This Article Abstract Full Text (PDF) Supplemental Material Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Francia, F. Articles by Venturoli, G. Search for Related Content PubMed PubMed Citation Articles by Francia, F. Articles by Venturoli, G.

Residual Water Modulates Q_A^--to-Q_B Electron Transfer in Bacterial Reaction Centers Embedded in Trehalose Amorphous Matrices

Francesco Francia ^*, Gerardo Palazzo , Antonia Mallardi , Lorenzo Cordone  and Giovanni Venturoli ^* ¶

^* Laboratorio di Biochimica e Biofisica, Dipartimento di Biologia, Università di Bologna, Bologna, Italy; Dipartimento di Chimica, Università di Bari, Bari, Italy and Dipartimento di Scienze e Tecnologie Agro-alimentari Ambientali e Microbiologiche, Università del Molise, Campobasso, Italy; Istituto per i Processi Chimico-Fisici, Consiglio Nazionale delle Recerche, Bari, Italy; Dipartimento di Scienze Fisiche ed Astronomiche and Istituto Nazionale per la Fisica della Materia, Palermo, Italy; and ^¶ Istituto Nazionale per la Fisica della Materia, Unità di Ricerca di Bologna, Bologna, Italy

Correspondence: Address reprint requests to Giovanni Venturoli, E-mail: ventur{at}alma.unibo.it.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS APPENDIX SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
The role of protein dynamics in the electron transfer from the reduced primary quinone, Q_A^-, to the secondary quinone, Q_B, was studied at room temperature in isolated reaction centers (RC) from the photosynthetic bacterium Rhodobacter sphaeroides by incorporating the protein in trehalose water systems of different trehalose/water ratios. The effects of dehydration on the reaction kinetics were examined by analyzing charge recombination after different regimes of RC photoexcitation (single laser pulse, double flash, and continuous light) as well as by monitoring flash-induced electrochromic effects in the near infrared spectral region. Independent approaches show that dehydration of RC-containing matrices causes reversible, inhomogeneous inhibition of Q_A^--to-Q_B electron transfer, involving two subpopulations of RCs. In one of these populations (i.e., active), the electron transfer to Q_B is slowed but still successfully competing with P⁺Q_A^- recombination, even in the driest samples; in the other (i.e., inactive), electron transfer to Q_B after a laser pulse is hindered, inasmuch as only recombination of the P⁺Q_A^- state is observed. Small residual water variations (7 wt %) modulate fully the relative fraction of the two populations, with the active one decreasing to zero in the driest samples. Analysis of charge recombination after continuous illumination indicates that, in the inactive subpopulation, the conformational changes that rate-limit electron transfer can be slowed by >4 orders of magnitude. The reported effects are consistent with conformational gating of the reaction and demonstrate that the conformational dynamics controlling electron transfer to Q_B is strongly enslaved to the structure and dynamics of the surrounding medium. Comparing the effects of dehydration on P⁺Q_A^-PQ_A recombination and Q_A^-Q_BQ_AQ_B^- electron transfer suggests that conformational changes gating the latter process are distinct from those stabilizing the primary charge-separated state.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS APPENDIX SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
Proteins display an extremely large number of conformational substates, which at physiological temperatures are reflected in complex conformational dynamics, intimately connected to protein reactivity (Frauenfelder et al., 1988, 1991; Frauenfelder and Wolynes, 1994, Frauenfelder and McMahon, 1998). The activation of this protein dynamics coupled to biological functionality appears to require a minimum water content (see e.g., Careri, 1992; Barron et al., 1997; Mattos, 2002). The possibility that internal motions determine the rate of long-range electron transfer catalyzed by redox proteins is well-recognized (see e.g., Hoffman and Ratner, 1987; Davidson, 1996; Kotelnikov et al., 1998; Sharp and Chapman, 1999; Balabin and Onuchic, 2000; Cherepanov et al., 2001). These motions, in principle, cover a wide temporal window (from femtoseconds to seconds), extending from fast vibrations of atoms to slow displacements of large protein domains (Woodbury and Parson, 1986; Peloquin et al., 1994; McMahon et al., 1998; Zhang et al., 1998; Izrailev et al., 1999).

The photosynthetic reaction center (RC) from purple bacteria, allowing the study of single-turnover reactions initiated by a pulse of light, provides an excellent laboratory for exploring the relationship between protein dynamics and electron transfer events. This integral pigment-protein complex catalyzes a light-induced charge separation across the membrane dielectric, thus promoting the primary event of photosynthetic energy transduction (Gunner, 1991). Within the photosynthetic RC from Rhodobacter sphaeroides a bacteriochlorophyll dimer (P) acts as the primary electron donor: after absorption of a photon, it delivers an electron (via a bacteriopheophytin molecule) to the primary ubiquinone acceptor Q_A, generating the primary charge-separated state P⁺Q_A^-. The photo-reduced Q_A^- in turn reduces a ubiquinone-10 molecule bound at the secondary acceptor Q_B site. When no physiological or artificial electron donor is available to re-reduce P⁺, the electron on Q_B^- recombines with the hole on P⁺, restoring the initial ground state of the RC (Feher et al., 1989; Okamura et al., 2000).

