7-(1-Acetoxymethylidene)Benzonorbornadiene: A Versatile Reagent for the Synthesis of 7-Formyl and 7-Hydroxymethyl Benzonorbornadienes

7-(1-Acetoxymethylidene)benzonorbornadiene 3, prepared in one step by the addition of benzyne to 6-acetoxyfulvene 2, is hydrolysed in acid solution to form a 3:2-epimeric mixture of syn- and anti- 7-formylbenzonorbornadienes 4 and 5, respectively; the corresponding 7-hydroxymethylbenzonorbornadienes 6 and 9 were produced by reduction of the formyl isomers with sodium borohydride.


Introduction
7-Substituted benzonorbornadienes are important reagents, yet simple routes for their synthesis are lacking. In principle, such compounds can be prepared by addition of benzyne to 5-substituted cyclopenta-1,3-dienes, however these isomers are themselves very unstable, being converted to the more stable 2-substituted and 3-substituted isomers during the course of benzyne addition. Further, the addition of benzyne to cyclopenta-2,4-dienones, which would circumvent the substrate isomerisation problem and provide a carbonyl group in the 7-position for subsequent modification, is frustrated by the thermal instability of benzonorbornadien-7-ones [1]. Benzyne additions to cyclopentadienyl anions [2] and cyclopentadiene ketals [3][4][5][6] have been reported and have been used to prepare 7-substituted benzonorbornadienes.
Our answer to this issue has been to use a fulvene reagent since they have a fixed structure and can carry latent functionality to the 7-position of benzonorbornadienes. Accordingly, we investigated the addition of benzyne to 6-acetoxyfulvene 2 [7] for the synthesis of 7-(1-acetoxymethylene)benzonorbornadiene 3, since hydrolysis of the enol acetate group should yield the versatile formyl derivatives 4 and 5 at the 7-position. Masking of the norbornene π-bond of 7-(1-acetoxymethylene)benzonorbornadiene as a cycloadduct has the potential to yield products which should be capable of ozonolysis to form stable benzonorbornen-7-one derivatives suitable for CO modifications. These latter aspects have been described separately [8], while this paper concentrates on the preparation of the title compound, its hydrolysis to 7-formyl-benzonorbornadienes and reduction to the 7-hydroxymethyl derivatives.

Results and Discussion
The reaction of benzyne 1, generated from diazotisation of anthranilic acid [9], to 6-acetoxyfulvene 2 proceeded without significant deacetylation to form 7-(acetoxymethylidene)-benzonorbornadiene 3 in 45% yield (Scheme 1). The 1 H-NMR of 3 confirmed the presence of the acetate group (methyl resonance at δ 2.40), while the unsymmetrical nature of the structure was evident from the bridgehead protons which appeared as well separated resonances (δ 4.16, 4.56). The aromatic and norbornene π-bond resonances overlapped to produce an unresolved multiplet at δ 6.80-7.40. The resonance for the exocyclic vinylidene proton in 3 occurred as a singlet at δ 6.58, considerably downfield from that observed for 7-methylidenebenzonorbornadiene 8 (δ 4.12), consistent with the electron-withdrawing capacity of the acetoxy substituent.

Scheme 1
Acid hydrolysis of 3 was conducted in acetone containing concentrated hydrochloric acid [10] and produced the syn-isomer 4 (stereochemical designations have been made relative to the benzene ring) and the anti-isomer 5 as an oily mixture (3:2, as determined by 1 H-NMR) in 69% yield. The individual isomers, which were separated by chromatography, each displayed a characteristic lowfield resonance for the aldehyde proton. Significantly, there was a difference of 0.42 ppm in the chemical shifts of the two isomers, and this was used to help make their structure assignments. The higher field resonance (δ 9.18) of the major isomer 4 also displayed a vicinal coupling (J = 2.5 Hz), which was not observed with its isomer 5 (slightly broadened singlet at δ 9.60). The shift to high field of the major syn-isomer was consistent with the aldehyde being positioned above the benzenoid ring and this was confirmed by the presence of an NOE between the formyl proton and the aromatic proton H a . The stabilisation afforded by the interaction of the formyl proton and the benzene ring fixed the formyl group in 4 in the transoid configuration. This proposal is supported by the 2.5 Hz coupling of the former proton with the bridge proton H b . The lack of coupling in the anti-isomer 5 suggested that free rotation was occurring. Neither aldehyde could be obtained crystalline, so the syn-isomer 4 was converted to its pale yellow-coloured 2,4-dinitrophenyl hydrazone derivative for characterisation.

Scheme 2
Reduction of the aldehydes 4 and 5 to their respective hydroxymethyl compounds 6 and 9 was achieved in quantitative yield using sodium borohydride. When conducted on a mixture, the generated alcohols were separated by chromatography on silica and the syn-isomer 6 eluted first, suggesting there might be some H-bonding with the aromatic. In keeping with the assignments made for the aldehydes, the methylene protons of syn-isomer 6 occurred at higher field (δ 3.08) than the anti-isomer 9 (δ 3.60) which reflected the fact that the methylene protons were positioned closer to the shielding zone of the aromatic ring. The oily alcohol 6 was characterised as its 3,5-dinitrobenzoyl derivative [11].
Tosylation of a mixture of 6 and 9 was achieved by treatment with tosyl chloride in pyridine and formed a mixture of 7 and 10 that was recrystallised slowly from alcohol to yield two types of crystals. The larger prisms were shown to be the syn-isomer 7, while the colourless needles, which were separated manually from the mixture, corresponded to the anti-isomer 10.

