microenvironment effects on the excited state properties of psoralens: a clue to their...

Photochemistry and Photobiology, 1996, 63(2): 182-186

Microenvironment Effects on the Excited State Properties of Psoralens: a Clue to Their Photobiological Activity

Cristina Sousa and Teresa Sa e Melo* Centro de Quimica Fisica Molecular, lnstituto Superior Tecnico, Lisboa, Portugal

Received 7 June 1995; accepted 1 November 1995

ABSTRACT

The singlet and triplet excited state parameters (af, T~

and aT) of psoralen (PSO) and derivatives 4,6,4'-trimethylangelicin (TMA) and 4,5',8-trimethylpsoralen (TMP) show an extreme sensitivity to solvation in dioxanelwater mixtures. These effects are attributed to the variation of the SI So internal conversion rate constant ki,, which is the nonradiative deactivation path dominating their photophysical behavior. Depending on the compound, ki, is very high, (-1 x 10'" s-l) in nonpolar solvents and then decreases to a low value (3 X los s-l), with increasing solvent polarity. This work shows that dioxanelwater mixtures display the same solvent-induced changes in the electronic structure of psoralens during solvation as those induced by the biological microenvironment sensed by the drug's localization. This mixture matches the photophysical parameters of psoralens observed in protic and aprotic pure solvents, in micelles, in liposomes and in human serum low-density lipoproteins (LDL). They can be used to probe the solvating ability of the interaction site in macrocyclic hosts. A particular localization site, i.e. the more (TMA and TMP) or less (PSO) lipophilic sites found when in interaction with LDL, determines the amount of the triplet reactive state of psoralens and the molecular mechanism available for photoreaction: oxic (type I and type 11) or anoxic (type HI) path- ways.

INTRODUCTION

Psoralen (PSO)? and the derivatives, 5-methoxypsoralen (5MOP), 8-methoxypsoralen (SMOP) and 4,5',8-trimethylpsoralen (TMP), currently used in PSO plus UVA (PUVA)

~~

*To whom correspondence should be addressed at: Centro de Qui- mica Fisica Molecular, Comp I, Instituto Superior Tecnico, Av. Rovisco Pais, 1096 Lisboa Codex, Portugal. FAX: 351-1-352 43 72.

tAbbreviations: DPPC, L-a-dipalmitoylphosphatidylcholine; DTAC, dodecyltrimethylarnmonium chloride; Dx, dioxane; em, triplet- triplet molar extinction coefficient; LDL, low-density lipoproteins; 5MOP, 5-rnethoxypsoralen; 8MOP, 8-methoxypsoralen; at, fluorescence quantum yield; QT, triplet formation quantum yield; PSO, psoralen; PUVA, psoralen plus UVA; SDS, sodium dodecyl sulfate; SPC, single photon counting; T , , fluorescence lifetime; TMA, 4,6,4'-trimethylangelicin; TMP, 4,5',8-trimethylpsoralen.

