Metal-ligand charge-transfer-promoted photoelectronic Bergman cyclization of copper metalloenediynes: Photochemical DNA cleavage via C-4′ H-atom abstraction

Article abstract of DOI:10.1021/ja020939f

Metal-to-ligand charge-transfer (MLCT) photolyses (λ ≥ 395 nm) of copper complexes of cis-1,8-bis(pyridin-3-oxy)oct-4-ene-2,6-diyne (bpod, 1), [Cu(bpod)₂]PF₆ (2), and [Cu(bpod)₂](NO₃)₂ (3) yield Bergman cyclization of the bound ligands. In contrast, the uncomplexed ligand 1 and Zn(bpod)₂(CH₃COO)₂ compound (4) are photochemically inert under the same conditions. In the case of 4, sensitized photochemical generation of the lowest energy ³π-π* state, which is localized on the enediyne unit, leads to production of the trans-bpod ligand bound to the Zn(II) cation by photoisomerization. Electrochemical studies show that 1, both the uncomplexed and complexed, exhibits two irreversible waves between Ep values of -1.75 and -1.93 V (vs SCE), corresponding to reductions of the alkyne units. Irreversible, ligand-based one-electron oxidation waves are also observed at +1.94 and +2.15 V (vs SCE) for 1 and 3. Copper-centered oxidation of 2 and reduction of 3 occur at E_1/2 = +0.15 and +0.38 V, respectively. Combined with the observed Cu(I)-to-pyridine(π*) MLCT and pyridine(π*)-to-Cu(II) ligand-to-metal charge transfer (LMCT) absorption centered near ～315 nm, the results suggest a mechanism for photo-Bergman cyclization that is derived from energy transfer to the enediyne unit upon charge-transfer excitation. The intermediates produced upon photolysis degrade both pUC19 bacterial plasmid DNA, as well as a 25-base-pair, double-stranded oligonucleotide. Detailed analyses of the cleavage reactions reveal 5′-phosphate and 3′-phosphoglycolate termini that are derived from H-atom abstraction from the 4′-position of the deoxyribose ring rather than redox-induced base oxidation.

Full text of DOI:10.1021/ja020939f

Published on Web 05/01/2003
Metal-Ligand Charge-Transfer-Promoted Photoelectronic
Bergman Cyclization of Copper Metalloenediynes:
Photochemical DNA Cleavage via C-4′ H-Atom Abstraction
†
‡
†
†
Pedro J. Benites, Rebecca C. Holmberg, Diwan S. Rawat, Brian J. Kraft,
†
†
,‡
,†
Lee J. Klein, Dennis G. Peters, H. Holden Thorp,* and Jeffrey M. Zaleski*
Contribution from the Department of Chemistry, Indiana UniVersity,
Bloomington, Indiana 47405, and Department of Chemistry, Venable Hall,
UniVersity of North Carolina, Chapel Hill, North Carolina 27599-3290
Received July 9, 2002; E-mail: zaleski@indiana.edu
Abstract: Metal-to-ligand charge-transfer (MLCT) photolyses (λ g 395 nm) of copper complexes of cis-
,8-bis(pyridin-3-oxy)oct-4-ene-2,6-diyne (bpod, 1), [Cu(bpod) ]PF (2), and [Cu(bpod) ](NO (3) yield
Bergman cyclization of the bound ligands. In contrast, the uncomplexed ligand 1 and Zn(bpod) (CH COO)
2
1
2
6
2
3 2
)
2
3
compound (4) are photochemically inert under the same conditions. In the case of 4, sensitized
3
photochemical generation of the lowest energy π-π* state, which is localized on the enediyne unit, leads
to production of the trans-bpod ligand bound to the Zn(II) cation by photoisomerization. Electrochemical
studies show that 1, both the uncomplexed and complexed, exhibits two irreversible waves between E
p
values of -1.75 and -1.93 V (vs SCE), corresponding to reductions of the alkyne units. Irreversible, ligand-
based one-electron oxidation waves are also observed at +1.94 and +2.15 V (vs SCE) for 1 and 3. Copper-
centered oxidation of 2 and reduction of 3 occur at E1/2 ) +0.15 and +0.38 V, respectively. Combined with
the observed Cu(I)-to-pyridine(π*) MLCT and pyridine(π*)-to-Cu(II) ligand-to-metal charge transfer (LMCT)
absorption centered near ∼315 nm, the results suggest a mechanism for photo-Bergman cyclization that
is derived from energy transfer to the enediyne unit upon charge-transfer excitation. The intermediates
produced upon photolysis degrade both pUC19 bacterial plasmid DNA, as well as a 25-base-pair, double-
stranded oligonucleotide. Detailed analyses of the cleavage reactions reveal 5′-phosphate and 3′-
phosphoglycolate termini that are derived from H-atom abstraction from the 4′-position of the deoxyribose
ring rather than redox-induced base oxidation.
1
. Introduction
without the need for external cofactors or reagents. The unusual
capability of these molecular structures to generate a toxic di-
radical intermediate unimolecularly warrants their continued
examination in an effort to develop well-controlled initiation
of Bergman cyclization reactivity.
The ability to control Bergman cyclization of enediyne com-
pounds photochemically with variable-wavelength electronic
excitation leads to potential applications of these unique mole-
cules in photodynamic therapy1-4 and biomedicine
5-9
or as
diagnostic tools for molecular biology.10 The chemical feature
that distinguishes these molecules from their biologically active
The use of electronic excitation to activate thermally stable
enediyne structures photochemically is a direct response to the
problem of controlled initiation. Indeed, it has been known for
counterparts is their unique 1,5-diyn-3-ene unit,1
1-16
which
17,18
some time that both enediyne natural products
and synthetic
rearranges to form a potent 1,4-phenyl diradical intermediate
†
Indiana University.
University of North Carolina.
(
11) Lee, M. D.; Dunne, T. S.; Siegal, M. M.; Chang, C. C.; Morton, G. O.;
‡
Borders, D. B. J. Am. Chem. Soc. 1987, 109, 3464-3466.
(
1) Boyle, R. W.; Dolphin, D. Photochem. Photobiol. 1996, 64, 469-485.
2) Mody, T. M., Sessler, J. L., Eds. Supramolecular Materials and Technolo-
gies; Wiley: Chichester, U.K., 1999; Vol. 4.
3) Ali, H.; van Lier, J. E. Chem. ReV. 1999, 99, 2379-2450.
4) Bonnett, R.; Martinez, G. Tetrahedron 2001, 57, 9513-9547.
5) Abramson, A. L.; Shikowitz, M. J.; Mullooly, V. M.; Steinberg, B. M.;
Amella, C. A.; Rothstein, H. R. Arch. Otolaryngol., Head Neck Surg. 1992,
18, 25-29.
6) Sieber, F.; O’Brien, J. M.; Gaffney, D. K. Blood Cells 1992, 18, 117-
28.
7) Sieber, F.; O’Brien, J. M.; Gaffney, D. K. Semin. Hematol. 1992, 29, 79-
7.
(12) Lee, M. D.; Dunne, T. S.; Chang, C. C.; Ellestad, G. A.; Siegal, M. M.;
Morton, G. O.; McGahren, W. J.; Borders, D. B. J. Am. Chem. Soc. 1987,
109, 3466-3468.
(
(
(
(
(13) Lee, M. D.; Dunne, T. S.; Chang, C. C.; Siegal, M. M.; Morton, G. O.;
Ellestad, G. A.; McGahren, W. J.; Borders, D. B. J. Am. Chem. Soc. 1992,
114, 985-997.
(14) Konishi, M.; Ohkuma, H.; Tsuno, T.; Oki, T.; VanDuyne, G. D.; Clardy,
J. J. Am. Chem. Soc. 1989, 112, 3715-3716.
1
(
(
(
(
(15) Golik, J.; Clardy, J.; Dubay, G.; Groenewold, G.; Kawaguchi, H.; Konishi,
M.; Krishnan, B.; Ohkuma, H.; Saitoh, K.; Doyle, T. W. J. Am. Chem.
Soc. 1987, 109, 3461-3462.
1
8
(16) Golik, J.; Clardy, J.; Dubay, G.; Groenewold, G.; Kawaguchi, H.; Konishi,
M.; Krishnan, B.; Ohkuma, H.; Saitoh, K.; Doyle, T. W. J. Am. Chem.
Soc. 1987, 109, 3462-3464.
8) Matthews, J. L.; Sogandares-Bernal, F.; Judy, M.; Gulliya, K.; Newman,
J.; Chanh, T.; Marengo-Rowe, A. Blood Cells 1992, 18, 75-89.
9) North, J.; Neyndorff, H.; Lewis, J. G. J. Photochem. Photobiol. B 1993,
(17) Uesawa, Y.; Kuwahara, J.; Sugiura, Y. Biochem. Biophys. Res. Commun.
1989, 164, 903-911.
1
7, 99-108.
(
10) Tullius, T. D. In Metal-DNA Chemistry; Tullius, T. D., Ed.; American
(18) Sugiura, Y.; Shiraki, T.; Konishi, M.; Oki, T. Proc. Natl. Acad. Sci. U.S.A.
1990, 87, 3831-3835.
Chemical Society: Washington, DC, 1989; Vol. 402, pp 1-23.
6434
9
J. AM. CHEM. SOC. 2003, 125, 6434-6446
10.1021/ja020939f CCC: $25.00 © 2003 American Chemical Society

