Transient intermediates in intramolecularly photosensitized pyrimidine dimer splitting by indole derivatives

Dynamics and mechanism of UV-damaged DNA repair in indole-thymine dimer adduct: molecular origin of low repair quantum efficiency

Many biomimetic chemical systems for repair of UV-damaged DNA showed very low repair efficiency, and the molecular origin is still unknown. Here, we report our systematic characterization of the repair dynamics of a model compound of indole-thymine dimer adduct in three solvents with different polarity. By resolving all elementary steps including three electron-transfer processes and two bond-breaking and bond-formation dynamics with femtosecond resolution, we observed the slow electron injection in 580 ps in water, 4 ns in acetonitrile, and 1.38 ns in dioxane, the fast back electron transfer without repair in 120, 150, and 180 ps, and the slow bond splitting in 550 ps, 1.9 ns, and 4.5 ns, respectively. The dimer bond cleavage is clearly accelerated by the solvent polarity. By comparing with the biological repair machine photolyase with a slow back electron transfer (2.4 ns) and a fast bond cleavage (90 ps), the low repair efficiency in the biomimetic system is mainly determined by the fast back electron transfer and slow bond breakage. We also found that the model system exists in a dynamic heterogeneous C-clamped conformation, leading to a stretched dynamic behavior. In water, we even identified another stacked form with ultrafast cyclic electron transfer, significantly reducing the repair efficiency. Thus, the comparison of the repair efficiency in different solvents is complicated and should be cautious, and only the dynamics by resolving all elementary steps can finally determine the total repair efficiency. Finally, we use the Marcus electron-transfer theory to analyze all electron-transfer reactions and rationalize all observed electron-transfer dynamics.

Thymine dimer photoreversal in purine-containing trinucleotides

Cyclobutane-pyrimidine dimer yields in UV-irradiated DNA are controlled by the equilibrium between forward and reverse photoreactions. Past studies have shown that dimer yields are suppressed at sites adjacent to a purine base, but the underlying causes are unclear. In order to investigate whether this suppression is the result of repair by electron transfer from a neighboring nucleobase, the yields and dynamics of the reverse reaction were studied using trinucleotides containing a cis-syn dimer (T<>T) flanked on the 5' or the 3' side by adenine or guanine. The probability of forming an excited state on T<>T or on the purine base was varied by tuning the irradiation wavelength between 240 and 280 nm. Cleavage quantum yields decrease by an order of magnitude over this wavelength range and are less than 1% at 280 nm, a wavelength that excites the purine base with more than 95% probability. Conditional quantum yields of cleavage for the trinucleotides given excitation of T<>T are similar in magnitude to the quantum yield of cleavage of unmodified T<>T. These results indicate that within experimental uncertainty all photoreversal in these single-stranded substrates is the result of direct electronic excitation of T<>T. Photolyase-like repair of T<>T due to electron transfer from an adjacent purine is negligible in these substrates. Instead, the observed variation in photoreversal quantum yields for adenine- versus guanine-flanked cis-syn dimer could be due to uncertainties in absorption cross sections or to a modest quenching effect by the purine on the excited state of T<>T. Pump-probe measurements reveal that the excited-state lifetimes of A or G in the dimer-containing trinucleotides are unperturbed by the neighboring dimer, indicating that electron transfer from purine base to T<>T is not competitive with rapid excited-state deactivation. Pump-probe measurements on unmodified T<>T in aqueous solution indicate that cleavage is most likely complete on a picosecond or subpicosecond time scale.

Photosensitized splitting of thymine dimer or oxetane unit by a covalently N-linked carbazole via electron transfer in different marcus regions

Although many similarities exist between the two classes of enzymes, cyclobutane photolyases and (6-4) photolyases have certain important differences. The most significant difference is in their repair quantum yields, cyclobutane photolyases with a uniformly high efficiency (0.7-0.98) and very low repair efficiency for (6-4) photolyases (0.05-0.1). To understand the significant difference, we prepared two classes of model compounds, covalently N-linked dimer- (1) or oxetane-carbazole (2) compounds with a dimethylene or trimethylene group as a linker. Under light irradiation, the dimer or oxetane unit of model compounds can be sensitized to split by the excited carbazole via an intramolecular electron transfer. The splitting reaction of dimer or oxetane unit in model compounds is strongly solvent dependent. In nonpolar solvents, such as cyclohexane or THF, no fluorescence quenching of the carbazole moiety of model compounds relative to a free carbazole, N-methylcarbazole, was observed and thus no splitting occurred. In polar solvents, two classes of model compounds reveal two reverse solvent effects on the splitting quantum yield. One is an inverse relation between the quantum yield and the polarity of the solvent for dimer-model systems, and another is a normal relation for oxetane-model systems. This phenomenon was also observed with another two classes of model compounds, covalently linked dimer- or oxetane-indole. Based on Marcus theory and thermodynamic data, it has been rationalized that the two reverse solvent effects derive from back electron transfer in the splitting process lying in the different Marcus regions. Back electron transfer lies in the Marcus inverted region for dimer-model systems and the normal region for oxetane-model systems. From repair solvent behavior of the two classes of model compounds, we gained some insights into the major difference in the repair efficiency for the two classes of photolyases.