Several independent experimental findings point toward structural changes accompanying light-induced charge separation within the reaction center (Arata and Parson, 1981; Woodbury and Parson, 1984; Kleinfeld et al., 1984a; Kirmaier et al., 1985; Parot et al., 1987; Nabedryk et al., 1990; Brzezinski and Andreasson, 1995; Kalman and Maroti, 1997; Stowell et al., 1997; Edens et al., 2000). The more direct evidences of a close coupling between electron transfer and protein conformational dynamics were obtained by comparing the rate of specific electron transfer processes in RCs frozen in the dark and under illumination. These pioneering studies have shown that RCs can be trapped at cryogenic temperatures in dark-adapted and light-adapted conformations drastically differing in the stability of the primary charge-separated state P⁺Q_A^- (Kleinfeld et al., 1984a); both these conformations consist of distribution of substates which result in a wide spectrum of electron transfer rates for charge recombination of the P⁺Q_A^- state (Kleinfeld et al., 1984a, McMahon et al., 1998). A second reaction, which has been intensively investigated in relation to RC dynamics, is the electron transfer from the primary reduced to the secondary ubiquinone acceptor. When RCs are frozen in the dark at temperatures <200 K, the electron transfer from Q_A^- to Q_B stops; by contrast, it persists even at 50 K in RC frozen under illumination and then allowed to return to the ground state (Kleinfeld et al., 1984a, Xu and Gunner, 2001). Consistently, at room temperature, the rate of electron transfer between the primary and secondary quinone was found to be independent of the associated redox free energy change (Graige et al., 1998). Such an independence of the driving force indicates that the reaction is not rate-limited by electron transfer itself, but gated by some other process. The reaction appears to be limited by proton binding at high pH (>8.0) and by conformational changes at pH < 8.0 (Paddock et al., 1989; Takahashi and Wraight, 1992; Xu and Gunner, 2002). Moreover, it has been shown that, at 120 K < T < 200 K, RCs frozen under illumination can relax to a Q_A^--to-Q_B inactive conformation, which is different from the dark-adapted one (Xu and Gunner, 2001). Recent temperature-dependence studies are beginning to shed some light on the complex energy landscape that governs this electron transfer process (Xu and Gunner, 2002; Xu et al., 2002).

Function-dynamics coupling is usually studied by hampering protein substate interconversion and relaxations at cryogenic temperatures (Austin et al., 1975; Ortega et al., 1996; McMahon et al., 1998). Alternatively, a severe conditioning of the internal protein dynamics can be achieved at room temperature by embedding the protein within a saccharide matrix and by dehydrating the resulting sample. In particular, the relationship between function and dynamics has been studied extensively by incorporating heme-proteins into dry trehalose glassy matrices (Hagen et al., 1995, 1996; Gottfried et al., 1996; Librizzi et al., 2002). Trehalose, a disaccharide found in large amounts in organisms that can survive conditions of extreme dehydration and high temperature, also exhibits peculiar properties in the preservation of biostructures (see e.g., Leslie et al., 1995; Uritani et al., 1995; Crowe et al., 1996). In trehalose matrices, nonharmonic contributions to internal motions (called protein-specific motions) are severely hindered, as shown by Mössbauer and optical absorption spectroscopy, neutron scattering, and molecular dynamics simulations (Cordone et al., 1998, 1999; Cottone et al., 2001). The analysis of function-dynamics coupling in trehalose matrices of different water content, at room temperature, offers a distinctive opportunity in relation to temperature-dependence studies performed at cryogenic temperatures: this approach in fact uncouples temperature from solvent effects. In principle, it can yield better insights on the extent to which protein internal dynamics is enslaved to the structure and dynamics of the external medium (Vitkup et al., 2000).

It has been reported that for carboxy myoglobin embedded in trehalose matrices of decreasing water content, the interconversion among conformational substates is progressively blocked at room temperature (Librizzi et al., 1999, 2002). In full agreement, we have recently shown that bacterial photosynthetic RCs can be functionally incorporated into trehalose matrices and that trehalose coating can be used to condition protein dynamics coupled to electron transfer in this large integral membrane protein (Palazzo et al., 2002). In extensively dehydrated trehalose matrices, RC relaxation from the dark-adapted to the light-adapted conformation and interconversion between conformational substates could be blocked at room temperature over the timescale of hundreds of milliseconds: the impairment was reflected in inhomogeneous kinetics of P⁺Q_A^- charge recombination. In the present work we exploited a similar strategy to assess the role of conformational dynamics in electron transfer from the primary photo-reduced (Q_A^-) to the secondary (Q_B) quinone acceptor. The reaction has been investigated by studying the kinetics of charge recombination after a laser pulse in progressively dehydrated trehalose matrices. We show that Q_A^--to-Q_B electron transfer can be slowed down by orders of magnitude at room temperature when the water content of the trehalose matrix is decreased over a limited range. By examining the kinetics of charge recombination after periods of continuous illumination the time course of the relaxation processes which control electron transfer has been determined as a function of the residual water.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS APPENDIX SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
The RCs were isolated and purified from Rb. sphaeroides R-26 according to Gray et al. (1990). This isolation procedure gives RCs with a Q_B content/activity of 50%. Reconstitution of the secondary quinone acceptor was achieved as in van Mourik et al. (2001) with minor changes, as described below. The purified RC was loaded on a column of DE-52 (Whatman, Maidstone, UK) previously equilibrated with buffer Tris-HCl 20 mM pH 8, 0.025% lauryl dimethylamine-n-oxide (LDAO) containing ubiquinone-10 (UQ-10; from Sigma Chemical, St. Louis, MO). The low solubility of UQ-10 in aqueous buffer was overcome by adding the quinone to the 30% LDAO stock solution used for preparing buffer. The RCs were extensively washed with the same buffer, eluted with 400 mM NaCl and dialyzed overnight. Q_B activity of the reconstituted samples was typically >95%.

Trehalose was from Hayashibara Shoij (Okayama, Japan). The concentration of trehalose in the samples is reported as trehalose weight percent (wt %), i.e.: (grams of trehalose/grams of trehalose plus water) %, neglecting the contribution of RC, buffer, and detergent to the sample weight.

Liquid samples (16 wt % and 53 wt % trehalose) were prepared by dissolving suitable amounts of trehalose in aqueous buffer (5 mM Tris pH = 7.5; 0.025% LDAO); the RC was added to this sugar solution from the concentrated stock solution. The sample composition was evaluated by weight.

Plasticized amorphous and glassy samples were prepared by drying under N₂ flow a thin layer of a trehalose solution containing RC on optical glass plates, as described in detail in Palazzo et al. (2002). Data were collected at 298 K at various times during the drying of trehalose matrices kept under dry nitrogen atmosphere.

The water content of the amorphous matrices was assayed by visible and near infrared (NIR) spectroscopy on a PerkinElmer (Fremont, CA) Lambda 19 spectrometer. To estimate the relative sample composition, the area of the combination band of water at 1950 nm was considered and the RC absorption at 802 nm was used as an internal standard (Palazzo et al., 2002).

