Cyclization kinetics and biological evaluation of an anticancer 1,2-dialkynylimidazole

Article abstract of DOI:10.1039/b925261d

1,2-Dialkynylimidazoles have been reported to undergo thermal cyclization/rearrangement to diradical and carbene intermediates. Optimization of the synthesis of the 1,2-dialkynylimidazole 3 has provided sufficient material for kinetic and biological studies. The 1,2-dialkynylimidazole 3 is cytotoxic against a wide range of cancer cells and induces apoptosis in A549 cells. Experimentally-determined kinetics of the thermolysis of 3 (E_a = 30.0 kcal mol^-1) are in excellent agreement with DFT calculations of the cyclization/rearrangement to diradical and cyclopentapyrazine carbene intermediates (E_a = 29.7 kcal mol^-1). Commensurate with the relatively high barrier for cyclization of 3, no evidence for cleavage of supercoiled DNA under physiological conditions was found; however, under aqueous conditions at 70°C 3 formed a covalent adduct with a model peptide. These studies indicate that if cyclization of 3 is involved in its anticancer activity, the cyclization must be facilitated, perhaps through initial protein binding, which could lead to covalent protein modification.

Full text of DOI:10.1039/b925261d

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Cyclization kinetics and biological evaluation of an anticancer
1,2-dialkynylimidazole†
Christophe Laroche,a Jing Li,a Cristina Gonzales,b Wendi M. Davidb and Sean M. Kerwin*a
Received 2nd December 2009, Accepted 3rd February 2010
First published as an Advance Article on the web 15th February 2010
DOI: 10.1039/b925261d
1,2-Dialkynylimidazoles have been reported to undergo
thermal cyclization/rearrangement to diradical and carbene
intermediates. Optimization of the synthesis of the 1,2-
dialkynylimidazole 3 has provided sufficient material for
kinetic and biological studies. The 1,2-dialkynylimidazole
3 is cytotoxic against a wide range of cancer cells and
induces apoptosis in A549 cells. Experimentally-determined
kinetics of the thermolysis of 3 (Ea = 30.0 kcal mol-1) are
in excellent agreement with DFT calculations of the cy-
clization/rearrangement to diradical and cyclopentapyrazine
carbene intermediates (Ea = 29.7 kcal mol-1). Commensurate
with the relatively high barrier for cyclization of 3, no evi-
dence for cleavage of supercoiled DNA under physiological
conditions was found; however, under aqueous conditions
at 70 ◦C 3 formed a covalent adduct with a model peptide.
These studies indicate that if cyclization of 3 is involved in its
anticancer activity, the cyclization must be facilitated, perhaps
through initial protein binding, which could lead to covalent
protein modification.
Scheme 1 Thermolysis of 1,2-dialkynylimidazoles.
The enediyne structural moiety is present in several naturally
occurring anticancer antibiotics.1 These antitumor enediyne nat-
ural products exhibit their cytotoxicity as a result of their
DNA cleavage ability based on the generation of 1,4-benzenoid
diradical intermediates arising from the Bergman cyclization
of the enediyne core.2 Numerous efforts to design cytotoxic
DNA-cleaving enediynes3 or compounds that undergo alternative
diradical-generating cyclizations4 have been reported. Certain
designed enediynes have been found to target proteins in addition
to, or instead of, DNA.5
More recently, a 1,2-dialkynylimidazole covalent inhibitor of
p38a kinase was described, although the roles of either the
diradical or carbene intermediates in this inhibition were not
addressed.8 Based on the potential of dialkynylimidazoles to
behave like anticancer enediynes, computational and experimen-
tal studies of the structure, cyclization, DNA- and protein-
interactions, and cancer cell cytotoxicity of a representative
1,2-dialkynylimidazole, 1-ethynyl-2-((4-methoxy-phenyl)ethynyl)-
1H-imidazole (3) (Scheme 2) were undertaken.