BLOCK Coupling of Syn-7-hydroxymethyl-benzonorbornadiene 6.
A special feature of the different isomers of the 7-hydroxymethyl-benzonorbornadienes 6 and 9 as substrates for BLOCK coupling reactions [12], is the accessibility of the π-bond for reagent attack at the exo-face. This difference is attributed to the screening effect of the hydroxymethyl group when it is on the same side as the π-bond (the anti-isomer 9) thereby precluding attack at that dienophilic centre. Thus, treatment of a mixture of alcohols 6 and 9 with the bis-epoxide 11 [13] gave only one coupled product 12, that derived from alcohol 6, while unchanged alcohol 9 was recovered from the reaction mixture (Scheme 3). The structure of the coupled product 12 was supported by spectroscopy in which the bridge methine proton (δ=3.51) was typically downfield-shifted owing to its proximity to the adjacent oxabridge. The NOE between protons H a and H b fully confirmed the exo,exostereochemistry.
For molecule 12 to be an effective spacer system, then attachment of functionality at the hydroxymethyl groups must be easy to achieve. In this case, acylation of the pendant hydroxyl groups was readily achieved by acetylation with acetic anhydride containing sodium acetate as base, to thus form the diacetate 13 in 93% yield. This derivatisation is a marked improvement over spacer units in which the hydroxyl group is directly attached to the methano-bridge [14].

Scheme 3 Conclusions
A direct entry to 7-formyl benzonorbornadienes by hydrolysis of 7-(1-acetoxymethylene)benzonorbornadiene has provided access to a versatile set of syn-and anti-starting materials that can be readily transformed to other 7-substituted benzonorbornadienes, many of which are difficult to prepare by other methods.

Acknowledgements
PF thanks CQU for the award of a PhD scholarship while GC acknowledges the award of a MSc scholarship from ANU. Funding support through the Centre for Molecular Architecture at CQU has enabled this work, which had its origins at ANU in the 1980s, to be completed. Several compounds described herein are presently being used in ARC-funded projects at CQU.

General
Microanalyses were carried out by the ANU Microanalytical Service. Melting points, which are uncorrected, were determined with a Reichert melting point apparatus. IR spectra were recorded on a Unicam SP200G infrared spectrophotometer. UV spectra were recorded with a Unicam SP800 spectrophotometer. Mass spectra were recorded on a Varian Mat CH7 or an A.E. I. MS902 spectrometer. EI or ESMS (electrospray mass spectrometry) were conducted on a Micromass Platroem II single quadripole mass spectrometer. 1 H-NMR (100 MHz) spectra were recorded on a Jeolco Minimar 100-MHz spectrometer. 1 H-NMR (400 MHz) and 13 C-NMR (75 MHz) were recorded on a Bruker Avance DPX 400-MHz spectrometer. All data was acquired using CDCl 3 solutions with TMS as an internal standard and are reported in ppm on the appropriate δ H and δ C scales. Coupling constants are reported in Hz. The silica gel used for column chromatography was silica gel 60 (230-400 mesh). TLC was performed on Merck aluminium sheets coated with silica gel 60 F 254 . Radial chromatography was carried out with a Chromatotron, Model No 7924T, using 1 mm plates coated with silica gel 60 F 254 . All preparative thin-layer chromatography was carried out on plates using 2 mm layers of silica gel (Merck GF 254 ). All solvents were removed under reduced pressure.

Syn-and anti-9-
A solution of enol acetate 3 (2 g, 9.4 mmol) in concentrated hydrochloric acid (4 mL) and acetone (16 mL) was refluxed for 0.5 h. After cooling, water (20 mL) was added and the mixture extracted with ether (2 x 20 mL). The ethereal extract was washed with water (30 mL), dried and freed of solvent. The 1 H-NMR spectrum of the crude product indicated a 3:2 mixture of the syn-isomer 4 and anti-isomer 5. The residue was distilled (approx. 80 °C, 0.1 mbar) to give the products as a colourless liquid (1.1 g, 69%). The two isomers were separated by preparative TLC (silica, 1:1 chloroform and petroleum spirit).

Syn-and anti
A mixture of the syn-aldehyde 4 and anti-aldehydes 5 (0.6 g, 3.5 mmol) in the ratio of 3:2 and sodium borohydride (0.2 g, 6.7 mmol) were stirred in ethanol (10 mL) for 1 h. Water was added and the solution extracted with ether. The ethereal extracts were dried, the solvent removed and the residual mixture of alcohols separated by preparative thin-layer chromatography (silica, 7% ethyl acetate in petroleum spirit, 10 developments).

Syn-and anti
A solution of the syn-alcohol 6 and anti-alcohol 9 (0.4 g, 2.3 mmol) in the ratio of 3:2 and tosyl chloride (0.36 g, 1.9 mmol) in anhydrous pyridine (3 mL) was stirred under nitrogen for 3 days. Water (10 mL) was added and the solution extracted with ether (3 x 5 mL). The combined ethereal extract was dried and freed of solvent. The residue was allowed to slowly crystallise from ethanol over a period of several days giving two crystal forms which were separated manually.