Biological deactivation and single-strand breakage of plasmid DNA by photosensitization using tris(2,2'-bipyridyl)ruthenium(II) and peroxydisulfate

Photochemically active DNA-intercalating ruthenium and related complexes - insights by combining crystallography and transient spectroscopy

Recent research on the study of the interaction of ruthenium polypyridyl compounds and defined sequence nucleic acids is reviewed. Particular emphasis is paid to complexes [Ru(LL)2(Int)]2+ containing potentially intercalating ligands (Int) such as dipyridophenazine (dppz), which are known to display light-switching or photo-oxidising behaviour, depending on the nature of the ancillary ligands. X-ray crystallography has made a key contribution to our understanding, and the first complete survey of structural results is presented. These include sequence, enantiomeric, substituent and structural specificities. The use of ultrafast transient spectroscopic methods to probe the ultrafast processes for complexes such as [Ru(TAP)2(dppz)]2+ and [Ru(phen)2(dppz)]2+ when bound to mixed sequence oligonucleotides are reviewed with particular attention being paid to the complementary advantages of transient (visible) absorption and time-resolved (mid) infra-red techniques to probe spectral changes in the metal complex and in the nucleic acid. The observed photophysical properties are considered in light of the structural information obtained from X-ray crystallography. In solution, metal complexes can be expected to bind at more than one DNA step, so that a perfect correlation of the photophysical properties and factors such as the orientation or penetration of the ligand into the intercalation pocket should not be expected. This difficulty can be obviated by carrying out TRIR studies in the crystals. Dppz complexes also undergo insertion, especially with mismatched sequences. Future areas for study such as those involving non-canonical forms of DNA, such as G-quadruplexes or i-motifs are also briefly considered.

Direct photo-induced DNA strand scission by a ruthenium bipyridyl complex

Irradiation of plasmid DNA in the presence of Ru(II)-2, a modified tris(2,2'-bipyridyl)Ru(II) complex, in which two hydroxamic acid groups are attached to one of the three bipyridyl ligands, results in total fragmentation of the DNA. The photo-chemical reaction products were analyzed by gel electrophoresis, which revealed complete fragmentation. Further evidence for the complete degradation of the DNA was obtained by imaging the pre- and post-treated plasmid DNA using atomic force microscopy (AFM). A mechanism for the reaction is proposed. It initially involves the photo-chemical generation of Ru(III) ions and superoxide radicals, as corroborated by absorbance difference spectroscopy and electron paramagnetic resonance (EPR). Consequently, Ru(III) preferentially oxidizes guanine, liberating superoxide radicals that yield OH radicals. The OH radicals were identified by observing the spectral change at 532 nm of a 5'-dAdG substrate forming a colored adduct with thiobarbituric acid. These radicals are associated with the major non-specific damage exerted to DNA.

Targeting of photooxidative damage on single-stranded DNA representing the bcr-abl chimeric gene using oligonucleotide-conjugates containing [Ru(phen)3](2+)-like photosensitiser groups

