Functionality of Photosynthetic Reaction Centers in Polyelectrolyte

Mar 8, 2007 - Such a system has been characterized from the functional point of view by ... Gerardo Palazzo , Francesco Francia , Antonia Mallardi , M...
0 downloads 0 Views 329KB Size
Download PDF
3304

J. Phys. Chem. B 2007, 111, 3304-3314

Functionality of Photosynthetic Reaction Centers in Polyelectrolyte Multilayers: Toward an Herbicide Biosensor Antonia Mallardi,*,† Mauro Giustini,‡ Francesco Lopez,§ Manuela Dezi,| Giovanni Venturoli,| and Gerardo Palazzo*,§ Istituto per i Processi Chimico-Fisici, CNR, Via Orabona 4, 70126 Bari, Italy, Dipartimento di Chimica, UniVersita` “La Sapienza”, I-00185 Roma, Italy, CSGI and Dipartimento di Chimica, UniVersita` di Bari, Via Orabona 4, I-70126, Bari, Italy, and Dipartimento di Biologia and CNISM, UniVersita` di Bologna, Italy ReceiVed: December 6, 2006; In Final Form: January 24, 2007

The bacterial reaction center (RC), a membrane photosynthetic protein, has been adsorbed onto a glass surface by alternating deposition with the cationic polymer poly(dimethyldiallylammonium chloride) (PDDA) obtaining as an end result an ordinate polyelectrolyte multilayer (PEM) where the protein retains its integrity and photoactivity over a period of several months. Such a system has been characterized from the functional point of view by checking the protein photoactivity at different hydration conditions, from extensive drought to full hydration. The kinetic analysis of charge recombination indicates that incorporation of RCs into dehydrated PEM hinders the conformational dynamics gating QA- to QB electron-transfer leaving unchanged the protein relaxation that stabilizes the primary charge separated state P+QA-. The herbicide-induced inhibition of the QB activity was studied in some detail. By dipping the PEM in herbicide solutions for short times, kinetics of herbicide binding and release have been determined; binding isotherms have been studied using PEM immersed in herbicide solution. QB functionality of RC has been restored by rinsing the PEM with water, thus allowing the reuse of the same sample. This last point has been exploited to design a simple optical biosensor for herbicides. A suitable kinetic model has been proposed to describe the interplay between forward and back electron-transfer processes upon continuous illumination, and the use of the PDDA-RC multilayers in herbicide bioassays was successfully tested.

1. Introduction The development of protein-friendly materials that can house enzymes or other proteins with retention of their functionality is receiving a growing attention by researchers in the fields of chemistry, biology, physics, and material science. The efficient immobilization of proteins on suitable matrices and their integration on precisely engineered nanoscale architectures is often a key step for their use in novel technical applications.1,2 Therefore, the study of protein functionality within non-natural, non-liquid environments plays an important role in the development of devices for sensing, energy storage/conversion, and bioelectronic applications.3,4,5,6,7 At the same time, the functional characterization of proteins within non-native environments can yield new insight into their mechanism of action, shedding light particularly on the relation between function and internal dynamics of the protein.8 Of course, these two aims of research (application- and protein-oriented) are fully interconnected. In the present work, we have extended our previous studies on the functionality of the photosynthetic bacterial reaction center (RC) embedded into amorphous matrices (reviewed in ref 8) to the assembly of this membrane protein and a cationic polymer, poly(dimethyldiallylammonium chloride) (PDDA), in ordinate multilayers. The alternate deposition of polyanions and poly* Corresponding authors. Address: Dipartimento di Chimica, via Orabona 4, I-70126, Bari, Italy. E-mail: [email protected] (G.P.); a.mallardi@ ba.ipcf.cnr.it (A.M.). † Istituto per i Processi Chimico-Fisici. ‡ Dipartimento di Chimica, Universita ` “La Sapienza”. § CSGI and Dipartimento di Chimica, Universita ` di Bari. | Dipartimento di Biologia and CNISM.

cations on a solid surface leads to the formation of films called polyelectrolyte multilayers (PEM). Such films are readily prepared by layer-by-layer assembly, a simple process based on the sequential adsorption, driven by electrostatic interactions, of cationic and anionic species on a charged substrate (for reviews, see refs 9 and 10). This technique has been used to prepare ordered protein films by alternating polyion layers with a number of different positively and negatively charged soluble proteins.11 The incorporation into PEM of an integral membrane protein has been also achieved, using as a guest model system the photosynthetic reaction center (RC) purified from purple bacteria.12,13 The bacterial RC is a pigment-protein complex that in ViVo spans the intracytoplasmic membrane. It promotes the primary events of photosynthetic energy transduction by catalyzing a light-induced charge separation across the protein dielectric through a sequence of electron-transfer events involving cofactors held at fixed distances by the polypeptide scaffolding.14 Upon photon absorption, a bacteriochlorophyll special pair (P), facing the periplasmic side of the intracytoplasmic membrane, acts as the primary electron donor by delivering an electron to the primary acceptor, QA, a ubiquinone molecule bound at a site close to the opposite, cytoplasmic side of the complex. This primary charge separation, accomplished in about 200 ps, is stabilized by electron transfer from QA- to a second ubiquinone molecule, bound at the QB site of the RC, which is located symmetrically to the QA site on the same cytoplasmic side of the membrane.14 In ViVo, the photoxidized donor, P+, is rapidly re-reduced by a soluble c-type cytochrome so that a second charge separation can take place across the RC, leading to the

10.1021/jp068385g CCC: $37.00 © 2007 American Chemical Society Published on Web 03/08/2007

PDDA-RC multilayers double reduction and protonation of QB, which leaves the RC in its quinol state, QH2. In this reaction, two protons are taken up from the cytoplasmic side of the membrane so that the RC catalyzes the first step in the generation of the transmembrane electrochemical potential for protons, which is the driving force for ATP synthesis.15 In Vitro, when no physiological or artificial electron donor is available to re-reduce flash generated P+, the electron on QB- recombines with the hole on P+, restoring the initial ground state of the RC. A striking functional homology exists between the electron acceptors of the bacterial RC and of photosystem II in plants.16 In both systems, the two acceptor quinones (QA and QB) act in series and function as a two electron gate that passes reducing equivalents out of the RC only in pairs, i.e., when QH2 is formed at the QB site. This homology is reflected by similar inhibitor sensitivities: many economically important herbicides, largely used in agriculture, known to act at PS II, are effective in the bacterial RC, where they compete with quinone for binding at the QB site. In view of these properties and of the crystallographic information available, the bacterial RC has become over the last decades the reference model in studying several features of the more complex PS II system.17 In the present paper, we show that bacterial RCs can be incorporated into PDDA multilayers, preserving the functionality of the QB site as an electron acceptor. Kinetics of charge recombination following a single-turnover photoexcitation are found to be affected by the hydration state of the PDDA-RC multilayers. This sensitivity has been put in relation with similar effects observed on RCs embedded in different non-liquid matrices of low water content. The PDDA-RC multilayer turned out to provide unique advantages in investigating the thermodynamics and kinetics of herbicide binding. An assay for herbicides, based on the response of the PDDA-RC multilayer to continuous illumination, has been extensively tested and modeled quantitatively. On the basis of these results, the properties of the system are discussed in view of its use in the design of a simple optical biosensor for herbicides. 2. Experimental N,N-Dimethyldodecylamine-N-oxide (LDAO) was purchased from Fluka as a 30% aqueous solution. n-Octyl-β-D-glucopyranoside (octylglucoside, OG), o-phenanthroline, terbutryn, and low molecular weight poly(dimethyldiallylammonium chloride) (PDDA, 20% water solution) were from Sigma-Aldrich. RCs from Rhodobacter (Rb.) sphaeroides were isolated and purified according to Gray et al.18 In all preparations, the ratio of the absorption at 280 and at 800 nm was between 1.2 and 1.3. This isolation procedure gives RCs with a QB content/activity of about 60%. In order to concentrate the sample, exchange LDAO with octylglucoside, and reconstitute the secondary quinone acceptor, a small (1-2 mL) DEAE Cellulose DE-52 (Whatman) anion exchange column, previously equilibrated with buffer (Tris-HCl 20 mM, pH ) 8, 0.025% LDAO) was loaded with the purified RC. Then, the column was washed first with buffer (Tris-HCl 20 mM, pH ) 8 (100 vol)) in order to eliminate LDAO and subsequentely with 100 vol of ubiquinone containing buffer (Tris-HCl 20 mM, pH ) 8, 0.8% OG (buffer TOG)). The low solubility of UQ10 in aqueous buffers was overcome by adding the quinone to a 30% OG stock solution before dilution for buffers preparation. Finally, UQ10 reconstituted RCs were eluted with the same buffer also containing 400 mM NaCl. The RC was dialyzed overnight against the TOG buffer (in the absence of NaCl) at 4 °C and in the dark. After this procedure, the QB activity of the reconstituted samples was typically higher

