22971
2006, 110, 22971-22974 Published on Web 10/21/2006
Intramolecular Vibrational Preparation of the Unfolding Transition State of ZnII-Substituted Cytochrome c Sanela Lampa-Pastirk and Warren F. Beck* Department of Chemistry, Michigan State UniVersity, East Lansing, Michigan 48824 ReceiVed: August 22, 2006; In Final Form: September 21, 2006
We show that an intramolecular vibrational excitation provided by the radiationless decay of a covalently bound electronic chromophore can be exploited to drive a protein from its native folded state to the transition state for unfolding. Using this approach, we examine the effect of the polarity and viscosity of the solvent medium on the unfolding and refolding reactions of ZnII-substituted cytochrome c at room temperature. The results show that the solvent polarity controls the activation energy for the unfolding and refolding reactions; the solvent viscosity further controls the rate by frictionally hindering the moving polypeptide. These findings suggest an important role for the solvent in the kinetic control of protein-folding trajectories on the energy landscape.
Current theories that account for the rapid folding of a protein into its native structure suggest a funnel-shaped topology1 for the potential-energy surface, which has been termed an energy landscape.2,3 The funnel provides a constraint over the range of possible trajectories so that the native fold is found on a relatively short (typically 100 ms to 10 s) time scale. If the ensemble searches the surface4 under a combination of kinetic and thermodynamic control,5 the native state may not be the structure at the global minimum on the surface. Misfolded structures that contribute to a number of diseases can arise spontaneously under physiological conditions upon unfolding the native state.6 In this work, we directly prepare the unfolding transition state of ZnII-substituted cytochrome c (ZnCytc)7 by exploiting the intrinsic ZnII-porphyrin as a source of an intramolecular vibrational excitation. These experiments test the hypothesis that an unfolding reaction will be initiated under the solution conditions that favor a stable native fold if a protein is provided a vibrational excitation that surpasses the local activation energy barriers that confine the native state. Having started a reaction, the ZnII-porphyrin is then exploited as a solvatochromic fluorescent probe8 for the resulting protein and solvent motions. Figure 1 shows the absorption and fluorescence spectra from ZnCytc and the corresponding energy level diagram for the π* excited states. Excitation of the Soret band is followed by radiationless decay on the 2-ps time scale. The initial result is a highly vibrationally excited, S1 state ZnII-porphyrin; the energy above the S1 state is converted into excitation of the normal modes of vibration of the porphyrin. This excess vibrational energy is transferred to the protein structure on the 2-4 ps time scale;9,10 the resulting vibrationally equilibrated S1 state ZnII-porphyrin usually relaxes to the ground state by emitting a photon (τF = 2 ns).7 The vibrational energy mostly propagates from the porphyrin to the protein via intramolecular vibrational redistribution, through the thioether linkages provided * Corresponding author.
10.1021/jp0654359 CCC: $33.50
by Cys 14 and Cys 17,11 rather than through contact of the porphyrin with its binding site. By tuning the excitation wavelength above the energy of the 0-0 transition for the S1 state, it is possible to vary the vibrational excitation provided to the protein structure. The preceding process produces a vibrationally excited protein within a few picoseconds of the optical absorption event. As soon as the vibrational excitation occurs, however, the protein starts transferring the excess vibrational energy to the translational and librational modes of the surrounding solvent. This process requires collisions between the protein and solvent molecules; the result is a dynamic bottleneck that allows the protein to propagate structurally at a vibrationally excited level for a short period of time, no more than 20 ps. At the end of this time period, the protein and its surrounding solvent have returned to equilibrium with the ambient solution temperature.12,13 We show in the following that if the vibrational excitation is large enough, this 20 ps of propagation time is enough to permit ZnCytc to overcome the native state’s unfolding activation energy barrier. In previous work, we studied the time evolution of the fluorescence spectrum from ZnCytc on the ps-ns time scale when the excitation laser is tuned to the Q-band 0-0 transition energy, where the absorption and fluorescence dipole-strength spectra cross. The S1 state is prepared in these experiments with very little excess vibrational energy. The dynamic Stokes shift, the shift to the red of the time-resolved fluorescence spectrum,8 reports under these conditions the characteristic time scales for the random protein and solvent motions that occur in the native folded state. Figure 2 shows that the time-resolved fluorescence spectrum exhibits a biexponential shift to lower energy when the Q-band is excited at 584 nm. The fast component arises from random motions of the hydrophobic core; its amplitude and time constant (250 ps) are insensitive to the viscosity and polarity of the external solvent. The time constant of the slower component, however, lengthens from 1.45 ns in water to 2.2 ns in 50% (v/v) glycerol. This component arises from random © 2006 American Chemical Society
22972 J. Phys. Chem. B, Vol. 110, No. 46, 2006
Letters
Figure 1. (Left) Electronic and vibrational energy levels for the intrinsic ZnII-porphyrin in ZnII-substituted cytochrome c. (Right) Absorption (A) and fluorescence (F) dipole-strength spectra at room temperature, A(ν)/ν and F(ν)/ν3, respectively. The vibrational structure in the Q-band absorption and fluorescence spectra arises from the 0-0 and 0-1 vibronic transitions.
