Bacterial Photosynthetic Reaction Centers in Trehalose Glasses

Sep 14, 2010 - Bacterial Photosynthetic Reaction Centers in Trehalose Glasses: ... of trehalose matrixes in restricting the RC motional degrees of fre...
0 downloads 0 Views 3MB Size
Download PDF
J. Phys. Chem. B 2010, 114, 12729–12743

12729

Bacterial Photosynthetic Reaction Centers in Trehalose Glasses: Coupling between Protein Conformational Dynamics and Electron-Transfer Kinetics as Studied by Laser-Flash and High-Field EPR Spectroscopies Anton Savitsky,† Marco Malferrari,‡ Francesco Francia,‡ Giovanni Venturoli,*,‡,§ and Klaus Mo¨bius†,| Max-Planck-Institut fu¨r Bioanorganische Chemie, 45470 Mu¨lheim an der Ruhr, Germany, Laboratorio di Biochimica e Biofisica, Dipartimento di Biologia, UniVersita` di Bologna, 40126 Bologna, Italy, Consorzio Nazionale InteruniVersitario per le Scienze Fisiche della Materia, c/o Dipartimento di Fisica, UniVersita` di Bologna, 40127 Bologna, Italy, and Fachbereich Physik, Freie UniVersita¨t Berlin, 14195 Berlin, Germany ReceiVed: June 23, 2010; ReVised Manuscript ReceiVed: August 26, 2010

The coupling between electron transfer (ET) and the conformational dynamics of the cofactor-protein complex in photosynthetic reaction centers (RCs) from Rhodobacter sphaeroides in water/glycerol solutions or embedded in dehydrated poly(vinyl alcohol) (PVA) films or trehalose glasses is reported. Matrix effects were studied by time-resolved 95 GHz high-field electron paramagnetic resonance (EPR) spectroscopy at room (290 K) and low (150 K) temperature. ET from the photoreduced quinone acceptor (QA•-) to the photo-oxidized donor (P865•+) is strongly matrix-dependent at room temperature: In the trehalose glasses, the recombination kinetics of P865•+QA•-, probed by EPR and optical spectroscopies, is faster and broadly distributed as compared to that of RCs in solution, reflecting the inhibition of the RC relaxation from the dark- to the light-adapted conformational substate and the hindrance of substate interconversion. Similarly accelerated kinetics was observed also in PVA at a water-to-RC molar ratio 10-fold lower than in trehalose. Despite the matrix dependence of the ET kinetics, continuous-wave (cw) EPR and electron spin echo (ESE) analyses of the photogenerated P865•+ and QA•- radical ions and P865•+QA•- radical pairs do not reveal significant matrix effects, at either 290 or 150 K, indicating no change in the molecular radical-pair configuration of the P865•+ and QA•- cofactors. Furthermore, the field dependences of the transverse relaxation times T2 of QA•- essentially coincide in trehalose and PVA at 290 K. T2 is similar in these two matrixes and in the glycerol/water system at 150 K, implying that the librational dynamics of QA•- are also unaffected by the matrix. We infer that the relative geometry of the primary donor and acceptor, as well as the local dynamics and hydrogen bonding of QA in its binding pocket, are not involved in the stabilization of P865•+QA•-. We suggest that the RC relaxation occurs rather by changes throughout the protein/solvent system. The control of the RC dynamics and ET by the environment is discussed, particularly with respect to the extraordinary efficacy of trehalose matrixes in restricting the RC motional degrees of freedom at elevated temperatures. Introduction Proteins display a manifold of internal thermal motions spanning an extraordinary large time scale that encompasses more than 16 orders of magnitude.1 These conformational dynamics reflect complex energy landscapes characterized by an extremely large number of minima (conformational substates), organized in a hierarchy of tiers according to the barrier heights separating quasi-isoenergetic conformational substates (see, e.g., refs 2 and 3). A protein’s function is believed to be intimately related to its ability, under physiological conditions, to structurally fluctuate, sampling the ensemble of conformational substates, as dictated by the energy landscape.4,5 This concept of fluctuations between a large number of different conformations in a hierarchically structured energy landscape has been thoroughly studied over the past few decades, both theoretically and experimentally, for well-characterized protein * Corresponding author. E-mail: [email protected]. Phone: +39-051-2091288. Fax: +39-051-242576. † Max-Planck-Institut fu¨r Bioanorganische Chemie. ‡ Dipartimento di Biologia, Universita` di Bologna. § Dipartimento di Fisica, Universita` di Bologna. | Freie Universita¨t Berlin.

complexes, such as myoglobin as a paradigm system for globular proteins (see, e.g., Frauenfelder and co-workers6 and references therein) and the reaction center of photosynthetic bacteria as a paradigm system for membrane proteins.7-9 In particular, the photosynthetic reaction center (RC) from the purple bacterium Rhodobacter (Rb.) sphaeroides provides a privileged model system for exploring the relationships between electron-transfer (ET) processes and protein conformational dynamics.7-10 This membrane-spanning pigmentprotein complex catalyzes the primary photochemical events that initiate solar-energy conversion in photosynthetic bacteria. Within the RC from Rb. sphaeroides (see Figure 1), following photon absorption, the primary electron donor P865 (a bacteriochlorophyll a dimer) delivers an electron in ∼200 ps to the primary quinone acceptor QA (a ubiquinone-10), located ∼30 Å (center-to-center) from P865, thus generating the primary charge-separated state P865•+QA•- (for reviews, see refs 11 and 12). From QA•-, the electron is then transferred, in a conformationally gated process,13 to a secondary quinone molecule QB (also a ubiquinone-10),12 acting as a two-electron, two-proton acceptor upon successive photochemical turnovers of the RC14 to form the dihydroquinone QBH2.15 When electron donors to

10.1021/jp105801q  2010 American Chemical Society Published on Web 09/14/2010

12730

J. Phys. Chem. B, Vol. 114, No. 39, 2010

Figure 1. X-ray structure of the RC from Rb. sphaeroides R2661 and solvent molecules used as matrixes [PVA stands for poly(vinyl alcohol)]. Cofactor abbreviations (indices A and B refer to the two electron-transfer protein branches in the RC): P865 bacteriochlorophyll a (BChl) dimer, B BChl monomer, H bacteriopheophytin a, Q ubiquinone-10, Fe nonheme iron Fe2+.

P865•+ are not available, in QB-deprived RCs (or in the presence of inhibitors of the QA•--to-QB electron transfer), the RC returns to its ground state by direct electron tunnelling16 from QA•- to P865•+ (primary charge recombination). A variety of independent observations are consistent with the notion that the RC/solvent system responds to the electric field generated through primary light-induced charge separation by undergoing a conformational change.7,17-20 This electric-field induced structural relaxation, in turn, affects the ET kinetics. The most direct evidence of a close coupling between charge separation and RC conformational changes (light-induced and/ or thermally driven) has been provided by low-temperature studies of P865•+QA•- recombination kinetics.7,18 Kleinfeld et al.18 showed that the recombination kinetics are about 5-fold accelerated in RCs frozen to 77 K in the dark, as compared to room temperature, whereas they are markedly slowed when the RCs are frozen under illumination. This behavior indicates that RCs can be trapped at cryogenic temperatures in a dark-adapted or a light-adapted conformation, significantly differing in the stability of the charge-separated P865•+QA•- state. From the nonexponential, strongly distributed kinetics, measured in both the dark- and light-adapted frozen states, it was inferred that both conformational states actually consist of large ensembles of substates, giving rise to a continuous distribution of rate constants. In a systematic study,7 a model of the coupling between ET and RC dynamics was developed in which protein relaxations and fluctuations between conformational substates were cast into a single parameter, the energy gap between the charge-separated (P865•+QA•-) and neutral (P865QA) state, which was mapped on a single conformational coordinate that accounts for the RC’s diffusive internal motions. At physiological temperatures, following primary charge separation to P865•+QA•-, the RC protein light-adapts by a rapid structural relaxation that solvates the altered charge distribution. The resulting decrease in the energy gap between the P865•+QA•- and P865QA states is reflected in a slowing of the recombination process, that is, a stabilization of charge separation. At the same time, at room temperature, the RC rapidly explores the distribution of conformational substates, thus averaging the corresponding distribution of ET rates over the time scale of charge recombination. This results in the almost exponential kinetics of P865•+QA•recombination observed under physiological conditions. Following an approach that complements low-temperature studies, we previously analyzed the effects of the protein environment on ET when embedding the RC into dehydrated glassy matrixes formed at room temperature by the disaccharide trehalose (R-D-glucopyranosyl R-D-glucopyranoside;21-23 see