J. Phys. Chem. B, Vol. 111, No. 12, 2007 3305 than 90%. RC-Octylglucoside complexes were frozen in liquid nitrogen and stored at -80 °C. The concentration of the RC was determined spectrophotometrically from the absorbance at 802 nm (802 ) 288 mM-1 cm-1). The protein assembly by layer-by-layer adsorption of the negatively charged RC and the positively charged PDDA was performed on a glass slide (1 × 5 cm, Hellma). Substrates were negatively charged by using an oxidative cleaning in “piranha” solution (oleum H2SO4 and 30% H2O2 in a 3:1 ratio) for 15 min on ice (care should be taken in preparing and handling this solution, as the reaction is exothermic and the solution is highly corrosive) and subsequently left in Millipore water for 5 min. Finally, the substrates were rinsed with acetone and dried with nitrogen. To adsorb a polycation layer, the slide was immersed in a 2 mg/mL PDDA/water solution. Subsequently, it was washed in Milli-Q water, and a layer of negatively charged RC was adsorbed by dipping the specimen into 10 mM Tris-HCl buffer, pH) 8.0, containing 0.8% OG and 3 µM RC. After rinsing with Milli-Q water, the above-described steps were repeated in order to obtain the required number of PDDA-RC multilayers. In this way, both sides of the slide are covered with the same number of polyelectrolite multilayers. In the following, we will define the procedure by the number of PDDA-RC alternate layers grown on one side of the glass plate (i.e., by the number of immersions in the RC solution). To reduce the possibility of RC desorption, the last layer was always made by PDDA; adsorption and washing times were 30 and 5 min, respectively; RC and polycation layers were grown in the dark at 4 °C. The solutions for inhibitor binding study have been all prepared by water dilution (performed into 10 mL measuring flask) of a concentrated ethanol stock solution of the inhibitors. The utmost care has been posed in keeping in all the samples the concentration of the ethanol constant to 2% (v/v). Using this procedure, solutions of terbutryn in the range 1 nM to 100 µM and for o-phenantroline in the range from 5 µM to 2.5 mM have been obtained. Rapid absorbance changes in single-turnover experiments were monitored by using one of two kinetic spectrophotometers of local design with exciting beams at right angles to the measuring beams: instrument A, described in detail in ref 19 allows the wavelength selection of the measuring beam in the range 400-700 nm; at variance, the measuring beam of instrument B was provided by a laser diode emitting at λ ) 870 nm (for further details, see ref 20). In both instruments, the exciting pulse (λ ) 532 nm) was provided by a frequency doubled Nd:YAG laser (Quanta System, Handy 710; 7 ns pulse width; 200 mJ pulse energy) and the detector was protected by a monochromator plus a suitable interference filter. The lightminus-dark spectrum in Figure 2 was recorded with instrument A using as exciting light a xenon flash lamp as described in ref 19. Time dependence of bleaching at 870 nm, induced by continuous illumination (see Section 3.4) was monitored with instrument B. In this experiment, the intensity of the measuring diode beam at 870 nm was high enough to photoexcite the RC; in this case, therefore, the same source was providing measuring and actinic light. Because reproducible positioning of the sample is difficult to achieve and absolute values of absorbance changes are crucial in these experiments, the measuring beam was defocused to illuminate a large part of the PDDA-RC PEM. The glass plate supporting the PDDA-RC multilayers was placed at 45° with respect to both the measuring beam and the actinic flash by inserting the plate along the diagonal of a 1 cm

3306 J. Phys. Chem. B, Vol. 111, No. 12, 2007

Figure 1. (A) Visible-NIR absorption spectra of PDDA-RC multilayers obtained following a different number of RC adsorption cycles (indicated in the labels). (B) Lower plot: absorbance (right linear ordinate) at 802 nm (closed circles) and 900 nm (open circles) as a function of the number of RC adsorption cycles. Upper plot (closed triangles): difference between absorbance measured at 802 and 900 nm (left logarithmic ordinate) as a function of the number of adsorption cycles. Solid lines represent exponential fits.

path length fluorescence cuvette. When required, the cuvette was filled with a suitable solution. The water content of the specimen was assayed by NIR spectroscopy on a Perkin-Elmer, Lambda 19, UV-vis-NIR spectrometer; UV-vis spectra were collected on the same instrument or on a UV-vis Cary 3 spectrophotometer. Fitting of experimental data to the relevant equation was performed by nonlinear least-squares minimization as described in ref 21. 3. Results and Discussion 3.1. Preparation of PDDA-RC Multilayers. The main driving forces for layer-by-layer assembly are electrostatic interactions. Due to their huge charge, polyelectrolytes strongly adsorb on an oppositely charged surface, and because of their large size, the electrostatic potential drop within the Stern layer leads to the charge reversal. As a consequence, immersion of a negatively charged surface (glass) into a solution of cationic polyelectrolyte (e.g., PDDA) results in the coverage of the surface by a positively charged polymer layer. Of course, immersion of this new cationic surface in a solution of

Mallardi et al.

Figure 2. Single-turnover experiments in dry PDDA-RC multilayers (25 cycles of RC adsorption). For each kinetic trace, 5000 data points were used in the fit; to increase readability, only 250 points are shown in the figures. (A) Charge recombination kinetics measured at three wavelengths in PDDA-RC PEM after a laser pulse fired at time t ) 0; dots represent experimental data and lines the fits according to eq 3. Best fit parameters are as follows: at 422 nm, Af ) 0.34, kf ) 9.3 s-1, λ ) 1.2 s-1, σ ) 1.3 s-1; at 600 nm, Af ) 0.39, kf ) 10.1 s-1, λ ) 1.4 s-1, σ ) 1.4 s-1; at 550 nm, Af ) 0.34, kf ) 9.3 s-1, λ ) 1.6 s-1, σ ) 1.6 s-1. (B) The same data of panel A normalized to the maximal absorbance change recorded immediately after the photoexcitation pulse and reported in a semilog plot. Inset: light-dark spectrum; each point at a given wavelength represents the absorbance change extrapolated at the time of the flash (t ) 0) by fitting the kinetic trace to eq 3.