0 1996 American Society for Photobiology 0031.8655196 $5.00+0.00

therapy for psoriasis are potent photosensitizing drugs to mammalian cells in culture (1). Their antiproliferative action is based on the photochemical reactions from their triplet excited state within the microenvironment of the human cells (2). DNA (3) , proteins (4) and fatty acids ( 5 ) have already been identified in vivo and in vitro, as possible biological targets. Historically, DNA is the researcher's first choice, due to the emphasis of PSO-DNA interstrand cross-linking formation on DNA repair and cell death. In this way, the intercalation power in DNA prior to UVA irradiation, the capacity for cross-link and/or monoadduct formation and the amount of triplet formation quantum yields have been ex- tensively reported in the last 10 years, for a great number of PSO derivatives (6-8). Recently, DNA-protein cross-linking adducts have been identified in cultured cells, as the molecular mechanism responsible for the observed 4,6,4'-trimethylangelicin (TMA) phototoxicity (9). Whatever efforts we may make, not too much is known about the molecular mechanisms (lo), oxic (type I and type 11) or anoxic (type III), or about the biological targets (nucleus macromolecules or membrane constituents) that are involved in the photobiological and photochemotherapeutic effects of PUVA therapy ( 1 1). The first attempt to correlate the primary photophysical processes with the in vivo phototoxicity activity has been made by the use of triplet yields (2-12) (and also singlet oxygen yields (1 3)) of PSO derivatives in homogeneous solutions. In the light of our recent studies (14,15), some of the above correlations are re-examined. Indeed, the extreme sensitivity to solvation of the singlet and the triplet excited state parameters (fluorescence quantum yield [af], fluorescence lifetime [T~] and triplet formation quantum yield [aT]) of 5MOP (14) and PSO (15) preclude the validity of a com- parative method based on a single solvent determination. In this work, the dioxanelwater mixtures were chosen because they have solvent composition properties that are the best known models to the cell situation. Using dioxane/water mixtures, we have recently described that 5MOP location, when incorporated in human serum low-density lipoproteins (LDL), determines the amount of its reactive triplet state and thus the triplet photoreaction occurring inside the LDL particle (14). In the present work, we have found that TMP and TMA (as 5MOP (14)) are localized in LDL in a more lipophilic interaction site than that found for PSO. We will show that a particular drug localization in LDL determines the amount of their triplet reactive state and the molecular mechanism pathway available for photoreaction. In this

182

Photochemistry and Photobiology, 1996, 63(2) 183

work we use dioxanelwater mixtures to describe the solvent- dependent photophysical parameters of PSO, TMA and TMP in biological media such as micelles (sodium dodecyl sulfate [SDS] and dodecyltrimethylammonium chloride [DTAC]), liposomes (L-a-dipalmitoylphosphatidylcholine [DPPC]) and human serum LDL. The singlet and triplet excited state parameters, determined for each compound in protic and aprotic pure solvents and in different biological systems, were found to be interrelated showing a localization in a single dielectric region of the dioxane/water mixtures.

MATERIALS AND METHODS Psoralen and TMP was purchased from Sigma and used without further purification. The PSO derivative TMA was kindly supplied by Dr. Dall’ Acyua from Padue. Spectroscopic grade “Uvasol” cyclohexane and ethanol were purchased from Merck. I ,4-Dioxane, methanol, ethanol and formamide were purchased from Riedel- detiaen, and BDH supplied spectroscopic grade acetonitrile. Pure water was obtained from deionized tap water that was first distilled and then evaporated in a quartz still. The human serum LDL was kindly separated by ultracentrifugation by Dr. G. Morais and co- workers in the Faculdade de CiEncias BiomCdicas (Lisboa), accord- ing to a method described elsewhere (16). Sodium dodecyl sulfate and DTAC were provided by Sigma and by Eastman, respectively. The lipid DPPC was a Sigma product.

