Excited State Proton Transfer and Solvent Relaxation of a 3

J. Phys. Chem. B 2008, 112, 11929–11935

11929

Excited State Proton Transfer and Solvent Relaxation of a 3-Hydroxyflavone Probe in Lipid Bilayers Ranjan Das,*,†,‡ Andrey S. Klymchenko,*,† Guy Duportail,† and Yves Me´ly† Photophysique des Interactions Biomole´culaires, UMR 7175 du CNRS, Institut Gilbert Laustriat, Faculte´ de Pharmacie, UniVersite´ Louis Pasteur, 67401 Illkirch, France, and Department of Chemistry, Bijoy Krishna Girls’ College, Howrah, West Bengal, India ReceiVed: June 5, 2008; ReVised Manuscript ReceiVed: July 22, 2008

The photophysics of a ratiometric fluorescent probe, N-[[4′-N,N-diethylamino-3-hydroxy-6-flavonyl]methyl]N-methyl-N-(3-sulfopropyl)-1-dodecanaminium, inner salt (F2N12S), incorporated into phospholipid unilamellar vesicles is presented. The reconstructed time-resolved emission spectra (TRES) unravels a unique feature in the photophysics of this probe. TRES exhibit signatures of both an excited-state intramolecular proton transfer (ESIPT) and a dynamic Stokes shift associated with solvent relaxation in the lipid bilayer. The ESIPT is fast, being characterized by a risetime of ∼30-40 ps that provides an equilibrium to be established between the excited normal (N*) and the ESIPT tautomer (T*) on a time scale of 100 ps. On the other hand, the solvent relaxation displays a bimodal decay kinetics with an average relaxation time of ∼1 ns. The observed slow solvent relaxation dynamics likely embodies a response of nonspecific dipolar solvation coupled with formation of probe-water H-bonds as well as the relocation of the fluorophore in the lipid bilayer. Taking into account that ESIPT and solvent relaxation are governed by different physicochemical properties of the probe microenvironment, the present study provides a physical background for the multiparametric sensing of lipid bilayers using ESIPT based probes. Introduction Phospholipid bilayers constitute the matrix of natural membranes in which proteins and enzymes display their activity.1 They are characterized by two fundamental physicochemical parameters, polarity and hydration, which exhibit a strong gradient across the bilayer. These two parameters are extremely low at the bilayer center, due to hydrophobic interactions of the fatty acid chains, and high at the bilayer interface, due to interactions of water molecules with the polar lipid head groups.2 The polarity and hydration gradients are responsible for the surface and dipole potentials3-5 which control membrane transport, ion conductance and insertion of proteins and other molecules into membranes and their translocation across the membrane.6,7 It is therefore, a prerequisite to have prior knowledge of these properties at precise bilayer depths to understand how the biological membranes function. Among the spectroscopic methods for studying the hydration and polarity in the lipid bilayers (such as NMR8,9 and EPR10,11), the fluorescence method using polarity-sensitive (solvatochromic) fluorescent probes e.g., Prodan12-14 and Laurdan15,16 is one of the most popular. However, these probes show some limitations as none of them discriminates polarity from hydration effects. To separately determine these effects, multiparametric fluorescent dyes are needed.17,18 In this respect, the most promising dyes are 3-hydroxyflavone derivatives that display an excitedstate intramolecular proton transfer reaction (ESIPT) resulting in dual fluorescence with two highly resolved emission bands * To whom correspondence should be addressed. E-mail: (R.D.) [email protected]; (A.S.K.) [email protected]. fr; Universite Louis Pasteur. Fax: +33 390 244313. Telephone: +33 390 244255. † Photophysique des Interactions Biomole ´ culaires, UMR 7175 du CNRS, Institut Gilbert Laustriat, Faculte´ de Pharmacie, Universite´ Louis Pasteur. ‡ Department of Chemistry, Bijoy Krishna Girls’ College.

Figure 1. (A) ESIPT transformation of 4′-(dialkyamino)-3-hydroxyflavone. (B) Chemical structures of F2N8 and F2N12S and schematic representation of the two ground-state species of F2N12S with respect to lipids in the bilayer.