negatively charged protein results in the protein adsorption and in the re-establishment of a negative charge density. Basically, the process can be repeated indefinitely as long as the charge of the cat- and anionic adsorbed polyelectrolytes remains unaffected by the immersion in the three relevant solutions, viz., cat- and anionic polymer solutions and water (in passing between the two adsorbing baths, the specimen must be rinsed with water; otherwise, free polyelectrolites will interact in solution forming neutral complexes). In the case of membrane proteins, what is relevant is the charge of the protein-detergent complex. RC is extracted from the bacterial intracytoplasmic membrane by using N,N-dimethyldodecylamine-N-oxide (LDAO), a neutral detergent. However, the N-oxide moiety of LDAO can be protonated at acidic pH leading to the formation of the cationic detergent N,N-dimethyldodecylammonium-N-hydroxide. The presence of negative charges on the RC surface shifts the pKa of protein-bound LDAO to physiological values so that for pH < 7.5 the protein detergent complex is neutral.22,23 In preliminary attempts to form polycation-RC multilayers using the LDAO-RC complex, we found a low adsorption efficiency, probably because the

PDDA-RC multilayers polycation adsorption and the rinsing steps were performed in unbuffered water characterized by slightly acidic pH. We have therefore exchanged LDAO with octylglucoside (a pH insensitive detergent) according to the procedure described in Section 2. It should be noted that the assembly of PDDA-RC PEM on gold substrates was previously achieved by Kong and coworkers using as detergent Triton X100 (another pH-insensitive surfactant).12 Using OG-solubilized RCs, the alternate deposition of PDDA and RC layers is very effective as shown by the absorption spectra of Figure 1. PDDA does not absorb in the spectral region considered, and the RC bacteriochlorins give rise to three intense absorbance bands in the region between 700 and 950 nm.24,25 The band at 760 nm is ascribable to the Qy transition of bacteriopheophytin molecules. The band at 802 nm is due to the Qy transitions of the monomeric bacteriochlorophylls, together with a contribution from the high-energy exciton component of the Qy transition of the bacteriochlorophyll molecules forming the special pair P. The long-wavelength band (at 865 nm in solution) is identified as the low-energy exciton component of the Qy transition of P. As shown in Figure 1, absorption spectra of PDDA-RC multilayers retain all these features superimposed to a background scattering, which is a strongly decreasing function of the wavelength. The position of absorption maxima can be safely determined from the first derivative of the absorption spectrum (not shown). Applying this procedure, we have found in the near IR region maxima at 760 and 802 nm, independently from the number of adsorbed RC-layers. At variance, the position of the long-wavelength maximum appeared to depend on the number of layers, ranging from 853 nm (15 and 19 RC deposition steps) to 843 nm (25 RC deposition steps). Therefore, the only difference between RC spectra measured in PDDA-RC PEM and in solution is a 15-22 nm blue shift of the special pair (P) long-wavelength absorption band. The addition of cat- and zwitterionic (betaines) detergents has been shown to induce a moderate (10 nm) blue shift in the position of the Qy-band of P.23,26-28 ENDOR/TRIPLE studies have correlated this effect with a change in the spin density distribution within the P special pair.26-28 A blue shift (to ∼850 nm) has also been observed at very high local concentrations of RC, as in various dry RC films21,29,30 or RC aggregates in solution.23,31 In the present case, the protein film is characterized by a large RC density and by strong electrostatic interactions with the cationic moiety of PDDA. Both these factors are likely to concur in determining the large blue shift observed in PDDA-RC PEM. The adsorption of RC during the PEM assembly can be probed from the absorbance of the band at 802 nm. Kong and co-workers reported a linear dependence of RC absorbance on the number of PDDA-RC multilayers.13 In our system, the optical density at all the wavelengths grows exponentially with the number of adsorption steps (see Figure 1). To probe the growth of protein layers we have subtracted the absorption at 900 nm (were the RC absorbance is very small) from the absorbance at 802 nm (maximum in RC absorbance spectrum). According to the result shown in Figure 1B, the RC absorbance grows exponentially over 3 decades with the number of adsorption cycles. Likely, the use of OG instead of Triton X100 is the rationale for different adsorption behavior found by us and Kong’s group. The PDDA-RC PEMs prepared with the above-described procedure, and stored in distilled water at 4 °C in the dark, display a striking stability, maintaining their spectral features as well as RC photoactivity over a period of several months. During the course of the present study, in many cases, the same

J. Phys. Chem. B, Vol. 111, No. 12, 2007 3307 specimen was exchanged between the different collaborating laboratories and extensively assayed; as described in the next sections, the assays often involved several washing steps. It turned out that, although the amount of RC adsorbed decreased to some extent over a period of a few months, the absorption spectra and the photoactivity (described in the following) retained the features of the native, functional protein. Although most of the work was done within 3-4 months from the multilayer preparation, selected specimen were assayed and found to be functional after 7 months. 3.2. Functionality of RC. The functional integrity of the RC when it acts as a constituent of PEM has been tested by studying light-induced charge separation and recombination reactions in single-turnover experiments. Following a flash of light, an electron is delivered in about 200 ps from the primary donor P to the primary quinone acceptor QA, forming the charge separated state P+QA-.14 When a second quinone molecule is bound at the secondary acceptor site QB, in about 100 µs, it receives the electron from QA-, yielding the P+QB-state.14 In the absence of electron donors to P+, the electron on QB (or QA) can recombine with the hole on the primary donor, as shown by the following equation, which describes charge separation and recombination processes occurring after a short flash of light: hν

kAB

AP

BA

PQAQB {\ } P+QA-QB {\ } P+QAQBk k

(1)

Under physiological conditions, kAP ≈ 10 s-1, kAB ≈ 6 × 103 s-1, and kBA ≈ 4.5 × 102 s-1.32 Charge recombination of the P+QAQB- state occurs essentially by thermal repopulation of the P+QA-QB state with an observable overall rate constant λ ≈ 1 s-1, which is determined by the values of kAP, kAB, and kBA. In the absence of the secondary acceptor QB, or when QAto QB electron transfer is blocked, direct recombination of the P+QA- state occurs so that P+ decays faster with kAP: hν

} P+QAPQA {\ k

(2)

AP

As a consequence, the decay of P+ following a short photoexcitation includes in general two kinetic components, a fast and a slow one, ascribed to RC subpopulations, which undergo P+QA- and P+QB- recombination, respectively. The relative contributions of the two components depend on the fraction of RC in which the final electron transfer from QA- to QB cannot take place. In principle, therefore, according to eqs 1 and 2, kinetic analysis of charge recombination yields information on P+QA- recombination as well as on electron transfer from QAto QB. Because the reduced and oxidized forms of P have different extinction coefficients in the visible spectral range, the transient flash-induced formation of P+ and its subsequent decay by charge recombination can be monitored spectrophotometrically at the appropriate wavelengths and time resolution. Kinetic and spectral features of absorbance changes induced by singleturnover excitation in PDDA-RC multilayers are shown in Figure 2. Absorbance changes detected immediately after photoexcitation over the 540-640 spectral range yield a lightdark differential spectrum (Figure 2B, inset), which coincides with that reported for RC in solution, confirming the capability of the RC in PDDA multilayers to perform charge separation. Figure 2A shows the kinetics of flash-induced absorbance changes (∆Abs) recorded at three different wavelengths for PDDA-RC multilayers allowed to dry in air for 10 min. The absorbance changes at 600 and 550 nm are due to the

3308 J. Phys. Chem. B, Vol. 111, No. 12, 2007

Mallardi et al.