Photo-Bergman Cyclization of Cu Metalloenediynes
A R T I C L E S
analogues1
9-26
can be photochemically activated to generate
transfer (MLCT) absorption bands in the optical spectrum. These
transitions are derived from significant spatial overlap between
the metal dσ or dπ molecular orbitals and bound ligand σ or π
molecular orbitals. The presence of such absorption features,
however, does not mandate that the participating metal- or
ligand-centered orbitals be the highest occupied molecular
orbital (HOMO) or the lowest unoccupied molecular orbital
(LUMO) of the complex. These localized states can therefore
serve as pathways to reactive electronic configurations that do
not absorb photons in the ground state (e.g., long-range charge-
separated states) and, therefore, are not populated directly by
fully cyclized Bergman products upon UV excitation. Recently,
2
2
the kinetics and electronic structures of enediyne excited
2
7
states giving rise to the observed product distributions, have
been more clearly elucidated. The emerging theme for photo-
Bergman cyclization of simple organic enediyne ligands is that
photoexcitation leads to an alkyne-localized excited π-π* state
that promotes an in-plane geometric distortion, leading to the
Bergman-cyclized 1,4-phenyl diradical intermediate. In con-
trast, population of the lowest-energy π-π* state, which is
localized on the ene functionality, leads to the potential for an
out-of-plane geometric distortion via rotation about the double
bond and localization of unpaired electron density at the alkyne
termini. This leads to sites for H-atom abstraction and the
subsequent formation of dien-yne photoproducts.
1
2
7
3
4
4-46
optical excitation.
Conceptually, this could lead to new
design strategies that utilize MLCT absorption to initiate
Bergman cyclization with visible-region photons photochemi-
cally, which would relax the requirement for high-energy,
alkyne-localized electronic excitation to drive such reactions.
Beyond the excited-state mechanistic details of photoinitiated
Bergman cyclization, the ability of the potent 1,4-phenyl
diradical intermediate to perform DNA degradation via H-atom
abstraction rather than redox reactions gives these metalloene-
diynes potential advantages over other metal-activated thermal
or photochemical agents. Thermal and photochemical DNA
Binding of transition-metal ions to simple enediyne ligands
influences thermal Bergman cyclization temperatures by induc-
ing a geometry that approaches the transition-state structure,
28-41
thereby lowering the activation barrier to cycloaromatization.
However, the potential role of metal ions in photo-Bergman
cyclization reactivity has not been examined. It is well
established that photochemical activation of both σ- and π-bound
metal complexes can lead to ligand dissociation and/or pho-
toreduction reactions.42 In the case of simple organic enediynes,
ligand-centered redox chemistry does not lead to 1,6-Bergman
cyclization. Rather, 1,5-cyclization and subsequent ion-radical
reactions are observed.43 Thus, it is unclear at first glance how
transition metals, especially those that are redox active, may
participate in driving photoinduced Bergman cyclization.
4
7,48
cleavage
by transition-metal complexes frequently occurs
by one of three general mechanisms: (i) one-electron guanine
4
9,50
50-54
oxidation (e.g., Ni(II) peptide
and macrocycle
complex
2
-
55
56-58
reactions with KHSO5, Cr2O7 /H2O2, CuCl2/H2O2,
ReO (4-OMepy) , Ru(bpy)3
O2 ), (ii) oxygen-atom transfer oxidation (e.g., [Ru(IV)(tpy)-
bpy)O] , [Ru(III)(tpy)(bpy)OH] , ) and sugar degradation
+
59
3+ 60,61
3+
2
4
,
and [Co(NH3)6]
via
1
62
2+ 63
2+ 64
(
6
5
by H-atom abstraction (e.g., Fe(II)-bleomycin, Rh-phi com-
One electronic property that redox-active transition-metal
complexes exhibit is the presence of metal-to-ligand charge-
6
6
+
67
plexes, and [Cu(phen)2] complexes ). The most common
of these is nucleobase oxidation at G-sites that is due to the
relatively low redox potential for guanine oxidation, as well as
(
19) von Nicolaou, K. C.; Dai, W.-M.; Wendeborn, S. V.; Smith, A. L.; Torisawa,
Y.; Maligres, P.; Hwang, C.-K. Angew. Chem., Int. Ed. Engl. 1991, 30,
1
032-1036.
(43) Ramkumar, D.; Kalpana, M.; Varghese, B.; Sankararaman, S.; Jagadeesh,
M. N.; Chandrasekhar, J. J. Org. Chem. 1996, 61, 2247-2250.
(44) Balzani, V.; Scandola, F. Supramolecular Photochemistry; Ellis Hor-
wood: New York, 1991.
(45) Vlcek, A. Coord. Chem. ReV. 1998, 177, 219-256.
(46) Ramamurthy, V., Schanze, K. S., Eds. Organic and Inorganic Photochem-
istry; Marcel Dekker: New York, 1998; Vol. 2.
(47) Burrows, C. J.; Muller, J. G. Chem. ReV. 1998, 98, 1109-1152.
(48) Armitage, B. Chem. ReV. 1998, 98, 1171-1200.
(49) Liang, Q.; Ananias, D. C.; Long, E. C. J. Am. Chem. Soc. 1998, 120, 248-
257.
(
20) Wender, P. A.; Zercher, C. K.; Beckham, S.; Haubold, E.-M. J. Org. Chem.
1
993, 58, 5867-5869.
(
21) Funk, R. L.; Young, E. R. R.; Williams, R. M.; Flanagan, M. F.; Cecil, T.
L. J. Am. Chem. Soc. 1996, 118, 3291-3292.
(
22) Evenzahav, A.; Turro, N. J. J. Am. Chem. Soc. 1998, 120, 1835-1841.
23) Kaneko, T.; Takahashi, M.; Hirama, M. Angew. Chem., Int. Ed. 1999, 38,
(
1
267-1268.
(
(
(
(
(
24) Choy, N.; Blanco, B.; Wen, J.; Krishan, A.; Russell, K. C. Org. Lett. 2000,
2
, 3761-3764.
25) Jones, G. B.; Wright, J. M.; Plourde, G., II.; Purohit, A. D.; Wyatt, J. K.;
Hynd, G.; Fouad, F. J. Am. Chem. Soc. 2000, 122, 9872-9873.
(50) Muller, J. G.; Hickerson, R. P.; Perez, R. J.; Burrows, C. J. J. Am. Chem.
Soc. 1997, 119, 1501-1506.
26) Purohit, A.; Wyatt, J.; Hynd, G.; Wright, J.; El-Shafey, A.; Swamy, N.;
Ray, R.; Jones, G. B. Tetrahedron Lett. 2001, 42, 8579-8582.
(51) Muller, J. G.; Zheng, P.; Rokita, S. E.; Burrows, C. J. J. Am. Chem. Soc.
1996, 118, 2320-2325.
27) Clark, A. L.; Davidson, E. R.; Zaleski, J. M. J. Am. Chem. Soc. 2001, 123,
2
650-2657.
(52) Chen, X.; Burrows, C. J.; Rokita, S. E. J. Am. Chem. Soc. 1992, 114, 322-
28) Warner, B. P.; Millar, S. P.; Broene, R. D.; Buchwald, S. L. Science 1995,
325.
2
69, 814-816.
(53) Muller, J. G.; Chen, X.; Dadiz, A. C.; Rokita, S. E.; Burrows, C. J. J. Am.
Chem. Soc. 1992, 114, 6407-6411.
(
(
(
29) Konig, B.; Pitsch, W.; Thondorf, I. J. Org. Chem. 1996, 61, 4258-4261.
30) Basak, A.; Shain, J. C. Tetrahedron Lett. 1998, 39, 3029-3030.
31) Basak, A.; Shain, J. C.; Khamari, U. K.; Rudra, K. R.; Basak, A. J. Chem.
Soc., Perkin Trans. 1 2000, 1955-1964.
(54) Burrows, C. J.; Rokita, S. E. Acc. Chem. Res. 1994, 27, 295-301.
(55) Misra, M.; Alcedo, K. J. A.; Wetterhahn, K. E. Carcinogenesis 1994, 15,
2911-2917.
(
(
32) Basak, A.; Rudra, K. R. Tetrahedron Lett. 2000, 41, 7231-7234.
33) Basak, A.; Bdour, H. M.; Shain, J. C.; Mandal, S.; Rudra, K. R.; Nag, S.
Bioorg. Med. Chem. Lett. 2000, 10, 1321-1325.
(56) Yamamoto, K.; Kawanishi, S. J. Biol. Chem. 1989, 264, 15435-15440.
(57) Yamamoto, K.; Kawanishi, S. J. Biol. Chem. 1991, 266, 1509-1515.
(58) Oikawa, S.; Kawanishi, S. Biochemistry 1996, 35, 4584-4590.
(59) Johnston, D. H.; Cheng, C.-C.; Campbell, K. J.; Thorp, H. H. Inorg. Chem.
1994, 33, 6388-6390.
(
34) Coalter, N.; Concolino, T. E.; Streib, W. E.; Hughes, C. G.; Rheingold, A.
L.; Zaleski, J. M. J. Am. Chem. Soc. 2000, 122, 3112-3117.
(
35) Benites, P. J.; Rawat, D. S.; Zaleski, J. M. J. Am. Chem. Soc. 2000, 122,
(60) Fedorova, O. S.; Podust, L. M. J. Inorg. Biochem. 1988, 34, 149-155.
(61) Gorner, H.; Stradowski, C.; Schulte-Frolinde, D. Photochem. Photobiol.
1988, 47, 15-29.
7
208-7217.
(
(
36) Konig, B. Eur. J. Org. Chem. 2000, 381-385.
37) O’Connor, J. M.; Lee, L. I.; Gantzel, P.; Rheingold, A. L.; Lam, K.-C. J.
Am. Chem. Soc. 2000, 122, 12057-12058.
(62) Yamamoto, K.; Inoue, S.; Yamazaki, A.; Yoshinaga, T.; Kawanishi, S.
Chem. Res. Toxicol. 1989, 2, 234-239.
(
(
(
38) Rawat, D. S.; Benites, P. J.; Incarvito, C. D.; Rheingold, A. L.; Zaleski, J.
(63) Carter, P. J.; Cheng, C.-C.; Thorp, H. H. Inorg. Chem. 1996, 35, 3348-
3354.
M. Inorg. Chem. 2001, 40, 1846-1857.
39) Schmitt, E. S.; Huffman, J. C.; Zaleski, J. M. Chem. Commun. 2001, 167-
(64) Cheng, C. C.; Goll, J. G.; Neyhart, G. A.; Welch, T. W.; Singh, P.; Thorp,
H. H. J. Am. Chem. Soc. 1995, 117, 2970-2980.
1
68.
40) O’Connor, J. M.; Friese, S. J.; Tichenor, M. J. Am. Chem. Soc. 2002, 124,
(65) Stubbe, J.; Kozarich, J. W. Chem. ReV. 1987, 87, 1107-1136.
(66) Pyle, A. M.; Long, E. C.; Barton, J. K. J. Am. Chem. Soc. 1989, 111, 4520-
4522.
3
506-3507.
(
(
41) Rawat, D. S.; Zaleski, J. M. Synth. Commun. 2002, 32, 1489-1494.
42) Horvath, O.; Stevenson, K. L. Charge-Transfer Photochemistry of Coor-
dination Compounds; VCH: New York, 1993; p 380.
(67) Sigman, D. S.; Bruice, T. W.; Mazumder, A.; Sutton, C. L. Acc. Chem.
Res. 1993, 26, 98-104.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6435