Rapid absorbance changes were monitored with a spectrophotometer of local design (Mallardi et al., 1997). RC photochemistry was elicited by a 20-ns pulse from a dye-laser (RDP-1; Radiant Dyes GmbH, Wermelskirchen, Germany) pumped by a frequency-doubled Q-switched Nd-YAG laser (Surelite 10, Continuum, Santa Clara, CA). Stryryl 9 was used as a dye (_max at 810 nm); only in double flash experiments, to increase pulse saturation, sulphorhodamine B was used (_max = 590 nm). The kinetics of charge recombination were monitored at 605 nm, 450 nm, and 422 nm (Feher and Okamura, 1978). The photomultiplier was protected from scattered excitation light by 0.01% blocking, 10-nm bandwidth interference filters centered at the corresponding wavelengths. Electrochromic transients associated with Q_A^--to-Q_B electron transfer were measured in the 750-nm region (Vermeglio and Clayton, 1977); in these measurements the photomultiplier was protected by a monochromator plus an interference filter. Continuous illumination was provided by a 200-W quartz tungsten halogen lamp collimated by an optical condenser and filtered by 8 cm of thermostated water and by a colored glass long-pass filter with a cut-on wavelength of 780 nm. In continuous illumination experiments we used a Uniblitz electro-programmable shutter system, characterized by a 3-ms closure time (Vincent Associates, Rochester, NY).

Nonlinear least-squares minimization was performed by computer routines based on a modified Marquardt algorithm (Bevington, 1969). Confidence intervals of fitting parameters were estimated numerically by an exhaustive search method (Beechem, 1992; Holzwarth, 1996). The confidence interval was obtained by using an F-statistics routine to determine the probability of a particular fractional increase in chi-square, as described in Palazzo et al. (2002).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS APPENDIX SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
Charge recombination in trehalose solutions and amorphous matrices
When isolated RCs are activated by a flash of light, the state P⁺Q_A^-Q_B is formed in <200 ps. Subsequent electron transfer events can be suitably described by the following equation:

(1)

The state P⁺Q_A^-Q_B can either recombine yielding the ground state PQ_AQ_B with a rate constant k_AP 10 s^-1 or, in the presence of ubiquinone bound at the Q_B site, yield P⁺Q_AQ_B^-. Under physiological conditions, the state P⁺Q_AQ_B^- is stabilized with respect to P⁺Q_A^-Q_B by 70 meV and, since (k_AB + k_BA) 10⁴ s^-1 >> k_AP, electron transfer to Q_B occurs with an extremely high quantum efficiency (Kleinfeld et al., 1984b). The state P⁺Q_AQ_B^- recombines slowly (lifetime 1 s) essentially via the P⁺Q_A^-Q_B state, with the direct route (k_BP) being negligible at room temperature (Kleinfeld et al., 1984b; Labahn et al., 1995). By contrast, the fast recombination of P⁺Q_A^-Q_B (lifetime 0.1 s) is observed whenever electron transfer to Q_B is blocked over this timescale. In general, since the quantum yield for Q_B^- formation and the consequent kinetics of recombination depend on k_AB and k_BA values relative to k_AP, information on the kinetics of electron transfer to Q_B can be obtained by following the kinetics of P⁺ decay after a flash of light (Kleinfeld et al., 1984b; Mancino et al., 1984). We have used primarily this approach to examine the kinetics of Q_A^--to-Q_B electron transfer in trehalose solutions and plasticized or solid matrices.

Fig. 1 shows the room temperature kinetics of charge recombination after laser excitation for RC in trehalose solutions (Fig. 1 A) and matrices progressively dehydrated (Fig. 1 B). Even in solution and in the absence of trehalose (Fig. 1 A), the decay of P⁺ is not strictly mono-exponential. A small fraction ( 5%) of P⁺ decaying in the tens of ms timescale (i.e., undergoing P⁺Q_A^- recombination) is present. Moreover, the dominating slow component slightly deviates from an exponential decay, suggesting a relatively narrow continuous distribution of rate constants. In fact, this slow kinetic phase can be accurately analyzed by a cumulant expansion truncated at the second term (Palazzo et al., 2000). Accordingly, an adequate and physically meaningful function fitting the P⁺ decays measured in liquid solutions, both in the presence and in the absence of trehalose, is (Ambrosone et al., 2002),

(2)

where A_f and (1 - A_f) are the fractions of fast and slow decay, respectively; k_f is the rate constant of the fast component; and k_s and _s are the average rate constant and distribution width for the slow decaying component. When fitting the kinetics recorded in solution, we fixed the rate constant of the fast phase to 8.2 s^-1 as measured for P⁺Q_A^- recombination in solution (Palazzo et al., 2002). This was done to avoid the effects of strong parameter correlation and in view of the low amplitude of the fast kinetic component. The validity of such an approach is confirmed by the fact that leaving k_f as an adjustable parameter yields values ranging between 8 s^-1 and 10 s^-1, although with large uncertainty.

View larger version (22K): [in this window] [in a new window]   FIGURE 1  Kinetics of charge recombination after flash excitation of RCs in trehalose solutions and matrices. (A) Normalized P+ decay after a flash at increasing trehalose concentrations (traces from bottom to top) in liquid samples. Labels indicate percent weight in trehalose. Continuous curves are best fit to Eq. 2, in which kf was fixed to 8.2 s-1 (see text for details). Kinetic parameters are given in the following; values in parentheses represent the extremes of confidence intervals within two standard deviations (see Materials and Methods). 0 wt %: Af = 6.5% (5.5, 7.5); ks = 0.647 s-1 (0.641, 0.653); and s = 0.14 s-1 (0.12, 0.16). 16 wt %: Af = 4.4% (2.1, 5.7); ks = 0.518 s-1 (0.512, 0.523); and s = 0.12 s-1 (0.11, 0.13). 53 wt %: Af = 3.9% (2.4, 5.5); ks = 0.49 s-1 (0.48, 0.50); and s = 0.15 s-1 (0.14, 0.16). (B) Charge recombination kinetics in trehalose matrices at increasing degree of dehydration (traces a–d). Trace a was recorded in a "wet" matrix, in a state very close to the glass transition (trehalose concentration = 90 wt %). Trehalose concentrations of 94.3%, 95.4%, and 96.2% were evaluated by NIR spectroscopy (see Materials and Methods) in matrices b, c, and d, respectively. Continuous curves are best fit according to Eq. 4. The corresponding kinetic parameters are given in Table 1. (C) Effects of rehydration of a trehalose glass on the kinetics of P+ decay. Trace c, solid matrix (95.4 wt % trehalose); trace c', after resuspension (liquid sample, 16 wt % trehalose); and trace c,'' resuspended sample plus 60 µM ubiquinone-10 in LDAO (final detergent concentration 0.06%).

