Encapsulation of 3-hydroxyflavone and fisetin in β-cyclodextrins: Excited state proton transfer fluorescence and molecular mechanics studies

www.elsevier.com/locate/cplett

Chemical Physics Letters 424 (2006) 379–386

Encapsulation of 3-hydroxyflavone and fisetin in b-cyclodextrins:Excited state proton transfer fluorescence

and molecular mechanics studies q

Anwesha Banerjee, Pradeep K. Sengupta *

Biophysics Division, Saha Institute of Nuclear Physics, 1/AF, Bidhannagar, Kolkata 700064, West Bengal, India

Received 11 April 2006; in final form 26 April 2006Available online 5 May 2006

Abstract

Excited-state intramolecular proton-transfer (ESIPT) and dual emission properties (emission profile, anisotropy and decay kinet-ics) of 3-hydroxyflavone (a synthetic, model flavonol) and fisetin (3,7,3 0,4 0-OH-flavone, a therapeutically active plant flavonol) havebeen exploited to study their encapsulation in nano-cavities comprising of natural and chemically modified b-cyclodextrins. In thepresence of b-CDs, both the flavonols show significantly enhanced relative yields (along with changes in other emission parameters)of the tautomer emission. In addition, for fisetin, large blue shifts are observed for the normal emission (which has significantcharge transfer character). From these we infer that the flavonols are encaged in predominantly hydrophobic micro-environments,where external hydrogen bonding perturbations (interfering with the intrinsic ESIPT), and dipolar relaxation effects, are minimized.This is further explained from results of molecular mechanics calculations which indicate selectivity in orientation of the encapsu-lated flavonols. Moreover, chemical modification of the b-CDs is found to profoundly influence the binding affinities of the guestflavonols.� 2006 Elsevier B.V. All rights reserved.

1. Introduction

In 1936 Rusznyak and Szent-Gyorgii first recognized thetherapeutically beneficial role of dietary flavonoids [1],which are polyphenolic compounds abundant in commonplant based food and beverages (e.g. citrus fruits, apple,soy products, onion, broccoli, tea and red wine). The lastdecade has witnessed a remarkable renascence of interestin this area, marked by an explosive growth of researchon various bioactive flavonoids, which have proved to be

0009-2614/$ - see front matter � 2006 Elsevier B.V. All rights reserved.

doi:10.1016/j.cplett.2006.05.006

q This research was presented at the National Symposium on ‘Mole-cules, Interactions and Design: A Biophysical Perspective (IBS 2006)’, heldunder the aegis of the Indian Biophysical Society Kolkata, India, January7–9, 2006, for which the younger author (A.B.) was recipient of the N.N.Saha Memorial award for best Poster presentation.

* Corresponding author. Fax: +91 33 23374637.E-mail addresses: [email protected], pradeepsinp@

yahoo.co.in (P.K. Sengupta).

effective against a wide range of free radical mediatedand other human diseases [2,3] (e.g. atherosclerosis, ische-mia, neuronal degeneration, cancers, tumors, allergies, car-diac problems, inflammation, AIDS, etc.). The highpotency and low systemic toxicity of these compoundsmake them viable alternatives to conventional therapeuticdrugs. However, since such applications are often limitedby poor water solubility, there is much current interest inusing suitable drug delivery vehicles (e.g. cyclodextrins)capable of ensuring increased hydrosolubility, and conse-quently improved bioavailability of these drugs.

On a different scenario, a major class of natural flavo-noids of widespread occurrence, namely flavonols (3-hydroxyflavones) have emerged as one of the best knownmolecular systems exhibiting excited state intramolecularproton transfer (ESIPT) and dual fluorescence behavior[4–6]. Consequently, there is enormous interest on suchcompounds as prototypes for mechanistic studies of

380 A. Banerjee, P.K. Sengupta / Chemical Physics Letters 424 (2006) 379–386

ESIPT, and as exquisitely sensitive fluorescence probesfor exploring their binding sites in various bio-relevanttargets e.g. proteins, DNA, bio-membranes and mem-brane mimetic organized assemblies (liposomes, normaland reverse micelles) [7–12], suggesting other promisingextensions.