A R T I C L E S
Benites et al.
the ability of transition metals to bind to N7. Reactions such as
were contained in either anaerobic (Kontes) quartz tubes or Pyrex
J-Young NMR tubes (the outer wall diameter was 0.38 mm, and
transmissions of >60% at 320 nm72 and >95% at wavelengths above
6
8
2+
69
•
FeEDTA/H2O2 or CuCl2 /H2O2 that generate OH or
Mn(V)-oxo intermediates7
0,71
formed by two-electron oxidation
3
60 nm were observed; the samples were placed in a temperature-
can perform both ribose and nucleobase oxidation. However,
controlled recirculating bath (0-3 °C). Electronic spectra were obtained
at ambient temperature on a Perkin-Elmer model Lambda 19 UV-
Vis/near-IR spectrometer, using 1-cm quartz cells retrofitted with a
Kontes Teflon stopcock assembly to ensure the integrity of air-sensitive
compounds.
•
the potential to generate freely diffusing OH with such reagents
can lead to reduced specificity. Thus, it would be potentially
beneficial to have molecules that can be activated with visible
excitation to produce localized radicals that could perform site-
directed H-atom abstraction rather than thermodynamically
determined base oxidation. Overall, such approaches may lead
to reagents with improved specificity, where recognition is
determined solely by the targeting element.
2
.1.3. Syntheses. The syntheses of cis-1,8-bis(pyridin-3-oxy)oct-
-ene-2,6-diyne (bpod, 1), as well as the [Cu(bpod) ]PF (2) and
[Cu(bpod) ](NO ) (3) analogues, have been reported elsewhere. The
4
2
6
35
2
3 2
corresponding Zn(bpod)
follows: A 50-mL round-bottom flask was charged with 41 mg (0.19
mmol) of Zn(CH COO) and dissolved in 5 mL of reagent-grade
2 3 2
(CH COO) compound (4) was prepared as
Within this theme, we report the photoelectronic Bergman
cyclization of copper metalloenediynes via MLCT excitation.
The photochemical products from these reactions show good
yields of the cycloaromatized ligand, indicating that the reaction
is not a consequence of irreversible redox chemistry. These
reactivities are in contrast to those observed upon photolysis of
3
2
methanol. Subsequently, 114 mg (0.39 mmol) of 1 was dissolved in 5
mL of methanol and added to the zinc acetate solution. The solution
was allowed to stir overnight as the desired complex was formed. To
this mixture, 20 mL of diethyl ether was added and the solution cooled
to 0 °C for 24 h, which produced a white solid. Yield: 67.5 mg (45%).
2+
the free ligand or the Zn analogue, suggesting that MLCT
excitation plays an important role in these photoreactions.
Moreover, we show that the intermediates produced upon
photolysis are potent to degrade both bacterial plasmid DNA
and a 25-base-pair (25-bp), double-stranded oligonucleotide. The
cleavage products derived from the oligonucleotide are indica-
tive of H-atom abstraction from the deoxyribose ring and not
base oxidation.
1
H NMR (400 MHz, CD
H), 5.97 (s, 4H), 7.36-7.43 (m, 8H), 8.23 (d, 4H), 8.34 (d, 4H).
NMR (CD CN): 22.75 (CH ), 57.56 (OCH ), 85.28 (Cquat), 92.10
3
3
CN) δ (ppm): 0.92 (s, 6H, 2CH ), 4.92 (s,
13
8
C
3
3
2
(Cquat), 120.86 (CH), 124.68 (CH), 125.69 (CH), 138.78 (CH), 143.52
+
(CH), 155.0 (Cquat), 179.69 (Cquat). ESI-MS m/z: 704 (M -CH
3
COO),
6
4
45, 473, 414, 353, 291. Anal. Calcd for C40
.49; N, 7.33. Found: C, 63.05; H, 4.16; N, 7.62.
The compound trans-1,8-bis(pyridin-3-oxy)oct-4-ene-2,6-diyne
34 4 8
H N O Zn: C, 62.88; H,
41
(t-bpod, 1a) was prepared using a literature procedure. The reaction
73
2. Experimental Section
of 3-prop-2-ynyloxy-pyridine with trans-dichloroethylene, in the
presence of Pd(0), CuI, and n-BuNH in benzene, led to the formation
of desired product in 45% yield. 1H NMR (400 MHz, CDCl
) δ
ppm): 4.87 (s, 4H, OCH ), 6.00 (s, 2H, CH), 7.25 (m, 4H), 8.27 (d,
H), 8.36 (s, 2H). 13C NMR (CDCl
) δ (ppm): 56.68 (OCH ), 85.49
2
2.1. Syntheses and Photochemistry. 2.1.1. Materials. All prepara-
3
tions were performed under a nitrogen atmosphere, using standard
Schlenk and drybox techniques. Acetonitrile and dichloromethane were
dried and distilled from calcium hydride. Dichloromethane and di-
methylformamide (DMF) were stored over 4 Å molecular sieves. The
acetonitrile used was HPLC grade, whereas the methanol, ether, and
isopropyl alcohol used were spectroscopic grade. NMR solvents
(
2
2
3
2
(
(
[
Cquat), 88.51 (Cquat), 121.08 (CH), 121.63 (CH), 123.15 (CH), 138.26
CH), 142.89 (CH), 155.56 (Cquat). HRMS (EI), m/z: 290.1067
HRMS Calcd for C18
.1.4. Photolysis of bpod (1). In a sealed J-Young tube, 20 mg (0.07
mmol) of 1 was dissolved in dry, degassed CD CN and a 10-fold excess
14 2 2
H N O : 290.1055].
2
3 6 6 3
(acetonitrile-d , dimethylsulfoxide-d (DMSO-d ), and chloroform-d )
3
were dried over activated 3 Å and 4 Å molecular sieves and degassed
by bubbling with nitrogen for 30 min. Degassed spectroscopic-grade
solvents were used for all UV-Vis measurements. All chemicals used
in the syntheses of all metalloenediynes were of the highest purity
available and were obtained from Aldrich and Fluka and used as
received. The organic enediyne compounds and photoproducts were
purified by flash chromatography with 60 Å silica gel (430-230 µm)
using HPLC-grade hexanes, ethyl acetate, dichloromethane, and
of 1,4-cyclohexadiene was added to achieve a total volume of 700 µL.
The solution was photolyzed with λ g 395 nm in a temperature-
controlled bath (0-3 °C) for 72 h, after which time no precipitate
formed, nor was any reaction detected by NMR. The experiment was
repeated at 320 and 345 nm and the same result was obtained. On a
preparative scale, in a sealed (Kontes) quartz tube, 35 mg (0.12 mmol)
of 1 was dissolved in dry, degassed CH
3
CN (8 mL). To this solution,
22 mL (287 mmol) of dry, degassed 2-propanol was added as a H-atom
methanol.
donor (9.5 M, >2300-fold excess). The resulting concentration of 1 in
solution was 4 mM. After photolysis for 72 h, the solvent was removed
1
2
.1.2. Physical Measurements. H and 13C NMR were recorded on
a Varian model VXR400 NMR spectrometer. The multiplicity of the
in vacuo and the remaining yellow oil was redissolved in CD
H NMR and C NMR spectra confirmed the identity of the yellow
3
CN. The
13
C signals was determined by the DEPT technique. High-resolution
1
13
electron ionization (EI) and fast ion bombardment (FAB) mass
spectroscopy (MS) spectra were acquired on a Kratos model MS-80
mass spectrometer that was interfaced with a Kratos model DS90 data
system. ESI-MS data were obtained at the University of Illinois with
a Micromass Quattro I mass spectrometer, and elemental analyses on
the samples were obtained from Atlantic Microlab, Inc. Photolyses were
performed using a 1000-W XeHg lamp (Oriel model 66021) at λ g
oil as starting material 1, indicating that, even with a high H-atom donor
concentration, no reactivity is observed.
2 6
2.1.5. Photolysis of [Cu(bpod) ]PF (2). Under anaerobic conditions,
a Schlenk tube was charged with 20 mg (0.03 mmol) of 2 and dis-
solved in ∼20 mL of dry, degassed (five freeze-pump-thaw cycles)
3
CH CN. To this solution, either a 10-fold excess of degassed 1,4-
cyclohexadiene or >5000-fold excess of 2-propanol was added as a
H-atom donor. In both cases, the final concentration of 2 in solution
was 1 mM. These solutions were photolyzed with λ g 395 nm within
a temperature-controlled bath (0-3 °C) for 4 h (2-propanol) to 12 h
320, 345, or 395 nm, where the cutoff wavelengths were selected using
a series of long-wavelength-pass Schott filters (Oriel 51872). Samples
(
68) Tullius, T. D.; Dombroski, B. A.; Churchill, M. E. A.; Kam, L. In Methods
in Enzymology, Vol. 155; Wu, R., Ed.; Academic Press: San Diego, CA,
1
987; pp 537-558.
(72) Murov, S. L. In Handbook of Photochemistry; 2nd ed.; Hug, G. L.,
Carmichael, I., Eds.; Marcel Dekker: New York, 1993; p 420.
(73) Brandman, H. A.; Freudewald, J. E.; Manowitz, M.; Nikawitz, E. J.;
Sharpell, F. H., Jr. Iodopropargyl Pyridyl and Picolinyl Ethers and
Thioethers as Paint Fungicides. U.S. Patent 4,170,704, October 9, 1979.
(
69) Shimizu, M.; Inoue, H.; Ohtsuka, E. Biochemistry 1994, 33, 606-613.
70) Mestre, B.; Pratviel, G.; Meunier, B. Bioconjugate Chem. 1995, 6, 466-
(
4
72.
(71) Gravert, D. J.; Griffin, J. H. Inorg. Chem. 1996, 35, 4837-4847.
6
436 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003
9

Photo-Bergman Cyclization of Cu Metalloenediynes
1,4-cyclohexadiene), depending upon the concentration of H-atom
A R T I C L E S
(
2.1.8. Photolysis of Zn(bpod)
2
(CH
3
COO)
2
(4). In a J-Young NMR
donor. Upon photolysis, the solutions change to a brown color, and a
precipitate begins to form within 1 h. To isolate the photoproduct,
larger-scale photolyses were performed (∼275 mg) using the same ratios
of H-atom donor. Upon completion of the photolyses, the solutions
were concentrated, stirred with a 1:3 mixture of degassed ether:
3
tube, 31 mg (0.04 mmol) of 4 was dissolved in dry, degassed CD CN.
To this solution, 37 µL (0.39 mmol) of 1,4-cyclohexadiene was added
to achieve a final volume of 800 µL (51.3 mM 4). The reaction mix-
ture was photolyzed at 0 °C for 24 h with λ g 395 nm, and the progress
1
of the reaction was monitored by H NMR. No change in the solu-
1
CH
was subsequently washed with ether and dried in vacuo to yield ∼135
mg (47%) of solid metal-containing photoproduct. H NMR and
NMR spectra of the solid were obtained in degassed CD CN and
3
CN to precipitate the product, and filtered anaerobically. The solid
tion color or composition was detected, and the H NMR spectrum of
the reaction mixture remained unchanged. The experiment was re-
peated with λ g 320 and 345 nm (using quartz and Pyrex cells,
respectively) and no reactivity was observed. Compound 4 was also
photolyzed under the same conditions, except for the presence of either
1
13
C
3
showed exclusive formation of two Bergman-cyclized ligands bound
to Cu(I). The NMR spectrum of this species matches that of the
excess Cu(NO
3
)
2
or Cu(CH
3
CN)
4
PF
6
(∼5-fold excess). Isolation of the
3
5 1
thermally cyclized product previously reported very well. H NMR
organic material by demetalation with EDTA as described previously
and subsequent NMR analyses revealed the presence of only the starting
material.
(
(
(
1
(
400 MHz, DMSO-d
6
) δ (ppm): 5.40 (s, 8H, OCH
2
Ph), 7.40-7.45
m, 8H), 7.56-7.59 (m, 4H), 7.69-7.71 (m, 4H), 8.39 (d, 4H), 8.54
1
3
d, 4H). C NMR (DMSO-d
25.67 (CH), 129.9 (CH, benzene), 130.38 (CH, benzene), 136.37
Cquat, benzene), 139.6 (CH), 143.54 (CH), 155.99 (Cquat). ESI-MS
6
) δ (ppm): 68.9 (OCH
2
Ph), 122.9 (CH),
2.1.9. Triplet-Sensitized Photolyses of 4. Triplet-sensitized pho-
tolyses of 4, using acetophenone, were performed both on the analytical
(J-Young tube) and preparative scales, using 1:9.1 molar ratios of 4:1,4-
cyclohexadiene, with molar acetophenone ratios in the range of 0.13-
1.09. In each case, the progress of the photoreaction was monitored by
63,65
+
m/z: 649.0, 647.2;
32 4 4
Cu [M Calcd for C36H N O Cu: 649.2/647.2],
3
57.1, 355.1, 293.2.
1
To analyze the filtrate, the solvent was removed under vacuum, and
H NMR. To fully characterize the reaction products, 248 mg of 4 (0.33
repeated washings of the viscous brown oil with ether gave a light-
brown solid. The solid was subsequently dissolved in degassed
3
mmol) was dissolved in dry and degassed CH CN, and 1,4-cyclohexa-
diene (0.28 mL, 3.0 mmol) and acetophenone (0.35 mmol) were added.
The resulting reaction mixture yields a concentration of 16.3 mM for
4 and a sensitizer concentration of 17.5 mM (A345 ) 2.4, ꢀ345 ) 139
CH
yield ∼90 mg of solid material. The H and C NMR spectra
confirmed the identity of the solid as unreacted starting material
38%-40% initial mass of 2).
.1.6. Stoichiometric Analysis of [Cu(bpod)
The solid inorganic photoproduct (∼90 mg) was dissolved in ∼25 mL
of DMF. A 10-fold excess of aqueous Na EDTA solution was added
3
CN and filtered. Solvent was then removed from the filtrate to
1
13
35
-
1
-1
M
cm ). Photolysis was conducted at 0 °C for 25 h with λ g 395
1
(
nm. New H NMR resonances at 4.94 and 6.06 ppm were observed
after 6.5 h. In addition, the 13C NMR spectrum revealed new alkyne
resonances, at 85.46 and 88.51 ppm. To allow completion of the
reaction, photolysis was continued for 6 days; however, maximum
conversion was observed after 72 h (40%). After photolysis, the solvent
was removed and the solid obtained was washed with ether. The crude
reaction material was dissolved in 20 mL of DMF and treated with a
2
2 6
]PF Photoproduct.
2
to sequester the metal, and the mixture was allowed to be stirred
gently at 0 °C for 12 h. The precipitate was filtered and the
uncomplexed organic material was extracted with dichloromethane. The
organic layer was then washed with water, using a separatory funnel,
1
0-fold molar excess of Na
resulting precipitate was filtered off and the organic material was
extracted with CH Cl and washed with excess H O. After metal
2
EDTA and stirred for 10 h at 0 °C. The
and dried over anhydrous Na
by flash chromatography (20% MeOH/CH
2
SO
4
. The resulting brown oil was purified
Cl , R ) 0.47) and had a
2
2
2
2
2
f
1
extraction, the H NMR spectrum of the reaction mixture shows the
presence of new features at 4.94 and 6.00 ppm. The 13C NMR spectrum
revealed the presence of two new alkyne resonances, at 85.46 and 88.51
ppm, as well as aromatic features, at 121.06, 121.60, and 142.88 ppm.
The new H and C NMR signals match those of the independently
synthesized, authentic t-bpod, 1a (vide supra). Other than unreacted
starting material, no additional species were detected in solution.
yield of 87% (based on 2 mol of ligand per mole of reacted com-
1
plex) of the dark-yellow Bergman-cyclized ligand. H NMR (400
MHz, CDCl
H), 7.52 (m, 2H), 8.20 (d, 2H), 8.39 (d, 2H). C NMR (CDCl
δ (ppm): 68.7 (OCH ), 121.24 (CH), 124.35 (CH), 129.24 (CH,
3
) δ (ppm): 5.24 (s, 4H, OCH
2
Ph), 7.24 (m, 4H), 7.41 (m,
1
3
2
3
)
1
13
2
benzene), 129.64 (CH, benzene), 134.75 (Cquat, benzene), 138.45
+
(
CH), 142.88 (CH), 155.04 (Cquat). FAB-MS, m/z: 293 (M + H),
2
.1.10. Electrochemistry. All measurements were made using dried
CN (distilled) or DMF (sieves) with tetramethylammonium hexaflu-
orophosphate (TMAPF , Aldrich) as the supporting electrolyte, which
1
16 2 2
98. HR-MS (EI), m/z: 292.1208 [HR-MS Calcd for C18H N O :
CH
3
2
92.1214].
6
2
.1.7. Photolysis of [Cu(bpod)
2
](NO
3
)
2
(3). The same anaerobic
PF was employed for
was stored in a vacuum oven at 80 °C to remove traces of water.
Electrodes were fabricated from a 3-mm-diameter glassy carbon rod
procedure used for the photolysis of Cu(bpod)
2
6
photoreaction of 3 in solution. Briefly, a degassed solution of 3 (200
mg, 0.30 mmol, 1 mM) containing a 10-fold excess of 1,4-cyclohexa-
diene was photolyzed at 3 °C for 12 h with λ g 395 nm. The green
solution changed to a yellow-brown color after only 4 h, indicating
reactivity of the complex in solution. The solution was concentrated
and, upon addition of ether, a dark-brown-green precipitate formed.
The solid was filtered and washed with ether (96 mg, 48%). Because
of the paramagnetism of divalent copper, a cyclized product could not
be straightforwardly confirmed by NMR; thus, the metal was extracted
using EDTA and the organic products were subsequently analyzed as
described for 2 (vide supra). The isolated yield of cyclized ligand
(Grade GC-20; Tokai Electrode Manufacturing Co., Tokyo, Japan) and
press-fitted into a Teflon shroud to provide a planar, circular electrode
2
with an area of 0.071 cm . Before use, the electrode was cleaned with
an aqueous suspension of 0.05-µm alumina (Buehler) on a Master-
Tex polishing pad (Buehler). For all experiments, the reference electrode
was a saturated cadmium amalgam in contact with DMF saturated with
both cadmium chloride and sodium chloride. This electrode has a
potential of -0.76 V, versus the aqueous saturated calomel electrode
4,75
(
SCE), at 25 °C.7 All quoted potentials have been converted to the
SCE scale, and E1/2 values were calculated from an average of Epa and
pc. Cyclic voltammograms were obtained in cells previously de-
scribed,76 with the aid of a Princeton Applied Research Corporation
PARC) model 175 universal programmer in combination with a PARC
E
product, based on 2 mol of ligand per mole of reacted complex, was
1
3
6 mg (81%). H NMR (400 MHz, CDCl
3
) δ (ppm): 5.24 (s, 4H,
(
OCH
2
Ph), 7.24 (m, 4H), 7.41 (m, 2H), 7.52 (m, 2H), 8.21 (d, 2H),
.39 (d, 2H). 13C NMR (CDCl
model 173 potentiostat-galvanostat and were recorded on a Yokogawa
model 3023 X-Y recorder.
8
1
3
) δ (ppm): 68.72 (OCH ), 121.24 (CH),
24.31 (CH), 129.24 (CH, benzene), 129.66 (CH, benzene), 134.75
2
(
Cquat, benzene), 138.45 (CH), 142.88 (CH), 155.04 (Cquat). FAB-
(
74) Marple, L. W. Anal. Chem. 1967, 39, 844-846.
+
MS, m/z: 293 (M + H), 198. HR-MS (EI), m/z: 292.1210 [HR-MS
Calcd for C18
(
75) Manning, C. W.; Purdy, W. C. Anal. Chim. Acta 1970, 51, 124-126.
H N
16 2
O
2
: 292.1214].
(76) Vieira, K. L.; Peters, D. G. J. Electroanal. Chem. 1985, 196, 93104.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6437

