9686
J. Phys. Chem. B 2007, 111, 9686-9692
“Switched-On” Flexible Chalcogenopyrylium Photosensitizers. Changes in Photophysical Properties upon Binding to DNA Tymish Y. Ohulchanskyy,† Michael K. Gannon, II,† Mao Ye,† Andrey Skripchenko,‡ Stephen J. Wagner,‡ Paras N. Prasad,*,† and Michael R. Detty*,† Institute for Lasers, Photonics, and Biophotonics, Department of Chemistry, UniVersity at Buffalo, The State UniVersity of New York, Buffalo, New York 14260, and the American Red Cross, Holland Laboratory for the Biomedical Sciences, 15601 Crabbs Branch Way, RockVille, Maryland 20855 ReceiVed: April 28, 2007; In Final Form: June 9, 2007
2,4-Bis(4-dimethylaminophenyl)-6-alkylthiopyrylium and selenopyrylium dyes are essentially nonfluorescent (φF < 0.001) and are poor generators of singlet oxygen in aqueous solution. However, upon complexation to calf thymus DNA, quantum yields for both fluorescence and generation of singlet oxygen increased dramatically. Irradiation of the dye-DNA complexes produced strand breaks in the DNA. The photodamage is not observed in the absence of oxygen and is suppressed by the addition of the singlet oxygen quencher imidazole. The inactivation of the pseudo-rabies virus upon treatment of oxygenated leukodepleted 20% hematocrit red blood cell suspensions with the chalcogenopyrylium dyes and light followed the same trend observed with quantum yields for the generation of singlet oxygen in the dye-DNA complexes.
Introduction Photodynamic therapy (PDT) uses the combination of a photosensitizer, light of a wavelength that is absorbed by the photosensitizer, and available oxygen to initiate a cytotoxic event.1 While PDT has been utilized primarily for the treatment of cancer, it also has been used to treat other diseases such as age-related macular degeneration2 and psoriasis.3 The advantages offered by PDT are the selectivity provided by preferential accumulation of a photosensitizer in the target tissue and the selective delivery of light to the target tissue. Photodynamic antimicrobial chemotherapy (PACT) is a form of PDT that is used ex vivo to target viral/bacterial pathogens that may infect the blood supply prior to transfusion.1,4,5 Ideally, PACT or any other method for pathogen reduction should target only the pathogen and be inert toward blood cells (red blood cells (RBCs) and platelets). A problem in developing a single method for pathogen reduction is that pathogens may be extracellular, intracellular, or bound to the cellular membrane. Molecules that bind to nucleic acids are an ideal starting point for developing a photosensitizer for PACT since platelets and RBCs lack genomic material. However, few materials bind to nucleic acids with 100% specificity, so some damage to RBCs or platelets may be unavoidable. While the blood supply is safer now than anytime in its more than 50 year history, there remains a small residual risk of pathogen transmission from infected donors whose blood was drawn during the window period before detectable levels of antigen, antibody, and/or nucleic acid for the pathogen were present.1,5 The application of a pathogen reduction procedure must inactivate the pathogen to 6 log inactivation) without significantly harming blood * Authors to whom correspondence should be addressed. (P.N.P.) Tel.: (716) 645-6800 ext. 2099; fax: (716) 645-6963; e-mail: pnprasad@ buffalo.edu. (M.R.D.) Tel.: (716) 645-6800 ext. 2200; fax: (716) 6456963; e-mail:
[email protected]. † University at Buffalo, The State University of New York. ‡ American Red Cross.
