COMMUNICATIONS TO THE EDITOR
April, 1951
1889
adenine synthesis. The possibility exists that this type of phenomenon may occur in other biochemical mutations. The initial stocks from which the yeasts used in these experiments were derived were kindly furnished by Dr. Carl C. Lindegren.
droxyketo-acid from (IV) reacted with one mole of periodate. Dihydropicrotoxinide (IV) was converted to its ethylene mercaptal (m.p. 250' ; calcd. fOr.C&&&: C, 58.50; H, 7.36. Found: C, 58.52; HI 7.38) desulfurized with Raney nickel to tetrahydrodesoxypicrotoxinide (V; m. p. 162' ; calcd. for C14H2203: C, 70.55; H, 9.31. Found: INSTITUTE OF RADIOBIOLOGY AND BIOPHYSICS 5.70 p ) . The latUNIVERSITY OF CHICAGO RICHARD ABRAMS C, 70.65; H, 9.41; infrared A, ter gave a benzoate (VI; m.p. 134'; calcd. for CHICAGO 37, ILLINOIS C21H2604: C, 73.66; H, 7.65. Found: C, 73.56; RECEIVED MARCH10, 1951 HI 7.88) which underwent smooth pyrolysis to benzoic acid, carbon dioxide and picrotoxadiene (VI1; THE SKELETON OF PICROTOXININ
I
Sir: In order to account for the formation of picrotic acid and related substances, Robertson, et al.,l have
I
proposed the partial carbon skeleton (I) for picrotoxinin, ClsHl606, one of the two components of the amaroid picrotoxin. However picrotoxinin possesses two carbocyclic rings, ;.e., one more carboncarbon bond must be drawn to complete the expression. Evidence now obtained defines the location of
j A
coo 1
0
the missing bond and requires that fiicrotoxinin be assigned the skeleton (11),one which lacks onlypve carbon atoms of a complete steroid nucleus. Dihydro-a-picrotoxininic acid2 underwent smooth pyrolysis with loss of carbon dioxide and water to a new substance, designated picrotoxinide, C14H1804, not crystalline (Am=. 254 mp, log E 4.0; Amax. 2.95, 5.70, 5.84 and 6120 p) formulated as (111). Hydrogenation of (111) gave 90% of dihydropicrotoxinide (IV, m.p. 187'; calcd. for C14H2004: C, 66.64; H, 7.99. Found: C, 66.82; H, 8.11) which formed a 2,4-dinitrophenylhydrazone (m.p. 209' dec.; calcd. for C20H240~N4: C, 55.53; H, 5.60.
yo
//O
I
I
VI1
b.p. 213'; Am, 256 mp, log E 3.6) characterized by its maleic anhydride adduct (m.p. 75'; calcd. for C17H22.03: C, 74.42; HI 8.08. Found: C, 74.51; H, 8.31)) the corresponding imide (m.p. 147-148'; [cu]~OD -78' [chloroform, c = 3-11; calcd. for ClrH2302N: C, 74.69; H, 8.48; N, 5.12. Found: C, 74.63; H I 8.55; N, 5.28) and the N-phenyl imide (m.p. 178'; [ ( u ] ~ O D -42' [chloroform, c = 2.51; calcd. for C23H2702N: C, 79.04; H, 7.79; N, 4.01; Found: C, 79.57; HI 7.80). Synthetic cis-5-isopropyl-8-methylhydrin-4,6diene (VII) was obtained by the action of isopropyl lithium on cis-S-methylhydrind-6-ene-5-0ne, (2,4dinitrophenylhydrazone m.p. 138-139' ; calcd. for CleHls04N4:C, 58.16; H, 5.49. Found: C, 57.93; H, 5.41) prepared by an unambiguous route from cis-2-methyl 2-carboxycyclopentane-1-aceticacid.4 Although the maleic anhydride adduct of the synthetic diene was not crystalline it gave an infrared spectrum identical with that of the natural adduct, and was converted to the crystalline imide (m.p. 158-159', mixed m.p. with the natural imide 147158'; found: C, 74.59; H, 8.36; N, 5.11) and the N-phenyl imide (m.p. 151-152', mixedm.p. with the natural N-phenyl imide 151-175'; found: C, 79.04; H I 7.83; N, 3.88). Both imides gave infrared spectra identical with those from the corresponding natural derivatives; picrotoxadiene is clearly an optically active form of cis-5-isopropyl-8methyl hydrin-4,6-diene (VII). DEPARTMENT OF CHEMISTRY HARVARD UNIVERSITY CAMBRIDGE 38, MASS. HAROLD CONROY~ RECEIVED DECEMBER 19,1950