(Figure 1A).19 The short-wavelength band corresponds to emission of the normal (N*) excited state, while the longwavelength band originates from the tautomer (T*, ESIPT product) state. The position of these bands as well as their

10.1021/jp804956u CCC: $40.75  2008 American Chemical Society Published on Web 08/27/2008

11930 J. Phys. Chem. B, Vol. 112, No. 38, 2008 relative intensities are highly sensitive to polarity and H-bonding effects, thus allowing the development of algorithms that discriminate between these two effects.17,20 On an increase in polarity, the relative intensity of the short-wavelength band increases and shifts to the red. Furthermore, intermolecular hydrogen bond with protic solvent (water or alcohols) inhibits the ESIPT and therefore strongly increases the relative intensity of the N* emission band.17,21 Over the past few years, the simple neutral dye 4′-N,N-dimethylamino-3-hydroxyflavone (F) was used to follow the interdigitation process in the bilayer22 and the effects of lipid composition on the bilayer surface charge,21 as well as to analyze simultaneously the polarity and hydration in lipid bilayers.23 Hydration of the probe is connected with the heterogeneity of its location, since the hydrated form (Hbonded) is located at the polar membrane interface, while the nonhydrated (H-bond free) form is more deeply embedded. To more precisely locate the fluorophore at a particular depth, we developed a 3-hydroxyflavone derivative, N-[[4′-N,N-diethylamino-3-hydroxy-6-flavonyl]-methyl]-N,N-dimethyl-1-octanaminium bromide (F2N8), containing a cationic ammonium group and an alkyl chain (Figure 1B).24 This probe showed ground-state hydrated and nonhydrated species (Figure 1B) that allow to evaluate simultaneously the hydration and polarity at a more precise location in the bilayer as compared to probe F.18,25 Recently, we developed an improved analogue of F2N8, N-[[4′-N,N-diethylamino-3-hydroxy-6-flavonyl]methyl]-N-methyl-N-(3-sulfopropyl)-1-dodecanaminium, inner salt (F2N12S), which having a zwitterionic group and longer alkyl chain (Figure 1), showed a remarkable selectivity to the cell plasma membrane. Due to its high sensitivity to the lipid composition, F2N12S showed a dramatic response to the loss of lipid transmembrane asymmetry, which was applied for detection of apoptosis.26 However, time-resolved studies of 4/-N,N-(dialkyamino)-3-hydroxyflavone derivatives in lipid vesicles have never been reported so far, and therefore it is still not clear how hydrated and nonhydrated species behave in the excited-state and whether they interconvert in the excited state. An even more intriguing question is related to the possible solvent relaxation around the highly dipolar N* species and how this relaxation interferes with the ESIPT transformation N* f T*. In aprotic low-viscous organic solvents, the ESIPT reaction in dyes F and F2N8 occurs on the scale of tens of picoseconds27 and thus is uncoupled with the solvent relaxation occurring on a much shorter time scale.28a It has also been recently demonstrated for 4′-N,N-diethylamino-3-hydroxyflavone in polar aprotic solvents by Chou29 et al. who attributed the observed fast subpicosecond component to dielectric relaxation of the solvent and the slower picosecond component to ESIPT of the solvated N* species. But the scenario is likely to be different in phospholipid vesicles, where solvent relaxation is slowed down to a nanosecond time scale.15,30-35 As a consequence, the ESIPT of our probes may be coupled with solvent relaxation in lipid bilayers. To check this, we carried out time-resolved fluorescence measurements of the F2N12S probe in lipid vesicles composed either of a natural mixture of phospholipids, egg yolk phosphatidylcholine (EYPC), or of a synthetic phospholipid, dioleoyl phosphatidylcholine (DOPC), both featuring similar properties of the liquidcrystalline phase. The present work is the first demonstration of an ESIPT coupled with solvent relaxation in the lipid bilayers.