A R T I C L E S
Benites et al.
2
.1.11. Photochemical Quantum Yield Determination. Quantum
2.2. DNA Photocleavage Assays. 2.2.1. Agarose Gel Electro-
phoresis. Bacterial plasmid pUC19 DNA was obtained from New
England Biolabs and used immediately as received. Quantitation of
yield measurements were performed via photolysis at 366 nm with a
000-W HgXe source, utilizing a series of long- (295, 345, 360 nm)
and short-wavelength-pass filters (450 nm), as well as a 366 nm
1
81
DNA was conducted using established procedures and yielded 250
-
4
6
notch filter. Freshly purified 1 (33 mg, 1.14 × 10 mol) was pumped
µg of pUC19 (MW ) 1.70 × 10 g/mol, 2686 base pairs (bp)). The
into the drybox and dissolved into 2.0 mL of dry degassed CD CN.
3
plasmid DNA was digested to linear (Form III) DNA by incubating a
mixture of 3 µL of Hind III buffer (10×) and 1.5 µL of Hind III enzyme
(20 000 units/mL, New England Biolabs), 6.6 µL of pUC19, and 18.9
µL of deionized water for 5 h at 37 °C.
-
5
To the resulting solution, a solution of 21.4 mg (5.74 × 10 mol)
Cu(CH CN) PF in 1.0 mL CH CN-d was added dropwise. A 460 µL
aliquot was then transferred to a J-Young tube and 340 µL (504 molar
eq) of dry, degassed 2-propanol-d was added. The resulting solution
11.0 mM 2 in a total volume of 800 µL) was then photolyzed in the
3
4
6
3
3
8
Thermal and photochemical cleavage reactions were performed under
deoxygenated conditions in airtight Pyrex conical vials with 50 µM bp
of pUC19 in 10 mM tris-acetate pH 7.6 buffer at λ g 395 nm, t ) 8
h, T ) 20 °C. All samples were quenched using liquid nitrogen after
photolysis, and samples were subsequently stored at -20 °C. The
concentration of compounds 2 and 3 ranged from 500 to 12.5 µM in
a total volume of 100 µL. The samples for compound 3 were prepared
and degassed individually by five freeze-pump-thaw cycles in airtight
Pyrex conical vials. Compound 2 was handled under a nitrogen
atmosphere in an inert-atmosphere box, and all samples were prepared
using degassed 10 mM tris-acetate buffer at pH 7.6. Reaction samples
(7 µL) were run on a on a 1% nondenaturing agarose gel (SeaKem
Gold agarose, FMC BioProducts Seakem and Genetic Technology
Grade) prepared with 1× TAE buffer (60 V, 3 h) and stained with EB
(10 mg/mL).
(
J-Young tube and changes in the concentration of 2 were monitored
by integration of the olefin resonance at 5.97 ppm in the H NMR
1
spectrum, relative to the residual CH
measurements were performed using the ferrioxalate method,72 using
00 µL sample volumes in sealed J-Young NMR tubes. To reduce the
error in the actinometry, a total of 15 measurements were made across
three groups of five samples at various times throughout the photolysis.
3
CN proton resonance. Actinometry
8
2
.1.12. DNA Binding Constant Evaluation. The binding of 3 to
calf thymus DNA (Sigma) was determined by a Scatchard analysis via
displacement and competitive inhibition of ethidium bromide (EB)
binding.
by monitoring the fluorescence intensity of EB upon titration of DNA
solutions containing varying concentrations of 3 with increasing
77-79
Specifically, the association constant for 3 was determined
concentrations of EB. The amount of bound EB (c
using eq 1:
b
) was calculated
As control experiments, 50 µM DNA was photolyzed alone at λ g
395 nm, T ) 20 °C (8 h), and complexes 2 and 3 (25 µM) were
incubated for 12 h with 50 µM bp plasmid DNA in darkness under
aerobic and anaerobic conditions (10 mM tris-acetate buffer, pH 7.6,
200 µL). In addition, inhibitor studies were also conducted according
7
8
It - I
0
cb )
(1)
(
V - 1)k
to a literature procedure82 with 0.5 M ethanol and 100 µg/mL superoxide
where I
0
is the fluorescence intensity of free EB, I
t
the total fluorescence
vs C (the
dismutase (SOD). For the SOD inhibition studies,82 1 mg of protein
was dissolved in 300 µL of 10 mM tris-acetate buffer (pH 7.6).83
intensity when DNA is present, k the slope of a plot of I
0
0
concentration of EB added), and V the fluorescence ratio of the bound
EB to free EB. The concentration of free EB and the ratio of bound
EB to the total nucleotide concentration (reb) was then calculated. From
these data, Scatchard plots were constructed using a competitive
Aliquots of the samples (7 µL) were stained with EB (10 mg/mL) and
analyzed on a 2% nondenaturing agarose gel (SeaKem Gold agarose,
FMC BioProducts Seakem and Genetic Technology Grade) prepared
with 1× TAE buffer (60 V, 3 h).
7
7,79
inhibition model (eq 2):
2.2.2. Radiolabeling/Non-Denaturing Polyacrylamide Gel Elec-
trophoresis. Synthetic oligonucleotides were purchased from the
Lineberger Comprehensive Cancer Center Nucleic Acids Core Facility
and purified by gel electrophoresis on a 20% polyacrylamide gel with
reb
Keb
)
(n - reb)
(2)
cf
[
(1 + K c )
]
m
m
7
M urea (Mallinckrodt).84 All experiments containing DNA (except
HPLC experiments) were performed using a 25-bp oligonucleotide with
where K
m
and c
m
are the association constant and concentration of bound
3
2
competitor, respectively, n is the number of binding sites, and c
concentration of free EB. Because Kobs ) [Keb/(1 + K )], the
) was determined from a plot of Keb/Kobs
f
is the
the following sequence ( P-labeled strand): d[5′-(ATG GCG TAA
TCA TGG TCA TAG CTG T)-3′]. Water was purified with a MilliQ
purification system (Millipore). Buffer salts were purchased from
Mallinckrodt. Gel electrophoresis reagents such as acrylamide, agarose,
and 10× TBE buffer (0.89 M tris, 0.89 M boric acid, 20 mM EDTA,
pH 8.3) were purchased from BioRad.
m m
c
association constant for 3 (K
vs c (eq 3).
m
m
7
9
Keb
)
K c + 1
(3)
m
m
Kobs
Solution concentrations of metal complex or DNA were determined
-
1
-1
by spectrophotometry (ꢀ395 ) 167 M cm for 3), using a Hewlett-
Packard model HP 8452 diode-array spectrophotometer. Oligonucleotide
strand concentrations were determined by absorbance at 260 nm with
Concentration ratios of bound EB to free EB were determined by
monitoring the fluorescence intensity of EB at λem ) 600 nm upon λex
8
5
)
510 nm at 25 °C in a 10 mM tris-acetate buffer at pH 7.6. The
a calculated extinction coefficient from the nearest-neighbor equation.
32
concentrations of 3 and EB were varied from 0 to 40 µM and from 4
to 12 µM, respectively, with a corresponding nucleotide concentration
of 36 µM, which was determined by the optical density at 260 nm
Single-stranded oligonucleotide was 5′-radiolabeled using γ P-ATP
(New England Nuclear). A 25 µL solution containing 1 µL of single-
stranded oligonucleotide (5 µM), 5 µL of forward reaction buffer (5×),
-
1
-1
80
(ꢀ
260 ) 6600 M cm per nucleotide). Samples were incubated for
(
81) Ausubel, F. M.; Brent, R.; Kingston, R. E.; Moore, D. D.; Seidman, J. G.;
Smith, J. A.; Struhl, K. In Current Protocols in Molecular Biology; 2nd
ed.; John Wiley & Sons: New York, 1994-1997.
1
0 h at 4 °C and, subsequently, 1 h at 25 °C prior to spectral analyses.
Control experiments revealed that neither 3 nor CuCl
on the fluorescence intensity of free EB. Similarly, CuCl
influence the fluorescence intensity of DNA-bound EB.
2
have an effect
does not
(82) Speetjens, J. K.; Parand, A.; Crowder, M. W.; Vincent, J. B.; Woski, S. A.
2
Polyhedron 1999, 18, 2617-2624.
(
83) SOD was purchased from Sigma and contained 3800 units per milligram
of solid. One unit corresponds to the amount of catalytic enzyme that inhibits
the autoxidation of pyrogallol by 50% at pH 8.2 and 25 °C.
(
(
(
77) Scatchard, G. Ann. N. Y. Acad. Sci. 1949, 51, 660-672.
78) Le-Pecq, J. B.; Paoletti, C. J. J. Mol. Biol. 1967, 27, 87-106.
79) Howe-Grant, M.; Wu, K. C.; Bauer, W. R.; Lippard, S. J. Biochemistry
(84) Sambrook, J.; Fritsch, E. F.; Maniatis, T. In Molecular Cloning:
A
Laboratory Manual; 2nd ed.; Cold Spring Harbor Laboratory Press:
Plainview, NY, 1989.
1
976, 15, 4339-4346.
(80) Reichmann, M. E.; Rice, S. A.; Thomas, C. A.; Dotty, P. J. Am. Chem.
(85) Fasman, G. D. Handbook of Biochemistry and Molecular Biology, Section
B; CRC Press: Cleveland, OH, 1976; Vol. I.
Soc. 1954, 76, 3047-3053.
6438 J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003

