摘要:
Org. 155'9 Spirans. Part W1.l Three Isomeric Grisan Derivatives from the Oxidative Cyclisation of Benzylidenebis-2-naphthol By D. 9. Bennett, F. M. Dean," and A. W. Price, The Robert Robinson Laboratories, University of Liverpool, Liverpool L69 3BX N.m.r. studies on the oxidation products [spirans (A) and ( B ) ] of benzylidenebis-2-naphthol show that hypo- bromite gives (2R",3S") -3-phenyl-4,5 : 2',3'-dibenzogrisa-2',4'-dien-6'-one (11) and not the (2R*,3R") stereo- isomer reported formerly; the latter compound was obtained from an oxidation with plumbic acetate in pyridine. A marked difference in the rates at which the related allylic alcohols (A) and (B) are converted by skeletal re- arrangement into 1.2 : 5.6-dibenzo-9-phenylxanthen has been confirmed and found to be greater for the acetates.It is now clear that there is steric compression in both series, and the difference is now ascribed to the fact that in the reaction in series (B) the relief from compression comes much earlier than in series (A). A third isomer, spiran (C), has been encountered and identified analytically as 3-phenyl-4,5 : 4',5'-dibenzogrisa- 2',4'-dien-6'-one (XVI), a conclusion confirmed by a synthesis of the dimethanesulphonate of the zinc reduction product, (1 -hydroxy-2-naphthyi) -(2-hydroxy-1 -naphthyl)phenylmethane (XVII). The origin of this isomer has been identified tentatively. OXIDATIVE cyclisation of the dinaphthol (I) gives the (racemic) grisan derivative (11) or the stereoisomer (111) depending on the reagent and conditions used.2 Hypo- removes this shielding so that the resonance moves downfield to T 4.22.Systematically, spiran A (11) is (2R *, 3s *)-3-phenyl-4,5: 2', 3'-dibenzogrisa-Y,4'-dien- 6'-one, while spiran (B) is the (2R*,3R*)-isomer. bromite oxidation affords spiran (A), while plumbic acetate in pyridine affords spiran (B), and the problem has been to assign to each the appropriate structure. Earlier, this could not be done directly because of the low solubilities in solvents suitable for n.m.r. spectro- scopy, but the instrumentation now available has re- moved that difficulty and made it clear that our original assignments2 were wrong, spiran A being, in fact, the so-called cis-isomer (11) and spiran (B) the trans t (111). A third isomer, spiran C, has also been encountered and is considered to result from a skeletal rearrangement occur- ring only in conditions suit able for hydroxylation followed by a Wagner shift.In the grisan derivative (IV) the proton HB resonates amongst the aromatic protons and so cannot be identified readily, but H, appears as a doublet ( J 10 Hz) centred at T 3-84. Spiran (A) behaves similarly, proton H, again appearing at 7: 3-84, but spiran (B) is different, proton HB being still overlaid but H, appearing at the much higher field T 4.57. Hence H, in spiran (B) is shielded and since the phenyl substituent must be responsible this spiran has structure (111). Independent support comes from the beliaviour of the methine protons Hp. In spiran (13) (111) this proton resonates at T 4.80 since it lies alongside ring E and is shielded by it. Borohydride reduction of the carbonyl group hardly affects this situation, and H, in the corresponding alcohol (V) resonates at T 4.92.On the other hand, in spiran (A) (11) proton H, appears at T 4.70 since it is shielded by the carbonyl group, but reduction to the allylic alcohol (V) =f These terms define the geometry about ring c. They relate the phenyl substituent to ring E. (Yl (YI) v Because the allylic alcohols (V) are more soluble in the requisite solvents than the parent ketones, they were the subjects of the orientational studies in the earlier work,2 the methods used including an examination