Savitsky et al. Figure 1). This sugar, which exhibits an extraordinary efficacy in the preservation of biostructures under adverse environmental conditions,24,25 is found in large amounts in organisms that can survive conditions of extreme drought and high temperatures, entering for a long time a state of reversibly suspended metabolism (anhydrobiosis).26-29 In particular, it was shown by optical absorption spectroscopy that, in strongly dehydrated trehalose matrixes, the kinetics of P865•+QA•- recombination accelerates and become broadly distributed,21,23 behaving at room temperature similarly to what was observed at cryogenic temperature in a glycerol/water mixture.7 From this similarity, it has been inferred that, in dehydrated trehalose matrixes, the thermal fluctuations between conformational substates, as well as relaxation from the dark-adapted to the light-adapted state, are strongly reduced even at room temperature over the time scale of charge recombination. Such an interpretation is supported by a large body of spectroscopic studies,30-33 as well as molecular-dynamics simulations,34 mainly performed on myoglobin (Mb), indicating a strong hindrance of the internal dynamics when the protein is incorporated into a dehydrated trehalose glass. In addition to reversibly affecting the kinetics of P865•+QA•- recombination, progressive drying of the RC/ trehalose sample also leads to the arrest of QA•--to-QB electron transfer in a progressively increasing fraction of the RC population,22 again mimicking at room temperature an effect observed in RCs frozen in the dark in water/glycerol mixtures.18 We have proposed that a water-mediated network of hydrogen bonds locks the RC surface to the trehalose matrix, thus coupling the internal degrees of freedom of the protein to those of the water/sugar matrix.35 The tightness of this coupling (“slaving”5) appears to depend on the nature of the matrix-forming molecule, because sucrose, despite its structural similarity with trehalose, was found to be almost ineffective in blocking the RC dynamics coupled to charge recombination and in protecting the RC against thermal denaturation.36 Effects similar to, although less pronounced than, those observed on ET in RC/trehalose samples, were also found in RCs incorporated into poly(vinyl alcohol) (PVA) matrixes, but only at significantly lower contents of residual water in the embedding matrix.37 This points to an important role that the specific interaction between the matrix and the protein hydration layer plays in modulating the slaving of the protein dynamics to those of the matrix. As to the present high-field electron paramagnetic resonance (EPR) experiments on the radical ions P865•+ and QA•-, as well as on the charge-separated radical-pair state P865•+QA•- in RCs, some introductory remarks might be appropriate: To understand the structure-dynamics-function relation of reacting proteins, the spatial and electronic structures not only of the initial and final states but also of the intermediate states are of primary concern. Standard X-band (9.5 GHz) EPR spectroscopy of protein systems often faces problems of spectral resolution and sensitivity. In this situation, higher Zeeman fields and resonance frequencies are needed to separate field-dependent from fieldindependent spin interactions and to select specific molecular orientations from the random distribution of the molecules in frozen-solution samples. For such disordered systems, millimeter and submillimeter high-field EPR methods, in both continuouswave (cw) and pulse modes, offer powerful tools for obtaining sufficient spectral and orientational selectivity of the radicals and radical pairs to provide the desired structural and electronic information (for an overview, see ref 38). In view of the present study, the merits of high-field EPR spectroscopy can be summarized as (i) high absolute detection sensitivity and time resolution because of the high resonance frequency, (ii) high

Bacterial Photosynthetic RCs in Trehalose Glasses spectral and orientational resolution of even small anisotropies of the g tensor or T2 relaxation time that become observable in high Zeeman fields, (iii) high sensitivity of the g and hyperfine tensor components toward polarity and hydrogen-bonding properties of the solvent matrix. This has been demonstrated for many protein systems, notably for light-induced ET intermediates of bacterial photosynthetic RCs (see, for example, refs 38-54). The information obtained goes beyond what is known from X-ray crystallography and provides pieces of information on structure and dynamics essential for an understanding of the biological transfer process. Hence, for the in-depth characterization of complex paramagnetic intermediates with nanosecond lifetimes in photoinduced reaction cycles, cw and time-resolved high-frequency EPR techniques are the methods of choice. In particular, the electron Zeeman, electron-nuclear hyperfine, and electron-electron dipolar interactions are excellent tools for probing subtle cofactor-protein interactions and their changes during the reaction cycle. Hyperfine interactions (in addition to g tensors, one of the prime sources for information on cofactor-protein interactions involving H-bonding) can be accurately measured by cw and pulsed electron-nuclear double resonance (ENDOR) and electron spin-echo envelope modulation (ESEEM); see, for instance, ref 55. Pulsed electron-electron double resonance (PELDOR) spectroscopy at the W-band (95 GHz) is very powerful for measuring distances and orientations of radicals in weakly coupled electron spin systems, for example, transient radical pairs P865•+QA•- with their large interspin distances.52 Magnetic resonance spectroscopy can probe not only the static structural details of a molecule but also details of its dynamic properties.50,56,57 If the motion is on the time scale of the EPR experiment, spin relaxation and, thereby, line broadening can be observed in the cw EPR spectrum. In many cases, the analysis of this effect is obscured by static (“inhomogeneous”) broadening effects from unresolved hyperfine interactions or g strain. Therefore, pulsed spin-echo techniques, which can separate dynamic and static contributions to the spectrum, are the proper choice to study molecular motion.58,59 At high Zeeman fields, one can spectroscopically select those molecules in the frozen-solution sample that are oriented with one of their principal g-tensor axes along the external magnetic field B0 to measure their orientation-dependent phase-memory time Tm.46 For monoexponential echo decays

S(2τ, B0) ) S0 exp[-2τ/Tm(B0)] Tm can be identified as the transverse relaxation time T2 at the magnetic-field values B0 corresponding to the gxx, gyy, and gzz orientations. In the fast-tumbling limit (“Redfield limit”), the relaxation rate 1/T2 is proportional to the square of the Zeeman field, making high-field ESE particularly sensitive to small changes of local fluctuations of the effective g values and corresponding Larmor frequencies. In other words, the T2 relaxation characteristics of a cofactor in the protein environment of its binding site, as revealed by high-field spin-echo detected by EPR spectroscopy in terms of both the magnitude and anisotropy of T2, sensitively probes the local dynamics of the binding site with its characteristic hydrogen-bonding network to the cofactor. In the present work, we describe a collaborative study of laserflash and high-field EPR spectroscopies on RCs from Rb. sphaeroides R26, in most experiments Fe2+ f Zn2+ substituted, embedded in different matrixes, specifically in PVA films and trehalose glasses with varying water contents. As outlined above,

J. Phys. Chem. B, Vol. 114, No. 39, 2010 12731 kinetic analysis of P865•+QA•- recombination by optical absorption spectroscopy indicates that the RC conformational dynamics at room temperature are severely restricted in dehydrated trehalose glasses and, to a lesser extent, also in PVA matrixes under extreme dehydration. It was concluded that, at room temperature, the dynamics of the protein was regulated by the dynamics of the external matrix (for a review, see Cordone and co-workers35). These observations prompted us to take advantage of the increased spectral and time resolution, as well as the increased orientation selectivity, of time-resolved 95 GHz highfield EPR spectroscopy with the aim to trace, beyond what is feasible with optical spectroscopy, potential changes of molecular conformation and dynamics of the cofactor-protein complex in RCs embedded in different solvent matrixes such as water, PVA film, and dehydrated trehalose saccharide with controlled residual water content. Specifically, photogenerated P865•+ and QA•- radical ions as well as P865•+QA•- radical pairs were studied to probe, at low and room temperatures, potential changes of the coupling between conformational protein dynamics and ET rates in these matrixes and to compare the results with those deduced from laser-flash spectroscopy. We aimed at an understanding of the connection between the astounding ability of trehalose glasses to inhibit biological RC functions at elevated temperatures, typical for hot deserts, and the “tiers” in the energetic hierarchy and functional relevance of the various “classes” of protein fluctuations5 for controlling light-induced and thermally activated photosynthetic processes. Such an inhibition of biological function does not occur (or occurs much less efficiently) by chemically very similar disaccharides such as sucrose.36 We wanted to determine, on the molecular level, whether any changes in structure and dynamics of the ET cofactor sites occur when the matrix is changed from water to dry PVA and trehalose or, alternatively, whether the inhibition mechanism is rooted in a trehalose-specific coating of the protein surface. Hence, from the combination of spectroscopic techniques, we expected new insights, at a nanoscopic scale, into the specific coupling between the dynamics of the trehalose bulk solvent, of the solvent shell of the whole RC, and of the local QA•binding site (e.g., quinone librations) and the ET kinetics of the RC/matrix system. In conventional matrixes (i.e., without trehalose or other sugars working similarly), a slowing of the protein dynamics is obtained by lowering the temperature. To keep the RC protein intact at low temperatures, cryoprotectants such as ethylene glycol or glycerol are commonly used in X-ray diffraction and spectroscopic studies. This procedure, however, might introduce artifacts into the RC structure with respect to cofactor and side-chain positions.60-64 In this respect, the use of trehalose matrixes at room temperature, to condition the RC protein dynamics, represents an attractive alternative to the lowtemperature approach for structure determination. Materials and Methods The RCs from Rb. sphaeroides R26 were purified according to the method of Gray et al.65 The RC concentration was determined spectrophotometrically by the absorbance at 802 nm (ε802 ) 288 mM-1 cm-1). The ratio of the absorption at 280 to that at 800 nm was 1.3. To avoid fast spin relaxation of the ubiquinone UQ-10 radical anions, the high-spin nonheme Fe2+ (S ) 2) was removed and replaced by diamagnetic Zn2+ as previously described,66 except for the following modifications: The RC used was NaCl-free, and only 1 equiv of UQ-10 was added during the first incubation step in the presence of o-phenanthroline. This procedure resulted in a partial replace-

