Influence of the A-Ring on the Redox and Nuclease Properties of the

Properties of the Prodigiosins: Importance of the. Bipyrrole Moiety in Oxidative DNA Cleavage. Matt S. Melvin, M. Wade Calcutt, Ronald E. Noftle, and ...
0 downloads 0 Views 107KB Size
Download PDF
742

Chem. Res. Toxicol. 2002, 15, 742-748

Influence of the A-Ring on the Redox and Nuclease Properties of the Prodigiosins: Importance of the Bipyrrole Moiety in Oxidative DNA Cleavage Matt S. Melvin, M. Wade Calcutt, Ronald E. Noftle, and Richard A. Manderville* Department of Chemistry, Wake Forest University, Winston-Salem, North Carolina 27109-7486 Received January 9, 2002

Prodigiosin (Prod, 1) is the parent member of a class of polypyrrole natural products that exhibit promising immunosuppressive and cytotoxic activity. They can facilitate copperpromoted oxidative double-strand (ds) DNA cleavage through reductive activation of Cu(II). This is triggered by oxidation of the electron-rich Prod molecule and may provide a basis for the cytotoxicity of the prodigiosins. To gain an understanding of this activity, we prepared several Prod analogues with various A-ring systems to examine their electrochemical properties in acetonitrile (MeCN) as a means to establish a basis for structure-reactivity relationships in copper-promoted nuclease activity. The intact bipyrrole (BP) chromophore is critical for the copper-mediated nuclease properties of the Prods. In fact, simple BP systems are shown to facilitate oxidative single-strand (ss) DNA cleavage. Replacement of the Prod A-pyrrole ring with alternative arenes (phenyl, furan-2-yl, or thiophen-2-yl) inhibits DNA strand scission and raises the half-peak oxidation potential (Ep/2) of the Prod free base [Ep/2 ) 0.44 V vs saturated calomel electrode (SCE) in MeCN] by ca. 200 mV. The same effect was achieved through attachment of an electron-withdrawing group (acetyl) at the 5′-position of the A-pyrrole ring. The structural modifications that inhibit DNA cleavage correlate with known structurereactivity relationships of Prods against leukemia and melanoma cancer cells. The implications of our findings with regard to the cytotoxicity of the Prods are discussed.

Introduction Prodigiosin (Prod,1 1, Figure 1) is the parent member of a family of naturally occurring red pigments produced by microorganisms such as Streptomyces and Serratia (1-3) that possess a pyrrolylpyrromethene skeleton with a C-4 methoxy group (4, 5). Other natural products within this structural class (Figure 1) include the tambjamine alkaloids (i.e., 2) that have been isolated from marine organisms (6-8) and the blue pigment 3 found in tambjamine (7) and Prod fractions (9). Collectively, 1-3 are derived from the bipyrrole (BP) aldehyde 4 (4, 5), which is the hydrolysis product of 2 (6, 10). A closely related derivative is roseophilin 5 that was isolated by Hayakawa et al. in 1992 from Streptomyces griseoviridis (11). Like the Prods (12-15), roseophilin 5 exhibits promising cytotoxicity in the submicromolar range (11).2 In the early 1960s, the Prods were studied as antimalarial and cytotoxic agents (1). However, during the past decade they have been shown to possess useful immunosuppressive activity when administered at nanomolar doses (16-19). Here they appear to reduce IL-2 signal transduction by inhibiting phosphorylation and * Address correspondence to this author at the Department of Chemistry, Wake Forest University, Winston-Salem, NC 27109-7486. Tel: (336)-758-5513, fax: (336)-758-4656, email: [email protected]. 1 Abbreviations: Prod, prodigiosin; BP, bipyrrole; ss, single-strand; ds, double-strand; CV, cyclic voltammetry; Ep/2, half-peak potential; SCE, saturated calomel electrode; TBAP, tetrabutylammonium perchlorate. 2 Information concerning the cytotoxicity of prodigiosin (NSC number 47147-F) and roseophilin (5‚HCl, NSC number 650718) against various human cancer cell lines in vitro is available from the NCI database on the Internet at http://www.dtp.nci.nih.gov.

Figure 1. Structure of prodigiosin (Prod, 1) and related natural products.

activation of JAK-3 at the IL-2 γ-chain (18). The resulting inhibition of T-cell proliferation occurs by a mechanism distinct from that of cyclosporin A, FK506, or rapamycin, suggesting that the Prods could be used synergistically in combination with existing drugs, or as alternative immunosuppressive therapies. As a result of this activity, D’Alessio and co-workers (20) undertook a medicinal

10.1021/tx025508p CCC: $22.00 © 2002 American Chemical Society Published on Web 04/17/2002