Photo-Bergman Cyclization of Cu Metalloenediynes
A R T I C L E S
Scheme 1. Synthesis of t-bpod Ligand (1a)
1
H
µL of T4 polynucleotide kinase (Life Technologies), and 17 µL MilliQ
2
O was heated in a water bath at 37 °C for 20 min. The T4
2.2.4. Fragment Modification. Procedures for modification of
oligonucleotide fragments were conducted in a manner similar to
previously published methods.86 Briefly, samples to be treated with
polynucleotide kinase was deactivated by heating the reaction solution
at 65 °C for 10 min. The P-labeled oligonucleotide (diluted to 50
µL) was then purified on a MicroSpin model G-50 column (Amersham
3
2
sodium borohydride (NaBH
for 20 min after the cleavage reaction, quenching, and ethanol
precipitation. The dried pellet was resuspended in 40 µL of NaBH
4
) were pelleted by centrifugation at 16 000g
3
2
Pharmacia Biotech). The resulting P-labeled oligonucleotide was
4
ethanol-precipitated and resuspended in MilliQ water (∼400 K cts/5
(0.2 M, Aldrich) and allowed to react for 15 min at room temperature.
The reaction was quenched with 8 µL of acetic acid and ethanol-precipi-
tated again. These samples were not subjected to piperidine treatment.
2.2.5. Denaturing Polyacrylamide Gel Electrophoresis. All samples
were pelleted by centrifugation at 16 000g for 20 min and resuspended
in 12 µL of gel-loading dye solution (80% formamide, 0.5× TBE buffer,
and orange G dye), and heated at 90 °C for 5 min. Each sample (6 µL)
was loaded into a well of 20% denaturing polyacrylamide gel containing
7 M urea and run at 65 W for 5 h. The gel was then transferred to a
phosphor imaging screen, exposed overnight, and scanned on a Storm
840 System PhosphorImager (Molecular Dynamics).
8
4
µL). Oligonucleotides labeled on the 3′ end were similarly generated
8
4
using published procedures.
Radiolabeled double-stranded oligonucleotide solutions were pre-
pared by mixing a 1:1.2 ratio of 20 µM stock solutions of unlabeled
oligonucleotide and complementary oligonucleotide in 20 mM sodium
3
2
phosphate buffer (pH 7). P-labeled oligonucleotide (5 µL, 400 cts)
was added to each 20 µL of reaction solution. The resulting mixture
was heated to 90 °C for 5 min and cooled to room temperature over 2
h to anneal the strands. Formation of the hybridized duplex was
confirmed using native gel electrophoresis.
Stock solutions of 3 (∼25 mM) were prepared in DMF (Fisher) and
3
. Results
kept at 4 °C until being diluted to 200 µM immediately before addition
3
2
32
to solutions of P-labeled oligonucleotide. Hybridized P-labeled
3
.1. Syntheses. The synthesis of cis-1,8-bis(pyridin-3-oxy)-
oligonucleotide (20 µL) was combined with the appropriate amounts
oct-4-ene-2,6-diyne (bpod, 1) can be accessed via the Sonash-
igura cross-coupling approach, using more than one route, as
have been previously described.
of MilliQ H O and 1.8 mM 3 to yield a total reaction solution (40 µL)
2
consisting of 10 mM sodium phosphate (pH 7), 10 µM oligonucleotide,
and varying concentrations (0, 90, 180, 270 µM) of 3. In a polyspring
insert tube (National Scientific Company), the reaction solution was
photolyzed for 8 h with a 300-W Hg lamp (Oriel) and a monochromator
set to 368 nm (60% transmission, 40 nm full width at half-maximum
3
5,41
The trans-1,8-bis(pyridin-
3-oxy)oct-4-ene-2,6-diyne (t-bpod, 1a) derivative was prepared
using a similar method, with substitution of trans-1,2-dichlo-
roethylene for the cis-isomer (Scheme 1). The corresponding
[Cu(bpod)2]PF6 (2) and [Cu(bpod)2](NO3)2 (3) complexes are
straightforwardly prepared using 2.1:1 bpod:Cu stoichiometry.35
The analogous Zn(bpod) (CH COO) compound (4) was simi-
larly prepared by dissolving Zn(CH3COO)2 in MeOH with a
subsequent addition of 2.1 equiv of ligand 1. The solution was
allowed to be stirred overnight and, upon addition of ether, was
then cooled to 0 °C for 24 h to produce the off-white solid in
45% isolated yield (Scheme 2).
(
fwhm)). Following photolysis, the reactions were quenched with 2
µL of EDTA (1 mM) and ethanol-precipitated with sodium acetate, as
84
described elsewhere. Select samples were treated with piperidine prior
2
3
2
8
4
to denaturing gel electrophoresis. Control reactions were also
3
2
performed with single-stranded P-labeled oligonucleotide in place
of the double-stranded 32P-labeled oligonucleotide and CuCl
in place
2
of 3.
2-
2.2.3. Hydroxyl Radical Reaction. Fe(EDTA) /H
2 2
O reactions with
68
oligonucleotides were performed according to published procedures.
3
.2. Electrochemistry. Reduction potentials for compounds
Briefly, equal amounts of a freshly prepared solution of 0.4 mM ferrous
ammonium sulfate (Aldrich) and 0.8 mM EDTA solution were
combined in an Eppendorf tube. The resulting 0.4 mM Fe(EDTA)2-
solution was diluted to 80 µM. In a clean Eppendorf tube, 5 µL of
1-3 are given in Table 1. In general, 1, in both the uncomplexed
and complexed states, exhibits two irreversible waves between
Ep values of -1.75 and -1.93 V (vs SCE), corresponding to
reductions of the alkyne units.
electron oxidation waves are also observed at +1.94 and +2.15
V (vs SCE) for 1 and 3, respectively, whereas for 2, this
2
-
43,87
ascorbate (8 mM, Sigma), 5 µL of Fe(EDTA) (80 µM), and 5 µL of
(0.024%, Mallinckrodt) were combined. Subsequently, 20 µL of
Irreversible, ligand-based one-
2 2
H O
radiolabeled oligonucleotide (20 µM) was added to the mixture. The
reaction solution was allowed to react for 12 min at room temperature,
quenched with 10 µL of thiourea (0.1 M, Mallinckrodt) and 1 µL of
EDTA (0.2 M, Mallinckrodt), and ethanol-precipitated. Selected
reactions were treated with piperidine prior to denaturing gel eletro-
(86) Breiner, K. M.; Daugherty, M. A.; Oas, T. G.; Thorp, H. H. J. Am. Chem.
Soc. 1995, 117, 11673-11679.
(
87) Martin, R. E.; Gubler, U.; Cornil, J.; Balakina, M.; Boudon, C.; Bosshard,
C.; Gisselbrecht, J.-P.; Diederich, F.; Gunter, P.; Gross, M.; Bredas, J.-L.
Chem.sEur. J. 2000, 6, 3622-3635.
8
4
phoresis.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6439