of the hydroxy-stretching frequencies near 3p, an interpreta- tion of the different fields at which the methyl groups of the derived acetates resonated, and kinetic studies of the acid-catalysed rearrangement of the alcohols to the (same) xanthen derivative (VI).Since the first two methods were never very secure on account of the com- plexity of the situation, we propose to disregard them; however, the kinetic study seemed convincing and we have, therefore, conducted a new examination. In structure (11) the conformation about ring D must be such as to avoid too close an approach of ring E to the phenyl substituent . The only reasonable conformation is shown in diagram (VII) ; in it, the dihydrofuran oxygen atom is pseudoequatorial, and since this side of ring D is much more accessible, the rapid attack by borohydride reduction would give an allylic alcohol of structure (VIII), the conformation being such as to make Hr pseudoaxial.The n.m.r. spectrum of allylic alcohol (A} 1 Part VI. F. M. Dean, Colin Fletcher, and H. D. LocksIey, 2 F. M. Dean and H. D. Locksley, J . Chew. Soc., 1963, 393. J . Chew. SOC., 1964, 5096.1558 J. Chem. SOC. (C), I970 A change from the conformation shown in (X) to that shown in (XIII) (dihydrofuran oxygen pseudoaxial) would reduce the steric compression in allylic alcohol (B) somewhat as it would increase the perpendicular distance from the (hydroxy) oxygen atom to the phenyl substi- tuent to 2-3 A and the n.m.r. spectrum agrees with this conformation since JG,, is now ca. 1-5-2 Hz consistent with an allylic angle of approximately 40", while JBr is ca. 3.7 Hz, consistent with a dihedral angle of ca. 30". A full analysis of the acetate gave for H,, Hp, and FI,, z values of 3.41, 4.02, and 4.50, respectively, and Jab 10, Jay -1.75, and JBy 4.5 Hz.In acetic acid, the acid-catalysed conversion of allylic alcohol (B) into xanthene derivative (XI) is ca. 70 times as fast as the conversion of (A).2 With the acetates the difference is even larger, (B) reacting ca. 250 times as fast as (A). As shown before,2 alcohol (B) seems t o be con- verted directly whereas alcohol (A) is first converted into the acetate, no doubt via the carbonium ion (XIV). Thus the difference in rates seems to be a consequence of the relative stability of this carbonium ion even though the compression between the phenyl substituent and ring E would appear to make this the less stable arrange- ment. However, Wagner migration in ion (XIV) at first causes the phenyl substituent to swing from a con- fined position alongside ring E to an equally confined position ring D, and there can be no relief from the associated compression until the migration is nearly com- plete.In contrast, as soon as migration begins in ion (XV) of the (B) series, the phenyl substituent swings into sparsely occupied space and the immediate relief from compression could well be responsible for the en- hanced migration rate. agree with this structure; Jar = ca. -3 Hz, consistent with an ' allylic ' angle near to 90"; and Jp,, = ca. 1.7 Hz, consistent with a dihedral angle near 90". The derived acetate exhibited very similar resonances, full analysis by comput er techniques defining resonances for Ha, Hp, and H, at T 3.55, 4-10, and 3-56 respectively, with JaO 10-75, Ja,, -3-0, and Jay 0.5 Hz.For structure (111) one conformation has the hetero- cyclic oxygen atom in the pseudoaxial position about ring D, a consequence of which is that the carbonyl oxygen atom is thrust towards the phenyl substituent (the per- pendicular distance to the plane of the phenyl group is 2.1 A). Probably the alternative conformation is adopted, as shown in diagram (IX). Again occurring at the more approachable side, borohydride