Thc reactions leading to the establishment of structure I for lycopodine are discussed. S t r u c t~~r e s are proposed for three cyclizatioll products derived from lycopodine.I n a short comlnunication (1) we proposed structure I for lycopodine. This structure has been confirmed by the worlt of Anet (2), who succeeded in interconverting several Lycopodiz~m alkaloids, among which was lycopodine. In this paper we wish to give a full account of our worlt on the elucidation of the structure of the alltaloid and t o discuss the structures of the tetracylic compounds obtained on elimination of hydrogen halide from lycopodine methiodide and a-and P-cyanobromolycopodine.

Acid hydrolysis of 5,6-dihydro-2-methyl-I,4-oxathiin-3-carboxanilide gives 2-(2-hydroxyethy1thio)-acetoacetanilide enol. 3-Carbonyl-substituted 5.6-dihydro-l,4-oxathiins were found to undergo ring cleavage by nucleophilic nitrogen attack on C-2. Thus the following reactions were observed: 3-acetyl-5,6-dihydro-2-methyl-1,4-oxathiin o n treatment with hydrazine gives 4-(2-hydroxyethylthi0)-3,5-dimethylpyrazole, instead of the hydrazone. The 1,4-oxathiins, N-(5,6-dihydro-2-methyl-1,4-oxathiin-3-ylcarbony1)-N'-phenylurea and 5,6-dihydro-2-methyl-1,4-oxathiin-3-carboxylic acid hydrazide rearrange to give 5-(2-hydroxyethy1thio)-6-methyl-I-phenylracil and 4-(2-hydroxyethy1thio)-3-methyl-2-pyrazolin-5-one, respectively. Finally, treatment of 5,6-dihydro-2-methyl-1,4-oxathiin-3-carbonyl chloride with 2-aminopyridine and with 2-aminopyrimidine affords 4H-3-(2-hydroxyethy1thio)-2-methylpyrido[l,2-nlpyrimidin-4-one and 4H-3-(2-hydroxyethylthio)-2-methylpyrimido[l,2-a]pyrimidin-4-one, respectively, a s the predominant products.

No abstract

Evidence is furnished which establishes the structure of a new allcaloid isolated from Lyro)odiz~nl flabellifornze.In the present publication we wish to report evidence which enables the assignment of structure I to a new allialoid obtained froin Lyco;bodiz~mJEabelliforme.The alkaloid, for ivhich the name flabellifornline seenls appropriate, is sparingly soluble in cold ether. Advantage was taken of this property in its isolatioil from the bases obtained, after the separation of lycopodine from the crude alkaloid mixture.Analysis of flabellifonnine, I , fits the formula C161-12j02N. The infrared spectrunl of the allialoid, measured in chloroform solution, shows lietonic carbonyl absorption a t 1705 cm-I and a hydroxyl band a t 3560 cm-l. Reduction of the carbonyl group with sodium borohydride affords the diol 11, C16H270?N. T h e hydroxyl group appears to be attached to a tertiary carbon atom, since flabelliforinine is recovered unchanged after treatment with chromium trioxide and its i1.m.r. spectrunl shows no absorption below a T-value of 6.5.Treatment of the allialoid with hydrogen iodide replaces the hydroxyl group with hydrogen; the reaction product is identical with lycopodine, I11 (1-4). Flabelliformine, therefore, is formally derived from lycopodine by replacement of a methine hydrogen by a hydroxyl group.Dehydration ol I , effected by both phosphoric acid and ;b-toluenes~~lphonic acid, affords a non-crystalline product, IV, which was converted to a crystalline methiodide, V, C17f1260NI. Compound IV has the properties of an a,p-unsaturated ketone. The ultraviolet spectrum of this substance shows I~~X~I~L I I~ absorption a t 245 mp and the infrared spectrum, measured in carbon tetrachloride solution, has bands of equal intensity a t 1685 and 1617 cm-l. The i1.m.r. spectrum of IV has a triplet a t a T-value of 3.01 which is attributed to a proton on the p-carbon of the a,p-unsaturated carboilyl system ( 5 , 6).

