Photochemical generation and reversible cycloaromatization of a nine-membered ring cyclic enediyne

Article abstract of DOI:10.1021/ja8077076

Irradiation of the nine-membered ring enediyne precursor, which has one of its triple bonds masked as cyclopropenone, efficiently(φ = 0.34) generates the reactive 4,5-benzocyclonona-2,6-diynol. The latter rapidly equilibrates with the corresponding 1,4-didehydronaphthalene diradical and then undergoes rate- limiting hydrogen abstraction to produce the ultimate product of the Bergman cyclization, benz[f]indanol.

Full text of DOI:10.1021/ja8077076

Published on Web 12/03/2008
Photochemical Generation and Reversible Cycloaromatization
of a Nine-Membered Ring Cyclic Enediyne
Dinesh R. Pandithavidana, Andrei Poloukhtine, and Vladimir V. Popik*
Department of Chemistry, UniVersity of Georgia, Athens, Georgia 30677
Received September 29, 2008; E-mail: vpopik@uga.edu
Abstract: Irradiation of the nine-membered ring enediyne precursor, which has one of its triple bonds masked
as cyclopropenone, efficiently (Φ ) 0.34) generates the reactive 4,5-benzocyclonona-2,6-diynol. The latter
rapidly equilibrates with the corresponding 1,4-didehydronaphthalene diradical and then undergoes rate-
limiting hydrogen abstraction to produce the ultimate product of the Bergman cyclization, benz[f]indanol.
scission.1 While natural enediynes are very powerful dDNA-
cleaving machines, the lack of antitumor selectivity results in a
Introduction
Natural enediyne antibiotics possess an unusual (Z)-3-ene-
1,5-diyne fragment (hence the name), which is responsible for
their extreme cytotoxicity.1 Upon binding to a double-stranded
DNA molecule, this enediyne moiety is usually positioned
within the minor grove of the host. Reductive or nucleophilic
activation of the natural enediynes removes stereochemical
blocking device and the (Z)-3-ene-1,5-diyne fragment undergoes
Bergman2 cyclization to produce a p-benzyne diradical,1 which
might exist in equilibrium with the enediyne precursor.3 By it
is size, p-benzyne (ca. 3.8 Å) is perfectly suitable to perform a
double hydrogen abstraction from carbohydrate backbones of
opposite DNA strands (1′H to 1′H distance is about 5.78
Å).Theoretical studies and experimental data suggest that this
diradical has relatively low reactivity.4,5 Initial hydrogen
abstraction from one strand, however, converts p-benzyne into
a 10-100 times more reactive phenyl radical,4,6 which im-
mediately abstracts the second hydrogen from the opposite DNA
strand. This process leads to an oxidative double-strand dDNA
very high general toxicity, which hampers clinical applications
of natural enediyne antibiotics.7 Photochemical triggering of the
cycloaromatization reaction might help to alleviate this problem
by allowing for the spatial and temporal control of enediyne
reactivity.8 The direct irradiation of acyclic9 and cyclic10
enediynes, as well as the natural enediyne antibiotic Dynemicin
A,11 demonstrated that the Bergman cyclization can be induced
photochemically, albeit with low efficiency. The quantum yield
of the photochemical Bergman cyclization can be substantially
improved by adjusting the electronic properties of substituents12
and/or using different modes of excitation energy transfer, for
example MLCT.13 In addition, several caged enediynes have
been prepared which undergo conventional chemical activation
after the photochemical uncaging step.14
(7) (a) Galm, U.; Hager, M. H.; Lanen, S. G. V.; Ju, J.; Thorson, J. S.;
Shen, B. Chem. ReV. 2005, 105, 739. (b) Gredicak, M.; Jeric, I. Acta
Pharm. 2007, 57, 133. (c) Shen, B.; Nonaka, K. Curr. Med. Chem.
2003, 10, 2317.
(1) (a) For recent reviews see: Enediyne Antibiotics as Antitumor Agents;
Borders, D. B., Doyle, T. W., Eds.; Marcel Dekker: New York: 1995;
(b) Jones, G. B.; Fouad, F. S. Curr. Pharm. Design 2002, 8, 2415. (c)
Nicolaou, K. C.; Dai, W.-M. Angew. Chem., Int. Ed. Engl. 1991, 30,
1387. (d) Nicolaou, K. C.; Smith, A. L. Acc. Chem. Res. 1992, 25,
497. (e) Danishefsky, S. J.; Shair, M. D. J. Org. Chem. 1996, 61, 16.
(f) Shao, R. G. Curr. Mol. Pharm. 2008, 1, 50–60. (g) Hamann, P. R.;
Upeslacis, J.; Borders, D. B. Anticancer Agents Nat. Prod. 2005, 451.
(h) Smith, A. L.; Nicolaou, K. C. J. Med. Chem. 1996, 39, 2103. (i)
Thorson, J. S.; Ahlert, J.; Shepard, E.; Whitwam, R. E.; Onwueme,
K. C.; Sievers, E. L.; Ruppen, M. Curr. Pharm. Des. 2000, 6, 1841.
(2) Bergman, R. G. Acc. Chem. Res. 1973, 6, 25.
(8) (a) Kar, M.; Basak, A Chem. ReV. 2007, 107, 2861. (b) Jones, G. B.;
Fouad, F. S. Curr. Pharm. Design 2002, 8, 2415.
