DNA Interaction and Photocleavage Properties of Porphyrins

Nov 7, 1998 - Binding at GC-rich sites is generally by intercalation and indicated by a negative peak upon binding to poly[d(G.C)].poly[d(G.C)], hence...
1 downloads 0 Views 343KB Size
94

Bioconjugate Chem. 1999, 10, 94−102

DNA Interaction and Photocleavage Properties of Porphyrins Containing Cationic Substituents at the Peripheral Position Sashikumar Mettath,† Benjamin R. Munson,† and Ravindra K. Pandey*,†,‡ Photodynamic Therapy Center, Department of Radiation Biology and Department of Nuclear Medicene, Roswell Park Cancer Institute, Buffalo, New York 14263. Received July 29, 1998; Revised Manuscript Received September 23, 1998

A series of mono- and disubstituted cationic porphyrins (1-8) were synthesized and investigated for their ability to bind and cleave DNA in the presence of light. In these porphyrins, the cationic substitutuents were introduced at various peripheral positions, i.e., the non-meso positions of the porphyrin system. The modes of binding of these porphyrins to DNA were investigated by UV-vis spectroscopy, circular dichroism, and an unwinding assay. The intrinsic binding constants Kb of these porphyrins to calf thymus DNA was found to be in the range 104-105 M-1. Two of the zinc(II) complexes of non-meso-substituted cationic porphyrins (5 and 8) were found to bind to DNA via intercalation, which is in contrast to the previously reported outside-binding mode for the Zn(II) complexes of meso-substituted cationic porphyrins. Except for monocationic porphyrin 1 and Ni(II) dicationic porphyrin 6, all the other porphyrins were found to be efficient photocleavers of DNA. The DNA photocleavage characteristics of this series of cationic prophyrins were found to depend on the structural characteristics of the poprhyrins such as (a) length of the side chain of the cationic substituents (2 vs 4), (b) the position of the side chain on the porphyrin ring (4 vs 7), and (c) the presence of the chelating metal in 3, 5, and 8 as compared to the nonmetallo porphyrins 2, 4, and 7, respectively.

INTRODUCTION

Chemical agents which bind and cleave nucleic acids are potentially useful as reagents for accessing structural and genetic information, as well as for development of efficient chemical nucleases (1). Such compounds may also be useful as drugs for treatment of cancer, genetic diseases, and viral infections. Among such agents, cationic porphyrins such as meso-tetra-(4-N-methylpyridyl)porphyrin [T4MPyP] and its related analogues are known to bind as well as cleave DNA (2). Binding studies by others have shown that, in general, the free bases and Cu2+ and Ni2+ complexes of meso-substituted cationic porphyrins intercalate between the base pairs of the DNA as well as bind in the minor groove by an outside binding mode (3, 4), whereas for Mn3+, Fe3+, Zn2+, and Co2+ complexes of these porphyrins, due to the presence of axially bound ligands, intercalation is blocked and only outside binding occurs (5, 6). DNA footprinting studies have shown that Mn3+, Fe3+, Zn2+, and Co2 metallocomplexes of T4MPyP bind to the AT rich regions of DNA in the minor groove. The Ni and Cu complexes of T4MPyP were found to intercalate in GC rich regions and bind in the AT rich regions by outside binding mode (7-10). The affinity of binding to these two regions were different. Fiel and co-workers found that Fe complexes of T4MPyP can be chemically activated to produce strand breaks in DNA (11). Studies by various groups have established that Fe3+, Mn3+, and Co2+ complexes of T4MPyP can be activated using a variety of oxidizing agents to produce * Author to whom correspondence should be addressed, Photodynamic Therapy Center. Phone: (716) 845-3203. Fax: (716) 845-8920. E-mail: [email protected]. † Photodynamic Therapy Center. ‡ Department of Nuclear Medicine.

strand breaks in the DNA (12-14). Conjugates of cationic manganese porphyrins with ellipticine (15, 16), spermine (17), bis(benzimidazole) dye (18), peptide nucleic acid (19), and oligonucleotides (20, 21) when activated by oxidizing agents have also been found to cleave DNA. Porphyrins can also be activated by light to produce an effect on biological molecules, and such a methodology has been used in photodynamic therapy for treatment of cancer (22, 23). Fiel and co-workers examined the light activated DNA cleavage property of meso-tetra(p-carboxyphenyl)porphine (T4MPyP) and found it to be an efficient cleaver (24). The DNA photocleavage properties of a series of cationic porphyrins including the zinc complexes, and porphyrin dimers were studied by Praseuth and co-workers (25). They found that the zinc complexes and the dimers were more efficient in cleaving DNA than the corresponding free bases and the monomers. Similarly, Munson et al. have demonstrated that cis and trans isomers of bis(N-methyl-4-pyridiumyl)diphenyl porphyrin intercalates into DNA and also cleaves DNA in the presence of light (26). A review of the literature showed that most of the cationic porphyrins studied for their DNA cleavage activity are metalloderivatives of porphyrins possessing cationic substituents at the meso positions and belong to the class of cleavers which needs an activating agent. The few exceptions in which the non-meso-substituted compounds have been studied for their DNA-cleaving abilities are the noncationic chlorins (27) and lutetium(III) texaphrin (28). The aim of this work was to investigate the photocleavage and DNA-binding properties of a series of non-meso-substituted cationic porphyrins to further understand the characteristics needed for efficient cleavage of DNA in the presence of light. The structural characteristics of this series of porphyrins that

10.1021/bc9800872 CCC: $18.00 © 1999 American Chemical Society Published on Web 11/07/1998

DNA Interaction and Photocleavage Properties of Porphyrins

Bioconjugate Chem., Vol. 10, No. 1, 1999 95

Figure 1. Structure of non-meso-substituted cationic porphyrins that were used in the study. Figure 3. General scheme for the synthesis of non-mesosubstituted cationic porphyrins.

