Origin of Solvent Dependence of Photosensitized Splitting of a

Apr 30, 2009 - ... of visible light to break the cyclobutane ring of the dimer (Figure 1). ..... of the Marcus theory are fully identical to Rose's co...
0 downloads 0 Views 877KB Size
Download PDF
J. Phys. Chem. B 2009, 113, 7205–7210

7205

Origin of Solvent Dependence of Photosensitized Splitting of a Cyclobutane Pyrimidine Dimer by a Covalently Linked Chromophore Wen-Jian Tang,†,‡ Qing-Xiang Guo,† and Qin-Hua Song*,† Department of Chemistry, UniVersity of Science and Technology of China, Hefei 230026, China, and School of Pharmacy, Anhui Medical UniVersity, Hefei 230032, China ReceiVed: July 7, 2008; ReVised Manuscript ReceiVed: March 20, 2009

In model studies involving the mechanisms of DNA photolyases, two reverse solvent effects on the quantum yield of photosensitized splitting of a cyclobutane pyrimidine dimer (CPD) by a covalently linked chromophore have been reported. One is an increase in the splitting efficiency in lower polarity solvents for model compounds with a short linker between the dimer and the chromophore. Another is more efficient splitting in higher polarity solvents for model compounds with a flexible and long linker. To unravel mechanisms of two opposite solvent effects, five covalently linked indole-dimer compounds with different-length linkers were prepared. Two solvent effects as described above were observed through measuring quantum yields of dimer splitting of these model compounds in four solvents. According to Marcus theory, back electron transfer in the splitting reaction was analyzed quantitatively in light of relative data of a model compound in four solvents. It was demonstrated that the dependence of the quantum yield on solvent polarity for the flexible long-linker system would derive from the change in the distance between a dimer unit (acceptor) and an indole moiety (electron donor) in different solvents. With increasing solvent polarity, a U-shaped conformation of the model compound would become a preferred conformation because of the hydrophobic interaction between indole and dimer moiety, and their distances would become closer. On the basis of Marcus theory, calculated results reveal that the rate of back electron transfer would be slowed down with increasing solvent polarity and the distance reduced, giving a more efficient splitting. Meanwhile, some new insights into mechanisms of DNA photoreactivation mediated by photolyases were gained. Introduction The major lesions formed in DNA by UVB radiation (280-320 nm) in sunlight are cyclobutane pyrimidine dimers (CPDs).1 DNA photolyases are flavin-containing repair enzymes which catalyze the efficient repair of cis-syn CPDs by utilizing the energy of visible light to break the cyclobutane ring of the dimer (Figure 1). The enzymes contain a redox active cofactor, flavin adenine dinucleotide (FAD), that operates in the fully reduced and deprotonated form (FADH-), and an auxiliary antenna chromophore. The latter activates the process of repair by absorbing near UV-vis radiation (300-500 nm) and transferring the energy to the flavin. Subsequently, the excited FADH- transfers one electron to the CPD to form the dimer radical anion, which cleaves spontaneously, and then back electron transfer restores the dipyrimidine and the functional form of flavin ready for a new cycle of catalysis.2,3 To unravel the mechanisms sketched above in detail, several model compounds were constructed to mimic the virtually intramolecular electron transfer from the enzyme-bound sensitizer to the enzyme-bound dimer, such as a chromophore attached to a CPD.4-6 These model systems have offered useful insights into electron transfer and bond-breaking processes involved in photosensitized dimer splitting. In the covalently linked chromophore-dimer systems, the photosensitized splitting of the dimer by the attached indole (1),7 arylamine (2),8 or methoxybenzene (3)9 (Chart 1) exhibited a strong solvent * To whom correspondence should be addressed. E-mail: qhsong@ ustc.edu.cn. Phone: +86-551-3607524. Fax: +86-551-3601592. † University of Science and Technology of China. ‡ Anhui Medical University.