(9) (a) 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. (b)
Turro, N. J.; Evenzahav, A.; Nicolaou, K. C. Tetrahedron Lett. 1994,
35, 8089. (c) Kaneko, T.; Takahashi, M.; Hirama, M. Angew. Chem.,
Int. Ed. 1999, 38, 1267. (d) Plourde II, G.; El-Shafey, A.; Fouad, F.;
Purohit, A.; Jones, G. Bioorg. Med. Chem. Lett. 2002, 12, 2985. (e)
Falcone, D.; Li, J.; Kale, A.; Jones, G. B. Bioorg. Med. Chem. Lett.
2008, 18, 934–7. (f) Spence, J. D.; Hargrove, A. E.; Crampton, H. L.;
Thomas, D. W. Tetrahedron Lett. 2007, 48, 725. (g) Sud, D.;
Wigglesworth, T. J.; Branda, N. R. Angew. Chem., Int. Ed. 2007, 46,
8071.
(3) (a) Iida, K.; Hirama, M. J. Am. Chem. Soc. 1995, 117, 8875. (b)
Hirama, M. Pure Appl. Chem. 1997, 69, 525. (c) Hirama, M.; Akiyama,
K.; Das, P.; Mita, T.; Lear, M. J.; Iida, K.-i.; Sato, I.; Yoshimura, F.;
Usuki, T.; Tero-Kubota, S. Heterocycles 2006, 69, 83.
(10) (a) Choy, N.; Blanco, B.; Wen, J.; Krishan, A.; Russell, K. C. Org.
Lett. 2000, 2, 3761. (b) Funk, R. L.; Young, E. R. R.; Williams, R. M.;
Flanagan, M. F.; Cecil, T. L. J. Am. Chem. Soc. 1996, 118, 3291–2.
(11) Shiraki, T.; Sugiura, Y. Biochemistry 1990, 29, 9795.
(12) (a) Alabugin, I. V.; Manoharan, M. J. Phys. Chem. A 2003, 107, 3363.
(b) Alabugin, I. V.; Manoharan, M.; Kovalenko, S. V. Org. Lett. 2002,
4, 1119. (c) Clark, A. E.; Davidson, E. R.; Zaleski, J. M. J. Am. Chem.
Soc. 2001, 123, 2650.
(4) (a) Logan, C. F.; Chen, P. J. Am. Chem. Soc. 1996, 118, 2113. (b)
Schottelius, M. J.; Chen, P. J. Am. Chem. Soc. 1996, 118, 4896. (c)
Hoffner, J. H.; Schottelius, M. J.; Feichtinger, D.; Chen, P. J. Am.
Chem. Soc. 1998, 120, 376. (d) Sander, W. Acc. Chem. Res. 1999,
32, 669–676. (e) Chen, P. Angew. Chem., Int. Ed. Engl. 1996, 35,
1478.
(5) (a) Wenk, H. H.; Winkler, M.; Amegayibor, F. S.; Nash, J. J.; Lee,
A. S.; Thoen, J.; Petzold, C. J.; Kentta¨maa, H. I. J. Am. Chem. Soc.
2002, 124, 12066. (b) Sander, W. Angew. Chem., Int. Ed. 2003, 42,
502.
(13) (a) Benites, P. J.; Holmberg, R. C.; Rawat, D. S.; Kraft, B. J.; Klein,
L. J.; Peters, D. G.; Thorp, H. H.; Zaleski, J. M. J. Am. Chem. Soc.
2003, 125, 6434. (b) Kraft, B. J.; Coalter, N. L.; Nath, M.; Clark,
A. E.; Siedle, A. R.; Huffman, J. C.; Zaleski, J. M. Inorg. Chem. 2003,
42, 1663. (c) Bhattacharyya, S.; Pink, M.; Baik, M. H.; Zaleski, J. M.
Angew. Chem., Int. Ed. 2005, 44, 592.
(6) Roth, W. R.; Hopf, H.; Wasser, T.; Zimmermann, H.; Werner, C. Leib.
Ann. 1996, 1691.
9
10.1021/ja8077076 CCC: $40.75
2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 351–356 351

A R T I C L E S
Pandithavidana et al.
Scheme 1
Here we report the synthesis and photochemistry of cyclo-
propenone precursors 1, as well as theoretical and experimental
studies of the reactivity of nine-membered ring enediyne 2
(Scheme 1).
