COMMUNICATIONS TO THE EDITOR
3914
lated benzylboron dichloride from these reactions or from the pyrolysis of the chloroborate. In the absence of details of their work we cannot account for the difference in products, but the isolation of 84.570 tropeniuin and recovered cycloheptatriene in our cyclohexane reaction indicates that aromatization is not a major factor under our conditions.
VOl. 83 !IXN/ CHzR
CHiR
&!\ N'
CH3
\
I1
&\CH3
I11
When R = CBH6the proton resonance spectrum in CDCla has a sharp methyl signal a t 2.37 ppm., DEPARTMENT OF CHEMISTRY KENNETH M. HARMON complex aromatics a t -7 ppni., and a t 16.5 pptn. the HARVEY MUDDCOLLEGE ANN B. HARMON broad signal from the acidic proton (the acidic proCLAREMONT, CALIFORNIA FRANK E. CUMMINCS'~ton in acetylacetone enol is a t 15.3 ppm. in CeH12a). RECEIVED JUNE 19, 1961 At 4.67 ppm. there is a field invariant doublet which collapses to a singlet upon deuteration of the A NUCLEAR MAGNETIC RESONANCE STUDY OF acidic proton and is thus ascribable to the benzyl KETO-ENOL EQUILIBRIA IN SCHIFF BASES. I1 methylene group split by N-H with J = 4.8 cps. When R = H the spectrum in either pyridine or Sir : A recent paper' presented evidence from proton CDClt is analogous to that of the benzyl base. I n magnetic resonance that bases derived from the 2: 1 CDC13 the methyl doublet is a t 3.08 ppiu. with J = condensation of a &diketone and a diamine are 4.4 cps. In both compounds the stability of the present in solution to an extent >95Y0 in the keta- keto-amine form with its chelated hydrogen-bonded mine form with only a small dependence of the equi- ring (inferred in both cases from the presence of a librium on solvent. The -NH-R structure was in- highly unshielded proton) is suflicient to destroy ferred from spin-spin splitting of the R methylene the aromatic structure of one of the naphthalene protons ( J = 6 cps.) by N-H. This study now has rings. These results present some of the most direct been extended to bases derived from other mono- physical evidence for such an effect. Previously, amines and @-diketonesin order to explore further similar tautomerism in certain arylazonaphthols the apparently high stability of the ketamine over had been inferred from ultraviolet and infrared ~ p e c t r a . ~ J .As ~ expected, salicylalbenzylamine the enol-imine tautomer. shows an unsplit methylene signal a t 4.77 ppm. in Bases formed from monoamines and simple @diketones such as acetylacetone give rise to proton CDCls. Preliminary studies of similar bases derived from resonance spectra similar to those derived from diamines. When K = H , CHI, G H s the com- 2-hydroxy-1-naphthone and 2-hydroxy-1-naphpounds in CDCla or CCld solution exist in the keta- thaldehyde indicate that the phenol-imine form predominates. When R = H in the former the mine form I as indicated by the vinyl signal a t -4.9 ppni.' and the splitting of the methyl or methylene N-methyl signal is a sharp singlet a t 3.21 ppm. in protons into a doublet with J = 5-6 cps. CDC13. The benzyl base of the aldehyde exhibits a t 4.73 ppm. in CDCll a broadened methylene resonance which persists in CC14 and pyridine. These and other bases derived from o-hydroxynaphthaldehydes and o-hydroxylnaphthones are presently under investigation and detailed results will be reported in the future. CH3 CHZR Spectra were taken on a Varian IIK-60 (at 01) I or 1-5.1Mc./sec.) or A-BO spectrometer using tetraThe benzyl base of dibenzoylmetliane behaves sitni- tnethylsilaue as an internal zero of reference. larly. No other signals attributable to the N-sub- Rases were formed from the carbonyl compound stituent were observable (concn. -0.4 M ) so that and amine by standard procedures. Compounds structure I is present to an extent >950/0. There were characterized by melting point or in the case of is no apparent solvent effect upon the equilibrium new