Drying (10 h) under N₂ flow of a suitable protein-trehalose-detergent solution (see Materials and Methods) layered on an optical glass plate results in an amorphous plasticized, transparent sample, which still contains residual loosely bound water that can be gradually removed by further drying under nitrogen (Palazzo et al., 2002). In the more "wet" matrix, obtained after 10 h drying (Fig. 1 B, trace a), the kinetics of P⁺ decay is similar to that observed in solution at high trehalose concentration (compare to Fig. 1 A). In such samples the slow decay fraction is essentially unchanged whereas the average rate constant k_s further decreases to 0.4 s^-1 (Table 1). Resuspension of this "wet" sample fully restores the kinetic parameters measured in the corresponding initial (16 wt %) trehalose solution (compare to Fig. 1 A and Table 1). A more prolonged exposure to N₂ atmosphere results in larger dehydration (Fig. 1 B and Table 1, samples b–d). The trehalose concentration in such samples, estimated on the basis of the water absorption at 1950 nm (Palazzo et al., 2002), increases from 94.3 wt % to 96.2 wt %. The progressive dehydration brings about a progressive increase of the amplitude of the fast component (A_f) of P⁺ decay, corresponding to P⁺Q_A^- recombination. This component accounts for >80% of the decay in the driest sample (Fig. 1 B and Table 1, sample d), thus suggesting that dehydration impairs electron transfer from Q_A^- to Q_B in a substantial fraction of the RC population. Moreover, nonexponential decay becomes evident for both fast and slow kinetic components: a reasonable description of the P⁺ decay requires two rate distributions (fast and slow). Accordingly, the kinetics can be fitted to either the sum of two cumulant expansions (Eq. 3), or the sum of two "power laws" (Eq. 4),

(3)

(4)

where the subscripts f and s identify the fast and slow components. Parameters and n in Eq. 4 are related to k and in Eq. 3 by the following equations:

(5)

Equations 3 and 4 fit the kinetics measured in dry samples equally well and yield comparable values of k and of the rate distribution functions, as was found in the analysis of P⁺Q_A^- recombination in dehydrated trehalose matrices (Palazzo et al., 2002). The fast kinetic component is constantly faster in dry samples than in solutions (average rate constant, k_f, ranging between 11 s^-1 and 14 s^-1) (Table 1). This acceleration is quantitatively consistent with that observed for the kinetics of P⁺Q_A^- recombination in Q_B-deprived RCs embedded in progressively dehydrated trehalose matrices (Palazzo et al., 2002). Rate constants ranging between 10 s^-1 and 14 s^-1 were obtained at a moderate dehydration, corresponding to trehalose concentrations between 92 wt % and 96 wt % (Palazzo et al., 2002), in line with NIR estimation of the water content of the present samples. The average rate constant of the residual slow phase, k_s, does not change over this range of dehydration, remaining close to the value of 0.4 s^-1 measured in the more "wet" amorphous matrix (Table 1).

The reversibility of the above effects was tested by redissolving samples at different degrees of drought (Fig. 1 C and Table 1). Equation 2, i.e., the same function used to describe kinetics in the initial trehalose solutions, was fitted to the kinetics of P⁺ decay in redissolved samples. Upon resuspension of non-extremely dried samples a complete recovery of the amplitude of the slow kinetic component (A_s) is observed (e.g., sample b in Table 1). Correspondingly, the average rate constant of the slow phase, k_s, which is 0.4 s^-1 in dry samples, upon resuspension increases to 0.7 s^-1, indicating a full recovery of the kinetic behavior measured in solution before drying. However, samples exposed longer to dry N₂ atmosphere exhibit, after resuspension, only partial recovery of the amplitude of the slow kinetic component (see Fig. 1 C, trace c', and Table 1, samples c and d). Based on the fact that Eq. 2, with k_f fixed at 8.2 s^-1, fits the kinetics in the redissolved samples well, one can infer that the effects of dehydration on P⁺Q_A^- recombination kinetics are fully reversible even in exhaustively dried samples. Analogous behavior has been observed in RC deprived of Q_B (Palazzo et al., 2002). The simplest explanation for the incomplete recovery of the slow component of P⁺ decay upon resuspension of the more dehydrated matrices is that, during prolonged exposition to dry N₂ atmosphere, an unknown electron donor reduces Q_B to Q_BH₂ in a fraction of the RCs. This population lacking oxidized secondary acceptor will undergo only primary charge separation (P⁺Q_A^-) and fast recombination. Several lines of evidence confirm this hypothesis:

Partial reduction to UQH₂ can be assessed by spectrophotometry when detergent trehalose suspensions of ubiquinone-10, in the absence of the RC, are extensively dehydrated by prolonged (from 3 to 7 days) exposure to dry N₂ atmosphere and are successively fully rehydrated (spectra are shown in Supplemental Material). An absorbance increase at 290 nm (diagnostic of UQH₂) is detected also upon resuspension of samples obtained by drying the detergent suspension under N₂ flux in the absence of trehalose (not shown);

Analogous results are obtained with ubiquinone-0; in this case a larger fraction of ubiquinone is reduced upon exposure to N₂ (see spectra in Supplemental Material);

Addition of an oxidant (1 mM K₃[Fe(CN)₆]), after resuspension of extensively dehydrated trehalose glasses, increases the relative amplitude of the slow phase of P⁺ decay (not shown);

When the exposure of the sample to N₂ is limited to a few hours and further dehydration is obtained under vacuum, a larger recovery of the slow kinetic component of charge recombination is observed upon rehydration of the solid matrix (not shown);

Addition of oxidized ubiquinone-10 after sample resuspension restores the original extent of the slow phase (>90%), even in the driest samples (see Fig. 1 C, trace c'', and Table 1, samples c and d). This shows that the partial irreversibility of the effects induced by prolonged dehydration of the trehalose samples is not due to irreversible damages of the Q_B site.