Results and Discussion
Synthesis of Cyclopropenone 1. A number of synthetic
methods are available for the preparation of the cyclopropenone
group, which is the key component of the photoactivatable
enediyne precursor 1.22 Few of them, however, are compatible
with a highly functionalized macrocycle 1, which forced us to
introduce cyclopropenone functionality on early stages of the
synthesis.23 The addition of difluorocarbene, which was gener-
ated by the thermolysis of trimethylsilyl fluorosulfonyldifluo-
roacetate (FO2SCF2CO2SiMe3, TFDA),24 to the acetylene 4,
produced 1,1-difluorocyclopropene. The latter was hydrolyzed
without isolation on wet silica gel to give cyclopropenone 5
(Scheme 2).23 Introducing cyclopropenone moiety early in a
multistep synthesis is difficult since this functionality is very
susceptible to a nucleophilic attack, which usually results in
ring opening.25 Complexation with Lewis acids, on the other
hand, produces relatively unreactive 2π-aromatic oxycyclopro-
penium cation. Masking cyclopropenone moiety as 2,2-dimethyl-
1,3-propanediyl acetal 6 allows us to avoid these complications
and broaden the range of reagents and reaction conditions that
can be employed for the preparation of target compound
(Scheme 2). It has to be noted, however, that cyclopropenone
acetals are extremely susceptible to acid hydrolysis. They have
life-times of only few milliseconds in an aqueous solution.26
Our group explores an alternative strategy for the phototrig-
gering of the cycloaromatization reaction: the in situ generation
of an activated enediyne system.15 Ideally, a photoswitchable
analogue of natural enediyne antibiotics should be stable in the
dark but undergo rapid and reversible cyclization to p-benzyne
after irradiation. We have recently shown that masking one of
the triple bonds in 10-membered cyclic enediynes with cyclo-
propenone moiety produces thermally stable enediyne precur-
sors. While photolysis of cyclopropenone precursors under
single-16 or two-photon excitation17 conditions results in efficient
generation of corresponding enediynes, the cycloaromatization
of the latter was not fast enough (τ40°C > 16 h) to achieve
temporal and spatial resolution of p-benzyne generation in
biological systems. To enhance the rate of Bergman cyclization
of photoswitchable enediynes, we decided to design a nine-
membered ring cyclopropenone-containing enediyne precursor
(Scheme 1). Highly strained nine-membered enediynes are
predicted to undergo very facile cycloaromatization under
ambient conditions.18 In fact, the only known monocyclic nine-
membered enediyne with measurable lifetime, 3-chloro-3-
cyclononene-1,5-diyne (τ40°C ≈ 8 min), is stabilized by chlorine
substituent in the vinylic position.19 Natural antibiotic C-1027
and its analogues, which possess a bicyclic nine-membered ring
enediyne core, are also extremely labile.3,20 To achieve revers-
ibility of the Bergman cyclization we decided to exploit the
fact that hydrogen abstraction often becomes a partially rate-
limiting step in the Bergman cyclization of benzannulated
enediynes.21
The second acetylenic substituent was introduced using
conventional Stille coupling condtions (7, Scheme 2). Simul-
taneous saponification of the acetate and removal of the TMS
group followed by iodination27 and Dess-Martin oxidation28
gave rise to the iodoaldehyde 10. Cyclization of the latter under
(21) (a) Zeidan, T. A.; Manoharan, M.; Alabugin, I. V. J. Org. Chem. 2006,
71, 954. (b) Stahl, F.; Moran, D.; Schleyer, P. R.; Prall, M.; Schreiner,
P. R. J. Org. Chem. 2002, 67, 1453. (c) Koseki, S.; Fujimura, Y.;
Hirama, M. J. Phys. Chem. A. 1999, 103, 7672. (d) Semmelhack,
M. F.; Neu, T.; Foubelo, F. J. Org. Chem. 1994, 59, 5038. (e) Boger,
D. L.; Zhou, J. J. Org. Chem. 1993, 58, 3018. (f) Semmelhack, M. F.;
Neu, T.; Foubelo, F. Tetrahedron Lett. 1992, 33, 3277.
(14) (a) 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, 1032. (b) Nicolaou, K. C.; Dai, W.-M. J. Am. Chem. Soc.
1992, 114, 8908. (c) Wender, P. A.; Zercher, C. K.; Beckham, S.;
Haubold, E.-M. J. Org. Chem. 1993, 58, 5867. (d) Wender, P. A.;
Beckham, S.; O’Leary, J. G. Synthesis 1994, 1278. (e) Basak, A.;
Bdour, H. M.; Shain, J. C.; Mandal, S.; Rudra, K. R.; Nag, S. Bioorg.
Med. Chem. Lett. 2000, 10, 1321.
(22) (a) Breslow, R.; Posner, J.; Krebs, A. J. Am. Chem. Soc. 1963, 85,
234. (b) Ko¨brich, G.; Flory, K.; Fischer, R. H. Chem. Ber. 1966, 99,
1793. (c) Chikos, J. C.; Patton, E.; West, R. J. Org. Chem. 1974, 39,
1647. (d) Wadsworth, D. H.; Donatelli, B. A. Synthesis 1981, 285.
(e) Musigmann, K.; Mayr, H. Tetrahedron Lett. 1987, 28, 4517. (f)
Weinder, C. H.; Wadsworth, D. H.; Knop, C. S.; Oyefesso, A. I.; Hafer,
B. H.; Hartman, R. J.; Mehlenbacher, R. C.; Hogan, S. C. J. Org.
Chem. 1994, 59, 4319. (g) Chiang, Y.; Grant, A. S.; Kresge, A. J.;
Paine, S. W. J. Am. Chem. Soc. 1996, 118, 4366. (h) Gleiter, R.;
Merger, M.; Altereuther, A.; Irngartinger, R. J. Org. Chem. 1996, 61,
1946. (i) Netland, K. A.; Gundersen, L.-L.; Rise, F. Synth. Commun.
2000, 30, 1767. (j) Poloukhtine, A.; Popik, V. V. J. Org. Chem. 2003,
68, 7833.
(15) Poloukhtine, A.; Karpov, G.; Popik, V. V. Curr. Trends Med. Chem.
2008, 8, 460.
(16) (a) Poloukhtine, A.; Popik, V. V. J. Org. Chem. 2005, 70, 1297. (b)
Poloukhtine, A.; Popik, V. V. Chem. Comm. 2005, 617.
(17) Poloukhtine, A.; Popik, V. V. J. Org. Chem. 2006, 71, 7417.
(18) (a) Nicolaou, K. C.; Ogawa, Y.; Zuccarello, G.; Schweiger, E. J.;
Kumazawa, T. J. Am. Chem. Soc. 1988, 110, 4866. (b) Snyder, J. P.
