Goal. Our group has published a series of papers intended to develop an understanding of how the electronic character of the double bond of an enediyne can influence the rate and activation energy of Bergman cyclization.1-7 The goal of the work proposed here is to use classic physical organic techniques to further clarify the issues of charge dependence and the interactions of aromatic solvents on Bergman cyclization, specifically in cases where the double bond is part of a heterocycle.
To reach the desired goal, the following specific aims are proposed:
1) Ten-membered ring enediynes (1-3) will be prepared using our previously established methodologies.
2) Each enediyne will be tested for thermal Bergman cyclization and its activation energy measured.
3) The following parameters will be studied for their effects on cyclization:
a) Charge dependence will be investigated on Bergman cyclization in organic and aqueous media using covalent and pH control (1•R+).
b) Aromatic interactions will be analyzed by performing cyclizations in carefully chosen aromatic solvents (2, 3).
Background. The Bergman cyclization is generally considered as the thermally allowed electronic rearrangement of a Z-3-ene-1,5-diyne (4) to a p-benzyne diradical (5, Scheme 1).8 The diradical is quenched by radical sources to afford new aromatic rings. It has been well established that both strain energy and c-d distance are important parameters for Bergman cyclization. However, only a handful of reports outside of our lab have demonstrated a direct electronic influence on cyclization. The most well understood effect is benzannulation.9 When the double bond of an enediyne is part of an aromatic ring retro-Bergman cyclization (k-1) becomes competitive with the trapping reaction (k2) and slower cyclization rates may be observed.
Our own work has demonstrated the following electronic effects
with acyclic enediynes:
1) Heteroatoms in the arene containing the double bond lower the
activation energy of Bergman cyclization;7 2) The solvent changes
the rate of cyclization;5 3) The tautomeric state of the ring
containing the double bond can alter the rate of cyclization;3,6
4) A linear correlation exists between the Hammett sm constant
and the rate of cyclization for
4-substitutited-1,2-diethynylbenzenes;2 5) Flavin-enediynes
exhibit electrochemical behavior consistent with theory.4
Experimental Design & Methods
Synthesis and Thermal Cyclization. The synthesis of 1-3 will be based on methodologies we have used to prepare the acyclic enediyne precursors of 2 and 3,7 and similar cyclic enediynes.1 Although synthetically more challenging, incorporating the enediyne into a ten-membered membered ring generally facilitates cyclization and provides better yields, translating into more reliable kinetic data. Bulk cyclizations will be performed for each enediyne to help optimize kinetic conditions, determine the reaction yield, and characterize the cyclization product(s). The authentic product will also be used to confirm the cyclization product from kinetic runs by co-injection.
Kinetics and Activation Energies. The appearance of product or disappearance of starting material for kinetic studies will be measured as a function of time using our HPLC assay.1-7 Since all of the enediynes proposed are aromatic we will perform [CHD]-dependent kinetics to test for benzannelation according to eq. 1 (CHD = 1,4-cyclohexadiene; radical trap). By plotting kobs vs. [CHD]-1 at a series of temperatures, k1 can be determined at each temperature and the activation energy taken from the Arrhenius relationship.10 If benzannelation is not involved, k1 = kobs and is rate limiting. The activation energy can be found directly according to the Arrhenius relationship as we have published.3
kobs = k1-[(k-1/k2)·(kobs/[CHD])] (eq. 1)
a. Charge Dependent Cyclization
One potential biological application of these studies would be in the development of enediyne anticancer agents triggered by the intracellular pH as a function of cell type. These experiments will provide evidence to show that introduction of a positive charge into a heterocycle will accelerate Bergman cyclization if an enediyne double bond is also part of the same heterocycle.
a1. pH control. The pH dependent cyclization of pyridine 3•H+ will be investigated in water (R = H, acylamino or -ammonium, as appropriate for solubility). When the alcohol is substituted, these compounds should have sufficient water solubility for our purposes. Target 3 can be prepared from the same precursor we used in the synthesis of 2,3-diethynylpyridine.1 As the pKa's of acetylenic pyridines have not been reported, the pKa of 3•H+ will be determined by potentiometric titration. Since a compound's pKa is a function of temperature, the pKa determined at room temperature will not be the same as that at the cyclization temperature. To compensate for this effect, the pKa change vs. temperature information for a series of pyridines will be extrapolated for 3•H+. This way the protonation state of the enediynes during cyclization can be accurately estimated. Kinetics will be measured in aqueous solutions at pH's above and below the pKa of 3•H+. We will adjust the temperature as necessary to get data on a reasonable time scale. Since the possibility exists that the propargylic substituents may undergo acid catalyzed elimination it may also be necessary to prepare the ketone.