Prodigiosins

Chem. Res. Toxicol., Vol. 15, No. 5, 2002 743

chemistry program in search of synthetic Prods with higher in vitro immunosuppressive activity:cytotoxicity ratios (selectivity index) than the natural derivative undecylprodigiosin (6, Figure 1). These efforts established that a nitrogen-containing heterocyclic A-ring, with the lone-pair nitrogen electrons in conjugation with the tricyclic frame, and a C-4 alkoxy [as reported earlier by Boger (5)] are important for biological activity. It was also found that the electron density of the A-ring plays an important role, as the presence of electron-donating substituents enhanced potency, whereas electron-withdrawing substituents affected a marked loss in activity. In our studies, we have demonstrated that tambjamine E (2, Figure 1) binds DNA effectively (10) and facilitates single-strand (ss) DNA cleavage in the presence of Cu(II) and O2 (21). We have also determined that Prod 1 (22) and the blue pigment 3 (23) induce oxidative doublestrand (ds) copper-mediated DNA cleavage. Fu¨rstner and co-workers suggested that the biological activity of synthetic Prod derivatives may be linked to their ability to complex Cu(II) (24), since roseophilin 5 (B-pyrrole ring replaced by a weaker Cu-ligating furan) did not facilitate DNA cleavage in the presence of copper (24). Similarly, we have shown that replacement of the A-pyrrole ring with a poorly coordinating thiophene suppressed nuclease activity and the ability to inhibit cell proliferation of leukemia (HL-60) cells (22). It is not clear, however, if these results can be rationalized solely based on metal chelation considerations. Abrupt changes in the tripyrrolic structure not only would influence the metalcoordinating ability of the drug but also would affect its electron-donating potential. To provide further insight into these issues, we have carried out an electrochemical study to provide a basis for the nuclease activity of the Prods. In the present paper, we report on the influence of the A-pyrrole ring on the electrochemical and nuclease activity of a Prod analogue. What has emerged from these studies is the importance of the bipyrrole (BP) moiety in ssDNA nicking, while the intact pyrrolylpyrromethene chromophore is critical for the more lethal dsDNA cleavage event.

Experimental Procedures Materials. Prodigiosin (1) was a gift from the Drug Synthesis & Chemistry Branch, Developmental Therapeutics Program, Division of Cancer Treatment and Diagnosis, National Cancer Institute (NCI). It was received as the free base, and confirmation of structure was obtained by 1H NMR spectroscopy and electrospray mass spectrometry (ES+): [M + H]+ ) 324.2. Stock solutions of 1 were prepared in MeCN, and concentrations were determined by UV-vis in 95% EtOH-HCl (535 ) 112 000 M-1 cm-1). The BP aldehyde 4 (Figure 1) was derived by base hydrolysis of tambjamine E (2), as described previously (10). It was obtained as a yellow residue, ES+: [M + H]+ ) 191.1. Supercoiled plasmid (Form I) DNA was a gift from Dr. Fred W. Perrino, Department of Biochemistry, Wake Forest University School of Medicine. The plasmid consisted of 2663 base pairs (bps) and was a derivative of pOXO4 containing the dnaQ gene (25). Starting materials and solvents, unless otherwise specified, were purchased from Aldrich (Milwaukee, WI), Sigma (St. Louis, MO), Fisher Scientific (Itasca, IL), Frontier Scientific (Logan, UT), or Alfa Aesar (Ward Hill, MA), and were used without further purification. Column chromatography was performed using ICN 60, 230-400 mesh silica gel, or 60-325 mesh alumina (Fisher Scientific, Brockman activity II-III). TLC was carried out on Analtech 250 µM layer, UV254 silica gel plates with glass backing. Distilled, deionized water from a Milli-Q

Scheme 1. Synthesis of Prod Analogues

system was used for all aqueous solutions and manipulations. Other solvents were purified and dried according to standard procedures. Agarose gel electrophoresis was carried out in 40 mM Tris-acetate buffer (pH 8.0), containing 5 mM EDTA. Agarose gel loading buffer was 40 mM Tris-acetate (pH 8.0), 5 mM EDTA, 40% glycerol, 0.3% bromophenol blue. Methods. Elemental analyses were carried out by Atlantic Microlab Inc. (Atlanta, GA). High-resolution mass spectra (HRMS) were carried out by either Mass Consortium (San Diego, CA) or the Duke University Mass Spectrometry Facility (Durham, NC). In-house low-resolution ES+ spectra were acquired using a Micromass Quattro II instrument. NMR spectra were recorded on a Bruker AVANCE 300DMX (300 MHz) spectrometer in CDCl3, and peaks were referenced to the residual CHCl3 peak, unless otherwise stated. Chemical shifts are given in ppm relative to TMS, and coupling constants (J) are reported in hertz (Hz). Synthesis of Prodigiosins (Scheme 1). The methods of syntheses of the Prod analogues are outlined in Scheme 1. Prod derivatives 7-12 were available in our laboratory and were prepared using the procedure outlined by D’Alessio and coworkers (20, 26). The N-Boc-protected pyrrole 13 was prepared in 92% yield from ethylpyrrole/Boc2O in the presence of 4-(dimethylamino)pyridine (DMAP). The pyrrole 13 was converted into 5-ethyl-(1-tert-butoxycarbonylpyrrol-2-yl)boronic acid using the method outlined by Fu¨rstner (27) and was allowed to react immediately with the triflate 7/Pd(PPh3)4 to afford the Prod analogue 14. 5′-Ethyl-4-methoxy-5-[(5-ethyl-2H-pyrrol-2-ylidene)methyl]-2,2′-bi-1H-pyrrole, R ) 5-Ethyl-1H-pyrrol-2-yl (14). N-BuLi (2 M in pentane, 3.32 mmol) was slowly added to a solution of 2,2,6,6-tetramethylpiperidine (2.82 mmol) in THF (10 mL) at -78 °C under argon. The mixture was allowed to warm to 0 °C and maintained at that temperature for 30 min. After cooling again to -78 °C, a solution of pyrrole 13 (2.56 mmol) in THF (10 mL) was added. The reaction mixture was stirred for 2 h at -78 °C prior to the addition of trimethyl borate (12. 8 mmol). The solution was allowed to react at ambient temperature overnight and was then used without further purification in the following coupling procedure. A solution of triflate 7 (0.11 mmol) in dimethoxyethane (DME, 20 mL), Pd(PPh3)4 (0.01 mmol) and the boronic acid solution (10 mL) were treated with aqueous Na2CO3 (0.85 mmol), and the resulting mixture was stirred at 80 °C for 16 h. A standard

744

Chem. Res. Toxicol., Vol. 15, No. 5, 2002

Melvin et al.