Interestingly, partial reduction of redox centers induced by drying has been recently observed by electron paramagnetic resonance in membrane fragments from a photosynthetic bacterium (Lieutaud et al., 2003).

The above points indicate that, when trivial effects due to the prereduction of Q_B in a fraction of RCs are taken into account, dehydration of the trehalose matrix reversibly impairs electron transfer to Q_B at room temperature.

Fig. 2 summarizes the results of the kinetic analysis of P⁺ decay performed in trehalose solutions and in a series of dehydrated amorphous matrices. Kinetic parameters, obtained from traces similar to those shown in Fig. 1, are plotted as a function of trehalose concentration. In Fig. 1 A, the amplitude of the slow component of P⁺ decay measured in dehydrated amorphous samples (trehalose concentration 90 wt %) has been normalized to that determined upon resuspension of the same matrix. This procedure corrects for the extent of the fast P⁺ recombination due to prereduction of Q_B (see above) and allows an estimate of the fraction of RC in which Q_A^--to-Q_B electron transfer is reversibly inhibited. The amplitudes measured in the liquid samples have been normalized to that measured at 16 wt % (i.e., at the trehalose concentration obtained upon resuspension of dehydrated samples). Fig. 2 corroborates and better defines the behavior indicated by the data of Fig. 1 and Table 1. The increase of trehalose concentration above 90 wt % results in the appearance of a fast rate distribution. The amplitude of this kinetic component increases with dehydration and predominantly accounts for P⁺ decay at 97 wt % trehalose (Fig. 2 A). The values measured for the average rate constant, k_f, and for the width, _f, of this fast rate distribution (Fig. 2 B) are quantitatively consistent with those measured for P⁺Q_A^- recombination at the corresponding trehalose concentrations in RC deprived of Q_B (Palazzo et al., 2002). The residual slow kinetic component (P⁺Q_B^- recombination) exhibits an average rate constant (k_s) which decreases when the trehalose concentration increases up to 90 wt % (Fig. 2 C). Under these conditions, Q_A^--to-Q_B electron transfer remains fast compared to the recombination of P⁺Q_A^- (i.e., k_AB + k_BA >> k_AP) (see below). Since the states Q_A^-Q_B and Q_AQ_B^- are in rapid equilibrium, the observed decrease in the P⁺Q_B^- recombination rate seems to reflect a slight stabilization of the charge-separated P⁺Q_B^- state with respect to P⁺Q_A^-. The width _s of the slow rate distribution tends to increase with trehalose concentration (Fig. 2 C). This slight broadening of the rate distribution likely reflects the decreased interconversion between protein conformational substates with increasing rigidity of the matrix.

View larger version (16K): [in this window] [in a new window]   FIGURE 2  Kinetic parameters of charge recombination as a function of trehalose concentration in solution and dehydrated amorphous samples. (A) Normalized amplitude AS of the slow kinetic component (circles). (B) Average rate constant kf (squares) and width f (triangles) of the fast kinetic component. At trehalose concentrations 90 wt % the fast kinetic component was fitted to a simple exponential (see Eq. 2) with kf fixed to 8.2 s-1. Correspondingly, no error bar is associated to kf and no f value is plotted at these concentrations. (C) Average rate constant ks (diamonds) and width s (inverted triangles) of the slow rate distribution. Error bars give the confidence intervals within two standard deviations. See text for details.

Comparison with low temperature data: homogeneous vs. heterogeneous inhibition of Q_A^--to-Q_B electron transfer in dehydrated trehalose matrices
When RCs are frozen in the dark, the yield of electron transfer from Q_A^- to Q_B diminishes with temperature, so that almost no reaction is seen at <200 K (Kleinfeld et al., 1984a; Xu and Gunner, 2001). By examining the fraction A_f of fast P⁺ decay after a flash, Xu and Gunner (2001) determined the temperature-dependence of the quantum efficiency of P⁺Q_B^- formation, defined as = (1 - A_f) k_AB/(k_AB + k_AP) (see Eq. 1); they obtained evidence that the decrease of at low temperatures reflects the competition within a homogeneous RC population between electron transfer from Q_A^- to Q_B (governed by the temperature-dependent rate constant, k_AB) and the return to the ground state of P⁺Q_A^- (determined by k_AP, assumed to be temperature-independent; see Eq. 1). Our data (Fig. 2 A) show that for trehalose-coated RCs the fraction A_s of slow P⁺ decay after a flash at room temperature decreases in parallel with dehydration of the matrix, practically vanishing in relatively dried matrices. This decrease of A_s upon dehydration resembles the decrease observed at cryogenic temperatures (Xu and Gunner, 2001). However, as mentioned in the previous paragraph, we interpret the decreased extent of P⁺Q_B^- recombination in dry samples as due to a block of electron transfer from Q_A^- to Q_B in a fraction of the RC population. Different observations, discussed in the following, suggest in fact that progressive dehydration of the trehalose matrix at room temperature does not slow progressively Q_A^--to-Q_B electron transfer homogeneously over the whole RC population, but rather prevents electron transfer in a progressively increasing subpopulation of the RCs, at least over the timescale of P⁺Q_A^- recombination. The increasing fraction of inhibited proteins in turn results in a progressively higher contribution of fast P⁺Q_A^- recombination to P⁺ decay.