J. Am. Chem. Soc. 1990, 112, 5367. (c) Magnus, P.; Fortt, S.; Pitterna,
T.; Snyder, J. P. J. Am. Chem. Soc. 1990, 112, 4986. (d) Jones, G. B.;
Warner, P. M. J. Am. Chem. Soc. 2001, 123, 2134. (e) Gaffney, S. M.;
Capitani, J. F.; Castaldo, L.; Mitra, A. Int. J. Qant. Chem. 2003, 95,
706. (f) Chen, W.-C.; Zou, J.-W.; Yu, C.-H. J. Org. Chem. 2003, 68,
3663. (g) Alabugin, I. V.; Manoharan, M. J. Phys. Chem. A 2003,
107, 3363.
(23) See the Supporting Information file for details.
(24) (a) Cheng, Z.-L.; Xiao, J.-C.; Liu, C.; Chen, Q.-Y. Eur. J. Org. Chem.
2006, 5581. (b) Cheng, Z.-L.; Chen, Q.-Y. Chin. J. Chem. 2006, 24,
1219.
(25) (a) Potts, K. T.; Baum, J. S Chem. ReV. 1974, 74, 189. (b) Halton, B.;
Banwell, M. G. In The Chemistry of Cyclopropyl Group; Rappoport,
Z., Ed.; Wiley: New York, 1987, 1300; (c) Ciabattoni, J.; Nathan,
E. C. J. Am. Chem. Soc. 1969, 91, 4766. (d) Dehmlow, E. V.; Neuhaus,
R.; Schell, H. G. Chem. Ber. 1988, 121, 569. (e) Dehmlow, S. S.;
Dehmlow, E. V. Z. Naturforsch. B 1975, 30b, 404. (f) Dehmlow, E. V.
Leib. Ann. 1969, 729, 64.
(19) Jones, G. B.; Plourde, G. W. Org. Lett. 2000, 2, 1757.
(20) (a) Yoshida, K.; Minami, Y.; Azuma, R.; Saeki, M.; Otani, T.
Tetrahedron Lett. 1993, 34, 2637. (b) Xu, Y. J.; Zhen, Y. S.; Goldberg,
I. H. Biochemistry 1994, 33, 5947. (c) Iida, K.; Hirama, M. J.Am.
Chem. Soc. 1995, 117, 8875. (d) Usuki, T.; Mita, T.; Lear, M. J.;
Das, P.; Yoshimura, F.; Inoue, M.; Hirama, M.; Akiyama, K.; Tero-
Kubota, S. Angew. Chem., Int. Ed. 2004, 43, 5249. (e) Inoue, M.;
Ohashi, I.; Kawaguchi, T.; Hirama, M. Angew. Chem., Int. Ed. 2008,
47, 1777.
(26) McClelland, R. A.; Ahmad, M. J. Am. Chem. Soc. 1978, 100, 7027.
(27) Crevisy, C.; Beau, J.-M. Tetrahedron Lett. 1991, 32, 3171.
(28) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4156.
9
352 J. AM. CHEM. SOC. VOL. 131, NO. 1, 2009

Nine-Membered Ring Cyclic Enediyne
A R T I C L E S
Scheme 2a
a Reagents and conditions: (a) (i) NaF, TFDA, diglyme, 120 °C, 2 h; (ii) wet silica gel, hexanes; (b) Et3OBF4, CH2Cl2, r.t., 1 h; then neopentylglycol,
Et3N, CH2Cl2, r.t., 12 h, 51% (after two steps); (c) trimethyl[(tributylstannyl)ethynyl]silane, Pd(PPh3)4, toluene, 90 °C, 2 h, 76%; (d) K2CO3, MeOH, 2 h,
88%; (e) I2, morpholine, benzene, 78%; (f) Dess-Martin periodinane, CH2Cl2, r.t., 3 h, 84%; (g) CrCl2, NiCl2, THF, 0 °C, 3 h, 36%; (h) Amberlyst, acetone,
r.t., 1 h, 77%.
Figure 1. UV spectra of ca. 2 × 10-5 M solutions of the cyclopropenone
Figure 2. Temperature rate profile for the Bergman cyclization of the
enediyne 2 generated in photolysis of 1 in 2-propanol. The line shown was
drawn using parameters obtained by least-squares fitting of Eyring equation.
The inset shows growth in absorbance at 227 nm after irradiation of 1 at
25 °C in neat 2-propanol. The line shown in the inset was drawn using
parameters obtained by least-squares fitting of single exponential equation.
1 (dash-dotted line), enediyne 2 (solid line), and benz[f]indan-1-ol (3,dashed
line) in 2-propanol.
Nozaki-Hiyama-Kishi conditions,29 followed by mild removal
of an acetal protection, produced the target nine-membered ring
cyclopropenone 1 (Scheme 2).23
Kinetics of the Bergman Cyclization of 4,5-Benzocyclonona-
2,6-diynol (2). The accurate rate measurements of cycloaroma-
tization of the enediyne 2 were conducted by UV spectroscopy
following the growth of the characteristic 227 nm band of
benz[f]indan-1-ol (3, Figure 1) in the photolysate at various
temperatures and solvent compositions. As shown on the inset
in Figure 2, the reaction follows first-order equation well. The
cyclization rates of the enediyne 2 were measured in 2-propanol
solutions in the temperature range from 15 to 55 °C with 0.1 °C
accuracy. The data so obtained are summarized in Table S123 and
are displayed as the temperature rate profile in Figure 2.