reduction H Ph H i-! Ph would now give an allylic alcohol of structure (X) in which the oxygen atom of the hydroxy-group is again thrust towards the phenyl substituent (perpendicular distance 2-2 A). The unfavourable nature of this stereochemistry must oppose any reaction in which the sp2 hybridisation of the carbonyl carbon changes to sfi3 and is manifest in the relative slowness of the borohydride reduction of spiran (B) to allylic alcohol (B), which allows time for appreciable reduction of the ethylenic bond and so for the appearance of the saturated alcohol (XI).It also accounts simply for the catalytic hydrogenation of spir-an (Bj to ketone (XII) in contrast to the formation from spiran (A) of the corresponding saturated alcohol.2 Again, it accounts for the formation of a 2,4-dinitro- phenylhydrazone by spiran (A) but not by spiran (B), the transition state requiring sp3 hybridisation at carbon. (XXm) R = H (XXIYI R = S02Me A third isomer, spiran (C), has sometimes been en- countered amongst the oxidation products of the di- naphthol (I) in yields of up to ca.8%. Spirans (A) and 3 S. Sternhell, Quart. Rev., 1969, 23, 236.Org. 1559 (B) exhaust the possibilities of stereoisomerism for their skeleton, so spiran (C) at once is recognisable as a re- arrangement product ; it has structure (XVI). The carbonyl groups absorbs at 1690 cm-1, at 13 cm.-l higher a frequency than either spiran (A) or spiran (B). The cis-ethylenic link is evidenced by two one-prot on doub- lets (J 10 Hz) centred at z 3.75 (Ha) and 4.25 (HB) These resonances are at higher fields than might have been expected, and we suggest, very tentatively, that the phenyl substituent lies close enough to shield them which is possible only for the stereochemistry indicated. (The other stereoisomer has not been encountered.) A one- proton singlet at T 4-88 corresponds to the methine proton and the other 15 protons, all aromatic, give a complex series of multiplets in the region 24--3*3.Chemical evidence for structure (XVI) depends upon the reduction by zinc in acetic acid to the dinaphthol (XVII) isomeric with the original (I). Unlike that, the new compound gives a strong positive test with Gibbs’ reagent showing it to be a phenol with a free para- position but, curiously, efforts to induce dehydration and cyclisation to the xanthen derivative (VI) were unsuccessful though the corresponding cyclisation of (I) is very easy. In the present case complex, highly coloured mixtures resulted. Indeed, since the di- naphthol seemed sensitive even to air, it was converted ir,to the stable dimesylate.Catalysed by boron fluoride, the isomerisation of 1-naphthyl benzoate gave ketone (XVIII) which was converted into the ester (XIX) by methanesulphonyl chloride and triethylamine as base. Pyridine was not an effective base, while sodium hydride caused the further cyclisation to the ‘ aldol ’ (XX), dehydration then giving the cyclic sulphonate (XXI). Reduction of the mesylate (XIX) with borohydride led to alcohol (XXII) which, condensed with 2-naphthol in the presence of phosphoric acid, afforded the monomesylate (XXIII) ; finally a dimesylate (XXIV) was obtained identical with the product from spiran (C). We explain the formation of spiran (C) as follows. The compound has been observed only when the oxidis- ing agent has a known capacity for hydroxylation or the < (Xxvrnl IXXVII) equivalent (mainly plumbic acetate, periodic acid, and related reagents) and the medium is weakly acidic (strongly acidic media cannot be used as they cyclise the dinaphthol to the xanthen too quickly).The first step can be represented therefore as insertion of a group OX where X is hydrogen or some group characteristic of the reagent. This gives (XXV), which can cyclise reversibly to the hemiacetal (XXVI). The acidic medium ensures the appearance of carbonium ion (XXVII) and the associ- ated Wagner shift indicated will then give a new carbon- ium ion (XXVIII) from which loss of X+ yields spiran (C> (XVI). EXPERIMENTAL M.p.s were determined on a Kofler block following the discovery that many compounds in this series are affected by the grinding necessary for the capillary method, especi- ally the acetate of the allylic alcohol (A) (VIII) ; generally, the m.p.’s quoted are higher than those given previously.N.m.r. spectra were determined on solutions in deuterio- chloroform. Molecular weights were determined by mass spectro- metry. (2R*, 3S*) -3-Phenyl-4,5: 2’, 3’-dibenzogrisa-2‘,4’-dien- 6’-one (11) [Spiran (A)].-This stereoisomer was formed by hypobromite oxidation and had m.p. 253-257’. Boro- hydride reduction gave (2R*,3S*, 6’R*)-3-phenyl-4,5:2’,3’- dibenzogrisa-2’,4’-dien-B’-01 [allylic alcohol (A)] (VIII) with m.p. 224-226’; Amx. (in acetic acid) 262infl., 270, 281, 293, 332, 343infl., and 346 nm. (log E 4.96, 4.99, 4.96, 4-75, 4.45, 4.48, 4.50). The derived acetate (acetic anhydride- pyridine) had m.p.185-186” but if crushed had m.p. 203- 204”: the i.r. spectra of samples of different m.p. were the same when examined in carbon tetrachloride. (2R*,3R*)-3-Phenyl-4,5:2‘, 3‘-dibenzogrisa-2’,4‘-dien-6‘-one (111) [Spiran (B)].-Prepared by oxidations with plumbic acetate in acetic acid,2 this compound had m.p. 265-266”. Treated with an excess of sodium borohydride during 1 hr., spiran (B) (0.5 g.) in tetrahydrofuran (15 ml.) and water (1.5 ml.) gave (2R*,3R*,6’RRs)-3-phenyl-2’,3’:4,5- dibenzogrisa-2’,4’-dien-B’-01 [allylic alcohol (B)] (XIII) which crystallised from ethanol as needles (0.1 g.), m.p. 226-228” ; A,, (acetic acid) 270, 278infl., 293, 331, 342infl., and 345 nm. (log E 5-08, 5.02, 4.82, 4.50, 4.51, 4.53): the acetate had m.p.206-209”. The mother liquors from the allylic alcohol (B) were concentrated and the residue was chromatographed on silica by the thick-layer technique with benzene-chloroform (1 : 1) as mobile phase. The slower running band gave further quantities of the allylic alcohol (B), the faster supplied (2R*, 311*,6’Ii*)-3-phenyl- 4,5: 2’, 3’-dibenzogrisa-2’-en- 6’-01 (XI) which separated from ethanol as needles (0.1 g.), m.p. 235-238’; a,, (acetic acid) 259infl., 271, 276, 283, 295, 332, 341infl., and 346 nm. (log E 4.70, 4-70, 4.67, 4.71, 4-57, 4.44, 4.46, 4-50). In the mass spectrometer, this alcohol exhibited a parent ion at m/e 378 (calc. for C,,H2,O,: M , 378). The com- pound was identical with a sample prepared by an alterna- tive route.2 Action of Hydvoclzlovic Acid o n (2R*,3S*,B’R*)-3-Phenyl- 4,5: 2’, 3’-dibenzogrisa-2’, 4‘-dien-6’-ol.-Treated with con- centrated hydrochloric acid (0.02 m1.) at 100’ for 6 hr., a mixture of the title alcohol (0.19 g.), acetic acid ( 5 ml.), and water (0.05 ml.) turned red; a red gum was left on con- centration of the mixture.This gum contained the alcohol, its acetate, and 9-phenyl-l,2: 5,6-dibenzoxanthen (VI) . The acetate and the xanthen co-crystallised from ethanol1560 J. Chem. SOC. (C), 1970 giving a product (0.12 g.) which was subjected to prepara- tive t.1.c. on silica. Obtained thus, the acetate crystallised from ethanol as needles (0.09 g.), m.p. 187-188". Since the m.p. varies with the treatment of the sample, the material was identified by i.r. spectroscopy with a sample prepared by the method above, and the earlier claim t o the isolation of a syn acetate 2 should be disregarded.Kinetic Study.