Synthesis of a protected 2-ethoxy-3-hydroxyethylfuran and its regioselectivity as a Diels–Alder diene on reaction with 3,5-dimethoxydehydrobenzene

As a potential route for the synthesis of naturally occurring benzisochromanequinones, 2-ethoxy- 3-(1'-methoxymethoxyethyl)furan 15 was prepared and shown to react regioselectively with in situ generated 3,5-dimethoxydehydrobenzene to afford, after hydrolysis and alkylation, the target naphthalene 1-(6',8'-dimethoxy-1'-ethoxy-4'-isopropoxy-2'-naphthalenyl)-1- (methoxymethoxy)ethane 27.


Introduction
3][4] Such hydroxyethylnaphthalenes are frequently obtained by the acetylation, followed by reduction, of a precursor naphthalene in additional sequential steps that reduce overall yields.Furthermore, the correct regioselectivity for the acetylation has to be achieved.A more convergent approach for the construction of the derived hydroxyethylnaphthalene might involve the use of a synthon already carrying this hydroxyethyl substituent.Having recently shown that oxygenated naphthalenes are readily synthesised with high regioselectivity and yields through addition of 3-methoxydehydrobenzenes to 2methoxyfuran, 5 the extension of these studies to the use of a 3,5-dialkoxydehydrobenzene 2 with its derivatives were used in subsequent experiments.In the 1 H NMR spectrum of each of these furans 11 and 12, the coupling constant between the two adjacent heterocyclic protons 4-H and 5-H was 2.3 Hz, consistent with reported values for 2,3-(ortho-) coupling. 12Both furans were unstable and decomposed significantly within several hours at room temperature, although both could be stored for several months at temperatures below -20 o C.
Attempted reduction of furan 11 with lithium aluminium hydride led to its decomposition, while with sodium borohydride there was no reaction.The alcohol 13 was, however, obtained in excellent yield (92%) by reduction of the ketone with sodium bis(2-methoxyethoxy)aluminium hydride (Red-Al) at room temperature in the presence of triethylamine.The structure of the alcohol was assigned on the basis of its 1 H and 13 C NMR and mass spectra.In the former, the acetyl methyl resonance of the starting material 11 at δ 2.34 was replaced by resonances at δ 1.42, δ 2.20 and δ 4.80 for the new methyl, hydroxy and methine protons in the product 13.In the 13 C NMR spectrum the carbonyl carbon signal at δ 191.6 in the starting material was lost in the product.In the mass spectrum, fragmentation of the molecular ion of the product 13 at m/z 156 afforded the base peak at m/z 139 and a second signal at m/z 138 (99%), through loss of a hydroxyl radical and water respectively.
Several protecting groups for the alcohol were examined.Dissolution in pyridine followed by addition of acetic anhydride did not give the acetate, nor could the t-butyldimethylsilyl ether be formed with either the silyl chloride and imidazole or by using t-butyldimethylsilyl trifluoromethanesulphonate, decomposition of the alcohol 13 taking place instead.The benzyl ether 14 was formed in 63% yield (as determined by 1 H NMR spectroscopy) using sodium hydride and benzyl bromide, although it was not possible to separate the product 14 completely from residual benzyl bromide.Protection of the alcohol 13 as the methoxymethyl derivative 15 was achieved in 70% yield with chloromethyl methyl ether and diisopropylethylamine in dichloromethane.In the mass spectrum of each of the ethers 14 and 15, the same fragment ion at m/z 139 was prominent, arising from loss of a benzyloxy or methoxymethoxy radical respectively.