Figure 1. UV-induced CPD photoproducts in DNA and their photorepair by DNA photolyases.

dependence with efficient splitting in low polarity solvents. The increase in splitting efficiency in lower polarity solvents has been interpreted in terms of a possible slowing of the highly exothermic back electron transfer due to Marcus inverted behavior. However, in the other two indole-dimer systems (4,10 511 in Chart 2) with long linkers, the solvent dependence of the quantum yield showed a reverse trend, which was more efficient splitting in higher polarity solvents. The reverse solvent dependence was also observed in covalently linked flavin-dimer systems (6 in Chart 2) by Carell et al.12 Through two sets of measurements in water/ethylene glycol mixtures and various organic solvents, they observed an increase in splitting efficiencies in higher polarity solvents. In a previous paper, we prepared covalently linked tryptophan-dimer compounds 7 (Chart 2) and investigated the photosensitized splitting of the thymine dimer by the attached tryptophan in various solvents.13 The quantum yield of dimer splitting was very sensitive to the polarity of solvents. Remarkably increased splitting efficiencies were observed in high

10.1021/jp805965e CCC: $40.75  2009 American Chemical Society Published on Web 04/30/2009

7206

J. Phys. Chem. B, Vol. 113, No. 20, 2009

Tang et al.

CHART 1: Model Compounds with a Short Linker

CHART 2: Model Compounds with a Relatively Long and Flexible Linker

polarity solvents. Thus, this observation was right, contrary to the earlier evaluations7-9 obtained from indole-dimer systems 1-3 but in agreement with the results from other indole-dimer systems 410 and 511 and flavin-dimer systems (6).12 By comparing these model compounds, we suggested that the length of linker may play an important role in the dependence of the splitting efficiency on solvent polarity. The length of linker, however, does not reflect accurately the center-to-center distance (RDA), which is a factor of determining the rate of electron transfer and not just the length of linker between electron donor (indole moiety) and acceptor (dimer unit), especially for the model systems with a long and flexible linker. To further explore the origin of solvent dependence of the dimer-splitting efficiency on the length of the linker, five model compounds, 8-12, with different-length linkers have been prepared and investigated in this work. Using compound 7a as an example, experimental results have been analyzed according to Marcus theory.14 For model systems with a long and flexible linker, the distance between the dimer and a chromophore moiety would change largely in different-polarity solvents, leading to change in the rate of electron transfer and giving different quantum yields. Experimental Details General Methods. Melting points were uncorrected. All materials were obtained from commercial suppliers and used as received. Solvents of technical quality were distilled prior to use. DMF was dried overnight with MgSO4 and distilled. Acetonitrile and methanol were spectroscopic grade from commercial suppliers and used without further purification.