Efforts to design alternative diradical-generating cycliza-
tions led to the preparation and study of 1,2-dialkny-
limidazoles (Scheme 1).6,7 Initial reports of the synthesis of
1,2-dialkynylimidazoles as well as their thermal rearrangement
to products derived from trapping of didehydroimidaza[1,2-
a]pyridine diradical intermediates or cyclopentylpyrazine carbene
intermediates have been reported (Scheme 1).6
Scheme 2 Synthetic route to 1,2-dialkynylimidazole 3. (a) n-BuLi, THF,
-78 ◦C 15 min., then I2, -78 ◦C 30 min (83%); (b) Pd(PPh3)4, CuI, RCCH,
Et3N, 50 ◦C (100%); (c) TBAF, THF, -78 ◦C 30 min (90%).
aDivision of Medicinal Chemistry, College of Pharmacy, The University
of Texas at Austin, Austin, Texas, 78712, USA. E-mail: skerwin@
mail.utexas.edu
bDepartment of Chemistry and Biochemistry, Texas State University, San
Marcos, TX, 78666, USA
The key step in 1,2-dialkynylimidazole synthesis involves the
formation of the 1-alkynylimidazole moiety. In the previously
reported synthesis of compound 3, an inefficient coupling between
imidazolate and a hypervalent iodine reagent was employed.6c
Recently, the synthesis of 1-alkynylimidazoles by coupling reac-
tion between imidazoles and bromoalkynes in the presence of
† Electronic supplementary information (ESI) available: General exper-
imental conditions, reaction optimization, compound characterization,
X-ray and DFT co-ordinates, and experimental details for biological and
kinetic studies. CCDC reference number 756647. For ESI and crystallo-
graphic data in CIF or other electronic format see DOI: 10.1039/b925261d
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 1535–1539 | 1535
©

View Article Online
catalytic copper salts and 2-acetylcyclohexanone (AcC) ligand9
or PEG40010 has been reported. The two-step conversion of 1-
alkynylimidazoles to 1,2-dialkynylimidazoles via 2-position io-
dination followed by Sonogashira coupling has recently been
described.11 However, while this approach affords both superior
yields and wider scope compared to previous routes, it suffers from
the requirement for an excess of base and bromoalkyne partner
in the ligand-mediated copper-catalyzed coupling reaction. Given
the need to produce sufficient quantities of 3 to study both its
chemistry and biology, the coupling reaction conditions were
optimized with the objective of decreasing the quantities of
bromoalkyne and base necessary for the preparation of the
alkynylimidazole 1 (Scheme 2). Thus imidazole and 1-bromo-
triisopropylsilylacetylene were coupled under various conditions
(see Electronic Supporting Information). These studies reveal
that when employing 5 mole% of CuI and a 20 mole% of
AcC ligand, the effect of decreasing the quantity of Cs2CO3
base from two to one equivalent was modestly unfavorable;
whereas, decreasing the quantity of bromoalkyne had a much more
pronounced deleterious effect on yield. However, the reaction
exhibited better yields when lower quantities of copper/ligand
were used. This observation corroborates the hypothesis of the
formation of catalytically inactive aggregates of the copper ion
and the imidazole.9 Reducing the quantity of copper disfavors the
presence of such polymorphic complexes. Thus, when employing
just 0.5 mole% of CuI and 2 mole% of ligand, with 1.1 equiv. of
both the bromoalkyne and Cs2CO3 the yield of 1 was 79%, which
is comparable to the yield employing the originally described pro-
cedure, but only requires a slight excess of bromoalkyne and base.
The 1-alkynylimidazole 1 was converted into the dialkynylim-
idazole 3 via iodination to 2 followed by Sonogashira coupling
and protodesilylation (Scheme 2). This sequence was carried out
on 5 mmole of alkynylimidazole 1 to afford a 59% overall yield
of 3. Notably, the Sonogashira coupling (Scheme 2, step b) was
realized in quantitative yield using only 1 mole% of Pd(PPh3)4
and 2 mole% of CuI. The structure of 3 was confirmed by X-ray
crystallography (Fig. 1). The structure reveals a relatively short
Fig. 2 Unit cell packing diagram for 3. The view is approximately
down the a axis. Dashed lines illustrate close C–H ◊ ◊ ◊ O contacts between
adjacent molecules. The geometry of this C14-H14◦◊ ◊ ◊ O1 interaction is:
˚
˚
C-O 3.231(3) A, H ◊ ◊ ◊ O 2.25(2) A, C–H ◊ ◊ ◊ O 171(2) .