Found: C, 55.42; HI 5.55) and a dibenzylidene derivative (m.p. 127-128'; calcd. for GsH2s04: C, 78.48; H, 6.59. Found: C, 78.43; H, 6.74) strikingly similar in infrared and ultraviolet spectra to 2,5-diben~ylidenecyclopentanone.~ The dihy(1) J. C. Harlaad and A. Robertson, J . Chcm. Soc., 937 (1939); D. Mercer, A. Robertson and R. S. Cahn. ibid., 997 (1935). (2) P. Horrmaaa, Ber., 46, 2793 (3) D. Vorlaader and K.Hobohm, i€id., 39, 1836 (1896).
(1913).
(4) K. D. Erringtoa and R. P. Liastead, J . Chcm. SOC., 666 (1938).
(5) National Institutes of Health Postdoctoral Fellow.
THE STRUCTURE OF PICROTOXININ
Sir :
The skeleton (I) for picrotoxinin, ClaHlaOe, has been proposed' to account for the formation of (1) H . Coaroy, THISJOURNAL, 73, 1889 (1951).
imn
I
A
11
picrotoxinide (11) i n the pyrolytic decomposition of hyclroxyl, for otherwise the formation (&de infra)of dihydro-a-picrotoxininic acid. Picrotoxinin ha.: the products (IV) and (V) obtained by Sutter and long been suspected213to be a monuhydroxy-dilac- Schlittlera wotild be prohibited. Coiiseqitently thr tone; the remaining oxygen was assumed to be CII,OTI present as an ether linkage, since no carbonyl derivatives could be prepared. The infrared absorption spectrum of picrotoxinin has now been examined, and it is consistent with these views. I n addition to an intense, fairly sharp band a t 2.90 p , due to the hydroxyl group, the only carbonyl absorption is a split peak with maxima a t 5.57 and 5.63 p, characcomplete structure of picrotoxinin may be repreteristic of two five-membered lactone^.^ Picrotoxic acid, C1~H1807, monobasic, formed di- sented as (VI), while a-picrotoxininic acid is picrectly by the action of one mole of base upon picro- tured as (VII), a view consistent with its infrared toxinin, produces infrared absorption a t 5.57 and spectrum and with the formation of picrotoxinide' 5.78 p . This evidence is consistent with the view in the pyrolysis of the dihydro acid. that in this acid only the lactone responsible for the 5.63 p band of picrotoxinin has been opened. The isomeric a-picrotoxininic acid, also monobasic, produces infrared bands at 5.70 and 5.78 p, with negligible absorption in the 5.6 p region, evidence consistent with the view that the single lactone system in this acid directly corresponds to neither of those present in the original picrotoxinin. Both of these isomeric, monobasic acids were treated with a known excess of neutral, aqueous sodium periodate solution for an extended period a t room temperature, and the remaining periodate deA VI A VI1 termined according to a well known p r ~ c e d u r e . ~ Neither picrotoxic nor a-picrotoxininic acid re- The formation6 of the products (IV) and (V) of acted to any extent with periodate under these con- Sutter and Schlittler from either a-dihydropicrotoxditions. On the other hand, picrotoxinin dicarbox- inin or dihydro-a-picrotoxininic acid is represented ylic acid,2 which can be prepared from either a- as a base catalyzed dealdolization with cleavage of picrotoxininic acid or picrotoxinin and which shows bonds a t C5-C6 and Cl-C2, followed by /3-eliminaonly carboxyl absorption a t 5.5 p, took up very tion of the oxide at C13. Picrotoxic acid, which nearly one mole of periodate under the same condi- does not undergo such cleavage, and is unchanged tions. These facts require the inclusion of the by the