Scheme 2. Bipyrrole Synthesis

extractive workup followed by flash chromatography on alumina using hexane/ethyl acetate (99:1 f 8:2) as eluent afforded 14 (12.1 mg, 38.3%). 1H NMR: δ 0.84 (m, 6H), 2.02 (m, 2H), 2.11 (m, 2H), 3.90 (s, 3H), 5.76 (m, 2H), 6.00 (s, 1H), 6.37 (d, J ) 3.5, 1H), 6.51 (d, J ) 3.5, 1H), 6.74 (s, 1H); 13C NMR: δ 13.7, 21.2, 58.9, 96.1, 107.6, 108.1, 114.1, 115.3, 120.7, 127.5, 128.9, 138.2, 141.7, 145.8, 160.5, 169.6; HRMS (FAB) calcd for C18H22N3O [M + H] + 296.1763, found 296.1767. 5′-Acetyl-4-methoxy-5-[(5-ethyl-2H-pyrrol-2-ylidene)methyl]-2,2′-bi-1H-pyrrole, R ) 5-Acetyl-1H-pyrrol-2-yl (15). To a CH2Cl2 solution of 9 (0.04 mmol) were added acetic anhydride (0.074 mmol) and AlCl3 (0.148 mmol) and allowed to react at ambient temperature for 3 h. Following standard workup, the crude material was purified over an alumina column (99:1 hexane/ethyl acetate). The collected fractions were concentrated to yield 15 (6.2 mg, 55%). 1H NMR: δ 1.26 (t, J ) 7.6, 3H), 2.54 (s, 3H), 2.90 (q, J ) 7.5, 2H), 3.92 (s, 3H), 6.07 (s, 1H), 6.20 (d, J ) 3.0, 1H), 6.76 (s, 1H), 6.85 (m, 2H), 7.03 (s, 1H); 13C NMR: δ 13.5, 22.4, 27.7, 59.5, 94.8, 114.0, 116.5, 117.2, 119.4, 121.4, 126.1, 127.4, 132.4, 137.9, 147.1, 158.0, 166.5, 188.8; HRMS (MALDI) calcd for C18H20N3O2 [M + H]+ 310.1556, found 310.1550. Bipyrrole (BP) Synthesis (Scheme 2). The 2,2′-bipyrrole methyl ester 17 (Scheme 2) was prepared using a procedure outlined by Rapoport (28). Condensation between pyrrole and methyl pyroglutamate in the presence of POCl3 generated the pyrrolinylpyrrole methyl ester 16, which, upon catalytic dehydrogenation [Pd(C)/xylene], gave the bipyrrole 17; 190.2 mg, 47%. 1H NMR (acetone-d6): δ 3.76 (s, 3H), 6.14 (m, 1H), 6.42 (m, 1H), 6.62 (m, 1H), 6.83 (m, 1H), 10.50 (bs, 1H), 10.80 (bs, 1H). 5′-Carboxaldehyde-1H,1′H-[2,2′]bipyrrolyl-5-carboxylic Acid Methyl Ester (18). The methyl ester 18 was obtained as a yellow powder (166.4 mg, 11.4%) upon Vilsmeier-Haack formylation of 17. 1H NMR: δ 4.05 (s, 3H), 6.86-6.90 (m, 2H), 7.10-7.12 (m, 1H), 7.23-7.25 (m, 1H), 9.71 (s, 1H); 13C NMR: δ 51.8, 109.8, 110.3, 116.7, 124.1, 128.5, 132.9, 134.5, 161.6, 178.8; HRMS (EI) m/z calcd for C11H10N2O3 218.0691, found 218.0695. Electrochemistry. Cyclic voltammetry (CV) was carried out in MeCN using a previously described (29) three-electrode minicell (2-3 mL) consisting of a glassy carbon working electrode, a Pt spiral counter-electrode, and a silver wire pseudoreference electrode. Following each experiment, potentials were calibrated against the SCE (EoSCE ) 0.241 vs NHE). Samples were close to 3 mM in MeCN with tetrabutylammonium perchlorate (TBAP, 0.4-0.5 M) as the supporting electrolyte. Data were collected using a Pine AFCBP1 computer-

Figure 2. Cyclic voltammetry of Prod (1, 6.8 mM) in MeCN (0.47 M TBAP) using a glassy carbon (diameter ) 1.5 mm) working electrode and a SCE reference, ν ) 0.100 V s-1. (A) Single CV scan. (B) Multiple CV scans. controlled bi-potentiostat/waveform generator (Pine Instrument Co., Grove City, PA) and PineChem 2.7 graphical interface software. DNA Cleavage Reactions. Reaction mixtures (20 µL total volume) contained 400 ng of Form I DNA, 10 mM MOPS buffer (pH 7.4), 25 mM NaCl, and 50 µM Prod + 1 equiv of Cu(OAc)2 (50 µM). Reaction mixtures were incubated at 37 °C for 30 min, and then quenched by the addition of 4 µL of loading buffer. Samples were loaded onto a 1% agarose gel containing ethidium bromide (1 µg/mL). The gel was run at 110 V for 2 h and visualized by UV illumination. Photographs of the gels were digitized with a Hewlett-Packard ScanJet 3300C and quantified using ScionImage v. 4.0 software. The amount of Form I DNA was multiplied by a factor of 1.22 to account for reduced ethidium intercalation. The relative copper-promoted nuclease efficiency (Rel. Nuc. Eff.) of the Prod analogues was calculated using the equation: Rel. Nuc. Eff. ) fII(analogue)/fII(Prod), where fII is the fraction of Form II present in the reactions.