12732

J. Phys. Chem. B, Vol. 114, No. 39, 2010

ment of Fe2+ by Zn2+: The ratio of the QA•--to-P865•+ spectral intensities in the Q-band EPR spectra (see Results, Figure 2) showed that Fe2+ was retained in about half of the RC population; evaluation of the Fe2+ and Zn2+ content by inductively coupled plasma atomic emission spectroscopy (ICPAES) before and after the replacement showed additionally that the ∼50% decrease of the Fe2+/RC ratio was compensated by an approximately corresponding increase of the Zn2+/RC ratio. The absence of a significant fraction of Fe-depleted RCs, lacking bound Zn2+, was confirmed by the comparable extents of photoxidized P865•+ induced by a saturating laser flash in the native and Zn2+-replaced RC preparation. On the contrary, a considerably reduced efficiency of light-induced charge separation is expected in Fe-depleted RCs that have no metal ion occupying the Fe site.67 In our preparation, the fraction of photoactive RC after the replacement was larger than 90%. This fraction was estimated from the absorbance change induced by a train of six laser pulses at 422 nm,68 assuming a differential extinction coefficient ∆ε422 ) 17.5 mM-1 cm-1. The Zn2+-RC preparation (at a concentration of about 50 µM) was supplemented with stigmatellin, one of the most potent inhibitors of the QA•--to-QB electron transfer,69-71 at a stigmatellin-to-RC molar ratio of 2. The suspension was subsequently concentrated by ultrafiltration using a 100 kDa cutoff cartridge (Amicon) up to a RC concentration of 335 µM. An aliquot, frozen in liquid nitrogen after addition of glycerol at a final concentration of 20% v/v, was stored at -80 °C for EPR and optical measurements of RC/water samples. The RC/trehalose glassy sample for EPR measurements was prepared according to the following procedure: A volume of 50 µL of the concentrated (335 µM) RC was mixed with 100 µL of 1.67 M trehalose to obtain a sugar-to-RC molar ratio of 104. Trehalose (>99% purity) was purchased from Hayashibara Shoij (Okayama, Japan). The RC/trehalose solution was layered on an optical window and dried in a desiccator under N2 flow at room temperature. The glassy sample obtained in about 4 h of flux was further dehydrated by alternating incubation under N2 atmosphere at room temperature for 18 h and direct N2 flux for 6 h; this drying cycle was repeated three times. The content of residual water in the dehydrated RC/trehalose and RC/PVA samples was estimated by visible and near-infrared spectroscopy on a Perkin-Elmer (Fremont, CA) Lambda 19 spectrometer. The molar H2O-to-RC ratio was determined from the area of the combination band of water at ∼1940 nm, using the RC absorption band at 802 nm as an internal standard (for details, see refs 21 and 72). The RC/PVA sample for EPR measurements was prepared by mixing 75 µL of the concentrated RC (335 µM) with an equal volume of 10% w/v PVA (Fluka, Mw ≈ 130000). To form the film, the same drying procedure as used for the RC/trehalose sample was employed. Given that optical measurements were not possible in these samples, because their absorbance was too high, RC/trehalose and RC/PVA samples, with 3- and 4-times lower RC concentrations, respectively, were prepared in parallel. These samples were characterized by the same trehalose-toRC and PVA-to-RC molar ratios as those prepared for EPR measurements. Also, the same drying procedure was applied for the two matrixes. RC/trehalose glassy matrixes were stored at room temperature, whereas RC/PVA films were kept at 4 °C. The kinetics of P865•+QA•- charge recombination was monitored by optical laser absorption spectroscopy at 422 nm68 with an apparatus of local design.36 Pulsed photoexcitation was provided by a frequency-doubled Nd:YAG laser (Handy 710,

Savitsky et al. Quanta System, Milano, Italy) that delivered 150 mJ pulses of 7 ns width. From 2 to 25 kinetic signals were averaged, depending on the sample, with a dark adaptation of at least 1 min between successive single photoexcitations. Nonlinear leastsquares minimization and numerical determination of confidence intervals of the kinetic parameters were performed as previously described.73 For EPR measurements, the dehydrated, fragile RC/trehalose glasses described above were crumbled into small flakes, which, by application of reduced pressure, could be quickly inserted into the cylindrical quartz capillaries (i.d. ) 0.6 mm) for the 95 GHz EPR cavity. The small strips from dehydrated RC/PVA film (1 mm in length, 0.4 mm in height) were fitted into the EPR capillary. The capillary was positioned in the EPR cavity with the strip surface perpendicular to the light beam direction to ensure the maximum excitation of the sample. High-field EPR and ESE measurements were performed on a laboratory-built W-band (95 GHz/3.4 T) spectrometer that had been optimized for a variety of cw and pulse experiments, as described previously.38,51 The spectrometer was equipped with a TE011 optical transmission microwave (mw) cavity with an unloaded quality factor QU ) 5000 (empty) for optimum detection sensitivity. The samples were contained in thin-walled quartz capillaries. For optical sample irradiation, the light was guided to the center of the cavity through a quartz fiber of 0.8-mm diameter. The electron transfer was initiated by singlet excitation of the primary donor at 532 nm using a frequency-doubled Nd:YAG laser assembled from various commercial components (5-ns pulse length, 1-10 Hz repetition rate, 0.5 mJ/pulse on the sample surface) or at 690 nm using a cw diode laser (25 mW output, 10 mW on the sample surface). The ESE measurements were performed using the standard mw pulse sequence for primary spin-echo generation: (tp)x,-x-τ-(2tp)-τ-echo. The pulse length tp of the π/2 mw pulses was generally set to 30 ns. To acquire field-swept ESE-detected spectra, the pulse separation time τ was fixed to 150 ns. The two-pulse echo decay traces were recorded by incrementing τ from a starting value τ0 ) 50 ns. The timeresolved cw W-band transient EPR (TREPR74) measurements of short-lived paramagnetic intermediates were performed without field modulation using the direct-detection technique with a time resolution of 10 ns. The P865•+QA•- chargerecombination kinetics data were obtained for radical pairs with thermally equilibrated spin polarization by recording their transient EPR absorption after a laser flash, using 30 kHz field modulation and lock-in detection with 1 ms time resolution. Temperature control was achieved by a gas-flow cryostat housing the cavity probehead. The accumulation time for an EPR spectrum was typically 300 s depending on the sample temperature and solvent matrix. All experimental spectral analyses and simulation procedures were performed using the EasySpin toolbox75,76 for the Matlab program package. Results In this section, we summarize the results of our W-band EPR investigations of Zn-substituted RCs from Rb. sphaeroides R26 in water/glycerol, PVA, and trehalose matrixes at 290 and 150 K. The aim of the study was to understand the role of the matrix properties, in terms of both composition and dynamics, for the hierarchy of coupling between the fluctuations of the QA•binding site, the solvent matrix, and the ET characteristics in charge-recombination process of P865•+QA•-. We anticipated new insights into this coupling between molecular dynamics and

Bacterial Photosynthetic RCs in Trehalose Glasses

J. Phys. Chem. B, Vol. 114, No. 39, 2010 12733

Figure 2. (a) W-band cw EPR spectra of dark-adapted Zn-RCs from Rb. sphaeroides in water at room temperature under continuous laser-diode illumination at 690 nm (ON, blue trace) and after the laser was switched off (OFF, green trace) recorded using field modulation at 8.2 kHz with an amplitude of 50 µT. (b) Difference spectrum (ON - OFF) overlaid with the spectrum simulation, giving g(QA•-) ) [2.00647 2.00525 2.00223]; g(P865•+) ) [2.00325 2.00240 2.00189]. The EPR-intensity ratio of QA•- to P865•+ is 0.52, revealing a Zn-substitution grade of 52%. Asterisks (*) mark the line positions of the field-calibration standard (Mn2+).