Measurement of Steady-State Fluorescence Emission. Fluorescence emission spectra were measured at room temperature on a luminescence spectrometer. To determine the extent of fluorescence quenching, Q, fluorescence intensities (Findol-D) of 8, 9, 10, 11, or 12 were compared to that (Findol) of the corresponding indole without a dimer attached, respectively, that is, Q ) 1 - Findol-D/Findol. The concentrations of indole moiety of the indol-dimer models and the free indole were controlled within 0.05 for absorbance at the wavelength of excitation of 295 nm, and fluorescence intensities were normalized with the absorbances. Measurements of Splitting Quantum Yields of Compounds 8-12. The 3 mL samples in quartz cuvettes with a Teflon stopper were irradiated with 295 nm light from a fluorescence spectrometer with a 10 nm slit. The exent of dimer splitting was monitored by the increase in absorbance at about 270 nm due to the regeneration of the 5,6-double bonds of the pyrimidines. After certain time intervals, the absorbance of the irradiated solutions was recorded by a UV-vis spectrometer. The quantum yields of splitting did not change with and without N2 bubbling prior to irradiation within an experiment error of (5%. Hence, the nondeaerated solution was employed in all measurements of quantum yield. Since the photoproducts, the pyrimidines, and indole moiety absorb light at 295 nm, to reduce competition of absorbing the irradiated light between the model compounds and the photoproducts, the splitting extent of model compounds was controlled within 5% in all of the measurements. To obtain the quantum yields for dimer splitting of the model compounds, [Φ ) (rate of dimer splitting)/(rate of photons absorbed)], the absorbance at about 270 nm (A270, the absorption peak of the pyrimidine) and 295 nm (A295) was measured at certain irradiated time intervals. The A270 change (∆A270) of the solution depends on the splitting extent of the model compounds. As the model compounds were fully splitting to form products, the change in mole extinction coefficients (∆ε270) was obtained from UV absorption spectra of the model compounds and the splitting products. The splitting concentration (cspl) of the model compound was obtained from ∆A270/∆ε270. The plot of cspl against the irradiation time (t, min) is well-fitted as a straight line. The slope of the straight line, B, is a splitting rate of the model compound. The intensity of the incident light I0 (unit: einstein min-1) was measured using ferrioxalate actinometry.15 The intensity of light absorbed (Ia) by solution was calculated in term of Beer’s law, Ia ) I0 (1 - 10-A295). These values allowed the calculation of the quantum yield, Φ ) BV0/Ia, wherein V0 was the volume of irradiation solution, 3 × 10-3 L. Results and Discussion The photophysical and photochemical processes of model compounds are illuminated with a simple mechanistic scheme (Figure 2). Since the dimer has no significant absorption under above 290 nm light, the chromophore unit of a model

Origin of Solvent Effects on CPD Repair

J. Phys. Chem. B, Vol. 113, No. 20, 2009 7207 TABLE 1: Dependence of the Splitting Quantum Yield of the Dimer Splitting (Φ) on Solvents for Model Compounds 8-12 solvent

8

9

10

11

12

7aa

THF acetonitrile methanol H2O/CH3CN (90:10)

0.33 0.26 0.24 0.18

0.20 0.19 0.18 0.12

0.13 0.14 0.12 0.07

0.12 0.16 0.23 0.21

0.04 0.06 0.10 0.13

0.016 0.041 0.072 0.093b

a

Figure 2. Photophysical and photochemical processes of model compounds.

CHART 3: Five Synthesized Model Compounds with Different-Length Linkers

compound absorbs a photon producing the excited state of the chromophore (1Ch*-D). The excited state may have relaxation pathways as follows: fluorescence (kf), internal conversion (kic), and electron transfer to the linked dimer (kfet). The charge-separated species (Ch•+-D•-) formed via the electron transfer can undergo two processes, CPD splitting (monomerization, kspl) or back electron transfer (kbet), resulting in an unproductive reversal. These processes undoubtedly have counterparts in the enzymatic repair process. There are two pairs of competitions in these processes: (1) the photophysical process (kf and kic) and electron transfer (kfet) and (2) the splitting (kspl) and back electron transfer (kbet). In the two pairs of competitions, kfet and kspl contribute to the observed quantum yield (Φ) of dimer splitting, Φ ) φfetφspl, while kbet, kf, and kic reduce the efficiency of dimer splitting. It has been widely accepted that back electron transfer leads to low splitting efficiencies in model systems and is efficiently suppressed in DNA photolyases achieving a high repair efficiency (Φ ) 0.7-0.98).3 Besides back electron transfer in flavin model systems, internal conversion (kic) plays a more important role for leading to a low splitting efficiency.16 To investigate the dependence of these processes on solvents, five model compounds with different-length linkers 8-12 were synthesized, shown in Chart 3; their covalent-bond numbers linking the dimer and indole moiety are 3-7, respectively, and these linkers are not rigid. The splitting properties of these model compounds were investigated in four solvents. Through irradiating model compounds in solutions under 295 nm monochromatic light, analysis of the photolysis mixture by reversed-phase HPLC confirmed that only photosensitized splitting of the dimer unit occurs, as no other products could be detected except the expected photoproducts. To measure the quantum yields (Φ) for the splitting of the model compounds, all sample solutions were prepared in

From ref 13. b From neat water.