photoxidation and subsequent re-reduction of the primary donor P.33 The change recorded at 422 nm includes contributions from both P+ and from the semiquinone formed on the acceptor complex. The time course of absorption changes is essentially the same at the three wavelengths as shown in Figure 2B, where traces are normalized in amplitude. This implies that the decays of P+ and semiquinone are concomitant and that, therefore, recovery kinetics recorded at these wavelengths are due to charge recombination processes only, without any interference from side reactions re-reducing P+ or reoxidizing the semiquinone. Charge recombination is clearly biphasic. Moreover, attempts to fit the data to the sum of two exponential decays, reveal that the slow phase deviates significantly from an exponential behavior. Incorporation of the RC into dehydrated amorphous matrices (i.e., sugar glasses8,21 and polyvinyl alcohol (PVA) films29) has been previously found to result in non-exponential, distributed kinetics of charge recombination. In the case of P+QBrecombination, this effect, which depends on the chemical nature and on the hydration state of the embedding matrix, occurs even in weakly interacting matrices (PVA) and relatively hydrated states.8,29,34 It has been interpreted as reflecting a matrix-induced inhibition of the RC dynamics, leading to a decreased rate of interconversion between different conformational substates of the RC, each characterized by its individual electron-transfer rate constant. When interconversions become slower than the electron-transfer process examined, kinetics of electron transfer is expected to exhibit a non-exponential character, describable by a continuous rate distribution. It has been shown that distributed kinetic phases of charge recombination can be adequately accounted for by a power law.21,35 Such a function accounts well for the distributed kinetics of the slow component of P+ decay also in RC multylayers, so that an adequate and physically meaningful function fitting charge recombination kinetics shown in Figure 2 is

P+(t) P+(0)

)

(

∆Abs(t) σ2 ) Afe-kft + (1 - Af)/ 1 + ‚t λ ∆Abs(0)

)

λ2/σ2

(3)

where the Af is the fraction of fast phase, λ is the average rate constant, and σ is the width of the rate distribution function characterizing the slow phase. The values obtained by this procedure for the kinetic parameters (listed in the caption of Figure 2) deserve some comments. First, very close values are obtained for Af, kf, λ, and σ at the three wavelengths considered, confirming quantitatively the essential coincidence of the kinetics, as inferred by visual inspection of normalized traces (Figure 2B). The rate constant of the fast exponential phase (kf) assumes, within the experimental uncertainty, the same values found in solution (≈10 s-1) for the charge recombination from P+QA- so that we can safely attribute this phase to RCs in which the final electron transfer from QA- to QB cannot take place (i.e., kf ≡ kAP, see eq 2), either because the protein lacks the quinone at the QB site or because this electron-transfer step is inhibited. On the other hand, the average rate constant characteristic of the slow phase (λ ) 1.2 s-1) is slightly larger than the λ value observed for QB-reconstituted RC in solution (λ ) 1.0 s-1),36 suggesting a destabilization of the P+QAQBwith respect to the P+QA-QB state. Interestingly, the width of the correspondent rate distribution is very high. The σ-value is actually larger than 1.2 s-1 in PDDA-RC multilayers, i.e., much higher than the values found for RC in solution (σ ) 0.19 s-1)34 and when incorporated in amorphous sugar matrices (σ ) 0.25 s-1)34 and polyvinylalcohol films (σ ) 0.20 s-1).29 Most likely,

Figure 3. (A) Charge recombination kinetics measured on the same PDDA-RC PEM at different degrees of hydration. “Dry” and “wet” refers to PEM allowed to desiccate for 10 and 3 min, respectively (black traces). The inset shows enlarged the first part of the decays in a semilog plot (closed squares and inverted triangles are data sampled in the “wet” and “dry” state, respectively. In both plots, open circles refer to measurements performed on the same PEM immersed in water. Because this last kinetics essentially coincide with that measured in the “wet” PEM, only a selected number of points is reported. Fits of kinetics according to eq 3 for wet and dry PEM are shown as gray lines. Best fit parameters are as follows: As ) 0.80, kf ) 10.4 s-1, λ ) 1.2 s-1, σ ) 1.1 s-1 for the wet state; As ) 0.62, kf ) 8.2 s-1, λ ) 0.91 s-1, σ ) 0.8 s-1 for the dry state. (B) NIR spectra of dry and wet PEM. The dashed curve shows the power-law dependence used to describe the background scattering in dry PEM. Inset: the RC spectrum obtained after correction for the background; the spike at 860 nm is an artifact due to the change in light source of the measuring beam.

such a huge polydispersity in charge recombination rates reflects some local heterogeneity in the electric charge distribution within the PEM matrix. The last parameter, characterizing the charge recombination kinetics in PDDA-RC, is the relative weight of the fast phase (Af). In dry PDDA-RC multilayers prepared from RCs with the QB site fully reconstituted, about 62% of the RC population undergoes slow charge recombination from P+QB- (Af ) 0.340.39, see legend of Figure 2). In the remaining RC population, recombination occurs from the P+QA- state. It is quite unlikely that this is due to quinone extraction during the multilayer assembly as demonstrated by the fact that a substantial recovery in the relative amplitude of the slow phase, (1 - Af), is observed upon re-hydration of multilayers (Figure 3A). When the hydration state of PDDA-RC multilayers was examined by NIR spectroscopy, it turned out that during exposition to air, after rinsing in water, the multilayer retains a substantial amount of water during the first 3 min. However, in the subsequent 6-8 min, the content of residual water

PDDA-RC multilayers decreases dramatically, becoming practically undetectable after 10 min, and leading to what we call a dry multilayer (see Figure 3B). In this state, the PDDA-RC multilayer is characterized, as stated above, by a consistent fraction of RCs undergoing fast P+QA- recombination. At variance, when charge recombination kinetics are measured in “wet” PDDA-RC multilayers (within about 3 min after rinsing), P+ decay is dominated by the slow component as shown in Figure 3A. It appears that in “wet” PEMs at least 80% of RCs are fully functional at the QB site, indicating that the amount of water strongly affects the yield of QA- to QB electron transfer in PDDA-RC multilayers. As shown in Figure 3A, charge recombination kinetics of RC in wet multilayers are quite similar to those observed in PEM directly dipped in water (immersion in water results only in a further 5% increase in the relative amplitude of the slow kinetic component). The absorption spectra in the visible-NIR region allow the quantification of the RCs involved in the building of multilayers and of the residual water present in wet specimens. As shown in Figure 3B, dry PEMs lack any absorption band in the region 1000-2200 nm. In both dry and wet PEMs, absorbance increases continuously at decreasing wavelengths. By subtracting this background, which can be fitted to a power-law, a corrected absorption spectrum is obtained (see the inset of Figure 3B) quite similar to that of the RC in solution (except for the blue shift of the long-wavelength band, discussed in Section 3.1). From the absorbance at 800 nm, assuming that the extinction coefficient determined for RC in solution holds also for RC in PDDA-RC, a surface density of 1.1 nmol RC/cm2 (for 25 RC layers on each side of the glass substrate) can be estimated. In wet PEMs, a water/RC molar ratio of about 7 × 104 can be evaluated from the area of the water absorption band centered at 1930, following the procedure described in ref 21. Because, as described above, further hydration of the RC multilayer causes only a limited (5%) increase in the fraction of slow charge recombination, it appears that such a water/RC ratio is sufficient to retrieve the function of QB as a secondary electron acceptor in most of the RC population. The sensitivity of QA- to QB electron transfer to the hydration level of RC multilayers is reminiscent of a similar behavior, previously observed by us in RC embedded in different nonliquid matrices. In particular, by studying the kinetics of charge recombination in trehalose coated RCs as a function of the content of residual water, we have shown that a progressive dehydration of the matrix blocks QA- to QB electron transfer in a progressively increasing fraction of the RC population, leading to complete inhibition in relatively wet matrices.8,34 A further reduction in the amount of residual water affects the kinetics of P+QA- recombination, which, in extensively dried matrices, is strongly accelerated and becomes distributed over a spectrum of rate constants.21 These effects have been taken to indicate a strong coupling between dynamics of the RC protein and of the trehalose matrix, the latter being progressively hampered upon dehydration.8 More specifically, the severe hardening of the trehalose matrix occurring at extremely low hydration levels suppresses the transition from a dark- to a lightadapted conformation, which stabilizes the primary charge separated state P+QA-.8,21,37 As a consequence, P+QA- charge recombination occurs from an unstabilized, unrelaxed RC conformation, with a rate (kAP) larger than that measured in solution. In more soft (wet) matrices, stabilization of the P+QAstate can take place, but the trehalose matrix heavily hinders the conformational changes coupled to QA- to QB electron transfer.8,29,34 This process has been shown indeed to be

