Syntheses with isoxazoles. Part IV. Ring-opening reactions of 4,5,6,7-tetrahydro-4-oxoisoxazolo2,3-apyridinium salts and related compounds under the influence of tertiary amines

摘要

1976 1241Syntheses with Isoxazoles. Part 1V.l Ring-opening Reactions of 4,5,6,7-Tetrahydro-4-oxoisoxazolo2,3-apyridinium Salts and Related Com-pounds under the Influence of Tertiary AminesBy Gurnos Jones and John R. Phipps, Department of Chemistry, University of Keele, Keele, StaffordshireST5 5BGOn treatment of the 4,5,6,7-tetrahydro-4-oxoisoxazolo2,3-a pyridinium salts ( I ) and (3) with triethylaminein methanol, methyl 4-(2-formyl-l -methoxyvinylamino) butyrate (7) and its homologue (9). respectively,are formed. The oxime ( I 3) and the phenylhydrazone (1 4). under similar conditions, give the methyl piperidin-2-ylideneacetates (1 6) and (1 7). The mechanisms of the ring-opening reactions are discussed. Compound (7)is converted spontaneously into N-(trans-3-methoxyacryloyl)-2-pyrrolidone (1 2), which has been synthesised.WE have reported 1-3 that the isoxazolopyridinium salts(1) and (2), treated with boiling acetic anhydride for afew minutes, are converted into the reduced furopyri-dinones (4) and (5) ; pyrrolopyridinones were similarlyY-1 2 (1) R1=R2=H,X =O,Y=Br (4) R = R =H,X=O(2) R' =H,R2=Me,X=0,Y= Br(3) R = Me, R2= H, X = 0,Y = Br(5) R1=Me,R2= H,X=O1formed from the oximes (X = NOH). The failure ofthe isoxazolopyridinium salt (3) to undergo a similar0H Base ~ bcof the mechanisms which we have suggested.Sinceanalogies exist for both mechanisms among the reactionsof isoxazolium salts, we have attempted to form andtrap the keten intermediate (6) required by Scheme 1,and thus establish the reaction pathway.In our first attempts, the salts (1) and (3) were treatedwith triethylamine in boiling methanol.Surprisingly,both salts gave non-quaternary products in satisfactoryyields, but the spectral data left little doubt that a moredrastic change had occurred than had been observed inthe acetic anhydride reaction. The compound from thesalt (1) showed in the lH n.m.r. spectrum a pair of widelyspaced doublets (both 1 H, J 4.0 Hz) at 6 8.85 and 4.55,whereas that from the salt (3) showed no signal near6 8.85, but a singlet (1 H) at 6 4.7. The product from thesalt (1) showed two 3 H singlets at 6 3.60 and 3.75assigned to methyl ester (or vinylogous ester) groups,amp;*'HSCHEME 1reaction we have ascribed to the absence of a hydrogenatom at position 2, essential to the conversion in either1 Part 111, G.Jones and J. R. Phipps, J.C.S. Perkin I, 1974,2 R. H. Good, G. Jones, J. R. Phipps, G. Ferguson, and W. C.and the rest of the spectrum showed a broad NH absorp-tion (vinylogous amide) at 6 10.5 and a trimethylenegrouping. The spectrum of the product from the saltR. H. Good, G. Jones, and J. R. Phipps, J.C.S. Pamp;in I,1972,2441.158.Marsh, Tetrahedron Letters, 1972, 6091242 J.C.S. Perkin I(3) was very similar; two satisfactory formulae can bewritten for each product: (7) and (8), and (9) and (lo),respectively. The i.r. and U.V. absorptions agreeequally well with either; thus v,, at 1 575 and 1625cm-l and A,, at 289.5 nm (log E 5.3) are comparable tothose reported for enamino-ketones v,, 1 575 and 1 635cm-l, A,, 278 (4.12).4 The 13C n.m.r.spectrum showedthree signals, at 183.9, 173.5, and 169 p.p.m. (fromMe,Si), in the low-field region expected for carbonylcarbon