lines with GI50 values from 10-8 to 10-6 M and a mean GI50
of 3 mM. A COMPARE analysis16 of the spectrum of activity
of 3 against that of standard anticancer drugs revealed no
similarities (correlation co-efficient < 0.5); however, within the set
of compounds selected by the NCI for in vivo testing, the activity
of 3 correlated most strongly (P = 0.727 to 0.602) to a series of
quinoxaline 1,4-dioxides. These specific quinoxaline 1,4-dioxides
have been shown to generate hydroxyl radicals and cleave DNA
under aerobic conditions through redox cycling involving the
NADPH/cytochrome P450 reductase system.17 Under anaerobic
conditions, the one-electron reduction of quinoxaline 1,4-dioxides
is accompanied by radical fragmentation to afford stoichiometric
hydroxyl radicals.18
In order to gain more insight into the mechanism of action of
the dialkynylimidazole 3, studies were conducted in the human
non-small cell lung cancer cell line A549,19 which is a well-
characterized standard human alveolar epithelial cell line. A549
cells were incubated with various concentrations of 3 and the
extent of apoptosis was determined by flow cytometry (Fig. 3).
1
1
12
˚
(N)Csp –C sp bond length of 1.167 A. This, combined with
exocyclic imidazo◦le C2–N1–alkyne (126.0◦) and imidazole N1–
˚
C2–alkynyl (120.8 ) bond angles leads to a relatively short 4.129 A
c,d distance13 between the two distal Csp1 atoms C4 and C15.14 In
addition, the crystallographic packing, shown in Fig. 2, reveals
intermolecular C–H ◊ ◊ ◊ O interactions between the ether oxygen
and the terminal Csp1-H of adjacent molecules.
Fig. 3 Apoptosis of A549 cancer cells in the presence of 3.
Fig. 1 View of 3 showing the atom labelling scheme. Displacement
ellipsoids are scaled to the 50% probability level.
After treating A549 cells with 5 mM of compound 3 for 24 h, the
proportion of apoptotic cells increased to 63%, 25-times higher
than the proportion of apoptotic cells in the absence of 3. Even at
concentrations as low as 1.25 mM, compound 3 caused a significant
increase in the proportion of apoptotic cells.
The cytotoxicity15 of 3 has been evaluated by the National
Cancer Institute against 60 human cancer cell lines. The di-
alkynylimidazole 3 displayed activity against a range of cell
1536 | Org. Biomol. Chem., 2010, 8, 1535–1539
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
The spectrum of anticancer activity of 3 analyzed by COM-
PARE correlates with the known DNA cleaving quinoxaline
1,4-dioxides. Agents of a wide variety of mechanistic actions,
including DNA-damaging agents, are able to induce apoptosis
in A549 cells.20 In order to probe more deeply the possible role of
DNA damaging intermediates arising from cyclization of 3 in the
anticancer activity of this compound, kinetic studies were under-
taken. While the cyclization of 3 is reported to produce diradical
and carbene intermediates (Fig. 1),6 the kinetics of this process
have not been previously investigated, either computationally or
experimentally.
the rate-limiting step for formation of carbene-trapping products
is the initial aza-Bergman cyclization.
Thermolysis of 1,2-dialkynylimidazole
3
in neat 1,4-
cyclohexadiene yielded a mixture of cyclopentapyrazine trapping
products (Scheme 3).6d Solutions of 3 and internal standard 5,6-
benzoquinoline in 1,4-cyclohexadiene were heated at 80, 90, 100,
120 or 150 ◦C, and the rate of disappearance of 3 was monitored
by LC-MS (Fig. 5).