Cyclodextrins (CDs) are cyclic oligomers composed ofa-D-glucose units connected to form doughnut shapedtruncated cones, having a hydrophobic cavity (‘nano-cages’ of ca. 0.45–0.8 nm diameter) capable of encapsulat-ing a wide range of compounds [13–17]. This has madetheir use extremely popular as efficient drug delivery vehi-cles, capable of enhancing the solubility, dissolution rateand membrane permeability of the encapsulated drugs[15,16]. Moreover, the reduced polarity and restrictedgeometry of the CD cavities provide interesting testinggrounds for studying the energetics and dynamics of var-ious photophysical processes of the encapsulated mole-cules [17,18].

The existing literature documents only a few ratherlimited attempts to characterize the interaction of flavo-noids with CDs [19–22], via spectroscopic (NMR andFTIR) and other biophysical techniques. In this Letter,we demonstrate the usefulness of the ESIPT and dualfluorescence of flavonols to explore their incorporationand binding in CD cavities. For this work, we chosetwo flavonols, namely 3-hydroxyflavone (3HF) and fisetin(3,7,3 0,4 0-OH flavone), and examined their complexationwith natural b-cyclodextrin (b-CD) and its chemicallymodified derivative (succinyl-(2-hydroxypropyl)-b-cyclo-dextrin (SHPb-CD)). While 3HF (a model flavonol ofsynthetic origin) has been the object of numerous spectro-scopic studies mainly as a prototype for intrinsic ESIPTreaction [4], fisetin (a naturally occurring flavonol), hasgained prominence for its interesting biological andrelated therapeutic (e.g. antioxidant, anti-cancer, and antiAIDS) properties [9], as well as remarkable fluorescencebehavior,with charge transfer (CT) emission from its nor-mal (non-proton transferred) form simultaneously occur-ring together with ESIPT tautomer emission, leading toits dual role as a CT and ESIPT fluorescence probe[9,12,23].

We report that on encapsulation into the b-CD hosts,conspicuous changes occur in the emission behavior of theflavonols, with appearance of unique dual fluorescenceemission with characteristic spectroscopic signatures,reflecting the influence of the confined geometries andmicro-environments of the respective CD nano-cavities onthe ESIPT process and resulting steady-state and dynamicfluorescence parameters. Binding constants, estimated fromthe emission data, permit comparison between b-CD andSHPb-CD with regard to their relative encapsulation effi-ciencies. We also present results of molecular mechanics cal-culations which provide useful insights regarding theprocess of formation of the flavonol-b-CD inclusion com-plexes and the orientation of the guest molecules (flavonols)in the CD nano-cavities.

2. Materials and methods

2.1. Experimental section

3HF was a product of Aldrich Chemical Company.Fisetin, b-cyclodextrin (b-CD) and succinyl-(2-hydroxy-propyl)-b-cyclodextrin (SHPb-CD) [degree of substitution(DS) = 4] were purchased from Sigma. Quartz distilledwater was used for preparing the b-cyclodextrin solutions.Concentrated stock solutions of the flavonols were pre-pared in spectrograde methanol (E. Merck) from whichsmall aliquots were added to the b-CD solutions to keepthe concentration of flavonols constant (15 lM).

Steady-state electronic absorption and fluorescencespectra were recorded with a Cecil 7500 spectrophotome-ter, and Hitachi F-4010 and Perkin–Elmer LS-55 spectro-fluorometers, respectively. The fluorescence spectra werecorrected for the wavelength dependence of the sensitivityof the apparatus. The fluorescence anisotropy (r) valueswere obtained using the expression r = (IVV � GIVH)/(IVV + 2GIVH), where IVV and IVH are the vertically andhorizontally polarized components of probe emission withexcitation by vertically polarized light at the respectivewavelength and G is the sensitivity factor of the detectionsystem [24]. Fluorescence lifetime (s) measurements wereperformed with time resolution �1.2–1.4 ns and spectralresolution �30 nm, using an Edinburgh Instruments nano-second time correlated single photon counting setup withnitrogen flash lamp excitation, and a similar setup with a370 nm nanosecond diode laser source (IBH, UK, nan-oLED-07). Data analysis was carried out by a deconvolu-tion method using a non-linear least-square fittingprogramme and fitted with a multi-exponential decay func-tion, F(t) =

PiAi exp(�t/si),

PiAi = 1. The goodness-of-fit

was estimated by using v2 values. All spectral measure-ments were carried out at ambient temperature (298 K).