cuvettes with a Teflon stopper and then irradiated with 295 nm light from a fluorescence spectrometer. After certain time intervals, the absorption spectra of the irradiated solution were recorded by a UV-vis spectrometer. The intensity of irradiation light was measured three times during the one-sample measurement, and an average of three measurements was employed. On the basis of these data, the quantum yields of splitting of model compounds were obtained, listed in Table 1. Obviously, data in Table 1 reveal two reverse solvent effects on the quantum yields. The quantum yields for model compounds with short linkers, 8, 9, and 10, show that one solvent effect is a decrease in the quantum yield as solvent polarity increases. Another reverse solvent effect is a change to a more polar solvent, causing an increase in values of Φ for the longlinker compounds, 11, 12, and 7a.13 Namely, the splitting efficiency of model compounds with different-length linkers could reveal different solvent behavior. In our previous paper, reverse solvent dependence of the splitting efficiency for model compounds with different-length linkers has been explained according to Marcus theory. On the rate of back electron transfer in the charge-separated state (Ch•+-D•-), effects of the linker length were analyzed qualitatively, and the calculation was not performed.13 According to Marcus theory,14 the rate of electron transfer is expressed by

ket ) A' exp(-G+ ⁄ kBT) +

(1)

G ) (∆G + λs) ⁄ 4λs 2

(2)

The equations allow the evaluation of the rate of back electron transfer in the photosensitized splitting reaction of the model compounds. The free energy of activation (G+) can be obtained from the change of free energy (∆G) and the solvent reorganization energy (λs) for back electron transfer. The former is the energy level of the charge-separated state, which can be estimated by using thermodynamic redox potentials. The free energy difference between the charge transfer state and the ground state is given by eq 3,17

-∆Gbet ) Eox - Ered + ∆Ecoul ∆Ecoul (eV) )

e 1 2 4πε0a ε 37.5

(

(3)

)

(4)

where Eox and Ered are the one-electron oxidation potential of a donor (tryptophan residue, 1.0 V (NHE)18) and the one-electron reduction potential of an acceptor (dimer, -2.2 V (saturated calomel electrode)19), respectively. ∆Ecoul and ∆E0,0 are the coulomb term and the energy level of the excited state, respectively. The energy of the excited state can be obtained from the fluorescence peaks of indole-dimer model systems. ε and a are the static dielectric constant of a solvent and the centerto-center distance between a donor and an acceptor, respectively, and ε0 ) 8.854 × 10-12C V-1 m-1.

7208

J. Phys. Chem. B, Vol. 113, No. 20, 2009

Tang et al.

Figure 3. (A) X-ray crystal structure of 7a.21 Oak Ridge thermal ellipsoid plots of the molecular structures. Displacement ellipsoids are shown at the 50% probability level. (B) Crystal packing of compound 7a.

TABLE 2: Calculated Values for Solvent Reorganization Energy (λs), Changes of Free Energy (∆Gbet), and Activation Energy (G+bet) for Back Electron Transfer in Compound 7a (Unit: eV) solvent

εsa

εopa

∆Gbet

λs

G+bet

RDA, Å

∆Gbet′

λ s′

G+bet′

THF acetonitrile methanol water

7.58 37.5 32.63 78.54

1.980 1.807 1.766 1.777

-3.15 -2.90 -2.90 -2.86

0.92 1.30 1.32 1.36

1.34 0.49 0.47 0.42

10 5 4.5 4

-3.07 -2.88 -2.89 -2.81

1.28 1.05 0.90 0.70

0.63 0.80 1.11 1.60

a

From ref 22.