atoms, the signal at lowest field being shown byanism for the formation of the enamino-aldehyde (7) andthe enamino-ketone (9) can be envisaged by considerationof the base-catalysed elimination of 3-acyl substituentsfrom isoxazolium salts. The intermediate ketenimine(11) is then analogous to that involved in the Mumm-Woodward-Olofson synthesis of peptides .5Compound (7) proved unstable to strong bases and toacid. The liquid compound (7) but not' compound (9)showed signs of slow crystallisation. A crystallinecompound was isolated in 20-30y0 yield and shown byOMe I R-C=C.CO - NH.CH21iC0,MeI H(10) R =Me(8) R = HMeO-H J !1 RC 0 .CH =C (OMe ) * N H. C Hamp;' COzMe(7) R = H(9) R =MeOHoff-resonance decoupling to be due to a formyl group.The small coupling constant between the low-field signalsin the lH n.m.r. spectrum, combined with the observationthat the 13C n.m.r. signal at lowest field is due to aformyl group, led us to prefer formulae (7) and (9) forthe two products, and a synthesis of compound (12) (seebelow) confirmed this preference. A satisfactory mech-analysis and mass spectral data to have been formed byelimination of methanol. The lowest field signal (1 H)in the lH n.m.r. spectrum of the new compound was at6 7.98, and showed a coupling to a signal at 6 5.12 (14 Hz) ;the i.r.absorptions in the carbonyl region were at 1 724,1 690, and 1 620 cm-l. A 13C n.m.r. spectrum showedtwo similar carbonyl carbon absorptions at 174.3 andE. J. Cone, R. H. Gamer, and A. W. Hayes, J . Org. Chem., 5 R. B. Woodward and R. A. Olofson, J . Amer. Chewz. SOC.,1972, 37,4436. 1961,83, 10071976167.6 p.p.m., two widely separated alkene carbon absorp-tions at 100.2 and 137.5 p.p.m., a methoxy-group andthree saturated methylene groups.These data agree with only one reasonable formula,(12), and this was confirmed by synthesis from thesodium salt of 2-pyrrolidone and trans-3-methoxyacryloylchloride. The transformation of compound (7) intocompound (12) is formally similar to that observed inthe peptide synthesis 5 with the important differencethat an alkyl rather than acyl group is migrating. Wesuppose that hydrolysis of the methoxy-group in com-pound (7) is followed by methylation of the terminalcarbonyl group and cyclisation of the y-aminobutyrateester to give the pyrrolidone, but we have not been ableto confirm any of these stages, nor, as yet, to acceleratethe transformation (Scheme 2).A referee has suggestedthat a 1,5 methyl shift is possible, but not probable, inthe light of the preferred H-bonded stereochemistry ofcompound (7), as shown in Scheme 2.Our first attempts to obtain a keten intermediate werethus frustrated by the highly electrophilic carbon atomat position 4 of the original isoxazolopyridinium salt (1).The electrophilic activity of this carbon atom wasexpected to be much less in the corresponding oxime orhydrazone.We have already reported the preparationof an oxime (13) from compound (l), and we have nowprepared a phenylhydrazone (14). When the oxime (13)n(13) R = OH, Y=Br( 1 4 ) R = NHPh,Y=Br(16) R=OH(17) R=NHPhSCHEME 3was treated with triethylamine in boiling methanol acompound was obtained whose spectral characteristicswere those expected for the tetrahydropyridine (16).The presence of a methyl ester was shown by carbonylabsorption (vmX 1640 cm-l; enamho-ester) and by a3 H lH n.m.r. singlet at 6 3.6. A sharp singlet (1 