Results Electrochemical Oxidation of Prod (1). Figure 2A shows a single CV scan of the free base form of Prod (1). The natural product exhibited three oxidation peaks (Ep/21 ) 0.44 V, Ep/22 ) 0.89 V, Ep/23 ) 1.54 V vs SCE) with peak 2 exhibiting a shoulder (Ep ∼ 1.06 V). The three peak currents were comparable (ip1 ) 32 µA, ip2 ∼ ip3 = 25 µA at ν ) 0.100 V s-1), and ip1 increased linearly with the square root of the scan rate (∂ip1/∂ν1/2 ) 45 µA s1/2 V-1/2), indicating that the process is diffusion-controlled. Repetitive scans revealed a sharp decrease in peak 1 and increases (broad features) at Ep ) 0.87 and 1.14 V (Figure 2B). The reduction peak (peak 4, Figure 2A) was observed on the reverse scan only when the anodic potential was scanned beyond peak 3. The large difference between peaks 3 and 4 and the absence of reduction peaks for peaks 1 and 2 showed that, under these experimental conditions, the oxidative processes are irreversible. The oxidative process at Ep/2 ) 0.89 V (i.e., peak 2) also resulted in broad secondary peaks at EpC ) 0.65 V, 0.10 V, and EpA ∼ -0.80 V (data not shown).

Prodigiosins

Chem. Res. Toxicol., Vol. 15, No. 5, 2002 745

Figure 3. Single CV scan (ν ) 0.100 V s-1) of Prod analogue 12 (3.3 mM) in MeCN (0.25 M TBAP) using a glassy carbon (diameter ) 1.5 mm) working electrode and a SCE reference. Table 1. Anodic Oxidation Potentials for Prod Analogues in MeCN compound

Ep/21 (V)a,b

Ep/22 (V)a,b

∆Ep/21 (mV)

∆Ep/22 (mV)

Prod (1) ProdH+ 8 9 10 11 12 14 15 17 18

0.44 0.62 0.72 0.51 0.67 0.74 0.75 0.49 0.69 0.62 1.02 (sh)

0.89 (sh) 1.04 0.87 1.12 1.29 1.30 1.00 1.15 0.93 -

180c 160d 230d 240d -2d 180d 400e

150c 250d 420d 430d 130d 280d -

a In volts vs SCE. b ν ) 0.100 V s-1. c Relative to values for Prod (1). d Relative to values for 9. e Relative to Ep/21 for 17. sh, shoulder.

Acidification (acetic acid) to generate the protonated species (ProdH+) affected an anodic shift of Ep/21 (0.62 V) and a slight decrease in ip1 (data not shown). These results were consistent with generation of the conjugate acid, which, as a positively charged species, is oxidized at a higher potential than the corresponding free base (30, 31). Structure-Activity Relationships. In general, the synthetic Prod analogues (9-12, 14, and 15; Scheme 1) exhibited two distinct irreversible (at scan rates up to 1 V s-1) oxidation peaks, as shown in Figure 3 for the phenyl derivative 12 (Ep/21 ) 0.75 V, Ep/22 ) 1.30 V). CV performed on the pyrromethene 8 showed a single oxidation peak with Ep/2 ) 0.72 V vs SCE. The BP derivatives (17 and 18, Scheme 2) showed one or two oxidation peaks in MeCN, depending on the nature of ring substitution. The disubstituted BP 18 showed a single oxidation peak (Ep/2 ) 1.02 V, with a small shoulder) while the BP ester 17 exhibited two distinct irreversible processes (Ep/2 ) 0.71, 1.09 V; data not shown). Attempts were made to determine the number of electrons (n) involved in the oxidations of 8, 12, 15, and 17 by calculations involving microelectrode steady-state currents and Cottrell slopes, as described previously (29). These calculations yielded diffusion coefficients (D) that were unreasonably low for this class of compounds in nonaqueous solution. However, if the diffusion coefficient of bipyrrole in MeCN [D ) 2.7 × 10-5 cm2 s-1 (32)] was assumed, then the Cottrell plot yielded a value of n ∼ 1 for the first oxidation process. The Ep/2 values (first and second oxidation process) of the various analogues are given in Table 1. The ∆Ep/2 values refer to differences in the first (Ep/21) and second

Figure 4. Relative copper nuclease efficiency of Prod analogues (50 µM) in the presence of 50 µM Cu(OAc)2. Reaction mixtures (20 µL total volume) contained 400 ng of Form I DNA in 10 mM MOPS buffer, pH 7.4, 20 mM NaCl and were incubated at 37 °C for 30 min.

Figure 5. Relaxation of Form I DNA by BP and pyrromethenes (50 µM) in the presence of 1 equiv of Cu(OAc)2 (50 µM). Reaction mixtures (20 µL total volume) contained 400 ng of Form I DNA in 10 mM MOPS buffer, pH 7.4, and were incubated at 37 °C for 30 min. Lane 1, DNA alone; lane 2, +Cu; lane 3, +4; lane 4, 4 + Cu; lane 5, 17 + Cu; lane 6, 18 + Cu; lane 7, 8 + Cu; lane 8, 16 + Cu.