Das et al. earlier.26 Large unilamellar vesicles (LUV) of 200 µM lipid concentration were obtained in 15 mM phosphate-citrate buffer, pH 7.0, by the classical extrusion method.36 Briefly, a suspension of multilamellar vesicles was extruded by using a Lipex Biomembranes extruder (Vancouver, Canada). The size of the filters was first 0.2 µm (7 passages) and thereafter 0.1 µm (10 passages). This generates monodisperse LUVs with a mean diameter of 0.11 µm as measured with a Zetamaster 3000 instrument (Malvern Instruments, Paris, France). LUVs were labeled by adding an aliquot of probe stock solution in DMSO to a final concentration of 2 µM. In LUVs labeled with the probe, DMSO was always e0.1% v/v. All the spectroscopic measurements were performed at 20 °C. Absorption and Fluorescence spectra were recorded on Cary 4 spectrophotometer (Varian) and FluoroMax 3.0 (Jobin Yvon, Horiba) spectrofluorometer, respectively. Time-resolved fluorescence measurements were performed with the time-correlated, single-photon counting technique using the frequency-doubled output of a Ti-sapphire laser (Tsunami, Spectra Physics) pumped by a Millenia X laser (Tsunami, Spectra Physics). The excitation wavelength was set at 430 nm. To obtain timeresolved spectra, fluorescence decays at wavelengths spanning the emission spectrum (460-610 nm) were collected at the magic angle (54.7°) of the emission polarizer in order to avoid any artifact due to vertically polarized excitation beam. The emission decays were recorded through a monochromator with a 4-nm bandpass, which ensured elimination of any contamination from the scattering of the laser excitation light. In addition, a long-wavelength path filter, model GG 455, from Schott (Mainz, Germany) was placed before the monochromator to cut off any interference from the scattered excitation light. The single-photon events were detected with a microchannel plate Hamamatsu R3809U photomultiplier coupled to a Philips 6954 pulse preamplifier and recorded on a multichannel analyzer (Ortec 7100) calibrated at 25.5 ps/channel. The instrumental response function was recorded with a polished aluminum reflector, and its full-width at half-maximum was 50 ps. The time-resolved decays were analyzed together with instrument response functions using an iterative reconvolution method27 which provided an effective time resolution of ∼15 ps. The goodness of the fit was evaluated from the χ2 values, the plots of the residuals and the autocorrelation function. Deconvolution of the emission spectra into three bands (N*, T*, and H-bonded form H-N*) was carried out as previously described18,23,25 by using the Siano software kindly provided by Dr. A.O. Doroshenko (Kharkov, Ukraine). The program is based on an iterative nonlinear least-squares method based on the Fletcher-Powell algorithm. The individual emission bands in this program were approximated by a log-normal function. Time resolved emission spectra (TRES) were generated from a set of emission decays (at least 16 wavelengths) recorded at 10 nm intervals spanning the fluorescence spectrum using the “Spectral Reconstruction” method as described elsewhere.28b The time evolution of the peak frequencies ν(t) in the TRES were fitted using a sum of two log-normal line shape functions: 2

F(ν,t) )

∑ g exp{-ln 2(ln[1 + 2b (ν - ν )/∆ ]/b ) }, 2

i

i

i

i

i

R>1

i)1

Materials and Methods The phospholipids egg yolk phosphatidyl-choline (EYPC) and 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) were from Sigma. The probe F2N12S was synthesized as described

) 0,

for R e 1,

where R ) 2bi(ν - νi)/∆i

(1)

where gi, νi, bi, and ∆i are the peak height, peak frequency, asymmetry parameter and width parameter respectively. The time-dependent fwhm (full width at half-maximum) function,

Proton Transfer and Solvent Relaxation

J. Phys. Chem. B, Vol. 112, No. 38, 2008 11931

Figure 2. Steady-state emission (A) and excitation (B) spectra of F2N12S in DOPC vesicles (pH 7.0) at 20 °C. Dashed curves in (A) represent the result of deconvolution of the emission spectrum into the three bands. F2N12S probe and DOPC concentrations were 2 µM and 200 µM, respectively. Excitation wavelength was 400 nm.

Γ(t) is related to the width and asymmetry parameter28b and is extracted from log-normal fitting of the TRES. Typical uncertainty in the position of the peak maximum obtained from the log-normal fit in the TRES was estimated to be ( 200 cm-1. The solvation response or the solvent correlation function C(t) is calculated according to

C(t) )

ν(t) - ν(∞) ν(0) - ν(∞)

(2)

where ν(0) and ν(∞) correspond to the peak frequency values obtained by “time-zero estimation” and infinity time, respectively. The peak frequency at t ) 0, which is not directly measurable, was calculated using the approximation proposed by Fee37 et al.

ν(0) ≈ νnp(em) - νnp(abs) + νp(abs)

(3)

where the indices “np” and “p” refer to the steady-state peak frequencies measured in nonpolar and polar solvents, respectively, and the terms “em” and “abs” in the parentheses denote the steady-state fluorescence and absorption peak frequencies respectively. For a quantitative solvent relaxation study the determination of ν(0) is crucial. The above method proposed by Fee37 has limitations for the fluorophores exhibiting complex photophysics, while glassy state of 2-methyltetrahydrofuran at 100 K was reported to serve better for ν(0) estimation.32 Nevertheless, in the present work we use the method proposed by Fee, because ESIPT in case a 3-hydroxyflavone dye is a reaction that convert the N* form into the tautomer (T*) exhibiting a well-separated emission band, so that ESIPT reaction has no direct influence on ν(0) value. Results and Discussion Steady-State Spectroscopy. The steady-state emission spectra of F2N12S exhibited a dual emission in large unilamellar vesicles (LUVs) (Figure 2A) with maxima at ∼510-520 nm and ∼570-580 nm, in line with earlier observations with this probe18 and the closely related probes F2N824 and F.23 Both the broad shape of the short-wavelength band and its limited