Photooxidative damage was induced predominantly at a single guanine base in a target DNA by irradiation (lambda > 330 nm) in the presence of complementary oligodeoxynucleotide conjugates (ODN-5'-linker-[Ru(phen)3]2+) (phen = 1,10-phenanthroline). The target DNA represents the b2a2 variant of the chimeric bcr-abl gene implicated in the pathogenesis of chronic myeloid leukaemia, and the sequence of the 17mer ODN component of the conjugate (3' G G T A G T T A T T C C T T C T T 5') was complementary to the junction region of the sense strand sequence of this oncogene. Two different conjugates were prepared, both of them by reaction of the appropriate succinimide ester with 5'-hexylamino-derivatised 17mer ODN. In Ru-ODN-1 (7) the linker was -(CH2)6-NHCO-bpyMe (-bpyMe = 4'-[4-methyl-2,2'-bipyridyl]), whereas in Ru-ODN-2 (13) it was -(CH2)6-NHCO-(CH2)3-CONH-phen. Photoexcitation of either of the conjugates when hybridised with the 32P-5'-end-labelled target 34mer 5'T G A C C A T C A A T A A G G A A G A A G21 C C C T T C A G C G G C C 3' (ODN binding site underlined) led to an alkali-labile site predominantly (> 90%) at the G21 base, which is at the junction of double-stranded and single-stranded regions of the hybrid. Greater yields were found with Ru-ODN-1 (7) than with Ru ODN-2 (13). In contrast to this specific cleavage with Ru-ODN-1 (7) or Ru-ODN-2 (13), alkali-labile sites were generated at all guanines when the 34mer was photolysed in the presence of the free sensitiser [Ru(phen)3]2+. Since [Ru(phen)3]2+ was shown to react with 2'-deoxyguanosine to form the diastereomers of a spiroiminodihydantoin derivative (the product from 1O2 reaction), 1O2 might also be an oxidizing species in the case of Ru-ODN-1 (7) and Ru-ODN-2 (13). Therefore to determine the range of reaction, a series of 'variant' targets was prepared, in which G21 was replaced with a cytosine and a guanine substituted for a base further towards the 3'-end (e.g. Variant 3; 5'T G A C C A T C A A T A A G G A A G A A C C G23 C T T C A G C G G32 C C3'). While it was noted that efficient reaction took place at distances apparently remote from the photosensitiser (e.g at G32, but not G23 for Variant 3), this effect could be attributed to hairpinning of the single-stranded region of the target. These results are therefore consistent with the photooxidative damage being induced by a reaction close to the photosensitiser rather than by a diffusible species such as 1O2.

Evidence of inter- and intra-molecular crosslinking of tyrosine residues of calmodulin induced by photo-activation of ruthenium(II)

Tris(2,2'-bipyridyl)ruthenium(n) upon illumination with light at a wavelength of 450 nm in the presence of an electron acceptor induces dityrosine crosslinking in proteins.

DNA photocleavage by a supramolecular Ru(II)-viologen complex

A novel Ru(II) complex possessing two sequentially linked viologen units, Ru-V(1)-V(2)(6+), was synthesized and characterized. Upon excitation of the Ru(II) unit (lambda(exc) = 532 nm, fwhm approximately 10 ns), a long-lived charge-separated (CS) state is observed (tau = 1.7 micros) by transient absorption spectroscopy. Unlike Ru(bpy)(3)(2+), which cleaves DNA upon photolysis through the formation of reactive oxygen species, such as (1)O(2) and O(2)(-), the photocleavage of plasmid DNA by Ru-V(1)-V(2)(6+) is observed both in air and under N(2) atmosphere (lambda(irr) > 395 nm).

Scope, limitations and mechanistic aspects of the photo-induced cross-linking of proteins by water-soluble metal complexes

BACKGROUND

Chemical cross-linking is a valuable tool with which to study protein-protein interactions. Recently, a new kind of cross-linking reaction was developed in which the photolysis of associated proteins with visible light in the presence of ammonium persulfate and tris(2,2'-bipyridyl)ruthenium(II) dication or palladium(II) porphyrins results in rapid and efficient covalent coupling (Fancy, D.A. & Kodadek, T. (1999). Proc. Natl. Acad. Sci. USA 96, 6020-6024 and Kim, K., Fancy, D.A. & Kodadek, T. (1999). J. Am. Chem. Soc. 121, 11896-11897). Here, mechanistic and practical aspects of the reaction of importance for its application to biochemical problems are examined.

RESULTS

It is shown that the photo-initiated cross-linking chemistry can be optimized for the analysis of protein-protein interactions in crude cell extracts. A number of commonly used epitope or affinity tags survive the reaction in functional form, allowing the simple visualization of the cross-linked products, or their isolation. It is shown that very little light-independent oxidation of protein residues occurs and that significant perturbation of complexes of interest prior to the brief photolysis period does not occur. Finally, evidence is presented that is consistent with a mechanistic model in which ammonium persulfate functions simply as an electron acceptor, facilitating the generation of the key high valent metal complex from the photoexcited species by electron transfer. In the absence of an electron acceptor, a much lower efficiency reaction is observed that appears to involve products resulting from reaction of the excited state metal complex with molecular oxygen.