This interpretation is supported by the outcome of a double flash experiment, similar to the one performed at low temperature by Xu and Gunner (2001). The experiment consists in photoexciting the RC sample with two closely spaced flashes, the second flash being fired when most of the P⁺Q_A^- RCs have returned to the ground state but the P⁺Q_B^- RCs have not. Information is obtained by comparing the kinetics of P⁺ decay after the second flash with that observed after a single photoexcitation in the dark-adapted sample. If the fast component of P⁺ decay is due to a complete inhibition of the Q_A^--to-Q_B electron transfer in a given subpopulation of RCs, the decay kinetics after the second flash will coincide with the one observed after the first flash. In this case, in fact, the second flash will re-excite the fraction of P⁺Q_A^- which has decayed in the time elapsed between the two flashes, thus restoring the situation produced by the first flash. As a consequence, the amplitude and the kinetics of fast phase decay will be the same as after the first flash. On the other hand, if the fast decay results from a decreased quantum yield of P⁺Q_B^-, a fraction of the P⁺Q_A^- reformed by the second flash will transfer the electron from Q_A^- to Q_B, thus increasing the slow decaying (P⁺Q_B^-) fraction. In such a case a lower fast phase amplitude will be observed after the second flash.

Measurements have been performed in a dehydrated trehalose matrix characterized by 50% fast P⁺ decay after a laser flash (Fig. 3). The energy of the laser pulse was increased to reach saturation of P photo-oxidation (as can be seen from the equal amplitude of P⁺ absorbance changes induced by the first and second photoexcitations in Fig. 3). The second flash was fired 380 ms after the first photoexcitation, i.e., when the fast phase of P⁺ decay after the first flash had essentially decayed. As shown in Fig. 3 (inset), the decay after the second flash overlaps with the one observed after the first flash, indicating that the fast decaying fraction does not decrease after the second photoexcitation. This coincidence was confirmed by fitting to Eq. 4 the recombination kinetics after the second flash and P⁺ decay after a single laser pulse recorded in the same dark-adapted sample. The relative amplitude of the fast phase was 46.4% and 45.9% after a single flash and after the second flash, respectively. Analogous results were obtained in similar measurements, in which the dark time between the two photoexcitations was reduced to 280 ms or increased up to 600 ms (not shown). These results clearly show that, at variance with that observed at low temperature by Xu and Gunner (2001), dehydration of the glassy matrix inhibits the Q_A^--to-Q_B electron transfer inhomogeneously.

View larger version (24K): [in this window] [in a new window]   FIGURE 3  Charge recombination kinetics after photo-oxidation of the primary donor P induced by two consecutive laser pulses fired 380-ms apart. The trace is the result of a single measurement (without averaging) in a dehydrated trehalose glass. In the inset the normalized P+ decay after the second (open squares) and the first (solid circles) laser pulse are superimposed.

The time-resolved electrochromic measurements described below are also consistent with the inhomogeneous character of electron transfer inhibition induced by dehydration.

Kinetics of electrochromic effects associated with Q_A^--to-Q_B electron transfer in trehalose solutions and solid matrices
Electrochromic effects induced in bacteriopheophytin and bacteriochlorophyll bands by the formation of Q_A^- and Q_B^- offer a valuable means to monitor electric charge displacements resulting from Q_A^--to-Q_B electron transfer (Vermeglio and Clayton, 1977; Shopes and Wraight, 1985; Tiede et al., 1996). NIR electrochromism of RC co-factors has revealed complex time- and wavelength-dependent responses to Q_A^--to-Q_B electron transfer in which the contribution of electron transfer and of relaxation processes, such as proton transfer or protein rearrangements, is not easily separated (Tiede et al., 1996, 1998; Li et al., 1998). Recent time-resolved measurements indicate that the kinetics of Q_A^--to-Q_B electron transfer includes at least two kinetic components (Tiede et al., 1996; Li et al., 1998, 2000). With the aim of gaining additional, independent information on the kinetics of Q_A^--to-Q_B electron transfer in trehalose solutions and solid matrices, we measured laser-induced electrochromic absorbance changes at 750 nm. Electrochromic transients recorded in solution and in two dehydrated trehalose matrices have been analyzed in terms of two exponential components. Results are summarized in Table 2. In the absence of trehalose, the lifetimes obtained for the fast and slow phase are respectively 20 µs and 160 µs, each component accounting for 50% amplitude of the kinetics. Considering the large uncertainty resulting from our confidence analysis of fitting (see Materials and Methods), these values compare well with the results of analogous measurements performed in solution at 757 nm (Tiede et al., 1996) and 398 nm (Li et al., 1998). In a 53 wt % trehalose solution, a sizable slowing of both kinetic components is observed; such an effect becomes progressively more pronounced when an amorphous trehalose matrix is formed upon dehydration (Table 2). At 95.4 wt % trehalose, the lifetimes of the fast and slow components increase to 160 µs and 2.4 ms, respectively. Further drying of the sample prevents a quantitative analysis of the kinetics recorded in the 750-nm region. This is due to a net decrease of the signal amplitude (see traces in Supplemental Material), with the consequent progressive decrease of the signal-to-noise ratio, and to a progressive red shift of the isosbestic point for P⁺Q_A^- spectral contributions (see Supplemental Material), which, from 750-nm, shifts to 756 nm in the driest samples. This notwithstanding, traces recorded in extensively dehydrated samples (not shown) indicate that: 1), the amplitude of electrochromic signals in plasticized samples decreases upon decreasing the water content and practically vanishes at a trehalose concentration close to 97 wt %; 2), in the driest matrices in which a residual electrochromic signal is detectable, it reaches the maximal amplitude within 5–8 ms from the laser pulse; and 3), upon rehydration a large fraction of the original electrochromic signal and the lifetimes of the two kinetic components measured in solution are essentially restored.