Figure 2. Structure of porphyrins that were used as starting materials for synthesis of non-meso-substituted cationic porphyrins.

were varied to understand their effect on the DNA cleavage activity were (i) number of cationic substituents, (ii) the length of the carbon chain between the cationic group(s) and the porphyrin, (iii) the position of the cationic substituents on the porphyrin ring, and (iv) the effect of the coordinated metal. MATERIAL AND METHODS

Synthesis of Porphyrins. The cationic porphyrins 1-8 (shown in Figure 1) were prepared from their respective porphyrins 9-16 (shown in Figure 2) by following the general scheme shown in Figure 3 (29). For the preparation of the moncationic porphyrin 1, 8-vinyl porphyrin 9 obtained from dueteroporphyrin via a series of reactions, was used as the starting material. Reaction of 9 with Eschenmoser’s salt [CH2dN+(Me)2I-] led to the formation of the dimethyl amino derivative, which on treatment with methyl iodide gave the desired cationic porphyrin 1. Cationic porphyrins 2-8 were similarly prepared from porphryins 10-16, respectively. Porphyrin 15 in which the cationic substituents were introduced at the vinyl groups in diagonal position (ring A and ring C) to form porphyrin 7 was prepared by following the scheme shown in Figure 4 (30). In brief, the acetoxy methyl pyrrole 17 was condensed with the R-free pyrrole 18 in the presence of Montmorillonite K 10 and pyrromethane 19 was obtained in 85% yield (31). Reaction of 19 with TFA/triethyl orthoformate gave the

monoformyl pyrromethane 20 in 70% yield. Hydrogenation of the benzyl ester functionality gave pyrromethane 21 as monocarboxylic acid, which on self-condensation under McDonald’s reaction conditions produced bis(2chloroethyl) porphyrin 22. The chloroethyl functionalities were then converted to vinyl substituents by refluxing in NaOH/pyridine solution to obtain 16. The divinyl free porphyin 15, obtained upon removal of the zinc metal from porphyrin 16 was converted to the cationic porphyrin 7 by reaction with Eschenmoser’s salt, followed by treatment with methyl iodide. All the new compounds were characterized by NMR and mass spectrometry analysis. DNAs. Calf thymus DNA was purchased from Sigma Chemical and dissolved in BPES buffer (without NaCl) for the binding studies. Plasmid DNA used was the 4752 bp circular pCAT-control from Promega, amplified in Escherichia coli strain DH5R. The DNA was isolated and purified using cesium chloride and dissolved in TE buffer. Polydeoxyguanylic deoxycytidylic acid (poly[d(G.C)].poly[d(G.C)]) and polydeoxyadenylic thymidylic acid (poly[d(A.T)].poly[d(A.T)]) were purchased from ICN and were dissolved in BPES buffer (without NaCl) (3). Hypochromicity Studies. All UV-vis spectrophotometric measurements were done on Genesys 5 spectrophotometer (5 nm spectral bandwidth). The hypochromicity in the Soret band of the cationic porphyrins caused by their interaction with calf thymus DNA and the polynucleotides poly[d(G.C)].poly[d(G.C)] and poly[d(A.T)].poly[d(A.T)] was determined as follows. Initially, visible absorption spectra of these porphyrins was determined at a concentration of ∼10-6 M. Calf thymus DNA, poly[d(A.T)].poly[d(A.T)] or poly[d(G.C)].poly[d(G.C)] was added in increments of 1-2 µL until no further decrease in the Soret band was observed. Buffer blanks were used to compensate for the dilution effect. The percentage hypochromicity was determined from the equation (ODf - ODb/ODf) × 100, where ODf and ODb represent the optical density of the free porphyrin and the mixture of free porphyrin and porphyrin bound to DNA, respectively. Intrinsic Binding Constant Kb. The intrinsic binding constant Kb was determined by using the formula

[DNA]/(A - F) ) [DNA]/(B - F) + 1/Kb(B - F) where A, F, and B correspond to Aobsd/[porphyrin], the extinction coefficient for the free porphyrin and the

96 Bioconjugate Chem., Vol. 10, No. 1, 1999

Mettath et al.

Figure 4. Synthetic strategy for the preparation of symmetrical porphyrins 15 and 16.

extinction coefficient for the porphyrin complex in fully bound form, respectively. A plot of [DNA]/(A - F) vs [DNA] will have a slope of 1/(B - F) and a y-axis intercept equal to 1/Kb (B - F). The binding constant Kb was calculated from the ratio of the slope to the y-axis intercept. Circular Dichroism (CD) Studies. CD spectra were recorded on a Jasco J-600 spectropolarimeter with a spectral bandwidth of 2 nm. The path length was 1 mm. The DNA (calf thymus DNA, poly[d(A.T)].poly[d(A.T)] or poly[d(G.C)].poly[d(G.C)]) was added to the porphyrins, and after an incubation period of 15 min, the samples were scanned in the visible region (320-480 nm). The r value, the ratio of the concentration of the porphyrins to the concentration of DNA, was less than 0.5. The baseline correction was applied using the buffer before scanning the samples. DNA Unwinding Experiments. DNA unwinding assays were performed in two steps. First, closed supercoiled DNA (form I) pCAT (0.5 µg) was relaxed with wheat germ topoisomerase I (Promega Corp., Madison, WI). This was followed by a second topoisomerasemediated relaxation in the presence of porphyrin. Specifically, topoisomerase I (2 units) was incubated at 37 °C for 30 min with 0.5 µg of form I pGS81 DNA first in the absence and then for 30 min in the presence of porphyrin in 25 µL of buffer containing 25 mM Tris-HCl, pH 7.9, 25 mM NaCl, 1 mM EDTA, 1 mM DTT, and 10% glycerol. The assay was stopped with 3 µL of 10% SDS and subjected to 1% agarose gel electrophoresis at 15 V for 20 h. After addition of the cationic porphyrin to DNA, care was taken to prevent exposure of the mixture to light.