continued action of aqueous base, retains one I I structural unit - CO--O- C-C-0-COin the of the original lactones intact, and is formulated as (VIII). The stereochemistry of the structure (VI) I i molecule of picrotoxinin, and taken with the evi- as well as possible mechanisms of the transformadence for the skeleton (I) indicate the partial formulation (111) for this substance. Any alternative o,\ C y/"\ 7 expression would embody either a four-membered or a six-membered lactone, both excluded by the infrared data4 (vide supra). The oxygen a t C6 cannot be present as the oxide function, but instead must represent the single free I I ( 2 ) P Horrmann, Bci ,46, 2793 (19131, 45, 2090 (1912). .4nn ,411, 073 (1916) (3) D Mercer, A Robertson and R S Cahn. J Chem Soc , 997 (1935) (4) R S. Rasniusseu aud It K Rrattaiu, THISJOURNAL, 71, 1076 (1949). (5) "Organic Reactions," Vol York. N Y ,1944, p 361.
IT. John Miley
nud
Sons. Tnr., New
(0) M. Sutter and E. Schlittler, Hclu. Chim. Acta, SO, 403 (1947); SO, 2102 (1947); 31, 1856 (1019); 32, 1860 (1944).
April, 1951
COMMUNICATIONS TO TNB EDITOR
1891
tions outlined will be discussed in a subsequent communication.
effecting the reduction of nitrobenzene without bringing about nuclear substitution. The presence of phenol along with biphenyl and terphenyl sugDEPARTMENT OF CHEMISTRY gests strongly that phenyl radicals axe also produced HARVARD UNIVERSITY HAROLD CONROY’ CAMBRIDGE 38, MASS. and that they, too, abstract oxygen rather than atRECEIVED JANUARY 29, 1951 tacking the nucleus of the nitro compounds. This 17) National T n s t i t i i t m of Health Postdoctoral Fellow is an entirely unexpected result in the light of previous work which indicates that phenyl radicals produced in the thermal decomposition of benzeneTHE REACTION OF TRIPHENYLMETHYL WITH diazoacetate6 and benzoyl peroxide6 attack nitroNITROBENZENE benzene to give p-nitrobiphenyl. This difference in Sir : behavior may be taken to indicate that pairs of It has been rather generally a c ~ e p t e d ’that ~ ~ ~ ~radicals produced in unimolecular decompositions aromatic nitro compounds are activated, relative are capable of making a concerted attack on solvent to their unsubstituted analogs, toward attack by molecules in the short time interval in which they free radicals leading to aromatic substitution. Re- are held in the same cavity in the solution. A simicent evidence concerning the inhibition of the per- lar discrepency involving the fate of benzoate radioxide induced polymerization of allyl acetate4 by cals? produced in different reactions has recently nitro compounds was taken to indicate that some been observed. other mode of reaction must be available to the This research was carried out under a contract growing chain radicals. A sequence of reactions involving attack on the nitro function was sug- with the Office of Naval Research. ( 5 ) W. S. M. Grieve and D. H . Hey, J . Chcm. SOC.,1803 (1934). gested as a likely alternative to nuclear substitution. D. F. DeTar and H . J. Scheifele, Absfructs, Chicago Mecling of We have undertaken a study of the reaction of tri- A m(6). Chcm. Soc., September, 1950. phenylmethyl with nitrobenzene in benzene solu(7) G. S Hammond, J. T Rudesill and F J Modic, THISTOURNAL, tion and find that this radical and others which may manuscript submitted. be formed in the course of the reaction react only by CHEMICAL LABORATORY abstracting oxygen from the nitro function. IOWA STATE COI,I.BC;E GEORGE S. HAMMOND