Fluorescence spectroscopy was carried out at 22°C with a SPEX Fluorolog 2 fluorometer with correction for excitation and emission spectra. Fluorescence quantum yields in pure organic solvents and in dioxaneiwater mixtures were determined using 5MOP in ethanol or in 25% vol/vol water in dioxane, as a reference ((Dl = 0.0255), and taking into account the change in refractive index of solvents and mixtures (14). 5-Methoxypsoralen is a better internal reference for PSO emission than the known fluorescence standards which have much higher fluorescence yields and emissions in different spectral regions. Fluorescence lifetimes were performed with a home-made set up (IST), using the time-correlated single photon counting (SPC) technique. The excitation source was a nitrogen-filled coaxial flash lamp, Edinburgh Instruments model 199, and the sample emission was detected with a Philips XP2020Q photomultiplier. Corrections for changes in pulse jitter and intensity were made by alternate col- lection of two samples, one of them being a reference standard, with 10000 counts at the maximum. The reference standard used in the analysis was a solution of TMP in acetonitrile (air equilibrated) using the lifetime of 0.30 ns, previously determined by Lai and co- workers (17). The (DT of PSO in DTAC micelles, in DPPC liposomes and in human serum LDL were determined as previously described (IS), using a value of 10500 M - ’ cm-’ for the PSO triplet-triplet n 1 0 h extinction coefficient, eTT. The QT of TMP in ethanol, acetonitrile and 3-methylpentane (or cyclohexane) were taken from relative values reported by Lai er al. (17). In the present work, a QT = 0.43 was observed for TMP in ethanol. Making use of the above value of eTT, a corrected value of QT = 0.31 for TMP in methanol is estimated, which is higher than that previously reported (aT = 0.093) by Beaumont et nl. (IS). The (DT = 0.21, reported for TMP in benzene (19) is assumed to be a corrected value for a dioxane solution, because both solvents display very close dielectric properties as already described (IS) . In the present work a QT = 0.07 was ohserved for TMA in dioxane, which is slightly different from that reported (0.13) in benzene (20). Again, the discrepancy between the present value and the reported one (20) is due to the different G~~ values used in the determination.

RESULTS The absorption and the fluorescence emission of TMP and PSO, respectively, in dioxanelwater mixtures are shown in Fig. 1. The TMA absorption spectra undergo a red shift and lose vibrational structure with increasing water concentration in the solvent mixture (Fig. I A). The same spectral behavior was also observed for PSO and TMP (not shown), being a

Q, 0 E Q

g v) P a

0.4

0.2

I 0

280 300 320 340 360 380

Wavelength (nm)

350 400 450 500 550 600

Wavelength (nm) Figure 1. (A) Absorption spectra of 45 p M of TMA in cyclohexane (-), dioxane (-.-), 50% (- - -) and 80% (........) vol/vol water in dioxane. (B) Top: Normalized fluorescence spectra (A,,, = 353 nrn) of TMA in dioxane (-) and in 10% (---) vol/vol water in dioxane. Bottom: Normalized fluorescence spectra of PSO in 20% (-) and 80% (-,-) vol/vol water in dioxane. The spectra are normalized to two different values for the sake of clarity.

common polarity-dependent phenomenon in aromatic molecules (14). The TMP and TMA are insoluble in water. The fluorescence wavelength is red-shifted with increasing water content in the dioxane/water mixture (Fig. 1 B), corresponding to an increase of both @, and aT (Table 2). The same photophysical behavior was also observed in the function of the increase in solvent polarity, in going from aprotic to protic pure organic solvents (Fig. 2). Figure 2 shows the T ~ ,

the af and OT of PSO and derivatives TMP and TMA, in the function of increasing water content in dioxanelwater mixtures. It can be observed that the fluorescence lifetimes display a monotonous increase with increasing water concentration in the mixture. This single line plot of the kinetic parameters can thus be correlated with those found when the sensitizers are incorporated in biological model systems such as SDS and DTAC micelles, DPPC liposomes and human serum LDL, as shown in Fig. 2. In addition, the QT of PSO, TMP and TMA were found to behave in the same way, i.e.

184 Cristina Sousa and Teresa Sa e Melo

0.1 0 1 - I

0.00 W 3

c W z k

1

0

0.6

0.4

0.2

0.0

p/ X J

0.0 0.2 0.4 0.6 0.8 1.0

vlv WATER FRACTION

Figure 2. Photophysical parameters (Qfr rf and GT) of PSO, TMP and TMA in dioxanefwater mixtures (*), in pure solvents, benzene (O), acetonitrile (O), methanol (A), formamide (0) and in 50% vol/ vol water in methanol (+). The above photophysical parameters in biological model systems such as DTAC (*) and SDS micelles (+), DPPC liposomes (X) and human serum LDL (*) are also shown.