CONCLUSIONS

These results provide useful practical information for chemists and biochemists who may wish to employ this new cross-linking chemistry for the analysis of protein complexes. They also shed new light on the mechanism of this interesting reaction.

Comparative study of Ru(bpz)3(2+) Ru(bipy)3(2+) and Ru(bpz)2Cl2 as photosensitizers of DNA cleavage and adduct formation

The efficiency of ruthenium complexes for photosensitizing DNA damage depends on the oxidizing character of their ligands. Here we report on the difference in behavior of tris(2.2'-bipyrazyl)ruthenium(II) (Ru[bpz]3(2+)), tris(2,2'-bipyridyl)ruthenium(II) (Ru[bipy]3(2+)) and cis-dichlorobis (2,2'-bipyrazyl)ruthenium(II) (Ru[bpz]2Cl2). Upon irradiation at 436 nm, Ru(bpz)3 (2+) was far less stable than Ru(bipy)3(2+). Ru(bpz)3(2+) in phosphate buffer containing NaCl undergoes a photoanation reaction leading to the formation of Ru(bpz)2Cl2, as previously reported also in organic media. In the presence of phage phi X174 DNA, Ru(bpz)3(2+) photosensitized the formation of single strand breaks with an efficiency that was, at the beginning of irradiation, similar to that of Ru(bipy)3(2+). After 8 min of irradiation, the cleavage efficiency of Ru(bpz)3(2+) reached a plateau that may correspond to its photode composition. For the same conditions, Ru(bpz)2Cl2 did not induce DNA breakage. Scavenging experiments showed that, in the presence of oxygen, DNA cleavage induced by Ru(bpz)3(2+) partly resulted from the formation of singlet oxygen and hydroxyl radical while in the absence of oxygen an additional mechanism involving electron transfer between the excited state of the ruthenium complex and DNA is proposed. The ICP measurement showed that Ru(bpz)3(2+) and Ru(bpz)2Cl2 gave rise to covalent binding onto DNA in contrast with Ru(bipy)3(2+), which did not bind to DNA under the experimental conditions. The results are discussed with regard to the potential use of these photosensitizers in phototherapy.

Inhibition of OH radical-induced strand break formation of poly(U) by Ru(bpy)32+ or Ru(phen)32+ attached to the polynucleotide

Reactions of OH radicals with poly(U) (polyuridylic acid) in the presence of Ru(bpy)32+ or Ru(phen)32+ in aqueous solutions were studied. OH radicals were produced by pulse radiolysis and their reactions with ruthenium complexes were measured spectrophotometrically under conditions were the complexes are attached to the polynucleotide. The OH radical adds to either the uracil moiety or the ruthenium complexes. The ratio of the radicals produced depends only on the ratio of their rate constants and the concentrations of poly(U) and ruthenium complexes. Similar results were obtained with uridine-5'-monosphosphate, where the ruthenium complexes are not attached to the nucleotide. Surprisingly, the yield of single-strand break formation from the OH adducts of uracil in poly(U) is much smaller than that expected on the basis of the yield measured in the absence of ruthenium complexes. Possible reasons for this behaviour are discussed.