Although the Bergman and related diradical-generating cy-
clizations present significant challenges to computational meth-
ods, seminal work by Kraka and Cremer has shown that
broken spin-symmetry unrestricted density-functional theory
(BS-UDFT) methods employing hybrid functionals can perform
as well as much more expensive methods on these systems.21 This
BS-UDFT method was applied to the computer design of new
enediyne warheads,22 although no experimental data for these
are available. BS-UDFT calculations were carried out on the
cyclization of 3 using the sum-corrected M1 method of Sherer
and co-workers.23 Briefly, these B3LYP/6-31G(d,p)//B3LYP/6-
31G(d,p) calculations were performed using the broken-spin-
symmetry, unrestricted method for the diradical and the transition
states to and from this species, and the energies of the resulting
singlets were corrected for triplet contamination using the sum
method employing triplet energies of the BS-UB3LYP singlet
geometries. These calculations indicate that the singlet diradical 5
resulting from aza-Bergman cyclization of 3 (Fig. 4) should rapidly
rearrange to the cyclic cumulene 7; the transition state 6 for this
process lies just 2.4 kcal mol-1 above the singlet diradical 5, and
the S-T gap for 5 is 9.3 kcal mol-1. There is a relatively low barrier
for rearrangement of 7 to the cyclopentacarbene 9. Despite an
unfavorable equilibrium between the carbene 9 and the cumulene
7, the reactivity of the former24 leads to its efficient trapping. Thus,
Scheme 3 Products obtained from the thermolysis of 1,2-dialkynylimi-
dazole 3 in neat 1,4-cyclohexadiene.
The kinetic experiments showed first-order disappearance of
starting material with half-lives ranging from 13.5 h to 7 min. An
Arrhenius plot of the first-order rate constants provides activation
parameters Ea = 30.0 kcal mol-1; A = 6 ¥ 1012 (Fig. 6). The
◦
calculated activation energy (Ea = DH‡ - RT) at 115 C for the
aza-Bergman cyclization of 3 to the diradical 5 from the DFT
calculations is 29.7 kcal mol-1, in excellent agreement with the
experimental value.
When incubated with supercoiled DNA at 37 ◦C for 24 h,
dialkynylimidazole 3 showed no evidence of DNA cleavage
ability. This is to be expected, since the extrapolated half-life for
cyclization of 3 at 37 ◦C is longer than 5 years. Thus, it does not
Fig. 4 Relative electronic energies + ZPE calculated at the BS-(U)B3LYP/6-31G(d,p) level for the thermal rearrangement of 3.
This journal is The Royal Society of Chemistry 2010 Org. Biomol. Chem., 2010, 8, 1535–1539 | 1537
©

View Article Online
Fig. 7 MALDI-TOF analysis of bradykinin incubated at 70 ◦C for
24 hours with 3.
Fig. 5 Data obtained from the thermolysis of 3 at 90 ◦C.
cyclization in the cytotoxicity of this compound. While 3 does not
cleave DNA at physiological conditions, at higher temperatures 3
can covalently adduct the peptide bradykinin. Thus, one intriguing
possibility for the origin of the anticancer activity of 3 involves
facilitated cyclization of 3 by protein binding, leading to the
formation of protein-interactive intermediates such as diradical
5 or carbene 9 under physiological conditions. Further work to
identify the specific molecular target(s) of 3 is underway.
Acknowledgements
This work was supported by grants from the Robert Welch Foun-
dation (F-1298) and The Texas Higher-Education Co-ordinating
Board. Prof. Shawn Bratton is acknowledged for assistance in
the flow cytometry experiments. Prof. Phillip Magnus is gratefully
acknowledged for his assistance to CL.
Fig. 6 Arrhenius plot from thermolysis of 3.
appear likely that the cytotoxicity or the apoptosis of cancer cells
due to 3 is a result of DNA strand scission. Perhaps the similarity
in the cytotoxicity profile of 3 with that of quinoxaline 1,4-dioxides
is due to the ability of both compounds to produce other types
of DNA damage, or to interact with key proteins involved in cell
proliferation and apoptosis.25
In light of the recent report of a 1,2-dialkynylimidazole covalent
p38a inhibitor,8 the ability of reactive intermediates derived from
3 to covalently interact with a model peptide under aqueous
conditions was explored. Incubation of 3 (1 mM) with the peptide
bradykinin [RPPGFSPFR] (1 mM) at 70 ◦C in 10 mM Tris buffer
(pH 7.0) containing 2.5% DMSO (to solubilize 3) for 24 h followed
by MALDI-MS analysis demonstrates the formation of a small
amount of an adduct corresponding to addition of 3 (or a reactive
intermediate derived from it, see Scheme 3) to the peptide (Fig. 7).