Exposure of DNA/Porphyrin to Light. The solution of cationic porphyrins was added to the plasmid DNA in a 1.5 mL eppendorf microtube such that the r value was 0.5 or less. The tubes were placed 4 cm from a 15 W fluorescent daylight lamp and exposed to light for times ranging 0-90 min. The photoexposure ranged from 4.0 × (10-5 to 10-4) J/cm2 among various experiments. After exposure of the DNA to light, 1/10 volume of a solution of 50% glycerol containing 0.25% bromophenol blue was added to the DNA sample and loaded on to a 1% agarose gel and run in a TAE buffer system at 80 V for approximately 1.5 h. The agarose gel was then stained in ethidium bromide (0.25 µg/mL) and photographed. The photographs were analyzed using Image Quant software to quantify the amount of plasmid DNA (form I) remaining after treatment with light. The singlet oxygen and hydroxyl radical quenching experiments were carried out in a manner similar to the cleavage reactions, with the addition of sodium azide (0.1 M) and mannitol (0.1 M), respectively. Sodium azide or mannitol was added at a final concentration of 0.1 M and exposed to light for varying amounts of time. RESULTS AND DISCUSSION

Binding Studies. UV-vis absorption spectroscopy, CD spectroscopy, and DNA unwinding assays were used to evaluate the binding of non-meso-substituted cationic porphyrins 1-8 to DNA. In general, the hypochromic shift in the Soret band of the porphyrins upon interaction with the DNA indicate binding of the porphyrin to DNA; the sign of the ellipticity obtained in the visible range of the CD scan can point to the nature of the binding.

DNA Interaction and Photocleavage Properties of Porphyrins

Bioconjugate Chem., Vol. 10, No. 1, 1999 97

Table 1 porphyrin

Soret band (nm)

% hypochromicity with CT.DNA

Kb

2 3 4 5 6 7 8

385 410 385 409 388 382 406

29.8 25.4 28.6 49.0 42.7 33.5 41.0

2.32 × 105 2.52 × 104 3.10 × 105 1.12 × 105 0.88 × 105 0.81 × 105 1.34 × 105

Table 2 porphyrin

Soret band (nm)

% hypochromicity with poly dA.dT

% hypochromicity with poly dG.dC

2 3 4 5 6 7 8

385 410 385 409 388 382 406

15.0 9.7 21.5 14.0 12.9 32.2 29.6

28.7 5.5 35.9 25.0 22.6 41.4 46.4

Similarly, the unwinding assay has been used to examine the nature of binding of porphyrins to DNA as described below (26, 32). (a) UV-Vis Absorption Spectroscopy. A summary of changes in the Soret band of the porphyrins upon interaction with calf thymus DNA is shown in Table 1. These studies were done at an ionic strength (µ) of 0.17. Sodium chloride was excluded in these solutions. Addition of NaCl to these solutions was seen to decrease the hypochromic effects. Porphyrin 1 did not show any hypochromic shift. Porphyrins 5, 6, and 8 exhibited hypochromic shifts greater than 35%. All the other porphyrins showed hypochromic shifts in the absorption band in the range 25-33% (Table 1). However, no change in the absorption wavelength of the Soret band has been observed. The intrinsic binding constant, Kb for the binding of these porphyrins to calf thymus DNA was determined by the method reported in the literature (33-35). The binding constant for all the compounds 2-8 except 4 were found to be in the range of ∼105; for 4 it was in the range of 104. However, because of the limitations imposed due to the aggregation of the cationic porphyrins, some variation in the value of the intrinsic binding constant, Kb cannot be ruled out. Attempts to find the number of binding sites by Scatchard plot proved to be difficult due to aggregation. Similarly, Sehlstedt et al. have also reported difficulty in finding the DNA binding constants of cobalt porphyrins due to aggregation effects (36). The hypochromic shift in the Soret band of the porphyrins upon interaction with copolymers poly[d(A.T)].poly[d(A.T)] and poly[d(G.C)].poly[d(G.C)] was also determined and are shown in Table 2. The hypochromic shifts in the Soret band in the range 9.7-29.6% were observed upon binding of these porphyrins to poly[d(A.T)].poly[d(A.T)] and in the range 5.5-46.4% upon unbinding to poly[d(G.C)].poly[d(G.C)]. (b) CD Studies. CD studies have been extensively used to study the binding of small molecules including porphyrins to DNA (3, 5, 35-37). In the visible range (320-480 nm), the porphyrins alone as well as the DNA by themselves do not show any ellipticity. However, an induced CD spectrum is obtained if there is a binding between the porphyrin and the DNA. The shapes of the curves obtained gave information about the nature of the binding. The mode of binding of the porphyrins 2-8 to calf thymus DNA, the copolymers poly[d(G.C)].poly[d(G.C)] and poly[d(A.T)].poly[d(A.T)] were hence investigated by CD. None of these porphyrins showed any

Figure 5. Visible CD spectra of CT. DNA in BPES/no salt buffer with porphyrins 2-8. (A) 2 (line a), 3 (line b), 4 (line c). (B) 5 (line b), 6 (line d), 7 (line c), 8 (line a). R < 0.5, µ ) 0.017.