2.2. Molecular mechanics studies

The calculations were performed for the 1:1 inclusioncomplexes of the flavonols and b-CD using molecularmechanics methods incorporated in HYPERCHEM 7.5[25]. The molecules were built on screen and fully optimizedin vacuo using MM+ force field. No cut-offs were usedand geometry optimization was carried out to an energyconvergence of 0.01 kcal A�1 mol�1 with the Polak-Ribiereconjugate gradient algorithm.

Low energy structures for the 1:1 inclusion complexes ofthe flavonols with b-CD were obtained in vacuo by follow-ing a docking procedure [26,27]. This consisted of translat-ing the guest molecule along the b-CD cavity axis step bystep so as to approach the secondary hydroxyl rim of theb-CD. Geometry optimizations were performed at eachstep in order to achieve the best possible conformationand each of these stochastically generated low energy struc-tures were then grouped so as to identify the nature of theinclusion complex formed with b-CD and the orientation

A. Banerjee, P.K. Sengupta / Chemical Physics Letters 424 (2006) 379–386 381

of the flavonols in the complex. Two distinct inclusion ori-entations of the flavonol molecules were considered, (I)with the chromone ring, and (II) with the phenyl ring,respectively inserted into the b-CD cavity. The stabilitiesof the inclusion complexes formed were judged from theirenergy of formation values (DE) calculated as the differencein the total energies of the complex and those of the freeflavonols (guest) and b-CD (host), respectively. The moststable complex among all the configurations correspondsto the greatest negative value of DE.

Solvation studies were then performed using a combina-tion of MM+ and AMBER force fields in a periodic box ofdimensions 25 A · 25 A · 25 A consisting of 517 watermolecules. The solvation energy (ES) was estimated as thedifference in energies of the hydrated complexes and theisolated molecules.

3. Results and discussion

Fig. 1a shows the effect of increasing concentrations ofSHPb-CD and b-CD (inset), respectively, on the emissionbehavior of 3HF. It can be seen that addition of the b-CDs induces dramatic changes in the emission behaviour

Fig. 1. Fluorescence emission spectra of (a) 3HF (kex = 340 nm) and (b) fiset(0! 5 mM), respectively. Insets display the spectra obtained in b-CD.

of 3HF. While the blue-violet fluorescence band can beassigned to the S1 (pp*)! S0 normal (non-proton-trans-ferred) emission, the large Stokes shifted green fluorescenceband (red shifted by�13 nm in b-CD and�16 nm in SHPb-CD relative to that in aqueous medium) (ref. Table 1) can beassigned to emission from the ESIPT tautomer species (gen-erated by an excited state proton transfer (ESIPT) processoccurring along the internal H-bond (i.e. C(4)@O� � �HOAC(3)) of the molecules) [4]. It may be noted that the ESIPTtautomer emission is highly sensitive to external hydrogen-bonding perturbation, which can compete with the intramo-lecular H-bond formation, leading to decrease in tautomeremission yield [5]. Thus the remarkably enhanced ESIPTtautomer fluorescence observed upon addition of theb-CDs indicates incorporation of 3HF into relativelyhydrophobic environments where such perturbations areminimized.

Fig. 1b demonstrates the changes in the emission behav-iour of fisetin on addition of the b-CDs. In aqueous medium,the fluorescence spectra of fisetin exhibits strong overlapbetween the normal and tautomer emission bands [12]. Onaddition of the b-CDs, conspicuous dual fluorescencebehavior is noted, with the appearance of a strong green

in (kex = 360 nm) in the presence of various concentrations of SHPb-CD

Table 1Steady-state fluorescence emission parameters of flavonols in SHPb-CD and b-CD

Probe Medium kabsmax kem

max

(blue-violet)nm

kemmax

(green)nm

Fluorescenceanisotropy (r)

II/III Bindingconstant,K (M�1)(298 K)

Gibbs free energy,DG0 (kcal/mol)

3HF Water 341.4 409 508 0.06a 2.86 – –b-CD 341.7 406.5 521.5 0.12a 4.02 496.26 �3.67SHPb-CD 343.2 406.5 524.5 0.13a 4.66 1523.14 �4.32

Fisetin Water 357 480 522 (sh) 0.03b 0.72 – –b-CD 361.3 463.5 524.5 0.19b 1.27 622.56 �3.80SHPb-CD 360.9 453 531 0.18b 2.57 2309.13 �4.57

a kex = 340 nm. kem = 525 nm.b kex = 360 nm, kem = 550 nm.

