We have discovered a novel class of (E)-3-acyloxy-4-(arylmethylidene)cyclodeca-1,5-diynes which exhibit promising enediyne-like DNA cleavage and cytotoxic activities. LCMS analysis of the incubation mixture (pH 8.5, 37℃) confirmed formation of 10-membered ring enediyne presumably via an allylic cation and suggested that the 1,4-benzenoid diradical might be one of the active species for DNA damage and cytotoxicity.

The naturally occurring enediynes1 represent a novel class of antitumor antibiotics that feature a (Z)-hex-3-ene-1,5-diyne moiety constrained in a 9-or 10-membered ring. Coupled with other structural domains responsible for drug activation and delivery, the enediyne antitumor antibiotics present challenging targets for chemical synthesis.1a-c Stimulated by the intriguing mechanism of action and promising biological activity, extensive chemical and biological investigations on enediynes have been carried out during the past decade.1 It is known that cycloaromatization2 of enediynes such as 2 will give the diradical species 3, which can damage DNA through hydrogen atom abstraction from the deoxyribose residue (Scheme 1).1a,d,e,3 This event is regarded as the origin of the biological activity of enediynes. However, the extreme lability of simple 9- or 10-membered ring enediynes presents an obstacle for the development of synthetic enediyne drugs. In our recent work, we have established an efficient methodology for conversion of the thermally stable 1,2-diynyl-substituted allyl alcohols into acyclic enediynes by rearrangement of the allylic double bond.4 A relatively unstrained 11-membered ring enediyne was synthesized similarly.4b Our methodology is conceptually related to the intramolecular allylic rearrangement proposed for the action of artifacts of the maduropeptin chromophore5,6 and represents one of the emerging strategies7 for enediyne prodrug design and synthesis. In this paper, we disclose the synthesis and DNA cleavage activity of the 10-membered ring enediyne

A novel synthesis of acyclic cis-enediynes 2 has been established by an acid-catalyzed rearrangement of 1,2-diyn-2-propen-1-ols 1 possessing a C3-aryl group in the presence of water, alcohols, or thiols. Reactivity of allyl alcohols and regio- and cis/trans diastereoselectivity of the allylic migration were examined. In the presence of (±)-10-camphorsulfonic acid (CSA), the parent allyl alcohol 5 and the C3-methyl-substituted 9 failed to give enediynes, whereas the C3-aryl-substituted 12 and 29 underwent the allylic rearrangement to provide predominantly cis-enediynes 16 and 31 at room temperature or below. Under similar acidic conditions, enediyne alcohol 13 produced 16b and 16d with the same regio- and cis/trans diastereoselectivity observed for 12. Allyl alcohol 30, an isomer of 29, also provided enediynes 31c and 32c after a prolonged reaction (90h) at room temperature in the presence of CSA and EtOH. These results suggested that the same allylic cations were obtained from allyl alcohols 12 and 13 or 29 and 30 even though the ease of ionization differed for each substrate. Involvement of allylic cations in the product-forming step was confirmed by the finding that chiral allyl alcohols (-)-12 and (-)-18c furnished racemic products. In general, the p-MeOPh-substituted allyl alcohol 29 gave a better regioselectivity than the Ph-substituted 12. In the reactions with alcohols, the regioisomeric ratios were 100:0 (31:33) for 29 and ca. 96:4 (16:17) for 12; the ratios decreased to ca. 90:10 (31:33) for 29 and ca. 70:30 (16:17) for 12 when thiols were used. The cis/trans diastereoselectivity is higher for allyl alcohol 12 (100% for 16 at 20℃) compared to that for 29 (31:32) 80:20-94:6 at 0℃). Computational calculations at the RHF/3-21G level, carried out on the model compounds and allylic cations, indicated that nucleophilic trapping takes place preferentially at the C3 carbon to form the thermodynamically much more stable enediynes. Under the best reaction conditions (1 equiv of CSA and 2 equiv of EtOH in CH2Cl2, 20℃), a number of acyclic cis-enediynes can be synthesized in three steps from the commercially available R-bromocinnamaldehyde (10).

ica gel in wet benzene. Kinetic studies used to determine the free energies of activation (AG*) for the cycloaromatization of 8 (22.6kcal/mol, 30℃) and 9 (25.7kcal/mol, 37℃) to products 32 and 33, respectively. Ab initio calculations regarding the reactivity of these systems were in agreement with the experimental findings. The isolation of compounds 8, 9, 30, and 31 provide strong support for the postulated intermediates in the dynemicin A reaction cascade. The physical, chemical, and biological profiles of the reported compounds may provide the basis for further applications in mechanistic, biological, and medical studies.

Continuing the theme of the preceding article, this paper describes the synthesis and chemical properties of designed enediynes related to dynemicin A. These model systems are equipped with triggering devices at C-3 of the aromatic nucleus. The design of these compounds (1 and 2) was based on the hypothesis that a C-3 phenolic group generated in situ would be capable of promoting epoxide opening and subsequent Bergman cycloaromatization according to the dynemicin A cascade. Compound 1 carrying a tert-butyl ester group at C-3 was synthesized from quinoline derivative 28 via the sequence 28 36→45→46→47→48→44→49→50→1. Compound 2 carrying the photoremovable (2-nitrobenzyl)oxy group at C-3 was constructed from quinoline 29 by a similar sequence. Exposure of I and 49 to aqueous LiOH in EtOH led to Bergman cycloaromatization products 58 and 57, respectively. Compounds 2 and 62 bearing the 2-nitrobenzyl group at C-3 were photolytically converted to free phenolic systems 63 and 64, respectively. Reaction of 63 and 64 with the nucleophiles EtOH, EtSH, or nPrNH2 under anaerobic conditions in basic buffer solutions led to aromatized products 66-70. Exposure of 63 and 75, on the other hand, with EtOH under aerobic conditions in basic buffer solutions furnished the novel quinone methide epoxide systems 71 and 76-77, respectively. The chemistry of compounds 63 and 64 combined with their DNA-cleaving capabilities provides support for the quinone methide mechanism of action of dynemicin A.