Annotine, C~J I~I O~N , is shown to be pentacyclic and to contain a tertiary hydroxyl group, a lactone function, a tertiary nitrogen atom, and a dialkylated double bond. The position of the double bond and the tertiary hydroxyl group relative t o the nitrogen atom has been established by Emde degradation of annotine methiodide. The presence of a lactone function is inferred from the reduction of annotine to dihydroannotinol, a hemiacetal, which reacts with 1 mole of ethyl mercaptan. T h e reduction of the lactone to a diol in a n annotine derivative has been carried out. T h e chemical studies and the examination of annotine and its derivatives by modern instrumental methods allow the assignment of a plausible structure to the alkaloid.Annotine, C16H2103N (I), was first isolated from L. annotinz~m of Canadian origin by hiIanslte and Marion and designated alkaloid L . l l (1). I t was subsequently found in European plant material by Bertho and Stoll (2), who assigned the name annotine t o it, and by Achmatowicz and Rodewald (3). An examination of the functional groups of annotine was carried out by Perry and MacLean (4), who concluded that the alltaloid contained a tertiary nitrogen atom, a double bond, a hydroxyl group, a ketonic carbonyl group, and a third inert oxygen function, which they ascribed t o a n ether linkage other than an epoxide.In this coinmunication we wish t o report further studies on annotine by chemical and modern instrumental methods. These studies have shown that annotine is a lactone and have thereby established the nature of the third inert oxygen function. They have also allowed a n assignment of the position of the double bond and the hydroxyl group in the n~olecule relative t o the nitrogen atom and have led t o the development of a plausible structure for the molecule.Nuclear magnetic resonance (n.m.r.) spectroscopy has been used extensively in this study and has provided useful information about the functional groups and their environment in annotine and its derivatives. The n.m.r. spectrum of annotine has a quartet centered a t T = 4.04 (J = 10 c.p.s.) with a n intensity corresponding to two protons. This H\ band indicates the presence of the grouping /C=C\ and is an example of an intermediate case of an AB spectrum in which the chemical shift between the two nuclei is comparable to the coupling constant between them (5). I t was possible t o resolve each component of the quartet into a triplet in each of which the splitting was the same. These data indicate the presence of a methylene group adjacent to the double bond, as in theA singlet occurring a t T = 8.57 with an intensity corresponding to three protons points to the presence of a methyl group attached t o a quaternary carbon. The low T value of the C-methyl group in this system can possibly be explained by the deshielding effect of a carbonyl group beta t o the methyl protons. I t has been postulated that the deshielding effect in such a system is due to the anisotropy of the carbonyl group and that the deshielding e...

The mass spectra of annotine and some of its derivatives are recorded and discussed. Fragmentation mechanisms are proposed to account for the formation of the major pealis in the spectra. T h e composition of the ions has been verified by measurement of the high-resolution spectra of four of the five compounds. The results lend support t o the structure previously proposed for this allraloid.In a recent publication (I) it was proposed that annotine (2-5) has structure I. This proposal was based upon a number of degradation experin~ents, upon the examination of annotine and some of its derivatives by infrared and nuclear magnetic resonance spectroscopy, and upon biogenetic considerations. In this con~lllunication we report on the mass spectrometric examination of annotine and some of its derivatives.The mass spectra of a representative group of Lycopodiunz alltaloids were reported several years ago (6); they sho\ved that the loss of the bridge carbon atoms was characteristic of all alkaloids examined. Since then mass spectrometric studies have been used as an aid in the elucidation of the structure of several ne\v allcaloids of this family, and in all cases this behavior has again been observed. Among other allcaloids, lycodoline (6,7), des-N-inethyl-hydroxy-a-obscurine (8), and lycofawcine (9), which carry a hydroxyl group a t C4 in the hydrojulolidine system, as has been proposed for annotine, \\-ere investigated. In the spectra of the first t\vo of these colnpounds a large PI-17 peak was observed as a result of loss of a hydroxyl radical. The All-17 ion so formed then under-\.\rent further fragmentation with loss of the bridge carbon atoms. Lycofawcine and its derivatives do not exhibit a prominent AfI-17 pealc, but the Cq hydroxyl is eliminated along with the bridge carbon atoms. One might expect that, if annotine has structure I, its fragmentation urould follo\\i a pattern sinlilar to that of other allcaloids bearing a C4-OEI group. This expectation was borne out, as the results reported below indicate.The major peaks in the mass spectrum of annotine, shown in Fig. l a , may be explained by involting two fragmentation processes. In one process the first step involves loss of the hydroxyl group, giving rise to an ion of mass 258. A metastable peak a t nz/e 242.0 is in accord with this change; the value of m* calculated for the transforination 275 -+ 258. is 242.0. The pealis observed a t m/e 172 and 144 may be considered to arise fro111 the ion of mass 258 as shown in Scheme la. A metastable pealc associated \\.it11 the change 258 -+ 172 is observed a t nz/e 115.0 (m* calcd. = 114.7). The last step, m/e 172 -+ 144, is analogous to the m/e 174 -+ 146 transfor~nation encountered in the fragmentatioil of dihydrolycopodine and other Lycopodiztnz allialoids (6).