showing the same shape in dioxanelwater mixtures, as the fluorescence data, T~ and Qf (Fig. 2). Thus, the microenvironment of the drug interaction site can be estimated because a single dielectric region in the solvents mixtures matches the singlet and triplet photophysical parameters observed in LDL. Indeed, a monotonous increase in the above three photophysical parameters was observed for PSO (15), TMP and TMA (Fig. 2), in contrast with the bell-shaped curve already described for 5MOP in dioxanelwater mixtures (14). How- ever, as shown in Fig. 2, the above photophysical data observed in pure organic solvents (aprotic and protic) are found to be interrelated with those in dioxane/water mixtures. This means that pure organic solvents like benzene, acetonitrile, methanol and formamide have dielectric properties analo- gous, respectively, to dioxane, lo%, 30% and 50% vol/vol water in dioxane. Thus, the microenvironment sensed by PSO, TMA and TMP in biological systems should display

Table 1. derivatives

The internal conversion rate constants (s ’) of PSO and

TMA TMP PSO

D i o x a n e 3.9 X lo9 6.8 X 10’ 8.4 X lo’$ Acetonitrile 2.4 x 109* 2.6 x 109 4.3 x 109 Methanol nd 0.8 x 109 2.8 x 109 70% vol/vol H,O:Dx nd 0.3 x 109t 1.2 x 109

*Assuming a value of (DT = 0.15 in acetonitrile. ?Assuming a value GT = 0.40 in 70% vollvol water in dioxane. *Assuming the same T( value of TMP in dioxane (see Fig. 2).

the same dielectric properties, as experimentally observed (Fig. 2). It can be observed that TMP and TMA are incorporated in a nonpolar lipophilic region of LDL, in contrast with PSO, which is located in a water-rich region, at the membrane interface of the LDL particle (Fig. 2). As expected, the same dielectric location was found in DPPC liposomes, i.e. 75% for PSO and 25% vol/vol water in dioxane for TMP and TMA (Fig. 2). Interestingly, it has been recently described (21) that the interface of DPPC liposomes should behave as a dielectric medium with II* = 0.88, the well-known Taft’s empirical polaritylpolarizability parameter (22). In DTAC cationic micelles, all derivatives are located in a dielectric medium analogue to 40% vol/vol water in dioxane. In SDS anionic micelles, TMP and TMA are in a medium like 55% vol/vol water in dioxane (Fig. 2). The same location in SDS micelles has already been found for 5MOP (14).

Table 1 shows the internal conversion rate constant k,,, determined by the expression:

k,, = (1/Tf) - [(@f/7f) + (Qd7f)I (S-I)

The S, 4 So nonradiative transition is the dominant deactivation path of the photophysical behavior of all PSO derivatives. The k,, rate constant is experimentally observed to fall within the range defined by the maximum value, -1 X 10’” s-I, in nonpolar and aprotic solvents, and the minimum value, 3 X lo8 s - I obtained for polar and protic solvents (Table 1). From Table 1, it can be observed that ki, of PSO, TMP and TMA decrease with increasing solvent polarity from dioxane to acetonitrile and from methanol to formamide. Table 2 shows the fluorescence wavelength maximum of TMP as a function of the solvent polarity. An increase in the solvent polarity corresponds to longer fluorescence wavelengths and thus to a decrease in the AE(S1-So) energy

Table 2. The fluorescence wavelength maximum, the fluorescence quantum yields and the triplet formation quantum yields of TMP in different solvents

Solvent

Cyclohexane Dioxane Acetonitrile Methanol Ethanol 70% vollvol H20:Dx

A, (nm)

420 425 433 43 8 435 -

rf (ns)

-

0.11 0.28 0.83 0.67 1.76

@ f @T

0.0008 0.08 0.0032 0.21* 0.0085 0.25 0.0290 0.31 0.0230 0.43 0.0630 0.40

~

* See Knox et al. (19)

Photochemistry and Photobiology, 1996, 63(2) 185

gap. These data are in agreement with a MI* nature of the first singlet excited state of psoralens (17). The molar extinction coefficient for this first absorption band of PSO at 360 nm (23) in cyclohexane (around 1000 M - ’ cm-’) shows increasing values with increasing solvent polarity. A shoul- der at 350 nm is clearly observed in the absorption spectra of TMA in cyclohexane (Fig. IA).