Figure 3. Time-dependent emission decays of F2N12S (2 µM) in EYPC and DOPC lipid vesicles (200 µM, pH 7.0) at 20 °C. Excitation wavelength was 430 nm.

separation (∼50-60 nm) from the T* band, indicate that the short-wavelength band is heterogeneous. Based on the strong homology of F2N12S with F2N8 and F,18,23,24,26 the short wavelength emission band can be assigned to an emission from the overlapped H-N* and N* (corresponding to H-bonded and H-bond free forms, Figure 1B) bands of F2N12S whereas the longer wavelength band is attributed to the emission from the excited-state T* (tautomer form). The H-N* band centered around 540-550 nm corresponds to the form of the dye H-bonded with water, which can be obtained by decomposition of its fluorescence spectra into three bands (Figure 2A), as it was previously done for F2N8.18 As previously shown for probe F2N8 in LUVs,18 the excitation spectra of F2N12S in DOPC vesicles monitored at the red-edge (540 nm) of the short-wavelength band is redshifted by ∼3 nm (Figure 2B) with respect to the excitation spectra monitored at the blue edge (470 nm) of the shortwavelength band or at the maximum of the T* band emission at 575 nm. This shift confirms the presence of two groundstate species of F2N12S in lipid bilayers: a H-bond free (N) species giving N* and T* emission bands and a H-bonded (H-N) species with red-shifted absorption, giving a single-band H-N* emission around 540-550 nm.18 Time-Resolved Fluorescence Spectroscopy. The photophysics of F2N12S in lipid bilayers was further explored using time-resolved fluorescence spectroscopy. Figure 3 displays the emission decays in EYPC and DOPC LUVs monitored at different wavelengths. The emission decays are strongly wavelength dependent. The fastest decays were observed at the blue edge of the emission spectrum at 460 nm, while the slowest decays were observed at the red edge (610 nm) of the T* band (Figure 3). At all of the recorded emission wavelengths for both EYPC and DOPC vesicles, the decays were multiexponential with 3-4 components (Table 1, see also Supporting Information). A fast decay component τ1 of ∼30-45 ps with a high amplitude was observed at the blue edge (460 nm) of the short-wavelength emission band. The amplitude of this component decreases at higher wavelengths and drops to 0 in the range 520-560 nm (Figure 4). At the wavelengths corresponding to the T* emission (e.g., 590 nm) a similar fast rise component of ∼30-45 ps

11932 J. Phys. Chem. B, Vol. 112, No. 38, 2008

Das et al.

TABLE 1: Time-Resolved Emission Data of F2N12S (2 µM) in LUVs Composed of EYPC or DOPC (200 µM, pH 7) at 20 °Ca R1

LUV

λ, nm

τ1

EYPC

460 480 510 540 590 460 480 510 540 590

0.045 0.073

0.607 0.431

0.043 0.033 0.058

-0.840 0.539 0.371

0.035

-0.790

DOPC

τ2

R2

τ3

R3

τ4

R4

χ2

0.245 0.336 0.250 0.300 0.290 0.181 0.265 0.184 0.220 0.170

0.280 0.347 0.328 -0.068 -0.150 0.335 0.335 0.250 -0.079 -0.210

0.92 1.11 0.89 0.81 1.20 0.700 0.795 0.746 0.600 1.00

0.073 0.122 0.302 0.283 0.228 0.087 0.150 0.324 0.178 0.160

2.80 2.84 2.90 2.90 3.00 2.50 2.63 2.80 2.80 2.70

0.040 0.100 0.370 0.717 0.772 0.039 0.144 0.426 0.822 0.840

1.20 0.96 1.04 1.09 1.20 1.02 1.01 1.07 1.09 1.20

λ ) recorded emission wavelength; τ1, τ2, τ3 and τ4 ) decay times (ns), R1, R2, R3 and R4 ) pre-exponential coefficients, χ2 ) goodness of the fit, obtained after deconvolution of the corresponding decay curves. The estimated error of the decay times (τi) is less than (10%. a

Figure 4. Wavelength dependence of the amplitudes of the four fluorescence decay times of F2N12S in EYPC (A) and DOPC (B) vesicles. The experimental conditions are the same as in Figure 3.