ellipticity in the visible range (320-480 nm) or in the UV range (220-300 nm). Interaction of the porphyrins 2-8 with calf thymus DNA and copolymers poly[d(A.T)].poly[d(A.T)] and poly[d(G.C)].poly[d(G.C)] resulted in the formation of induced CD spectrums in the visible range (320-400 nm). The results of these interactions are shown in Figures 5 (interaction with calf thymus DNA), 6 (interaction with poly[d(A.T)].poly[d(A.T)]), and 7 (interaction with poly[d(G.C)].poly[d(G.C)]). Porphyrin 2 (line a, Figure 5A) upon binding to calf thymus, DNA showed a strong positive peak centered at 396 nm, whereas compound 3 (line b, Figure 5A), the zinc complex, showed a weak positive peak and a weak negative peaks centered at 413 and 394 nm, respectively. Porphyrin 4 (line c, Figure 5A) and its metallocomplexes 5 (line b, Figure 5B) and 6 (line d, Figure 5B) all showed negative ellipticity; however, the intensity of the peaks as well as the wavelength at maximum peak height differed. Similarly, compound 7 (line c, Figure 5B) showed a strong negative peak upon interaction with calf thymus DNA. For the zinc complex 8, a very weak ellipticity pattern was obtained, and changes in the r value did not cause any change in the shape of the spectra. The visible CD spectrums obtained upon binding of porphyrins 2-8 with poly[d(A.T)].poly[d(A.T)] are shown in Figure 6. Porphyrin 2 upon interaction with poly[d(A.T)].poly[d(A.T)] (line a, Figure 6a) showed a strong positive peak centered at 393.0 nm, whereas for porphyrins 3 and 4, very weak and broad negative peaks (lines b and c in Figure 6A, respectively) were obtained. Compound 6 showed strong negative peaks centered at 402 nm (line b, Figure 6B), whereas compounds 5 and 8 exhibited broad peaks centered at 396 and 412 nm, respectively (lines a and d, Figure 6B). The interaction of compound 7 with poly[d(A.T)].poly[d(A.T)] did not result in any ellipticity.

98 Bioconjugate Chem., Vol. 10, No. 1, 1999

Figure 6. Visible CD spectra of poly[d(A-T)].poly[d(A-T)] in BPES/no salt buffer with porphyrins 2-8. (A) 2 (line a), 3 (line b), 4 (line c). (B) 5 (line d), 6 (line b), 7 (line c), 8 (line a). R < 0.5, µ ) 0.017.

Figure 7. Visible CD spectra of poly[d(G-C)].poly[d(G-C)] in BPES/no salt buffer with porphyrins 2-8. (A) 2 (line b), 3 (line a), 4 (line c). (B) 5 (line d), 6 (line b), 7 (line a), 8 (line c). R < 0.5, µ ) 0.017.

The interaction of porphyrins 2-8 with poly[d(G.C)]. poly[d(G.C)] is shown in the Figure 7. A strong positive peak was obtained in the visible CD spectra for compound 3 (line a, Figure 7A), upon binding to poly[d(G.C)].poly[d(G.C)]. In the visible range, the binding of all the other porphyrins to poly[d(G.C)].poly[d(G.C)] resulted in nega-

Mettath et al.

tive ellipticity. These interactions are represented by lines b and c in Figure 7A, for 2 and 4, respectively, and lines a-d in Figure 7B for the compounds 7, 6, 8, and 5, respectively. (c) Unwinding Assay. The ability of compounds to intercalate to DNA can be checked by the unwinding assay using topoisomerase I (26, 32). The DNA that is used in this assay is negatively supercoiled DNA. The supercoiled DNA is first converted to relaxed supercoiled DNA using topoisomerse I. The compounds were then added to interact with the DNA, and the assay was continued for 30 min in the presence of topoisomerase I. Compounds that intercalate into DNA induce supercoiling of DNA and, beyond a certain r value, complete supercoiling to negatively supercoiled DNA is obtained. The binding of porphyrins 2-8 were studied using unwinding assay and are shown in Figure 8 a-d. Porphyrins 2, 3, 4, 6, and 7 [porphyrins 2, 3, 4 (Figure 8, panels a-c), 6 (Figure 8D, lanes 11-13), and 7 (Figure 8D, lanes 2-4)], exhibit only a very weak unwinding of DNA, whereas porphyrins 5 (lanes 8-10, Figure 8D) and 8 (lanes 5-7, Figure 8D) caused complete unwinding to form supercoiled DNA. These unwinding assays indicate that compounds 5 and 8 are able to intercalate into DNA. Photosensitized Cleavage of DNA. A convenient way to assess the ability of a cationic porphyrin to photocleave DNA is to measure the conversion of supercoiled DNA (form I) to form II (26). In the presence of light, all the cationic porphyrins, except 6 and 1, were able to nick closed circular DNA (form I) to convert it into relaxed closed circular DNA (form II) in varying amounts. The amount of form II created depended upon the amount of exposure to visible light, the porphyrin that was used, and ratio of the concentration of the porphyin to that of the DNA. DNA (form I) was quantified by densitometry studies using Image Quanta software. Upon exposure to light for 90 min, all the porphyrins (r ) 0.5), except 1 and 6, were able to convert over 80% of DNA (form I) to DNA (form II) and not much difference in cleavage efficiency was observed. However, upon exposure to light for shorter time periods up to 60 min, differences in the efficiency of DNA cleavage by various porphyrins were observed. In the case of compound 5 (Figure 9b, lanes 13-16), within 15 min of exposure to light (lane 13), over 95% of the DNA (form I) was converted to form II, and 100% conversion was seen upon exposure to light for 60 min (lane 15). For compounds 2, 3, and 4, the amount of DNA converted to form II in 15 min was 38.33, 49.81, and 50.93%, respectively, as shown in lanes 2, 7, and 12 of Figure 9a, respectively. Compounds 7 and 8 caused the conversion of 33.44 and 39.20% of the DNA form I to form II in 15 min (lanes 3 and 8 of Figure 9b, respectively). Compound 6 (lanes 1820), the nickel complex, did not cause any cleavage of DNA (form I) even after exposure to light for an extended period of time (90 min). Similarly, the monocationic compound 1 also did not cause any cleavage of DNA (results not shown). The comparison of the DNA cleaving efficiencies of these porphyrins is shown in Table 3 (37). The species responsible for causing the DNA cleavage is believed to be singlet oxygen. Singlet oxygen is generated upon transfer of energy from an excited porphyrin in the triplet state to oxygen in ground state (triplet state). The transfer of energy causes the conversion of oxygen to excited singlet state. The Ni complex 6 does not generate singlet oxygen, and this could explain the noncleavage of the DNA by this compound. For the other porphyrins, if singlet oxygen is the primary agent responsible for the cleavage of DNA, irradiation of the

