- to 3-fold decrease in k_max compared with the unlabeled S, an effect similar in magnitude to that of changing the length or base identity of the 3´-portion of S (11). Additionally, the Cy5 dye attached to the tether had no effect on the cleavage reaction rate.

Docking equilibrium in solution

The docking equilibrium constant, K_dock, for the all-ribose S was determined from the equation, K_dock = (K_d^undocked · k^s_on/k^s_off) - 1 (8). k^s_off was measured using pulse-chase experiments (2-7). k_on was measured under conditions in which S binding is rate-limiting (2-6). The equilibrium constant for binding of S to the ribozyme in the undocked state, K_d^undocked, was determined previously (9, 10).

FRET changes in overall folding study

Supplemental Figure 1. Histograms of the FRET signals of individual ribozyme molecules: (A) The FRET value for the ribozyme prefolded in solution to the native state and then immobilized to surface is centered at 0.9. (B) Upon removal of Mg²⁺ to unfold the ribozyme, FRET decreased to ~0.1. (C) Subsequent refolding by the addition of Mg²⁺ resulted in two populations of molecules: 12% of the molecules had FRET ~0.9 and reached the native state [see (D)], the remaining 88% showed a FRET ~0.3, indicating misfolding. (D) After further washing the sample with a buffer containing 1 mM G to induce cleavage of S, the molecules with FRET ~0.9 disappeared. Thus, these molecules were enzymatically active and indeed in the native state. These results are consistent with the previous identification of two folding pathways: 88% of the ribozyme molecules fold to a misfolded state M that lacks substrate cleavage activity and convert to the native state after many hours, the remaining 12% avoid M and fold to the native state more quickly (12). Approximately 16% of the ribozyme molecules were labeled with inactive Cy5. These molecules gave a peak centered at FRET ~0 in each of the histograms. This peak was removed by fitting it with a Gaussian and then subtracting the Gaussian from the histogram. The fluorescence properties of Cy3 and Cy5 depend only slightly on Mg²⁺ concentration.

Medium version | Full size version

Ensemble experiment on overall folding of ribozyme free in solution with and without bound S

Binding of S to the unfolded ribozyme is ~1000-fold weaker than binding to the folded ribozyme in the undocked state (unpublished results). Therefore, to measure the folding kinetics of ribozyme bound to S, we followed the general method used in the single molecule experiments. First, a trace amount of S was bound to folded ribozyme (2 mM). The ribozyme was then unfolded in a 50mM NaMOPS (pH 7), 2 mM EDTA and 0.2 mM MgCl₂ for 10 s or 60 s. Both conditions are sufficient to completely unfold the ribozyme (see below). After 10 s and 60 s under such conditions, the fraction of S remaining bound to the ribozyme is ~80% and ~30%, respectively (data not shown). The ribozyme was then refolded in 50 mM NaMOPS (pH 7), 5 mM MgCl₂, 0.2 mM EDTA, and 2 mM G. A fraction of ribozyme folded rapidly with a rate constant >0.2 s^-1, indicated by the fast formation of oligonucleotide cleavage product. Of the ribozyme that folded to the native state the fractions that folded rapidly were 0.6 ± 0.3 and 0.2 ± 0.1 for the 10 s and 60 s unfolding time, respectively. Folding of the ribozyme molecules that were not bound to S at the time of initiation of folding was also detected by the cleavage of S due to the rebinding of S to the folded ribozyme. Being much slower than the rate constants for S binding, G binding, and cleavage, the folding rate constant of these molecules was thus faithfully determined as well (12). In an otherwise identical experiment in which S was not bound to the ribozyme before unfolding but added after refolding, no fast folding molecule was observed. These above results strongly suggest that population of the fast folding pathway is due to the presence of bound S.

Movies of cleavage reaction and photobleaching of individual ribozyme molecules.

To view these movies, download a QuickTime viewer.

• Movie 1
  Real time images of individual ribozyme molecules with Cy3-labeled S. Right panel: cleavage buffer containing 1mM G was added to the sample at time = 0. After cleavage occurred, the Cy3-labeled cleavage product was released from the ribozyme, leading to the disappearance of fluorescence spots. Left panel: images taken under the same condition except that the buffer added at time = 0 contained no G. The disappearance of fluorescence spots was slower by more than 10-fold, indicating that photobleaching was negligible.

Control experiments for complete unfolding

Time-resolved single-molecule FRET measurements to determine folding kinetics were performed following four unfolding conditions: 10 s in buffer lacking Mg²⁺, 60 s in buffer lacking Mg²⁺, 60 s in buffer containing 1 mM EDTA and no Mg²⁺, and 60 s in containing 5 mM EDTA and no Mg²⁺. Each gave identical partitioning between the misfolded and native states and between the slow and fast folding paths, indicating complete unfolding of ribozyme under all of these conditions.

REFERENCES

1. A. J. Zaug, C. A. Grosshans, T. R. Cech, Biochemistry 27, 8924 (1988).

2. D. Herschlag and M. Khosla, Biochemistry 33, 5291 (1994).

3. D. Herschlag, J. A. Piccirilli, T. R. Cech, Biochemistry 30, 4844 (1991).

4. D. Herschlag, F. Eckstein, T. R. Cech, Biochemistry 32, 8312 (1993).

5. T. S. McConnell and T. R. Cech, Biochemistry 34, 4056 (1995).

6. D. S. Knitt and D. Herschlag, Biochemistry 35, 1560 (1996).

7. D. Herschlag and T. R. Cech, Biochemistry 29, 10159 (1990).

8. G. J. Narlikar and D. Herschlag, Nature Struct. Biol. 3, 701 (1996).

9. G. J. Narlikar, L. E. Bartley, M. Khosla, D. Herschlag, Biochemistry 38, 14192 (1999).

10. G. J. Narlikar, M. Khosla, N. Usman, D. Herschlag, Biochemistry 36, 2465 (1997).

11. R. Russell and D. Herschlag, RNA 5, 158 (1999).

12. R. Russell and D. Herschlag, J. Mol. Biol. 291, 1155 (1999).

To Advertise   |   Find Products