Origin of solvent dependence of photosensitized splitting of a cyclobutane pyrimidine dimer by a covalently linked chromophore

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.

A deoxyribozyme, Sero1C, uses light and serotonin to repair diverse pyrimidine dimers in DNA

An in vitro selection search for DNAs capable of catalyzing photochemistry yielded two distinctive deoxyribozymes (DNAzymes) with photolyase activity: UV1C, which repaired thymine dimers within DNA using a UV light of >300 nm wavelength and no extraneous cofactor, and Sero1C, which required the tryptophan metabolite serotonin as cofactor in addition to the UV light. Catalysis by Sero1C conformed to Michaelis-Menten kinetics, and analysis of the action spectrum of Sero1C confirmed that serotonin did indeed serve as a catalytic cofactor rather than as a structural cofactor. Sero1C and UV1C showed strikingly distinct wavelength optima for their respective photoreactivation catalyses. Although the rate enhancements characteristic of the two DNAzymes were similar, the cofactor-requiring Sero1C repaired a substantially broader range of substrates compared to UV1C, including thymine, uracil, and a range of chimeric deoxypyrimidine and ribopyrimidine dimers. Similarities and differences in the properties of these two photolyase DNAzymes suggest, first, that the harnessing of less damaging UV light for the repair of photolesions may have been a primordial catalytic activity of nucleic acids, and, second, the broader substrate range of Sero1C may highlight an evolutionary advantage to coopting amino-acid-like cofactors by functionality-poor nucleic acid enzymes.

Efficient photosensitized splitting of the thymine dimer/oxetane unit on its modifying beta-cyclodextrin by a binding electron donor

Two modified beta-cyclodextrins (beta-CDs) with a thymine dimer and a thymine oxetane adduct respectively, TD-CD and Ox-CD, have been prepared, and utilized to bind an electron-rich chromophore, indole or N,N-dimethylaniline (DMA), to form a supramolecular complex. We have examined the photosensitized splitting of the dimer/oxetane unit in TD-CD/Ox-CD by indole or DMA via an electron-transfer pathway, and observed high splitting efficiencies of the dimer/oxetane unit. On the basis of measurements of fluorescence spectra and splitting quantum yields, it is suggested that the splitting reaction occurs in a supramolecular complex by an inclusion interaction between the modified beta-CDs and DMA or indole. The back electron transfer, which leads low splitting efficiencies for the covalently-linked chromophore-dimer/oxetane compounds, is suppressed in the non-covalently-bound complex, and the mechanism has been discussed.

The efficiency of photolyase and indole complexes to repair DNA containing dimers of pyrimidine: A theoretical analysis of the electron transfer reactions

We analyze the effects of competing reactions to the efficiency of enzymatic splitting of pyrimidine dimers formed in DNA by the incidence of ultraviolet radiation. This is accomplished with the aid of a formula that expresses the efficiency of the repair in terms of parameters that regulate the reaction rates for primary and for back long-range electron transfers taking place in the process. Comparison of experimental data with estimations on account of this formula supports early conjectures in the literature that attribute the relative high performance of the enzymatic complexes of photolyase to its ability to suppress the back reaction.

Model studies of DNA photorepair: enthalpy of cleavage of a pyrimidine dimer measured by photothermal beam deflection calorimetry

The enzyme DNA photolyase mediates the repair of pyrimidine dimers. This repair step, a net retro [2 + 2] reaction, proceeds through either the cation or anion radical of the pyrimidine dimer. In order to understand how electron transfer makes the repair process possible, its energetics have been examined by photothermal beam deflection calorimetry, fluorescence quenching and quantum yield studies. The enthalpy for the cleavage reaction of cis-syn 1,3-dimethylthymine dimer itself was found to be -19 kcal/mol. In addition, from the redox potentials, the enthalpies for the cleavage reactions of the dimer cation radical and the anion radical were determined to be -19 kcal/mol and -28 kcal/mol, respectively.