A R T I C L E S
Benites et al.
Scheme 2. Synthesis of Zn(bpod)2(CH3COO)2 (4)
Table 1. Electrochemical Data for 1-3 (vs SCE) at 25 °C
E
p
bpod (V)
compound
E1/2 Cu, 2+/+ (V)
E
p
Cu, +/0 (V)
+/0
0/−
−/2−
bpod(1)a
+1.94(ir)
-1.75(ir)
-1.80(ir)
-1.76(ir)
-1.93(ir)
-1.92(ir)
-1.93(ir)
[
[
Cu(bpod)2]PF6 (2)b
+0.15
+0.38
-0.95(ir)
-0.85(ir)
a
Cu(bpod)2](NO3)2 (3)
+2.15(ir)
a
Redox potentials recorded in CH3CN. b Redox potentials recorded in DMF. The uncertainty in the reported potential is (0.02 V.
Table 2. Electronic Absorption Spectral Data for 1-4 Obtained in
CH3CN at 25 °C
λ )
max/nm (ꢀ/M cm-1
a
-1
compound
bpod (1)
264 (21 800), 276 (20 300)
[
[
Cu(bpod)2]PF6 (2)
Cu(bpod)2](NO3)2 (3)
265 (27 400), 276 (25 900), 315 (2800)
265 (31 900), 276 (31 500), 311 (4700)
264 (18 900), 276 (17 900)
Zn(bpod)2(CH3COO)2 (4)
a
The uncertainty in the reported λmax is (2 nm.
Binding of 1 to redox-active Cu2 results in a marked increase
+
-
1
in the extinction of the transitions at 265 nm (ꢀ ) 31 900 M
-
1
-1
-1
cm ) and 276 nm (ꢀ ) 31 500 M cm ), while, to a lower
energy, a broad absorption feature is observed at 311 nm (ꢀ )
-
1
-1
2+
4
700 M cm ). To a first approximation, pyridine-to-Cu
Figure 1. Electronic absorption spectra of 1-4 in CH3CN at 25 °C.
ligand-to-metal charge-transfer (LMCT) transitions can be
identified as promotions from the pyridine σ (formally the
nonbonding lone pair) and π and π (perpendicular to the ring
oxidation process occurs outside the solvent (DMF) window
1
2
9
0
2
2
(
+1.2 V, vs SCE) and is not detected. These values compare
favorably with those obtained for the 1,8-bis(tetrahydropyran-
-yloxy)oct-4-ene-2,6-diyne precursor to 1, identifying these
plane) orbitals to the half-occupied d
x -y
orbital centered on
2
+
the Cu cation, as have been described for Cu-imidazole
9
1,92
The strongly allowed ring N(σ)-to-Cu(dx2-y2)
2
complexes.
88
waves as enediyne-localized redox processes. Two metal-
centered redox processes are also observed for 2 and 3,
corresponding to 2+/+ (reversible) and +/0 (irreversible) redox
couples. For 2, the Cu(II)/Cu(I) couple occurs at E1/2 ) +0.15
V, whereas for 3, reduction to Cu(I) occurs at a noticeably
positive potential (E1/2 ) +0.38 V), indicating the relative ease
of reduction to the Cu(I) state. Not surprisingly, the Cu(I)/
Cu(0) redox couple is irreversible and is observed at more-
negative potentials (Ep values from -0.85 to -0.95 V, vs SCE).
LMCT transition occurs at high energy (<300 nm) and is
obscured by the prominent ligand-centered transitions. Descent
x2-y2
in symmetry due to deviations from a rigorous d
square
35
plane and rotation of the pyridine rings about the metal-ligand
3
5,93
2
2
bond
permit mixing of dπ character into the Cu(d
x -y
)
HOMO. This generates pyridine(π)-Cu(HOMO) overlap and
thus pyridine(π)-to-Cu LMCT transitions of moderate intensity
9
1,92
in the 300-375 nm range
that are not affected strongly by
9
4
the enediyne unit.
10
+
35
3.3. Electronic Spectra. Electronic absorption spectra of 1-4
For the d Cu complex 2 in a pseudo-tetrahedral geometry,
the same two strong ligand-centered π-π* transitions are
observed at 265 nm (ꢀ265 ) 27 400 M cm ) and 276 nm
at ambient temperature in dry, degassed CH3CN are shown in
Figure 1, and the peak maxima are given in Table 2. The
uncomplexed enediyne ligand 1 possesses two absorption
-
1
-1
-
1
-1
(89) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric Identification
of Organic Compounds; 4th ed.; John Wiley & Sons: New York, 1981.
features, at 264 nm (ꢀ ) 21 800 M cm ) and 276 nm (ꢀ )
-
1
-1
2
0 300 M cm ), that are comprised of primarily 3-alkoxy-
pyridine π-π* transitions. Complexation of 1 to Lewis-acidic
Zn produces no shifts in the energies and only a slight decrease
(
(
(
90) Innes, K. K.; Ross, I. G.; Moomaw, W. R. J. Mol. Spectrosc. 1988, 132,
89
492-544.
91) Bernarducci, E.; Schwindinger, W. F.; Hughey, J. L.; Krogh-Jespersen,
K.; Schugar, H. J. J. Am. Chem. Soc. 1981, 103, 1686-1691.
92) Fawcett, T. G.; Bernarducci, E. E.; Krogh-Jespersen, K.; Schugar, H. J. J.
Am. Chem. Soc. 1980, 102, 2598-2604.
2+
in the molar extinction of the transitions at 264 nm (ꢀ ) 18 900
-1
-1
-1
-1
M
cm ) and 276 nm (ꢀ ) 17 900 M cm ).
(93) Pradilla, S. J.; Chen, H. W.; Koknat, F. W.; Fackler, J. P., Jr. Inorg. Chem.
1
979, 18, 3519-3522.
93
4 3 2
(94) We have synthesized and characterized [Cu(OMePy) ](NO ) that exhibits
(
88) Redox potentials for the compound 1,8-bis(tetrahydropyran-2-yloxy)oct-
ligand-centered π-π* absorption between 250 and 300 nm (ꢀ ) 15 800
-
-
1
1
-1
4
-ene-2,6-diyne are comparable (E1/2 ) +2.0, -2.05, -2.25) to the ligand-
M
M
cm ) and LMCT absorption between 300 and 350 nm (ꢀ ) 1900
-1
centered redox potentials reported in Table 2, thereby establishing the redox
process as localized on the enediyne portion of the molecule.
cm ), whereas uncomplexed OMePy absorbs strongly only between
-1
-1
250 and 300 nm (ꢀ ) 3800 M cm ).
6440 J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003

Photo-Bergman Cyclization of Cu Metalloenediynes
A R T I C L E S
Scheme 3. Photo-Bergman Cyclization of [Cu(bpod)2]+ (2) with λ g 395 nm
Scheme 4. Photo-Bergman Cyclization of [Cu(bpod)2]2 (3) with λ
+
ꢀ276 ) 25 900 M- cm ), along with a distinct Cu (dπ)-to-
1
-1
+
(
g 395 nm
pyridine(π*) MLCT envelope that is observed at 315 nm (ꢀ315
-
1
-1 95-102
)
2800 M cm ).
The pseudo-tetrahedral geometry of
the Cu(I) center, combined with free rotation of the pyridine
+
rings, once again permits strong Cu (dπ)-pyridine(π*) overlap
and thus charge-transfer transitions with π-character. Overall,
2
+
+
for both the Cu and Cu cases, these ligand-centered and
charge-transfer (CT) absorption features parallel in energys
and, more coarsely, in intensitysthose generally observed for
the corresponding tetrakis(N-heterocycle) complexes of copper.
is characteristic of two cyclized enediyne ligands bound to
Cu(I) (5, Scheme 3). The metal was isolated from the organic
transformation product by dissolving the crude solid in a DMF/
aqueous EDTA solution. Filtration and subsequent isolation by
column chromatography afforded the pure organic photoproduct.
3
.4. Photochemistry of 1-3. Photochemical reactivities of
enediynes 1-3 were studied with λ g 395 nm. Compound 1 (4
mM) was anaerobically photolyzed for 72 h at 0-3 °C in a
constant-temperature bath in the presence of 1,4-cyclohexadiene
1
The H NMR spectrum of the demetalled organic photoproduct
(CHD) (10-fold excess) or 2-propanol (>2300-fold excess).
exhibited resonances at 5.24 (OCH2Ph), 7.24, 7.41, 7.52, 8.20,
and 8.39 ppm that match those of uncomplexed, Bergman-
After photolysis for 72 h, the solvent was removed under
1
vacuum, and the resulting yellow oil analyzed by NMR ( H,
35
cyclized compound 6 (87% yield). The diamagnetic Cu(I)
1
3
1
13
C). The H and C NMR spectra confirmed the identity of
1
13
photoproduct 5 was also fully characterized ( H, C NMR,
the yellow oil as starting material, indicating that no reactivity
was observed under these reaction conditions.
HRMS) and, through comparison to the isolated thermal product
35
previously reported, was determined to possess two Bergman-
In contrast, the Cu(I) complex 2 (1 mM) was photolyzed using
CHD (10-fold excess) or 2-propanol (> 5000-fold excess) as a
H-atom donor in degassed CH3CN under anaerobic conditions.
The H NMR spectrum indicates a new resonance at 5.40 ppm
and a change in the chemical shifts in the aromatic region after
cyclized ligands bound to Cu(I) (Scheme 3).
Photolysis of [Cu(bpod)2](NO3)2, 3, changes the color of the
bright-green starting solution to yellow-brown after only 4 h.
The solution was photolyzed for a total of 12 h; however, unlike
the Cu(I) photolysis reaction, no precipitate formed. The solvent
from the reaction mixture was removed and, upon the addition
of ether, a dark brown-green powder was obtained in 48%
yield. The same workup used to isolate the organic photoproduct
for 2 was employed to remove the metal and extract the reacted
1
1
h, concurrent with the formation of a brown precipitate. The
new resonances located at 5.40 ppm and in the aromatic region
became constant after 4 h (2-propanol) and 12 h (CHD) of
photolysis, respectively. After the reaction mixture was con-
centrated, an oily material was isolated and stirred in a mixture
of ether and acetonitrile (1:3), to remove the unreacted starting
material. The amount of unreacted starting material was
confirmed to be 38-40 wt %. Filtration and washing of the
1
13
organic ligand. The NMR ( H, C) and mass spectral data of
the organic photoproduct after demetalation (81% based on two
ligands/Cu) are identical to those of the Bergman-cyclized
product 6 obtained from photolysis of 2, as well as from the
thermal cyclization of 3 in solution, thereby confirming the
photo-Bergman cyclization reactivity of 3 (Scheme 4).
brown precipitate with a mixture of ether and CH3CN gave
1
4
5%-50% of the solid product. The H NMR spectrum of the
solid showed a new resonance at 5.40 ppm for the -OCH2Ph
3
.5. Photochemistry of 4. Zn(bpod)2(CH3COO)2, 4, was
unit, as well as two multiplets at 7.57 (4H) and 7.68 (4H) ppm
1
3
photolyzed using reaction conditions comparable to those
previously discussed. Photolysis of 4 using a 10-fold excess of
CHD as a H-atom donor with λ g 395 nm was conducted at a
temperature of 0-3 °C. The reaction mixture was photolyzed
from the disubstituted benzene framework. Furthermore, a
C
NMR spectrum revealed the presence of a disubstituted benzene
unit with resonances at 129.90, 130.38, and 136.37 ppm, which
1
for 24 h, and the progress of the reaction monitored by H NMR
(
95) Kitagawa, S.; Munakata, M.; Higashie, A. Inorg. Chim. Acta 1982, 59,
1
2
19-223.
spectroscopy. After 24 h, the H NMR spectrum of the reaction
(
96) Ichinaga, A. K.; Kirchhoff, J. R.; McMillin, D. R.; Dietrich-Buchecker, C.
O.; Marnot, P. A.; Sauvage, J. P. Inorg. Chem. 1987, 26, 4290-4292.
97) Sauvage, J.-P. Acc. Chem. Res. 1990, 23, 319-327.
mixture remained unchanged and revealed the presence of only
starting material, with no indication of decomposition. The same
experiment was also conducted with λ g 320 and 345 nm under
identical conditions, and, once again, no reactivity was observed.
To determine if free copper could contribute to Bergman
cyclization, 4 was photolyzed under the same conditions in the
presence of either Cu(NO3)2 or Cu(CH3CN)4PF6 (∼5-fold
(
(
98) Benniston, A. C.; Grossheny, V.; Harriman, A.; Ziessel, R. Angew. Chem.,
Int. Ed. Engl. 1994, 33, 1884.
(
(
99) Juris, A.; Ziessel, R. Inorg. Chim. Acta 1994, 225, 251-254.
100) Grosshenny, V.; Harriman, A.; Ziessel, R. Angew. Chem., Int. Ed. Engl.
1
995, 34, 1100-1102.
(
101) Simon, J. A.; Palke, W. E.; Ford, P. C. Inorg. Chem. 1996, 35, 6413-
421.
102) McMillin, D. R.; McNett, K. M. Chem. ReV. 1998, 98, 1201-1220.
6
(
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6441