appears with a negative amplitude, indicating an excited-state reaction that can be attributed to the ESIPT transformation N* f T*. The emission decays at both ∼ 460 and 590 nm exhibit also an identical long-lived decay component (τ4) of ∼ 2.5-3 ns, which likely corresponds to the population decay (radiative plus nonradiative) time of the N* and T* forms. The amplitude corresponding to this longest decay time (τ4) is positive at all wavelengths and increases in magnitude upon going towards to the red edge of the monitored wavelength range (Figure 4). Moreover, the two pre-exponential factors describing the growth and decay of T* at 590 nm are similar in magnitude but opposite in sign (Table 1), suggesting a reversible ESIPT reaction, 38,39 as observed earlier for the parent dye F in organic solvents.27,40 A reversible ESIPT is quite reasonable given that the population decay time of the ESIPT tautomer is roughly 100-fold larger than the fast component attributed to ESIPT. Consequently, there is a fast equilibrium between the normal (N*) and tautomer (T*) excited states, in which the rates of both forward and reverse ESIPT are much faster than that of the population decay rate for both states. The other two decay components τ2 and τ3 of ∼180-250 ps and ∼0.8-1.2 ns, respectively, were observed in the whole range of monitored emission wavelengths, but their amplitudes were wavelength-dependent. The shorter component, τ2, decays with a positive amplitude at 460-510 nm wavelengths range, while at 540 nm and above, it evolves as a rise component with

a negative amplitude (Figure 4 and Table 1). The evolution of τ2 as a rise component at 540 nm, where the emission is mainly from the solvent relaxed H-N* species, likely describes a photophysical mechanism related to the formation of the H-N* species from the N* species. In this respect, a possible explanation would be the formation of the H-N* species through solvent relaxation of the N* species. Indeed, the N* species possesses a large dipole moment due to intramolecular charge transfer (Figure 1A), which renders the 4-carbonyl group a strong H-bond acceptor.18,20 Consequently, the N* species likely induces dipolar relaxation in the lipid bilayer and formation of a hydrogen-bond of the 4-carbonyl group (Figure 1B) with a water molecule, leading to the formation of the H-N* species. Since the water molecules are bound quite strongly to the phospholipids, the overall relaxation is likely to be slow. Finally, the amplitude of the longer component (τ3) is positive at all the wavelengths and is most pronounced in the range 510-540 nm (Figure 4), suggesting that this component probably describes the population decay of the solvent relaxed H-N* species. However, we should mention that due to the kinetic complexity introduced by the ESIPT equilibrium and the solvent relaxation, the obtained decay constants cannot be directly used to calculate the kinetic constants of the corresponding excited-state processes, but only provide a rough estimate for their time-scale. Time-Resolved Emission Spectra (TRES) and Solvent Relaxation. Solvent relaxation around the F2N12S dye in LUVs was further substantiated by the continuous time dependent Stokes shift (TDSS) to the red of the peak maximum of the short-wavelength emission band in the TRES (Figure 5). In contrast, no such dynamic Stokes shift was observed for the long-wavelength emission band, indicating absence of solvent relaxation of the less polar T* species. Recently, our semiempirical calculations41 for analogous N,N-dialkylamino-3-hydroxyflavones have shown that the T* species possesses a significantly lower dipole moment than the N* species and its dipole moment is found to be more or less similar to that of the ground-state N species. The TRES also reveals that the dynamic Stokes shift takes place on a nanosecond time scale, indicating a slow solvent relaxation dynamics. To obtain quantitative information on solvent relaxation, the solvent correlation function C(t) was extracted from the TRES according to eq 2 in the Experimental Section of ref 42. The obtained C(t) functions (Figure 6) for EYPC and DOPC bilayers could be satisfactorily fitted with a biexponential equation:

C(t) ) {a1 exp(-t/τs1) + a2 exp(-t/τs2)}

(4)

where a1 and a2 are pre-exponential coefficients and τs1 and τs2 are solvent relaxation time constants.

Proton Transfer and Solvent Relaxation

J. Phys. Chem. B, Vol. 112, No. 38, 2008 11933

Figure 5. Time-resolved emission spectra (TRES) of F2N12S in EYPC and DOPC vesicles. The experimental conditions are the same as in Figure 3.

Figure 6. Kinetics of solvent relaxation in EYPC and DOPC LUVs evaluated from the time-resolved emission of F2N12S. The experimental conditions are the same as in Figure 3.

TABLE 2: Decay Components in the Solvent Relaxation of F2N12S in LUVsa LUV

τs1, ns

a1

τs2, ns

a2

EYPC 0.15 ( 0.01 0.49 1.73 ( 0.05 0.51 DOPC 0.15 ( 0.01 0.43 1.70 ( 0.04 0.57

〈τs〉, ns ∆ν, cm-1 0.95 1.03

2305 2267

a 〈τs〉 ) ∫∞0 C(t) dt, average solvent relaxation time; τs1, τs2 ) solvent relaxation components; a1 and a2 ) pre-exponential coefficients; ∆ν ) dynamic Stokes shift. The experimental conditions are the same as in Figure 3.