The solvent reorganization energy of back electron transfer (λs) can be estimated using the equation20 as follows:

λs ) (e ⁄ 4πε0)[(2rD)-1 + (2rA)-1 - (RDA)-1](εop-1 - εs-1)

(5) where rD and rA are the ionic radii of the donor and the acceptor, respectively, and RDA is the distance between a donor and an acceptor, that is, a in eq 4. εop and εs are the optical and static dielectric constants of the solvent, respectively, with εop ≈ n2, n being the solvent refractive index. As an example, solvent effects on back electron transfer in the dimer splitting of compound 7a could be analyzed quantitatively in terms of the above equations. The X-ray crystal structure21 of compound 7a is shown in Figure 3A. The values of rD, rA, and RDA can be obtained from the structure; they are 2.2 Å, 4.5 Å, and 3.6 Å, respectively. Using the values of rD and rA and if, in solution, RDA ) 6 Å, ∆Gbet and λs can be estimated in four solvents, as listed in Table 2. The values of -∆Gbet decrease, and those of λs increase, with increasing solvent polarity. The former is much higher than the latter for all four solvents, that is, the back electron-transfer reaction lies in the so-called Marcus inverted region. In terms of eq 2, the values of the free energies of activation, G+bet, were calculated and decreased with increasing solvent polarity. Thus, according to eq 1, the rate of back electron transfer, kbet, would increase to lead to a decrease in the splitting quantum yield with increasing solvent polarity. The expectation is just right contrary to the experimental results from compound 7a,13 but it is in accord with the observations from the model systems with short linkers, 1-37-9 and 8-10. Furthermore, the same solvent dependence of G+bet would be deduced from the calculations with any certain value of RDA in a possible range 3.5-10 Å. In other words, when RDA is a constant, the expectations in terms of calculated results are fully identical to Rose et al.’s observations and interpretation to the short-spacer systems. The driving force for back electron transfer in low polar media would become so exothermic as to lie in the Marcus inverted region, where it would be significantly slowed down. A fast and unproductive charge recombination

by back electron transfer within the charge-separated species can be suppressed in low polarity solvents.7-9 However, the distance RDA is not constant in different solvents, especially the model systems with a long and flexible linker like compound 7a. As showed in Figure 3A, the crystal structure is a U-shaped conformation of compound 7a, whose single crystal was obtained from methanol solution, and the distance RDA would be much shorter than the length of linker connected the dimer and indole moiety. If the U-shaped conformation of the model compound also exists in a solution, then the distance of electron transfer would become shorter because of a through-space rather than through-bond pathway. The rate of electron transfer depends strongly on the distance, RDA, and the distance is determined by the molecular conformation of the model compound in the solution. In contrast with the model systems with a long and flexible linker, the U-shaped conformation would be formed with difficulty for the short-linker systems because of intramolecular strain. In other words, the change of the distance would be small for model systems with a short linker in different solvents. This small change in the distance would not play an important role in the solvent effects on the quantum yield of splitting. Hence, short-linker model systems would give similar acceptor-donor distances in different solvents, and the solvent dependence on the splitting quantum yield can be well explained in terms of Marcus theory. The cis-syn cyclobutane thymine dimer is a specific structure with an asymmetric polarity, in which the cyclobutane ring is hydrophobic and the opposite edges of the thymine bases have nitrogens and oxygens capable of forming hydrogen bonds. Two types of regions in the crystal packing of 7a show also this specific structure of the dimer. One is the hydrophobic region including the cyclobutane ring, indole, and pentyl, and the other is the H-bond region, forming between N-H and oxygens of pyrimidine bases and solvent methanol, shown in Figure 3B.21 Although it is formed in its single crystal, this U-shaped conformation of compound 7a would also be a preferred conformation in polar solvents because of the specific asymmetric polarity and hydrophobic interaction between the cyclobutane ring and the indole moiety. With increasing solvent

Origin of Solvent Effects on CPD Repair

J. Phys. Chem. B, Vol. 113, No. 20, 2009 7209

Figure 4. Fluorescence emission spectra of compound 7a and N-acetyl tryptophan methyl ester (Trp) in four solvents, excitation wavelength λex ) 295 nm.