DNA Interaction and Photocleavage Properties of Porphyrins

Bioconjugate Chem., Vol. 10, No. 1, 1999 99

B

A

C

D

Figure 8. Agarose gel electrophoresis of supercoiled DNA relaxed by topoisomerase I in the presence of cationic non-meso-substituted porphyrins 2-8. (A) Lane 1 is DNA form I and relaxed DNA. Lane 2 is control, DNA relaxed in the presence of topoisomerase I. Lanes 3 and 4, DNA relaxed in the presence of 19 and 38 µM of porphyrin 2, respectively. (B) Lane 1 is DNA form I and relaxed DNA; lane 2 is control, i.e., DNA relaxed in the presence of topoisomerase I. Lanes 3-5 contain DNA relaxed in the presence of 14, 28, and 56 µM of porphyrin 3, respectively. (C) Lane 1 is DNA form I. Lanes 2 and 6 are control, DNA relaxed in the presence of topoisomerase I. Lanes 3-5 DNA relaxed in the presence of 28, 42, and 60 µM of porphyrin 4, respectively. (D) Lane 1 is control, DNA form I treated with topoisomerase I, lanes 2-4, DNA relaxed in the presence of 14, 10, and 7 µM of porphyrin 7, respectively. Lanes 5-7, DNA relaxed in the presence of 14, 10, and 7 µM of porphyrin 8, respectively. Lanes 8-10, DNA relaxed in the presence of 14, 10, and 7 µM of porphyrin 5, respectively. Lanes 11-13 DNA relaxed in the presence of 14, 10, and 7 µM of porphyrin 6, respectively.

A

B

Figure 9. Agarose gel electrophoresis of plasmid DNA (form I) photosensitized with porphyrins 2-8. Plasmid DNA (10 µL, 1 µg/10 µL) was added to eppendorf microtubes containing the porphyrins (10 µL, 75 µmol) and exposed to light for durations ranging 15-90 min and subsequently separated on 1% agarose gel. (A) Photocleavage of plasmid DNA (form I) by porphyrins 2, 3, and 4. Lanes 1, 6, and 11 are controls and contain no porphyrins. Lanes 2-5: DNA + porphyrin 2, treated with light for 15, 30, 60, and 90 min, respectively. Lanes 7-10: DNA + porphyrin 3, treated with light for 15, 30, 60, and 90 min, respectively. Lanes 12-15: DNA + porphyrin 4, treated with light for 15, 30, 60, and 90 min, respectively. (B) Photocleavage of plasmid DNA (form I) by porphyrins 5, 6, 7, and 8. Lanes 1, 7, 12, and 17 are controls and contains no porphyrins. Lanes 2-6: DNA + porphyrin 7 treated with light for 0, 15, 30, 60, and 90 min, respectively. Lanes 8-11: DNA + porphyrin 8 treated with light for 15, 30, 60, and 90 min, respectively. Lanes 13-16: DNA + porphyrin 5 treated with light for 15, 30, 60, and 90 min, respectively. Lanes 18-20: DNA + porphyrin 6 treated with light for 15, 60, and 90 min, respectively.

porphyrin-DNA complex in the presence of a singlet oxygen quencher should result in the decrease of the amount of DNA cleavage (16, 24). As expected, a decrease in the cleavage of DNA was observed upon irradiation in the presence of sodium azide (100 mM), a well-known 1O2 quencher (see Figure 10). No decrease in the cleavage was observed when the studies were done

in the presence of equivalent concentration of NaCl (100 mM). DNA cleavage was also studied in the presence of hydroxyl radical scavenger mannitol (100 mM), and no discernible decrease in the cleavage of the DNA was observed (results not shown). These results indicate that singlet oxygen is the major species responsible for the cleavage of DNA.

100 Bioconjugate Chem., Vol. 10, No. 1, 1999

Mettath et al.

Figure 10. Agarose gel electrophoresis of plasmid DNA (form I) photosensitized with porphyrins 2-8 in the presence of 0.1 M sodium azide for 45 min. Lane 1 contains only DNA. Lanes 2 and 3: DNA phostosenstized with porphyrin 2 in the presence and absence of 0.1 M sodium azide, respectively. Lanes 4 and 5: DNA phostosensitized with porphyrin 3 in the presence and absence of 0.1 M sodium azide, respectively. Lanes 6 and 7: DNA phostosenstized with porphyrin 4 is the presence and absence of 0.1 M sodium azide, respectively. Lanes 8 and 9: DNA phostosensitized with porphyrin 5 is the presence and absence of 0.1 M sodium azide, respectively. Lanes 10 and 11: DNA phostosensitized with porphyrin 7 is the presence and absence of 0.1 M sodium azide, respectively. Lanes 12 and 13: DNA phostosensitized with porphyrin 8 in the presence and absence of 0.1 M sodium azide, respectively. Table 3 porphyrin 1 2 3 4 5 6 7 8