J. Phys. Chem. B, Vol. 111, No. 12, 2007 3309 conformationally gated.38 When the RC is incorporated in a weakly interacting matrix, i.e., in a PVA film, kAP is affected to a much lower extent, even when the content of residual water is below the detection limit of NIR spectroscopy.29 Also in PVA, dehydration primarily affects the functionality of the secondary quinone acceptor, but a complete inhibition of QA- to QB electron transfer over the whole RC population is never attained, even under extreme dehydration.29 Taken together, these observations can help in interpreting the behavior observed in RC multilayers. In PEM, the rate constant kf of the fast phase is unchanged with respect to RC in solution, suggesting that in PDDA-RC multilayers the relaxation of the RC protein from the dark- to the light-adapted conformation is still faster than P+QA- recombination. At variance, incorporation in PEMs and their dehydration appears to block QA- to QB electron transfer in a significant fraction of the RC population. The effect is qualitatively similar to that observed in trehalose matrices and PVA films but it is much weaker in RC-multilayers. Even after extensive drying (dry PDDA-RC multilayers), more than 60% of RC displays P+QB- recombination, and in “wet” trehalose matrices, only P+QA- recombination is observed. Full hindering of interquinone electron transfer appears to require, in addition to dehydration, a structural and dynamical protein-matrix coupling, which is peculiar of trehalose glasses but not of PVA films and PDDA-RC PEM. In summary, analysis of charge recombination in PDDA-RC multilayers indicates that hydration has a non-negligible, but minor effect on the intraprotein electron transfer within RCs. From a practical point of view, layer-by-layer assemblies have an additional advantage over the other non-liquid systems discussed above: they can be dipped repeatedly in water without destroying the matrix. This allowed us to test extensively the reversibility of hydration/dehydration processes. Hydration/ dehydration effects can be observed reproducibly over a period of several months on the same multilayer. In view of their stability, RC-multilayers provide a powerful tool in studying the interaction of herbicides with the QB site. 3.3. Effect of PSII-Herbicides. A wide variety of agents are found to be specific inhibitors of the electron transfer between the quinone electron acceptors of the photosystem II complex (PSII) in higher plants and of bacterial RCs.16 Due to their use for the control of weeds, these agents are termed PSII-herbicides. PSII-herbicides bind, in ViVo, to the D1 protein of the PSII complex and displace the secondary electron acceptor (a plastoquinone in higher plants). Due to the structural and functional analogies of the quinone acceptor complex in bacterial RCs and plant PSII, this class of herbicides includes components that act efficiently as competitive inhibitors also at the QB site of bacterial RCs. The binding of herbicides to bacterial RCs has been studied in solution by analyzing the kinetics of charge recombination following a short flash of light.39 In fact, the relative amplitude of the fast phase of recombination kinetics can be taken as a measure of the fraction of the RC population in which the herbicide is bound at the QB site and inhibits electron transfer to the secondary acceptor.40 As shown in Figure 4, dipping of PDDA-RC PEM into a concentrated (0.8 mM) solution of o-phenantroline, a classical competitive inhibitor of the QB site in bacterial RCs, has a dramatic effect on the charge recombination kinetics: the flashgenerated P+ decays in a few hundreds of milliseconds instead of several seconds. In PDDA-RC multilayers exposed to o-phenantroline, most of the P+ decay occurs exponentially with a rate constant that essentially coincides with kAP found in solution, indicating that this herbicide successfully competes

3310 J. Phys. Chem. B, Vol. 111, No. 12, 2007

Figure 4. Charge recombination kinetics measured in dry PDDA-RC PEM after immersion in water (control) and 0.8 mM o-phenanthroline solution.

for the QB site of the protein in PDDA-RC PEM as well. The decrease in the maximal absorbance change reached immediately after the flash indicates that exposure to o-phenantroline at very high concentration displaces also the quinone at the QA site (as reported for RC in liposomes).41 Quite surprisingly, the inhibitory effects described above are caused by exposure to the herbicide for a relatively short time: a few seconds of dipping in the herbicide solution are sufficient to affect the kinetics as illustrated in Figure 4. The acquisition of charge recombination kinetics characterized by a good signal-to-noise ratio requires the averaging of several transients separated by a 1 min time interval during which the sample is dark-adapted to allow full relaxation of photogenerated species. Because of these constraints, the kinetics of herbicide binding and release were probed on dry PDDA-RC films. The procedure was the following. PDDA-RC PEM previously stored in pure water was dipped in 6 mL of herbicide solution for a given time. Then, the specimen was removed from the solution, placed vertically on absorbing paper and dried for 10 min in air. Finally, the charge recombination kinetics was recorded. The same sample was then dipped again in the herbicide solution for a longer time, and the procedure was repeated until the relative amplitude of the fast phase in the measured P+ decay reached a maximum, constant value. To study the kinetics of herbicide de-binding, the same fully inhibited sample was rinsed with pure water. The procedure was analogous to that used for the binding kinetics; the specimen was dipped in 100 mL water for different times and the charge recombination kinetics were measured subsequently in the dry PEM. The results obtained in a consecutive binding and de-binding experiment are shown in Figure 5, where the fraction of the fast phase of the P+ decay is reported as a function of the immersion time. Both processes of binding and release are accomplished in a few seconds and are reasonably described by first-order kinetics. Binding appears to be faster, with a characteristic time of 2.5 ( 0.5 s, as compared to a time of 13 ( 2 s for de-binding. As to the decrease in the absorbance change measured immediately after the flash (∆Abs(0)), we note that it is also fully reversible by rinsing the PEM with water; difficulties in the exact repositioning of the specimen preclude a detailed kinetic analysis of the decrease and recovery observed in ∆Abs(0); however, the o-phenantroline-induced drop in ∆Abs(0) reaches a steady state within 100 s of incubation in the herbicide solution, and 50 s of rinsing with water are enough to revert this effect.

Mallardi et al.

Figure 5. Inhibitor binding and release kinetics (see text for details). (A) Relative fast phase amplitude of charge recombination kinetics measured on PDDA-RC PEM as a function of the incubation time in a 0.8 mM o-phenanthroline solution. (B) Relative fast phase amplitude of charge recombination kinetics for fully inhibited PDDA-RC PEM (last point of panel A) after rinsing with water for different times. Lines are fitting according to first-order kinetics.

