Excited-State Intermolecular Proton Transfer of Firefly Luciferin III

Dec 3, 2010 - Steady-state and time-resolved techniques were employed to study the excited-state proton transfer (ESPT) from d-luciferin, the natural ...
0 downloads 0 Views 2MB Size
Download PDF
J. Phys. Chem. A 2010, 114, 13337–13346

13337

Excited-State Intermolecular Proton Transfer of Firefly Luciferin III. Proton Transfer to a Mild Base Itay Presiado, Yuval Erez, and Dan Huppert* Raymond and BeVerly Sackler Faculty of Exact Sciences, School of Chemistry, Tel AViV UniVersity, Tel AViV 69978, Israel ReceiVed: August 5, 2010; ReVised Manuscript ReceiVed: NoVember 4, 2010

Steady-state and time-resolved techniques were employed to study the excited-state proton transfer (ESPT) from D-luciferin, the natural substrate of the firefly luciferase, to the mild acetate base in aqueous solutions. We found that in 1 M aqueous solutions of acetate or higher, a proton transfer (PT) process to the acetate takes place within 30 ps in both H2O and D2O solutions. The time-resolved emission signal is composed of three components. We found that the short-time component decay time is 300 and 600 fs in H2O and D2O, respectively. This component is attributed either to a PT process via the shortest water bridged complex available, ROH · · H2O · · Ac-, or to PT taking place within a contact ion pair. The second time component of 2000 and 3000 fs for H2O and D2O, respectively, is attributed to ROH* acetate complex, whose proton wire is longer by one water molecule. The decay rate of the third, long-time component is proportional to the acetate concentration. We attribute it to the diffusion-assisted reaction as well as to PT process to the solvent. SCHEME 1

Introduction D-Luciferin (shown in Scheme 1), the natural substrate of the firefly luciferase, first isolated and purified by McElroy and co-workers,1,2 has been the subject of much research. Although D-luciferin does not appear to be the emitting species of the firefly bioluminescence, oxyluciferin, the chromophore responsible for the bioluminescence emission, shows UV-vis spectroscopic characteristics similar to those of D-luciferin.3 Courtship in many species of fireflies relies on flash communication through precisely timed rapid flashes of light. The duration of a single flash is ∼100 ms and is composed of a number of shorter pulses.4 Neural activity stimulates release of octopamine, a neurotransmitter that subsequently triggers light emission from the firefly lantern. In the firefly, oxidation of D-luciferin (shown in Scheme 1) to oxyluciferin is catalyzed by luciferase in the presence of ATP and magnesium ions. The following deactivation of oxyluciferin from its electronically excited state is accompanied by light emission. The on-off switch of this reaction is oxygen.5 The bioluminescence produced by the firefly is purely in the visible, considered very efficient, and accompanied by very low thermal radiation. Although all species of fireflies use D-luciferin, the color of light emitted by different species varies greatly in the range of 552 nm (green-yellow) to 582 nm (orange). These color differences are attributed to changes in pH and Zn2+ ionic concentration, as well as enzyme differences between the species.6 In the past decades, intermolecular excited-state-protontransfer (ESPT) to a solvent or to a base via a solvent-solute complex in a liquid solution, and more recently in ice, has been extensively studied.7-19 For many years, we conducted thorough research on the reversible and irreversible photoprotolytic cycle of a photoacid. We used a PT model that accounts for the diffusion-influenced geminate recombination of the transferred proton with the deprotonated form of the photoacid.13,20,21

* Corresponding author. Phone: 972-3-6407012. Fax: 972-3-6407491. E-mail: [email protected].

SCHEME 2

In a previous study,22 we measured the steady-state (timeintegrated) emission, excitation, and absorption spectra, as well as the time-resolved emission properties of excited luciferin in water. D-Luciferin has dual emission bands when excited from its neutral form (the protonated form, NROH*). The bands correspond to emission from the NROH* and from the deprotonated species NRO-* of D-luciferin. The decay of the timeresolved emission of the NROH* in neat water is nonexponential and fast, that is, about 25 ps at room temperature. The timeresolved emission signal of the NRO-* is complex because the proton geminate recombination process is irreversible and thus enhances the quenching of the NRO- fluorescence. We interpret these observations as arising from an ESPT to the solvent process. Scheme 2 describes the photoprotolytic cycle processes that D-luciferin undergoes after excitation of the ground-state protonated form. For more details, see the Supporting Information. In a more recent work,23 we focused our attention on the systematic study of the photoprotolytic process that D-luciferin undergoes in water-methanol mixtures with varying molar fraction of methanol, ranging from 0 to 1. We found that in the range of 0 < χMeOH < 0.8, the PT rate decreases exponentially with an increase of the methanol concentration. At high methanol concentrations, that is, χMeOH > 0.8, there is a steep drop in the PT rate. In neat methanol, we estimate that the PT rate is roughly 3 orders of magnitude slower than in water.

