Article pubs.acs.org/JPCB
Dynamic Stokes Shift of the Time-Resolved Phosphorescence Spectrum of ZnII-Substituted Cytochrome c Lynmarie A. Posey,* Ryan J. Hendricks,† and Warren F. Beck Department of Chemistry, Michigan State University, East Lansing, Michigan 48824, United States ABSTRACT: The dynamic phosphorescence Stokes shift (PSS) response of ZnII-substituted cytochrome c (ZnCytc) was detected using the time-resolved phosphorescence spectrum of the intrinsic ZnII-porphyrin chromophore, which senses the motions of the surrounding protein and hydration shell. The phosphorescence spectrum of ZnCytc exhibits resolved vibronic structure arising from in-plane deformations of the porphyrin macrocycle, as is also observed in the absorption and fluorescence spectra. As the emission time increases, the phosphorescence spectrum shifts to the red without incurring a significant change in vibronic structure or line shape, so the shift arises from dynamic solvation, the reorganizational motions of the protein and solvent that occur in response to formation of the first excited triplet state. A correlation time of 294 ± 14 μs was obtained from a single-exponential fit to the time dependence of the mean emission frequency of the T(0,0) peak in the phosphorescence spectrum. This time scale is consistent with a diffusive sampling of the native structure’s minimum due to global or collective conformational fluctuations. We suggest that studies of the PSS response sensed in proteins by an intrinsic probe will be informative of protein and hydration-shell dynamics over the microsecond− millisecond time regimes associated with biological function.
■
INTRODUCTION Proteins possess richly contoured potential energy surfaces surrounding the global minimum corresponding to the native state. Structural fluctuations allow proteins to sample multiple low-lying, nearby states on this multidimensional energy landscape1 separated by thermally accessible barriers. Low barriers between large numbers of conformational states with comparable energies make proteins inherently dynamic species, with equilibrium fluctuations playing an important role in biological function.2−4 The surrounding hydration shell further impacts the energy landscape. Coupling of protein and solvent motions strongly influences the underlying protein dynamics.5−12 Protein motions can be classified on the basis of the height of the barrier between conformational substates,1,2 the amplitude of the motion, and the characteristic time scale. Generally, collective or large-amplitude protein motions associated with biological function occur on time scales of microseconds to milliseconds.13 In the hierarchy of protein motions established by Frauenfelder and co-workers,6,10 these motions cause transformations between tier 0 conformational states, which are separated by barriers on the order of several kT.3 The height of the barrier separating tier 0 conformational states permits isolation of these states. Motions occurring on time scales shorter than microseconds are associated with local motion within the protein, such as side-chain rotation, loop motion,3 and coupled protein−hydration-shell motions.10,11 As the activation barrier decreases in magnitude, the rate of the equilibrium fluctuations between substates increases. Frauenfelder and co-workers6 have posited that nonequilibrium “functionally important motions” leading from one conforma© 2013 American Chemical Society
tional state to another structurally similar state are related to the equilibrium fluctuations between similar states and further that the fluctuation−dissipation relation14 provides a connection between “functionally important motions” and equilibrium fluctuations. The fluorescence Stokes shift (FSS) response15−19 senses protein fluctuations in terms of the dynamic Stokes shift of the time-resolved fluorescence spectrum from an intrinsic or extrinsic electronic probe. In condensed phases, the polar part of the FSS response arises from the dielectric relaxation of the medium in response to an optically induced change in a probe’s permanent dipole moment.15,20 The nonpolar part of the FSS response arises from changes in the probe’s size that accompany its optical excitation; its time scale is largely controlled by the viscosity of the medium.21 In the linearresponse limit, the FSS response provides an estimate for the probe’s ground-to-excited-state energy-gap time-correlation function. This function describes the characteristic loss of memory that arises from the fluctuations of the medium.20,22 Beck and co-workers have characterized the FSS response in the small proteins ZnII-substituted cytochrome c (ZnCytc)23 (Figure 1) and metal-free (or free-base, fbCytc) cytochrome c.24 In ZnCytc, the FSS response includes two distinct time scales. A fast response with a time constant of 250 ps was attributed to motion within the hydrophobic core of the protein, and a slow response characterized by a time constant of 1.45 ns was assigned to motion of the protein regions Received: June 6, 2013 Revised: October 17, 2013 Published: November 14, 2013 15926