The photophysical parameters observed in the biological model systems have important consequences in the molecular mechanism of the antiproliferative action of PSO derivatives.

DISCUSSION Because both the singlet and triplet excited state parameters present the same shape in dioxane/water mixtures, and these mixtures have similar dielectric properties of pure organic solvents, the photophysical parameters in biological media should display the same dielectric behavior. Dioxane/water mixtures can thus be used as a solvent model for the biological microenvironment sensed by the photosensitizer. This is only true in the absence of any physical or chemical quenching of the singlet and/or triplet excited states. This means that any observed decrease of fluorescence lifetime and/or triplet amount may be assigned to a possible photoreaction event, only if one knows the dielectric region where the above photophysical parameters should fall upon. The present results demonstrate that the triplet formation quantum yield of PSO in LDL is the only parameter that shows an inhibition (aT = 0.20), relative to its expected value in the dielectric region of incorporation (aT = 0.44) ( 1 S), i.e. around 75% vol/vol water in dioxane (Fig. 2). Here we shall recall thst a similar triplet inhibition was already described for SMOP, this time located in a lipophilic region of LDL, around 10% vol/vol water in dioxane (14). This lipophilic region is also found to be the interaction site for TMA and TMP in LDL particles (Fig. 2). Unfortunately, their triplet amounts in LDL are not yet available. However, both derivatives are expected to present a triplet inhibition when incorporated in LDL particles, showing the same triplet reac- tivity with the apoprotein residues as already observed for 5MOP (14).

In nonpolar solvents, the S,(nn*) and the S,(nn*) excited states of PSO are isoenergetic (24), showing the maximum values for the internal conversion rate constants, k , (Table I ) . With increasing solvent polarity, the S,(nn*) state energy is raised and S , 4 So transition is more readily allowed. Thus, an increase in Qf is observed, corresponding to longer fluorescence wavelengths (Table 2). At high solvent polarity and longer fluorescence wavelengths, the singlet to triplet energy gap, AE(S,-T,) decreases, and intersystem crossing is promoted (Fig. 2). Because Qf are very low for all com- pounds, both in polar and in nonpolar solvents (Fig. I ) , the minimum and the maximum solvent-dependent ki, values de- termine, respectively, the maximum and the minimum triplet reactive state amounts.

CONCLUSIONS Dioxanelwater mixtures provide a self-consistent description of the susceptibility of psoralens to solvation. It gives an accurate picture of the solvating ability of the interaction site

in macrocyclic hosts. Because the photophysical parameters (Df, T~ and (DT present the same behavior in these solvent mixtures, they can be used to model the molecular mechanisms of photoreactions from the triplet state. Because the same solvent-dependent photophysical behavior was also observed in cyclohexane/ethanol and acetonitrile/water mixtures (not shown), we believe that this phenomenon is not dependent on specific hydrogen bonding but rather more general and not particular to the molecular system of psoralens. The importance of the hydrophobic environment is underlined by the kinetic results in the presence of micelles, liposomes and human serum LDL. With the present results we may conclude that PSO is able to photoreact at the interface of the cell membranes by an oxygen-dependent photodynamic mechanism pathway. With a higher (DT = 0.44 ( 1 S), PSO is able to produce singlet oxygen by energy trans- fer (type 11) at the cell membrane interface. Indeed, PSO has already been reported to be much more erythemogenic than the PUVA derivatives (SMOP, 8MOP and TMP), preventing its clinical use. In contrast, TMP and TMA (as SMOP) are able to induce oxygen-independent photoreactions producing photoadducts inside the cytoplasmic organelles of the cell, as already observed in vivo (9). In the case of XMOP in dioxane/water mixtures, our preliminary studies (25) have predicted that the observed water-rich interaction site in membranes should be sufficient for a photoreaction from its triplet state to occur. With a much lower QT = 0.04 (M. Bazin, personal communication) value than PSO, in the same dielectric region of incorporation (25), 8MOP should be able to form oxidation photoproducts at cell membranes by an oxic (type I) photodynamic mechanism, as already reported (26).