10.1021/jp107360d  2010 American Chemical Society Published on Web 12/03/2010

13338

J. Phys. Chem. A, Vol. 114, No. 51, 2010

In the current work, we studied the reaction of acetate ions (Ac-) in H2O and D2O solutions with the excited NROH form of D-luciferin. The protonated form of a photoacid generally reacts with mild bases such as Ac-, H2PO4-, HCO3-, and many others. This PT reaction produces the conjugated acid of these bases. In recent studies published jointly by Pines and Nibbering10,11,16 and later by Bakker and co-workers,15 this type of reaction was examined using a UV-pump mid-IR probe femtosecond laser-based optical setup. The 8-hydroxy-1,3,6pyrenetrisulfonate (HPTS) photoacid, which has a large absorption cross-section around 400 nm, is often used for these experiments. Both groups found that at acetate concentrations higher than 1 M a sizable share of the ROH population of HPTS reacted with Ac- ions within a few tens of picoseconds. The Smoluchowski model describes diffusion-assisted chemical reactions in solution, where one of the reactants is in large excess. The assumption is that the reaction takes place only when the reactants are in a contact distance. In spherical symmetry terms, we can discuss a sphere surface of radius a. Usually, this radius is taken to be 5-7 Å in aqueous solutions. For a mildly viscous solvent, such as water, with a viscosity of ∼1 cP, the diffusion constant of ions of the size of the acetate anion is ∼5 × 10-6 cm2/s. The diffusion-controlled rate constant limits the long-time rate of the reaction, when the reactants need first to diffuse toward each other to form the encounter pair prior to chemical reaction. For 1 M solution of Ac-, the pseudo first-order reaction rate constant is ∼5 × 109 s-1. Therefore, the reaction of 1 M of Ac- with HPTS should take place at times longer than 20 ps at an average rate of 5 × 109 s-1. However, it was found that a high percentage of the molecules in the ROH* form reacts with Ac-* at times shorter than 10 ps. Pines and Nibbering found that at very high concentrations of acetate a substantial part of the ROH population reacted within an even shorter time than the time resolution of their experimental setup, that is, less than 150 fs. The explanation they provided for this observation is that an ultrafast PT reaction rate occurs between the ROH and Ac- in contact ion pairs, ROH* · · · Ac-, which pre-existed the short pulse excitation, τpulse = 150 fs. They also found an intermediate rate of 1/6 ps-1 that they attributed to a HPTS acetate complex, in which one water molecule bridges between the two reactants. Bakker’s data on this system are comparable with those of Pines and Nibberinig, but their interpretations are different. Both interpretations are briefly explained in the Discussion. In the current study, we found that the NROH* of D-luciferin efficiently reacts with Ac- ions in aqueous solutions. We found that D-luciferin and HPTS exhibit similar behavior, in that a large portion of the NROH* population of the former reacts at short times faster than predicted by the Smoluchowski model. We found three time components in the complex decay profile of the excited NROH in 1 M aqueous solution of sodium acetate or higher. We attribute the short-time component of ∼300 and 600 fs for H2O and D2O, respectively, to the PT reaction of a contact ion pair or to a water-bridged complex. We attribute the intermediate time component of 2000 and 3000 fs for H2O and D2O, respectively, to a complex, where the bridge is longer by one water molecule. The long-time component we attribute to the diffusion-assisted reaction, in which the proton donor (the NROH*) and the acceptor (Ac-) diffuse to produce either a contact pair or a water bridged complex. According to the Smoluchowski model, the time-dependent reaction rate constant, k(t), depends on both the intrinsic reaction rate and the diffusioncontrolled rate.

Presiado et al.

Figure 1. Steady-state emission spectra of D-luciferin in sodium acetate solutions. Samples were excited at 340 nm, the absorption maximum of the protonated form.

Experimental Section Fluorescence up-conversion technique was employed in this study to measure the time-resolved emission of D-luciferin. The laser used for the fluorescence up-conversion was a cavity dumped Ti:Sapphire femtosecond laser, Mira, Coherent, which provides short, 150 fs, pulses at around 800 nm. The cavity dumper operated with a relatively low repetition rate of 800 kHz. The second harmonic of the laser, operating over spectral ranges of 380-420 nm, was used to excite the D-luciferin in the liquid samples. The up-conversion system (FOG-100, CDP, Russia) operated at 800 kHz. The samples were excited by pulses of ∼8 mW on average at the SHG frequency. The time-response of the upconversion system is evaluated by measuring the relatively strong Raman stokes line of water shifted by 3600 cm-1. It was found that the full-width at half-maximum (fwhm) of the signal is 280 fs. Samples were placed in a rotating optical cell to avoid degradation. (4S)-2-(6-Hydroxy-benzothiazole-2-yl)-4,5-dihydrothiazole4-carboxylic acid (D-luciferin) 99.5% was purchased from Iris Biotech (Germany). Deionized water had a resistance of >10 MΩ. Methanol and sodium acetate of HPLC grade were purchased from Fluka. All chemicals were used without further purification. Results Figure 1 shows the steady-state emission spectrum of D-luciferin samples containing sodium acetate. The samples were

excited at 340 nm, where the NROH* band is at its maximum absorption. The position of the NRO-* band depends on the NaAc concentration. This phenomenon was also found in the steady-state spectrum of HPTS. We found that at 1 M the blue band shifts by ∼200 cm-1 with respect to its position in neat water. The shift of the blue band indicates that Ac- ions form complexes or that, generally, the difference in the strength of interaction between the two molecules in the ground state and excited state shifts the emission even when the average distance is longer than 10 Å, which is about four water molecules. The weak intensity of the NROH* band, whose maximum is at 450 nm, decreases as the NaAc concentration increases. This

Proton Transfer of Firefly Luciferin III

Figure 2. Fluorescence up-conversion signals of d-luciferin, measured at several wavelengths, in sodium acetate solutions. (a) 10 mM. (b) 1 M. (c) Same as (b) but expanded to 220 ps.

phenomenon is explained by the sharp increase in the ESPT in the presence of NaAc in the solution. Figure 2a shows several fluorescence up-conversion signals of D-luciferin in a 10 mM aqueous solution of NaAc, monitored in the range of 460-540 nm. The sample was excited by a train of ∼150 fs laser pulses at 800 kHz. The second harmonic of a cavity-dumped Ti:sapphire mode-locked laser was at 383 nm. The up-conversion system had an overall instrumental response function (IRF) of ∼350 fs at full-width at half-maximum (fwhm). Each fluorescence curve was monitored at a different wavelength, designated in the plot. The NROH* band maximum is at 450 nm, whereas that of the deprotonated form is at 530 nm. The short wavelength signals, that is, λ < 510 nm, decay nonexponentially. In a previous study,22 we interpreted these signals as arising from the photoprotolytic reactions and the radiative decay (see Scheme 2). The short wavelength signals’