For both EYPC and DOPC, a fast relaxation time of ∼150 ps and a slower one of ∼ 1.7 ns were obtained (Table 2). The average relaxation time 〈τs〉 for F2N12S in LUVs is ∼1 ns for both lipids and close to that observed with PRODAN in EYPC vesicles.30 This mean relaxation time in the nanosecond regime is much slower than solvent relaxation in pure water which is commonly in the subpicosecond time scale.43 The slow dynamics in restricted environments like phospholipid vesicles30-35 are attributed to a collective relaxation of the dye environment, i.e., to a time dependent reorientation of the molecular domains of phospholipids and water molecules surrounding the fluorescent dye.31,33,44 Another contributing factor for the observed nanosecond relaxation dynamics is likely to be the relocation of the fluorophore in the lipid bilayer, since previous parallax data on F2N818 showed a shallower location of the H-N* species as compared to the N* species. The N* species may thus diffuse from the deeper into the shallower region close to the interface, leading also to relaxation of the dye environment.

Figure 7. Time dependence of the full width at half-maximum, Γ(t), for the short-wavelength emission band of F2N12S in DOPC and EYPC LUVs. The experimental conditions are the same as in Figure 3.

SCHEME 1: Photophysical Processes of F2N12S in Lipid Bilayersa

a Key: H-N*, solvent-relaxed hydrogen-bonded normal form; N*, non-hydrogen bonded normal form; T*, ESIPT tautomer form.

The time dependence of the spectral half-widths (fwhm) of the dynamically Stokes shifted band in TRES can additionally be used to describe the solvent relaxation. For a pure solvent relaxation process, the fwhm function, Γ(t) generally increases at initial times, reaches a maximum around the average relaxation time, and then decreases.45 The Γ(t) obtained for the dynamically Stokes shifted shortwavelength emission band (Figure 7) in both of the LUVs is fully consistent with solvent relaxation. The time corresponding to the maximal values of Γ(t) can serve as a rough index of the kinetics of solvent relaxation. This time (about 1 ns) is slightly shorter in EYPC than in DOPC vesicles (Figure 7), in agreement with the trends in the 〈τs〉 values (Table 2). ESIPT and Solvent Relaxation. To rationalize the photophysics of F2N12S in lipid bilayers, we propose a model in which we hypothesize that apart from the population decay (radiative and nonradiative), two other decay channels govern the photophysics of the hydrogen-bond free N* form in the S1 state. One is the fast and reversible ESIPT (N*TT*) and the second one is the solvent relaxation (N* f H-N*) (Scheme 1). Both processes are evidenced from the TRES (Figure 5), where ESIPT is observed as a fast growth of the T* emission on a time scale around 100 ps, and the solvent relaxation is manifested as the time dependent Stokes shift of the shortwavelength band to the red on a nanosecond time scale. We should also take into account the ground-state heterogeneity, namely the presence of H-bond free N and H-bonded H-N forms,18,23 observed in the excitation spectra. In the ground state, the N form dominates, as evidenced by the blue-shifted position of the short-wavelength band at the early times of emission (50