Thus, in water, at temperatures where its rate of cyclization is
significant, 3 is capable of covalently attacking peptides.
In summary, an improved procedure for the synthesis of
1-alkynylimidazole derivatives has been employed to prepare
sufficient quantities of 3 for biological evaluation. The 1,2-
dialkynylimidazole 3 is cytotoxic against a wide range of cancer
cells and induces apoptosis in A549 cells. DFT calculations are
in excellent agreement with experimentally determined kinetic
analysis of the thermolysis of 3; however, the rate of conversion of 3
to the diradical 5 or cyclopentapyrazine carbene 9 at physiological
temperature is insignificant, ruling out any role for unassisted
Notes and references
1 (a) A. L. Smith and K. C. Nicolaou, J. Med. Chem., 1996, 39, 2103–
2117; (b) S. G. Van Lanen and B. Shen, Curr. Top. Med. Chem., 2008,
8, 448–459.
2 R. R. Jones and R. G. Bergman, J. Am. Chem. Soc., 1972, 94, 660–661.
3 (a) A. Basak, S. K. Roy, B. Roy and A. Basak, Curr. Top. Med. Chem.,
2008, 8, 487–504; (b) I. Alabugin, B. Breiner and M. Manoharan,
Advances in Physical Organic Chemistry, 2008, Vol. 42, pp 1–33; (c) A.
Polukhtine, G. Karpov and V. V. Popik, Curr. Top. Med. Chem., 2008,
8, 460–469; (d) D. S. Rawat and J. M. Zaleski, Synlett, 2004, 393–
421; (e) M. Klein, T. Walenzyk and B. Konig, Collect. Czech. Chem.
Commun., 2004, 69, 945–965; (f) G. B. Jones and F. S. Fouad, Curr.
Pharm. Des., 2002, 8, 2415–2440.
4 (a) W. M. David and S. M. Kerwin, J. Am. Chem. Soc., 1997, 119,
1464–1465; (b) L. Feng, D. Kumar and S. M. Kerwin, J. Org. Chem.,
2003, 68, 2234–2242; (c) L. Feng, D. Kumar, D. Birney and S. M.
Kerwin, Org. Lett., 2004, 6, 2059–2062; (d) W. Tuntiwechapikul, W. D.
David, D. Kumar, M. Salazar and S. M. Kerwin, Biochemistry, 2002,
41, 5283–5290; (e) B. Tuesuwan and S. M. Kerwin, Biochemistry, 2006,
45, 7265–7276.
5 (a) F. S. Fouad, J. M. Wright, G. Plourde, A. D. Purohit, J. K. Wyatt,
A. El-Shafey, G. Hynd, C. F. Crasto, Y. Q. Lin and G. B. Jones, J. Org.
Chem., 2005, 70, 9789–9797; (b) Y. M. Du, C. J. Creighton, Z. Y. Yan,
D. A. Gauthier, J. P. Dahl, B. Zhao, S. M. Belkowski and A. B. Reitz,
Bioorg. Med. Chem., 2005, 13, 5936–5948.
6 (a) A. K. Nadipuram and S. M. Kerwin, Org. Lett., 2002, 4, 4543–4546;
(b) A. K. Nadipuram and S. M. Kerwin, Synlett, 2004, 8, 1404–1408;
(c) A. K. Nadipuram and S. M. Kerwin, Tetrahedron Lett., 2006, 47,
353–356; (d) A. K. Nadipuram and S. M. Kerwin, Tetrahedron, 2006,
62, 3798–3808.
7 For studies of 4,5-dialkynylimidazoles, see: (a) Z. Zhao, Y. Peng, N. K.
Dalley, J. F. Cannon and M. A. Peterson, Tetrahedron Lett., 2004, 45,
1538 | Org. Biomol. Chem., 2010, 8, 1535–1539
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
3621–3624; (b) Z. G. Zhao, J. G. Peacock, D. A. Gubler and M. A.