J. Phys. Chem. A, Vol. 114, No. 51, 2010 13339 decay is complex and consists of several components. At short wavelengths, below 460 nm, the amplitude of shortest component of the signal is large. We attribute the short component, whose ultrafast decay falls within the time response of the fluorescence up-conversion, to the strong Raman signal of H2O. In pure water, the Raman signal is even stronger because the absorption of the D-luciferin molecule dampens it. In D2O samples, the Raman signal, whose maximum in H2O is at 446 nm, is shifted to 426 nm because the O-H stretching frequency is shifted from 3600 to ∼2600 cm-1. Therefore, we consider the fluorescence signals in D2O at λ g 440 nm as free of any contribution from the Raman Stokes band. This short-time component interferes with the fast components, seen in the presence of high concentrations of NaAc. In the same work,22 we found that the NROH* form decays nonexponentially with an average decay time, 〈τ〉, of 30 and 75 ps for H2O and D2O, respectively. A short decay time component of ∼1 ps exists at short wavelengths. It is attributed to the solvation process of the excited D-luciferin molecule. A plausible second explanation to the short lifetime component is that it arises from a fast PT from the hydroxyl group to the heterocyclic nitrogen compound via a solvent bridge, containing several solvent molecules. In Figure 2a, the amplitude of the solvation component at 460 nm is 0.09 of the total normalized signal. At longer wavelengths, the amplitude decreases as the longer the measured wavelength becomes. The signals at λ g 520 nm have distinctive rise components with time constants of τ > 200 fs. The longer is the monitored wavelength, the larger the amplitude of the rise component becomes. At 560 nm, the amplitude of the rise component is about 0.65 of the normalized signal. The average rise-time is somewhat shorter than the average decay time of the protonated NROH* form. The discrepancy between the decay of the NROH and the rise of the NRO-* form may arise from a fast and effective proton quenching process: NRO- * + H+ fNROH(g). The geminate proton recombines with the heterocyclic nitrogen compound. This recombination brings the molecule back to the ground state, hence shortening the effective lifetime of its excited state. Figure 2b shows the fluorescence up-conversion signal of d-luciferin in a 1 M NaAc aqueous solution at several wavelengths. Only the first 25 ps of the luminescence of the signals were drawn to better display the short-time components, which are significant in the presence of high concentrations of acetate, a weak base. In previous studies on the effect of acetate ions on the photoprotolytic process of 8-hydroxy-1,3,6-pyrenesulfonate (HPTS), it was found that the acetate ions react with the ROH* of the photoacid in two major channels. The long-time process is described as a diffusion-assisted deprotonation reaction by the Ac- ion: ROH* + Ac- f RO- * + HAc. The diffusion constant of the bulky Ac- ion is relatively small, that is, DAc- = 5 × 10-6 cm2/s. Therefore, a sizable share of the reaction occurs on long-time scales, that is, τ > 10 ps, even in Ac- concentrations as high as several molars. Pines, Nibbering, and co-workers16 and Bakker and co-workers15 focused their studies on another kind of reaction between Acand the ROH occurring within less than 10 ps. They found that a part of the ROH population of HPTS reacts at short times. Pines et al. found that in 4 M (and higher) aqueous solutions of potassium acetate, the PT rate is ultrafast; that is, it is shorter than the time resolution of their transient IR ultrafast system response of ∼150 fs. They also observed that part of their transient IR signal decay and rise indicates that water bridged complexes are also capable of transferring protons efficiently from the hydroxyl group of the ROH* form to the acetate via

13340

J. Phys. Chem. A, Vol. 114, No. 51, 2010

Figure 3. Fluorescence up-conversion signal of the NROH* form of D-luciferin, measured at 460 nm. (a) Short time scale of 25 ps. (b) Long time-window of 100 ps.

the water bridge. The reaction rate of PT via water bridged complexes is slower than that of the contact ion pair complex and depends on the number of water molecules between the proton donor and acceptor. Figure 2b shows that the up-conversion fluorescence signal decay of the NROH* form in a 1 M NaAc aqueous solution is significantly faster than that from a 10 mM NaAc solution. Additional fast components with an amplitude of 0.6 are responsible for the fast decay rate of the NROH* signal. We attribute these fast components to PT processes, occurring between the hydroxyl group and the acetate via water-bridged complexes, NROH · · (H2O)n · · Ac-, or even a contact ion pair, ROH · · · Ac-. In the Discussion, we analyze the fluorescence up-conversion signal and determine the PT rate constant of these type of reactions. This figure also shows the NRO-* signal in a 1 M aqueous solution of NaAc. The rise time, monitored at λ g 530 nm, fits nicely to the decay of the NROH* form, which is monitored at λ e 480 nm. Figure 2c shows fluorescence up-conversion signals monitored at several wavelengths on a relatively wide time-window, expanded to ∼220 ps (∼10 times longer than the time-window shown in Figure 2b). The long-time decay of the NROH* signals (λ e 480 nm) is faster in the 1 M solution of NaAc than in the 10 mM solution. The long-time decay is attributed to the diffusion-assisted PT reaction of the Ac- with the NROH*: NROH* + Ac- f NRO- * + HAc. Figure 3a shows the fluorescence up-conversion signal of the NROH* band, monitored at 460 nm, upon excitation of aqueous solutions of NaAc in various concentrations. The time-window of 25 ps allows a better observation of the short-time components of the decay of the signal. As seen in the figure, the amplitude of the fast component is nearly invariant under change

Presiado et al.

Figure 4. Fluorescence up-conversion signals of the deprotonated NRO-* form of D-luciferin, measured at 560 nm, in several concentrations of sodium acetate. (a) Short time scale of 25 ps. (b) Long time scale of 220 ps.

of concentration above 1 M. Furthermore, the decay rate of the fast component is also invariant under change of the NaAc concentration. At 250 mM or lower, the amplitude of the shorttime components is small, less than 0.25, but at 1 M or higher, the amplitude is 0.65. The decay rate is quite similar to the higher NaAc concentrations. Figure 3b shows the same results shown in Figure 3a but on an stretched time-window of 100 ps. This allows us to clearly see that the decay rate of the long-time component is almost unaffected at the concentration range of 0-250 mM, whereas at high concentrations the decay rate is faster. Figure 4a shows the fluorescence up-conversion signal of the NRO-* band, from various aqueous solutions of NaAc, monitored at 560 nm, and displayed on a narrow time-window of 25 ps. The rise time of the signal at 250 mM or at lower concentrations depends on the NaAc concentration. The higher the concentration, the faster the rise time becomes. At 1 M or higher, the rise time of the signal is ∼3 ps long, and it is nearly independent of the concentration. Figure 4b shows the same results as in Figure 4a but on a wider time-window of 220 ps. The wider time-window allows us to see that the long-time component of the signal depends on the NaAc concentration. We explain this finding as follows: in previous studies,22,23 we found that the geminate recombination with the proton is irreversible, and therefore it quenches the NRO-* band. When the diffusion-assisted recombination of the NRO-* with the transferred proton occurs, it probably recombines with a heterocyclic nitrogen (see Scheme 2). The product of this type of recombination process is the ground-

Proton Transfer of Firefly Luciferin III

J. Phys. Chem. A, Vol. 114, No. 51, 2010 13341

Figure 6. Fluorescence up-conversion signals of the NRO-* form of D-luciferin, measured at 540 nm in 1 M sodium acetate H2O and D2O solutions.