Diels-Alder reactions
The reactivity of 3-acetyl-2-ethoxyfuran as a Diels-Alder diene was investigated.Attempted reaction of the furan 11 with the common dienophile maleic anhydride under various conditions of solvent and temperature (including in a sealed ampoule at 150 o C in toluene) afforded no products of cycloaddition.On prolonged treatment with the strongly dienophilic dimethyl acetylene dicarboxylate in benzene under reflux, however, the known 7 furan 16 was obtained in moderate yield.This product had been obtained previously by Grigg on reaction of the oxazole 8 with this latter dienophile.
It was recognized at the outset that production of a benzyne (Scheme 1) in situ under basic conditions in the presence of the acetylfuran 11 could lead instead to removal of an acetyl proton.Use of 2-bromo-3,5-dimethoxyphenyl p-toluenesulfonate 19 5 as the precursor to 3,5dimethoxybenzyne 20 and butyl lithium (one equivalent) as the base might, however, lead to preferential lithium -halogen exchange, and therefore generation of the benzyne 20 in the presence of the acetylfuran.The successful regioselective reaction of these would lead to the formation of acetylnaphthols and thence acetylnapthoquinones related to those used by the Brimble route 13 to benzisochromanequinones.In practice, however, only products of decomposition were observed.The in situ generation of benzyne itself under neutral conditions (from anthranilic acid and isoamyl nitrite 14 ) in the presence of the acetylfuran 11 also gave the same result.This suggested that the acetylfuran 11 was a poor diene in comparison with 2methoxyfuran. 5 No reaction took place between the hydroxyethylfuran 13 and either maleic anhydride or the butynone 9.This diene did react readily, however, with dimethyl acetylene dicarboxylate to afford the unstable adduct 17 in 72% yield, which was identified by 1 H NMR spectroscopy.This rearranged under the influence of hydrochloric acid to give the phenol 18 in high yield.The diene 13 was not necessarily expected to undergo a Diels-Alder reaction with the benzyne 20, since alcohols are known to react as nucleophiles with methoxybenzynes, 15 and, in practice, no product was isolated.
The reactions of the protected furan alcohols with 3,5-dimethoxybenzyne 20 produced in situ were examined.When the benzyl protected derivative 14 was reacted with the bromotosylate 19 in the presence of butyl lithium, the presumed Diels-Alder adduct 21 was treated with trifluoroacetic acid, and the derived mixture was reacted with acetic anhydride and pyridine, the naphthyl acetate 24 was isolated in 24% yield (Scheme 3).In addition, a small amount (8%) of the naphthol 23 was obtained.The formation of this additional naphthol may have arisen through acidic work-up of the acetate 24 to remove pyridine used in its formation, either through hydrolysis of the acetate or from residual adduct 21 remaining from its incomplete ring-opening prior to acetylation of the naphthol 23.
When the same benzyne 20 was produced in situ in the presence of the methoxymethyl protected alcohol 15, the adduct ring-opened and the derived naphthol 25 acetylated (Scheme 3), the naphthyl acetate 26 was obtained in 41% overall yield based on furan 15.This represented an average yield for each of the three steps of approximately 75%.Alternatively, when the naphthol 25 was protected as the 2-propyl ether 27, this was isolated in 23% yield, with a further 10% of the precursor naphthol 25 also being recovered.Proof was sought for the anticipated regiochemistry of the naphthalene 27 arising from the cycloaddition reaction between the protected hydroxyethylfuran 15 and the benzyne 20.Initially, the conversion of 27 into a known compound was examined.As a preliminary step, the removal of the methoxymethyl protecting group was attempted, first with concentrated hydrochloric acid and, subsequently, with trifluoroacetic acid.This led in both cases only to products of decomposition.Milder conditions were then sought.With trimethylsilyl bromide 16 in methylene dichloride the vinylnaphthalene 28 was isolated as the sole product in a yield of 14%.The same product 28 was also isolated in 55% yield using pyridinium p-toluenesulfonate in 2-butanone. 17 This presumably arises through attachment of the electrophile (X = H or TMS, Scheme 4) to a methoxymethyl oxygen (as in 29) followed by expulsion of the side-chain, this being promoted by the ethoxy and two aromatic methoxy substituents.
Support for the assigned regiochemistry in compound 27 was obtained from a two-dimensional NOESY spectrum.The proximities between the relevant protons are shown in red in Figure 1.In particular, the ethoxy group was in close proximity to the 8-methoxy substituent, the methylene protons of the methoxymethyl group and the benzylic methine proton of the alkoxyethyl sidechain.The isopropoxy substituent was found to be adjacent to the aromatic protons H-3 and H-5.
From these results it can be seen that the furan 11, and therefore furan 12, are the 2,3disubstituted structural isomers assumed above and also that cycloaddition of the furan 15 (and therefore furan 14) and the benzyne 20 occurred with the anticipated regiochemistry for the only naphthalene isolated.