A R T I C L E S
Benites et al.
excess). Isolation of the organic material and subsequent NMR
analyses revealed the presence of only starting material.
In the absence of direct photochemical reactivity of 4, a
triplet-state sensitizer (acetophenone) was employed to deter-
mine if reactivity could be induced via bimolecular energy
transfer. Photolysis of complex 4 (16.3 mM) in the presence of
CHD (9-fold excess) and acetophenone (17.5 mM) yielded new
1
H resonances at 4.94 and 6.06 ppm, as well as a multiplet in
the aromatic region after 6.5 h with no evidence of unbound
ligand or uncomplexed acetate counterions. The reaction was
allowed to continue for six days, but maximum conversion
(35-40%) was observed after 72 h. After photolysis, the solvent
was removed and the solid obtained was washed with ether.
The resulting solid was dissolved in DMF and treated with an
aqueous EDTA (excess) solution. Subsequent filtration and
extraction of the organic products with CH2Cl2 produced a light-
1
yellow oil. The H NMR spectrum of the organic reaction
mixture exhibited the presence of new resonances at 4.94 and
6
.00 ppm, and the 13C NMR spectrum showed the presence of
Figure 2. Agarose gel (1%) electrophoretic analysis of the photoinduced
cleavage of 50 µM pUC19 plasmid DNA (2686 bp) by 2 and 3 upon λ g
395 nm photolysis at 20 °C for 8 h. Form I, supercoiled plasmid; Form II,
nicked plasmid; Form III, linear plasmid. For panel a, the lanes are identified
as follows: 1, DNA (marker); 2, Hind III digest of DNA (marker); 3, DNA;
two new alkyne resonances, at 85.46 and 88.51 ppm, as well
as aromatic features at 121.06, 121.60, and 142.88 ppm. These
features are identical to those of authentic t-bpod, 1a, therefore
indicating that the triplet sensitization of 4 leads to ligand
isomerization and not Bergman-cyclized product.
4
, DNA + 12.5 µM 2; 5, DNA + 50 µM 2; 6, DNA + 100 µM 2; 7, DNA
250 µM 2; 8, DNA + 500 µM 2. For panel b, the lanes are identified as
follows: 1, DNA (marker); 2, Hind III digest of DNA (marker); 3, DNA;
+
3
.6. Quantum Yield For Photocyclization. The quantum
yield for photocyclization of 2 was performed via photolysis at
66 nm with 1000-W HgXe source utilizing a series of long-
4
2
, DNA + 10 µM 3; 5, DNA + 25 µM 3; 6, DNA + 50 µM 3; 7, DNA +
50 µM 3; 8, DNA + 500 µM 3.
3
and short-pass filters as well as a 366 nm notch filter. From a
total of 15 measurements made across three groups of five
samples at various times throughout the photolysis, the quantum
yield for photocyclization of 2 (φ366) was determined to be 4.2
those of the more strongly binding extended aromatic porphyrin
or DPPZ-containing metal complexes.
3
.8. Plasmid DNA Photocleavage Reactions. Anaerobic
photolysis (λ g 395 nm) of 2 in aqueous solution (10 mM tris-
acetate, pH 7.6) in the presence of pUC19 at 20 °C for 8 h
-
5
(
( 0.6) × 10 . The absolute magnitude of this value is low
and is consistent with the need for photolysis times exceeding
h before product is detected.
.7. DNA Binding. The association constant for 3 with calf
(Figure 2a) leads to the generation of nicked plasmid (Form II)
2
at low concentrations (12.5 µM 2, lane 4). At higher concentra-
tions of 2 (50-500 µM, lanes 5-8), a mixture of linear (Form
III) and nicked plasmid is detected, with complete consump-
tion of the supercoiled Form I. Quantitations of the band
densities of lanes 5-8 reveal an increasing linear/nicked ratio
(0.25-0.68) with increasing concentration of 2; however,
between 250 and 500 µM, the ratio is essentially constant. These
results suggest that linear Form III DNA likely is derived from
multiple nicking.
Analogous results are obtained upon photolysis of 3 under
the same conditions (Figure 2b). At low concentrations of 3
(10-25 µM, lanes 4-5), a mixture of unreacted supercoiled
Form I and nicked Form II DNA is observed. At concentrations
of 3 from 50 to 500 µM (lanes 6-8), supercoiled Form I DNA
is completely consumed and a mixture of linear Form III and
nicked Form II DNA remains. Very similar to the cleavage
results at higher concentrations of 2, the ratio of linear/nicked
DNA increases only modestly, from 0.7 to 0.98, as the
concentration of 3 increases, culminating in essentially a 1:1
ratio of the two forms. Once again, at low concentrations of 3,
the presence of nicked Form II DNA in the absence of linear
Form III (lanes 4-5) suggests that a significant contribution of
3
thymus DNA was determined by a Scatchard analysis via
displacement and competitive inhibition of EB binding at pH
7-79
7
.6 in 10 mM tris-acetate buffer.7
Binding isotherms (eq 2)
converged at a single abssica value (reb ) 0.18), confirming
competitive inhibition of EB binding by 3. These plots, as well
2
as that of Keb/Kobs vs cm (eq 3), yielded excellent linear fits (R
>
0.98) and, consequently, an association constant Km for 3 of
4
-1
2
.6 (( 0.2) × 10 M . The association constant is significantly
2+
decreased, relative to Os(DPPZ)(4,4′-amino-2,2′-bipyridine)2
7
-1 103
(
1
1
Km ) 3.1 × 10 M ), the meso-pyridal porphyrins (Km ≈
5
7
-1 104
+
0 -10 M ), and the Pt/Pd(terpy)(X) complexes (Km ≈
5 -1 79,105 x+
0 M ),
but is comparable to M(phen)y motifs such as
+
4
-1 106
3+
[
2
Cu(phen)2] (Km ) 5 × 10 M ), [Co(phen)3] (Km )
4
-1 107
2+
4
-1 107
.6 × 10 M ), [Co(phen)3] (Km ) 1.6 × 10 M ),
2+ 3 -1 107 3+
[
1
Fe(phen)3] (Km ) 7.1 × 10 M ), [Fe(phen)3] (Km )
4
-1 107
2+
3
.46 × 10 M ),
and [Ru(phen)3] (Km ) 6.2 × 10
-1 108
M ). Clearly, the minimal planar π-system of 3 yields an
intercalative association constant that is more comparable to
those of classical phenanthroline and terpy metal complexes than
(
(
(
103) Maruyama, K.; Mishima, Y.; Minagawa, K.; Motonaka, J. Anal. Chem.
2
002, 74, 3698-3703.
(106) Veal, J. M.; Rill, R. L. Biochemistry 1991, 30, 1132-1140.
(107) Carter, M. T.; Rodriguez, M.; Bard, A. J. J. Am. Chem. Soc. 1989, 111,
8901-8911.
(108) Barton, J. K.; Danishefsky, A.; Goldberg, J. J. Am. Chem. Soc. 1984,
106, 2172-2176.
104) Sari, M. A.; Battioni, J. P.; Dupre, D.; Mansuy, D.; Le Pecq, J. B.
Biochemistry 1990, 29, 4205-4215.
105) Jennette, K. W.; Lippard, S. J.; Vassiliades, G. A.; Bauer, W. R. Proc.
Natl. Acad. Sci. U.S.A. 1974, 71, 3839-3843.
6442 J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003

Photo-Bergman Cyclization of Cu Metalloenediynes
A R T I C L E S
(
Figure 3b, lane 4).64,110 Subsequent treatment of the photore-
action products with a reductant such as NaBH4 produces no
change in the quantity of either product (lane 5), thus ruling
out the 3′-phosphoglycaldehyde product that is derived from
C-3′ H-atom abstraction as the faster migrating species.1
Comparison of the migrations of the photoreaction products from
11
2
-
the well-characterized Fe(EDTA) /H2O2 thermal DNA deg-
radation reaction upon treatment with piperidine and NaBH4
(Figure 3b, lanes 6-8) shows similar behavior, indicating that
the faster migrating species is the 3′-phosphoglycolate that is
6
4,86
derived from C-4′ activation.
The absence of a slower-
moving band containing a (-3′-furanone-) suggests that the
mechanism does not involve abstraction of the 1′ H-atom.6
This result is supported by HPLC analyses of the photoreaction
mixtures, which do not show the presence of 5-MF, which is a
4,86
6
4,86,111
signature of 1′ H-atom abstraction.
Photolysis experiments
3
2
with 3′( P)-labeled oligonucleotide resulted in a single band
corresponding to the phosphate termini. A second band due to
the 5′-aldehyde product that is derived from activation at the 5′
6
4,86
position was not detected.
Combined, these results identify
the origin of DNA degradation by photolysis of 3 as C-4′
H-atom abstraction and not base-centered redox chemistry.
4. Discussion
The reactivities of 1-4 illustrate two novel processes in
fundamental enediyne photochemistry. The first is the ability
for copper coordination to promote photochemical Bergman
cyclization while the uncomplexed ligand is unreactive under
the same conditions. Coupled to this unusual result is the second
intriguing aspect of the observed photo-Bergman reactivity,
namely, the use of MLCT excitation as a conduit to cyclized
product formation.
Photolysis of 1 with λ g 395 nm in the presence of excess
H-atom donor results in no reaction on either the NMR tube or
preparative scales. Examination of the electronic absorption
spectrum of 1 (Figure 1, Table 2) reveals that 1 has virtually
no extinction at λ > 300 nm, and thus would not be photo-
electronically reactive at longer wavelengths. In contrast, both
the Cu(I) and Cu(II) complexes, 2 and 3, respectively, exhibit
good yields of Bergman-cyclized product (5, 6) upon photolysis
with λ g 395 nm. In either case, both bound enediyne lig-
ands are determined to have reacted (87% and 81% yields
from 2 and 3, respectively, on the basis of stoichiometric analy-
sis of the demetalated organic products extracted following
photolyses, as well as the characterization of the inorganic
photoproduct 5.
Figure 3. Polyacrylamide sequencing gels (20%) of the photoinduced (λ
)
368 nm, 8 h) cleavage of 5′-radiolabeled ds oligonucleotide d[5′-(ATG
GCG TAA TCA TGG TCA TAG CTG T)-3′]. Panel a shows the
concentration dependence (0-270 µM 3) of the cleavage of 10 µM
oligonucleotide in 10 mM sodium phosphate, pH 7.0. Lanes are identified
as follows: 1, Maxam-Gilbert G lane (cold [DNA] ) 5 µM); 2, 10 µM ds
oligonucleotide; 3, lane 2 + 90 µM 3; 4, lane 2 + 180 µM 3; 5, lane 2 +
2
70 µM 3. Panel b depicts the analysis of the oligonucleotide cleavage
products by chemical modification (piperidine, NaBH4) and a comparison
2-
to the C-4′ H-atom abstraction products from Fe(EDTA) /H2O2 treatment.
Lanes are identified as follows: 1, Maxam-Gilbert G lane (cold [DNA]
)
5 µM); 2, 10 µM ds oligonucleotide; 3, lane 2 + 270 µM 3; 4, lane 3 +
2-
piperidine treatment; 5, lane 3 + NaBH4 treatment; 6, lane 2 + Fe(EDTA)
/
H2O2; 7, lane 6 + piperidine treatment; 8, lane 6 + NaBH4 treatment; 9,
lane 2 + 30 µM CuCl2.
linear DNA likely is derived from multiple nicking. The
insensitivity of the DNA degradation patterns upon photolysis
of either 2 or 3 in the presence of inhibitors109 such as superoxide
dismutase (SOD) or hydroxy radical scavengers such as ethanol
suggests that DNA degradation does not result from residual
oxygen-derived radical species.
3
2
The determination that both enediyne ligands bound to copper
are Bergman-cyclized following photolysis leads to intriguing
questions about potential photochemical mechanisms by which
these photoreactions occur. Excitation of both 2 and 3 with λ
g 395 nm accesses MLCT absorption bands (MLCT for 2,
LMCT for 3) of π-character. Using Table 2, and the ∆E(redox)
approximation to estimate the charge-separated state energies
upon electron transfer between the metal-ligand HOMO and
3
.9. 5′( P)-Labeled Oligonucleotide DNA Photocleavage
Reactions. Anaerobic photolysis of 3, as a function of con-
centration (100-300 µM) in the presence of 5′-radiolabeled
d[5′-(ATG GCG TAA TCA TGG TCA TAG CTG T)-3′],
generates nonspecific cleavage sites distributed systematically
along the strand, with the extent of cleavage increasing with
increasing concentration of 3 (lanes 3-5, Figure 3a). Closer
inspection of the gel reveals a ladder of two bands, indicating
the presence of two distinct cleavage products. Treatment of
the photoreaction (270 µM 3, Figure 3b, lane 3) with piperidine
yields enhancement of the slower-migrating band, thereby
identifying this product as containing 5′-phosphate termini
1
12
•+
•-
LUMO, the Cu(I)-L (1.77 eV) and Cu(II)-L (1.95 eV)
species are thermodynamically accessible via the e3.14 eV
excitation energy. However, preparative-scale photoreactions of
(
110) Breiner, K. M.; Daugherty, M. A.; Oas, T. G.; Thorp, H. H. J. Am. Chem.
Soc. 1995, 117, 11673-11679.
(
109) Parand, A.; Royer, A. C.; Cantrell, T. L.; Weitzel, M.; Memon, N.; Vincent,
(111) Pogozelski, W. K.; Tullius, T. D. Chem. ReV. 1998, 98, 1089-1108.
(112) Vlcek, A. A. Electrochim. Acta 1968, 13, 1063-1078.
J. B.; Crowder, M. W. Inorg. Chim. Acta 1998, 268, 211-219.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6443