Photochemistry, photophysics, and mechanism of pyrimidine dimer repair by DNA photolyase

DNA photolyases photorepair pyrimidine dimers (Pyr < > Pyr) in DNA as well as RNA and thus reverse the harmful effects of UV-A (320-400 nm) and UV-B (280-320 nm) radiations. Photolyases from various organisms have been found to contain two noncovalently bound cofactors; one is a fully reduced flavin adenine dinucleotide (FADH-) and the other, commonly known as second chromophore, is either methenyltetrahydrofolate (MTHF) or 8-hydroxydeazaflavin (8-HDF). The second chromophore in photolyase is a light-harvesting molecule that absorbs mostly in the near-UV and visible wavelengths (300-500 nm) with its high extinction coefficient. The second chromophore then transfers its excitation energy to the FADH-. Subsequently, the photoexcited FADH- transfers an electron to the Pyr < > Pyr generating a dimer radical anion (Pyr < > Pyr.-) and a neutral flavin radical (FADH.). The Pyr < > Pyr.- is very unstable and undergoes spontaneous splitting followed by a back electron transfer to the FADH.. In addition to the main catalytic cofactor FADH-, a Trp (Trp277 in Escherichia coli) in apophotolyase, independent of other chromophores, also functions as a sensitizer to repair Pyr < > Pyr by direct electron transfer.

The photo repair of pyrimidine dimers by DNA photolyase and model systems

Pyrimidine dimers are eliminated from DNA by a number of different mechanisms known as DNA repair. Photoreactivation, the reversal of the harmful effects of short wavelength radiation by subsequent exposure to longer wavelengths, is one such mechanism. In photoreactivation, the enzyme DNA photolyase utilises light in order to catalyse the cleavage of the cyclobutane ring of the pyrimidine dimer. The results of recent studies of E. coli DNA photolyase and model systems using techniques such as steady state and flash photolysis, time resolved fluorescence and photo CIDNP are surveyed. A mechanism is proposed for the in vitro reaction of E. coli DNA photolyase which involves photoreduction of the FAD radical cofactor followed by electron donation to the dimer from the excited singlet state of reduced FAD.

Detection of radical ion intermediates in flavin-photosensitized pyrimidine dimer splitting

Photosensitized splitting of cis-syn- and trans-syn-1,3-dimethyluracil dimers by 2',3',4',5'-tetraacetylriboflavin in acetonitrile containing a trace of perchloric acid was studied by laser flash photolysis. Protonation of the flavin prior to excitation resulted in excited singlet and triplet states that abstracted an electron from the dimers and yielded the protonated flavin radical (FlH2.+), which was detected by absorption spectroscopy. Electron abstraction by the excited singlet state predominated over abstraction by the triplet state. Approximately one-third to one-half of the excited states quenched by the trans-syn dimer yielded FlH2.+, the balance presumably undergoing back electron transfer within the geminate radical ion pair generated by the initial electron transfer. A covalently linked dimer-flavin exhibited very inefficient flavin radical ion formation, consistent with the known low efficiency of dimer splitting in this system. These results constitute the first identification of a flavin radical ion intermediate in photosensitized pyrimidine dimer splitting.

Flavin-sensitized photochemically induced dynamic nuclear polarization detection of pyrimidine dimer radicals

A photochemically induced dynamic nuclear polarization (photo-CIDNP) study of carboxymethyllumiflavin-sensitized splitting of pyrimidine dimers has been carried out. In aqueous solution at high pH, an emission signal (delta 3.9 ppm) was observed from the dimer C(6)- and C(6')-protons of an N(1), N(1')-trimethylene-bridged thymine dimer (1). The dimer photo-CIDNP signal was seen only above pD 11.6 and was most intense at pD 12.9. Also observed were weak enhanced absorption signals from the product of splitting, trimethylenebis(thymine) (delta 1.7 and 7.2 ppm). In contrast, cis, syn-thymine dimer (3) gave no photo-CIDNP signals from the dimer. An enhanced absorption at 1.8 ppm, however, due to the product of splitting (thymine) was observed. It was found that dimer 1 and, to a lesser extent, dimer 3 quenched flavin fluorescence. An N(3),N(3')-dimethylated derivative of 1, however, failed to quench flavin fluorescence. Comparison of the pD profile of the dimer photo-CIDNP signal to the pKa values for thymidine dimer suggested that principally the dideprotonated dimer undergoes electron abstraction by the excited flavin.