% of DNA form I cleaved at 15 min

% of DNA form I cleaved at 60 min

38.33 49.81 50.73 95.86

88.31 93.67 91.22 100

33.44 39.20

80.44 80.27

Binding Characteristics. The hypochromicity values obtained for the interaction of compounds 5, 6, and 8, with calf thymus DNA are 49.0, 42.7, and 41.0%. The hypochromicity values obtained for the other porphyrins are in the range 25.4-33.5%. In the literature, hypochromicity values greater than 35% have been considered as one of the characteristics of intercalators (1, 3, 5). Interaction of all the porphyrins, with poly[d(A.T)].poly[d(A.T)] and poly[d(G.C)].poly[d(G.C)], also resulted in hypochromic shifts in the absorbance of the Soret band. However, for all the porphyrins except 3, higher hypochromicity value was obtained upon interaction with poly[d(G.C)].poly[d(G.C)] than poly[d(A.T)].poly[d(A.T)], probably indicating a slight preference for GC rich sites. The nature of the binding of the porphyrins to the DNA has been generally interpreted from the shape of the induced CD bands obtained upon the interaction of the porphyrins with calf thymus DNA, poly[d(G.C)].poly[d(G.C)] and poly[d(A.T)].poly[d(A.T)]. In general, intercalation of the porphyrins is usually seen at GC-rich sites and is denoted by negative peaks in the CD spectrums with calf thymus DNA and poly[d(G.C)].poly[d(G.C)] (3, 5, 38-40), whereas outside binding has usually been observed at AT-rich sites and is denoted by positive peaks in the CD spectrums with calf thymus DNA and poly[d(A.T)].poly[d(A.T)] (5, 38, 39). The interpretations of the CD spectrums obtained for all porphyrins upon interaction with calf thymus DNA, poly[d(A.T)].poly[d(A.T)] and poly[d(G.C)].poly[d(G.C)] as well as the results from the unwinding assay can be interpreted as follows. (a) Porphyrin 2 shows a positive peak with calf thymus DNA (Figure 5a) and poly[d(A.T)].poly[d(A.T)] (Figure 6a) and weak positive and negative peaks with poly[d(G.C)].poly[d(G.C)] (Figure 7a), indicating an outside binding mode with preference to AT. The negative peak with poly dG.dC does point toward a small amount of binding at GC sites. The unwinding assay also indicates that this compound does not intercalate (Figure 8a). (b) Porphyrin 3 shows a weak positive peak and a weak negative peak with calf thymus DNA (Figure 5a),

a weak negative peak with poly[d(A.T)].poly[d(A.T)] (Figure 6a), and a strong positive peak with poly[d(G.C)].poly[d(G.C)] (Figure 7a). Binding at GC-rich sites is generally by intercalation and indicated by a negative peak upon binding to poly[d(G.C)].poly[d(G.C)], hence the positive peak upon interaction with GC-rich sites could indicate an outside binding mode. (c) Porphyrins 4, 6, and 7; all three porphyrins show a strong negative peak in the visible CD with CT.DNA (Figure 5), a weak negative peak with poly[d(A.T)].poly[d(A.T)] (Figure 6), and a strong negative peak with poly[d(G.C)].poly[d(G.C)] (Figure 7). The nature of the peaks obtained indicate an intercalative mode of binding with preferential binding to GC-rich sites, with some binding to AT sites. However, the unwinding assays (Figure 8, panels c and d) indicate that these compounds are able to unwind the DNA to only a smaller extent. The combined results from these two assays point toward an outside binding mode with at GC-rich sites. (d) Porphyrin 5 shows a broad negative peak with calf thymus DNA and poly[d(A.T)].poly[d(A.T)] and a strong negative peak with poly[d(G.C)].poly[d(G.C)]. The CD spectrums indicate an intercalative mode of binding at GC-rich sites and also a weak binding at AT sites. However, the negative CD pattern upon interaction with poly[d(A.T)].poly[d(A.T)] is unusual as compared to meso-substituted porphyrins, wherein a positive ellipticity is obtained for binding at AT sites. The unwinding assay also indicates that this porphyrin does intercalate into DNA. (e) Porphyrin 8 shows very weak negative peak in the visible range spectrums with calf thymus DNA (Figure 5b) and poly[d(A.T)].poly[d(A.T)] (Figure 6b). A strong negative peak as well as a small positive peak is obtained in visible range with poly[d(G.C)].poly[d(G.C)] (Figure 7b). The negative peak obtained upon binding of 8 to poly[d(G.C)].poly[d(G.C)] indicates intercalation at GC sites. The small positive peak does point toward a small amount of outside binding at GC-rich sites. The unwinding assay also indicates an intercalative mode of binding. DNA Cleavage Characteristics. The amount of DNA cleaved by the various porphyrins upon exposure to light for 15 and 60 min is shown in Table 3. The factors affecting the efficiency of DNA cleavage seem to be complex (37), however, certain porphyrins, structural characteristics needed for efficient cleavage of DNA can be elicited from these data. Among the factors that affect the cleavage efficiency are (a) the number of charges, (b) length and position of the side chain connecting the cationic substituent to the porphyrin macrocycle, and (c) the presence of chelated metal at the core of the porphyrin ring.

DNA Interaction and Photocleavage Properties of Porphyrins

(a) Number of Charges. Our studies with various cationic porphyrins show that a minimum of two charges is needed for the binding and cleavage of DNA by cationic porphyins. For example, montocationic porphyrin 1 did not show any cleavage of DNA even after incubation with light 90 min under the condition used for all the other porphyrins. (b) Length of the Side Chain of the Cationic Substituents. Dicationic porphyrins 2 and 4 are both linked to the porphyrin macrocycle by side chains at positions 3 and 8. However, in porphyrin 4, the length of the side chain is extended by two more carbon atoms. This increase in the length of the side chain is seen to increase the efficiency of cleavage. Porphyrin 2 is able to convert 38.33% of DNA form I to form II and form III in the presence of light for 15 min; however, under the same conditions, porphyrin 4 is able to cleave 50.73% of DNA form I to DNA form II. (c) Position of the Cationic Substituent on the Porphyrin System. In our attempts to understand the effect of the position of the cationic charges, porphyrins 4 and 7 were used as sensitizers. In porphyrin 4, the charged group is substituted at positions 3 and 8, whereas in compound 7, the charged groups are present at positions 3 and 13 of the porphyrin ring system. Porphyrin 4 converted 50.73% of the DNA form I to form II and form III in 15 min (Figure 9a, lane 12), and under the same conditions, porphyrin 7 was able to convert 33.44% of the DNA form I to form II (Figure 9b, lane 3). (d) Effect of Chelated Metal. Compared to the nonmetalated analogues 2, 4, and 7, the zinc analogues 3, 5, and 8 were found to have a higher efficiency of photocleavage of DNA. CONCLUSION