Following the determination of binding and de-binding kinetics, we have built an equilibrium binding isotherm by measuring the charge recombination kinetics after a 5 min equilibration of the RC-multilayer with o-phenantroline solutions at increasing herbicide concentration. In these assays, rather than using dried RC-multilayers, measurements were performed directly on PDDA-RC PEM immersed into the o-phenantroline solution because 5 min of incubation are a time long enough to reach the equilibrium. This procedure minimizes the scattering artifacts and maximizes the effect of the herbicide because in fully hydrated PDDA-RC PEM the QA-QB f QAQB- electron transfer works efficiently over a larger RC population (see previous section). The data, shown in Figure 6, indicate that the fraction of RCs in which the electron transfer to QB is inhibited increases with the herbicide concentration reaching a plateau for concentrations higher than 400 µM. Under these conditions, the relative amplitude of the fast phase exceeds 80% of the total, showing that the inhibitor blocks QA- to QB electron transfer in most of the RCs. This indicates that o-phenantroline has access to a major fraction of the RC population, including those deeply embedded in the multilayer, and implies that the inhibitor can diffuse freely through the multilayer. It is not clear, a priori, whether the interaction of a ligand with proteins assembled to form a film should be treated as a classical binding process or as the adsorption to a surface. In any case, both processes are expected to obey formally the same equation

X)

K[I] 1 + K[I]

(4)

where [I] is the concentration of the ligand/adsorbate I. X represents the fraction of protein bound to I, and K is the binding constant (the reciprocal of the dissociation constant). In the case of adsorption to a surface, eq 4 identifies with the Langmuir’s isotherm, where X denotes the fraction of adsorption sites occupied and K is the Langmuir constant (the ratio between the adsorption and desorption rate constants). Whichever of the processes is considered, in our case, X can be taken as the fraction of the RC population undergoing P+QA- recombination (i.e., the relative amplitude Af of the fast phase in the decay of the absorbance change after a light pulse). From the data of Figure 6, it is clear that in the absence of o-phenantroline a fast

PDDA-RC multilayers

J. Phys. Chem. B, Vol. 111, No. 12, 2007 3311 SCHEME 1: A Simple Kinetic Scheme Used to Quantitatively Describe Steady-State Behavior under Continuous Illumination

Figure 6. Relative amplitude of the fast phase of charge recombination kinetics measured on a PDDA-RC PEM as a function of o-phenanthroline concentration. The data (stars) were fitted to eq 5 (continuous line) under the assumption that the herbicide concentration in solution is unaffected by the binding process (o-phenanthroline is always in large excess compared to RC). The best fitting binding constant is K ) 6.9 ( 0.7 mM-1.

phase of non-negligible relative amplitude is already present in the recombination kinetics and that, even at high herbicide loading, Af does not reach unity (the latter feature is observed also in solution and is generally attributed to the competition between herbicide and quinone).39 Accordingly, we have modified eq 4 in the form below by adding two adjustable parameters representing the values X assumes when [I] ) 0 and [I] ) ∞ (Xmin and Xmax, respectively)

X ) Xmin + (Xmax - Xmin)

K[I] 1 + K[I]

(5)

As illustrated in Figure 6, eq 5 nicely fits the experimental data for K ) 6.9 ( 0.7 mM-1. This value corresponds to a dissociation constant (145 ( 15 µM) of the same order of magnitude of the one reported for RCs in solution.39 Extensive rinsing with water allows the recovery of the QB activity so that the same specimen can be used repeatedly; the effect of the inhibitor is reproducible within the uncertainty associated with fitting of the charge recombination kinetics. These features are promising for the use of PDDA-RC PEM as a biomediator in herbicide bioassays. 3.4. An Optical Biosensor. Herbicides are widely used on a variety of crops for the control of broad leaf weeds but can be highly toxic for human and animal health. Their wide use in agriculture has resulted often in the herbicide pollution of water and the level of herbicides allowed in drinking water is subject to regulation, at least in the industrialized countries. Different attempts have been made to introduce biological detection systems in order to overcome the high cost of time-consuming HPLC analysis. Several immunoassays have been proposed, mainly ELISA42-44 and immunosensors.45-49 However, antibodies can detect only one or few cross-reacting substances. To overcome this limit, the analytical use of PSII as a receptor protein in bioassays has been investigated.50 The practical application of herbicide biosensors based on PSII preparations

is seriously hampered by their instability (in particular when illuminated).50 At variance, the bacterial RC is remarkably stable against denaturation. Its use in optical bioassays was proposed already in 1993 by exploiting phenomenologically the response to herbicides of the extent of primary donor P photoxidized under continuous illumination.51 According to eqs 1 and 2, the steady-state attained by RCs under continuous light is different in the presence and absence of an inhibitor of QA--to-QB electron transfer. Intuitively, the level of photoxidation reached by P under continuous photoexcitation will be determined by the competition between charge separating processes and charge recombination reactions. The latter are faster when occurring from the P+QA- state (i.e., in the presence of inhibitor) than from the P+QB- state (in the uninhibited RC); a progressively lower level of P oxidation is therefore expected in steady state at increasing herbicide concentrations. Steady-state behavior under continuous illumination can be described quantitatively with the aid of Scheme 1: Acceptors prior to QA are incorporated into the light-dependent rate constant kL, which is actually the product of an absorption cross-section and photon flux. Equilibration of the electron between QA and QB has been assumed much faster than the other considered processes;40 under this condition, recombination from the P+(QAQB)- state in the uninhibited complex can be described by a single rate constant λ ) kAP/(1 + LAB), where LAB ) kAB/kBA is the equilibrium constant for interquinone electron transfer (see eq 1 and ref 40). The steady-state solution of Scheme 1 (see Appendix) yields the following relation for the fraction, FO, of photoxidized RC, as a function of the inhibitor concentration [I]:

FO )

[P+(QAQB)-] + [P+QA- ] [N]

) kL(kAP + λK[I])

kAP(λ + kL) + λ(kAP + kL)K[I]

(6)

where [N] is the total concentration of RCs and K ) kI/k-I is the binding equilibrium for the herbicide. From eq 6, by considering that kAP > λ (kAP ≈ 10 s-1, λ ≈ 1 s-1),40 it is clearly seen that the photoxidation level of P varies from a maximum, FO,max ) kL/(kL + λ), for I ) 0, to a minimum value, FO,min ) kL/(kL + kAP), for I f ∞. Actually, as these relationships show under extreme conditions, the steady-state level of P+ is simply determined by the competition between charge separation (rate constant kL) and recombination from P+QA- (rate constant kAP) and P+(QAQB)- (rate constant λ), respectively. It appears therefore that FO can be taken as a simple measure of herbicide inhibition. We have tested the behavior predicted by eq 6 in PDDA-RC PEM by measuring the time course of P photoxidation from the absorbance bleaching induced at 870 nm by continuous illumination. The measuring beam at 870 nm was sufficiently

3312 J. Phys. Chem. B, Vol. 111, No. 12, 2007

Mallardi et al.

intense to act also as an excitation (actinic) light source (see Section 2). Figure 7A shows the absorbance changes recorded over 5 s for a PDDA-RC PEM dipped in solutions of different o-phenantroline concentration. The steady maximal extent of this bleaching reached after prolonged illumination decreases upon increasing the inhibitor concentration in equilibrium with the multilayer, as expected from eq 6 (see Figure 7B). The fractional extent of P photoxidation (F0) reached under continuous illumination can be compared with the fraction X of RC bound to the inhibitor, evaluated from the relative amplitude of fast P+QA- recombination in single-turnover experiments as described in the previous section. Starting from eq 6, a useful index of inhibition FI, based on the measured bleaching data, can be defined as

FI )

FO,max - FO aK[I] ) FO,max - FO,min 1 + aK[I]