In conclusion, the dioxane/water mixtures can be used as a solvent model to estimate the dielectrical microenvironment of photosensitizer localization in LDL and to predict the amount of its triplet reactive state. The molecular mechanism pathway available for the triplet photoreaction can thus be estimated from these experimental data.

Acknowledgements-This work was supported by JNICT, with a Ph.D. grant (no. 857/94) for C.S. and by Instituto Superior TCcnico (CQFM) funds. We are grateful to Dr. R. Santus, M. Bazin, Prof. S. B. Costa and P. Coutinho for the flash photolysis facilities and valuable scientific discussions. Thanks are due to Dr. Mapnita for his interest and helpful discussions on the accuracy of fluorescence lifetime measurements. Moreover, we thank Dra. A. T. Reis e Sousa and Dr. J. M. G. Martinho for scientific advice and facilities for the SPC equipment. Trimethylangelicin and human serum LDL were kindly supplied by Dr. Dall’Acqua (Padue) and Dr. G. Morais (Lis- bon), respectively.

REFERENCES Averbeck, D. (1989) Recent advances in psoralen phototoxicity mechanism. Photochem. Photohid. 50, 859-882. Bensasson, R. V., E. J. Land and C. Salet (1978) Triplet excited state of furocoumarins: reaction with nucleic acid bases and aminoacids. Photochem. Photobiol. 27, 273-280. Musajo, L., G. Rodighiero, G. Colombo, V. Torlone and F. Dall’ Acqua ( 1 965) Photosensitizing furocoumarins: interaction with DNA and photo-inactivation of DNA containing viruses. Experientiu 21, 3451-3456. Hornicek, F. J., G. I. Malinin, W. B. Glew, U. Awret, J. V. Garcia and T. P. Nigra (1985) Photochemical cross-linking of erythrocytes ghost proteins in the presence of S-methoxypsoralen and trimethylpsoralen. Photochern. Photobiol. 9, 263-269.

186 Cristina Sousa and Teresa Sa e Melo

5. Midden, W. R. (1988) Chemical mechanisms of bioeffects of furocoumarins: role of reactions with proteins, lipids and other cellular constituents. In Psorulen DNA Photobiology, Vol. 11, pp. 1-50, CRC Press, Boca Raton, FL.

6. Bensasson, R. V., E. J. Land and T. G. Truscott (1992) Excited states and free radicals in biology and medicine. Contributions from flash photolysis and pulse radiolysis. Oxford Science Pub- lications, pp. 265-282.

7. Craw, M., R. V. Bensasson, J. C. Ronfard-Haret, T. SB e Melo and T. G. Truscott (1983) Some photophysical properties of 3- carbethoxypsoralen, 8-methoxypsoralen and 5-methoxypsoralen triplet states. Photochem. Photobiol. 37, 61 1-615.

8. Ronfard-Haret, J. C., D. Averbeck, R. V. Bensasson, E. Bisagni and E. J. Land (1982) Some properties of the triplet excited state of the photosensitizing furocoumarin: 3-carbethoxypsoralen. Photochem. Photobiol. 35, 479-489.

9. Bordin, F., F. Carlassare, L. Busulini and F. Baccichetti (1993) Furocoumarin sensitization induces DNA-protein cross-links. Photochem. Photobiol. 58, 133-1 36.

10. Foote, C. S. (1976) Photosensitized oxidations and singlet oxygen: consequences in biological systems. Free Radiculs Biol. 2, 85-133.