Figure 3. (a) W-band cw EPR spectra of dark-adapted Zn-RCs in PVA at room temperature under prolonged continuous illumination at 690 nm and after the diode laser was switched off. (b) Laser OFF spectrum (green) overlaid with its simulation (red), resulting in g(QA•-) ) [2.00652 2.00531 2.00224]. Asterisks (*) mark the Mn2+ line positions of the field-calibration standard.

biological function to be obtained that widen our knowledge deduced from previous optical spectroscopic studies. Because the sample requirements for optical and EPR measurements are quite different, for example, in terms of concentration and paramagnetic perturbations and/or impurities, in a first step, we characterized the various RC/matrix samples according to their EPR behavior. Equivalent samples, in terms of relative matrix/ protein composition, were subsequently analyzed by laser optical spectroscopy and used for estimating the hydration level by visible-NIR (near-infrared) spectroscopy. W-Band cw EPR Spectroscopy at 290 K. Figure 2a shows the cw EPR spectrum of Zn-substituted RCs in water solution at room temperature during continuous illumination (690 nm) and after illumination. Under illumination, the EPR spectra of the QA•- and P865•+ radical ions appear. After the diode laser is switched off, only very small photoaccumulated signals of QA•and P865•+ remain (95% at both 290 and 150 K. Apparently, the RC-preparation procedures for the EPR measurements leave the structure and dynamics of the protein intact. (2) Illumination of the RC/PVA and RC/trehalose samples at 290 K results in a strong steady-state signal of a QA-type radical anion. (At present, our EPR experiments cannot reveal the exact structure of the photoaccumulated radical species, but additional ENDOR experiments are planned for further structure information.) After about 1 h of illumination, only about 20% of the RCs are still intact, that is, show cyclic ET behavior. The process of QA-type radical photoaccumulation is independent of RC concentration, that is, not due to photoheating of the sample. After the addition of water (∼50%), the steadystate signal of the QA-type radical disappears, and normal behavior as in RC/water solutions is restored. The photoaccumulation of a QA-type radical anion is reflected in the comparable decrease of the optically detected P865•+ signal recorded in room-temperature trehalose glasses following repetitive laser excitation. (3) At 150 K, the behaviors of all three RC/matrix samples are similar. The QA-type radical photoaccumulation process (in RC/PVA and RC/trehalose samples) does not take place. The P865•+QA•- charge-recombination kinetics are similar for all three matrixes. (4) At 290 K, the P865•+QA•- recombination kinetics probed by EPR and optical absorption are in excellent agreement. They differ significantly in the RC/trehalose and RC/PVA matrixes, as compared to the water/glycerol RC solution, which is in line with previous optical measurements.21,37,73 (5) At both 290 and 150 K, there is no significant difference in the electron spin T2 relaxation dynamics due to librational

12738

J. Phys. Chem. B, Vol. 114, No. 39, 2010

fluctuations of QA•- in all three RC/matrix samples (except for a factor of about 2 in the T2 time for the RC/water sample at 150 K). (6) The shapes of the spin-correlated P865•+QA•- radical-pair spectrum in all three samples are similar at 150 and 290 K. Thus, from the high-field EPR and laser-flash absorption measurements, it must be concluded that the motional dynamics of QA•- in its binding site remains the same in the three solvent matrixes. This suggests that the observed matrix dependence of the P865•+QA•- recombination kinetics in dark-adapted RCs at room temperature (RC/water versus RC/PVA and RC/ trehalose; see Figure 5) is not due to changes in the local environment of the QA•- cofactor in its binding site. Rather, it probably originates in the high rigidity of the dry trehalose glassy matrix firmly coating the surface of the RC already at room temperature. Discussion As outlined in the Introduction, a rich body of independent experimental evidence indicates that, in room-temperature solutions, the RC protein rapidly fluctuates among an extremely large number of conformational substates.5 This is essential for biological function because such a highly structured energy landscape with its multifrequency fluctuations among the substates provides the reservoir of entropy that dictates the extent to which a given quantity of thermal energy is available for doing biologically useful work. Following light-induced primary charge separation, the RC relaxes from a dark-adapted to a lightadapted state, thereby stabilizing the P865•+QA•- state. Despite the fact that a central goal in photosynthesis research over the past few decades has been to relate structure, dynamics, and function, the structural and dynamical basis of the relaxation by which the RC responds to the light-induced generation of an electric field around P865•+ and QA•- remains elusive at present. To fit the nonexponential kinetics observed in light-frozen and dark-frozen RCs, Kleinfeld and co-workers18 assumed a distribution of distances between P865•+ and QA•- and proposed that the RC relaxation from the dark-adapted to the light-adapted state involves an increase of the distance between P865•+ and QA•-, estimated to be in the range of 1 Å. At variance with this assumption, on the basis of FTIR studies,19 it was suggested that highly localized conformational changes of the protein matrix near the QA site might play a key role in assisting the stabilization of the primary charge-separated state; alternatively, it has been proposed that the protonation of amino-acid residues located at a large distance from QA (>16 Å) can participate in the stabilization of the QA•- quinone anion.89 This appears reasonable because the one-electron states of the quinones in the primary photosynthetic ET chain, QA•- and QB•-, are accompanied by protonation reactions, but the H+ binding targets are protein residues rather than the quinone cofactors themselves.90 A time-resolved optical spectroscopy study performed on wild-type RCs and four site-specific mutants with widely modified free-energy gaps for P865•+QA•- recombination has led to the suggestion to associate conformational relaxation with subtle rearrangements of the cofactors within their cavities.9 Finally, the involvement of water molecules, weakly bonded to the RC and perturbed by light-induced QA reduction (see below), has recently been invoked as playing a major role in the relaxation process that stabilizes the primary charge separation.91 With the aim to contribute toward the clarification of this confusing scenario of strategies for stabilizing the charge-

Savitsky et al. separated primary radical pair P865•+QA•-, we have compared the behavior of RCs embedded in water solution and incorporated into dehydrated matrixes; the latter were shown to hinder the RC relaxation following P865•+QA•- formation and to trap the RC distributed over a large ensemble of conformational substates.21,37,73 In fact, the embedding of the RC in a dehydrated trehalose glassy matrix or in a strongly dried PVA film gives rise, at room temperature, to significantly accelerated P865•+QA•recombination kinetics, as compared to water-glycerol solutions. In the present work, such matrix effects were consistently probed in parallel by laser optical absorption spectroscopy and by direct-detection EPR microwave absorption; it is satisfying that similar kinetics were obtained for such vastly different spectroscopic methods (see Table 1). Both measurements show additionally that, whereas the P865•+QA•- decay is essentially exponential with a unique rate constant in aqueous solutions, a distribution of rate constants is needed to fit the recombination kinetics in the dehydrated trehalose and PVA matrixes. This indicates that, in the latter matrixes, over the time scale of P865•+QA•- recombination, trapping of conformational substates occurs already at room temperature. We observed slightly faster and rate-distributed kinetics in the RC/trehalose sample, as compared to the RC/PVA one, despite the much lower content (by 1 order of magnitude) of residual water in the PVA matrix. The values measured for the kinetic parameters and for the residual water content in the two matrixes fit well the previously observed relationship between the kinetics of charge recombination and the hydration level.21,37,73 Interestingly, analysis of the charge-recombination kinetics, probed by EPR spectroscopy at 150 K, shows that the average rate constant and the width of the rate distribution are further increased in the same trehalose and PVA matrixes at cryogenic temperatures (see Table 2). These values are significantly larger than those obtained for the water-glycerol system at the same temperature, which, in turn, are in good agreement with the values measured by optical laser spectroscopy at cryogenic temperatures.7 It appears, therefore, that the observed inhibition of relaxation from the dark-adapted to the light-adapted state in the trehalose and PVA matrixes, although quite significant, was not complete at room temperature. Apparently, some conformational dynamics, coupled to the ET process, survived at room temperature in the examined samples, most likely because of their residual water content. Indeed, we have shown that a more extensive dehydration of RC/trehalose matrixes leads at room temperature to larger k and σ values, which are comparable to those measured in glycerol-water systems at cryogenic temperatures.21,73 As another important result, we point out that W-band EPR experiments performed on both RC/PVA and RC/trehalose samples at room temperature revealed the photoaccumulation of a stable QA-type radical (probably a quinone anion radical; see above). A simple explanation for the photoaccumulation of a QA-type radical is that, in the subpopulation of RCs in which photoaccumulation takes place, reduction of the flash-oxidized P865•+ by an exogenous electron donor occurs faster than recombination between P865•+ and QA•-. Such an adventitious electron donor is expected to eventually be present in the liquid sample at very low concentrations, thus being unable to reduce P865•+ through a collisional process at a rate that efficiently competes with P865•+QA•- recombination. Correspondingly, no photoaccumulation of a QA-type radical is detected in water/ RC solutions. In the solid PVA and trehalose matrixes, because of the extensive dehydration, the concentration of the putative electron donor to P865•+ would be greatly increased, possibly