The temperature dependence of the cyclization rate is
surprisingly weak: every 10 K of temperature rise produces only
about 2-fold rate increase (Figure 2). This observation suggests
that the energy barrier (∆G‡) for the cyclization reaction grows
with the temperature. In fact, least-squares fitting of the data to
the Eyring equation30 gives the following activation parameters:
∆H‡) 13.63 ( 0.22 kcal M-1 and ∆S‡ ) -30.66 ( 0.61 cal
M-1 K-1. The substantial negative entropy of activation indicates
that in the transition state for the rate-determining step degrees
Photochemical Generation and Reaction of Nine-Membered
Ring Enediyne 2. The UV spectrum of cyclopropenone 1 shows
two close lying absorbance bands at 231 nm (log ε ) 4.5) and
238 nm (log ε ) 4.6, dash-dotted line in Figure 1). Photolysis
of 1 in 2-propanol with 254 nm light resulted in efficient
decarbonylation (Φ254nm ) 0.34 ( 0.03)23 and the formation of
the 4,5-benzocyclonona-2,6-diynol (2, Scheme 1). The latter
undergoes spontaneous but relatively slow Bergman cyclization
with a lifetime of ca. 2 h at 25 °C in 2-propanol. This process
is surprisingly clean for a radical reaction producing quantitative
yields of the ultimate product, benz[f]indan-1-ol (3, Scheme 1).
The structural assignment of the product isolated from the
photolysate of 1 was supported by the direct synthesis of
benzindanol 3.23
(29) (a) Takai, K.; Kuroda, T.; Nakatsukasa, S.; Oshima, K.; Nozaki, H.
Tetrahedron Lett. 1985, 26, 5585. (b) Aicher, T. D.; Kishi, Y.
Tetrahedron Lett. 1987, 28, 3463. (c) Brandstetter, T.; Maier, M. E.
Tetrahedron 1994, 50, 1435. (d) Harwig, C. W.; Py, S.; Fallis, A. G.
J. Org. Chem. 1997, 62, 7902. (e) Dai, W.-M.; Wu, A.; Hamaguchi,
W. Tetrahedron Lett. 2001, 42, 4211.
9
J. AM. CHEM. SOC. VOL. 131, NO. 1, 2009 353

A R T I C L E S
Pandithavidana et al.
Scheme 3
of freedom in the system are significantly reduced. The
activation energy for the cyclization of the enediyne 2 to
diradical 12 (Scheme 3) should have a low entropy component
and is not consistent with the experimental value of ∆S‡. The
next step in the process, i.e., hydrogen abstraction, is a
bimolecular reaction requiring proper orientation of the diradical
12 and a hydrogen donor. Such reactions are usually character-
ized by pronounced negative entropy of activation. In other
words, the temperature dependence of the rate of Bergman
cyclization of enediyne 2 suggests that hydrogen abstraction
rather than formation of 12 is the rate limiting step of the
reaction.
positions. 2-Propanol is a good hydrogen donor with an R-C-H
33
bond dissociation energy of 89 kcal mol-1
.
This value is
similar to the C-H bond strength in tetrahydrofuran (92 kcal
mol-1),34 a model compound for DNA deoxyribose. Water, on
the other hand, is a poor hydrogen donor (D0(H-OH) ) 118 kcal
mol-1).35 The rate of the cyclization of the enediyne 2 to
benz[f]indan-1-ol (3) in 2-propanol-water mixtures was found
to be linearly proportional to the concentration of alcohol up to
neat 2-propanol solutions. (Figure 3, Table S223). This observa-
tion indicates that equilibration between enediyne 2 and
p-benzyne 12 is much faster than hydrogen abstraction even at
the highest concentration of 2-propanol (Scheme 4).
Further support for the rate-limiting hydrogen abstraction
comes from the value of the solvent isotope effect on the rate
of cyclization. When enediyne 2 was generated in 2-propanol-
d8 at 25 °C, the observed rate of its conversion to 3 was
substantially lower than in 2-propanol (kD ) (2.669 ( 0.013)
× 10-5 s-1 vs kH ) (1.497 ( 0.019) × 10-4 s-1). The
pronounced primary kinetic isotope effect, kH/kD ) 5.61, clearly
indicates that formation of a carbon-hydrogen bond happens
on the rate-limiting step. This value, however, agrees well with
both rate-limiting hydrogen transfer, as shown in Scheme 3,
and rate-limiting proton transfer, which has been shown to
induce cyclization of some enediyne compounds.31 The use of
2-propanol-d1 (i-PrOD) as a hydrogen donor provides a good
method for a discrimination of the two pathways. Carbon-
centered radicals are expected to abstract protium from 2-pro-
panol-d1 by cleaving the weakest C-H bond of a methine group,
while proton (deuteron) transfer should involve a more acidic
OH (OD) group. The rate of the cycloaromatization of 2 in
2-propanol-d1, kOBS ) (1.383 ( 0.022) × 10-4 s-1, was
essentially the same as in an all-protium solvent. This observa-
tion clearly shows that the hydrogen, but not proton, transfer is
involved in the rate-limiting step of the Bergman cyclization
of 2. Proton transfer-initiated reactions are also expected to be
catalyzed by acid, while conventional Bergman cyclization
shows no acid catalysis.32 Enediyne 2 generated in the presence
of 0.025 M hydrochloric acid or absence of acid in 2-propanol-
water mixture (4:1) at 25 °C undergoes conversion to 3 with
the same rate (kOBS ) (1.03 ( 0.024) × 10-4 vs (9.87 ( 0.82)
× 10-5 s-1, respectively).