-The rates of conversion of the acetates of allylic alcohols (A) and (B), (VIII) and (X), was studied by the procedure already described for the alcohols them- selves.2 The product was 1,2: 5,6-dibenzo-9-phenylxan- then (VI) in each case, and no other product was detected except that, after several days, a pink colour indicated slight oxidation to a xanthylium salt. The rates were determined at 20" in acetic acid by follow- ing the change in optical density at 306 nm. of a solution of each acetate (1-121 x 10-4 g. mole/l.) and sulphuric acid (30 ml./l.). The rate constant for the acetate (A) was 3.2 x 10-6 min.-l; and for the acetate (B) was 843 x min.-l.(2R*, 3R*)-3-Phenyl-4,5: 3', 4'-dibenxogrisa-3', 5'-dien-2'-one [Spiran (C)] (XVI) .-Bis-( 2-hydroxy-1-naphthy1)phenyl- methane (1 g.) in acetic acid (75 ml.) was treated with finely powdered plumbic acetate (1.2 g.). After 24 hr. the mixture was poured into water; the solid product was dried in air, and chromatographed on silica. Benzene- light petroleum (b.p. 60-80") (1 : 1 v/v) first eluted the spiran (C) and then a mixture (0.29 g.) of spirans (A) and (B) . The spiran (C) separated from methanol-chloroform as faintly yellow needles (0.015 g.), m.p. 214-217"; A,,, (ethanol) 269, 280, 292, 329, and 342 nm. (log 4.93, 4-89, 4.74, 4-60, 4.63) (Found: C, 85.9; H, 5.4%; M , 374-130092. C,,H,,O, requires C, 86.6; H, 4.85%; M , 374.130672). Numerous experiments were carried out in an effort to increase the yield of the spiran (C), but that recorded above was the best that could be obtained regularly.Addition of mineral acids to the mixture caused too rapid a formation of the xanthen derivative (VI) for oxidation to be effective, while after additions of base (sodium acetate, pyridine) the spiran (C) could not be detected though (A) and (B) were produced normally. Experiments with periodic acid and other oxidising agents will be reported separately. 2-Benzoyl-1-naphthyl MethanesuZphonate (XIX) .- 2-Benzoyl- 1-naphthol was prepared as before and purified from light petroleum (b.p. 60-80") whereafter it formed large, golden yellow crystals m.p. 76". The naphthol (6.7 g.) and triethylamine (7 ml.) in ether (50 ml.) were stirred for 4 hr.with methanesulphonyl chloride (10 g.). The solution was diluted with ether and washed successively with dilute hydrochloric acid and water and dried (Na,SO,) ; the solvent was evaporated off to leave a solid which, purified from propan-2-01, gave 2-benzoyl-1-naphthyZ methanesdphonate as needles (6-4 g.), m.p. 129-131.5", devoid of hydroxylic absorption in the i.r. region but absorbing at 1665 cm.-l (diaryl ketone) and resonating at z 6.99 (OS0,Me) (Found: 66.1; H, 4.2; S, 19.0. C,,H,,O,S requires C, 66-2; H, 4.3; S, 19.1%). Methanesulphonate (XXII).-The above ketone ( 5 g . ) in tetrahydrofuran (50 ml.) and ethanol (50 ml.) was reduced at 0" by the addition in small quantities of sodium borohydride (0.5 g.). After 3 hr. the excess of the reagent was destroyed by acetic acid and the solution was poured into water and extracted with dichloromethane.Isolated in the usual way and purified from propan-2-01, the product furnished the Ynethanesulphon- ate as needles (4.5 g.), m.p. 118-119" showing no carbonyl 2- (cc-Hydvoxy benzyl) - 1-naphthy 1 absorption but a hydroxylic band near 3300 cm.-l (Found: C, 65.8; H, 5.1; S, 10.0. CI8H,,O4S requires C, 65.9; H, 4.9; S, 9.8%); z 6.70 (OSO,Me), 6-95br (OH), and 3.54 (Ar,CNOH) . 2-(2-Hydvoxy-l-naphthyZ~henylmethyZ)-l-naphthyZ nleth- anesulphonate (XXIII) .-The foregoing compound (0.66 g.) was condensed with 2-naphthol (0.30 g.) in refluxing acetic acid (30 ml.) containing phosphoric acid (0.3 ml.) during 24 hr.; the mixture was poured into water and the product was extracted into ether ; it was purified by chromatography on silica (40 g.) from chloroform.The main eluate gave a gum which, crystallised from ethanol, supplied the waethane- sulphonate as prisms (0.4 g.), m.p. 201-203" (decomp.) (Found: C, 73.8; H, 5.1; S, 7.3%; M , 454. CZ8H2,SO4 requires C, 74.0; H, 4.8; S, 7.05%; M , 454). The methyl group resonated at z 5.40 and the metliine proton at 3-22. phonyloxy- l-naphthyl)#henyZmethane (XXIV) .