Figure 5. Fluorescence up-conversion signal of the NROH* form of D-luciferin, measured at 460 nm, in both H2O and D2O in a 2 M aqueous solution of sodium acetate. Note the large kinetic isotope effect. (a) Short time scale of 25 ps. (b) Long time scale of 220 ps.

state NROH, whereas the product of the reversible recombination process is the excited-state NROH*. Figure 5a shows a comparison between the up-conversion signals of the NROH* form, monitored at 460 nm, from 2 M solutions of NaAc in H2O and D2O on a 25 ps time-window. The decay of the signal depends on the isotopic composition of the solution. In D2O, the decay of both the short- and the long-time components is significantly slower than in the H2O solution. Figure 5b shows the fluorescence up-conversion signals shown in Figure 5a on a stretched time-window of 220 ps. Clearly, both figures show that a kinetic isotope effect (KIE) exists for both the rate of PT to the solvent and the rate of PT to the Ac-, whether by bridged complexes or by diffusionassisted reaction. Figure 6 shows a comparison of the fluorescence upconversion signals of the NRO-* emission band from 1 M solutions of NaAc in H2O and D2O, measured at 540 nm. The figure clearly shows that the rise component of the signal is significantly longer in D2O than in H2O. Figure 7 shows the fluorescence up-conversion signals of the NROH* band in several deuterated NaAc solutions, monitored at 460 nm. In D2O, as in H2O (see Figure 3), both the shortand the long-time decay components are affected by the presence of NaAc in the solution. The short decay is invariant under change of the NaAc concentration. The intermediately long and long-time decay rates depend on the NaAc concentration. The higher the concentration, the faster its decay rate becomes.

Figure 7. Fluorescence up-conversion signals of the NROH* form of d-luciferin in various concentrations of sodium acetate in D2O, measured at 460 nm.

Discussion In the current study, we focused our attention on PT reactions from the excited D-luciferin photoacid to the weak acetate base, introduced into aqueous solutions at a fairly wide concentration range, 0-4 M. In previous studies on this kind of reaction, 8-hydroxy-1,3,6-pyrenetrisulfonate (HPTS) was the photoacid used.10,11,15,16 The advantage of HPTS in the study of this reaction is that its ESPT to water is a relatively slow process with a rate constant of kPT ) 1010 s-1, which opens a time-window of ∼200 ps, through which the direct PT to the acetate ions is observed. D-Luciferin is a stronger photoacid, and that is why its ESPT rate is faster, kPT ) 3 × 1010 s-1, which limits the observable time-window to ∼50 ps. We used the fluorescence upconversion technique, which has a time resolution of about 100 fs, to monitor the time-resolved emission of the protonated NROH* form and the deprotonated NRO-* form in the presence of NaAc in H2O and D2O. The main findings of this study are as follows: (1) We find large changes in the time-resolved emission decay of both the NROH* and the NRO-* bands in aqueous solutions containing NaAc concentration of 1 M or higher. (2) The decay profile consists of three distinguishable time components. The short one is ∼300 fs, nearly independent of the NaAc concentrations in the range of 1-4 M. Its relative amplitude increases with the concentration.

13342

J. Phys. Chem. A, Vol. 114, No. 51, 2010

(3) The intermediate component is ∼2000 fs long in H2O, and its relative amplitude depends on the NaAc concentration, whereas its decay time is almost independent. (4) The long-time decay component in a neat D-luciferin H2O solution is 30 ps long, and it designates the PT process to water. In the presence of high concentrations of NaAc, that is, 1 M or higher, this components constricts as a function of the concentration. At 4 M, the decay time is ∼18 ps. (5) In D2O, the decays of all three time components are longer: ∼600 and 3000 fs for the short and the intermediate components, respectively. (6) The long-time decay in neat D2O is ∼60 ps, and it reduces to ∼27 ps in 4 M solution of NaAc. (7) The relative amplitudes of the three time components in D2O vary only slightly with respect to their values in H2O. Previous Studies on the HPTS Acetate Reaction. Weller and co-workers24 have already studied this type of reactions in the early 1960s by means of steady-state emission spectroscopy. Pines and co-workers used the TCSPC technique, which has a time resolution of ∼10 ps, to estimate the PT rate from the ROH* to the Ac- in very high Ac- concentrations of up to 8 M.25 They found that the rate constant of the reaction is ∼7 × 1010 s-1, and it is independent of the transport of the two reactants toward each other, because at this high concentration the average distance is ∼5 Å. More recently, we used the UV pump-vis probe technique, which is based on an amplified Ti: sapphire laser system and whose time resolution is 100 fs, to monitor the Ac- reaction with the ROH* of HPTS. In our study,26 we focused on the intermediately long and long-time scales, that is, 3 ps or more. At these long times, the pump-probe signal decay, monitored at 540 nm, could be explained by the Smoluchowski model, in which a reaction takes place after the reactants approach encounter radius, a, via transport. This limits the reaction only to pairs that at t ) 0 are at r > a. One of the fitting parameters of the Smoluchowski model is the reaction rate constant at contact. We found that its value is 1.5 × 1011 s-1, a somewhat higher value than the one Pines and co-workers reported in their study.25 However, we missed the more interesting fast and intermediate components, described below, that were the focus of the most recent studies conducted on this system. Later, the Pines and Nibbering research groups collaborated to extensively study the HPTS-Ac- system.20,21 They used 400 nm femtosecond laser pulses to pump the HPTS to S1, and midIR femtosecond pulses to probe it. The IR probing allows the monitoring of not only HPTS, for which visible probing is sufficient, but also the acetate and the H3O+. This improvement in the detection enhanced the amount of detailed information retrieved from the experiment. In summary, they found that in addition to the long-time reaction, controlled by the diffusion process of both species, there are two kinds of ROH*-Accomplexes reacting within the short-time regime. One kind of complex is a contact pair, where one of the oxygen atoms of the acetate is in close proximity to the hydrogen atom of the hydroxyl group of HPTS, and it reacts within 150 fs, which is the time resolution of the experimental system. The other kind of complex is a water-bridged complex, ROH · · (H2O)n · · Ac-, for which the reaction rate is much slower, and depends on the number of water molecules in the “wire”. When there is only one water molecule in this complex, they found that the PT rate is 1/6 ps-1. This value fits nicely to the intrinsic rate constant of 1.5 × 1011 s-1 we found.26 Moreover, they also noticed that there is a time lag between the detachment of the proton from the hydroxyl group of the ROH* and the formation of the HAc.