Charge recombination after continuous illumination of RCs in dehydrated trehalose matrices
The results presented above strongly suggest that reducing the water content of the trehalose matrix below a critical level inhibits the conformational changes gating the Q_A^--to-Q_B electron transfer, which are likely triggered by the formation of the primary charge-separated state P⁺Q_A^-. As a result, progressive dehydration of the matrix impairs the Q_A^--to-Q_B electron transfer in an increasing fraction of RCs. To evaluate to which extent the transition from the inactive to the active conformation is slowed down in this inhibited RC population, we studied the kinetics of charge recombination after continuous illumination of the sample (from a few ms to seconds). The underlying idea is that continuous illumination, keeping the RC in the primary charge-separated state (P⁺Q_A^-) for a sufficiently long time, will eventually allow the slowed conformational transition to occur in the "inhibited" fraction of the RCs. In this case, we expect that the kinetics of P⁺ decay, after a sufficiently long and intense photoexcitation, will show a larger fraction of the slow phase, as compared to the kinetics monitored after a short laser pulse.

The results of such measurements, performed in three samples labeled 1, 2, and 3, characterized by an increased dehydration and exhibiting an increasing inhibition of Q_A^--to-Q_B electron transfer, are shown in Fig. 4. Fig. 4 A compares kinetic traces of charge recombination measured in a moderately dehydrated sample (sample 2 in Fig. 4 B) after a laser pulse (trace a), a 300-ms (trace b), and a 3.5-s (trace c) period of continuous illumination. In this sample, the amplitude of the residual slow phase of P⁺ decay, 57% of the total when the RC is photoexcited by a laser pulse, increases to 67% and 76% after continuous photoexcitations of 300 ms and 3.5 s, respectively. Under all tested conditions (different dehydration of the sample, different duration of photoexcitation) P⁺ decay could be accurately fitted to the sum of two power laws (Eq. 4). The relative amplitude of the slow kinetic component obtained in samples 1–3 is plotted in Fig. 4 B, as a function of the duration of photoexcitation. The other kinetic parameters of P⁺ decay (not shown) do not exhibit any dependence upon the period of continuous illumination: the values of k_f range between 9 s^-1 and 13 s^-1, increasing from sample 1 to sample 3; the corresponding values of k_s, _s, and _f fluctuate at 0.4 s^-1, 0.2 s^-1, and 5 s^-1, respectively, in agreement with fitting of P⁺ decay after a laser pulse (see Fig. 2).

View larger version (27K): [in this window] [in a new window]   FIGURE 4  Kinetics of charge recombination after continuous illumination of RCs embedded in dehydrated trehalose matrices. (A) Normalized P+ decay after a laser pulse (trace a) and continuous illumination of 300 ms (trace b) and 3.5 s (trace c) duration in a moderately dehydrated matrix (sample 2 in B). Continuous curves are best fit to the sum of two power laws (Eq. 4). (B) Amplitude of the slow component of charge recombination as a function of the duration of photoexcitation at increasing degree of dehydration of the trehalose matrix. Data were obtained by fitting P+ decay to Eq. 4 as shown in A. The trehalose concentration was 94.8 wt %, 95.6 wt %, and 96.3 wt % in samples 1, 2, and 3, respectively. Continuous curves represent best fit to exponential kinetics characterized by rate constants of 4.9 s-1, 1.7 s-1, and 0.38 s-1, for samples 1, 2, and 3, respectively.

We conclude that the recovery kinetics of the slow phase of charge recombination after a period of continuous illumination reflect the kinetics of the conformational change(s) gating Q_A^--to-Q_B electron transfer in the inhibited population of RCs. The transition, indeed, does not occur in this fraction of RCs when a short laser pulse excites P, since it would be too slow to compete successfully with P⁺Q_A^- recombination. Consistently, the rate constant that characterizes the recovery kinetics of the slow component of charge recombination (Fig. 4 B), is always smaller than that (k_AP) of P⁺Q_A^- recombination (from a factor 2 in "wet" sample 1, to almost two orders of magnitude in sample 3). After continuous illumination of increasing duration, the transition from the inactive to the active conformation can be monitored: its kinetics are markedly slowed down upon dehydration, likely due to the hardening of the glassy matrix. By comparing the recovery kinetics of the amplitude of the slow recombination component shown in Fig. 4 B with the kinetics measured for Q_A^--to-Q_B electron transfer in solution, exploiting the electrochromic effects, one can estimate the extent to which the incorporation into dehydrated matrices slows down the transition from the inactive to the active conformation. Even assuming that only the slow phase of the electrochromic effects associated with Q_A^--to-Q_B electron transfer reflects the gating conformational change, it appears that in the more extensively dehydrated trehalose matrix the transition from the inactive to the active conformation is slowed by at least four orders of magnitude. In fact, it can be evaluated that the rate constant k of the gating conformational change decreases from 6 x 10³ s^-1 in solution (_s 160 µs, in the absence of trehalose; see Table 2) to 3.8 x 10^-1 s^-1 (Fig. 4 B, sample 3).