-(i) The above phenolic methanesulphonate (1 g.) was esterified by methanesulphonyl chloride (1.5 g.) in pyridine (10 ml.) during 3 hr. at 22" ; the mixture was poured into water and extracted with ether. The product was recovered in the usual way and purified on a column of silica with benzene as eluant ; the naphthylphenyZmetharte crystallised from ethanol as small prisms (0.55 g.), m.p.157-158" (Found: C, 65.7; H, 4.8; S, 12.4%; M , 532. C,,H,,O,S, requires C, 65.4; H, 4.5; S, 12.0%; M , 532). (ii) Zinc dust was gradually sifted into a solution of the spiran (C) (0.1 g.) in acetic acid (3 ml.) and chloroform (1.5 ml.) until t.1.c. indicated completeness of reaction. The mixture was poured into water and the product was collected into chloroform. The extract was dried (Na,SO,) and the solvent was removed; the product was purified from chloroform on a column of alumina (7.5 g. ; Grade 11) giving ( l-hydroxy-2-naphthyl)-( 2-hydroxy- 1-naphthy1)- phenylmethane (XVII) as a glass (0-065 g.) showing no carbonyl absoption in the i.r. region and only aromatic and hydroxylic bands in the n.m.r. spectrum.This material (0.15 g.) became coloured when kept and was at once esteri- fied with methanesulphonyl chloride (0.20 g.) in pyridine ( 5 ml.) by the technique described in (i); it gave the naphthylmethane (XXIV) (0.10 g.), m.p. and mixed m.p. 157-158" (Found: C, 65.4; H, 4.7; S, 12.2%). Further identification was effected by i.r. and mass spectroscopy. S-Phenylna;bhtho[2,1-e]-l,2-oxathiin 2,2-Dioxide (XXI) .- To 8-benzoyl-1-naphthol (2-5 g.) in dimethylformamide (distilled from calcium hydride; 50 ml.) was added sodium hydride (dispersion in oil, 0-5 g.). After being stirred at 23" for 2 hr., the mixture was warmed to 70" for 45 min. to complete the formation of a deep red solution; the mixture was then cooled. Methanesulphonyl chloride (0.80 ml.) in dimethylformamide (10 ml.) was added to the mixture; the colour soon faded. After 35 min. the solution was greatly diluted with water and the organic materials were extracted into dichloromethane and purified by chroma- tography on silica from benzene. Crystallised from propan- 2-01, the product gave 3,4-dilzydro-4-hydroxy-4-phenyl- naphtho[Z, l-e]-1,2-oxathiin 2,2-dioxide (XX) as needles (2.0 g.), m.p. 173", absorbing strongly at ca. 3500 cm.-l but not at all in the carbonyl region (Found: C, 66.0; H, 4.6; S, 9.9%; M , 326. C,,H,,O,S requires C, 66.2; H, 4-3; S, 9.8%; M , 326). The n.m.r. spectrum showed a band at z 5.42 (OH) removed by deuteriation and an AB quartet centred at 6.01 (CH,) for which first order analysis gave ( 1-DimethylsuZ;bhonyloxy - 2-naphthyl) - (2'-dimethyls.ul- HA 5.95, HB 6.24, JAB 15 Hz.Org . 1561 This compound (2.2 g.) was kept at 100' in acetic acid (40 ml.) containing concentrated hydrochloric acid (0-4 ml.) for 4 hr. Precipitated by adding water and crystallised (:CHSO,.). from ethanol, 3-phenylna~htlzo[2,1-e]-1,2-omthiiuz 2,2-di- hydroxylic absorption (Found: C, 69.7; H, 4.0; S, 10.7%; 'pectra Of -44, 308. C,,H,,O,S requires C, 70.1 : H, 3.9; S, 10.4% ; M , 308). Apart from resonances due to aromatic protons, the n.m.r. spectrum exhibited only a sharp peak at T 3.23 oxide formed prisms (1.4 g.), m.p. 192-1930, showing no we thank Dr+ A. Thomas for the Of the ll.m*r. (*) and (B)* [9/2225 Received December 31st, 19691
ISSN:0022-4952
DOI:10.1039/J39700001557
出版商:RSC
年代:1970
数据来源: RSC