(Ep/22) oxidation processes. Here, the free base form of Prod (1) is compared to the protonated species (ProdH+), while the values for the synthetic Prod analogues 1012, 14, and 15 are with respect to the tripyrrolic prodigiosin 9; BP 18 is compared to BP 17. Copper-Mediated DNA Cleavage. The nuclease efficiency of the synthetic analogues was compared to the natural Prod (1) using agarose gel electrophoresis and supercoiled plasmid (Form I) DNA. As presented previously (22), Prod (1) facilitates oxidative dsDNA cleavage in the presence of equimolar Cu(II) in the range of 1050 µM. For the present experiments, where the focus was placed on the relative nicking activity of the Prod analogues, reaction mixtures contained 50 µM Prod analogue-Cu(OAc)2 and were incubated for 30 min at 37 °C. Under these conditions, the most extensive cleavage [∼90% conversion of Form I DNA into nicked circular (Form II) DNA] was found for the natural derivative 1 (Figure 4). The tripyrrolic synthetic analogues 9 and 14 showed almost equivalent activity (∼80% conversion) and were only slightly less efficient than 1 at promoting strand scission. Virtually no nuclease activity was found for 10-12 and 15. These results demonstrated the strict requirement for the A-pyrrole ring, as derivatives 1012 failed to promote strand scission. That 14 showed nuclease activity, while 15 did not, also demonstrated that an alkyl substituent at the 5′ position of the A-ring (see Figure 1 for numbering of Prod) is tolerated, while an electron-withdrawing substituent (acetyl) is not. Further insight was obtained by examining the Cunuclease activity of the pyrromethene 8, and BP fragments (4, Figure 1; and 17, 18, Scheme 2). Figure 5 shows a gel highlighting relaxation of Form I DNA by these analogues. Here, BP 4 (lane 4), 17 (lane 5), and 18 (lane 6) were found to facilitate the conversion of Form I DNA

746

Chem. Res. Toxicol., Vol. 15, No. 5, 2002

Melvin et al. Scheme 3. Oxidation of 9-Methyl Dipyrrinone

Figure 6. Inhibition studies on cleavage of Form I DNA by 17/Cu(OAc)2 (50 µM). Reactions contained 400 ng of Form I DNA and were carried out for 30 min as described in the caption of Figure 5. Lane 1, DNA alone; lane 2, +Cu; lane 3, Cu + 17; lanes 4-9, Cu + 17: +100 mM NaN3, +1 M tert-butyl alcohol, +1 M DMSO, +100 mM EDTA, +1000 units/mL catalase, +1000 units/mL SOD, respectively.

into nicked-circular (Form II), with the extent of cleavage being 4 > 17 > 18. The pyrromethene 8 (lane 7) and the pyrrole derivative 16 (lane 8) were inactive. A single pyrrole ring (2-ethylpyrrole) was also incapable of inducing copper-promoted DNA cleavage at 50 µM (data not shown). Inhibition studies on strand scission by BP 17 (Figure 6) showed that the cleavage was oxidative, as the enzyme catalase, which lowers solution concentrations of hydrogen peroxide, almost completely inhibited strand scission (lane 8). Superoxide dismutase (SOD) had little effect (lane 9), indicating that O2•- is not required for the reduction of Cu(II), nor is O2•- responsible for strand scission (22). The hydroxyl radical scavengers tert-butyl alcohol (lane 5) and DMSO (lane 6) also failed to inhibit cleavage, which argued against participation of the freely diffusible hydroxyl radical. As noted for DNA cleavage by the natural products, 1-3 (21-23), the singlet oxygen scavenger NaN3 inhibited formation of Form II DNA (lane 4). Generally, the results for BP 17 were similar to that noted for tambjamine E (2), which also affects oxidative ssDNA cleavage with no Form III (linear DNA) DNA forming prior to complete cleavage of Form I DNA (21). These results are contrasted by copper-promoted DNA cleavage by Prod (1), which facilitates the dsDNA cleavage event (22).

Discussion Although the Prods exhibit promising cytotoxicity (1215), their mode of action has not been firmly established. However, we have demonstrated that they facilitate dsDNA cleavage in the presence of equimolar Cu(II) (22, 23). Oxidative dsDNA cleavage creates damage, that, for obvious reasons, might be considered much more difficult for a cell to repair than ssDNA cleavage. Thus, dsDNA cleavage has been proposed to be of importance therapeutically (33), and is believed to be responsible (at least in part) for the cytotoxicity of the clinically used bleomycins (34, 35). Thus, it is conceivable that dsDNA cleavage represents the mode of Prod cytotoxicity, and since this activity is triggered by oxidation of the electronrich polypyrrole molecule, we prepared several Prod analogues and measured their electrochemical properties and nuclease activity in the presence of Cu(II). It was anticipated that these efforts would help shed light on the structural properties that govern the nuclease activity of the prodigiosins. The redox properties of the various Prod analogues were measured in MeCN using cyclic voltammetry (CV). For pyrrolic systems it has been established that the anodic oxidation of simple pyrroles (36), bipyrroles (32, 37), pyrromethenes (30), and linear oligopyrrole bile