TABLE 3: Dependence of the Fluorescence Quenching Efficiency (Q) and the Splitting Efficiency of the Dimer Radical Anion (Ospl) on Solvents for Model Compounds 8-12 solvent

8

9

10

11

12

7aa

THF acetonitrile methanol H2O/CH3CN (90:10)

0.95 (0.35) 0.98 (0.26) 0.99 (0.24) 0.99 (0.18)

0.68 (0.29) 0.82 (0.23) 0.89 (0.20) 0.85 (0.14)

0.64 (0.20) 0.92 (0.15) 0.68 (0.17) 0.86 (0.08)

0.63 (0.19) 0.69 (0.23) 0.85 (0.27) 0.94 (0.22)

0.52 (0.07) 0.59 (0.10) 0.70 (0.14) 0.82 (0.16)

0.36 (0.04) 0.48 (0.08) 0.65 (0.11) 0.81b(0.12)

a

From ref 13. b From neat water.

polarity, the proportion of the U-shaped conformer would increase in a solution. The fluorescence emission spectra of 7a provided a support for existence of the U-shaped conformation in methanol or water. Besides a long-wavelength peak, there is a short-wavelength shoulder peak in the spectra of 7a in methanol or water, and the wavelengths of the shoulders are similar to the fluorescence peak in THF (Figure 4). The shoulder peaks imply that a part of the indole moieties of 7a are in hydrophobic surroundings by forming the U-shaped molecular conformation. When the model compound with a long and flexible linker forms a U-shaped conformation, the acceptor-donor distance would become closer with increasing solvent polarity. If values of RDA are given in terms of solvent polarity (see Table 2), new values of ∆Gbet′, λs′, and G+bet′ would be obtained. Data in Table 2 show that the values of G+bet′ increase with solvent polarity, and this is just right contrary to the change trend in values of G+bet. The change would cause kbet to decrease and give an increase in the quantum yield of dimer splitting, with increasing solvent polarity. This explained well the reversal in the dependence of the quantum yield on solvent polarity for the flexible long-linker systems compared with the short-linker systems. For the same solvent, the quantum yields decrease gradually from 8 to 12 only in THF (Table 1). This implied that their conformation may be spreading in the low-polarity solvent, and the distance between the dimer and the indole increases gradually from 8 to 12, with covalent-bond numbers of linkers from 3 to 7, giving a gradual decrease in the splitting efficiency. In other polar solvents, the change in the quantum yield is not regular because of different changes in the conformation of model compounds with different-length linkers. For chromophore-dimer model systems, the fluorescence quenching efficiency (Q) of the chromophore can reflect the efficiency of forward electron transfer, Q ) φfet.13,23 In the model studies reported, similar solvent effects on forward electron transfer have been observed; that is, the efficiency increases with solvent polarity. Data in Table 3 come to a

conclusion that is completely in agreement with results reported. The efficiency of forward electron transfer should accord with the distance between the electron donor (indole) and the acceptor (dimer); that is, a higher value of Q implies a closer distance for those similar model compounds (only a different linker) in the same solvent. A change to a more polar solvent causes an increase in values of Q for compounds 8, 11, and 12. This is in agreement with observations from 39 and 7.13 For the lowpolarity solvent THF, the values of Q decrease with increasing the covalent-bond number from 8 to 12, similar to the quantum yields Φ, and in other polar solvents, the values of Q did not wholly decrease. For example, in the water/acetonitrile mixture solvent, values of Q are 0.99, 0.85, 0.86, 0.94, and 0.82 for compounds from 8 to 12, respectively. The value 0.94 for compound 11 implied that the molecule may be a U-shaped conformation in the high polarity solvent (90:10 H2O/CH3CN). For the same compound in different solvents, values of Q allow the evaluations of solvent effect on one pair of competition: photophysical processes (kf and kic) and forward electron transfer (kfet), that is, efficiency of kfet increases with increasing solvent polarity. Furthermore, the change extent of Q values increases with the covalent-bond number from 8 to 12; that is, the difference in Q values is small for the short-linker compounds such as 8 (ranging from 0.95 to 0.99) and large for long-linker compounds such as 11 (0.63-0.94) and 12 (0.52-0.82). These results show that a small change in distance results from small changes in molecular conformation for shortlinker compounds and a large change in the distance for the long-linker compounds such as the change from spreading to the U-shaped conformation, in different solvents. Values of Q for compounds 12 and 713 show that the U-shaped conformation may be formed in higher polarity solvents, and thus the distances become closer with increasing solvent polarity. Thus, the pathways of electron transfer for the flexible long-linker systems may be through-bond in a nonpolar solvent and through-space in a high polarity solvent.