Treatment of hydrastine with p-nitrophenylchloroformate in the presence of diethylisopropylamine resulted in opening of the isoquinoline ring with the formation of an ene-lactone carbamate. The latter compound rearranged upon treatment with methanolic sodium methoxide to an indan-1,3-dione carbamate which gave, on basic hydrolysis, the corresponding amine. This was ring closed to a spirobenzylisoquinoline. In addition, acid-catalysed dehydration of the amino indandione led to the formation of two isomeric benzazepines, synthetic precursors of the rhoeadine alkaloids. The predominately formed isomer has a substitution pattern identical with that of the rhoeadines.

large chain lengths, first order in H20,. Table 1 and Fig. 4 demonstrate that these are experimentally realized. The apparent first-order rate constant includes a square-root dependence on dose rate which is again in agreement with the experimental results. Ass~lming a value of 2/c1, = 2.4 x lo9 (12) for the bimolecular termination reaction, a rate constant for the chain propagating ~.eaction may be estimated as k,, = (4.0 f 0.4) x lo4 M-' s-'. This is somewhat less than the rate constant found by Seddon and Allen (4) for the propagating reaction in the ethano1/H20, chain reaction. It Inay be noted that the rate co~lstant for reduction of ferricyanide by CH,OH is slightly less than that for reduction by CH,CHOH (8).The small yield of ethylene glycol found,of the termination reaction. In the photolysis of n~ethanol-water solutions (13), the major product of radical combination was said to be ethylene glycol. At high concentrations of either alcohol the yield of the chain reaction is decreased. This phenomenon has previously been reported for the photo-induced reaction in mixed wateralcohol systems (5) and in the radiation-induced oxidation of 2-propanol (1). I11 the latter case the decreased yield was attributed to a decrease in the rate of reaction [7], as a res~ilt of the decreasing dielectric constant of the medium at alcohol concentrations s~~fficiently high to make reaction [7] rather than reaction [8] the ratedetermining propagation step. SummaryThe principal n~~merical results are summarized in Table 2. We have demonstrated that the two-radical model for the chain oxidation of alcohols by H,O,, as developed for 2-propanol, is applicable also to ethanol but that a oneradical model accounts satisfactorilv for the features of the oxidation of ~nethanol. Treatment of 1-thioacetoacetanilide with sulfuryl chloride or thionyl chloride produced a crystalline compound identified as 3.5-diacetonylidene-4-phenyl-l,2,4-ditliiazolidine. The absence of carbonyl absorption in the infrared spectrum of the conipo~~nd indicated that it s h o~~l d be represented by a resonance structLlrc (2) involving non-carbonyl components.Canadian Journal o r Chemistry. 48, 2632 (1970) In connection with our synthetic work on instead of the desired product, a chlorine-free pesticides, the preparation of 2-chloro-I-thio-crystalli~le compound (1n.p. 195-195.5') was acetoacetanilide was required. Accordingly I -obtained. thioacetoacetanilide (1) was treated at roomIn the present p~~blication we provide evidence temperature with ail equinlolar quantity of establishing the structure of this compound. sulfi~ryl chloride. Reaction occurred readily; butThe n~o l e c~~l a r formula Cl,Hl,NO,S,, estabCan. J. Chem. Downloaded from www.nrcresearchpress.com by 34.218.44.141 on 05/11/18For personal use only.