pigments (31) is irreversible, with the resulting π-radical cations either polymerizing or reacting with added or adventitious nucleophiles to give substitution products. The oxidation of pyrromethenes is also pH-dependent, with the pyrromethene free base exhibiting oxidation potentials lower than the corresponding halogen salts (30). Pyrrolic species such as dipyrrinones containing alkyl substituents at C9 undergo a 2e/1H+ transformation to an azafulvenic structure, as shown in Scheme 3. In dipolar aprotic solvents, the CV of dipyrrinones shows two 1e oxidative processes, with the first appearing between 0.57 and 0.8 V (SCE) and the second between 0.98 and 1.30 V (31). The electrochemical oxidation of the free base of Prod (1) in MeCN displayed three successive oxidative processes (Figure 2A) with the first two occurring at relatively low potentials (Ep/21 ) 0.44, Ep/22 ) 0.89 V vs SCE; Table 1). Protonation of the Prod free base to generate ProdH+ caused an anodic shift in the first two oxidative processes (∆Ep/2 ∼ 150 mV, Table 1), as anticipated (30). When the electrode potential was scanned repetitively beyond Ep/22, a dark film appeared on the surface of the electrode, and broad redox waves became apparent at E < Ep/22 (i.e., EpC ) 0.65, 0.10 V; data not shown). This is consistent with the deposition of a layer of redox-active polymer or oligomer (32) that resulted from the coupling of Prod oxidation products. It is of interest to note that film formation (polymerization) was not observed upon repetitive scans just beyond the first oxidative process, Ep/21. The synthetic Prod analogues also displayed two successive oxidative processes (e.g., Figure 3) with Ep/21 appearing between 0.49 and 0.75 V (SCE) and Ep/22 between 0.97 and 1.30 V (Table 1). These values are remarkably similar to the ones published for the dipyrrinones (31), which undergo two successive 1e oxidations to generate the azafulvenic species shown in Scheme 3. It is informative that the first oxidative process was determined to be monoelectronic for derivatives 8, 12, 15, and 17 [using the diffusion coefficient of bipyrrole in MeCN (32)], suggesting that Ep/21 represents π-radical cation formation. Here, replacement of the A-pyrrole ring in 9 with alternative arenes (i.e., 10, 11, and 12) raised Ep/21 by ∼ 200 mV (Table 1). The same effect was achieved through attachment of the electron-withdrawing acetyl group at the 5′-position to generate 15. However, attachment of the electron-withdrawing aldehyde functionality to BP 17 to generate 18 had a much greater impact and raised the anodic potential by 400 mV (Table 1). A 500 mV increase for the same transformation in a single pyrrole ring has been reported (36). Thus, the results for the synthetic Prod analogues suggested that the first oxidative process occurred at the ethylpyrrole C-ring. Subsequent oxidation of this species at Ep/22 could generate an azafulvenic structure (i.e., Scheme 3), which would be expected to polymerize or react with adventitious water in the MeCN solvent. Note that generation of the azafulvenic structure and subsequent polymerization lead to release of protons. Acidification of the electrode surface would yield ProdH+, which undergoes oxidation at a higher potential than the free base form. This factor

Prodigiosins

may account for the anodic shift in Ep/21 upon repetitive scans above Ep/22 (i.e., Figure 2B). The information from the CV experiments provided a basis for the relative nuclease properties of the Prods. The analogues with the lowest anodic oxidation potentials in MeCN (1, 9, and 14) were the only ones capable of promoting DNA cleavage in the presence of Cu(II) (Figure 4). The importance of the relatively low oxidation potential was also supported by the results obtained with the BP analogues (Figure 5). Here BP 4 and 17 were found to be much more efficient than BP 18 at promoting ssDNA cleavage in the presence of equimolar Cu(II). The anodic oxidation potential of 17 is 0.62 V in MeCN (Table 1), and attachment of the aldehyde functionality to yield 18 raised the potential to 1.01 V and generated a species with diminished nuclease activity. These results were similar to the finding that conversion of Prod 9 into 15 inhibited strand scission. While the anodic oxidation potential is one parameter deemed important in the strand scission chemistry, the number of electrons (n) involved in the oxidation also plays a key role. Here the BP moiety is known to undergo a 1e oxidation (32, 37); a chemical step is then required to regenerate a neutral species that can donate additional electrons (37). We have demonstrated that the neutral BP 4 lacks the ability to bind DNA effectively (10), nor does it bind Cu(II) at physiological pH. This suggests that BP 4 and 17 provide the reducing equivalents to reductively activate “free” or DNA-bound Cu, which is known to generate ss breaks and oxidative base damage (3840). However, the Prods can undergo two successive 1e oxidation processes. Thus, one molecule of Prod could potentially provide the reducing equivalents to activate 2 equiv of Cu(II). The anticipated copper coordination properties of the Prods make it possible for the generation of a redox-active metalloprodigiosin species that directs site-specific dsDNA cleavage, as noted for the bleomycins (34, 35). In terms of biological relevance, it is especially noteworthy that the relative nuclease activity of the Prods (Figure 4) exhibits striking correlations with inhibition of colony formation by HL-60 cancer cells (22) and the relationships presented by D’Alessio on the cytotoxicity of undecylprodigiosin 6 (Figure 1) against melanoma (20). Thus, replacement of the A-pyrrole ring of a Prod analogue with an alternative arene (phenyl, thiophen2-yl, furan-2-yl) inhibits Cu-promoted strand cleavage (24) and cytotoxicity (20, 22). The same effect is also achieved by attaching an electron-withdrawing substituent (acetyl group) to the 5′-position of the A-pyrrole ring. In conclusion, the A-pyrrole ring of the Prods influences the redox properties of the pyrromethene. The BP moiety is capable of promoting ssDNA cleavage, while the intact pyrrolylpyrromethene chromophore of the Prods is required for the more lethal copper-promoted dsDNA cleavage event. The BP moiety lacks the ability to coordinate copper effectively at physiological pH, and undergoes 1e oxidations to generate π-radical cations. In contrast, the Prods can undergo successive 1e processes and coordinate Cu(II), properties deemed important for dsDNA cleavage. Since Prod possesses alkyl substituents on the C-ring, it may also form an azafulvenic structure upon oxidation (31). Such a species would be expected to alkylate DNA, based on the known DNA alkylation properties of 3-methylindole upon oxidative transformation into the corresponding azafulvenic 3-methylene-

Chem. Res. Toxicol., Vol. 15, No. 5, 2002 747

indolenine species (41). That roseophilin 5 (Figure 1) is incapable of promoting DNA cleavage (24) can be ascribed to the furan ring system, which lacks the metalcoordinating properties of the pyrromethene and, based on our electrochemical studies, can be predicted to raise the oxidation potential by ∼200 mV relative to the analogous pyrrolic species. It is also noteworthy that Prod (1) and roseophilin 5 show different cytotoxicity profiles against the NCI 60-cell line panel. Use of the NCI COMPARE algorithm also suggests that these agents act by different modes (42). Extension of the present work to include the copper coordination properties of the Prods and to identify their products of oxidation and DNA alkylation properties is currently underway.