7210

J. Phys. Chem. B, Vol. 113, No. 20, 2009

The variations of the quantum yield (Φ) of dimer splitting and the degree of fluorescence quenching (Q) with solvent polarity allow the evaluations of solvent effect on another pair of competition: back electron transfer (kbet) and the splitting of the dimer (kspl), that is, φbet + φspl ) 1. The splitting efficiencies of dimer radical anion φspl (φspl) Φ/Q, listed in Table 3) decrease with increasing solvent polarity for the short-spacer compounds, 8, 9, and 10. As the short-linker systems 3,9 this result can be explained as the inverted behavior. Highly exothermic back electron transfer (kbet) was gradually slowed with decreasing solvent polarity, giving gradual increase in the quantum yields. The values of φspl for the long-linker compounds 12 and 713 increase with solvent polarity. In the competition of the splitting and back electron transfer of the dimer radical anion, the efficiency of the former increases because the latter is slowed down with increasing solvent polarity. Thus, back electron transfer by the through-space pathway or even direct van der Waals contact between the indole and the dimer would be suppressed efficiently by the solvation of the charge-separated species. The van der Waals contact suppressing back electron transfer has been demonstrated in the complex of a modified β-CD by dimer with indole or N,N′-dimethylaniline.24 Hence, the process where van der Waals contact between the flavin and the dimer in the complex of photolyase bonding the substrate retards back electron transfer by active-site solvation25 should be one important factor for achieving highly efficient repair. In summary, two reverse solvent effects have been observed from the quantum yields of dimer splitting of five synthesized covalently linked indole-dimer compounds with different-length linkers 8-12, and a mechanism for interpreting the solvent effects has been suggested. One solvent effect is an increase in the splitting efficiency in lower polarity solvents for model compounds with a short linker, 8-10, and another is more efficient splitting in higher polarity solvents for model compounds with a flexible and long linker, 11 and 12. If the distance between the dimer and the indole is a constant or a small change for a model system in various solvents, calculated results in terms of the Marcus theory are fully identical to Rose’s conclusions that the increase in splitting efficiency in lower polarity solvents derives from a possible slowing of the highly exothermic back electron transfer due to Marcus inverted behavior. This can explain well the solvent effect of the shortlinker compounds 8-10. In the case of model systems with a long flexible linker, the distance would become much less from a spreading to a U-shaped conformation with increasing solvent polarity because of their hydrophobic interaction and the specific structure of dimer. On the basis of Marcus theory, calculated results reveal that the rate of back electron transfer would be slowed down with increasing solvent polarity, giving a more efficient splitting. These results can clarify the reversal solvent dependence of the quantum yield for long-linker compounds 11, 12, and 7, compared with the short-linker compounds. It can be deduced that the solvent dependence from flavin-dimer systems12 with a long flexible spacer also derive