for transfusion. For RBCs, the latter condition is stringent: hemolysis must involve less than 1% of red cells after 42 days of storage at 1-6 °C since regulatory rules do not permit the clearance of products with excessive hemolysis.5 Many of the more promising photosensitizers6 for PACT, such as methylene blue (MB, Chart 1), methylene violet, and dimethylmethylene blue, bear a positive charge and have a high affinity for nucleic acids.7-11 Several studies have shown that MB and related compounds when bound to DNA can damage DNA by processes that involve the generation of singlet oxygen (1O2) and by oxygen-independent processes that involve direct interaction of the excited dye with DNA.8,12,13 Unfortunately, since MB and related compounds are rigid molecules via the three fused rings of the phenothiazine core, they are efficient generators of 1O2 in solution, and collateral damage to RBCs from unbound MB is pronounced when red cell units are treated with MB and light.6,7 An alternative approach for developing a photosensitizer for PACT is to use a flexible photosensitizer that dissipates excitedstate energy via rotational/vibrational relaxation in solution but that becomes rigid upon binding to genomic materials. Under these conditions, the rate of internal conversion would be significantly reduced in the bound state in comparison to that of free dye in solution, and other singlet excited-state deactivation processes, such as fluorescence and/or intersystem crossing, would then predominate. Flexible dyes such as malachite green (MG, Chart 1) and related structures are virtually nonfluorescent in solution in the absence of a biopolymer but have been shown to “switch-on” fluorescence in viscous or cold environments or by binding to oligonucleotides, such as RNA aptamers.14 The binding of MG to RNA increases the MG fluorescence quantum yield, φF, more than 2000-fold (φF,free ) 7.9 × 10-5 and φF,bound ) 0.187) relative to free dye in aqueous solution. The MG binding aptamer was originally conceived in the hope that irradiation of aptamer bound dye might inactivate an attached mRNA by the generation of hydroxy radicals.15 More recently,
10.1021/jp073270d CCC: $37.00 © 2007 American Chemical Society Published on Web 07/24/2007
Flexible Chalcogenopyrylium Photosensitizers
J. Phys. Chem. B, Vol. 111, No. 32, 2007 9687
CHART 1
CHART 2
tetramethylrosamine derivatives have been shown to increase their fluorescence upon binding to biomolecules.16 More applied applications of this dye switching approach to PACT are found with thiopyrylium dye, 1-S,17 and the cyanine dye thiazole orange (TO, Chart 1).18 Dye 1-S is effective against viral and bacterial pathogens and has acceptable hemolysis in the presence of a competitive inhibitor of RBC membrane binding.17 TO also is effective against some bacterial and viral pathogens and has acceptable hemolysis without additives.18 However, 1-S requires 10-4 M dye to be effective against some viral and bacterial pathogens,17 while TO’s electronic absorption bands overlap those of hemoglobin.18 Like MG, in solution and not bound to a biopolymer, both 1-S and TO dissipate energy from the excited state by internal conversion via rotational and vibrational modes, as shown in Chart 1. The positive charge on 1-S and TO increases binding to nucleic acids, while binding of the flexible photosensitizers to nucleic acids reduces rotational degrees of freedom, which might in turn increase the relative yield for the phototoxic process(es). In spite of the potential of these dyes for use in PACT, no detailed photophysical studies of the photophysics of 1-S or TO bound to DNA have been described. Single- and doublestranded cleavage of DNA has been described upon irradiating DNA complexes of the related oxazole yellow (YO, Chart 1) as well as dimeric structures of both YO and TO.19,20 In these studies, intercalated dyes appear to damage DNA via generation of hydroxy radicals, while externally bound dye appears to damage DNA via generation of singlet oxygen. Similar studies of the DNA-1-S complex have not been described, although an increase in fluorescence intensity upon intercalation of 1-S into DNA has been reported.21 In related work, we examined the binding of 1-S and the thiopyrylium (2-S-6-S) and selenopyrylium (6-Se) analogues, shown in Chart 2, to DNA using a topoisomerase I DNA unwinding assay, competition dialysis experiments with [poly(dGdC)]2 and [poly(dAdT)]2, ethidium displacement studies, circular dichroism studies, and isothermal titration calorimetry to determine binding constants.22 The results of these assays with 1-S and other chalcogenopyrylium dyes