(7)

where a ) λ(kL + kAP)/kAP(kL + λ) and 0 e FI e 1. Equation 7 has the form of a binding isotherm (see eq 4), where the binding constant is replaced by (aK). Being kAP > λ (see above), it follows that a e 1; i.e., FI will titrate with an apparent binding constant lower or equal to the real one. According to eq 5, the corresponding fractional inhibition based on X is F′I ) (X Xmin )/(Xmax - Xmin) ) K[I](1 + K[I])-1 and depends only on the true binding constant K. The inhibition index FI obtained from continuous light measurements and the fractional inhibition F′I of the QB site evaluated from single-turnover recombination kinetics are compared in Figure 8. It is clear that the two approaches, probing on different time scales and illumination regimes the interaction between the herbicide and RC, yield very close binding isotherms. In fact, the dependence of FI upon [I] fits eq 7 for (aK) ) 6 ( 1 mM-1; i.e., a ) 0.9 ( 0.2 ≈ 1. Because a is so close to unity, under our experimental conditions, kL , λ ∼ 1 s-1. This is somehow expected because photobleaching reaches a steady level only after a long illumination (several seconds, see Figure 7A). A remarkable implication of the results shown in Figure 8 is that photobleaching measurements performed on PDDA-RC with a low enough intensity of the measuring beam successfully probe the inhibition of QB site within the RC protein. The main advantage of this approach is that it requires a simple instrumentation (single beam at fixed wavelength and very low time resolution) and it does not require any deconvolution of kinetic traces (only the steady-state value FO is required). In agreement with our observations, previous studies had shown that titration-like curves could be obtained in solutions of RCs by plotting the extent of bleaching against the herbicide concentration.51,52 We note, however, that no attempt was made in these studies to quantitatively account for the obtained titrations in terms of the binding constant and of the kinetic parameters of the system. We show here that, under appropriate conditions, a true binding constant can be easily evaluated from photobleaching induced by continuous illumination. Several limitations have precluded a practical use of bacterial RC solutions in herbicide bioassays. A first drawback is the low intrinsic affinity of most herbicides for the RCs from purple bacteria, which limits considerably the sensitivity of any RCbased bioassay. Other limitations are related to the cost of the protein and to the repeatability and reproducibility of the bioassay test. Although the purification protocols for RC are well assessed, its purification suffers of all the problems common to membrane proteins: high cost and low yield (compared to hydrosoluble proteins). In addition, the inhibitor/

Figure 7. (A) Time course of RC photobleaching induced by continuous illumination at 870 nm and measured on the same PDDARC PEM immersed in solutions of o-phenanthroline at different concentration (indicated in the labels). (B) Values of absorbance change after 5 s of illumination versus herbicide concentration (values of absorbance change are from the traces of panel A, averaged over the time interval bracketed by the dashed lines). Closed circles refer to measurements done by dipping the specimen in solutions of increasing o-phenanthroline concentration; the specimen was then extensively washed with water and selected solutions were assayed (open circles). Fitting the data to eq 7 (dashed line) yields aK ) 6 ( 1 mM-1.

quinone competition at the QB site implies that the binding equilibrium of quinone itself has to be considered. Although the mechanisms of the herbicide/quinone competition are reasonably well understood,40 a practical strategy to control the quinone occupancy at the QB site is still lacking. Native ubiquinone is very hydrophobic, and its loading requires the use of detergents or organic solvents. On the other hand, hydrophilic quinones are relatively labile in solution. In both cases, a high controlled level of QB occupancy is not easily obtained and maintained upon storage in solution RCs. For analytical purposes, this means that a relatively large amount of RCs (reconstituted at the QB site) must be used to determine the calibration curve (bleaching versus herbicide concentration) and to test in close sequence the unknown samples. Most importantly, the protein in solution cannot of course be

PDDA-RC multilayers

J. Phys. Chem. B, Vol. 111, No. 12, 2007 3313 concentration is well described by eq 7, yielding a binding constant aK ) 0.39 ( 0.06 mM-1. It should be stressed that the two isotherms in Figure 9 have been obtained using the same PDDA-RC PEM (the time elapsed between the assays on different herbicide was 1 month). Moreover, experiments with a given herbicide were performed during several days and some measurements were carried out after extensive washing (see caption in Figure 9) thus simulating the assay of an unknown sample. The consistent, reproducible results shown in Figure 9 denote that this approach allows us to assay many samples by using a single PDDA-RC over an extended period of time. The use of a small solid support within a flow cell might further reduce the amount of RC and improve the reproducibility of the assay (many fluctuations in the measured bleaching are due to difficulties in positioning the PDDA-RC plate).

Figure 8. A comparison between the fractional extent of QB inhibition determined from single-turnover flash (stars) and continuous illumination (closed circles) experiments. See text for details.

Acknowledgment. The financial support of MIUR of Italy is acknowledged (grants COFIN-PRIN/2003 prot. 2003022158 and COFIN-PRIN/2005 prot. 2005027011). G. P. and M. G. were partially supported by the Consorzio Interuniversitario per lo sviluppo dei Sistemi a Grande Interfase (CSGI-Firenze). M. G. wishes to thank Mr. G. Gervasoni for the technical support in building the sample compartment of instrument B. Appendix Steady State Attained by the RC under Continuous Illumination in the Presence of an Inhibitor I of Electron Transfer to QB. The rate equations describing the electrontransfer kinetics and binding equilibrium summarized in Scheme 1 are given by

( ) ( ( )

[PQAQB] + d [P (QAQB) ] ) dt [PQAI] [P+QA-I] -(kL + kI[I]) kL kI[I] 0

Figure 9. Titration-like curves (extent of photobleaching vs herbicide concentration) obtained by continuous illumination of the same PDDARC PEM: Circles, o-phenanthroline; diamonds, terbutryn. In both cases open, closed, and shadowed symbols refer to different runs performed after extensive rinsing with water. Continuous curves are fit to the data according to eq 7 (o-phenanthroline, aK ) 6 ( 1 mM-1; terbutryn, aK ) 0.39 ( 0.06 mM-1).

reemployed in subsequent assays. All these aspects make quite unpractical the use of RC solutions for large scale herbicide screenings. The incorporation of RC into layer-by-layer assemblies overcomes many of these weaknesses. Because the herbicides can be efficiently washed out, the PDDA-RC multilayers can be reused for several samples. This property, coupled with the exceptional stability of the RC when embedded into layer-bylayer assemblies (up to 7 months when stored at 4 °C), make possible to perform complete runs (calibration curve plus unknowns) with a small amount of protein (1-2 nmol with our plates). As an example, in Figure 9, we show titration curves obtained with the same specimen for two different herbicides (o-phenantroline and terbutryne). Also in the case of terbutryne, the dependence of steady-state photobleaching upon the inhibitor

λ -λ 0 0

k-I 0 -(kL + k-I) kL

)( )

[PQAQB] 0 [P+ (QAQB)-] 0 ‚ kAP [PQAI] -kAP [P+ QA-I] (A1)

The steady-state solutions of eq A1 are

[PQAQB] [P+(QAQB)-] [PQAI] [P+QA-I]

)

tf∞

( )

λkAPk-I kAPk-IkL N ‚ kAPk-I(λ + kL) + λkI[I](kAP + kL) λkAPkI[I] λkLkI[I]

(A2)

where N ) [PQAQB]t + [P+(QAQB)-]t + [PQAI]t + [P+QA-I]t is the total concentration of reaction centers. References and Notes (1) Gill, I.; Ballesteros, A. Trends Biotechnol. 2000, 18, 282. Gill, I.; Ballesteros, A. Trends Biotechnol. 2000, 18, 469. (2) Lebedev, N.; Trammell, S. A.; Spano, A.; Lukashev, E.; Griva, I.; Schnur, J. J. Am. Chem. Soc. 2006, 128, 12044. Trammell, S. A.; Pano, A.; Price, R.; Lebedev, N. Biosens. Bioelectron. 2006, 21, 1023. (3) Hampp, N. Chem. ReV. 2000, 100, 1755.