a2. Covalent modification. Experiments with 3•CH3+ will provide additional evidence of activation of Bergman cyclization by decreasing the electron density within the heterocycle. Furthermore, and biologically significant, the results should demonstrate that cyclization is faster in organic solvents than in water, in direct accord with our previous results.5 This would suggest that enediynes might be activated by intercalation. Many important DNA intercalators are quaternized heterocyclic salts.11 Unlike 3•H+, compound 3•CH3+, will have a “permanent” positive charge on the pyridine nitrogen. Target 3•CH3+ can be synthesized as the iodide salt by stirring 3 (R = H) in CH3I and isolating the precipitate. Should the alcohol not methylate during the process, the free alcohol will be used. The remainder of the compounds can be prepared by ion exchange. Cyclization rates of 3•CH3+ will be measured in organic (e.g. C6H5Cl) and aqueous systems using appropriate counter ions (e.g. PF6– or BF4– for organic solvents and Cl– or I– for H2O). Different counter ions will be used to ensure that they are not involved in the rate. Activation energies will be measured as previously discussed where possible.
b. Aromatic effects: Polarity and p-Stacking.
It has been postulated that Bergman cyclization has a very late transition state,12 suggesting that the transition state will be stabilized by non-polar or p-stacking solvents. Aromatic solvents can potentially affect Bergman cyclization by p-stacking of the solvent with the developing aromatic ring as in the stabilization of flavin radicals.13 Alternatively, the formation of p-complexes between the radical intermediates and the solvent are possible. The stability of these complexes correlates with each solvent's Hammett sm value.14 We have also observed differences in the cyclization rates for arenediynes in various solvents.5 A correlation between solvent dielectric constant and cyclization rate was found with more polar solvents giving slower reactions. These experiments will examine the effects of aromatic solvents on Bergman cyclization with respect to polarity and p-stacking.
To address both p-stacking and polarity, the cyclization of 2 and 3 will be compared in benzene, nitrobenzene, and N,N-dimethylaniline. In each case 100 equiv. CHD will be used as the radical trap. Both 2 and 3 may be influenced differently because of their number of aromatic rings. Based on our previous work we would expect rates to be related to each solvent's dielectric constant5 [C6H6 (e = 2.84) > C6H5N(CH3)2 (e = 4.90) > C6H5NO2 (e = 20.8); T = 20 ºC].15 However, based on the sm values a different order of reactivity would be expected, C6H5NO2 (sm = 0.71) > C6H6 (sm = 0.00) > C6H5N(CH3)2 (sm = -0.16).16 Since the dielectrics predict one trend, and the sm values another, it will be important to determine which effect is more important. While these experiments will not rule out the other possible arene-arene interactions, 17 they will allow us to determine if there are bulk effects from different aromatic solvents other than simple polarity.
References: (1) Choy, N.; Blanco, B.; Wen, J.; Krishan, A.; Russell, K.C., Org. Lett. 2000, 2, 3761-3764; (2) Choy, N.; Kim, C.-S.; Ballestero, C.; Artigas, L.; Diez, C.; Lichtenberger, F.; Shapiro, J.; Russell, K.C., Tetrahedron Lett. 2000, 41, 6955-6958; (3) Kim, C.-S.; Diez, C.; Russell, K.C., Chem. Eur. J. 2000, 6, 1555-1558; (4) Choy, N.; Russell, K.; Alvarez, J.; Fider, A., Tetrahedron Lett. 2000, 41, 1515-1518; (5) Kim, C.-S.; Russell, K.C., Tetrahedron Lett. 1999, 40, 3835-3838; (6) Choy, N.; Russell, K.C., Heterocycles, 1999, 51, 13-16; (7) Kim, C.-S.; Russell, K.C., J. Org. Chem. 1998, 63, 8229-8234; (8) Bergman, R.G., Acc. Chem. Res. 1973, 6, 25-31; (9) Roth, W.; Hopf, H.; Wasser, T.; Zimmerman, H.; Werner, C., Liebigs Ann. 1996, 1691-1695; (10) Huisgen, R., Angew. Chem. Int. Ed. Engl. 1970, 9, 751-762. (11) Staerk, D.; Hamed, A.A.; Pedersen, E.B.; Jacobsen, J.P., Bioconjugate Chem. 1997, 8, 869-87; (12) Lindh, R.; Lee, T.J.; Bernhardsson, A.; Persson, B.J.; Karlstrom, G., J. Am. Chem. Soc. 1995, 117, 7186-7194; (13) Breinlinger, E.C.; Rotello, V.M., J. Am. Chem. Soc. 1997, 119, 1165-1166; (14) Walling, C., J. Org. Chem. 1988, 53, 305-308; (15) Marsh, K.N. in Recommended Reference Materials for the Realization of Physiochemical Properties; Blackwell Scientific Publication: Oxford, 1987; (16) Hansch, C.; Leo, A.; Taft, R.W., Chem. Rev. 1991, 91, 165-195; (17) Jorgensen, W.L.; Severence, D.L., J. Am. Chem. Soc. 1990, 112, 4768-4774.ÿoe ÌCP