The third chromophore of DNA photolyase: Trp-277 of Escherichia coli DNA photolyase repairs thymine dimers by direct electron transfer

Photolyases repair pyrimidine dimers in DNA by converting the light energy of 300- to 500-nm photons into chemical energy. Enzymes from various organisms contain two chromophore cofactors (FADH2 and either methenyltetrahydrofolate or 8-hydroxy-5-deazaflavin) that absorb the low-energy photons and initiate splitting of the cyclobutane ring by a radical mechanism. Here, we show that, in addition to these two chromophores, in the far UV range, direct excitation of one specific tryptophan residue (out of 15 total) in the polypeptide chain of Escherichia coli photolyase leads to splitting of the cyclobutane ring with high quantum yield (phi = 0.56), independent of the other chromophores. The specific tryptophan residue responsible for photosensitized repair was identified as Trp-277 by site-specific mutagenesis.

Pyrimidine dimer splitting in covalently linked dimer-arylamine systems

Cyclobutadipyrimidines (pyrimidine dimers) undergo photosplitting which is sensitized by electron donors. We prepared a series of compounds in which a dimer is directly linked to an arylamine, which acts as sensitizer for dimer splitting. Two diastereomers of the dimer-arylamine exhibited very different splitting efficiencies. Also studied were N-methyl, ring methoxy, as well as deuterated derivatives of the sensitizer. These dimer-arylamines had an absorption band with lambda max approximately 300 nm. In each case intramolecular photosensitization of dimer splitting was highly dependent on the solvent, ranging in one instance from phi spl = 0.02 in water to a high value of 0.31 in the least polar solvent mixture examined (1,4-dioxane: isopentane, 1:99). A mechanism is proposed which involves photoinduced electron transfer from arylamine to dimer and splitting of the dimer radical anion. The dependence of splitting on the solvent was rationalized on the basis of retardation of back electron transfer due to Marcus inverted behavior of the charge-separated species. Photolyases might achieve their high efficiency of dimer splitting in part by employing a hydrophobic active site to slow back electron transfer in a similar manner.

Solvent dependence of pyrimidine dimer splitting in a covalently linked dimer-indole system

Cyclobutadipyrimidines (pyrimidine dimers) undergo splitting that is photosensitized by indole derivatives. We have prepared a compound in which a two-carbon linker connects a dimer to an indolyl group. Indolyl fluorescence quenching indicated that the two portions of the molecule interact in the excited state. Intramolecular photosensitization of dimer splitting was remarkably solvent dependent, ranging from phi spl = 0.06 in water to a high value of phi spl = 0.41 in the least polar solvent mixture examined, 1,4-dioxane-isopentane(5 : 95). A derivative with a 5-methoxy substituent on the indolyl ring behaved similarly. These results have been interpreted in terms of electron transfer from the excited indolyl group to the dimer, which would produce a charge-separated species. The dimer anion within such a species could split or undergo back electron transfer. The possibility that back electron transfer is in the Marcus inverted region can be used to rationalize the observed solvent dependence of splitting. In the inverted region, the high driving force of a charge recombination exceeds the reorganization energy of the solvent, which is less for solvents of low polarity than those of high polarity. If this theory is applicable to the hypothetical charge-separated species, a slower back electron transfer, and consequently higher splitting efficiencies, would be expected in solvents of lower polarity. Photolyases may have evolved in which a low polarity active site retards back transfer of an electron and thereby contributes to the efficiency of the enzymatic dimer splitting.

Photo-CIDNP detection of pyrimidine dimer radical cations in anthraquinonesulfonate-sensitized splitting

Anthraquinone-2-sulfonate (AQS) photosensitizes pyrimidine dimer splitting. Electron abstraction from the dimer is thought to induce dimer splitting, but direct evidence for the existence and intermediacy of dimer radical cations has been lacking. By employing photochemically induced dynamic nuclear polarization, we have found emission signals in the NMR spectra of dimers upon photolysis of dimers in the presence of anthraquinone-2-sulfonate. The two dimers employed were cis, syn-thymine dimer in which the N(1)-positions were linked by a three-carbon bridge and the N(3), N(3')-dimethyl derivative of that compound. The anthraquinone-2-sulfonate sensitized photochemically induced dynamic nuclear polarization spectrum of the methylated derivative exhibited an emission signal from the dimer-C(6) hydrogens. This result implied the existence of a dimer radical cation (mD+.) formed by electron abstraction by excited anthraquinone-2-sulfonate and nuclear spin sorting within a solvent caged radical ion pair [mD+. AQS-.]. Product pyrimidine photochemically induced dynamic nuclear polarization signals were also seen [enhanced absorption by C(6)-hydrogens and emission by C(5)-methyl groups]. Nuclear spin polarization in the product resulted from spin sorting in one or more of its precursors, including mD+. The results support the conclusion that dimer radical cations not only exist but are intermediates in the photosensitized splitting of pyrimidine dimers by anthraquinonesulfonate.