Bacterial Photosynthetic RCs in Trehalose Glasses leading to a close complex between the exogenous electron donor and the RC and enabling a fast electron donation to P865•+. If this is the case, we have to assume that the electron transfer is thermally activated, because no photoaccumulation of a QAtype radical was observed in the RC/PVA and RC/trehalose samples at 150 K. The photoaccumulation of a QA-type radical as revealed by W-band EPR measurements is consistent with previous observations by flash absorption spectroscopy: In progressively dried RC/trehalose glasses, repetitive photoexcitation led, in fact, to a decreased amplitude of the flash-induced optically detected P865•+ signal.21 This decrease had been attributed to a fraction of RCs in which P865•+QA•- was not formed, and charge recombination occurred on a subnanosecond time scale from the photoreduced bacteriopheophytin. Because this decay was much faster than the time resolution of the optical measurements, it resulted in a decrease of the signal of the photo-oxidized P865•+. The present EPR experiments are fully consistent with this observation by optical spectroscopy, because photoaccumulation of a radical in a subpopulation of RCs will clearly prevent the formation of the primary charge-separated state P865•+QA•-, leading to photoreduced bacteriopheophytin in that subpopulation. Upon dissolving the glassy sample, the EPR-detected photoaccumulated QA-type radical was no more observed in the dark. Consistently, the dissolution of dehydrated RC/trehalose glasses fully restored the original extent of P865•+ photo-oxidized by a laser flash, as detected by optical spectroscopy.21 Rather unexpectedly, the W-band EPR and ESE experiments on the photogenerated P865•+ and QA•- radical ions as well as on P865•+QA•- radical pairs when performed on the same glycerol/water, trehalose, and PVA samples, which are characterized by different recombination kinetics at room temperature (see Table 1), did not show any significant matrix-induced effects, in either the spectra or the echo decays. The transient W-band EPR signals of the spin-correlated P865•+QA•- radical pair, measured at 290 K in the dehydrated PVA and trehalose matrixes and at 150 K in the glycerol-water system, did not exhibit significant differences when normalized in amplitude. This suggests that incorporation of RCs into the dehydrated trehalose and PVA matrixes does not significantly change the radical-pair configuration of the P865•+ and QA•- cofactors as compared to RCs in water-glycerol; that is, the glassy matrixes do not cause structural distortions of the cofactor binding sites of the RC. Therefore, the effects observed on the kinetics of electron transfer cannot be simply traced back to matrix-induced structural changes, but rather have to be related to the dynamics and/or energetics of the protein/solvent system, with its characteristic fluctuations between conformational substates of the energy landscape, as was originally proposed by Palazzo et al.21 and Francia et al.23 From these observations, we also infer that the RC relaxation from the dark-adapted to the light-adapted state occurring under physiological conditions7,18 is quite unlikely to involve any change in the geometry of the cofactors. Moreover, the ESE experiments showed that the field dependences of the anisotropic T2 spin-spin relaxation times are essentially coincident in the trehalose and PVA matrixes at 290 K and quite similar in these two matrixes and in the glycerol-water system at 150 K. This indicates that the relaxation dynamics of QA•- (dominated, under high-field conditions, by its g-tensor anisotropy) are essentially the same in the three solvent matrixes, implying, in turn, that the hydrogen-bond network in the QA binding pocket is also unaffected by the different matrixes. Therefore, this hydrogenbond network does not appear to be involved in the conforma-

J. Phys. Chem. B, Vol. 114, No. 39, 2010 12739 tional response of the RC to the light-induced electric field generated by the formation of the charge-separated radical pair P865•+QA•-. Taken together, the results imply that the relative geometries of the primary donor and acceptor, as well as the local environment of QA in its binding pocket, do not differ significantly in the light-adapted state, which is observed at room temperature in solution RCs, with respect to the dark-adapted state, which is trapped in both the dehydrated trehalose and PVA matrixes at room temperature and in water-glycerol mixtures at cryogenic temperature. These findings suggest that the structural and dynamical basis of the RC relaxation processes following charge separation resides rather in changes throughout the protein-solvent system that do not involve the geometry or local environment of the cofactors. Thus, it appears possible that the stabilizing relaxation process consists mainly in a reorientation of amino-acid side chains in response to the lightinduced electric field around P865•+ and QA•-. However, if this is the case, then these side chains are unlikely to be in the close vicinity of the QA binding pocket. Otherwise, one would expect that, at variance with what was observed by the high-field ESE experiment on T2 anisotropy, the QA•- environment and dynamics would be significantly altered by incorporation of the RC into solid matrixes, which substantially inhibit the relaxation processes, as is shown by the kinetics of charge recombination. As an alternative explanation, in addition to solvation effects due to residues of the RC protein, a major energetic contribution to stabilization of the primary charge-separated state is also expected from internal water molecules interacting with the RC cofactors. In a recent work, Iwata et al.91 observed light-induced FTIR spectral changes, associated with the photoreduction of QA, in the O-H stretching region (3700-3500 cm-1), which is characteristic of weakly hydrogen-bonded water. The authors proposed that the orientation of these “free” water molecules plays a major role in stabilizing the P865•+QA•- state by dielectric screening of the light-induced electric dipole of the chargeseparated radical pair. As compared to previous crystallographic RC structures, such as that of Ermler et al. with 2.65-Å resolution,92 crystal structures with significantly better resolution have become available by now, allowing increasing numbers of unbound water molecules in the protein to be identified. Recently, Koepke et al. reported a Rb. sphaeroides RC structure with a resolution of 1.87 Å, which is the best resolution obtained so far,64 and indeed, the number of modeled water molecules that was included in the structure could be extended from 16092 to 430: Many newly assigned unbound water molecules are clustered at the cytoplasmic surface of the RC (where the quinone acceptors are located), and several unbound water molecules are buried in its membrane-spanning region. Hence, unbound water molecules are expected to be able to reorient in the electric field of the cofactor ions following charge separation, thereby giving rise to a significant “dynamic screening”. The notion that a water-mediated dynamic screening of the light-generated P865•+QA•- radical pair can largely contribute to the relaxation from the dark-adapted to the light-adapted RC conformation is in line with the anchorage model,36 which was proposed to explain the tight dynamic coupling that characterizes glassy trehalose/water/protein matrixes. This model is based on the extrapolation of molecular-dynamics simulations performed on carboxymyoglobin (MbCO) embedded in trehalose/water systems and is corroborated by several pieces of experimental evidence obtained in trehalose-coated MbCO, RC, and GFP (green fluorescent protein) (see Francia et al.36 and references therein). According to the anchorage model, in trehalose/water

12740

J. Phys. Chem. B, Vol. 114, No. 39, 2010

systems of low water content, the residual water preferentially surrounds the protein. Upon drying, an increasing fraction of this hydration shell becomes involved in a H-bond network that bridges protein groups and sugar molecules, thus anchoring the protein surface to the surrounding trehalose matrix. Such a locking of the protein surface (and of the protein hydration shell) to the dynamics of the solid matrix is expected to hinder protein fluctuations and conformational changes that are related to amino-acid side-chain reorganization as well as to the reorientation of water molecules interacting with the RC, thus contributing to the stabilization of the charge-separated P865•+QA•- state. This scenario is consistent with the present EPR results, because such a matrix-protein interaction does not necessarily involve either any alteration of the radical-pair configuration of the P865•+ and QA•-cofactors or any dynamic-coupling change on the local scale of the QA binding pocket. Changes in these parameters are, therefore, not necessarily involved in the stabilization mechanism of charge separation by protein/solvent relaxation. The control exerted on the RC dynamics and ET kinetics by the fluctuations of the cofactor environment can be further clarified within the framework of a “unified model of protein dynamics” developed by Frauenfelder et al.,5 which summarizes decades of experimental and theoretical investigations focused on myoglobin (Mb). Three types of protein motions are essentially identified as relevant for biological function:93 solvent-coupled (R-slaved) processes (class I); hydration-shell coupled (β-slaved) processes (class II); and inner-molecular processes (class III), such as molecular vibrations in the forcefield potential of inner-molecular atom-atom interactions, that are higher in energy and nonslaved to solvent fluctuations. Except for the latter processes, which do not appear to be controlled by external thermal fluctuations, the dominant internal protein motions and concomitant protein functions are driven by the dynamics of the solvent. Class I, solvent-slaved motions, follow the dielectric (R-) fluctuations in the bulk solvent; they involve large-scale conformational changes, such as those controlling the entrance and exit of ligands in Mb or of diprotonated QB2- (QBH2) in RCs. Because these motions are controlled by the bulk-solvent viscosity η(T), they are considered to be absent in rigid environments (η f ∞).93 Class II motions are slaved to fast (β-) fluctuations in the hydration shell (one to two layers of solvent molecules). They are controlled by the degree of hydration of the protein and are expected to be absent in fully dehydrated proteins. Such β-fluctuations are supposed to involve side chains of the protein and to govern processes such as the movement of ligands between different cavities inside Mb. In the case of Mb, the distinction between class I and II processes (R- and β-fluctuations) has been put in relation with two hierarchically organized tiers of the energy landscape, CS2 (the lowest) and CS1 (the next higher one), both of which are well below the tier involving class III processes.5 In an attempt to apply this concept to RC dynamics coupled to light-induced ET, we propose to ascribe the structural relaxation from the dark-adapted to the light-adapted state to class II processes, slaved to the β-fluctuations of the hydration shell. According to the anchorage model outlined above, in a sufficiently dehydrated RC/trehalose/water matrix, we expect that the H-bond network involving the hydration shell substantially reduces its motional degrees of freedom (i.e., restricts the β-fluctuations), thus hindering the β-slaved relaxation from the dark-adapted to the light-adapted state. In line with the concept of a β-slaved relaxation, we recently found that an extensive drying of the RC in the absence of trehalose sugar, which leads to a substantial depletion of the hydration shell, is also able to