11934 J. Phys. Chem. B, Vol. 112, No. 38, 2008 ps, see Figure 5). In contrast, at later emission times, we observe the transformation of the N* species into the solvent relaxed H-N* species, resulting in the predominant red-shifted emission of the H-N* species. The N* f H-N* transformation through solvent relaxation may involve a number of parallel processes in the excited state: dipolar relaxation, localized H-bond formation, and relocation of the fluorophore in the lipid bilayer. These processes connected with relaxation of the dye environment probably contribute to the observed bimodal relaxation dynamics. (Table 2). Similar multiexponential relaxation dynamics was also observed for other environment-sensitive probes, such as Prodan, Laurdan and Patman.30 The peculiarity in the case of F2N12S is that the ESIPT reaction competing for the depopulation of the solvent nonrelaxed N* form is reversible, giving the T* tautomer which does not undergo any solvent relaxation, due to its low dipole moment. Therefore, despite the irreversible N* f H-N* transformation, the solvent nonrelaxed N* state is continuously repopulated by the reversible ESIPT reaction (Scheme 1). This continuous repopulation of the N* state explains the limited decrease of the fwhm in TRES, after reaching the maximum at ca. 1 ns (Figure 7), so that fwhm preserves its large values due to the continuous presence of the solvent nonrelaxed (H-bond free) N* species. The obtained results provide a physical background for the previously reported multiparametric analysis of the membrane environment based on the N*/T* and H-N*/(N* + T*) intensity ratios.17,23 According to the proposed model, the ESIPT equilibrium N* T T* is established on the time scale of 100 ps, while the solvent relaxation N* f H-N* is a much slower process that takes place on a time scale of 1 ns. Therefore, the N*/T* ratio, is mainly controlled by the ESIPT equilibrium, and should not be strongly affected by the slow solvent relaxation. Thus, the N*/T* ratio should describe the static dielectric properties of the lipid bilayers (such as electric fields). In contrast, the intensity ratio H-N*/(N* + T*) should describe the hydration water in lipid bilayers that contributes to both the formation of H-bonded species (H-N) in the ground-state and the slow solvent relaxation (N* f H-N*) in the excited state. Therefore, the N*/T* and H-N*/(N* + T*) intensity ratio parameters are intrinsically independent and describe different physicochemical properties of lipid bilayers. This constitutes the basis of the multiparametric characterization. Conclusions In the present work, the ratiometric fluorescent dye F2N12S was found to exhibit unique photophysics in the lipid bilayers. The TRES demonstrate the existence of both ESIPT and solvent relaxation. ESIPT reaction, being fast, establishes an equilibrium between the two tautomeric forms on a time scale of ca. 100 ps, while solvent relaxation is slow with an average relaxation time of ∼1.0 ns. The slow and multiexponential relaxation kinetics is connected with dipolar relaxation of the dye environment, localized probe-water H-bond formation, and relocation of the fluorophore in the lipid bilayer. To the best of our knowledge, this is a first demonstration of an environmentsensitive fluorescent dye exhibiting both ESIPT and solvent relaxation in LUVs. The obtained results provide a physical background for multiparametric physicochemical characterization of lipid bilayers using ESIPT based probes. Acknowledgment. R.D. is a recipient of the Visiting Scientist fellowship of CNRS, France. This work was supported by a Conectus Alsace grant.