compounds by elemental analysis. Financial (CDC13, C6He, pyridine). support from the National Science Foundation is These results, together with corroborative in- acknowledged. frared studies on the above and related compounds,2 (3) L. W. Reeve% Con. J . Chcm., 96, 1351 (1957). offer further evidence of the greater stability of the (4) A. Burawoy and A . R . Thompson, J . Chcm. SOC.,1443 (1963) (5) E. Sawicki, J . Org. Chem., 99, 743 (1957). ketamine form in aliphatic compounds as compared ( 8 ) K . J. Morgan, J . Chcm. SOC.,2151 (l96l), and references therein. to either the ketimine or enol-imine tautomers. OF CHEMISTRY Accordingly, a search was made for other systems in DEPARTMENT UNIVERSITY G. 0. DUDEK which an alternate tautomer would be of compa- HARVARD R. H. HOLM CAMBRIDGE 38, MASS. rable stability, RECEIVED AUGUST16, 1961 Derivatives of o-hydroxynaphthones were investigated. The tautomers JI and 111 are possible STRUCTURE OF a-LUMICOLCHICINE-SOME for the base derived from I-hydroxy-2-naplithone THE EXAMPLES OF DIAMAGNETIC SHIELDING BY THE and a monoamine. CARBON-OXYGEN DOUBLE BOND (1) G. 0. Dudek and R. H . Holm, J . A m . Chcm. SOC.,88, 2089 Sir: (1981). a-Lumicolchicine, which is formed with 8(2) N. H. Cromwell, F. A. Miller, A. R. Johnson, R. L. Frank and D. J. Wallace, {bid., T i , 3387 (1949). lumicolchicine and y-lurnicolchicine in the irradia(10) American Chemical Society-Petroleum Research Fund Scholar, 1900.
q;;
COMMUNICATIONS TO THE EDITOR
Sept. 20, 1961
tion of ~ o l c h i c i n e , * ~has ~ ~ 3remained a structural enigma although structures were assigned to 8lumicolchicine and ylumicolchicirie several years ago.’,’ The recent suggestion that a-lumicolchicine is I6 prompts a description of our results which are in substantial disagreement with the formulation I and which lead unambiguously to structure I1 for a-lumicolchicine. A central feature in the establishment of structure I1 was the demonstration that a-lumicolchicine is a dimer rather than a monomer.‘ OMe
OMe
3915
sodium borohydride in aqueous tetrahydrofuran gives a diol (IVa; mol. wt., 850 (Rast), 788 (Signier); 227, 27s rnp) which gives a diacetate (IVb; mol. wt., 786; 5.73 p ;228,280 mp) and a monoacetate (IVc, mol. wt., 810; 5 . 7 4 ~ ; 228,279mp; C-methyl, 5.13%, calcd. 5.34%). Oxidation of diol IVa regenerates 11. Reduction of a-lumicolchicine with one equivalent of sodium borohydride gives a ketoalcohol (Val 5.82 p7; 217, (228). 279 mp) which gives a ketoacetate (Vb, 5.74, 5.84 p’; mol. wt., 748; 217, (228), 281 mp). Oxidation of the ketoalcohol (Va) gives IT and reduction of Va gives diol IVa. Oxidation of the monoacetate (IVc) gives the ketoacetate (Vb). Theformation of a monoace-
late, a ketoalcohol and a ketoacetate can only be interpreted i n terms of a dimeric structure for a-lumicolchicine. The nuclear magnetic resonance spectra of a-
6
Me0m 0
I
I11
#OMe Me0
OMe
M e O q TT
Me0
Me
LL
lumicolchicine and its derivatives not only confirm the assigned structures but provide rare examples of diamagnetic shielding by a carbon-oxygen double bond. Theorys predicts that protons in a roughly conical area above and below the plane of a carbonoxygen double bond will experience diamagnetic shielding while protons in the vicinity of the carbonyl group but outside this conical area experience paramagnetic shielding (e.g., the paramagnetic shift of aldehydic protons). The nuclear magnetic resonance spectrum of a-lumicolchicine is remarkably similar to that of 8-lumicolchicine; the most notable differences are the absence of the 3.32 7 doublet from the olefinic hydrogen of I11 and the shift of one methoxyl resonance from 6.28 to 6.98 7. Methoxyl groups a- t o ketones usually appear in the 6.4-6.6 r region (VI, 6.60s;VIIa, 6.40s; VIIb, 6.587; VIIc, 6 . 5 7 ~ ) . The ~ resonance a t 6.98 7 is OMe
IVa. R = R ’ = H b.R=R’=Ac C.R = H; R’= AC