Figure 2. Dynamic Stokes shift of the fluorescence spectrum of ZnIIsubstituted cytochrome c in water at room temperature (22 °C) following excitation at 420 nm, in the Soret absorption band, or at 584 nm, in the Q absorption band (see Figure 1). The center frequency of the 0-0 vibronic transition from the time-resolved fluorescence spectrum, ν00, is plotted in each case as a function of the time delay following the excitation pulse.
motions of the protein in the solvent contact region, which is frictionally damped by the surrounding bulk solvent.14,15 The fluorescence spectrum from ZnCytc exhibits the same time evolution when the excitation laser is tuned as far as 700 cm-1 above the S1 state. In contrast, Figures 2 and 3 show that a multiphasic, bidirectional response is observed with excitation of the Soret band at 420 nm, ∼6700 cm-1 above the S1 state. This unusual dynamic Stokes shift reports a change in the protein structure. The spectrum initially shifts to lower energy about twice as rapidly as was observed with Q-band excitation. At the 180-ps delay point, the spectrum stops shifting to the red and reverses its course; a biexponential shift to higher energy is then observed over the subsequent 5 ns. The spectrum reaches a quasi-stationary point 168 cm-1 to the blue of the final shift observed with Q-band excitation. We assign the structure at the limit of the blue shift to an unfolded or partially unfolded state in which the axial ligands to the ZnII ion, from His 18 and Met 80, are dissociated. This assignment is consistent with the observation that the continuous fluorescence spectrum from ZnCytc shifts to the blue by ∼80 cm-1 as the temperature is ramped over the thermal unfolding transition.16 In an experiment with an integrating detector, the instantaneous spectrum is
Figure 3. Dynamic fluorescence Stokes shift at room temperature from ZnII-substituted cytochrome c following excitation at 420 nm in three solvent mixtures: water, 50% (v/v) glycerol-water, and 40% (v/v) methanol-water. (The 0-2-ns portion of the signal in water is shown in Figure 2.)
weighted by the decay of population from the S1 state according to the fluorescence lifetime, so the shift observed in the thermal unfolding experiment is smaller than the total shift observed in the time-resolved experiment. A somewhat larger blue shift, 120 cm-1, is observed if one compares the continuous fluorescence spectrum observed with Soret-band excitation to the one obtained with Q-band excitation (as in Figure 1). We conclude, then, that the intramolecular vibrational excitation that results from Soret-band excitation of ZnCytc at room temperature prepares the unfolding transition state of the protein and that an unfolded state is reached in only 5 ns. The rate at which the structure of ZnCytc unfolds after Soretband excitation is strongly controlled by the viscosity and polarity of the surrounding bulk solvent medium. Figure 3 compares the dynamic Stokes shift from ZnCytc observed in
Letters
J. Phys. Chem. B, Vol. 110, No. 46, 2006 22973
Figure 4. Fitted and extrapolated models for three phases of dynamics for the unfolding and refolding of ZnII-substituted cytochrome c following Soret-band excitation (see Figures 2 and 3). (Top) Comparison of the models obtained from the dynamic fluorescence Stokes shift observed in water with Soret-band (blue curve) and Q-band (red curve) excitation. (Bottom) Difference between the responses observed with Soret- and Q-band excitation. Dashed curves indicate the extrapolated time regions that lie before and after the time region observed in the picosecond time-resolved fluorescence experiments. The initial refolding component shown here is just an estimate obtained by scaling the refolding time constant in methanolwater by the ratio of the time constants for the blue shifts observed in water and methanol-water.