ABRAHAM RAVVE When one mole equivalent of nitrobenzene is AMES, IOWA RECEIVED MARCH11,1951 added to a benzene solution of triphenylmethyl, the solution immediately develops a red color. If such a solution is allowed to stand exposed to dif- RADIATION CHEMISTRY OF FERROUS SULFATE fuse daylight for 24 hours, nearly a quantitative SOLUTIONS yield of triphenylcarbinol precipitates. The su- Sir: pernatant liquid is fractionated by adsorption on a Emphasis has been placed recently on R, the a 5 : 1 Magnesol-Celite mixture followed by elution yield ratio, Fe$$+/Fe;:+, obtained by X-ray or with mixtures of Skellysolve B, benzene and acetone or ethanol. Products isolated and characterized y-ray irradiation of dilute aqueous solutions of ferare unreacted nitrobenzene, azobenzene, azoxyben- rous sulfate in the presence and absence of air.lo2 zene, nitrosobenzene, biphenyl and p-terphenyl. These values vary from 2.55 to 4.0, as is shown in The nitrogen balance is quantitative but oxygen Table I.132J*4 The present communication demonand hydrogen have not been entirely accounted for. strates that theoretical values less than 4.0 are posPhenol is detected as its tribromo derivative in alka- sible on the basis of the accepted ferrous sulfate line extracts of the original solutions. When the oxidation mechanism. According to present concepts reaction (1) occurs reaction is carried out in the dark a slower reaction produces a quantitative yield of ditriphenylmethyl on the passage of ionizing particles through water. peroxide and the same nitrobenzene reduction Reaction ( 2 ) also proceeds through free radical inproducts. There is no evidence for attack on termediates and occurs where the rate of energy loss solvent in the dark reaction. Nitrosobenzene, is high and results from pairwise recombination of which is found only in small amounts, reacts rap- hydrogen atoms and hydroxyl radicals in zones of idly with triphenylmethyl, in separate experiments, high free radical concentration. The mechanism to produce azobenzene and a trace of azoxybenzene far ferrous sulfate oxidation is: (substantial amounts are produced from nitrobenH2O = H + OH (1) zene) along with triphenylcarbinol, biphenyl and €I20 = ‘/2H2 ‘/2H202 (2) terphenyl. Azoxybenzene does not react to any H + 0 2 = H01 (3) appreciable extent in 72 hours. Fe++ + OH = Fe++++ OH(4) It is not our intention to consider the mechanisms F e + + + H 0 2 = F e + + ++ €IOn(5) of these transformations in detail a t this time since Hot- + H + = H202 (6) further study is in progress. However, certain imFe + H202 = F e + + + + OH + OH- ( 7 ) portant conclusions are apparent. It is clear that I3 + H = Hz (8) a t least one radical, triphenylmethyl, is capable of In the absence of air the oxidation follows reac(1) W. A. Waters. “The Chemistry of Free Radicals,” Oxford Unitions (l),( 2 ) , (4), (7) and (8) and in the presence of versity Press, Oxford, 1946, p. 151.
+
(2) L. P. Hammett, “Physical Organic Chemistry,” McCraw-Hill C o . , New York, N. Y . , 1940, p. 383. (3) G. W. Wheland, “Advanced Organic Chemistry,” Jobn Wiley and Sons, Inc,, New York, N. Y . , 1949, p. 670. (4) G S. Hainrnond and P. D. Bartlett. J . Polymer Sci., in press.
(1) N. Miller, J . Chcm. Phys., 18, 79 (1950). (2) F. H. Krenz and H. A. Dewhurst, ibid., 11, 1337 (1949). (3) H.Fricke and S. Morse, Am. J . Roenf. a%d Rad, T h a . , 18, 430 (1927); H. Frickeand E. J. Hart, J . C k m . Phys., 8,60 (1986). (4) N. A. Shishacow, Phil. Mug., 14, 198 (1932).