Tang et al. from the hydrophobic interaction between the flavin ring and the dimer, resulting in a closer distance in a higher polarity solvent and giving a higher splitting efficiency. Acknowledgment. This work was supported by the National Natural Science Foundation of China (Grant Nos. 30870581, 30470444, and 20802003). Supporting Information Available: Synthesis procedures, spectral characterization data, and copies of the 1H (300 MHz) and 13C NMR (75 MHz) spectra, for compounds 8-12. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Friedberg, E. C.; Walker, G. C.; Siede, W. DNA Repair and Mutagenesis; ASM Press: Washington, DC, 1995. (2) Sancar, A. Biochemistry 1994, 33, 2–9. (3) Sancar, A. Chem. ReV. 2003, 103, 2203–2237. (4) Begley, T. P. Acc. Chem. Res. 1994, 27, 394–401. (5) Heelis, P. F.; Hartman, R. F.; Rose, S. D. Chem. Soc. ReV. 1995, 24, 289–297. (6) Carell, T.; Burgdorf, L. T.; Kundu, L. M.; Cichon, M. Curr. Opin. Chem. Biol. 2001, 5, 491–498. (7) Kim, S.-T.; Hartman, R. F.; Rose, S. D. Photochem. Photobiol. 1990, 52, 789–794. (8) Kim, S.-T.; Rose, S. D. J. Photochem. Photobiol., B 1992, 12, 179– 191. (9) Hartzfeld, D. G.; Rose, S. D. J. Am. Chem. Soc. 1993, 115, 850– 854. (10) Young, T.; Kim, S.-T.; Van Camp, J. R.; Hartman, R. F.; Rose, S. D. Photochem. Photobiol. 1988, 48, 635–641. (11) Rosep, S. D. In CRC handbook of organic photochemistry and photobiology; Horspool, W., Song, P.-S., Eds.; CRC Press: New York, 1995; pp 1332-1346. (12) Epple, R.; Wallenborn, E.-U.; Carell, T. J. Am. Chem. Soc. 1997, 119, 7440–7451. Carell, T.; Epple, R. Eur. J. Org. Chem. 1998, 1245– 1258. Butenandt, J.; Epple, R.; Wallenborn, E.-U.; Eker, A. P. M.; Gramlich, V.; Carell, T. Chem. Eur. J. 2000, 6, 62–72. (13) Song, Q. H.; Tang, W. J.; Hei, X. M.; Wang, H. B.; Guo, Q. X.; Yu, S. Q. Eur. J. Org. Chem. 2005, 1097–1106. (14) Marcus, R. A. J. Chem. Phys. 1956, 24, 966–978. (15) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochemistry, 2nd ed.; Marcel Dekker Inc.: New York, 1993. (16) Song, Q. H.; Wang, H. B.; Tang, W. J.; Guo, Q. X.; Yu, S. Q. Org. Biomol. Chem. 2006, 4, 291–298. Song, Q. H.; Tang, W. J.; Ji, X. B.; Wang, H. B.; Guo, Q. X. Chem. Eur. J. 2007, 13, 7762–7770. (17) Weller, A. A. Z. Phys. Chem. (Muenchen, Ger.) 1982, 133, 93–98. (18) Scannell, M. R.; Fenick, D. J.; Yeh, S. R.; Falvey, D. E. J. Am. Chem. Soc. 1997, 119, 1971–1977. (19) DeFelippis, M. R.; Murthy, C. P.; Broitman, F.; Weinraub, D.; Faraggi, M.; Klapper, M. H. J. Phys. Chem. 1991, 95, 3416–3419. (20) Marcus, R. A. Can. J. Chem. 1959, 37, 155–163. Marcus, R. A. Annu. ReV. Phys. Chem. 1964, 15, 155–196. (21) Tang, W. J.; Song, H. B.; Song, Q. H. Chin. J. Struct. Chem. 2007, 26, 381–384. (22) Reichardt, C. Chemical Rubber. In SolVents and SolVent Effects in Organic Chemistry, 2nd ed.; VCH: Weinheim, 1988; Handbook of Chemistry and Physics, 75th ed.; Lide, D. R., Ed.; CRC Press: Boca Raton, FL, 1994. (23) Kim, S.-T.; Rose, S. D. J. Phys. Org. Chem. 1990, 3, 581–586. (24) Tang, W. J.; Song, Q. H.; Wang, H. B.; Yu, J. Y.; Guo, Q. X. Org. Biomol. Chem. 2006, 4, 2575–2580. (25) Kao, Y.-T.; Saxena, C.; Wang, L. J.; Sancar, A.; Zhong, D. P. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 16128–16132.

JP805965E