Va,R=H b. R=Ac
a-Lumicolchicine (11, 215, (230), 282 mp) is quantitatively converted t o &lumicolchicine (111) on heating to the melting point or heating above 100’ in solution. Conversely, a-lumicolchicine (11) can be obtained by irradiation of 111 in Pyrex vessels. a-Lumicolchicine is inert to hydrogen under conditions in which 6-lumicolchicine absorbs two equivalents of hydrogen. a-Lumicolchicine fails to give carbonyl derivatives’ but does show a strong 5 . 7 8 ~band in the infrared. The head to head dimer structure (11) is required by the facile thermal reversion t o I11 and by the nuclear magneticresonancespectra of 11, IVa,b,c, and Va,b which also require that the head to head dimer be transfused and specifically eliminate the head to tail dimer. Reduction of a-lumicolchicine with excess (1) R. Grewe and W. Wulf, Bcr., 84, 621 (1961). The a-lumicolchicine which we prepared wan identical with a sample provided by Professor R. Grewe. (2) F. Santavf, Bioi. Listy, 31, 246 (1960). (3) E. J. Forbes, J . Chcm. SOC.,3884 (1955). (4) P. D. Gardner, R. L. Brandon and G. R. Haynes, J . Am. Chrm. Soc., 79, 5334 (1957). (6) G. 0. Schenck, H.J. Kuhn and 0.-A.NeumUller, Tetrahedron Lctfcrs, No. 1, 12 (1961). ( 6 ) a-Lumicholchidne bas been reported to give monomeric molecular weight.‘
cHx R
R’
VIIa.R=R’=H b.R=Me; R ’ = H c.R=H; R’=Me
VI
beyond even the value for dimethyl ether (6.767). This shift must result from diamagnetic shielding within the molecule. The shielding cannot arise solely from the carbonyl adjacent to the methoxyl group (cf. VI and VII) and thus must involve the ketonic carbonyl in the other half of the dimer. This would not be possible in a head to tail dimer or in a cis-fused head to head dimer but is accounted for neatly by structure I1 (see stereoformula VIII). The rotating methoxyl groups a- to the ketone carbonyls must pass directly over the ketone carbonyl of the other half of the dimer. Evidence that diamagnetic shielding is responsible for the observed (7) This carbonyl absorption is a t slightly higher wave length than that of XI, but rearrangement Is precluded by the interconvenion of Va and I1 and VP and IVa and the oxidation of IVc t o Vb. ( 8 ) L. M. Jackman, “Nuclear Magnetic Resonance Spectroscopy.” Pcrgamon Press, New York, N . Y., 1959, pp. 122-124. (9) W. G. Dauben, K. Koch, 0. L. Chapman and S. L. Smith, J . Am. Chcm. Soc., 88, 1768 (1961).
e
,
VI11
0
unstable “syn” series compounds were rarbinolamines, not oximesa Recent studies in these Laboratories have established that both geometrical isotners of isonicotinaldehyde oxime may be i ~ o l a t e d . ~The syn isomer
\17
shift comes from examination of the resonance posi499 tions of the niethoxyl protons in diol IVa (normal, N 6.75 T), ketoalcohol Va (normal 6.81 and shifted 6.907) and ketoacetate Vb (normal 6.75 and shifted I1 I 6 . 8 8 ~ ) . The duality of the methoxyl resonance in V u and Vb offers unequivocal evidence of the dimeric OH I nature of these compounds. The N-H resonance in 529H N ‘c? the diol IVa is strongly shifted to low field (1.95 T ) from the N-H resonance in IT (3.28 T ) . This shift is shown by all derivatives of I1 which possess an hydroxyl group and is the result of hydrogen bonding between the acetamido nitrogen and the nearest CHj I CH3 1 hydroxyl group. This is possible only if these IV I11 groups are cis- thus defining the stereochemistry of Fig. 1.-Xuclear magnetic resonance peaks (c.P.s.) of IVa,b,c and Va,b. The duality of the N-H reso- isonicotinaldehyde vxirries and 4-formyl-1-methylpyrinance in the monoacetate (1.76, 3.37 T ) , the keto- dinium iodide oximes from spectra obtained by Varian Asalcohol (1.95, 3.60 T ) and the ketoacetate (3.15, sociates in deuterium oxide at 60 Mc. The ring proton 3.86 7) emphasizes again the dimeric nature of these resonances refer to doublets. Tetrarnethylsilane was used as internal reference at 0 with respect to observed resocompounds. nance. Acknowledgment.