A R T I C L E S
Benites et al.
Scheme 5. Triplet-Sensitized Photolysis of Zn(bpod)2(CH3COO)2 (4)
2
and 3, as well as NMR-scale photolyses of 2, reveal no
detectable formation of redox-derived Cu(II) (from 2) or Cu(I)
from 3) photoproducts. In addition, cyclization of both enediyne
in the presence of 4 leads to formation of the trans-bpod
enediyne species (t-bpod, 1a) that is bound to Zn(II) (4a, Scheme
5) via bimolecular triplet-triplet energy transfer. Identification
of the product is confirmed by demetalation with EDTA and
spectroscopic comparison to authentic 1a (Scheme 1). Although
not isolated and characterized as a Zn(II) complex, coordination
of the product to the metal is determined by thin-layer
chromotography and NMR shifts relative to authentic, unbound
(
ligands bound to the same copper center rules out an irreversible,
intramolecular redox reaction. Photolysis of the Zn(II) complex
4
in the presence of excess Cu(NO3)2 or Cu(CH3CN)4PF6 also
does not lead to the generation of Bergman-cyclized photo-
products, indicating that cyclization is not derived from either
free copper or a trace-metal-induced redox reaction. Finally,
bulk electrolysis of 3 at 2.20 V failed to give clean conversion
to a single species and, by NMR analysis, yielded a complex
mixture devoid of the characteristic resonances of the Bergman-
cyclized product. Although our findings are consistent with the
redox lability of the enediyne moiety, taken together, these
results suggest that photo-Bergman cyclization of 2 and 3 does
not result from irreversible, metal-enediyne-localized photoredox
chemistry. Our observations are consistent with redox lability
of alkyne or enediyne compounds, but the inability of the
latter to generate Bergman-cyclized product upon electrochemi-
cal oxidation.
Insight into the photoreactivities of 2 and 3 can be gained
from the photochemistry of the Zn(II) analogue 4. Figure 1
3
1a. These results show that ligand-localized π-π* excitation
2
7
(ET ≈ 3.12 eV) of a closed-shell enediyne leads to cis-trans
isomerization about the ethylene unit, which has been described
2
7
114
both theoretically and experimentally.
Because neither redox-initiated Bergman cyclization nor
π-π*-derived products are detected upon the photolysis of 2
3
or 3, and because both photoreactions generate comparable
yields of cyclized product whereas the free ligand 1 is inert
under the same conditions, the photochemical pathways by
which these copper metalloenediynes perform Bergman cy-
clization can only be considered to be atypical. This is especially
8
7
43
4
3
10
apparent in light of the photosensitized photolysis of d ,
3
diamagnetic 4, which yields traditional π-π* product as
1
14
observed for simple organic enediynes. In this sense, the
Zn(II) center in 4 serves a structural role in drawing the alkyne
termini into closer proximity for cyclization, as evidenced by
the moderate DSC temperature for the complex (163 °C),
(Table 2) reveals that 4 has almost the identical electronic
spectrum as 1 (i.e., no CT transitions) and, thus, the same
inertness to direct electronic excitation at λ g 320, 345, or 395
nm. However, these results do not preclude photoreactivity of
3
4,35,39,116
relative to other metalloenediynes.
Energetically, this
4
(or 1). Many simple enediynes are known to be photochemi-
lowers the barrier to enediyne cyclization but, from an electronic
1
0
cally reactive either by direct photolysis with λ < 320
nm,
excited-state perspective, the diamagnetic d metal center is
essentially an innocent spectator and, thus, does not influence
the character of the excited electronic manifold.
21,22,33,113
22
or by an energy-transfer mechanism. Under these
conditions, three types of photoreaction pathways predominate,
namely (i) the formation of Bergman-cyclized product via
In contrast, the role of the copper centers in the Bergman
cyclization of 2 and 3 is markedly different. Although difficult
to verify experimentally, plausible mechanistic avenues can be
1
π-π* excitation, (ii) reduction of an alkyne unit to yield
3
22
dien-yne species upon π-π* population, or (iii) cis-trans
isomerization of the ethylene unit in the absence of an annulated
skeleton.1
The photochemical inertness of 4 to direct excitation for λ
g 320 nm provides an opportunity to use photoinduced triplet
energy transfer to activate the enediyne moiety, and to inves-
tigate the product obtained via ligand-centered triplet population.
proposed by considering established photochemical properties
14
95-102
of Cu(I) and Cu(II) bis- or dipyridyl complexes.
For the
Cu(I) complex 2, the d10 electronic configuration and redox
activity of the metal center leads to partial one-electron, MLCT
promotions from the metal dπ orbitals to the empty π* orbitals
of the pyridine ring,101,102 which generates the 3MLCT state102
115
Excitation of a triplet sensitizer (acetophenone; ET ) 3.21 eV)
rapidly via spin-orbit coupling. Free rotation of the pyridine
rings relative to the enediyne unit, combined with the lower-
lying π orbitals of enediyne motif (Table 1), permit delocal-
ization of the excited state onto the enediyne fragment through
(
(
(
113) Choy, N.; Blanco, B.; Wen, J.; Krishan, A.; Russell, K. C. Org. Lett.
2
000, 2, 3761-3764.
114) Kagan, J.; Wang, X.; Chen, X.; Lau, K. Y.; Batac, I. V.; Tuveson, R. W.;
Hudson, J. B. J. Photochem. Photobiol., B 1993, 21, 135-142.
115) Turro, N. J. Modern Molecular Photochemistry; The Benjamin/Cummings
Publishing Company, Inc.: Menlo Park, CA, 1978.
(116) Rawat, D. S.; Zaleski, J. M. Chem. Commun. 2000, 2493-2494.
6444 J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003

Photo-Bergman Cyclization of Cu Metalloenediynes
A R T I C L E S
potential influence of 2CT vs 1,3π-π* excited states, are
currently under investigation.
The triplet energy-transfer-derived isomerization of the bpod
ligand bound to Zn(II) in 4 is consistent with previous energy-
2
2,114
transfer studies of enediynes,
(
and density functional theory
DFT) calculations for nonbenzannulated structures.27 Excited
states derived from a molecular-orbital approach using ethylene
and acetylene fragments are thermodynamically ordered such
that the lowest-energy triplet state is localized on the double
2
7
bond and is of π-π* parentage. Population of this state by
bimolecular Dexter energy transfer would be expected to lead
to the isomerized trans-enediyne product, as observed for 4.
In addition to the fundamentally intriguing photochemistry
of these metalloenediynes, the abilities of 2 and 3 to effect DNA
6
5,110,125
degradation by H-atom abstraction
rather than by the
more-common base oxidation mechanism of inorganic agents
are rare. At low concentrations of 2 and 3, nicked DNA is
prevalent, whereas at higher concentrations (>100 µM), an
almost constant mixture of linear:nicked DNA exists that begins
to approach 1:1. The low quantum yield for photoreactivity of
Figure 4. Relative energies of copper and ligand frontier orbitals and
their involvement in the photoelectronic Bergman cyclization reactions of
-
5
2
(φ366 ) 4.2 (( 0.6) × 10 ) suggests that the length of
2
and 3.
photolysis time is directly a result of the quantum efficiency
and, as the moderate association constant of 3 (Km ) 2.6 ((
the π-system (Figure 4). Although not equivalent to activating
4
-1
0
.2) × 10 M ) suggests, is not a result of poor binding to the
the enediyne unit for reactivity via an electrochemical redox
4
3
117
DNA duplex.
reaction, this Dexter-type (exchange) energy-transfer mech-
The 32P-labeled oligonucleotide results provide two important
insights. First, the high-resolution gels confirm that the funda-
mental photochemical intermediate produced upon the photoly-
ses of 2 and 3 reacts via H-atom abstraction and not base
oxidation, which is consistent with the 1,4-phenyl diradical
proposed in the formation of the Bergman-cyclized products.
4
4,118,119
118
anism,
which is analogous to ligand-field or charge-
transfer-to-intraligand4
5,120,121
excited-state quenching, could
induce small but sufficient geometric perturbation to permit
Bergman cyclization. Thus, the photoreaction could be viewed
as occurring through an enediyne-centered state, despite initial
population of the formal Cu-pyridine CT manifold.122 Charge-
transfer-to-intraligand photochemical dynamics of this type are
2
+
Moreover, introduction of free Cu cations does not yield the
cleavage pattern observed upon photoactivation of 3, confirming
that trace metals are not responsible for these results. Although
no sequence specificity is observed for the photoreactivity of
uncommon but not unprecedented.1
20,121,123,124
On the basis of
the lack of a strongly radiative chromophore, and the absence
of prominent dipole-allowed absorption bands for 1 at λ > 300
nm, a unimolecular exchange energy-transfer mechanism is very
3
, this is not surprising in the absence of any sequence-selective
functionality. Second, the two-band pattern observed in Figure
, as well as subsequent identification of the faster-migrating
plausible.44,118,119
However, because of the mixed-spin character
3
of the CT excited states, and their ability to interact with ligand-
centered singlet/triplet states, a F o¨ rster singlet energy-transfer
species as a 3′-phosphoglycolate, indicates that copper metal-
loenediyne 3 is indeed cleaving DNA by ribose degradation
and not by base-centered, redox-induced strand scission. In the
metalloenediyne field, photoelectronic H-atom abstraction re-
activity has not been previously demonstrated. These results
place 2 and 3 in a category with other H-atom-abstracting metal-
1
23
scheme cannot be completely ruled out.
9
Photolysis of the d Cu(II) complex 3 with λ g 395 nm leads
to charge transfer from the pyridine π-orbitals to the empty hole
2
2
2
in the dx -y orbital ( LMCT), creating a partial π-hole on the
bound pyridine moiety. The modest enediyne oxidation potential
is reflective of high-energy π-orbitals containing considerable
enediyne character that could serve as a thermodynamically
accessible site for excited-state-partial-hole transfer onto the
enediyne unit (Figure 4). Such electronic modulation could
generate sufficient geometric distortion to promote Bergman
cyclization. The details of these distortions, as well as the
2
+
65
66
based agents such as Fe -bleomycin, Rh-phi complexes,
and [Pt(pop)4] (pop ) P2O4H2 ),110,125 of which there are a
4
-
2-
limited number of examples.
5. Conclusion
We have described the unique photoelectronic Bergman
cyclization of metalloenediynes of Cu(I) and Cu(II) and shown
that the intermediates perform double-stranded DNA cleavage.
The photocyclization mechanism and consequential deoxyribose
chemistry do not result from direct redox reactions with the
metal centers, but rather are derived from the photoelectronic
Bergman cyclization of the enediyne unit and subsequent C-4′
H-atom abstraction. The results show that metal-ligand charge-
transfer transitions can serve as conduits for population of
(
(
(
117) Dexter, D. L. J. Chem. Phys. 1953, 21, 836-850.
118) Endicott, J. F. Coord. Chem. ReV. 1985, 64, 293-310.
119) Ferraudi, G. F. Elements of Inorganic Photochemistry; John Wiley and
Sons: New York, 1988.
(
(
(
120) Yam, V. W.-W.; Lau, V. C.-Y.; Cheung, K.-K. Chem. Commun. 1995,
2
59-261.
121) Ford, P. C.; Wink, D.; Di Benedetto, J. Prog. Inorg. Chem. 1983, 30,
2
13-271.
122) Time-dependent DFT calculations to evaluate the progression of the
cyclization reaction along the excited-state reaction coordinate for 2 and
3
are ongoing and will be described in a future report.
(
123) Boyde, S.; Strouse, G. F.; Jones, J. W. E.; Meyer, T. J. J. Am. Chem.
Soc. 1989, 111, 7448-7454.
(125) Kalsbeck, W. A.; Gingell, D. M.; Malinsky, J. E.; Thorp, H. H. Inorg.
Chem. 1994, 33, 3313-3316.
(
124) Meyer, T. J. Pure Appl. Chem. 1990, 62, 1003-1009.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6445

A R T I C L E S
Benites et al.
enediyne-localized excited states that lead to Bergman cycliza-
tion and, thus, diradicals capable of performing H-atom abstrac-
tion from the deoxyribose backbone of DNA. Overall, the study
reveals a role that transition metals can play in influencing not
only thermal metalloenediyne cyclization, but also in the
photocyclization of metalloenediynes.
Acknowledgment. The generous support of the National
Institutes of Health (R01 GM62541-01A1) is gratefully ac-
knowledged. We thank Dr. Sibprasad Bhattacharyya for as-
sistance in ligand preparation.
JA020939F
6446 J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003