Tuning the mechanism of DNA cleavage photosensitized by ruthenium dipyridophenazine complexes by varying the structure of the two non intercalating ligands

The influence of the nature of ligands on the efficiency of ruthenium complexes for photosensitizing DNA cleavage was investigated. Ru(bipy)2dppz2+ and Ru(bpz)2dppz2+ were selected as DNA breakers on the basis of their high affinity for DNA due to the presence of a dppz ligand which can partially intercalate in the major groove of DNA. Their photosensitizing properties were compared to those of Ru(bipy)3(2+), a complex which binds to DNA with a far lower constant. Upon irradiation, these complexes promoted the formation of single strand breaks in supercoiled phi X 174 DNA. Unexpectedly, Ru(bipy)2dppz2+ was found to be less efficient than Ru(bipy)3(2+) whatever the dye concentration or the [base pair]/[Ru] molar ratio r. Scavenging experiments have shown that the oxidative DNA cleavage induced by Ru(bipy)2dppz2+ mainly results from a Type II mechanism. The behavior of Ru(bipy)2dppz2+ was different: this compound was clearly more efficient than Ru(bipy)3(2+) as DNA breaker and its efficiency was not modified by the presence of oxygen or by addition of scavengers of reactive oxygen species. In this case, a mechanism involving electron transfer between the excited state of the ruthenium complex and the guanine residue was proposed in agreement emission lifetime measurements. The change in mechanism observed between Ru(bipy)2dppz2+ and Ru (bipy)2dppz2+ results from an increase of the reduction potential of the ruthenium complexes in the excited state, which appears to be the main factor controlling the efficiency.

Photoaddition of ruthenium(II)-tris-1,4,5,8-tetraazaphenanthrene to DNA and mononucleotides

Formation of adducts between Ru(TAP)3(2+) (TAP = 1,4,5,8-tetraazaphenanthrene) and DNA has been monitored by gel electrophoresis, UV-vis spectroscopy and dialysis methods. Adduct formation is found for both single- and double-stranded nucleic acids. The reaction with double-stranded DNA is found to be insensitive to solution pH or aeration. Spectroscopic changes similar to those for DNA are found with GMP in oxygen-free pH 5 solution. However, different reactions occur with GMP at higher pH or when the solution contains oxygen. Comparative experiments with double-stranded poly[d(G-C)] or poly[d(A-T)] indicate that the adduct with DNA involves binding to the guanosine moiety. It is proposed that the product is formed by the reaction of the reduced ruthenium complex and oxidised guanine species produced by photo-induced electron transfer.

Are enzymatically produced single-strand breaks involved in UV-induced inactivation of plasmid DNA?

pBR322 plasmid DNA was exposed to 254 nm UV radiation and examined for enzymatically produced single-strand break (sbb) and double-strand break (dsb) formation by treatment with an extract containing the proteins of Escherichia coli (AB1157 (uvrA+ recA+) and AB2480 (uvrA- recA-)). Enzymatic conversion of the 254 nm-induced lesions into ssbs on treatment with an extract from AB1157 was observed, but not conversion into dsbs. The rate of enzymatic ssb formation in the AB1157 extract is initially higher than in the AB2480 extract, the sbb formation levels off leading to plateau values with increasing incubation time. The rate of ssb formation in the AB2480 extract is initially lower, but does not level off, and the ssb yield becomes larger at longer incubation times than that with the AB1157 extract. The biological inactivation of the plasmids was measured as a function of 254 nm fluence by transformation of E. coli AB1157 and AB2480. Inactivation with AB2480 is mainly due to a single photoproduct, a cyclobutane-type pyrimidine dimer, per DNA molecule. Inactivation with AB1157 occurs with a quantum yield which is virtually identical with that of the plateau values of enzymatic ssb formation, as measured by incubation in the AB1157 extract. A possible interpretation is that the formation of irreparable ssbs is the lethal step in the sequence of events leading to inactivation of plasmid DNA in the repair wild-type strain. The quantum yield of inactivation is 10-20 times smaller for transformation of AB1157 than for AB2480, indicating that enzymatic repair of photolesions of the plasmid occurs in AB1157.