water with that observed in 50% (v/v) glycerol and 40% (v/v) methanol over the entire fluorescence time window (100 ps12 ns). The glycerol-water and methanol-water mixtures are identically less polar than pure water but have greatly different viscosities. Compared to the response in water, the least viscous of the solvent systems, the response in the glycerol-water mixture exhibits a smaller initial shift to the red, a delayed turnaround point (372 ps), and then a slower shift to the blue that apparently continues on after the end of the recording. The overall response in the glycerol-water mixture is slowed by a factor of 2 to 3 compared to that in pure water. The response in the methanol-water mixture is greatly accellerated, by at least a factor of 5, compared to the response in pure water, and the overall blue shift is about half of that observed in pure water. The initial red shift observed in pure water and the glycerolwater mixture is not observed in the methanol-water mixture most likely because it occurs on the < 100-ps time scale, prior to the first point in the recording; the subsequent blue shift lasts only until the 2-ns time point, where a slow red shift begins. This final component is apparently too slow in water and in the glycerol-water mixture to be observed during the fluorescence time window. Given that the equilibrium structure of ZnCytc under the solvent conditions and temperature used in these experiments is that of the native state, the final red shift reports the beginning of a thermally driven refolding trajectory. The end point would be marked by a fluorescence spectrum from the ZnII-porphyrin that is isoenergetic with that observed at the end of the Q-band response. As indicated in Figure 4, which compares the models from Figures 2 and 3 for the dynamic Stokes shift following Soretband and Q-band excitation, the response from ZnCytc after the internal vibrational excitation from the ZnII-porphyrin has
three phases: (1) an initial fast red shift of the time-resolved fluorescence spectrum that accompanies the crossing and relaxation from the unfolding transition state, the initial unfolding motions of the polypeptide, and the entry of water molecules to the hydrophobic core; (2) a shift of the spectrum to the blue as the protein, solvent, and ZnII-porphyrin transiently assume a structure comparable to that obtained at equilibrium in the thermally unfolded state;16 and (3) a final recovery of the spectrum to the red that accompanies refolding of the protein and relaxation to the native state. The effect of the surrounding solvent on the structural dynamics in ZnCytc following Soretband excitation arises from two independent factors: the height of the activation-energy barrier between the native and unfolded states, and the hydrodynamic friction between the moving polypeptide and the surrounding solvent. Lowering the polarity from that of pure water lowers the activation energy, which is predominantly controlled by the difference in polarity between the hydrophobic core of the native fold and that of the surroundings.17 The slower dynamics in the glycerol-water mixture arise mostly from the increased solvent friction, whereas the faster dynamics in the methanol-water mixture are dominated by the decreased activation energy. The results suggest an important role for the solvent in limiting the rates of a protein’s unfolding and refolding trajectories. The unfolding trajectory’s initial stages involve the motions of an condensed structure that largely excludes the surrounding solvent, so solvent friction retards the reaction’s progress later on as the unfolding polypeptide is increasingly exposed to the bulk solvent.18,19 The effect of the viscosity on the rates of the first and second phases of response of ZnCytc following Soret-band excitation, then, indicates that even on the 100-ps time scale the protein assumes large-amplitude