Effects of dehydration and role of trehalose
To better understand the role of trehalose in the effects described above and to examine to what extent dehydration per se inhibits Q_A^--to-Q_B electron transfer, we studied the kinetics of P⁺ decay in dehydrated RC films, in the absence of the sugar. As already observed with RCs deprived of Q_B (Palazzo et al., 2002), in the absence of trehalose, a few hours of exposure to a flow of dry N₂ causes exhaustive sample dehydration. Indeed, no absorption is detected in the NIR band at 1950 nm, which we exploited for estimating the sample water content (Fig. 5 B, lower spectrum). Kinetic analysis of P⁺ decay (Fig. 5 A, trace a) shows that, under these conditions, in 50% of the RC population fast recombination occurs from the P⁺Q_A^- state. Resuspension of the film results in a considerable recovery of the slow phase amplitude (A_s = 0.8, not shown), which is essentially complete upon addition of ubiquinone-10 (Fig. 5 A, trace b). In the absence of trehalose, therefore, dehydration reversibly inhibits Q_A^--to-Q_B electron transfer in a fraction of the RC population. However, the behavior observed in trehalose matrices is substantially different. Upon dehydration, a comparable inhibition of Q_A^--to-Q_B electron transfer, resulting in 50% fast P⁺ recombination, occurs in the saccharide glass at a higher water content (see Fig. 5 B, upper spectrum) as compared to the RC dried in the absence of trehalose. Actually an almost complete inhibition of the reaction (see Fig. 2 A) is observed in trehalose samples at 97 wt % trehalose, corresponding to >10⁴ water molecules per RC. By contrast, if RC films are dehydrated in the absence of trehalose, inhibition of Q_A^--to-Q_B electron transfer can be achieved only in a limited fraction of the RC population, despite extremely low water contents. Although extensive dehydration per se affects markedly the yield of Q_B^- formation, a much more effective inhibition of conformational dynamics assisting Q_A^--to-Q_B electron transfer occurs in moderately dehydrated trehalose-coated RCs as compared to RCs dried in the absence of trehalose. Analogous evidences of a specific role of the trehalose matrix in affecting the protein dynamics were obtained when studying P⁺Q_A^- recombination in RCs deprived of the secondary quinone acceptor (Palazzo et al., 2002). In this system extensive RC dehydration in the absence of the saccharide had almost no effect on P⁺Q_A^- recombination kinetics, which exhibited a very narrow rate distribution ( 5 s^-1) centered at k 11 s^-1, a value close to the one measured in solution. In agreement with this observation, in dry films of RCs in the presence of the secondary acceptor Q_B, the fast phase of P⁺ decay was characterized by a narrow spectrum of rate constants (_f = 2.3 s^-1) centered at k_f = 8.7 s^-1 (Fig. 5 A, fit to trace a). These results suggest that in the absence of trehalose, at a very low water content, the RC protein retains some conformational flexibility; interestingly (see below), under such conditions, whereas P⁺Q_A^- recombination kinetics is scarcely affected, electron transfer to Q_B is hampered reversibly in almost half of the RC population.

View larger version (22K): [in this window] [in a new window]   FIGURE 5  Kinetics of charge recombination in RC dried without trehalose. (A) Normalized P+ decay in a dry film (trace a) and after resuspension of the film and addition of 60 µM ubiquinone-10 in LDAO (final detergent concentration 0.06%; trace b). The solid line through trace a is a fit to Eq. 4, yielding the following values of parameters according to Eq. 5: As = 0.49 (0.46, 0.51); ks = 0.42 s-1 (0.37, 0.43); s = 0.22 s-1 (0.18, 0.25); kf = 8.7 s-1 (7.6, 10.4); and f = 2.3 s-1 (0, 5.1). Continuous curve through trace b represents a fit according to Eq. 2 with kf fixed to 8.2 s-1; best fitting yields Af = 0.097 (0.085, 0.108); ks = 0.628 s-1 (0.624, 636); and s = 0.19 s-1 (0.18, 0.20). Values in parentheses indicate the extremes of the confidence intervals of the fitting parameters within two standard deviations. (B) The visible NIR spectrum recorded in the dry RC film (lower spectrum) is compared to the spectrum recorded in a dehydrated trehalose matrix (95.4 wt % trehalose). The kinetics of charge recombination recorded in this sample is shown in Fig. 1 B as trace c. The position of the water combination band at 1950 nm is indicated by the arrow. The two spectra have been normalized at the maximal amplitude at 802 nm and the spectrum of the trehalose matrix has been offset by 0.2 absorbance units for the sake of visual clarity.

The relationship between conformational relaxation of the P⁺Q_A^- state and electron transfer to Q_B
The kinetic analysis of charge recombination in trehalose samples enables one to monitor at the same time the effect of dehydration on two electron transfer processes. The average rate constant k_f and the distribution width _f of the fast component of P⁺ decay give information on the distributed kinetics of recombination between P⁺ and Q_A^-. The kinetics of this process has been considered as a probe of RC internal dynamics (Kleinfeld et al., 1984a; Feher et al., 1987; McMahon et al., 1998; Palazzo et al., 2002). On the other hand, the relative amplitude of the slow phase of P⁺ decay (A_s) measures the fraction of the RC population in which Q_A^--to-Q_B electron transfer occurs at a rate high enough to compete successfully with P⁺Q_A^- recombination. The relationship between these two parameters, obtained from the data of Fig. 2, A and B, is presented in Fig. 6. When Q_A^--to-Q_B electron transfer is inhibited in a increasing fraction of RCs, i.e., A_s decreases, the average rate of P⁺Q_A^- recombination, k_f, never exceeds values of 12–14 s^-1, even at an almost complete inhibition of Q_A^--to-Q_B electron transfer. The corresponding values of the rate distribution width, _f, obtained in these matrices range between 4 s^-1 and 6 s^-1 (see Fig. 2 B).

View larger version (21K): [in this window] [in a new window]   FIGURE 6  Average rate constant of the fast phase of charge recombination (kf) as a function of the corresponding amplitude of the slow kinetic component (As) in a series of trehalose samples (see Fig. 2, A and B) characterized by a different degree of dehydration (solid squares) and in a dehydrated RC film in the absence of trehalose (open square). The relative amplitude As is normalized to the amplitude of the slow phase determined upon resuspension of the dehydrated matrix. The arrows pointing to the right and to the left indicate the average rate constant of P+QA- recombination measured in the absence of QB in solution, and in a solid trehalose matrix, extensively dehydrated, respectively (Palazzo et al., 2002). Error bars give the confidence intervals within two standard deviations.

A variety of structural changes have been proposed as responsible for the gating of electron transfer to Q_B. The accumulated x-ray data show several positions and orientations for Q_B, which can be brought back essentially to two possible binding sites (Walden and Wheeler, 2002). In the crystal structure of the RC cooled to cryogenic temperatures under illumination, i.e., trapped in an active state, Q_B was found 2.7 Å closer to Q_A than in the protein frozen in the dark and flipped by 180° around the isoprenoid chain (Stowell et al., 1997). The movement of Q_B from an inactive-distal to an active-proximal site was proposed as the major structural change involved in the conformational gating step (Stowell et al., 1997). However, a mutated RC (L209PY), in which Q_B was found in the crystal structure in the proximal position even in the dark (Kuglstatter et al., 2001), exhibited practically the same rate of Q_A^--to-Q_B electron transfer as the wild-type (Baciou and Michel, 1995). Mor