Presiado et al. TABLE 1: Fitting Parameters of Up-conversion Fluorescence of D-Luciferin in H2O with NaAc at Several Wavelengthsa,b,c NaAc λ conc [M] [nm] 0.01 0.05 0.25 1 2 4

460 470 480 460 470 480 460 470 480 460 470 480 460 470 480 460 470 480

a1

τ1d

a2

τ2d

a3

τ3e

a4

0.15 0.09 0.08 0.19 0.17 0.16 0.11 0.077 0.059 0.266 0.198 0.193 0.31 0.24 0.21 0.275 0.216 0.168

600 600 600 600 650 450 600 600 600 270 300 400 300 300 300 250 250 250

0.024 0.024 0.022 0.019 0.018 0.017 0.13 0.077 0.052 0.266 0.261 0.244 0.35 0.32 0.29 0.327 0.335 0.291

3000 3000 3000 1900 1900 1900 3000 3000 3000 2200 1900 1900 3500 2900 2100 1900 1900 1900

0.8 0.8 0.74 0.74 0.72 0.68 0.72 0.77 0.74 0.426 0.432 0.336 0.29 0.31 0.29 0.314 0.287 0.279

26.5 25.2 27 23 25 25 28 26 26 23 21 20 27 23 20 16.5 16.5 14.5

0.024 0.08 0.15 0.046 0.082 0.144 0.036 0.077 0.148 0.041 0.108 0.227 0.05 0.13 0.21 0.085 0.162 0.263

y ) ∑ai exp[-(t/τi)Ri]. b R1 ) 0.8; R2 ) 0.85; R3 ) 0.9; R4 ) 1. τ4 ) 250 ps. d Values are in femtoseconds. e Values are in picoseconds. a

c

They concluded, therefore, that in the water-bridged complex the reaction is stepwise, and the proton first resides on the water molecule, which becomes H3O+. In the very recent past, Bakker and Cox27 also used the UV pump mid-IR probe femtosecond technique to study the HPTS acetate system. In short, their results show short, intermediate, and long time components. Their interpretation of the experimental data differs from that of Pines and Nibbering. At high base concentrations, the PT process to the Ac- occurs via proton wires of various lengths comprised of hydrogen-bonded water molecules. The typical length of these wires varies from 1 to 4 water molecules. The proton moves along the wires by a concerted Grotthus like mechanism. This long-range PT requires an activated solvent configuration to operate as a proton wire. In a recent paper by Bakker, Agmon, and co-workers,28 the experimental data were fitted using an extended Smoluchowski model, in which the PT reaction occurs not only at the contact radius a, but at all distances. The dependence on the distance is in the form of a Gaussian rate function, k(r). Thus, this model includes both PT reaction in water-bridged complexes and diffusion of the proton donor (the photoacid) and the proton acceptor (the acetate) to distances in which the water-bridged complexes are formed and the PT reaction takes place. Data Analysis of the Fluorescence Up-conversion Experimental Results. We used a multiexponential fitting procedure to fit the fluorescence up-conversion signal of D-luciferin in NaAc solution, and it was measured at short and long wavelengths. For the best fit of the short wavelength signal, we used four exponential terms. We used stretched exponents for better results in the process of fitting the fast and intermediate time components. The fitting parameters of the best fit of the short wavelength signals attributed to the NROH* decay are given in Table 1. In H2O, we found that the short-time component is 300 fs long and nearly independent of the Ac- concentration in 1 M solutions or higher. Furthermore, we found that the higher is the Ac- concentration, the larger is the amplitude of this component. The intermediate time component in H2O has a time constant of 2000 fs. The lifetime of the intermediate time

Proton Transfer of Firefly Luciferin III

J. Phys. Chem. A, Vol. 114, No. 51, 2010 13343

component much like that of the short-time component is only slightly affected by variation in the Ac- concentrations; the higher is the concentration, the larger is the amplitude. The relative amplitude of each of the two components in a 4 M solution is on the order of 0.3 at 460 nm. The third time component is long, and the decay time depends on the Acconcentration. In a neat solution, this component is ∼30 ps long and attributed to the PT to the solvent. When Ac- is introduced into the solution, another reaction takes place between the Acand the NROH*: NROH* + Ac- f NRO-* + HAc. This reaction can be described in the terms of the Smoluchowski model briefly described below. The Smoluchowski Model. The Smoluchowski model is used to describe the diffusion-assisted irreversible reaction A + B f AB, where the concentration of B is in a great excess over A. In this study, it is used to fit the time-resolved emission decay of the acidic form, NROH*, of D-luciferin in a highly concentrated Ac- aqueous solution. We assumed that the Ac- transport toward the NROH* is the rate-limiting step. The mathematical and computational details of the Smoluchowski model are given elsewhere.29 According to the Smoluchowski model, the survival probability of a single (static) donor, an excited NROH molecule (the A particle), due to its irreversible reaction with a c ) [Ac-] (B is the Ac- ion in the liquid) is given by30

S(t) ) exp(-c

∫0t k(t') dt')

(1)

where k(t) is the time-dependent rate coefficient for the donor-acceptor pair

k(t) ) kap(a, t)

(2)

whose intrinsic proton transfer rate constant is ka. The pair (NROH*/Ac-) density distribution, p(r, t), is governed by a three-dimensional Smoluchowski equation (diffusion in a potential U(r)). When U(r) ) 0, the above equations are analytically solvable for k(t). Szabo30 found an approximate expression for the time-dependent rate constant for the instances when U(r) * 0. When a potential is introduced, it behaves correctly at both t ) 0 and t ) ∞, that is:

k(0) ) kPTe-βU(a), k(∞) ) [k(0)-1 + kD-1]-1

(3)

kD ) 4πDae

(4)

where

is the diffusion-controlled rate constant, and ae is an effective radius that depends on the Coulomb pair attraction potential. The mutual diffusion constant of D-luciferin and Ac- is estimated to be around 1.10-5 cm2/s for a dilute aqueous solution with a viscosity of 1 cP. In concentrated NaAc solutions of several molars, the viscosity rises with the concentration, whereas the mutual diffusion constant decreases, because it scales inversely with the viscosity. The effective long-time reaction rate constant is given by

k(∞) ) (k(0)-1 + kD-1)-1

(5)

where k0 and kD are the intrinsic and diffusion-controlled rate constants, respectively. If k0 . kD, then k∞ = kD, and the reaction is nearly diffusioncontrolled. Because the diffusion constant is small, the pseudo first-order reaction rate constant is given by

kD′ ) 5 × 109 · cAc- s-1

(6)