22974 J. Phys. Chem. B, Vol. 110, No. 46, 2006 motions that accompany its expansion and solvation. The observed dynamics are consistent with the structure of the transition state resembling more the unfolded state than the native state. In contrast, the refolding trajectory begins with a fully solvated polypeptide, and the collapse to form the anhydrous hydrophobic core involves desolvation;20 solvent friction retards the early stages of the folding reaction, but the approach to the native fold is sterically hindered as the native protein-protein contacts increase in number. At each point in the unfolding or folding trajectory, however, solvent friction with the exposed portion of the protein would contribute to kinetic control over the formation of the product structure. This effect should be considered in a discussion of the roles of chaperones and crowding in the folding of proteins in the cellular environment.21 This work opens up a number of new possibilities for the study of protein folding and unfolding dynamics. The short time scale observed here for the unfolding reaction after preparation of the unfolding transition state permits the study of refolding dynamics under physiological solution conditions and with complete control of the temperature, so it should be possible to determine Arrhenius activation-energy parameters. The usual approach of perturbing the ensemble with either a change in temperature22 or solution composition23 involves, in contrast, a macroscopic change of the state of the surrounding solvent. An additional probe of the structure of the energy landscape can be obtained by tuning the excitation energy for the electronic chromophore; the amount of vibrational energy required to stimulate an unfolding reaction is a direct measure of the activation-energy barrier’s height from the native state. Acknowledgment. This work was supported by grants MCB-0091210 and MCB-0520002 from the Molecular Biophysics program of the National Science Foundation. Additional support for instrumentation was provided by the Michigan Structural Biology Center at Michigan State University, which is supported by the Michigan Life Sciences Corridor. The ZnCytc samples used in this work were prepared by Wayne Wright and Professor Jane Vanderkooi (School of Medicine, University of Pennsylvania). We thank Professor Gary Blanchard (Michigan State University) for the use of his laboratory’s picosecond fluorescence spectrometers. Supporting Information Available: Experimental procedures and methods of data analysis, Figure S1 (structure and
Letters solvent-contact surface of ferricytochrome c), Table S1 (polarity and viscosity parameters for the solvent systems used in ZnCytc solutions), Table S2 (fit parameters for the dynamic Stokes shift response shown in Figures 2 and 3). This material is available free of charge via the Internet at http://pubs.acs.org. Note Added after ASAP Publication. Paragraphs 2 and 5 of this paper have been corrected as a result of production errors. This paper was published ASAP on 10/21/2006. The revised version was reposted on 10/24/2006. References and Notes (1) Lazaridis, T.; Karplus, M. Science 1997, 278, 1928-1931. (2) Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. Science 1991, 254, 1598-1603. (3) Onuchic, J. N.; Luthey-Schulten, Z.; Wolynes, P. G. Annu. ReV. Phys. Chem. 1997, 48, 545-600. (4) Dill, K. A. Protein Sci. 1999, 8, 1166-1180. (5) Baker, D.; Agard, D. A. Biochemistry 1994, 33, 7505-7509. (6) Dobson, C. M. Nature (London) 2003, 426, 884-890. (7) Vanderkooi, J. M.; Adar, F.; Erecı´nska, M. Eur. J. Biochem. 1976, 64, 381-387. (8) Stratt, R. M.; Maroncelli, M. J. Phys. Chem. 1996, 100, 1298112996. (9) Miller, R. J. D. Annu. ReV. Phys. Chem. 1991, 42, 581-614. (10) Owrutsky, J. C.; Raftery, D.; Hochstrasser, R. M. Annu. ReV. Phys. Chem. 1994, 45, 519-555. (11) Bushnell, G. W.; Louie, G. V.; Brayer, G. D. J. Mol. Biol. 1990, 214, 585-595. (12) Miller, R. J. D. Acc. Chem. Res. 1994, 27, 145-150. (13) Lian, T.; Locke, B.; Kholodenko, Y.; Hochstrasser, R. M. J. Phys. Chem. 1994, 98, 11648-11656. (14) Lampa-Pastirk, S.; Lafuente, R. C.; Beck, W. F. J. Phys. Chem. B 2004, 108, 12602-12607. (15) Lampa-Pastirk, S.; Beck, W. F. J. Phys. Chem. B 2004, 108, 16288-16294. (16) Lesch, H.; Stadlbauer, H.; Friedrich, J.; Vanderkooi, J. M. Biophys. J. 2002, 82, 1644-1653. (17) Kauzmann, W. AdV. Protein Chem. 1959, 14, 1-63. (18) Ansari, A.; Jones, C. M.; Henry, E. R.; Hofrichter, J.; Eaton, W. A. Science 1992, 256, 1796-1798. (19) Ballew, R. M.; Sabelko, J.; Gruebele, M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 5759-5764. (20) Levy, Y.; Onuchic, J. N. Annu. ReV. Biophys. Biomol. Struct. 2006, 35, 389-415. (21) van den Berg, B.; Ellis, R. J.; Dobson, C. M. EMBO J. 1999, 18, 6927-6933. (22) Callender, R. H.; Dyer, R. B.; Gilmanshin, R.; Woodruff, W. H. Annu. ReV. Phys. Chem. 1998, 49, 173-202. (23) Roder, H.; Maki, K.; Cheng, H. Chem. ReV. 2006, 106, 18361861.