H)at 6 5.18 (=CH) and broad singlets at 6 8.8 and 9.3 wereall removed in D,O. The exchange of the alkene protonprovides evidence for the tautomeric equilibrium(15) f-c (16), shown in Scheme 3.The product (17)obtained after treatment of the phenylhydrazone (14)6 G. J . Karabatsos and R. A. Taller, J . Amer. Chem. SOC.,1963, 85, 3624 give 6 6.8 for acetophenone phenylhydrazone.with triethylamine in boiling methanol showed similarcharacteristics; a singlet due to an exchangeable protonat 6 5.42, and two broad NH signals at 6 7.55 and 8.95(tentatively assigned to the phenylhydrazone NH andto H-1, by comparison with similar systems 6). We feelthat the isolation of these two products confirms theintermediacy of a keten in the opening of the isoxazolering. When the ring opening reagent is acetic anhydride,the acetylation of the piperidine nitrogen atom increasesthe electron deficiency at the keten carbon atom andpossibly enhances also the contribution from ' enol 'tautomers which can cyclise to give the lactone or pyr-rolidone ring.EXPERIMENTAL1.r.spectra were measured on a Perkin-Elmer 257 spectro-meter and 1H n.m.r. spectra on a Perkin-Elmer-HitachiR24 instrument ; 13C n.m.r. spectra were determined byJEOL (U.K.) Ltd., with an FT-60 spectrometer. M.p.swere determined on a Kofler hot-stage apparatus. Columnchromatography was performed on Woelm alumina, activity4.Methyl 4-( 2-Forwyl- 1-methoxyvi~zyla~~zino) butyrate (7) .-The isoxazolium bromide (1) (1 g) was dissolved in boilingmethanol (30 ml) and triethylamine (1.5 ml, ca. 2 equiv.)was added.After further boiling (3-5 min) the solventwas removed under reduced pressure and the residueextracted with dry acetone (5 nil) and filtered from pre-cipj tated triethylamine hydrobromide. Evaporation ofthe filtrate gave an oil which was adsorbed on alumina andeluted with chloroform. The first band eluted (pale yellow)was almost pure ester (7) (0.51 g, 56). The ester couldbe further purified by preparative layer chromatography,but gave inconsistent analyses (Found : M+, 201.1003.C,H,,NO, requires M , 201.1001), LX. 292 nm (log E 4.31),vmX. (film) 1 739, 1 625, and 1 575 cm-l, 6H(ccl,) 1.9 (2 H,q), 2.28 (2 H, m), 3.30 (2 H, q), 3.60 (3 H, s), 4.55 (1 H, d,J 4 Hz), 8.85 (1 H, d, J 4 Hz), and 10.5br (1 H), Gc(CDC1,)25.29, 31.15, 39.28, 51.64, 56.03, 78.0, 169.15, 173.47, and183.88.Methyl 4-( l-ll.i7ethoxy-3-oxobut-l-enylaunino)butyrate (9) .-Prepared in 67 yield as described for compound (7), fromthe isoxazolium salt (3), the ester (9) had b.p.100-115"at 0.1 mmHg (bulb-tube) (Found: C, 55.5; H, 7.7; N,6.4. C1,H1,N04 requires C, 55.8; H, 7.95; X, 6.5),Amx 289.5 nm (log E 5.3), vmx. (film) 1 735 and 1 625 cm-1,a~(cDC1,) 1.9 (2 H, m), 2.02 (3 H, s), 2.3 (2 H, m), 3.3(2H,q),3.65(3H,s),3.75(3H,s),4.72(1H,s),and10.75br(1 H), m/e 215 (M+), 200, 172, 142, 128, 116, 101, 99, and 59.( 1 2) .-(a)The ester (7), kept (4 weeks) in the dark, gave a mixture;the pywolodinone (12), isolated by p.l.c., had m.p. 78-78.5"from petroleum (b.p. 40-60deg;) (Found: C, 57.0; H, 6.7;N, 8.45.C,H,,NO, requires C, 56.8; H, 6.55; N, 8.3), LX. 271 nm (log E 4.34), vmax. (Nujol) 1 724, 1690, and1 620 cm-l, G=(CDCl,) 2.32 (4 H, m), 3.5 (2 H, t), 3.67 (3 H,s), 5.12 (1 H, d, J 14 Hz), and 7.98 (1 H, d, J 14 Hz),Gc(CDC1,) 17.46 (t), 30.94 (t), 44.99 (t), 51.41 (q), 100.22 (d),137.5 (d), 167.63 (s), and 174.3 (s) (off-resonance decoupledmultiplicity in parentheses), ?n/e 169 (M+), 138, 110, 82, and41.