Savitsky et al. inhibit the conformational change from the dark-adapted to the light-adapted state at room temperature, as probed by the strongly accelerated and distributed kinetics of P865•+QA•recombination in superdry RC films following a laser pulse.100 A reduction of the β-fluctuations in solid matrixes is expected to depend on the specific molecular interactions between the matrix and the hydration shell of the protein (e.g., the ability to form H-bond networks). This interaction appears to be weaker in PVA matrixes than in trehalose, resulting in a weaker protein-matrix dynamic coupling.33 Accordingly, in agreement with previous observations,37 an almost comparable acceleration of the charge-recombination kinetics has been observed in the RC/PVA sample at hydration levels 1 order of magnitude lower than those prevailing in the RC/trehalose sample. Because of the low content of residual water, it is likely that the observed inhibition of the transition from the dark- to the light-adapted conformation is due, in this case, not only to a reduction of the RC hydration-shell dynamics (β-fluctuations) as caused by interaction with the PVA matrix, but also to a depletion of the protein hydration layer where the β-fluctuations occur. A different mechanism of matrix-protein coupling in the RC/PVA and RC/trehalose sample is further supported by the observation that, in the dehydrated trehalose matrix, the RC can withstand temperatures as high as 50 °C for several weeks without undergoing any denaturation (a prominent example of anhydrobiosis) whereas, in PVA, a significant fraction of the RC population undergoes thermal denaturation under the same conditions during the period of a few days (data not shown). Interestingly, the librational fluctuations of QA•- in the H-bond network of its binding site turned out to be nonslaved to the matrix environment, because the observed anisotropy of the transversal spin-relaxation time T2 is essentially independent of the chosen matrix, even in the extreme case of dehydrated trehalose glass. Although, in general, librational fluctuations of the whole cofactor might be expected to belong to the energetic tier of class II processes (β-fluctuations), they turned out to be not controlled externally by the solid protein-coating trehalose matrix but rather internally by temperature-dependent amplitudes of fluctuations affecting the hierarchy of H-bond strengths between the quinone and local amino-acid residues.48 In the frame of the unified model of protein dynamics proposed by Frauenfelder and co-workers,5,93 we tend, therefore, to ascribe the RC cofactor dynamics, as revealed by the present high-field EPR measurements at 290 and 150 K, to class III processes. The cofactor dynamics, nonslaved to the microenvironment, appears to be coupled to the process of light-induced charge separation, which proceeds even in the dehydrated RC and in solid matrixes at room and cryogenic temperatures. Finally, important questions remain to be answered: Given that, in the light-induced primary charge-separation processes in the RC/trehalose matrix the EPR experiments show that the internal characteristics of the cofactors in their binding sites (i.e., the electronic and structural properties of the radical ions P865•+ and QA•- and their radical pairs P865•+QA•-, as well as the T2 spin-relaxation times and their anisotropies) remain unchanged upon variation of the solvent to water or dehydrated PVA, what then is the biologically relevant process in illuminated RCs that gets inhibited in dry trehalose glass at room temperature and above, and what are the external control mechanisms for this inhibition? Analyses of the optically and EPR-detected kinetics of P865•+QA•- recombination after pulsed photoexcitation consistently show that the incorporation of RCs into a dehydrated trehalose glass strongly inhibits the conformational relaxation that stabilizes the primary charged separated

Bacterial Photosynthetic RCs in Trehalose Glasses state, as revealed by the increased rate constant of P865•+QA•recombination. Inhibition of this conformational relaxation, however, cannot be critical per se for the global function of the RC (i.e., for the overall photochemical RC quantum yield), because the rate constant of the subsequent electron transfer step, from QA•- to QB, is on the order of 104 s-1 in roomtemperature solutions of RCs, so that the forward secondary electron-transfer step could, in principle, efficiently compete with recombination from P865•+QA•-, even when its rate constant increases from ∼10 s-1 (in solution) to ∼30 s-1 (in extensively dried trehalose glasses). In fact, it was shown earlier22 that the secondary electron-transfer step QA•- f QB, which is generally considered to be rate-limited by conformational gating,13 is also strongly inhibited in dry trehalose glass even at elevated temperatures. This implies that this secondary electron-transfer step is one of the crucial processes, for which the dehydrated trehalose glass matrix leads to a large, inhomogeneous increase of the energy barriers that govern the conformational gate and, simultaneously, to strongly reduced molecular mobility and conformational fluctuations of the protein, thereby inhibiting the biological function at the level of the RC protein. In this anhydrobiosis strategy, trehalose is distinguished, in comparison to other disaccharides such as sucrose, by its high glass-transition temperature, Tg, of ∼380 K for dry trehalose, dropping sharply with increasing water content,94 which is about 40 °C higher than the Tg of sucrose.95 The higher the Tg value of the amorphous protein matrix, the higher the temperature to which the stability of the cell and its functional domains can be extended. When one progressively dehydrates the RC/trehalose sample, the inhibition of QA•--to-QB electron transfer also progressively increases and becomes inhomogeneous; that is, in a progressively increasing fraction of the RC population, the electrontransfer step is blocked.22 A similar behavior has been observed in RC/PVA films,37 where a complete inhibition of QA•--to-QB electron transfer is obtained at hydration levels scarcely affecting the P865•+QA•- recombination kinetics, which is accelerated only in extremely dried films (cf. the present work). Apparently, the QA•--to-QB electron-transfer step is significantly more sensitive to dehydration of the matrix than the conformational relaxation that stabilizes the P865•+QA•- state. This suggests that the conformational gate, which governs QA•--to-QB electron transfer, is also controlled by (slaved to) class II fluctuations of the hydration shell of the RC protein,5,93 which we expect to be drastically reduced in extensively dried trehalose glasses. Although the molecular origin of the conformational gate is not yet clear, buried water molecules in the RC protein91 would certainly be reasonable candidates for playing a role in this process. FTIR bands, attributed to weakly hydrogen-bonded water molecules near the quinone sites,91 appear to be perturbed when the electron is transferred to QB. Moreover, the dependence of the interquinone ET equilibrium on temperature and water activity has been recently interpreted by assuming that the QA•-to-QB ET is associated with the release of about three water molecules by the RC protein.96 Outlook The reported results show that incorporation of the RC into a dehydrated, glassy trehalose matrix affects neither the electronic and structural properties of the radical ions P865•+ and QA•- and their radical pairs P865•+QA•- nor the local environment and dynamics of QA•- in its binding pocket. This finding is considered to be of particular relevance because it implies that the strongly accelerated and distributed P865•+QA•- recombination