Acknowledgment. R.A.M. acknowledges support from the Petroleum Research Fund (ACS-PRF 37177AC4,3). The electrochemical portion of this work (R.E.N.) was supported in part by the NIH (R15-GM59628). NMR spectra were recorded on instruments purchased with the partial support of the NSF (CHE-9708077) and NCBC (9703-IDG-1007). We are very grateful to the Drug Synthesis & Chemistry Branch, Developmental Therapeutics Program, Division of Cancer Treatment and Diagnosis of the NCI, for the sample of 1, and to Dr. Fred Perrino (Department of Biochemistry) for the plasmid DNA.

References (1) Gerber, N. N. (1974) Prodigiosin-like pigments. Crit. Rev. Microbiol. 3, 469-485. (2) Bennett, J. W., and Bentley, R. (2000) Seeing red: The story of prodigiosin. Adv. Appl. Microbiol. 47, 1-32. (3) Manderville, R. A. (2001) Synthesis, proton-affinity and anticancer properties of the prodigiosin-group natural products. Curr. Med. Chem.: Anti-Cancer Agents 1, 195-218. (4) Rapoport, H., and Holden, K. G. (1962) The synthesis of prodigiosin. J. Am. Chem. Soc. 84, 635-642. (5) Boger, D. L., and Patel, M. (1988) Total synthesis of prodigiosin, prodigiosene, and desmethoxy-prodigiosin: Diels-Alder reactions of heterocyclic azadienes and development of an effective palladium(II)-promoted 2,2′-bipyrrole coupling procedure. J. Org. Chem. 53, 1405-1415. (6) Carte´, B., and Faulkner, D. J. (1983) Defensive metabolites from three nembrothid nudibranchs. J. Org. Chem. 48, 2314-2318. (7) Lindquist, N., and Fenical, W. (1991) New tambjamine class alkaloids from the marine ascidian Atapozoa sp. and its nudibranch predators. Origin of the tambjamines in Atapozoa. Experientia 47, 504-506. (8) Blackman, A. J., and Li, C. (1994) New tambjamine alkaloids from the marine Bryozoan Bugula dentata. Aust. J. Chem. 47, 16251629. (9) Wasserman, H. H., Friedland, D. J., and Morrison, D. A. (1968) A novel dipyrrolydipyrromethene prodigiosin analogue from Serratia marcescens. Tetrahedron Lett., 641-642. (10) Melvin, M. S., Ferguson, D. C., Lindquist, N., and Manderville, R. A. (1999) DNA binding by 4-methoxypyrrolic natural products. Preference for intercalation at AT sites by tambjamine E and prodigiosin. J. Org. Chem. 64, 6861-6869. (11) Hayakawa, Y., Kawakami, K., Seto, H., and Furihata, K. (1992) Structure of a new antibiotic, roseophilin. Tetrahedron Lett. 33, 2701-2704. (12) Yamamoto, C., Takemoto, H., Kuno, K., Yamamoto, D., Tsubura, A., Kamata, K., Hirata, H., Yamamoto, A., Kano, H., Seki, T., and Inoue, K. (1999) Cycloprodigiosin hydrochloride, a new H(+)/ Cl(-) symporter, induces apoptosis in human and rat liver hepatocellular cancer cell lines in vitro and inhibits the growth of hepatocellular carcinoma xenografts in nude mice. Hepatology 30, 894-902. (13) Yamamoto, D., Kiyozuka, Y., Uemura, Y., Yamamoto, C., Takemoto, H., Hirata, H., Tanaka, K., Hioki, K., and Tsubura, A. (2000) Cycloprodigiosin hydrochloride, a H+/Cl- symporter, induces apoptosis in human breast cancer cell lines. J. Cancer Res. Clin. Oncol. 126, 191-197.