Conclusions
Syntheses for the 3-acetyl-2-alkoxyfurans 11 and 12 were readily developed, together with those of the related 3-hydroxyethylfuran 13 and its protected derivative 15.
Conditions have yet to be established for the use of the latter as an appropriate synthon for the synthesis of benzisochromanequinones.The acetylfurans were poor Diels-Alder dienes, reacting slowly and in moderate yield with the strongly dienophilic dimethyl acetylene dicarboxylate, but not with either maleic anhydride or benzyne.The hydroxyethylfuran 13 reacted rapidly and in very good yield with the first of these dienophiles.Its methoxymethyl derivative 15 reacted with the dienophile 3,5-dimethoxybenzyne 20 with the desired regioselectivity, although the yields of the derived naphthalenes were somewhat lower than for the analogue 2-methoxyfuran. 5Presumably this arose on account of the greater sensitivity of the substituents in the present study to the acid treatment required to convert the Diels-Alder adduct into naphthalenes.Experimental Section General Procedures.Nuclear magnetic resonance spectra were recorded using either a Hitachi R24B spectrometer ( 1 H 60 MHz), a Bruker AM-300 or a Bruker Avance DPX-300 spectrometer ( 1 H 300 MHz, 13 C 75.5 MHz).All spectra were run on the latter two instruments unless otherwise stated.These spectra were recorded at ambient temperature in deuterochloroform (CDCl 3 ) using tetramethylsilane (TMS) as an internal standard.In the 13   for 25 min then allowed to warm to room temperature over 20 min.Trifluoroacetic acid (10 drops) was added and the mixture stirred for 10 min.The reaction was quenched with saturated aqueous sodium hydrogen carbonate and extracted with diethyl ether.Dry triethylamine (2 mL) was added and the organic solution dried (magnesium sulfate) and then concentrated.The residue was dissolved in dry pyridine (2 mL) and acetic anhydride (0.204 g, 2.00 mmol) was added.The reaction was stirred at room temperature overnight (14 h), quenched with water and extracted with dichloromethane.The organic extracts were washed with hydrochloric acid (1M) and water.The extracts were then dried (magnesium sulfate), concentrated and chromatographed (radial, 15% ethyl acetate−petroleum ether) to give two products.The higher R f product identified as 1-benzyloxy-1-(6',8'-dimethoxy-1'-ethoxy-4'-hydroxy-2'-naphthalenyl)ethane 23 (     give the title compound 28 (14 mg, 14%).

Scheme
Scheme

( b )
Trimethylsilyl bromide (0.184 g, 1.20 mmol) was added to a solution of compound 27 (0.114 g, 0.30 mmol) and 4Å molecular sieves (0.5 g) in dry dichloromethane (10 mL) at -78 o C. The reaction was stirred at -78 o C for 25 min, quenched with saturated aqueous sodium hydrogen carbonate and extracted with dichloromethane.The extracts were dried (magnesium sulfate) and concentrated.The residue was chromatographed (radial, 10% ethyl acetate−petroleum ether) to