382 A. Banerjee, P.K. Sengupta / Chemical Physics Letters 424 (2006) 379–386

tautomer fluorescence band and a blue-violet normal emis-sion. Interestingly, the normal fluorescence of fisetin (whichhas significant CT character, resulting in prominent solventdipolar relaxation effects, and consequently, utility as apolarity monitor [12,23]) exhibits large blue shifts in thepresence of the b-CDs (�17 nm in b-CD and �27 nm inSHPb-CD) (see Table 1). The enhanced tautomer emissionas well as the strongly blue shifted normal fluorescenceband, indicate that the guest (fisetin) molecules experience

Fig. 2. Plots showing the intensities of the green (II) (d) and blue-violet (III) (sb-CD concentrations.

relatively hydrophobic environments, in the presence of b-CDs.

Fig. 2 shows the intensities of the green (II) and blue-violetemission (III) bands plotted against the b-CD concentra-tion. The tautomer emission yields are notably higher inSHPb-CD as compared to b-CD indicating that the aver-age environment of the encaged flavonols is more hydro-phobic in the former case. The relatively modest increasein III may be presumably due to decrease in non-radiative

) fluorescence bands of 3HF (a and b) and fisetin (c and d) as a function of

A. Banerjee, P.K. Sengupta / Chemical Physics Letters 424 (2006) 379–386 383

deactivation rates within the confined b-CD cavities. Theratio of the intensities of the two bands (II/III) (listed inTable 1) is an especially useful parameter for monitoringthe enhancement in relative yield of the ESIPT tautomeremission and provides a convenient indicator of the hydro-phobicity of the microenvironment of flavonols as empha-sized in our previous studies [8,11]. This ratio increasesdramatically on incorporation into the b-CDs.

For a 1:1 complex formation between fluorescent guestmolecule and b-CDs the binding constant can be obtainedfrom the fluorescence data by using the modified Benesi–Hildebrand equation [28]:

1=DF ¼ 1=DF maxK½b-CD� þ 1=DF max ð1Þwhere DF = Fx � F0, Fx and F0 represent the fluorescenceintensities of the flavonols in the presence and absence oftotal added b-CD concentrations, respectively. DFmax isthe maximum change in fluorescence intensity and K isthe binding constant for the 1:1 complex.

Plots of 1/DF vs. 1/[SHPb-CD] and 1/[b-CD] concentra-tions displayed in Fig. 3 show good linearity. This indicatesformation of inclusion complexes between the host (b-CD)and the guests (3HF and fisetin) with a stoichiometry of 1:1with both substituted and native b-CDs. The binding con-

Fig. 3. Double reciprocal plots for 3HF (a and b) and fisetin (

stants estimated from the ESIPT tautomer emission inten-sities (ref. Table 1) indicate that the binding affinities of theflavonols in the SHPb-CD system are significantly higherthan that in the native b-CDs. Negative values of the Gibbsfree energy change (DG0) indicate spontaneous binding ofthe flavonols with the b-CDs.

Fig. 4 presents typical fluorescence anisotropy (r) data forthe ESIPT tautomer fluorescence of 3HF. Interestingly,upon incorporation in the b-CDs, appreciably high anisot-ropy values are obtained (ref. Table 1) for both the flavo-nols. This indicates that the flavonol molecules areconfined in the motionally constrained cavities of b-CDwhere rotational diffusions are reduced. In order to ensurethat the observed anisotropies are free from lifetime inducedartifacts, the rotational correlational times (hr) were esti-mated from a hydrodynamic model [24] according to which,

hr ¼ gV =kBT ð2Þwhere V is the molecular volume of the probe, g is the vis-cosity of the microenvironment, kB is the Boltzmann con-stant and T is the absolute temperature. From themolecular volumes (�663.8 and �733.5 A3 for 3HF andfisetin, respectively as estimated from their dimensions)and the effective microviscosity of the b-CD cavity [29],

c and d) complexed to SHPb-CD and b-CD, respectively.

Fig. 4. Variation in fluorescence anisotropy of 3HF (m) (kex = 340 nm,kem = 525 nm) and fisetin (d) (kex = 360 nm, kem = 550 nm) with increas-ing concentrations of (a) SHPb-CD and (b) b-CD, respectively.