This rate is on the order of 1010 s-1, and therefore it only slightly influences the long-time component of the fluorescence upconversion signal of D-luciferin in the presence of sodium acetate. Data analysis reveals that the long-time component decreases from 30 ps in a neat solution to ∼23 ps in a 1 M solution of NaAc and further decreases to ∼16.5 ps in a 4 M solution. In D2O, the long-time component decreases from 60 ps in a neat solution to ∼35 ps in a 1 M solution of NaAc and decreases even further to ∼27 ps in a 4 M solution. This dependence on the NaAc concentration is in accord with the Smoluchowski formalism, because at short-times k(t) approaches k(0), which has a very high value. According to the Smoluchowski model, the overall effective rate constant is time-dependent, and therefore denoted by k(t). At short times, it assumes the value of the reaction rate constant of the contact pair, k(0). At long times, the overall rate constant, k∞, is dominated by the transport of the reactants that leads to the formation of a contact ion pair. At very high acetate concentrations, we found in the present and in previous studies of others10,11,27,28 by using HPTS that the protonated photoacid and the acetate anion form a contact ion pair or an ion pair bridged by one or more water molecules already in the ground state.28 The proton transfer in the preexisting ground-state “contact ion pairs” is seen in the decay of the signal of the protonated form as short-time components that cannot be explained by the Smoluchowski model. The average decay time of the long-time component according to the Smoluchowski model depends on the Ac- concentration, and the decay at longtimes should depend on kD, the diffusion-controlled rate constant. From data analysis, we find that indeed the decay time depends on the Ac- concentration. However, its effect is relatively weak because the diffusion constant of large organic compounds is small, that is, ∼10-5 cm2/s, whereas the ESPT to the solvent is fast. The fourth time component in the data fitting is very long, ∼230 ps, and it is attributed to the fluorescence of the NRO-* of D-luciferin, generated by the ESPT reaction to the solvent and the NROH* reaction with Ac-. Furthermore, the longer is the wavelength at which the signal was measured, the larger the amplitude of the NRO-* signal becomes. This is a consequence of the overlap of the NROH* and the NRO-* emission bands. In the table, we provide the analysis of the NROH* signal, measured at 460, 470, and 480 nm. Table 2 provides the fitting parameters of the up-conversion signals of the NROH* in a deuterated NaAc solution. As mentioned throughout this Article, the decay times of all four components are longer in D2O. We also used the multiexponential function to fit the experimental data of the NRO-* up-conversion signal measured at 560 nm. Tables 3 and 4 give the fitting parameters of the NRO-* signals from acetate solutions of different concentrations. In general, we observe three rise time components in the NRO-* signal. The two short-time components are commensurate with the decay times seen in the NROH* signal. The fastest rise time is 300 fs in H2O. The amplitude of this

13344

J. Phys. Chem. A, Vol. 114, No. 51, 2010

Presiado et al.

TABLE 2: Fitting Parameters of Up-conversion Fluorescence of D-Luciferin in D2O with NaAc at Several Wavelengthsa,b,c NaAc conc [M] λ [nm] 0

460 460 450 460 460 450 460 460 470 480 440 450 460 460 470 480

0.04 1 2 4

a1

τ1d

a2

τ2d

a3

τ3e

a4

0.144 0.233 0.288 0.222 0.203 0.305 0.24 0.22 0.21 0.16 0.384 0.308 0.301 0.319 0.259 0.247

900 900 900 900 550 600 600 450 550 600 300 300 300 350 520 520

0.019 0.017 0.016 0.063 0.078 0.13 0.13 0.29 0.28 0.275 0.295 0.322 0.315 0.353 0.271 0.258

3000 3000 3000 3000 2400 3000 3000 3000 3600 3000 2400 2400 2400 2700 3000 3000

0.77 0.69 0.64 0.635 0.625 0.519 0.53 0.43 0.41 0.375 0.295 0.336 0.329 0.28 0.235 0.225

60 60 48 49 44 38 35 37 38 40 23 24 26 27 27 23

0.067 0.06 0.056 0.079 0.094 0.045 0.093 0.06 0.1 0.19 0.025 0.035 0.055 0.049 0.235 0.27

a y ) ∑ai exp[-(t/τi)Ri]. b R1 ) 0.8; R2 ) 0.85; R3 ) 0.9; R4 ) 1. τ4 ) 350 ps. d Values are in femtoseconds. e Values are in picoseconds.

c

TABLE 3: Fitting Parameters of Up-conversion Fluorescence of the RO- Form of D-Luciferin in H2O with NaAc Measured at 560 nma,b NaAc conc [M]

a1

τ1c

a2

τ2c

a3

τ3c

τ fd

y1e

y2e

0.01 0.05 1 2 4

0.197 0.167 0.333 0.354 0.404

300 300 300 400 300

0.197 0.167 0.333 0.354 0.356

1900 1900 1900 1950 1850

0.606 0.667 0.333 0.292 0.24

15 500 15 500 12 500 10 500 8500

0.29 0.29 0.56 0.9 1.9

0 0 0.157 0.157 0.157

1 1 0.843 0.843 0.843

a y2 ) (∑{1 - ai exp[-(t/τi)Ri]}) exp-t/τf. b R1 ) 0.9; R2 ) 0.9; R3 ) 0.75. c Values are in femtoseconds. d Values are in nanoseconds. e y ) a′y1 + (1 - a′)y2; y1 is NROH fitting signal.