3314 J. Phys. Chem. B, Vol. 111, No. 12, 2007 (4) Stuart, J. A.; Marcy, D. L.; Wise, K. J.; Birge, R. R. Synth. Met. 2002, 127, 3. (5) Holden, M. A.; Bayley, H. J. Am. Chem. Soc. 2005, 127, 6502. (6) O’Neill, H.; Greenbaum, E. Chem. Mater. 2005, 17, 2654. (7) Das, R.; Kiley, P. J.; Segal, M.; Norville, J.; Yu, A. A.; Wang, L. Y.; Trammell, S. A.; Reddick, L. E.; Kumar, R.; Stellacci, F.; Lebedev, N.; Schnur, J.; Bruce, B. D.; Zhang, S. G.; Baldo, M. Nano Lett. 2004, 4, 1079. Benson, D. E.; Conrad, D. W.; de Lorimer, R. M.; Trammell, S. A.; Hellinga, H. W. Science 2001, 293, 1641. (8) Cordone, L.; Cottone, G.; Giuffrida, S.; Palazzo, G.; Venturoli, G.; Viappiani, C. Biochim. Biophys. Acta 2005, 1749, 252. (9) Decher, G., Schlenoff, J. B., Eds. Multilayer Thin Films: Sequential Assembly of Nanocomposite Materials; Wiley-VCH: Weinheim, Germany, 2003. (10) Schonhoff, M. Curr. Opin. Colloid Interface Sci. 2003, 8, 86-95. Schonhoff, M. J. Phys.: Condens. Matter 2003, 15, R1781. (11) Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. J. Am. Chem. Soc. 1995. 117, 6117. (12) Kong, J.; Lu, Z.; Lvov, Y. M.; Desamero, R. Z. B.; Frank, H. A.; Rusling, J. F. J. Am. Chem. Soc. 1998, 120, 7371. (13) Zhao, J.; Liu, B.; Zou, Y.; Xu, C. Kong, J Electrochim. Acta 2002, 47, 2013. (14) Feher, G.; Allen, J. P.; Okamura, M. Y.; Rees, D. C. Nature 1989, 33, 111 and references therein. (15) Ort, D. R.; Melandri, B. A. In Photosynthesis. Energy ConVersion by Plants and Bacteria; Govindjee, Ed.; Academic Press: New York, 1982; p 537. (16) Draber, W.; Tietjen, K.; Kluth, J. F.; Trebst, A. Angew. Chem., Int. Ed. Engl. 1991, 3, 1621. (17) Deisenhofer, J.; Epp, O.; Miki, K.; Huber, R.; Michel, H. Nature 1985, 318, 618. Allen, J. P.; Feher, G.; Yeates, T. O.; Komiya, H.; Rees, D. C. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 5730. (18) Gray, K. A.; Wachtveitl, J.; Breton, J.; Oestherhelt, D. EMBO J. 1990, 9, 2061. (19) Mallardi, A.; Palazzo, G.; Venturoli, G. J. Phys. Chem. B 1997, 101, 7850. (20) Giustini, M.; Castelli, F.; Husu, I.; Giomini, M.; Mallardi, A.; Palazzo, G. J. Phys. Chem. B 2005, 109, 21187. (21) Palazzo, G.; Mallardi, A.; Hochkoeppler, A.; Cordone, L.; Venturoli, G. Biophys. J. 2002, 82, 558. (22) Piazza, R.; Pierno, M.; Vignati, E.; Venturoli, G.; Francia, F.; Mallardi, A.; Palazzo, G. Phys. ReV. Lett. 2003, 90, 208101. (23) Palazzo, G.; Mallardi, A.; Francia, F.; Dezi, M.; Venturoli, G.; Pierno, M.; Vignati, E.; Piazza, R. Phys. Chem. Chem. Phys. 2004, 6, 14391445. (24) Steffen, M. A.; Lao, K.; Boxer, S. G. Science 1994, 264, 810. (25) Cherepy, N. J.; Shreve, A. P.; Moore, L. J.; Boxer, S. G.; Mathies, R. A. J. Phys. Chem. B 1997, 101, 3250.

Mallardi et al. (26) Mu¨h, F.; Schulz, C.; Schlodder, E.; Jones, M. R.; Rautter, J.; Kuhn, M.; Lubitz, W. Photosynth. Res. 1998, 55, 199. (27) Mu¨h, F.; Rautter, J.; Lubitz, W. Ber. Bunsen-Ges. 1996, 100, 1974. (28) Mu¨h, F.; Rautter, J.; Lubitz, W. Biochemistry 1997, 36, 4155. (29) Francia, F.; Giachini, L.; Palazzo, G.; Mallardi, A.; Boscherini, F.; Venturoli, G. Bioelectrochemistry 2004, 63, 73. (30) Clayton, R. K. Biochim. Biophys. Acta. 1978, 504, 255. (31) Gast, P.; Hemelrijck, P. W.; van Gorkom, J.; Hoff, A. J. Eur. J. Biochem. 1996, 239, 805. (32) Kleinfeld, D.; Okamura, M. Y.; Feher, G. Biochim. Biophys. Acta 1984, 766, 126-140. (33) Feher, G.; Okamura, M. Y. In The Photosynthetic Bacteria; Clayton, R. K., Sistrom, W. R., Eds.; Plenum Press: New York, 1978; p 349. (34) Francia, F.; Palazzo, G.; Mallardi, A.; Cordone, L.; Venturoli, G. Biophys. J. 2003, 85, 2760. (35) Kleinfeld, D.; Okamura, M. Y.; Feher, G. Biochemistry 1984, 23, 5780-5786. (36) Francia, F.; Dezi, M.; Rebecchi, A.; Mallardi, A.; Palazzo, G.; Melandri, B. A.; Venturoli, G. Biochemistry 2004, 43, 14199. (37) Francia, F.; Palazzo, G.; Mallardi, A.; Cordone, L.; Venturoli, G. Biochim. Biophys. Acta Bioenerg. 2004, 1658, 50. (38) Graige, M. S.; Feher, G.; Okamura, M. Y. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 11679. (39) Stein, R. R.; Castellvi, A. L.; Bogacz, J. P.; Wraight, C. A. J. Cell. Biochem. 1984, 24, 243-259. (40) Shinkarev, V. P.; Wraight, C. A. In The Photosynthetic Reaction Center; Deisenhofer, J., Norris, J. R., Eds.; Academic Press: New York, 1993; Vol. 1, p 193 and references therein. (41) Wright, C. A. Isr. J. Chem. 1981, 21, 348. (42) Schneider, P.; Goodrow, M. H.; Gee, S. J.; Hammock, B. D. J. Agric. Food Chem. 1994, 42, 413. (43) Piletskaya, E. V.; El’skaya, S. A.; Sozinov, A. A.; Marty, J. L.; Roulillon, R. Anal. Chim. Acta 1999, 398, 49. (44) Banks, K. E.; Hernandez, S. Talanta 2003, 61, 257. (45) Harris, R. D.; Luff, B. J.; Wilkinson, J. S.; Piehler, J.; Brecht, A.; Gauglitz, G.; Abuknesha, R. A. Biosens. Bioelectron. 1999, 14, 377. (46) Brecht, A.; Piehler, J.; Lang, G.; Gauglitz, G. Anal. Chim. Acta 1995, 311, 289. (47) Bier, F. F.; Jockers, R.; Schmid, R. D. Analyst 1994, 119, 437. (48) Skl’adal, P. Biosens. Bioelectron. 1999, 14, 257. (49) Barzen, C.; Brecht, A.; Gauglitz, G. Biosens. Bioelectron. 2002, 17, 289. (50) Giardi, M. T.; Pace, E. Trends Biotechnol. 2005, 23, 257 and references therein. (51) Jockers, R.; Bier, F. F.; Schmid, R. D.; Wachtveitl, J.; Oesterhelt, D. Anal. Chim. Acta 1993, 274, 185. (52) Baldini, F.; Domenici, C.; Giannetti, A.; Masci, D.; Mencaglia, A. Sens. Actuators 2003, B90, 198.