Issue in Honor of Dr. Douglas Lloyd ARKIVOC 2002 (iii) 149-165 ISSN 1424-6376 Page 157
C NMR spectra, assignments of signals with the same superscript are interchangeable.Mass spectra were recorded on either a Hewlett Packard 5986 spectrometer at 35 eV, or on a Perkin-Elmer ITD IonTrap Detector spectrometer at 55 µA with automatic gain control.High resolution mass spectra were recorded on a VG Autospec High Resolution Mass Spectrometer.Infrared spectra were The mixture was cooled to 0 o C and the solution of ethyl alaninate 6 in tetrahydrofuran was added dropwise.The reaction was stirred at 0 o C for 1 h 30 min and then at room temperature overnight (18 h).The reaction was concentrated, the residue dissolved in ethyl 10corded as thin films between KBr plates for oils and as KBr discs for solids using a Perkin-Elmer 1720−X Fourier Transform Spectrometer.Melting points are uncorrected and were recorded on a Reichert hot stage apparatus.Column chromatography was carried out on columns prepared as slurries of Merck silica gel 60 (70−230 mesh) in the eluent.Preadsorption was carried out on Merck silica gel 60 (35−70 mesh).Radial chromatography was performed using Merck silica gel 60 PF 254 .Preparative layer chromatography was performed on glass plates coated with Carmag silica gel as a 0.3 mm thick layer, while thin layer chromatography was carried out on aluminium plates coated with Merck Kieselgel 60 F 254 .Petroleum ether refers to the fraction of boiling point 65−70 o C. All solvents were purified by distillation and, if necessary, were dried according to standard methods.The amount of residual water present in solvents was determined using a Metrohm Karl Fischer Coulometer 684.Ethyl N-formyl alaninate (7).(a)A solution of DL-alanine (1.00 g, 11.2 mmol) and formic acid (2.58 g, 56.1 mmol) in ethanol (13 mL) was heated in a sealed glass ampoule at 150 o C overnight (16 h).The mixture was then cooled and distilled on a Kugelrohr at 130 o C/0.12 mmHg to give the known title compound 7 (1.20 g, 74%) (lit.,10100oC/0.1 mmHg); δ H (60 MHz) 1.28 (3H, t, J (26.21 g, 256.7 mmol) was added to a solution of sodium formate (17.46 g, 256.7 mmol) in dry tetrahydrofuran (40 mL).The reaction was cooled to 0 o C and formic acid (11.82 g, 256.8 mmol) added dropwise.The reaction was then warmed to 35−40 o C for 1 h and then cooled to room © ARKAT USA, Inc temperature for 10 min.underrefluxovernight(19h).The mixture was cooled and saturated aqueous sodium hydrogen carbonate (300 mL) was added carefully.The solution was filtered and the resulting layers separated.The aqueous layer was extracted with dichloromethane.The combined organic layers were washed with water and dried (magnesium sulfate).Careful removal of the solvent under vacuum gave the oxazole 8 as a pale yellow oil [δ H (60 MHz) 1.32 (3H, t, J 7.2, CH 2 CH 3 ), 2.00 (3H, s, 4-CH 3 ), 4.05 (2H, q, J 7.2, CH 2 CH 3 ), 7.25 (1H, s, 2-H)] which was immediately dissolved in dry toluene (40 mL).To this solution was added hydroquinone (40 mg) and 3-butyn-2-one 9 (1.54 g, 22.6 mmol), and the mixture, under nitrogen, stirred under reflux for 5 h 30 min.Cooling, followed by concentration, gave a yellow oil which was immediately purified by column chromatography(10−20% ethyl acetate-petroleum ether) to give the product 11 as pale yellow crystals (1.34 g, 46%) mp 45−46.5 o C (from hexane) (Found: C, 62.0; H, 6.55.C 8 H 10 O 3 requires C, 62.35; H, 6.55%); ν max /cm -1 2988 (C-H), 1644 (C=C), 1581 (C=O), 1450 (C=C), 1124 (C-O); δ H 1.42 (3H, t, J 7.1, CH 2 CH 3 ), 2.34 (3H, s, COCH 3 ), 4.42 (2H, q, J 7.1, CH 2 ), 6.62 (1H, d, J 2.3, 4-H), 6.79 (1H, d, J 2.3, 5-H); δ C 14.Issue in