TABLE 4: Fitting Parameters of Up-conversion Fluorescence of the RO- Form of D-Luciferin in D2O with NaAc Measured at 560 nma,b NaAc conc [M] 2 4

a1

τ1c

a2

τ2c

a3

τ3c

τ fd

y 1e

y2e

0.364 400 0.318 2800 0.318 14 500 1.6 0.157 0.843 0.333 400 0.333 2500 0.333 14 500 1.6 0.157 0.843

a y2 ) (∑{1 - ai exp[-(t/τi)Ri]}) exp-t/τf. b R1 ) 0.9; R2 ) 0.9; R3 ) 0.75. c Values are in femtoseconds. d Values are in nanoseconds. e y ) a′y1 + (1 - a′)y2; y1 is NROH fitting signal.

component increases with the Ac- concentration. At 4 M, the relative amplitude is ∼0.4. Part of the signal at 560 nm is assigned to the NROH* form because of the spectral overlap between the two bands. At 520 nm, the overlap is too great, so we opted for analyzing the data gathered at 560 nm, where the NRO-* signal is still considerable, but the overlap is small. We estimate that 50% of the shortest time component arises from contribution of direct excitation of the ground-state NROform or from the NROH* emission. The second rise time component is ∼1900 fs long, and its amplitude in Acconcentration of 1 M or higher is ∼0.35. The third rise time component is long and depends on the Ac- concentration. In a 50 mM acetate solution, it is ∼15.5 ps long, whereas in 4 M it reduces to roughly 8.5 ps. The amplitude of this component decreases as the Ac- concentration increases. In 50 mM the

amplitude is ∼0.6, while in 4 M it is only 0.24. We assign the third rise time component to the formation of the NRO-* form by the diffusion-assisted reaction with Ac- (the Smoluchowski model) and by direct PT to the solvent (Scheme 2). As aforementioned, the time constant of this component depends on the Ac- concentration. The time constant of this component is about twice as short as the one measured for the decay of the NROH* signal. The reason for this inconsistency probably arises from a nonexponential nonradiative decay of the NRO-* form, which we already reported.22 Limitations on the Experimental Conditions and Other Noteworthy Points. The cavity-dumped Ti:sapphire laser has a limited tunability, and the shortest wavelength, at which it operates with high intensity and stability, is 770 nm. Therefore, the shortest second harmonic wavelength is 385 nm. The peaks of the NROH* and NRO-* absorption bands are positioned at 340 and 400 nm, respectively, and the isosbestic point is at 355 nm. Photoprotolytic experiments on a photoacid require that the optimal excitation wavelength will be near the peak of protonated form. As above-mentioned, we were unable to meet the optimal excitation requirement for the D-luciferin molecule. Therefore, the direct excitation of the NRO- should also be considered. It may affect the NROH* measurements at 460, 470, and 480 nm, as well as the NRO-* signals measured at 520 nm or longer. The solubility of D-luciferin in water is low. This fact in conjunction with the relatively short optical path length of the sample cell, that is, 0.8 mm, renders the up-conversion signalto-noise ratio relatively large. Another problem we faced was the large contribution of the relatively strong Stokes Raman line of water to the upconversion signal. When excited at 385 nm, the Raman line is at 446 nm. The wide NROH* band peak is at 450 nm, near the Raman signal. To avoid any overlap of the two signals, we measured the fluorescence up-conversion signal at λ g 460 nm. In D2O, the Stokes Raman line blue-shifts to ∼426 nm, so there is no overlap with the NROH* fluorescence. The last point of concern is the relatively broad fwhm of the IRF of ∼350 fs. We found that the IRF at longer excitation and detection wavelengths is narrower. For example, at 400 nm excitation and 470 nm detection the fwhm of the IRF is 280 fs. The relatively broad IRF reduces the accuracy in the determination of the decay time of the short-time component of the NROH* up-conversion signal decay in the presence of NaAc at times shorter than 200 fs. In our measurements, we found that the short-time component was 300 and 600 fs in H2O and D2O, respectively. We are confident that the decay of this component is indeed not much shorter than 300 fs. Assignment of the Various Time Components of the Fluorescence Up-conversion Signal. In general, we adopt the interpretation of the previous studies on PT from a photoacid to a weak base.10,11,15,16,27,28 When a base is in close proximity to a photoacid, a proton is transferred on an ultrafast time scale via a solvent bridge. In a solution with a NaAc concentration of several molars, the average distance between a proton donor and acceptor is rather short. At NaAc concentrations as high as 4 M, the average distance is on the order of two water molecules, for which it is possible that water-bridged complexes pre-exist10,11 photoacid excitation, or perhaps that solvent molecules reorientation, occurring within a few hundreds of femtoseconds, is needed to form such complexes.15,27 We assign the short-time component of 300 and 600 fs in H2O and D2O, respectively, to the PT process occurring in the smallest complex between the OH group and Ac-. The KIE is

Proton Transfer of Firefly Luciferin III fairly large, which may indicate that water molecules are involved in the transfer process. However, the analysis of the collected data does not provide enough clues so as to infer the structure of the complex, whether it is a contact pair or a single water wire, connecting the OH and the Ac-. We assign the intermediate time component to a PT within a solvent bridge complex, one water molecule longer than the complex, whose fast PT is seen as the short-time component. The large difference in the decay times of the two time components was also observed in the experiments of Pines, Nibbering, and co-workers.10,11 Calculations on Green Fluorescent Protein (GFP) revealed that the PT in a one-dimensional “proton wire”, consisting of a water molecule and a serine residue that acts as a bridge between the OH group of the chromophore and the glutamate, is concerted, and occurs within tens of femtoseconds.31 Voth and co-workers32 calculated the properties of an aqueous system containing an excess proton and confined to hydrophobic cylindrical channels. They found that when the aqueous proton systems are sufficiently constricted, there is a 10-fold increase in the proton diffusion constant. This result supports the existence of proton wires, which may be relevant to PT from the photoacid to the acetate. This scenario is unlikely in the disordered three-dimensional hydrogen-bond network of water molecules, acetate anions, and sodium or potassium cations. We assign the long-time component of the decay of ∼30 ps in length to two PT processes as well as to a nonradiative decay process unrelated to the PT process. In the absence of NaAc in the aqueous solution, a direct PT to the solvent takes place as should happen upon excitation of a photoacid. The primary process of ESPT to the solvent in the case of D-luciferin takes place with time constants of 30 and 60 ps in H2O and D2O, respectively. These short times prevent the detailed detection of longer diffusion-assisted processes of PT to acetate. The Smoluchowski model formulates the problem and suggests a survival probability of the donor in the presence of many acceptors. Summary In the present work, we studied the reaction of a mild base, the acetate ion, with the protonated NROH form of D-luciferin in its excited state. For that purpose, we used time-resolved emission techniques to monitor both the protonated and the deprotonated forms. Previous studies on this subject mainly focused on the reaction between the commonly used HPTS photoacid and acetate ions. Pines, Nibbering, and co-workers jointly studied this system by UV (400 nm) pump mid-IR probe ultrafast technique.10,11 They theorized that PT occurs via preexisting water-bridged complexes between the OH of HPTS and one of the oxygen atoms of the acetate. The rate is the fastest when there are no water molecules between the donor and acceptor, which form a contact ion pair. Bakker and co-workers15,27 used similar experimental techniques, but their interpretation of the data was different. In a highly concentrated solution of sodium acetate, the distances between the donors and acceptors are short. This means that at any moment protons may be transferred via a solvent bridge of 1-4 water molecules or longer. The important experimental fact in the current study is that short, intermediate, and relatively long time components exist in the decay rate of the excited donor. We found that in 1 M aqueous solutions of sodium acetate or higher PT from D-luciferin occurs on three time scales: 300 fs, 2000 fs, and τ > 10 ps. The relative amplitude of all three components is ∼1/3