Photolesions in DNA upon 193 nm excitation

Aqueous solutions of plasmid (pBR322 and pTZ18R) and calf thymus DNA were excited by 20 ns laser pulses at 193 nm. The quantum yields of single-and double-strand break formation, interstrand cross-links, locally denatured sites, (6-4)photoproducts and biological inactivation (phi ssb, phi dsb, phi icl, phi lds, phi 6-4 and phi ina, respectively) were measured. The quantum yields are virtually independent of intensity, demonstrating a one-quantum process. The obtained values in aerated neutral solution in the absence of additives are phi ssb approximately 1.5 x 10(-3), phi dsb approximately 0.06 x 10(-3) (dose: 10-200 J m-2), phi icl approximately phi lds approximately 0.1 x 10(-3) and phi 6-4 = 0.5 x 10(-3). Both phi ssb and phi dsb decrease strongly with increasing concentrations of TE buffer (0.01-10 mM). Biological inactivation of the pTZ18R plasmid was determined from the transformation efficiency of Escherichia coli bacteria strains AB1157, AB1886 uvr and AB2480 uvr rec; the phi ina values are 1.4 x 10(-3), 2.1 x 10(-3) and 3 x 10(-3), respectively. The monoexponential survival curves in all cases show that a single damage site leads to inactivation (one single hit). The biological consequences of different photoproducts are discussed.

Photoinduced interaction of Ru(bpy)3 2+ with nucleotides and nucleic acids in the presence of S2O8 2-; a transient conductivity study

The photochemical reactions of Ru(bpy)3(2+) with single- and double-stranded DNA, polynucleotides and purine-containing nucleotides in argon-saturated aqueous solution in the presence of S2O8(2-) were studied using time-resolved absorption and conductivity methods. The conversion of Ru(bpy(3(3+) to Ru(bpy)3(2+), monitored spectroscopically either after rapid mixing with substrate or after laser flash excitation (lambda exc = 353 nm) is quantitative at nucleotide-to-sensitizer ratios [N]/[S] of 1-2 for DNA and other guanine-containing compounds. Conductivity measurements following the laser pulse revealed a fast conductivity increase (rise time, less than 0.1 ms) due to the formation of protons and, to a lesser degree, to charged species of much lower ion mobility. A slower component in the 0.01-1 s range was observed for nucleic acids; its amplitude is markedly reduced at pH 6-9. In buffered neutral solution the signal is replaced by a slight decrease in conductivity. Electronically excited Ru(bpy)3(2+) bound to DNA reacts with S2O8(2-) to form Ru(bpy)3(3+) and SO4(.-) as primary oxidizing species both of which react with bases. The resulting base radicals react subsequently with Ru(bpy)3(3+) and Ru(bpy)3(2+) or the ligands in the ruthenium complex, producing protons which give rise to the slower conductivity increase. The formation of single-strand breaks and the ensuing release of condensed counterions does not appear to contribute significantly to the slow component. The transient conductivity behaviour is sensitive to the single- or double-stranded nature of DNA.

Photo-induced electron transfer from nucleotides to ruthenium-tris-1,4,5,8-tetraazaphenanthrene: model for photosensitized DNA oxidation

The luminescence quenching of ruthenium-tris-1,4,5,8-tetraazaphenanthrene [Ru(tap)3(2+)] by nucleotides approaches the diffusion rate only with guanosine-5'-monophosphate (GMP), the most reducing nucleotide, and leads to an electron transfer with the production of the monoreduced complex and the oxidized base. The resulting deprotonated GMP(-H).radical recombines with the monoreduced complex according to a bimolecular equimolar process. The pH dependence of the decay of the transient reduced complex, in the presence of an oxidant (oxygen or benzoquinone) indicates the formation of Ru(tap)2(tapH)2+, i.e. the reduced protonated species, subsequent to the electron transfer, with a pKa of 7.6 as confirmed from pulse radiolysis experiments. As the non-protonated reduced complex, Ru(tap)2(tap-.)+, has a higher reducing power than the protonated one, oxygen is able to reoxidize only the non-protonated species, whereas benzoquinone reoxidizes both species but with different rate constants. The flash photolysis of Ru(tap)3(2+) in the presence of DNA and the effect of Mg2+ ions and GMP as supplementary additives also show the existence of a photo-induced electron transfer with the nucleic acid, which can be correlated to the photosensitized cleavage of DNA by this complex.