Figure 3. Observed rates of the Bergman cyclization of the enediyne 2 at
various concentrations of 2-propanol in water at 25.0 ( 0.1 °C.
Scheme 4
The kinetics of the Bergman cyclization can be described by
eq 1, where k1 is the rate constant for the cycloaromatization of
2 to 12, k-1 represents the rate of cycloreversion, kH is a second
order rate constant for hydrogen abstraction, and [H] is a
concentration of a hydrogen donor.
As was mentioned in the introduction, the rate of the Bergman
cyclization of enediyne compounds often depends on the
concentration and the nature of hydrogen donor.3 This effect is
especially prominent in case of benzannulated enediynes.6,21 We
have measured the rate of the cyclization of 4,5-benzocyclonona-
2,6-diynol (2) in 2-propanol-water mixtures of various com-
k1 × kH × [H]
kOBS
)
(1)
(2)
k-1 + kH × [H]
kOBS ) Keq × kH × [H]
When the rate of cycloreversion is much faster than the rate
of hydrogen abstraction (i.e., k-1 . kH × [H]), the observed
rate is linearly proportional to the hydrogen donor concentration,
(30) The transmission coefficient in the Eyring equation was set to unity.
The nonlinear least square fitting calculation were conducted using
Origin 7.5 by OriginLab. The correlation coefficient for this fit was
R2 ) 0.999.
(31) (a) Karpov, G.; Kuzmin, A.; Popik, V. V. J. Am. Chem. Soc. 2008,
130, 11771. (b) Poloukhtine, A.; Popik, V. V. J. Am. Chem. Soc. 2007,
129, 12062.
(33) Kanabus-Kaminska, J. M.; Gilbert, B. C.; Griller, D J. Am. Chem.
Soc. 1989, 111, 3311.
(34) Golden, D. M.; Benson, S. W Chem. ReV. 1969, 69, 125.
(35) Ruscic, B.; Feller, D.; Dixon, D. A.; Peterson, K. A.; Harding, L. B.;
Asher, R. L.; Wagner, A. F. J. Phys. Chem. A 2001, 105, 1.
(32) Perrin, C. L.; Rodgers, B. L.; O’Connor, J. M. J. Am. Chem. Soc.
2007, 129, 4795.
9
354 J. AM. CHEM. SOC. VOL. 131, NO. 1, 2009

Nine-Membered Ring Cyclic Enediyne
A R T I C L E S
as shown in eq 2, where Keq ) k1/k-1. The rate of hydrogen
abstraction from methanol by 1,4-dehydronaphthaline is 4 ×
105 M-1 s-1.6 Since 2-propanol is ∼2.7 times more reactive
than methanol as a hydrogen donor in radical reactions,36 we
can estimate the rate of hydrogen abstraction by the diradical
12 from 2-propanol as 1.1 × 106 M-1 s-1. Using this value and
eq 2 we can evaluate the enediyne 2 T p-benzyne 12
equilibrium constant at Keq ≈ 10-11. The free energy difference
between 2 and 12 should be ∆Go ≈ 15 kcal mol-1. At high
hydrogen donor concentration the hydrogen abstraction step
often becomes faster than the cycloreversion (i.e., kH × [H] .
k-1) and observed rate levels off at kOBS ) k1. Such rate
saturation has been reported for the Bergman cyclization of
acyclic6,21a and cyclic21d benzannulated enediynes in the pres-
ence of 1,4-cyclohexadiene or methanol. The rate of the
cycloaromatization of 2 shows no signs of leveling off even in
neat 2-propanol (Figure 3). This observation allows us to
conclude that cycloreversion reaction (k-1, Scheme 4) is at least
an order of magnitude faster than the reaction of 12 in neat
2-propanol. The lower limit for the cycloreversion reaction can
be therefore estimated at 1.4 × 108 s-1 and for the cycloaro-
matization reaction (k1, Scheme 4) at 1.4 × 10-3 s-1. The upper
limit for the activation barrier for the Bergman cyclization of 2
Figure 4. Schematic potential energy profile for the Bergman cyclization
of 4,5-benzocyclonona-2,6-diynol (2) in the presence of 2-propanol.
Table 1. ZPVE-Corrected Relative Electronic Energies Calculated
at the B3LYP/6-31++G(d,p) Level
∆E (kcal /mol)
2
0.00
22.30
2‡
at 25 °C is 21-22 kcal mol-1
.