(b) A solution of 2-pyrrolidone (4.3 g) in benzene (100ml) with sodium hydride (50 dispersion; 2.4 g) was boiledN- (trans- 3-Met h oxy acry Zoy 2) py r.yo Zidi n - 2 -0 n J.C.S. Perkin I(2 h). To the cooled mixture was added, dropwise, trans-3-methoxyacryloyl chloride 7 5.5 g; prepared from trans-3-methoxyacrylic acid *79 and thionyl chloride in boilingether; b.p.77-79" (20 mmHg). The mixture wasstirred a t room temperature (2.5 days), treated with i c ewater, and filtered, and the benzene layer was separated,dried, and evaporated. Chromatography on alumina(elution with benzene) gave first an unidentified methoxy-acryloyl compound, and secondly the pyrrolidinone (12),identical with that obtained as in (a).Methyl (3-Hydroxyiminopiperidin-2-ylidene) acetate (1 6) .-To the oxime bromide (13) (0.5 g) in hot methanol (30 ml;dry) was added triethylamine (0.7 ml). The solution wasallowed to cool to room temperature (ca. 0.5 h), thenevaporated under reduced pressure, and the residue wastreated with chloroform and filtered from triethylaminehydrobromide. The chloroform extract was evaporated onto alumina, and the coated alumina was added to a column.Elution with benzene gave the fliperidinyzideneacetate (1 6),m.p.138.6-139.5deg; (0.2 g, 52) (Found: C, 51.6; H, 6.45;N, 14.7. C,Hl,N20, requires C, 52.15; H, 6.55; N,15.2), ha 230 and 338 nm (log E 5.15 and 4.90), vmX(Nujol) 3 290, 3 220, 1 640, 1 610, and 1 570 cm-l, GH(CDC~,)1.9 (2 H, m), 2.7 (2 H, t) 3.3 (2 H, m), 3.6 (3 H, s), 5.18(1 H, s exch.), 8.8br (1 H, exch.), and 9.3br (1 H, exch.).4,5,6,7-Tetrahydro-4-oxoisoxazolo2,3-apyridinium Browide Phenylhydrazone (14) .-A solution of phenylhydrazine7 I. I. Kolodkina, K. V. Levshina, S. I. Sergievskaya, andA. I. Kravchenko, Zhur. org. Khim., 1966,2,66 (Chem. Abs., 1966,64, 14,087).hydrochloride (0.36 g) in absolute ethanol (10 ml) was addedto a solution of the isoxazolium bromide (1) (0.5 g), also inabsolute ethanol (10 ml).The mixture was boiled (20 min)and evaporated; the residue was crystallised from ethanol,and the salt so obtained passed through a column ofAmberlite IRA 400 resin (Br-). The eluate was crystallisedfrom ethanol to give the phenylhydrazone bromide (14), m.p.300deg; (Found: C, 50.95; H, 4.6; N, 13.5. C13H14BrN,0requires C, 50.65; H, 4.6; N, 13.65), kx. 245, 300sh,and 380 nm (log E 4.13-4.27), G=(CF,-CO,H) 2.5 (2 H, m),2.8 (2 H, m), 4.55 (2 H, t), 7.4 (6 H, m, Ph + isoxazoliumH), and 8.52 (1 H, d, J 2 Hz).Methyl (3-Phenylhydrazinofliperidin-2-ylidene)acetate (17).-The phenylhydrazone (14) (1 g) was treated as describedfor compound (13), and gave the piperidinyl acelate (17),m.p. 122-124' (0.05 g, 6) (Found: C, 64.55; H, 6.55;N, 16.15. C,,H,,N,O2 requires C, 64.85; H, 6.6; K,16.2), vmX. (film) 3 280, 1 620, and 1 595 cm-l; A, 242,and 330 nm (log E 4.15 and 4.22), s~(cDC1,) 1.92 (2 H, m),2.4 (2H, m), 3.22 (2 H, m), 3.62 (3 H, s), 5.42 (1 H, s, exch.),7.15 (5 H, m), 7.55br (1 H, exch.), and 8.95br (1 H, exch.),Other products were formed but were not characterised,decomposing during chromatography.We thank JEOL (U.K.) Ltd., for determining the 13C,n.m.r. spectra, and the P.C.M.U. for the exact mass deter-mination.5/2078 Received, 24th Octobev, 197518 E. Winterfeld and H. Preuss, C h m . Ber., 1966, 99, 460.9 K. Bowden, Canad. J . Chem., 1966, 44, 661