were consistent with mixed binding modes to DNA, presumably involving intercalation. These chalcogenopyrylium dyes have characteristics of a good photosensitizer for the photoinactivation of blood-borne pathogens; they (a) bind to DNA and (b) have a strong absorbance above 600 nm when bound to DNA, which is red-shifted relative to the longest wave absorption band of hemoglobin (band I with λmax of 575 nm). Herein, we examine the photophysical properties of the thiopyrylium dye 1-S and the related thio- (6-S) and selenopyrylium (6-Se) analogues in aqueous solution and in aqueous solution bound to DNA. We demonstrate that not only the fluorescence yield but also the quantum yield for the production of 1O2 in these systems greatly increases upon binding to DNA. This increase correlates with the ability of the chalcogenopyrylium dyes to inactivate the extracellular virus in red blood cells and to damage DNA upon irradiation. Of the many flexible dyes that have been examined for PACT, the selenopyrylium dye 6-Se described herein has photophysics clearly dominated by the triplet state when DNA bound. Experimental Procedures Materials. Compounds 1-S-6-S and 1-Se were prepared by literature methods.22 Rhodamine 6G (R6G) and Rose Bengal were purchased from Fisher Scientific and were used as received. 1,3-Diphenylisobenzofuran (DPBF), HPLC-grade methanol, and certified methylene blue were used as received from Sigma-Aldrich Co. meso-Tetra(4-sulfonatophenylporphyrin) (TPPS4) was purchased from Frontier Scientific, Inc. and was used as received. Ultrapure calf thymus DNA was obtained from Invitrogen. DNA duplex concentrations (in bp) were determined spectrophotometrically using 260 ) 12 824 M-1 cm-1 for calf thymus DNA. Instrumentation. Steady-state measurements at room temperature were performed using a Shimadzu UV-3101PC spectrophotometer or a PerkinElmer Lambda 12 spectrophotometer for absorption spectra and a Fluorolog 3 spectrofluorometer (Jobin Yvon) for fluorescence spectra. The photolysis radiation was supplied by a 100 W QTH lamp (Oriel). The output of the lamp was passed through 500 nm long-pass and 800 nm shortpass filters (both from Oriel) and focused through a 1 cm cell of water onto the sample cuvettes. Emission Quantum Yields. Aqueous solutions of dyes 1-S, 6-S, and 6-Se (8 × 10-6 M) were prepared in 0.05 M Tris-HCl buffer at pH 7.5 in the presence and absence of ctDNA (8 × 10-5 M bp). Excitation was at 522 nm, where the absorbance for all dye-DNA solutions was essentially close at this concentration at 293 ( 1 K. For determination of values of φF, the absorbance of R6G in methanol at 522 nm was matched to the absorbance of the dye-DNA complexes, and the emission from the dye-DNA complex was compared to the emission from R6G (φF of 0.93 in methanol) at 293 ( 1 K.23
9688 J. Phys. Chem. B, Vol. 111, No. 32, 2007 Singlet Oxygen Quantum Yields. Quantum yields for singlet oxygen generation in air-saturated methanol were determined by indirect methods monitoring the dye sensitized photooxidation of DPBF24 in a stopped-flow spectrophotometer (Applied Photophysics, Ltd.; ×18) at 293 ( 1 K using techniques that we have previously described.25,26 Chemical oxidation of anthracene-9,10-dipropionic acid (ADPA) in dye-buffer dispersions with and without ctDNA was used as an independent method to characterize 1O2 generation efficiency. ADPA is bleached by a singlet oxygen to its corresponding endoperoxide, and this reaction can be monitored spectrophotometrically.27 In this case, a decrease in the absorbance of the ADPA added to dye buffer dispersions with and without ctDNA was monitored as a function of time, following irradiation with a 514 nm laser line from an argon ion laser (Spectra-Physics) at 293 ( 1 K. Absorbance of the compared samples was matched at 514 nm for irradiation. Measurements were performed using the TPPS4 buffer solution as a standard (for TPPS4 in H2O φ(1O2) ) 0.74).28 During experiments, all samples were in quartz cuvettes with a 1 cm path length. Viral Assays for Pseudo-rabies Virus (PRV). Oxygenated leukodepleted 20% hematocrit RBC suspensions in a citratecontaining additive solution, erythrosol, were deliberately inoculated with extracellular PRV, a double-stranded DNA virus used as a model for cytomegalovirus, incubated with various concentrations of 1-S through 6-S or 6-Se, and subsequently illuminated with 1.1 J cm-2 red (660-680 nm) light. VERO cells (isolated from African green monkey