748

Chem. Res. Toxicol., Vol. 15, No. 5, 2002

(14) Diaz-Riuz, C., Montaner, B., and Perez-Tomas, R. (2001) Prodigiosin induces cell death and morphological changes indicative of apoptosis in gastric cell line HGT-1. Histol. Histopathol. 16, 415-421. (15) Montaner, B., and Perez-Tomas, R. (2001) Prodigiosin-induced apoptosis in human colon cancer cells. Life Sci. 68, 2025-2026. (16) Magae, J., Yamashita, M., and Nagai, K. (1993) Suppression of alloantigen presentation by prodigiosin, a T cell-specific immunosuppressant. Ann. N.Y. Acad. Sci. 685, 339-340. (17) Lee, M. H., Yamashita, M., Tsuji, R. F., Kataoka, T., Magae, J., and Nagai, K. (1998) Suppression of T cell stimulating function of allogeneic antigen presenting cells by prodigiosin 25-C. J. Antibiot. 51, 92-94. (18) Songia, S., Mortellaro, A., Taverna, S., Fornasiero, C., Scheiber, E. A., Erba, E., Colotta, F., Mantovani, A., Isetta, A. M., and Golay, J. (1997) Characterization of the new immunosuppressive drug undecylprodigiosin in human lymphocytes: retinoblastoma protein, cyclin-dependent kinase-2, and cyclin-dependent kinase-4 as molecular targets. J. Immunol. 158, 3987-3995. (19) Stepkowski, S. M., Erwin-Cohen, R. A., Behbod, F., Wang, M. E., Qu, X., Tejpal, N., Nagy, Z. S., Kahan, B. D., and Kirken, R. A. (2002) Selective inhibitor of Janus tyrosine kinase 3, PNU156804, prolongs allograft survival and acts synergistically with cyclosporine but additively with rapamycin. Blood 99, 680-689. (20) D’Alessio, R., Bargiotti, A., Carlini, O., Colotta, F., Ferrari, M., Gnocchi, P., Isetta, A., Mongelli, N., Motta, P., Rossi, A., Rossi, M., Tibolla, M., and Vanotti, E. (2000) Synthesis and immunosuppressive activity of novel prodigiosin derivatives. J. Med. Chem. 43, 2557-2565. (21) Borah, S., Melvin, M. S., Lindquist, N., and Manderville, R. A. (1998) Copper-mediated nuclease activity of a tambjamine alkaloid. J. Am. Chem. Soc. 120, 4557-4562. (22) Melvin, M. S., Tomlinson, J. T., Saluta, G. R., Kucera, G. L., Lindquist, N., and Manderville, R. A. (2000) Double-strand DNA cleavage by copper‚prodigiosin. J. Am. Chem. Soc. 122, 63336334. (23) Melvin, M. S., Wooton, K. E., Rich, C. C., Saluta, G. R., Kucera, G. L., Lindquist, N., and Manderville, R. A. (2001) Coppernuclease efficiency correlates with cytotoxicity for the 4-methoxypyrrolic natural products. J. Inorg. Biochem. 87, 129-135. (24) Fu¨rstner, A., and Grabowski, E. J. (2001) Studies on DNA cleavage by cytotoxic pyrrole alkaloids reveal the distinctly different behavior of roseophilin and prodigiosin derivatives. ChemBioChem 9, 706-709. (25) Ardus, J. A., Gillman, I. G., and Manderville, R. A. (1998) On the role of copper and iron in DNA cleavage by ochratoxin A. Structure-activity relationships in metal binding and coppermediated DNA cleavage. Can. J. Chem. 76, 907-918. (26) D’Alessio, R., and Rossi, A. (1996) Short synthesis of undecylprodigiosine. A new route to 2,2′-bipyrrolyl-pyrromethene systems. Synlett., 513-514. (27) Fu¨rstner, A., Grabowski, J., and Lehmann, C. W. (1999) Total synthesis and structural refinement of the cyclic tripyrrole pigment nonylprodigiosin. J. Org. Chem. 64, 8275-8280. (28) Rapoport, H., and Bordner, J. (1964) Synthesis of substituted 2,2′bipyrroles. J. Org. Chem. 29, 2727-2731.

Melvin et al. (29) Calcutt, M. W., Gillman, I. G., Noftle, R. E., and Manderville, R. A. (2001) Electrochemical oxidation of ochratoxin A: correlation with 4-chlorophenol. Chem. Res. Toxicol. 14, 1266-1272. (30) Tabba, H. D., Cavaleiro, J. A. S., Jeyakumar, D., Graca, M., Neves, P. M. S., and Smith, K. M. (1989) Electrochemical study of the nonaqueous oxidation of dipyrrolic compounds. J. Org. Chem. 54, 1943-1948. (31) Ribo, J. M., Farrera, J.-A., Claret, J., and Grubmayer, K. (1992) Reactivity of pyrrole pigments. Part 14. The electrochemical oxidation and reduction of bile pigments. Bioelectrochem. Bioenerg. 29, 1-17. (32) Raymond, D. E., and Harrison, D. J. (1993) Observation of soluble pyrrole oligomers and the role of protons in the formation of polypyrrole and polybipyrrole. J. Electroanal. Chem. 355, 115131. (33) Povirk, L. F. (1983) in Molecular aspects of anti-cancer drug design (Neidle, S., and Waring, M., Eds.) pp 157-181, MacMillan, London. (34) Keck, M. V., Manderville, R. A., and Hecht, S. M. (2001) Chemical and structural characterization of the interaction of bleomycin A2 with d(CGCGAATTCGCG)2. Efficient, double-strand DNA cleavage accessible without structural reorganization. J. Am. Chem. Soc. 123, 8690-8700. (35) Hoehn, S. T., Junker, H.-D., Bunt, R. C., Turner, C. J., and Stubbe, J. (2001) Solution structure of Co(III)-bleomycin-OOH bound to a phosphoglycolate lesion containing oligonucleotide: implications for bleomycin-induced double-strand DNA cleavage. Biochemistry 40, 5894-5905. (36) Tabba, H. D., and Smith, K. M. (1984) Anodic oxidation potentials of substituted pyrroles: derivation and analysis of substituent partial potentials. J. Org. Chem. 49, 1870-1875. (37) Guyard, L., Hapiot, P., and Neta, P. (1997) Redox chemistry of bipyrroles: further insights into the oxidation polymerization mechanism of pyrrole and oligopyrroles. J. Phys. Chem. B 101, 5698-5706. (38) Li, Y., and Trush, M. A. (1993) DNA damage resulting from the oxidation of hydroquinone by copper: role for a Cu(II)/Cu(I) redox cycle and reactive oxygen generation. Carcinogenesis 14, 13031311. (39) Rodriguez, H., Holmquist, G. P., D’Agostino, R., Jr., Keller, J., and Akman, S. A. (1997) Metal ion-dependent hydrogen peroxideinduced DNA damage is more sequence specific than metal specific. Cancer Res. 57, 2394-2403. (40) Liang, Q., and Dedon, P. C. (2001) Cu(II)/H2O2-induced DNA damage is enhanced by packaging of DNA as a nucleosome. Chem. Res. Toxicol. 14, 416-422. (41) Regal, K. A., Laws, G. M., Yuan, C., Yost, G. S., and Skiles, G. L. (2001) Detection and characterization of DNA adducts of 3methylindole. Chem. Res. Toxicol. 14, 1014-1024. (42) Deng, J.-Z., Newman, D. J., and Hecht, S. M. (2000) Use of COMPARE analysis to discover functional analogues of bleomycin. J. Nat. Prod. 63, 1269-1272.

TX025508P