dx.doi.org/10.1021/jp405611w | J. Phys. Chem. B 2013, 117, 15926−15934
The Journal of Physical Chemistry B
Article
phosphorescence emission spectrum with respect to time arising from a dynamic solvation response analogous to that observed for the fluorescence. In fact, the phosphorescence Stokes shift (PSS) response has been used to study solvation dynamics in viscous liquids and glasses near the glass transition temperature.45−47 Proteins have a number of properties common to glasses1,48,49 because they have comparably rugged energy landscapes with vast numbers of thermally accessible substates. Consequently, it is quite likely that the PSS response sensed in proteins by an intrinsic probe will be informative of protein and hydration-shell dynamics over the time regimes associated with biological function. In this Article, we report the first time-resolved phosphorescence spectra for ZnCytc and the first application of the PSS response to study equilibrium fluctuations in a protein. This work extends the previous FSS studies on ZnCytc of Beck and co-workers23 to the microsecond−millisecond regime that is relevant to collective protein motions, and it builds on the previous studies by Vanderkooi and co-workers40−42 of the phosphorescence of ZnCytc and metal-free cytochrome c. The PSS response in ZnCytc evidences a single correlation time in the microsecond regime that is likely to correspond to a global conformational motion involving crossing of a thermally accessible barrier in the native fold’s minimum on the energy landscape.
Figure 1. X-ray crystal structure of horse-heart ferricytochrome c (1hrc) displayed using ribbon (left) and surface (right) representations. Stick and space-filling representations are incorporated in the ribbon and surface renderings, respectively, to highlight the porphyrin, the methionine (Met 80) and histidine (His 18) that coordinate to the FeIII as axial ligands, and the cysteines (Cys 14 and 17) that anchor the porphyrin to the protein via thioether linkages. Cooperative unfolding/folding units within the protein are color-coded in order of increasing free energy (red → yellow → green → blue) following the scheme of Englander and co-workers.25−27
contacting the hydration shell.23 Removal of the metal ion from cytochrome c decreased the stability of the folded structure and increased the time constants for the two components of the FSS response to 1.4 and 9.1 ns, respectively.24 The longer time constants were attributed to an increase in the amplitude of motion in the destabilized metal-free cytochrome c based on application of a Brownian diffusion model over a onedimensional energy landscape with thermally accessible barriers.24 While the FSS response provides a sensitive probe of equilibrium protein fluctuations associated with local flexibility and the hydration shell, the intrinsic lifetime of the singlet state prevents this method from accessing the time regime that is sensitive to collective protein motions. The fact that many proteins exhibit room-temperature phosphorescence28 suggests the possibility of exploiting the long lifetime of the triplet state to extend optical methods to the longer time scales needed to access collective motions associated with biological function. Earlier phosphorescence studies involving proteins have largely exploited the triplet state lifetimes of tryptophan (Trp) residues29−39 or (metallo)porphyrin chromophores40−43 to detect motion via triplet quenching kinetics. Trp phosphorescence lifetimes, τp, detected at a single emission wavelength in the absence of quenchers have been used as probes of the local flexibility of proteins,30,31 while rate constants for quenching, kQ, by O2 and acrylamide32−35 have been correlated with the accessibility of the Trp residue to the bulk solvent and with conformational fluctuations that permit diffusion of the quencher through the protein. Changes in τp have also been used to follow H/D exchange in alkaline phosphatase.44 The use of Trp phosphorescence is best suited to proteins containing a single Trp residue; otherwise, the measured τp and kQ have contributions from multiple Trp environments. Even in proteins with just a single Trp residue, nonexponential phosphorescence decays are observed, which Steel and coworkers36 suggest is an “intrinsic property of the pure protein”. The presence of multiple phosphorescence decay components on the 1-ms and longer time scale typical of these measurements has been attributed to the presence of multiple conformers.36−39 We suggest in this contribution that the nonexponential behavior of single-wavelength phosphorescence transients in proteins may arise instead from a shift of the