J. Phys. Chem. A, Vol. 114, No. 51, 2010 13345 at Ac- concentration of 1 M or higher. We attribute the two shortest decay components to PT between the hydroxyl group and the acetate in water-bridged complexes. The shortest time constant is somewhat larger than the one found by Pines and Nibbering’s groups. Because we use visible emission spectroscopy, we are unable to define the bridged complex, in which proton is transferred. It can be either a contact ion pair or a complex bridged by one water molecule. We also cannot measure the time of arrival of the proton to the acetate, and thereby determine whether the process is stepwise or concerted. We attribute the intermediate time constant of ∼2000 fs to a PT process in a water-bridged complex longer than the waterbridged complex, in which proton is transferred during the fastest decay component, by one water molecule. We found that the higher is the concentration of Ac-, the faster is the decay rate of the third decay component. We attribute this component to a fast diffusion-assisted reaction between the Ac- ions at long distances and the D-luciferin molecule. This reaction occurs at t g 10 ps because the diffusion constant of large ions in aqueous solutions is small, that is, D e 5 × 10-6 cm2/s. We found a distinctive isotope effect on all three time constants. In D2O solution, the time constants of the two fast components are 600 and 3000 fs. These time constants are significantly longer than the ones found for H2O solution. This fact rules out the possibility that these decay components arise from other processes such as preferential solvation of D-luciferin by Acions. Acknowledgment. This work was supported by grants from the Israel Science Foundation and from the James-Franck German-Israeli Program in Laser-Matter Interaction. Supporting Information Available: Reversible and irreversible photoprotolytic cycle of photoacids. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Strehler, B. L.; McElroy, W. D. J. Cell. Physiol. 1949, 34, 457. (2) White, E. H.; McCapra, F.; Field, G. F.; McElroy, W. D. J. Am. Chem. Soc. 1961, 83, 2402. (3) Jung, J.; Chin, C. A.; Song, P. S. J. Am. Chem. Soc. 1976, 98, 3949. (4) Barua, A. G.; Hazarika, S.; Saikia, N. M.; Baruah, G. D. J. Biosci. 2009, 34, 287. (5) Trimmer, B. A.; Aprille, J. R.; Dudzinski, D. M.; Lagace, C. J.; Lewis, S. M.; Michel, T.; Qazi, S.; Zayas, R. M. Science 2001, 292, 2486. (6) Seliger, H. H.; Buck, J. B.; Fastie, W. G.; McElroy, W. D. J. Gen. Physiol. 1964, 48, 95. (7) Ireland, J. E.; Wyatt, P. A. AdV. Phys. Org. Chem. 1976, 12, 131. (8) (a) Gutman, M.; Nachliel, E. Biochim. Biophys. Acta 1990, 391, 1015. (b) Pines, E.; Huppert, D. J. Phys. Chem. 1983, 87, 4471. (9) Tolbert, L. M.; Solntsev, K. M. Acc. Chem. Res. 2002, 35, 19. (10) Rini, M.; Magnes, B. Z.; Pines, E.; Nibbering, E.T. J. Science 2003, 301, 349. (11) Mohammed, O. F.; Pines, D.; Dreyer, J.; Pines, E.; Nibbering, E. T. J. Science 2005, 310, 5745. (12) Tran-Thi, T. H.; Gustavsson, T.; Prayer, C.; Pommeret, S.; Hynes, J. T. Chem. Phys. Lett. 2000, 329, 421. (13) Agmon, N. J. Phys. Chem. A 2005, 109, 13. (14) Spry, D. B.; Fayer, M. D. J. Chem. Phys. 2008, 128, 084508. (15) Siwick, B. J.; Cox, M. J.; Bakker, H. J. J. Phys. Chem. B 2008, 112, 378. (16) Mohammed, O. F.; Pines, D.; Nibbering, E. T. J.; Pines, E. Angew. Chem., Int. Ed. 2007, 46, 1458. (17) Mondal, S. K.; Sahu, K.; Sen, P.; Roy, D.; Ghosh, S.; Bhattacharyya, K. Chem. Phys. Lett. 2005, 412, 228. (18) Prasun, M. K.; Samanta, A. J. Phys. Chem. A 2003, 107, 6334. (19) Bhattacharya, B.; Samanta, A. J. Phys. Chem. B 2008, 112, 10101. (20) Pines, E.; Huppert, D.; Agmon, N. J. Chem. Phys. 1988, 88, 5620. (21) Agmon, N.; Pines, E.; Huppert, D. J. Chem. Phys. 1988, 88, 5631. (22) Erez, Y.; Huppert, D. J. Phys. Chem. A 2010, 114, 8075. (23) Presiado, I.; Erez, Y.; Huppert, D. J. Phys. Chem. A 2010, 114, 9471.

13346

J. Phys. Chem. A, Vol. 114, No. 51, 2010

(24) Weller, A. Prog. React. Kinet. 1961, 1, 187. (25) Pines, E.; Magnes, B. Z.; Lang, M. J.; Fleming, G. R. Chem. Phys. Lett. 1997, 281, 413. (26) Genosar, L.; Cohen, B.; Huppert, D. J. Phys. Chem. A 2000, 104, 6689. (27) Cox, M. J.; Bakker, H. J. J. Chem. Phys. 2008, 128, 174501. (28) Cox, M. J.; Timmer, R. L. A.; Bakker, H. J.; Park, S.; Agmon, N. J. Phys. Chem. A 2009, 113, 6599.

Presiado et al. (29) Cohen, B.; Huppert, D.; Agmon, N. J. Phys. Chem. A 2001, 105, 7165. (30) Szabo, A. J. Phys. Chem. 1989, 93, 6929. (31) Vendrell, O.; Gelabert, R.; Moreno, M.; Lluch, J. M. J. Phys. Chem. B 2008, 113, 5500. (32) Brewer, M. L.; Schmidtt, U. W.; Voth, G. A. Biophys. J. 2001, 80, 1691.

JP107360D