Peterson, Tetrahedron Lett., 2005, 46, 1373–1375.
8 J. Li, T. S. Kaoud, C. Laroche, K. N. Dalby and S. M. Kerwin, Bioorg.
Med. Chem. Lett., 2009, 19, 6293–6297.
9 C. Laroche, J. Li, M. W. Freyer and S. M. Kerwin, J. Org. Chem., 2008,
73, 6462–6465.
17 B. Solano, V. Junnotula, A. Marin, R. Villar, A. Burguete, E. Vicente,
S. Perez-Silanes, I. Aldana, A. Monge, S. Dutta, U. Sarkar and K. S.
Gates, J. Med. Chem., 2007, 50, 5485–5492.
18 B. Gonley, G. Chowdhury, J. Bhansali, J. S. Daniels and K. Gates,
Bioorg. Med. Chem., 2001, 9, 2395–2401.
19 D. J. Giard, S. A. Aaronson, G. J. Todaro, P. Arnstein, J. H. Kersey, H.
Dosik and W. P. Parks, J. Natl. Cancer Inst., 1973, 51, 1417–1423.
20 (a) K. Erdelyi, P. Bai, I. Kovacs, E. Szabo, G. Mocsar, A. Kakuk, C.
Szabo, P. Gergely and L. Virag, FASEB J., 2009, 23, 3553–3563; (b) H.
Zhu and N. Gooderham, Toxicol. Sci., 2006, 91, 132–139; (c) J. Ren, N.
Agata, D. S. Chen, Y. Q. Li, W. H. Yu, L. Huang, D. Raina, W. Chen,
S. Kharbanda and D. Kufe, Cancer Cell, 2004, 5, 163–175.
21 J. Gra¨fenstein, E. Kraka, M. Filatov and D. Cremer, Int. J. Mol. Sci.,
2002, 3, 360–394.
22 E. Kraka and D. Cremer, D., J. Am. Chem. Soc., 2000, 122, 8245–8264.
23 E. C. Sherer, K. N. Kirschner, F. C. Pickard IV, C. Rein, S. Feldgus and
G. C. Shields, J. Phys. Chem. B, 2008, 112, 16917–16934.
24 Related cyclopentadienylidene carbenes have negative predictedDH‡
for C–H insertion, see for example: J. L. Mieusset and U. H. Brinker,
J. Org. Chem., 2008, 73, 1553–1558.
10 G. A. Burley, D. L. Davies, G. A. Griffith, M. Le and K. Singh, J. Org.
Chem., 2010, 75, 980–983.
11 C. Laroche and S. M. Kerwin, Tetrahedron Lett., 2009, 50, 5194–5197.
12 The corresponding bond lengths from X-ray crystallographic structures
˚
˚
of other 1-alkynylimidazoles are: 1.190 A (ref. 8) and 1.199 A
(ref. 10).
13 K. C. Nicolaou, G. Zuccarello, C. Riemer, V. A. Estevez and W. M.
Dai, W. M., J. Am. Chem. Soc., 1992, 114, 7360–7371.
˚
14 The corresponding c,d distance in a 4,5-dialkynylimidazole is 4.62 A,
see reference 7a.
15 M. C. Alley, D. A. Scudiero, A. Monks, M. L. Hursey, M. J. Czerwinski,
D. L. Fine, B. J. Abbott, J. G. Mayo, R. H. Shoemaker and M. R. Boyd,
Cancer Res., 1988, 48, 589–601.
16 K. D. Paull, R. H. Shoemaker, L. Hodes, A. Monks, D. A. Scudiero,
L. Rubinstein, J. Plowman and M. J. Boyd, J. Natl. Cancer Inst., 1989,
81, 1088–1092.
25 H. U. Gali-Muhtasib, M. Diab-Assaf and M. J. Haddadin, Cancer
Chemother. Pharmacol., 2005, 55, 369–378.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 1535–1539 | 1539
©