12 singlet
12 triplet
12a‡
12b‡
13a
13b
3
10.74
15.64
12.41
12.65
-7.97
-7.68
-30.06
It is also interesting to note that no additional products were
observed in the reaction of enediyne 2 at low 2-propanol
concentration. In wholly aqueous solutions UV spectrum of 2
remains virtually unchanged for four days at ambient temper-
atures. In other words, an enediyne 2-diradical 12 equilibrated
system has a relatively long lifetime in the absence of good
hydrogen donors.
produces a moderate ring strain of ∼11 kcal mol-1 according
to B3LYP/6-31++G(d,p) calculations.23 This value is similar
to the ring strain of the isopropyl 2-oxo-5,6-benzocyclodeca-
3,7-diyne-1-carboxylate, a benzannulated 10-membered ring
enediyne which undergoes Bergman cyclization with a 5 h
lifetime in 2-propanol at 36 °C.39 The strain of the nine-
membered ring is apparently reflected in a reduced endother-
micity of enediyne 2 cycloaromatization. Thus, 12 lies only 10.7
kcal mol-1 above 4,5-benzocyclonona-2,6-diynol (2, Table 1),
while the energy gap between the acyclic o-diethynylbenzene
Theoretical Analysis. Theoretical analysis of the cycloaro-
matization reaction of enediyne 2 was conducted using a hybrid
DFT B3LYP method with 6-31++G(d,p) orbital set. All
geometries were optimized using either the restricted (2 and
3), unrestricted (13a,b), or broken-spin unrestricted (12 and
transition states 2‡, 12‡a,b) calculations (Figure 4). The BS-
UB3LYP approach has been shown to provide good description
of the energetics of Bergman reaction at reasonable computa-
tional costs.37 The optimized geometries, as well as the details
of theoretical procedures, are provided in the Supporting
Information file. The relative energies of species involved in
the Bergman cyclization of enediyne 2 are listed in the Table 1
and presented graphically in the Figure 4. Radical centers in
4,9-didehydrobenz[f]indanol (12) are different and there are two
transition states for hydrogen transfer: 12a‡ leading to 9-dehy-
drobenz[f]indanol radical (13a, this pathway is shown in Figure
4); and 12b‡ leading to 4-dehydrobenz[f]indanol radical (13b).
The energies of isomeric transition states 12a‡ and 12b‡, as well
as radicals 13a and 13b, are very close (Table 1).23
and 1,4-didehydronaphthaline is about 7 kcal mol-1 higher
6
(∆H0exp) 17.8 kcal mol-1
,
∆E(BCCD(T)/cc-pVDZ) ) 17.6 kcal
mol-1 40). Frequency calculations indicate that entropy change
accompanying the cyclization of 2 to 12 is quite small and free
energy change at 25 °C (∆G0298°) is 11 kcal mol-1. This value
is 5 kcal mol-1 lower than an estimate based on experimental
data (vide supra). We put more trust in experimental estimate
because BS-DFT method is known for overestimation of the
diradical stability. Thus, various BS-DFT calculations produce
energy gap between o-diethynylbenzene and 1,4-didehydronaph-
According to BS-UB3LYP/6-31++G(d,p) calculations, the
4,9-didehydrobenz[f]indanol (12) is a ground-state open-shell
singlet. The triplet state of this diradical is 4.9 kcal mol-1 higher
in energy. Enhanced stability of singlet diradicals over triplet
state is associated with their reduced reactivity in comparison
to monoradicals.4,38 Incorporation of the enediyne fragment into
a nine-membered ring in 4,5-benzocyclonona-2,6-diynol (2)
thaline in the range of 11-14.4 kcal mol-1 38b,40 vs experimental
6
value of 17.8 kcal mol-1
.
The activation energy predicted by BS-UB3LYP/6-
31++G(d,p) calculations (∆E‡ ) 22.3 kcal mol-1, Table 1)
agrees with the upper estimate for the rate of the diradical 12
formation (k1 < 10-3 s-1, vide supra). Surprisingly, ring strain
(38) (a) Clark, A. E.; Davidson, E. R. J. Am. Chem. Soc. 2001, 123, 10691.
(b) Pickard, F. C.; Shepherd, R. L.; Gillis, A. E.; Dunn, M. E.; Feldgus,
S.; Kirschner, K. N.; Shields, G. C.; Manoharan, M.; Alabugin, I. V.
J. Phys. Chem. A 2006, 110, 2517.
(36) Janzen, E. G.; Nutter, D. E., Jr.; Evans, C. A. J. Phys. Chem. 1975,
79, 1983.
(37) (a) Grafenstein, J.; Hjerpe, A. M.; Kraka, E.; Cremer, D. J. Phys. Chem.
A 2000, 104, 1748. (b) Zeidan, T. A.; Kovalenko, S. V.; Manoharan,
M.; Alabugin, I. V. J. Org. Chem. 2006, 71, 962. (c) Schreiner, P. R.;
Navarro-Vasquez, A.; Prall, M. Acc. Chem. Res. 2005, 38, 29.
(39) Karpov, G.; Popik, V. V. J. Am. Chem. Soc. 2007, 129, 3792.
(40) Prall, M.; Wittkopp, A.; Schreiner, P. R. J. Phys. Chem. A 2001, 105,
9265.
9
J. AM. CHEM. SOC. VOL. 131, NO. 1, 2009 355

A R T I C L E S
Pandithavidana et al.
destabilization of the starting enediyne 2 has limited effect on
the height of the activation barrier. Thus, for 10-membered ring
analogue, 3,4-benzocyclodeca-1,5-diyne, experimental estimate
puts activation energy at 25 kcal mol-1.21c For the parent acyclic
enediyne, o-diethynyl-benzene, experimental enthalpy of activa-
6
tion from the NO-trapping experiments is 25.2 kcal mol-1
,
which agrees with BS-DFT value of 24.6 kcal mol-1 but high
accuracy BCCD(T)/cc-pVDZ method put the barrier at 29.2 kcal
mol-1 40 This phenomenon can be apparently explained by the
.