Table 3Energy of formation values (DE) and the solvation energies of theFlavonol-b-CD inclusion complexes, as obtained by molecular mechanicscalculations

GuestFlavonol

Orientationof guest

DE (kcal/mol) Solvation energy,ES (kcal/mol)

3HF I �26.86 3.69II �18.99 18.29

Fisetin I �24.63 �6.79II �19.03 �0.44

384 A. Banerjee, P.K. Sengupta / Chemical Physics Letters 424 (2006) 379–386

the rotational correlation times were calculated as �645and �710 ps for 3HF and fisetin, respectively. As these val-ues are comparable to the observed fluorescence lifetimes,lifetime induced artifacts could be ruled out. The substan-tial difference in the anisotropy values between the CDcomplexes of the two flavonols (for which the differencein mass and size are not so significant) is noteworthy. Thismay be related to the presence of additional OH groups infisetin, which could significantly affect the structure of thelocal solvation shell, H-bonding characteristics of the fla-vonol and geometrical arrangement of the guest inside

Table 2Fluorescence decay parameters of 3HF and fisetin in SHPb-CD and b-CD

Probe Medium s1 (n

3HF (kex = 337 nm, kem = 550 nm) b-CD 0.261SHPb-CD 0.634

Fisetin (kex = 370 nm, kem = 550 nm) b-CD 0.96SHPb-CD 0.881

the b-CD cavities, thereby increasing its anisotropy. Thisexplanation, however, requires additional confirmationfrom further studies e.g. by NMR.

We also carried out fluorescence lifetime measurementsto examine the excited state dynamics of the flavonolstrapped in the b-CDs. It is evident from such studies (seeTable 2) that in the presence of b-CDs the emission is pre-dominantly from a sub-nanosecond component (decaytime �261, 634, 960 and 881 ps for 3HF-b-CD, 3HF-SHPb-CD, fisetin-b-CD and fisetin-SHPb-CD, respec-tively), which we attribute to the ESIPT tautomer speciesof the flavonols encapsulated in the nano-cavities of theb-CDs. It is clearly evident that such behavior contrastssharply with decay characteristics obtained in aqueousmedium (e.g. s1 � 433 ps; A1 = 0.41, s2 � 5.516 ns; A2 =0.59 for 3HF, the chromophoric moiety of flavonols [30]).The long lifetime component (s = 3.36 ns (A = 0.21)) notedin case of fisetin in the presence of b-CD could presumablyarise from uncomplexed fisetin molecules. However, suchan assignment would require further confirmation fromdetailed analyses of the emission decays (obtained usingbetter spectral resolution) for different flavonol:CD con-centration ratio and other relevant parameters.

Molecular mechanics calculations were performed inorder to gain insight into the thermodynamic and struc-tural features of the 1:1 complexes of the flavonols withb-CD. Judging from the energy of formation values ofthe inclusion complexes formed and their solvation ener-gies (see Table 3), it is evident that the inclusion complexeswith the chromone ring inserted into the b-CD cavity andthe phenyl ring exposed (complexes I) are more stable(compared to complex II, where the phenyl ring is insidethe cavity) in case of both the flavonols (ref. Fig. 5). Suchan arrangement of the flavonols inside the hydrophobic b-CD cavity, where the chromone ring is favorably shieldedfrom the bulk water could explain the enhanced tautomer

s) A1 s2 (ns) A2 v2

0.97 2.64 0.03 1.1211.0 – – 0.563

0.79 3.36 0.21 1.051.0 – – 0.567

Fig. 5. Low energy structures calculated for the complexes between b-CD and (a) 3HF and (b) fisetin.

A. Banerjee, P.K. Sengupta / Chemical Physics Letters 424 (2006) 379–386 385

emission (for both 3HF and fisetin) as well as blue shifts inthe normal emission (for fisetin).

Promising extensions of our findings can be foreseen inthe analyses of therapeutically active flavonols encapsu-lated in cyclodextrin based drug delivery vehicles, whichmight be potentially useful to the pharmaceutical industry.

Acknowledgements

A.B. thanks the Council for Scientific and Industrial Re-search (CSIR), India for Senior Research fellowship (CSIRGrant Award No. 9/489 (44)/2002-EMR-I). We thankProf. Nitin Chattopadhyay, Jadavpur University, andProf. S. Basak of our institute for use of the fluorescencelifetime setups. We are greatly indebted to the reviewerfor valuable suggestions and useful comments.