kidney, CCL81, ATCC) were propagated in medium (RPMI-1640 supplemented with glutamine, Biofluids) supplemented with 10% fetal bovine serum. Cells were seeded into six-well culture plates and allowed to grow to confluency. Control and phototreated samples were then serially diluted 10-fold, plated onto confluent monolayers, incubated for 60 min, and gently rocked at 310 K for virus adsorption to cells. The inoculum was then removed by aspiration and washed with PBS. A semiliquid agar layer (0.2%) was added to each well, and infected monolayers were incubated at 310 K in air containing 5% CO2. The incubation periods for PRV at 310 K is 4 days. After incubation, the agar layer was removed by aspiration, and the monolayer was stained with 0.1% crystal violet in ethanol for at least 15 min. The stain was removed by aspiration, the plates were washed with water, and the plaques were then enumerated. The extent of viral inactivation was determined by subtracting the log viral titer of phototreated samples from the log of their respective unilluminated controls containing the dye. Photocleavage of Dye-DNA Complexes. pUC19 is a plasmid DNA vector, which is highly supercoiled, with over 90% of the plasmid being in the supercoiled form.29 Ten microliters of an 8 × 10-5 M solution of 6-S or 6-Se in pH 7.4 sodium phosphate buffer was added to 10 µL of purified pUC19 plasmid DNA (8 × 10-4 M bp, roughly 2700 bp per strand) and diluted with 50 µL of pH 7.4 sodium phosphate buffer in an 8 mm high Beckman microcell cuvette and irradiated with 25 mW of filtered 500-800 nm light for varying time periods at 293 ( 1 K. Ten microliter aliquots were taken at various time points and diluted with 10 µL of loading buffer (at pH 7.5) and mixed via centrifugication. Gel electrophoresis was run on each time point on an agarose gel. The gel was transferred to a phosphoimager slide and imaged using a Bio-Rad Fx phosphoimager. The supercoiled and nicked (strand broken) plasmid DNA moved at different rates on the gel. The
Ohulchanskyy et al. phosphoimager was used to give quantitative densities to allow the determination of a time course for DNA damage. The dye-DNA complex, prepared as described previously, was degassed by bubbling a stream of argon bubbles through the solution in a cuvette under sonication for 10 min. The cuvette was sealed under an argon atmosphere and was irradiated with 25 mW filtered 500-800 nm light over a 2 h period at 293 ( 1 K. Ten microliter aliquots were taken at various time points and diluted with 10 µL of loading buffer (at pH 7.5) and mixed via centrifugication. Control samples included 8 × 10-5 M pUC19 DNA in pH 7.4 sodium phosphate buffer allowed to stand for 1 h in the dark or irradiated with 25 mW filtered 500800 nm light for 60 min. A third control was the dye-DNA complex described previously, which was allowed to stand in the dark for 1 h. The dye-DNA complex was prepared as described previously with the singlet oxygen quencher imidazole30 dissolved in the phosphate buffer to give final concentrations of 0.008 and 0.08 M in imidazole. The imidazolecontaining solutions were irradiated with 25 mW filtered 500800 nm light over a 2 h period. Ten microliter aliquots were taken at various time points and diluted with 10 µL of loading buffer (at pH 7.5) and mixed via centrifugication. Statistical Analyses. All statistical analyses were performed using the Student’s t test for pairwise comparisons. A p value of 95% of both 6-Se and 6-S (95% of the dye is bound to DNA (with values of Kb of (4.2 ( 0.6) × 105 M-1 for 6-S and (3.9 ( 0.5) × 105 M-1 for 6-Se, concentration of dye of 8 × 10-6 M, and a concentration of ctDNA of 8 × 10-5 M in bp), the values of φ(1O2) increase to 0.064 ( 0.003 for 1-S, 0.21 ( 0.01 for 6-S, and 0.35 ( 0.02 for 6-Se (Table 1).
The increase in singlet oxygen yield also represents only that fraction of singlet oxygen that is not quenched by DNA and that survives to reach the ADPA in aqueous solution. The actual values of φ(1O2) should be higher than those observed in the reaction with ADPA. As illustrated in Chart 1, binding to DNA should reduce the rate of the rotational relaxation of excited-state energy via the various substituents around the chalcogenopyrylium core. The net result is to decrease the rate of internal conversion relative to the rates of fluorescence and intersystem crossing to the triplet. The triplet then reacts with ground-state oxygen to generate a singlet oxygen. Not only is φ(1O2) enhanced in the dye-DNA complex, but values of φF in solution are miniscule (
