A new 3-ethoxycarbonylisoxazolopyridone as a precursor to acylpyridones

A new 3-ethoxycarbonylisoxazolo[4,5-c ]pyridone has been prepared as a potential precursor to 3-acyl-4-hydroxypyridones, by a route involving dipolar cycloaddition of carboethoxyformonitrile with an enamino-ester derived from β -alanine.

During a programme of synthesis towards metabolites containing the enolised heterocyclic tricarbonyl motif 1 (Figure 1), we have reported dipolar cycloaddition strategies to access dihydroisoxazolo- [4,5-c]pyridin-4-one 2 (Scheme 1; enamine from a β-ketoester plus nitrile oxide from a nitro compound) and the [4,3-c] isomer 3 as masked non-polar synthons for the 3acyl-4-hydroxypyridin-2-one nucleus 4.This unit is found in a group of metabolites with a range of biological activities, exemplified by pigments tenellin 5a and bassianin 5b isolated from insect pathogenic fungi, and by the elfamycin antibiotics.In order to exploit the strategy of Scheme 1, it was essential to elaborate the bicycle 2 at C-3.We have reported on simple 4alkoxycarbonylisoxazoles with electrophilic carbon substituents at C-3, and now wish to communicate our studies towards a 3-ethoxycarbonyl analogue of 2. The β-ketoester 6 was prepared from N-benzyloxycarbonyl-β-alanine as we have reported, 2 and converted into the pyrrolidine enamine 7 (pyrrolidine, toluene, Dean-Stark trap for water removal).Attempted reaction with nitrile oxides derived from 2-(2-nitroethyoxy)tetrahydro-2Hpyran or 2,2-diethoxynitroethane (POCl 3 , Et 3 N, 0°C) 6 did not lead to any isoxazole cycloadduct even with a ten-fold excess of nitro-compound, in various solvents and using other protocols for dipole generation.We thus resorted to generating a 3-substituent at the carboxylate oxidation level.The required dipole, carboethoxyformonitrile oxide (CEFNO), was generated in situ by base treatment of ethyl chlorohydroxyiminoacetate 8 (Et 3 N, Et 2 O, 25°C), itself available from ethyl glycinate hydrochloride (NaNO 2 , aq.HCl).Reaction of CEFNO with the enamine 7 (Scheme 2) afforded the 3,4-bis(ethoxycarbonyl)isoxazole 9 (56%).CEFNO is recognised as a reactive dipole, so success with this reagent was not unexpected.

Scheme 2
Removal of the benzyloxycarbonyl protection was efficiently accomplished to afford the amine hydrobromide salt 10 (HBr-AcOH, 25°C, 4 h; 85%), and subsequent basification (aq.Na 2 CO 3 , 25°C, 16 h) resulted in cyclisation to give 3-ethoxycarbonyl-4,5,6,7-tetrahydroisoxazolo [4,5-c]pyridin-4-one 11 (70%).The NMR spectral data for 11 are interesting, in that two sets of signals are observed in the 1 H and 13 C spectra.We have reported a related phenomenon in the isoxazolopyridine 12 (Figure 2) and suggested it may be attributed to a pyridone-hydroxypyridine tautomer mixture. 2 It is possible this may be the case also for 11, with the hydroxy-tautomer stabilized by hydrogen bonding to the ester carbonyl group.The proportion of the two sets varies with concentration, but we have no further evidence at this stage.

Figure 2
A variety of hydride reagents (LiBH 4 , NaBH 4 , NaBH(OMe) 3 , DIBAL) was employed without success in an attempt to reduce the 3-ethoxycarbonyl function (to, for example, an aldehyde for C-C bond forming processes).We have since learnt of reductions to isoxazolidines and ring cleavages of 4-acylisoxazoles under related conditions, which are consistent with our observations.
Finally we unmasked the tricarbonyl functionality by hydrogenolysis (H 2 , Pd-C, MeOH) to afford a product of N-O bond cleavage (Scheme 3).NMR data indicated the product to be a 1:1 tautomer mixture in CD 3 OD solution, which we draw as the enolised dicarbonylimine 14 and a dicarbonylenamine such as 15, although we have no definitive evidence.This functional array is closely related to that observed in the herbicide GRASP 13 (Figure 2).
We thus report a 3-ethoxycarbonylisoxazolopyridone with potential as a precursor to 4hydroxy-3-acylpyridones of biological significance.

Scheme 3
Experimental Section General Procedures.Melting points were determined using a Kofler hot stage apparatus and are uncorrected.Infrared spectra were recorded on a Perkin-Elmer 1720X FT spectrometer.Nuclear magnetic resonance (NMR) spectra were recorded using the following instruments: 1 H spectra at 250 MHz on a Bruker WM250 PFT or at 400 MHz on a Bruker AM400 PFT; 13 C spectra on a Jeol JNM-EX270 at 68 MHz or a Bruker AM400 PFT at 100 MHz and multiplicities were determined using DEPT sequences.NMR Spectra were recorded for solutions in deuteriochloroform with tetramethylsilane as an internal standard unless otherwise stated.Chemical shifts are reported in parts per million (ppm) with the following abbreviations: ssinglet, d -doublet, t -triplet, q -quartet and br -broad; coupling constants (J) are quoted in Hz.Mass spectra were recorded on AEI MS902, VG 7070E or VG Autospec spectrometers using electron impact as the ionisation technique.Microanalytical data were obtained using a Perkin-Elmer 240B elemental analyser.Solvents were distilled prior to use; methanol was dried over magnesium turnings and diethyl ether over sodium.Column chromatography was carried out at medium pressure using Merck Kieselgel 60 silica (ART.7729), and thin layer chromatography (tlc) was carried out using silica G plates F254 (Merck 5554).Solvent extracts were dried (MgSO 4 ) for 10-30 min.before filtration and the solvent removed using a Büchi rotary evaporator.