Photosensitized reactions of poly(U) with tris(2,2'-bipyridyl)ruthenium(II) and peroxydisulfate

The reactions of polyuridylic acid [poly(U)] with Ru(bpy)3(3+) [Ru(III)] and SO4.-, following UV and visible light irradiation of Ru(bpy)3(2+) [Ru(II)] in the presence of S2O8(2-), were studied in an argon-saturated aqueous solution using time-resolved absorption and conductivity methods. The kinetics of the Ru(III) conversion to Ru(II) in the presence of poly(U) was monitored spectroscopically either in the absence of SO4.- [rapid mixing with Ru(III)] or in its presence (after laser flash excitation, lambda exc = 353 nm). The conversion of Ru(III) to Ru(II) is complete at a [nucleotide]/[sensitizer] (N/S) ratio greater than or equal to 10 (rate constant k = 12 s-1) for rapid mixing and at N/S greater than or equal to 6 (k = 15 s-1 at N/S = 10) after laser pulsing. Conductivity measurements following the laser pulse revealed a fast conductivity increase (risetime less than 10 micros), due to the formation of charged species and protons. A slower increase in the 0.1-0.5 s range was observed for poly(U) but it is considerably smaller for poly(dU) and absent in uracil containing monounits. The slow increase is unaffected by pH changes in the 3.5-7 range, markedly reduced in the 7-9 range and is replaced by a slight decrease in conductivity in buffered solutions. An explanation is that poly(U)-bound excited Ru(II) reacts with S2O8(2-) forming Ru(III) and SO4.- as oxidizing species both of which react with poly(U) bases. The resulting base radicals react with Ru(III) or the ligands in the ruthenium complex, producing protons which give rise to the slow conductivity increase (k = 15 s-1 at N/S = 10). The formation of single-strand breaks and the ensuing release of condensed counterions does not appear to contribute significantly to the slow conductivity signal. At N/S less than 10 the observed rate and extent of Ru(III)--Ru(II) conversion and of the slow proton production vary markedly with the N/S ratio.

A study of some polypyridylruthenium(II) complexes as DNA binders and photocleavage reagents

The nature of the binding of several ruthenium polypyridyl complexes containing 2,2'-bipyridine (bipy), 4,4'-dimethyl-2,2'-bipyridine (DMB), 1,10-phenanthroline (phen), 4,7-diphenyl-1,10-phenanthroline (DPP), 2,2',2"-terpyridine (terpy), 2,2'-biquinoline (biq), 1,4,5,8-tetraazaphenanthrene (TAP) and 1,4,5,8,9,12-hexaazatriphenylene (HAT), with calf thymus DNA, poly[d(A-T)] and poly[d(G-C)] were studied by absorption and emission spectroscopy, DNA melting techniques, and emission lifetime measurements. In low ionic strength phosphate buffer, spectroscopic changes and DNA stabilization depended on the polypyridyl ligands present, and indicated binding that varied from substantially electrostatic to intercalative. Ru(bipy)2(HAT)2+ and Ru(phen)3(2+), which bind by partial intercalation, also show a strong preference for poly[d(A-T)]. The emission quantum yields for most complexes were increased in the presence of DNA. An exception was Ru(TAP)3(2+) which has a markedly reduced emission quantum yield and lifetime in the presence of poly[d(G-C)] or CT-DNA, due to photoredox interaction with quanines. Emission decays of the complexes generally showed multiexponential behaviour. The ability of the ruthenium complexes to sensitise DNA cleavage was determined using pBR322 plasmid DNA. Ru(TAP)3(2+) is the most efficient sensitiser while uncharged complexes and complexes with very short-lived excited states do not cleave DNA.