The absorption, circular dichroism, unwinding assays, and the DNA photocleavage evidence from the present study shows that non-meso-substituted cationic porphyrins bind, as well as photocleave, DNA efficiently. On the basis of the DNA-binding studies and unwinding assays, it is suggested that the nature of binding of nonmeso-substituted cationic porphyrins is different from that of the cationic meso-substituted pyridyl porphyrins. The results from the absorption spectroscopy, CD, and unwinding studies point toward intercalation as a mode of binding for the Zn(II) porphyrins 5 and 8. The nature of the binding obtained for zinc porphyrins 5 and 8 is surprising, due to the presence of the nonplanar substituents as well as the presence of metal zinc, both of these characteristics are thought to block intercalation (5, 41, 42). Furthermore, the negative ellipticity obtained in the visible CD spectrums with calf thymus DNA and poly[d(G.C)].poly[d(G.C)] for porphyrins 4, 6, and 7 indicate an intercalative mode of binding; however, this binding mode is not supported by unwinding assay. Again, the positive peak with poly[d(G.C)].poly[d(G.C)] and negative peak with poly[d(A.T)].poly[d(A.T)] for porphyrin 3 completely differ from the ellipticity obtained under similar conditions for meso-substituted cationic porphyrins. These studies show that the binding of porphyrins is complex, and from the presently existing model with mesosubstituted cationic porphyrins, it is difficult to come to any precise conclusion regarding the binding mode obtained for non-meso-substituted cationic porphyrins. Clearly, further research in this area is needed to understand the binding characteristics of these types of porphyrins to DNA. To shed more light on the exact

Bioconjugate Chem., Vol. 10, No. 1, 1999 101

nature of the binding of these porphyrins to DNA, molecular modeling and DNA footprinting studies are underway. ACKNOWLEDGMENT

We thank Drs. R. J. Fiel, T. J. Dougherty, and R. Shibata for their suggestions during this project. We also thank Drs. M. J. Levine and N. Ramasubbu (Research Center in Oral Biology, SUNY, Buffalo) for allowing us the use of Jasco J-600 spectropolarimeter (supported by Grant DE 08240). LITERATURE CITED (1) Sigman, D. S. (1990) Chemical Nucleases. Biochemistry 29, 9097-9105. (2) Fiel, R., Jenkins, B., and Alderfer, J. (1990) Cationic Porphyrins - DNA Complexes. in 23rd Jerusalem Symposium in Quantum Chemistry and Biochemistry (B. Pullman and J. Jortner, Eds.) Kluwer Academic Press, Netherlands. (3) Fiel, R. J., Howard, J. C., Mark, E. H., and Datta-Gupta, N. (1979) Interaction of DNA with porphyrin ligand: evidence for intercalation. Nucleic Acids Res. 6, 3093-3118. (4) Banville, D. S., Marzill, L. G., and Wilson, W. D. (1983) 31P NMR and viscometric studies of the interaction of meso-tetra(4-N-methylpyridyl) porphine and its Ni(II) and Zn(II) derivatives to DNA′. Biochem. Biophys. Res. Commun. 113, 148154. (5) Pasternack, R. F, Gibbs, E. J., and Villafranca, J. J. (1983) Interaction of porphyrins with Nucleic Acids. Biochemistry 22, 2406-2414. (6) Fiel, R. J. (1989) Porphyrin Nucleic Acid Interactions: A review. J. Biomol. Struct Dyn. 6, 1259-1274. (7) Ward, B., Skorobogaty, A and Dabrowiak, J. C. (1986) DNA cleavage specificity of a group of cationic metalloporphyrins. Biochemistry 25, 6875-6883. (8) Ford, K. G., and Neidle, S. (1995) Perturbations in DNA structure upon interaction with porphyrins revealed by chemical probes, DNA footprinting and molecular modelling. Biorg. Med. Chem. 3, 671-677. (9) Bromley, S. D., Ward, B. W., and Dabrowiak, J. C. (1986) Cationic porphyrins as probes of DNA structure. Nucleic Acids Res. 14, 9133-9147. (10) Ward, B., Skorobogaty, A., and Dabrowiak, J. C. (1986) DNA binding specificity of a series of cationic metalloporphyrin complex. Biochemistry 25, 7827-7833. (11) Fiel R. J, Beerman, T. A, Mark, E. H, and Datta-Gupta, N. (1982) DNA strand scission activity of metalloporphyrins. Biochem. Biophys. Res. Commun. 107, 1067-1074. (12) Byrnes, R. W., Fiel, R. J., and Datta-Gupta, N. (1988) DNA strand scission by iron comples of meso-Tetra(N-Pyridyl)Porphines Chem.-Biol Interact. 67, 225-241. (13) Fouquet, E., Pratviel, G., Bernadou, J., and Meunier, B. (1987) Nuclease activity of a water soluble manganese porphyrin associated with potassium hydrogen persulphate: oxidative cleavage of DNA. J. Chem. Soc., Chem. Commun. 1169-1170. (14) Pratviel, G., Bernadou, J., Ricci, M., and Meunier, B. (1989) Magnesium monoperoxophthalate: An efficient single oxygen atom donor in DNA cleavage catalysed by mettaloporphyrins. Biochem. Biophys. Res. Commun. 160, 1212-1218. (15) Ding, L., Etemad-Moghadam, G., Cros, S., Auclair, C., and Bernard, M. (1989) Cytotoxic hybrid molecules Metalloporphyrin - Ellipticine having a high affinity for DNA. J. Chem. Soc., Chem. Commun. 1711-1713. (16) Sentagne, C., Meunier, B., and Paillous, N. (1992) DNA cleavage photoinduced by new water - soluble zinc porphyrins linked to 9-methoxyellipticine. J. Photochem. Photobiol. B 16, 47-59. (17) Jakobs, A., Bernadou, J., and Meunier, B. (1997) Preparation of teracationic metalloporpyrin-spermine conjugates. J. Org. Chem. 62, 3505-3510. (18) Frau, S., Bernadou, J., and Meunier, B. (1997) Nuclease activity and binding characteristics of a cationic “Manganese