fact that Bergman cyclization has an “electronically late”
transition state. The geometry of the transition state for
cycloaromatization of the parent (Z)-hex-3-ene-1,5-diyne is
“80% product-like”, while electronically it is “70% reactant-
like”.41 This means that shortening of transannular distance
results in a steady growth of a repulsive interaction between
filled in-plain acetylenic π-orbitals, while stabilizing σ-bond
formation and aromatization of the out-of-plain π-system lags
behind.12a Benzannulation of the enediyne system results in
further reduction of the aromatic stabilization energy for the
transition state and the diradical.21b Subsequent abstraction of
hydrogen by the diradical 12 from 2-propanol to give mono-
radicals 13a,b is a highly exothermic process (∆E° ) -18.7
kcal mol-1, Table 1), which makes this step practically irrevers-
ible. The barrier for the initial hydrogen abstraction is mostly
of entropic nature (∆E‡ ) 1.7 kcal mol-1, Table 1). Radicals
13a,b should be even more reactive and the second hydrogen
abstraction from 2-propanol has even higher exothermicity, ∆E°
Figure 5. Cleavage of ꢀX174 plasmid DNA by the photogenerated
enediyne 2. Lane 1: DNA alone incubated for 16 h at 25 °C; lane 2 and 4:
DNA incubated in the dark with cyclopropenone precursors 1, 1 and 5 mM;
lane 3 and 5: DNA is incubated with the irradiated solution of 1, 1 and 5
mM.
structural modification that would allow us to shift the absor-
bance of a cyclopropenone precursor to 350-400 nm.
) -22.1 kcal mol-1
.
Conclusions
DNA Cleavage Experiments. An evaluation of the photoge-
nerated nine-membered ring enediyne 2 nuclease activity was
carried out using supercoiled plasmid DNA cleavage assays.
Three forms of this DNA: native (RF I), circular relaxed (RF
II, produced by single-strand cleavage), and linear (RF III,
formed by scission of both strand in close proximity) are readily
separated by the agarose gel electrophoresis.23 To produce
reactive enediyne 2, 1 and 5 mM aqueous solutions of
cyclopropenone 1 were irradiated with low-pressure mercury
lamp until ca. 90% conversion. A solution of ꢀX174 supercoiled
circular DNA (10 ng/µL) in TE buffer was added to photolysate
and incubated for 16 h at 25 °C.
Incubation of the DNA with cyclopropenone precursor 1
(lanes 2 and 4, Figure 5) does not induce any detectable DNA
cleavage. The photogenerated enediyne 2, on the other hand,
induces substantial single strand cleavage of ꢀX174 DNA (RF
II), while linearized form (RF III) becomes prominent only at
higher (5 mM) concentration of the cleaving agent (lanes 3 and
5, Figure 5). Integration of fluorescence of bands on the gel
shown in Figure 5, allowed us to evaluate the relative abundance
of the native, circular, and linearized forms of ꢀX174 DNA.
Thus, incubation of the latter with the 1 mM of enediyne 2
produces 45% of single strand cleavage (RF II) and less than
5% of the double strand cleavage (RF III, lane 3, Figure 5). At
5 mM concentration of 2 DNA-cleavage efficiency grows to
67% and 10%, respectively (lane 5, Figure 5). The relatively
high concentrations of 2, which are required to achieve double-
strand DNA photoscission indicate that 2 has low affinity to a
dDNA molecule. In order to improve the photonuclease activity
of the enediyne 2, we are working on the design and synthesis
of enediyne 2 analogues containing a dDNA minor-groove
binding or a dDNA-intercalating moiety. We also explore
Irradiation of the thermally stable cyclopropenone precursor
1 produces the first known example of a benzannulated nine-
membered ring enediyne, 4,5-benzocyclonona-2,6-diynol (2).
The enediyne 2 exist in a rapid equilibrium with a corresponding
p-benzyne analogue, 4,9-didehydrobenz[f]indanol (12), even at
25 °C (Keq ≈ 10-11). In the presence of hydrogen donors, 12
undergoes double hydrogen abstraction quantitatively producing
benz[f]indanol (3). Kinetic data indicates that the latter process
is a rate-limiting step for the Bergman cyclization of the
enediyne 2. While 2 is rapidly consumed in neat 2-propanol
(τ ) 111 min), the lifetime in aqueous solution is much longer,
τH2O > 96 h. In other words, if 2 is generated in a cellular
environment, it will survive in active from until it encounters
an appropriate hydrogen donor (DNA, thiol, protein, etc.).
Enediyne 2 induces single and double strand cleavage of dDNA
molecules, albeit with moderate efficiency. Cyclopropenone
precursor 1, on the other hand, does not produce DNA damage.
Therefore, cyclopropenone 1 equipped with a DNA-binding
moiety is a promising candidate for the development of an
efficient in vivo photonuclease. Synthetic efforts in this direction
are underway in our laboratory.
Acknowledgment. Authors thank the National Science Founda-
tion (CHE-0449478), Georgia Cancer Coalition, and donors of the
ACS Petroleum Research Fund (434444-AC4) for the support of
this project.
Supporting Information Available: Experimental details;
synthetic procedures for the preparation of compounds 1 and
3; photochemical generation of 2; details of kinetic measure-
ments and DNA-cleaving experiments; Cartesian coordinates
of DFT-optimized structures. This material is available free of
charge via the Internet at http://pubs.acs.org.
(41) Galbraith, J. M.; Schreiner, P. R.; Harris, N.; Wei, W.; Wittkopp, A.;
Shaik, S. Chemistry 2000, 6, 1446.
JA8077076
9
356 J. AM. CHEM. SOC. VOL. 131, NO. 1, 2009