J. Phys. Chem. B, Vol. 114, No. 39, 2010 12741 kinetics observable in such trehalose matrixes is related to neither structural distortions affecting the cofactor geometry nor alterations in the local cofactor dynamics. These kinetic effects are rather a genuine probe of the protein/solvent conformational dynamics that govern the stabilization of the primary RC chargeseparated state, as we had previously proposed.21,73 Because the kinetic parameters of this electron-transfer process (the average rate constant k and the rate distribution width σ) are unambiguously related to the content of residual water in the trehalose matrix,35,36,73 it is argued that the protein/solvent conformational dynamics in trehalose/water/RC matrixes can be modulated continuously and reversibly at room temperature by varying the hydration level of the system. It appears, therefore, that, in general, trehalose glasses provide a suitable tool for roomtemperature investigations of the molecular mechanisms of the coupling between protein/solvent dynamics and electron-transfer processes that are supposed to be slaved to class II β-fluctuations of the protein hydration shell. The use of strongly dehydrated, solid trehalose matrixes offers the additional advantage of preserving, at room temperature, the structural and functional integrity of the cofactor-protein complex over a period of at least several months. Within the framework of the anchorage hypothesis,36 the tight protein-matrix dynamic coupling arising in trehalose glassy systems at very low water contents is ascribed to the ability of trehalose to form water-mediated hydrogen-bond networks that connect residues at the protein surface with the surrounding solid matrix. To further test this hypothesis, also in view of results from molecular-dynamics simulations of trehalose-protein interactions in aqueous solution,97 future experiments will need to focus on the structural and dynamic properties of the residual water at the RC surface, as well as on the coupled fluctuations of surface residues of the protein as a function of the hydration level of the embedding trehalose matrix. We expect that direct information on the dynamics of the RC protein surface and on the possible involvement of exposed RC residues in watermediated hydrogen bonding with the matrix can be obtained by combining site-specific NO• spin labeling of the RC H-subunit (e.g., labeling of the native cysteine at position 15698,99) with high-field EPR spectroscopy38 and incorporating the labeled RC complex into trehalose matrixes at different hydration levels. Experiments along these lines are underway in our laboratories, and we believe that this approach could significantly contribute to a better understanding of the molecular mechanisms by which the RC relaxes from a dark-adapted to a light-adapted state. Acknowledgment. K.M. thanks the Institute for Advanced Studies of the University of Bologna for a Senior Fellowship in 2009 and 2010, particularly Stefano Ciurli for having catalyzed this Bologna-Berlin cooperation project with his advice and encouragement and Barbara Cimatti for her help in solving upcoming problems. K.M. also thanks the members of the Department of Experimental and Evolutionary Biology of the University of Bologna for their grand hospitality, in particular, Giovanni Venturoli and Paola Turina for enlightening discussions on science and beyond. K.M. and A.S. thank Martin Plato (FU Berlin) for helpful discussions concerning the relation between protein dynamics and electron transfer and relaxation. We also thank Wolfgang Lubitz (Max Planck Institute, Mu¨lheim, Germany) for his critical reading of the manuscript. Financial support from the Deutsche Forschungsgemeinschaft (DFG) in the framework of the Priority Program SPP 1051, the Collaborative Research Center SFB 498, and the Group Project MO 132/19-2 is gratefully acknowledged. G.V. is grateful to

12742

J. Phys. Chem. B, Vol. 114, No. 39, 2010

Lorenzo Cordone (University of Palermo) for stimulating, clarifying discussions on the structural and dynamic properties of protein-water-trehalose systems. M.M and G.V. acknowledge financial support of MIUR of Italy (Grant PRIN 2008ZWHZJT). References and Notes (1) Kamen, M. D. Primary Processes in Photosynthesis; Academic Press: New York, 1963. (2) Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. Science 1991, 254, 1598. (3) Frauenfelder, H.; Wolynes, P. G. Phys. Today 1994, 47, 58. (4) Henzler-Wildman, K.; Kern, D. Nature 2007, 450, 964. (5) Frauenfelder, H.; Chen, G.; Berendzen, J.; Fenimore, P. W.; Jansson, H.; McMahon, B. H.; Stroe, I. R.; Swenson, J.; Young, R. D. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 5129. (6) Frauenfelder, H.; McMahon, B. H.; Fenimore, P. W. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 8615. (7) McMahon, B. H.; Mu¨ller, J. D.; Wraight, C. A.; Nienhaus, G. U. Biophys. J. 1998, 74, 2567. (8) Kriegl, J. M.; Forster, F. K.; Nienhaus, G. U. Biophys. J. 2003, 85, 1851. (9) Kriegl, J. M.; Nienhaus, G. U. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 123. (10) Frauenfelder, H.; McMahon, B. H. Ann. Phys. 2000, 9, 655. (11) Feher, G.; Allen, J. P.; Okamura, M. Y.; Rees, D. C. Nature 1989, 33, 111. (12) Hoff, A. J.; Deisenhofer, J. Phys. Rep. 1997, 287, 2. (13) Graige, M. S.; Feher, G.; Okamura, M. Y. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 11679. (14) Okamura, M. Y.; Paddock, M. L.; Graige, M. S.; Feher, G. Biochim. Biophys. Acta 2000, 1458, 148. (15) Koepke, J.; Krammer, E. M.; Klingen, A. R.; Sebban, P.; Ullmann, G. M.; Fritzsch, G. J. Mol. Biol. 2007, 371, 396. (16) McElroy, J. D.; Mauzerall, D. C.; Feher, G. Biochim. Biophys. Acta 1974, 333, 261. (17) Arata, H.; Parson, W. W. Biochim. Biophys. Acta 1981, 636, 70. (18) Kleinfeld, D.; Okamura, M. Y.; Feher, G. Biochemistry 1984, 23, 5780. (19) Nabedryk, E.; Bagley, K. A.; Thibodeau, D. L.; Bauscher, M.; Ma¨ntele, W.; Breton, J. FEBS Lett. 1990, 266, 59. (20) Brzezinski, P.; Andreasson, L. E. Biochemistry 1995, 34, 7498. (21) Palazzo, G.; Mallardi, A.; Hochkoeppler, A.; Cordone, L.; Venturoli, G. Biophys. J. 2002, 82, 558. (22) Francia, F.; Palazzo, G.; Mallardi, A.; Cordone, L.; Venturoli, G. Biophys. J. 2003, 85, 2760. (23) Francia, F.; Palazzo, G.; Mallardi, A.; Cordone, L.; Venturoli, G. Biochim. Biophys. Acta 2004, 1658, 50. (24) Crowe, L. M.; Reid, D. S.; Crowe, J. H. Biophys. J. 1996, 71, 2087. (25) Crowe, J. H.; Carpenter, J. F.; Crowe, L. M. Annu. ReV. Physiol. 1998, 60, 73. (26) Crowe, J. H.; Hoekstra, F. A.; Crowe, L. M. Annu. ReV. Physiol. 1992, 54, 579. (27) Clegg, J. S. Comp. Biochem. Physiol. B 2001, 128, 613. (28) Crowe, L. M. Comp. Biochem. Physiol. A 2002, 131, 505. (29) Sakurai, M.; Furuki, T.; Akao, K.; Tanaka, D.; Nakahara, Y.; Kikawada, T.; Watanabe, M.; Okuda, T. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 5093. (30) Hagen, S. J.; Hofrichter, J.; Eaton, W. A. Science 1995, 269, 959. (31) Cordone, L.; Ferrand, M.; Vitrano, E.; Zaccai, G. Biophys. J. 1999, 76, 1043. (32) Giuffrida, S.; Cottone, G.; Librizzi, F.; Cordone, L. J. Phys. Chem. B 2003, 107, 13211. (33) Giachini, L.; Francia, F.; Cordone, L.; Boscherini, F.; Venturoli, G. Biophys. J. 2007, 92, 1350. (34) Cottone, G.; Cordone, L.; Ciccotti, G. Biophys. J. 2001, 80, 931. (35) Cordone, L.; Cottone, G.; Giuffrida, S.; Palazzo, G.; Venturoli, G.; Viappiani, C. Biochim. Biophys. Acta 2005, 1749, 252. (36) Francia, F.; Dezi, M.; Mallardi, A.; Palazzo, G.; Cordone, L.; Venturoli, G. J. Am. Chem. Soc. 2008, 130, 10240. (37) Francia, F.; Giachini, L.; Palazzo, G.; Mallardi, A.; Boscherini, F.; Venturoli, G. Bioelectrochemistry 2004, 63, 73. (38) Mo¨bius, K.; Savitsky, A. High-Field EPR Spectroscopy on Proteins and Their Model Systems; RSC Publishing: Cambridge, U.K., 2009. (39) Burghaus, O.; Plato, M.; Rohrer, M.; Mo¨bius, K.; MacMillan, F.; Lubitz, W. J. Phys. Chem. 1993, 97, 7639. (40) Klette, R.; To¨rring, J. T.; Plato, M.; Mo¨bius, K.; Bo¨nigk, B.; Lubitz, W. J. Phys. Chem. 1993, 97, 2015. (41) Prisner, T. F.; Rohrer, M.; Mo¨bius, K. Appl. Magn. Reson. 1994, 7, 167.