102 Bioconjugate Chem., Vol. 10, No. 1, 1999 Porphyrin-Bis(benzimidazole) Dye (Hoechst 33259)”. Bioconjugate Chem. 8, 222-231. (19) Bigey, P., Sonnichsen, S. H., Meunier, B., and Nielsen, P. (1997) DNA binding and cleavage by a cationic manganese porphyrin-peptide nucleic acid conjugate. Bioconjugate Chem. 8, 267-270. (20) Mestre, B., Jakobs, A., Pratviel, G, and Meunier, B. (1996) Structure/nuclease activity relationships of DNA cleavers based on cationic metalloporphyrin-oligonucleotide conjugates. Biochemistry 35, 9140-9149. (21) Bigey, P., Pratviel, G., and Meunier, B. (1995) Cleavage of double stranded DNA by metalloporphyrin - linker - oligonucleotide, molecules: influence of the linker. Nucleic Acid Res. 23, 3894-3900. (22) Dougherty, T. J., Gomer, C. J., Henderson, B. W., Jori, G., Kessel, D., Korbelik, M., Maon, J., Peng, Q. (1998) Photodynamic Therapy. J. Natl. Cancer Inst. 90, 889-905. (23) Henerson, B. W., and Dougherty, T. J., Eds. (1992) Photodynamic Therapy: Basic principles and clinical applications, New York, Marcel Dekker. (24) Fiel, R. J., Datta-Gupta, N., Mark, E. H., and Howard, J. C. (1981) Induction of DNA damage by porphyrin photosensitizers. Cancer Res. 41, 3543-3545. (25) Prasueth, D., Gaudemer, A., Verlhac, J., Kraljic, I., Sissoeff, I., and Guille, E. (1986) Photocleavage of DNA in the presence of synthetic water soluble porphyrins. Photochem. and Photobiol. 44, 717-724. (26) Munson, B. R. and Fiel, R. J. (1992) DNA intercalation and photosensitization of cationicmeso-substituted porphyrin. Nucleic Acid Res. 20, 1315-1319. (27) Boutorine, A. S., Brault, D., Takasugi, M., Delgado, O., and Helene, C., (1996) Chlorin-oligonucleotide conjugates: synthesis, properties and red light-induced photochemical sequence-specific DNA cleavage in duplexes and triplexes. J. Am. Chem. Soc. 118, 9469-9476. (28) Magda, D., Wright, M., Miller, R. A., Sessler, J. L., and Sansom, P. I. (1995) Sequence specific photocleavage of DNA by an expanded porphyrin with irradiation above 700 nm. J. Am. Chem. Soc. 117, 3629-3630. (29) Pandey, R. K., Shiau, F., Smith, N. W., Dougherty, T. J., and Smith, K. M. (1992) Syntheses of water soluble cationic porphyrins and chlorins. Tetrahedron 48, 7591-7600. (30) Pandey, R. K., Shiau, F. Y., Ramachandran, K., Dougherty, T. J., and Smith, K. M. (1992) Long wavelength photosensitizers related to chlorins and bacteriochlorins for use in photodynamic therapy. J. Chem. Soc., Perkin Trans. 1 13771385. (31) Jackson, A. H., Pandey, R. K., Roberts, E., and Rao, K. R. N. (1985) Reaction on solid supports Part II: A conveninet

Mettath et al. method for synthesis of pyrromethanes using a Montmorillonite clay as catalyst. Tetrahedron Lett. 26, 793-796. (32) Keller, W., (1975) Determination of the number of spherical turns in simian virus 40 DNA by gel electrophoresis. Proc. Natl. Acad. Sci. U.S.A. 72, 4876-4880. (33) Wolfe, A., Shimer, G. H., and Meehan, T., (1987) Polycyclic aromatic hydrocarbon physically intercalate into duplex region of denatured DNA. Biochemistry 26, 6392-6396. (34) Pyle, A. M., Rehmann, J. P., Mesoyrer, R., Kumar, C. V., Turro, N. J., and Barton, J. K. (1989) Mixed-ligand complexes of ruthenium(II): factors governing binding to DNA. J. Am. Chem. Soc. 111, 3051-3058. (35) Bhattacharya, S., and Mandal, S. S. (1996) DNA cleavage by intercalatable cobalt-bispicoylamine complexes activated by visible light. Chem. Commun. 1515-1516. (36) Sehlstedt, U., Kim, S. K., Carter, P., Goodisman, J., Vollano, J. F., Norden, B., and Dabrowiak, J. C. (1994) Interaction of cationic porphyrins with DNA. Biochemistry 33, 417-426. (37) The factors influencing the relative photocleavage efficiencies of the different chromophores, other than their structural differences, are complex and could reflect affects such as binding affinities and singlet oxygen quantum yield. As suggested by one of the reviewers, another important factor that can effect the DNA cleaving efficiency is the exact number of porphyrin molecules bound to double stranded DNA, which is difficult to establish, due to aggregation effects. (38) Carvlin, M. J., Datta-Gupt, N., and Fiel, R. J. (1982) Circular dichroism spectroscopy of a cationic porphyrin bound to DNA. Biochem. Biophys. Res. Commun. 108, 66-73. (39) Carvlin, M. J., and Fiel, R. J. (1983) Intercalative and nonintercalative binding of a large cationic porphyrin ligand to calf thymus. Nucleic Acid Res. 11, 6121-6140. (40) Carvlin, M. J., Mark, E., and Fiel, R., Intercalative and nonintercalative binding of a large cationic porphyrin ligands to polynucleotides. Nucleic Acids Res. 11, 6141-6154. (41) Sari, M. A., Battioni, J. P., Dupre, D., Mansuy, D., and Lepecq, J. B. (1996) Mode of interaction and apparent binding constants of meso-tetraaryl porphyrins bearing between one and four positive charges with DNA. Biochem. Biophys. Res. Commun. 141, 643-649. (42) Sari, M. A., Battioni, J. P., Dupre, D., Mansuy, D., and Lepecq, J. B. (1990) Interaction of cationic porphyrins with DNA: Importance of the number and position of the charges and minimum structural requirements for intercalation. Biochemistry 29, 4205-4215.

BC9800872