Das et al. Supporting Information Available: Extended version of Table 1. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Zimmerberg, J.; Gawrish, K. Nat. Chem. Biol. 2006, 2, 564– 567. (2) Cevc, G.; Marsh, D. Phospholipid Bilayers. Physical Principles and Models; Wiley: New York 1987. (3) Franklin, J. C.; Cafiso, D. S. Biophys. J. 1993, 65, 289. (4) Gawrisch, K.; Ruston, D.; Zimmerberg, J.; Parsegian, V. A.; Rand, R. P.; Fuller, N. Biophys. J. 1992, 61, 1213. (5) Gabdoulline, R. R.; Zheng, C.; Vanderkooi, G. Chem. Phys. Lipids 1996, 84, 139. (6) Cladera, J.; O’Shea, P. Biophys. J. 1998, 74, 2434. (7) Voglino, L.; McIntosh, T. J.; Simon, S. A. Biochemistry 1998, 37, 12241. (8) Volke, F.; Eisenblatter, S.; Klose, G. Biophys. J. 1994, 67, 1882. (9) Nagle, J. F.; Liu, Y.; Tristram-Nagle, S.; Epand, R. M.; Stark, R. E. Biophys. J. 1999, 77, 2062. (10) Marsh, D. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 7777. (11) Bartucci, R.; Guzzi, R.; Marsh, D.; Sportelli, L. Biophys. J. 2003, 84, 1025. (12) Krasnowska, E. K.; Gratton, E.; Parasassi, T. Biophys. J. 1998, 74, 1984. (13) Massey, J. B. Biochim. Biophys. Acta 1998, 1415, 193. (14) Bagatolli, L. A.; Gratton, E.; Fidelio, G. D. Biophys. J. 1998, 75, 331. (15) Parasassi, T.; Di Stefano, M.; Loiero, M.; Ravagnan, G.; Gratton, E. Biophys. J. 1994, 66, 763. (16) Bagatolli, L. A.; Parasassi, T.; Fidelio, G. D.; Gratton, E. Photochem. Photobiol. 1999, 70, 557. (17) Klymchenko, A. S.; Demchenko, A. P. Phys. Chem. Chem. Phys. 2003, 5, 461. (18) Klymchenko, A. S.; Me´ly, Y.; Demchenko, A. P.; Duportail, G. Biochim. Biophys. Acta 2004, 1665, 6. (19) Sengupta, P. K.; Kasha, M. Chem. Phys. Lett. 1979, 68, 382. (20) Klymchenko, A. S.; Pivovarenko, V. G.; Demchenko, A. P. J. Phys. Chem A. 2003, 107, 4211. (21) Duportail, G.; Klymchenko, A. S.; Me´ly, Y.; Demchenko, A. P. FEBS Lett. 2001, 508, 196. (22) Bondar, O. P.; Pivovarenko, V. G.; Rowe, E. S. Biochim. Biophys. Acta 1998, 1369, 119. (23) Klymchenko, A. S.; Duportail, G.; Demchenko, A. P.; Me´ly, Y. Biophys. J. 2004, 86, 2929. (24) Klymchenko, A. S.; Duportail, G.; Oztu¨rk, T.; Pivovarenko, V. G.; Me´ly, Y.; Demchenko, A. P. Chem. Biol. 2002, 9, 1199. (25) M’Baye, G.; Me´ly, Y.; Duportail, G.; Klymchenko, A. S. Biophys. J. 2008, 95, 1217–1225. (26) Shynkar, V. V.; Klymchenko, A. S.; Kunzelmann, C.; Duportail, G.; Muller, C. D.; Demchenko, A. P.; Freyssinet, J.-M.; Me´ly, Y. J. Am. Chem. Soc. 2007, 129, 2187. (27) (a) Shynkar, V. V.; Me´ly, Y.; Duportail, G.; Pie´mont, E.; Klymchenko, A. S.; Demchenko, A. P. J. Phys. Chem. A 2003, 107, 9522. (b) Roshal, A. D.; Organero, J. A.; Douhal, A. Chem. Phys. Lett. 2003, 379, 53–59. (28) (a) Reynolds, L.; Gardecki, J. A.; Frankland, S. J. V.; Horng, M. L.; Maroncelli, M. J. Phys. Chem. 1996, 100, 10337–10354. (b) Horng, M. L.; Gardecki, J. A.; Papazyan, A.; Maroncelli, M. J. Phys. Chem. 1995, 99, 17311. (29) (a) Chou, P. T.; Pu, S. C.; Cheng, Y. M.; Yu, W. S.; Yu, Y. C.; Hung, F. T.; Hu, W. P. J. Phys. Chem. A 2005, 109, 3777. (b) Chou, P. T.; Yu, W. S.; Cheng, Y. M.; Pu, S. C.; Yu, Y. C.; Lin, Y. C.; Huang, C. H.; Chen, C. T. J. Phys. Chem. A 2004, 108, 6487. (30) Sykora, J.; Kapusta, P.; Fidler, V.; Hof, M. Langmuir 2002, 18, 571. (31) Jurkiewicz, P.; Olzynska, A.; Langner, M.; Hof, M. Langmuir 2006, 22, 8741. (32) Sykora, J.; Slavicek, P.; Jungwirth, P.; Barucha, J.; Hof, M. J. Phys. Chem. B 2007, 111, 5869. (33) Sommer, A.; Paltauf, F.; Hermetter, A. Biochemistry 1990, 29, 11134. (34) Parasassi, T.; Conti, F.; Gratton, E. Cell. Mol. Biol 1986, 32, 103. (35) Parasassi, T.; De Stasio, G.; d’Ubaldo, A.; Gratton, E. Biophys. J. 1990, 57, 1179. (36) Mayer, L. D.; Hope, M. J.; Cullis, P. R. Biochim. Biophys. Acta 1986, 858, 161. (37) Fee, R. S.; Maroncelli, M. Chem. Phys. 1994, 183, 235. (38) Laws, W. R.; Brand, L. J. Phys. Chem. 1979, 83, 795. (39) Das, R.; Mitra, S.; Mukherjee, S. J. Phys. Chem. 1996, 100, 14514.

Proton Transfer and Solvent Relaxation (40) Roshal, A. D.; Organero, J. A.; Douhal, A. Chem. Phys. Lett. 2003, 379, 53–59. (41) Yesylevskyy, S. O.; Klymchenko, A. S.; Demchenko, A. P. J. Mol. Structure: Theochem. 2005, 755, 229. (42) Bakshiev, N. G. Opt. Spectrosc. 1965, 19, 196. (43) Jimenez, R.; Fleming, G. R.; Kumar, P. V.; Maroncelli, M. Nature 1994, 369, 471.

J. Phys. Chem. B, Vol. 112, No. 38, 2008 11935 (44) Nilsson, L.; Halle, B. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 13867. (45) Hof, M. SolVent relaxation in biomembranes (Applied Fluorescence in chemistry, biology, and medicine); Rettig,W., Ed.; Springer-Verlag: Berlin, 1999, 439.

JP804956U