■
EXPERIMENTAL SECTION Preparation of ZnCytc. Cytochrome c from horse heart (Sigma C2506) was demetalated and reconstituted with ZnII following the procedure developed by Vanderkooi and coworkers.40,50 The demetalation reaction was performed in liquid anhydrous hydrogen fluoride (Airgas) in a Teflon vessel on a home-built Teflon gas-handling system. UV−visible absorption spectroscopy was used to confirm the removal of the iron to produce the free-base form. Reconstitution of the metal-free porphyrin with ZnII was carried out in a 50 mM ammonium acetate buffer solution (pH 5.0) at 55 °C with a 10fold molar excess of zinc acetate (Sigma 379786). Progress of the ZnII-substitution reaction was tracked by monitoring the Qband region in the UV−visible absorption spectrum of the protein. The products of the demetalation and ZnIIreconstitution reactions were worked up with strong cationexchange chromatography on Whatman CM-52 media using procedures derived from those developed by Winkler and coworkers51 and Kostić and co-workers.52 The ZnII-substituted product was then equilibrated with a 25 mM sodium phosphate buffer solution (pH 7.0) by repeated dilution and concentration steps over a Millipore/Amicon YM10 ultrafiltration membrane. Approximately 100-μL aliquots of the concentrated samples were flash frozen in liquid nitrogen and stored at −78 °C. Sample Preparation. For use in phosphorescence spectroscopy, frozen ZnCytc samples were thawed and then diluted with 25 mM sodium phosphate buffer solution (pH 7.0) to obtain 2.5-mL samples with an absorbance of 0.30 at the maximum of the 0−0 peak of the Q-band (584 nm) in a 1-cm path length cell. Samples were handled under low-light conditions throughout the sample preparation process to prevent oxygen-promoted photodegradation.40 The samples were prepared in a quartz fluorescence cuvette sealed with a septum screw cap (NSG, Type 43) under a continuous argon purge (Airgas Ultra Pure Carrier, >99.9995%, < 0.5 ppb O2). An enzyme-based oxygen-scavenging system consisting of glucose oxidase (Type VII, Aspergillus niger, Sigma-Aldrich 15927
dx.doi.org/10.1021/jp405611w | J. Phys. Chem. B 2013, 117, 15926−15934
The Journal of Physical Chemistry B
Article
intensity was monitored by a reference channel consisting of a photodiode (Thorlabs PDA55) positioned at 180° to the fluorescence/phosphorescence collection axis and a Stanford Research Systems SR250 gated integrator. Normalization of the phosphorescence signal using the total fluorescence signal on a pulse-by-pulse basis compensated for fluctuations in the laser intensity and for any photodegradation of the sample that occurred during the experiment. LabVIEW (National Instruments) programs controlled the instrumentation and data acquisition.
G2133), catalase (bovine liver, Sigma-Aldrich C9322), and glucose42,53 was then introduced to further minimize the concentration of dissolved oxygen using gastight syringes. Measurement of a phosphorescence lifetime of ≥9 ms at 22 °C with excitation at 584 nm and detection at 730 nm was taken as an indication that the sample had been adequately deoxygenated. During phosphorescence experiments, a magnetic stir bar gently stirred the sample, and an argon purge of the sample headspace was maintained at a rate of 1−2 bubbles per second. Each sample was used for a maximum of two experimental runs. The absorption spectrum was recorded before and after each run to assess sample integrity; no broadening of the sharp Q-band features due to irreversible photodecomposition40 was observed. Absorption and Fluorescence Spectroscopy. Absorption spectra were acquired using a Hitachi U-2000 spectrophotometer (2-nm bandpass). Continuous-wave fluorescence spectra were obtained using a home-built spectrometer54 consisting of a Jobin-Yvon AH10 100-W tungsten−halogen light source, Jobin-Yvon H10 excitation monochromator (4-nm bandpass), Acton SP-150 emission spectrograph (2-nm bandpass), and a Jobin-Yvon Symphony liquid-N2-cooled CCD detector. A Quantum Northwest TLC50F sample holder with Peltier-effect temperature regulation held sample cuvettes in the fluorescence spectrometer. When fluorescence spectra are presented as a function of wavenumber, the fluorescence intensities collected at fixed spectral bandpass in wavelength units are multiplied by the square of the wavelength to preserve the area under the emission curve.55 Phosphorescence Spectroscopy. Time-resolved phosphorescence spectra were obtained from single-wavelength phosphorescence transients acquired using a home-built spectrometer. Excitation pulses were obtained from a QuantaRay PDL-2 dye laser (bandwidth