Savitsky et al. (42) Prisner, T. F.; Van der Est, A.; Bittl, R.; Lubitz, W.; Stehlik, D.; Mo¨bius, K. Chem. Phys. 1995, 194, 361. (43) Huber, M.; To¨rring, J. T. Chem. Phys. 1995, 194, 379. (44) Huber, M.; To¨rring, J. T.; Plato, M.; Finck, U.; Lubitz, W.; Feick, R.; Schenck, C. C.; Mo¨bius, K. J. Sol. Energy Mater. Sol. Cells 1995, 38, 119. (45) Rohrer, M.; Plato, M.; MacMillan, F.; Grishin, Y.; Lubitz, W.; Mo¨bius, K. J. Magn. Reson. A 1995, 116, 59. (46) Rohrer, M.; Gast, P.; Mo¨bius, K.; Prisner, T. F. Chem. Phys. Lett. 1996, 259, 523. (47) Rohrer, M.; MacMillan, F.; Prisner, T. F.; Gardiner, A. T.; Mo¨bius, K.; Lubitz, W. J. Phys. Chem. B 1998, 102, 4648. (48) Schnegg, A.; Fuhs, M.; Rohrer, M.; Lubitz, W.; Prisner, T. F.; Mo¨bius, K. J. Phys. Chem. B 2002, 106, 9454. (49) Fuchs, M. R.; Schnegg, A.; Plato, M.; Schulz, C.; Mu¨h, F.; Lubitz, W.; Mo¨bius, K. Chem. Phys. 2003, 294, 371. (50) Kirilina, E. P.; Prisner, T. F.; Bennati, M.; Endeward, B.; Dzuba, S. A.; Fuchs, M. R.; Mo¨bius, K.; Schnegg, A. Magn. Reson. Chem. 2005, 43, S119. (51) Mo¨bius, K.; Savitsky, A.; Schnegg, A.; Plato, M.; Fuchs, M. Phys. Chem. Chem. Phys. 2005, 7, 19. (52) Savitsky, A.; Dubinskii, A. A.; Flores, M.; Lubitz, W.; Mo¨bius, K. J. Phys. Chem. B 2007, 111, 6245. (53) Schnegg, A.; Dubinskii, A. A.; Fuchs, M. R.; Grishin, Y. A.; Kirilina, E. P.; Lubitz, W.; Plato, M.; Savitsky, A.; Mo¨bius, K. Appl. Magn. Reson. 2007, 31, 59. (54) Flores, M.; Savitsky, A.; Abresch, E.; Lubitz, W.; Mo¨bius, K. Photosynth. Res. 2007, 91, 155. (55) Schweiger, A.; Jeschke, G. Principles of Pulse Electron Paramagnetic Resonance; Oxford University Press: Oxford, U.K., 2001. (56) Thomann, H.; Dalton, L. R.; Dalton, L. A. In Biological Magnetic Resonance; Berliner, L. J., Ed.; Plenum: New York, 1984; Vol. 6, p 143. (57) Saxena, S.; Freed, J. H. J. Phys. Chem. A 1997, 101, 7998. (58) Dzuba, S. A.; Tsvetkov, Y. D.; Maryasov, A. G. Chem. Phys. Lett. 1992, 188, 217. (59) Millhauser, G. L.; Freed, J. H. J. Chem. Phys. 1984, 81, 37. (60) Pokkuluri, P. R.; Laible, P. D.; Crawford, A. E.; Mayfield, J. F.; Yousef, M. A.; Ginell, S. L.; Hanson, D. K.; Schiffer, M. FEBS Lett. 2004, 570, 171. (61) Stowell, M. H.; McPhillips, T. M.; Rees, D. C.; Soltis, S. M.; Abresch, E.; Feher, G. Science 1997, 276, 812. (62) Breton, J. Biochemistry 2004, 43, 3318. (63) Nabedryk, E.; Breton, J. Biochim. Biophys. Acta 2008, 1777, 1229. (64) Koepke, J.; Krammer, E. M.; Klingen, A. R.; Sebban, P.; Ullmann, G. M.; Fritzsch, G. J. Mol. Biol. 2007, 371, 396. (65) Gray, K. A.; Farchaus, J. W.; Wachtveitl, J.; Breton, J.; Oesterhelt, D. EMBO J. 1990, 9, 2061. (66) Utschig, L. M.; Greenfield, S. R.; Tang, J.; Laible, P. D.; Thurnauer, M. C. Biochemistry 1997, 36, 8548. (67) Debus, R. J.; Feher, G.; Okamura, M. Y. Biochemistry 1986, 25, 2276. (68) Sloten, L. Biochim. Biophys. Acta 1972, 275, 208. (69) Oettmeier, W.; Preusse, S. Z. Naturforsch. 1987, 42c, 690. (70) Giangiacomo, K. M.; Robertson, D. E.; Gunner, M. R.; Dutton, P. L. In Progress in Photosynthesis Research; Biggins, J., Ed.; Martinus Nijhoff Publishers: Dordrecht, The Netherlands, 1987; Vol. 2, p 409. (71) Ginet, N.; Lavergne, J. Biochemistry 2001, 40, 1812. (72) Bonner, O. D.; Choi, Y. S. J. Phys. Chem. 1974, 78, 1723. (73) Francia, F.; Malferrari, M.; Sacquin-Mora, S.; Venturoli, G. J. Phys. Chem. B 2009, 113, 10389. (74) Kim, S. S.; Weissman, S. I. J. Am. Chem. Soc. 1979, 101, 5863. (75) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42. (76) Stoll, S.; Schweiger, A. In ESR Spectroscopy in Membrane Biophysics; Hemminga, M. A., Berliner, L. J., Eds.; Springer: New York, 2007; Vol. 27, p 299. (77) Isaacson, R. A.; Lendzian, F.; Abresch, E. C.; Lubitz, W.; Feher, G. Biophys. J. 1995, 69, 311. (78) Tang, J.; Utschig, L. M.; Poluektov, O.; Thurnauer, M. C. J. Phys. Chem. B 1999, 103, 5145. (79) Hales, B. J.; Case, E. E. Biochim. Biophys. Acta 1981, 637, 291. (80) Lubitz, W.; Abresch, E. C.; Debus, R. J.; Isaacson, R. A.; Okamura, M. Y.; Feher, G. Biochim. Biophys. Acta 1985, 808, 464. (81) Feher, G.; Isaacson, R. A.; Okamura, M. Y.; Lubitz, W. In Antennas and Reaction Centers of Photosynthetic Bacteria; Michel-Beyerle, M. E., Ed.; Springer: Berlin, 1985; pp 174-189. (82) Steinbach, P. J.; Chu, K.; Frauenfelder, H.; Johnson, J. B.; Lamb, D. C.; Nienhaus, G. U.; Sauke, T. B.; Young, R. D. Biophys. J. 1992, 61, 235. (83) Flores, M.; Savitsky, A.; Abresch, E. C.; Lubitz, W.; Mo¨bius, K. In Photosynthesis. Energy from the Sun; Allen, J. F., Osmond, B., Golbeck, J. H., Gantt, E., Eds.; Springer: Heidelberg, Germany, 2008; p 59. (84) Stehlik, D.; Bock, C. H.; Petersen, J. J. Phys. Chem. 1989, 93, 1612.

Bacterial Photosynthetic RCs in Trehalose Glasses (85) Hore, P. J. In AdVanced EPR, Applications in Biology and Biochemistry; Hoff, A. J., Ed.; Elsevier: Amsterdam, 1989; p 405. (86) Flores, M.; Savitsky, A.; Paddock, M. L.; Abresch, E. C.; Dubinskii, A. A.; Okamura, M. Y.; Lubitz, W.; Mo¨bius, K. J. Phys. Chem. B, manuscript submitted. (87) Parak, F.; Frolov, E. N.; Kononenko, A. A.; Mo¨ssbauer, R. L.; Goldanskii, V. I.; Rubin, A. B. FEBS Lett. 1980, 117, 368. (88) Eaton, G. R.; Eaton, S. S. In Distance Measurements in Biological Systems by EPR; Berliner, L. J., Eaton, G. R., Eaton, S. S., Eds.; Kluwer Academic/Plenum Publishers: New York, 2000; Vol. 19, p 29. (89) Kalman, L.; Maroti, P. Biochemistry 1994, 33, 9237. (90) Maroti, P.; Wraight, C. A. Biophys. J. 1997, 73, 367. (91) Iwata, T.; Paddock, M. L.; Okamura, M. Y.; Kandori, H. Biochemistry 2009, 48, 1220. (92) Ermler, U.; Fritzsch, G.; Buchanan, S. K.; Michel, H. Structure 1994, 2, 925.

J. Phys. Chem. B, Vol. 114, No. 39, 2010 12743 (93) Fenimore, P. W.; Frauenfelder, H.; McMahon, B. H.; Young, R. D. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 14408. (94) Simperler, A.; Kornherr, A.; Chopra, R.; Jones, W.; Motherwell, W. D.; Zifferer, G. Carbohyd. Res. 2007, 342, 1470. (95) Frank, G. A. J. Phys. Chem. Ref. Data 2007, 36, 1279. (96) Palazzo, G.; Francia, F.; Mallardi, A.; Giustini, M.; Lopez, F.; Venturoli, G. J. Am. Chem. Soc. 2008, 130, 9353. (97) Lins, R. D.; Pereira, C. S.; Hunenberger, P. H. Proteins 2004, 55, 177. (98) Poluektov, O. G.; Utschig, L. M.; Dalosto, S.; Thurnauer, M. C. J. Phys. Chem. B 2003, 107, 6239. (99) Borovykh, I. V.; Ceola, S.; Gajula, P.; Gast, P.; Steinhoff, H. J.; Huber, M. J. Magn. Reson. 2006, 180, 178. (100) Malferrari, M.; Francia, F.; Venturoli, G., in preparation.

JP105801Q