-The authors acknowledge financial support of this work by research grant of isonicotinaldehyde oxime, I , n1.p. 132-133’, was (CY-4253) from the Cancer Division, National reported previously.6 The other geometrical isomer, 11, m.p. 165--167’ (Anal. Found for C ~ H G N ~ O : Institutes of Health, Public Health Service. C, 58.8; H, 4.9; 0, 13.3) was isolated in yields of DEPARTMENT O F CHEMISTRY less than 5% from isonicotinaldehyde and hydroxylIOWASTATE UNIVERSITY 0. L. CHAPMAN AMES,IOWA H. G. SMITH amine in basic aqueous media a t 10-15’ by fractional crystallization from the syn isomer. This RECEIVED AUGUST2, 1961 procedure was not satisfactory because of inconsistent reproducibility, but a better method was not CONFIGURATIONAL ANALYSIS OF 4-FORMYL-1- found. Alkylation of I and I1 with methyl iodide METHYLPYRIDINIUM IODIDE OXIMES AND ITS yielded isomeric 4-formyl-1-methylpyridiniumioRELATIONSHIP TO A MOLECULAR dide oximes, 111 (m.p. 182-183’; reported6 181COMPLEMENTARITY THEORY ON THE 183’, p K a 8.6) and IV, m.p. 169-172’ ( A n a l . Found REACTIVATION OF INHIBITED ACETYLCHOLINESTERASE for C7HSIN20:C, 32.3; H , 3 . 3 ; O,6.3; neut. equiv., Sir: 246; pKa 9.0), respectively. On comparing the nuclear magnetic resonance has offered the hypothesis that chemstherapeutic activity of 2-formyl-1-methylpyridin- spectra of I and I1 it was noted that the signal from ium iodide oxime in the treatment of “nerve gas” the hydrogen on the oximino carbon was shifted poisoning depends both on its ability to associate chemically to lower field in the spectrum of I. This with inhibited enzyme a t the site of phosphoryla- suggests by analogy with propionaldoxime,6 that the tion and a proper orientation of the reactive oxi- hydrogen is syn to the oxime oxygen (Fig. 1). mino function for enzyme reactivation by nucleo- Additional strength for this assignment is found in philic displacement of the phosphate grouping. The the signals from the ring protons ortho to the oxianti configuration in the 2(and 4)-formyl-l-methyl- mino function which are chemically shifted further pyridinium oximes satisfied geometrical disposi(3) E. J. Posiomek, D. N. Kramer, B. W. Fromm and W. A tions defined for the displacement reaction2; sup- Mosher, J . Org. Chem., 26, 423 (1961). (4) E. J. Poziomek, 13. N.Kramer and W. A. Mosher, Abstr. 138th port for the contribution of geometric conforma- Meeting, Am. Chem. SOC.,1960. tion regarding the reactivation rates of inhibited (5) S . Cinsburg and 1. B. Wilson, J . A m . Chem. Soc.. 79, 481 (19ri7). acetylcholinesterase was obtained from observa(6) W . D. Phillips, A n n . N . Y . Acad. Sci., 70, 817 (1058). Cited tions of enhanced activity of the “oximes” of the by J. A. Pople, W. G.Schreiner and H. J. Berstein, ”High-resolution Nuclear Magnetic Resonance,” McCraw-Hill Book Co., Inc., New “anti” configuration .z Only 2-formyl-1-methyl- York, N. Y., 1959, p. 374. T o check Phillips’ assignments. of the pyridinium iodide oxime was claimed sufficiently two -CH=NON multiplets arising from syn-anfi isomerism, Lustigab stable in the “syn” series for study and was not examined the spectra of the two p-chlorobenzaldoximes in dimethyl found active in reactivating inhibited acetylcholine- sulfoxide solution. The -CH=NOH resonance of the syn oxime appeared at lower field and Phillips’ assignmenth is substantiated. sterase.2 Subsequently, it was discovered that the The structures of both isomers of p-chlorobenzaldoxime had been de-
CL6
(1) I . B . Wilson and S. Ginsburg, Biochim. cf Bioghys. Acto, 18, 168 (1955). (2) I . B . Wilson, Federation Proc., 18, 752
(1959).
termined previously through X-ray dieraction studies.60 (b) E. Lustig. J . Phys. Chem., 66, 491 (1961). ( c ) B. Jerslev, Nature, 188, 741 (1950); 180, 1410 (1958).