11. Psoralens (1992) Les Nouvelles Dermatologiques 11 Supple- ment (Edited by T. B. Fitzpatrick, P. Forlot, M. A. Pathak, F. Urbach). Intercontinental Medicine, Strasbourg.

12. Sa e Melo, T., D. Averbeck, R. Bensasson, E. Land and C. Salet ( 1 979) Some furocoumarins and analogs: comparison of triplet properties in solution with photobiological activities in yeast. Photochem. Photobiol. 30, 645-65 1.

13. Joshi, P. and M. Pathak (1983) Production of singlet oxygen and superoxide radicals by psoralens and their biological sig- nificance. Biochem. Biophys. Res. Commun. 112, 638-646.

14. SB e Melo, T., A. Maqanita, M. Prieto, M. Bazin, J. Ronfard- Haret and R. Santus (1988) Photophysical behaviour of 5-methoxypsoralen in dioxane-water mixtures. Photochem. Photobiol. 48, 429437.

15. SB e Melo, T., M. Bazin, J. Ronfard-Haret and R. Santus (1993) Quantification of the solvent effects on the triplet quantum yield

of psoralen by the linear solvation energy relationship. Photo- chem. Photobiol. 58, 19-27.

16. Havel, R. J., H. A. Eder and J. Bragdon (1955) The distribution and chemical composition of ultracentrifugally-separated lipoproteins in human serum. J. Clin. Invest. 34, 1345-1353.

17. Lai, T., B. Lim and E. Lim (1982) Photophysical properties of biologically important molecules related to proximity effects: psoralens. J. Am. Chem. SOC. 104, 7631-7635.

18. Beaumont, P., B. Parsons, G. Phillips and J. Allen (1979) A laser flash photolysis study of the reactivities of the triplet states of 8-methoxypsoralen and 4,5’,8-trimethylpsoralen with nucleic acid bases in solution. Biochim. Biophys. Actu 562, 214-221.

19. Knox, C., E. J. Land and T. G. Truscott (1986) Singlet oxygen generation by furocoumarin triplet states. I. Linear furocoumarins (psoralens). Photochem. Photobiol. 43, 359-363.

20. Knox, C., E. J. Land and T. G. Truscott (1988) Triplet state properties and triplet-state-oxygen interactions of some linear and angular furocoumarins. J . Photochem. Photobiol. Biol. 1, 315-321.

21. Tsukamoto, I. and K. Ogli (1992) Anesthetic-membrane interactions evaluated by Taft’s polarity parameters. J. Colloid. In- terface Sci. 154, 17-23.

22. Reichardt, C. (1990) Solvents and Solvent E’ects in Organic Chemistry, 2nd ed. VCH Verlag, Weinheim, Germany.

23. Becker, R. S., S. Chakravorti, C. A. Gartner and M. G. Miguel (1993) Photosensitizers: comprehensive photophysics/photochemistry and theory of coumarins, chromones, their homo- logues and thione analogues. J . Chem. Soc. Faraduy Trans. 89, 1007-1 0 19.

24. Lim, E. C. (1977) Vibronic interactions and luminescence in aromatic molecules with non-bonding electrons. In Excited Stares, Vol. 3 (Edited by E. C. Lirn), pp. 305-337. Academic Press, New York.

25. SB e Melo, T., L. Ribeiro, A. MaGanita, M. Bazin, J. Ronfard- Haret and R. Santus (1990) Photochemical processes of psoralen and derivatives (8MOP and 5MOP) in biological systems. Int. Conference on Luminescence, Lisbon, Portugal, 1&20 July, Abstract 3P-7.

26. Zarebsla, K. (1994) Cell membrane, a target for PUVA therapy. J . Photochem. Photobiol. B Biol. 23, 101-109.

microenvironment effects on the excited state properties of psoralens: a clue to their...

Documents