References

[1] St. Rusznyak, A. Szent-Gyorgyi, Nature 138 (1936) 27.[2] S.W. Lamson, M.S. Brignall, Altern. Med. Rev. 5 (2000) 196.[3] F.A.A. van Acker, O. Schouten, G.R.M.M. Haenen, W.J.F. van der

Vijgh, A. Bast, FEBS Lett. 473 (2000) 145.[4] P.K. Sengupta, M. Kasha, Chem. Phys. Lett. 68 (1979) 382.[5] M. Kasha, J. Chem. Soc., Faraday Trans II 82 (1986) 2379.[6] A.P. Demchenko, S. Ercelen, A.D. Roshal, A.S. Klymchenko, Polish

J. Chem. 76 (2002) 1287.[7] J. Guharay, B. Sengupta, P.K. Sengupta, Proteins: Struct. Funct.

Genet. 43 (2001) 75, and references cited therein.[8] B. Sengupta, A. Banerjee, P.K. Sengupta, J. Photochem. Photobiol.

B: Biol. 80 (2005) 79.[9] B. Sengupta, A. Banerjee, P.K. Sengupta, FEBS Lett. 570 (2004) 77,

and references cited therein.[10] A.S. Klymchenko, G. Duportail, A.P. Demchenko, Y. Mely,

Biophys. J. 86 (2004) 2929.[11] M. Sarkar, P.K. Sengupta, Chem. Phys. Lett. 179 (1991) 68.

386 A. Banerjee, P.K. Sengupta / Chemical Physics Letters 424 (2006) 379–386

[12] J. Guharay, S.M. Dennison, P.K. Sengupta, Spectrochim. Acta A 55(1999) 1091.

[13] G. Krishnamoorthy, S.K. Dogra, J. Phys. Chem. A 104 (2000) 2542.[14] S.L. Wang, T.W. Yeh, T.I. Ho, Chem. Phys. Lett. 418 (2006) 397.[15] T. Loftsson, M. Masson, H.H. Sigurdsson, Int. J. Pharm. 232 (2002)

35.[16] K. Miyake, H. Arima, F. Hirayama, M. Yamamoto, T. Horikawa,

H. Sumiyoshi, S. Noda, K. Uekama, Pharm. Dev. Technol. 5(2000) 399.

[17] A. Banerjee, B. Sengupta, S. Chaudhuri, K. Basu, P.K. Sengupta, J.Mol. Struct. (in press) Available online from 29 March 2006.

[18] A. Douhal, T. Fiebig, M. Chachisvilis, A.H. Zewail, J. Phys. Chem. A102 (1998) 1657.

[19] M. Christoff, L.T. Okano, C. Bohne, J. Photochem. Photobiol. A:Chem. 134 (2000) 169.

[20] R. Ficarra, S. Tommasini, D. Raneri, M.L. Calabro, M.R. Di Bella,C. Rustichelli, M.C. Gamberini, P. Ficarra, J. Pharm. Biomed. Anal.29 (2002) 1005.

[21] D. Haiyun, C. Jianbin, Z. Guomei, S. Shaomin, P. Jinhao, Spectro-chim. Acta A 59 (2003) 3421.

[22] Y. Zheng, I.S. Haworth, Z. Zuo, M.S.S. Chow, A.H.L. Chow, J.Pharm. Sci. 94 (2005) 1079.

[23] A. Sytnik, D. Gormin, M. Kasha, Proc. Natl. Acad. Sci. 91 (1994)11968.

[24] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, second edn.,Plenum Press, New York, 1999 (Chapter 10).

[25] Hyperchem, Hypercube Inc., USA, 2002.[26] D.J. Barbiric, E.A. Castro, R.H. de Rossi, J. Mol. Struct. (Theo-

chem.) 532 (2000) 171.[27] L.M.A. Pinto, M.B. de Jesus, E. de Paula, A.C.S. Lino, J.B. Alderete,

H.A. Duarte, Y. Takahata, J. Mol. Struct. (Theochem.) 678 (2004) 63.[28] H.A. Benesi, J.H. Hildebrand, J. Am. Chem. Soc. 71 (1949) 2073.[29] D. Zhong, A. Douhal, A.H. Zewail, Proc. Natl. Acad. Sci. 97 (2000)

14056.[30] B. Sengupta, P.K. Sengupta, Biopolymers (Biospectroscopy) 72

(2003) 427.