43SG
RAPHAEL~~ECHOU1,ARI AND FRANZ SONDI1EIBfER
Inspection of the absorption spectra of the cytoclironie c-deoxycholate mixtures did not reveal any significant shifts. This was not completely unexpected, however, since the binding of the iron to the protein in cytochrome c is very stable from p H :3.0 to 11.0. The absence of any effect is in contrast to the action of some other surface active agents on the prosthetic group of cytochrome c and of other conjugated protein~.l63~~ Our studies did, howcver, demonstrate the existence of a 315 mp peak which was evident after reduction with H2 and platinum black, ie., in the absence of hydrosulfite. This peak had been observed by Theorell18 and Schales and Behrnts- J e n ~ e n 'who ~ concluded that it indicated the presence of a sulfoxide group in cytochrome c. Lavin, et a1.,20did not observe the peak, and Lemberg and LeggeZ1consider it as be(16) hl.Rabinovitz and P. D. Boyer,
J . B i d . Chem., 183, 111 (1950).
(17) D. Keilin and E. F. Hartree, Naluue, 146, 934 (1940). 118) I I . Theorell, Biockem. Z . , 286, 207 (19363). l!l) 0. Schales and €1. Belirnts-Jensen, I l o p p e - S e y 1 r i . s Z. physiol. ( ' h e i r i , 257, 106 (1939). ( 2 0 J (;. I. Lavin, C. L. Eloagland and S. M. Ward, Proc. Soc. Bxpti. l i d . .Vd., 43, 737 (1940). (21) l?. Lemberg and J. W. Legge, "Hematin Compounds and Bile
[CONTRIBUTION FROM THE
L-01 PO
ing due to the presence of sodium hydrosulfite used as a reducing agent. Nevertheless, i t is present after reduction with hydrogen and platinum black and in the absence of any sodium hydrosulfite. The concentration of sodium deoxycholate used in these experiments is well above the "critical" concentrations for the formation of micelles; Ekwa1lz2 reported it to be 0.014 .M and McBain?3reported it to be 0.005 AT. Because the experimental evidence indicates that the binding is electrostatic, the inicellar structure is probably not involved in the primary binding found here. The apparent binding of sodium deoxycholate by cytochrome c above its isoelectric point may, however, indicate a secondary interaction involving the micelles of sodium deoxycholate. Pigments, Their Cunstitution, Metabolism and Funclion," Iuterscience Publishing Co., New York, ii. Y., 1949, p. 349. ( 2 2 ) P. Ekwall, "Koninkl. Vlaam. Acad. Vetenschap., en Schoue Munsten Belgic. KI. Vetcnschap., Intern. Colloquim Biochem., P r r h l m l Lipiden." (23) J. W. SlcBain, K. C. XIerrill, J r . , lirirl J I < . Viuc,l,rail, . I ' I I ~ ? JOGKNAL, 63, 0 7 0 (19-41).
NEWBRUSSWICK, XEWJERSEY
DANIELSIEFF RESEARCH IXSTITUTE, THE
LVEIZMASN INSTITUTE O F S C I E N C E ]
The Wittig Reaction with Fluorenone. Formation of Cyclopropane Derivatives BY RAPHAEL MECHOULAM AND FRANZ SONDHEIMER RECEIVED FEBRUARY 24, 1958 The Wittig reaction between fluorenone (I) and excess of triphenylphosphine-n-butylidene(IIa) has been found t o yield spiro-(2,3-dipropylcyclopropane-1,9'-fluorene) (IXa). With triphenylphosphinemethylene (IIb), fluorenone gives the previously described spiro-(cyclopropane-I,9'-fluorene) (IXb). The reactions presumably proceed via the 9-alkylidenefluorenes (V), since treatment of 9-n-butylidenefluorene (Va) with triphenylphosphine-n-butylidene(IIa) leads t o the same spiro compound I X a as obtained from fluorenone.
In connection with another investigation,' we were interested in determining how well the Wittig reaction2 proceeds between ketones and the triphenylphosphine-alkylidene derived from a saturated primary bromide containing several carbon atoms. With this aim, we studied the reaction between the highly crystalline ketone fluorenone (I) and triphenylphosphine-n-butylidene (IIa) . The results of this and related experiments are described in this paper. Treatment of n-butyl bromide with triphenylphosphine yielded n-butyltriphenylphosphonium b r ~ m i d em.p. , ~ 243", which was dehydrobrominated with ethereal butyllithium. The resulting solution of triphenylphosphine-n-butylidene (IIa) in ether then was allowed to react with fluorenone (I) first a t room temperature and then in refluxing tetrahydrofuran, the reagent I I a being in threefold excess. Chromatographic separation produced successively three different substances, the last two of which proved to be, respectively, triphenylphosphine and triphenylphosphine oxide. (1) Cf.F. Sondheimer and R. Mechoulam, THISJOURNAL,79, 5029 (1957); SO, 3087 (1958). (2) Cf. G. Wittig, Experientio, 12, 41 (1956); Angew. C h e m . , 68, 505 (19.561,a n d earlier references cited there. (3) Badische Aailia nnd Soda-Fabrik, German Patent 824,047 (C A . 49, 6998 (192;)).
The first substance, m.p. 45", obtained in 50% yield, was found to differ from the expected known4 9-nbutylidenefluorene (Va) of m.p. 5.5'. It proved to be a hydrocarbon, C21H24, i.e., it was derived from the reaction of fluorenone with two molar equivalents of triphenylphosphine-n-butylidene. The fluorene carbon skeleton has been preserved, since oxidation with chromium trioxide regenerated fluorenone. The ultraviolet spectrum (two highest wave length maxima a t 292 and 303 mp) was very similar to that of a 9,9-dialkylfluorene (cf. 9,9diethylfluorene, which has the corresponding maxima a t 292 and 303 m ~ ) .It~ was significantly different from that of fluorene (Amax 288.5 and 300 mp)6 and of a 9-monoalkylfluorene (9-methyl290 and 301 mp)Eand quite different fluorene, A,, from that of a 9-alkylidenefluorene (9-n-butylidene301 and 311 mp).7 That the 9-posifluorene, A,, tion was disubstituted was confirmed by the fact that no red color was produced with Triton B in pyridine solution.8 These facts lead us to propose (4) R. F. Schultz and C. F. Smnllin, THIS JOURNAL,62, 2904 (1940); Y. Sprinzak, ibid., 74, 2116 (1952). (5) E. J. Greeahow a n d D. McNeil, J. Chcm. SOC.,3204 (1956). (6) E. J. Greenhow, D. McNeil and E. N. White, ibid., 986 (1952). (7) S. Wawzonek and E. Dufek, THISJOURNAL, 78, 3530 (1956). (8) Fluorene derivatives containing one or two hydrogen atoms a t C-9 give a red color under these conditions (Y. Syrinzak, Cunzress
WITTIGREACTION WITH FLUORENONE
Aug. 20, 1058
4387
ing of course in the opposite sense than in the corthe spiro- (2,3-di-n-propylcyclopropane-l,9’-fluoresponding ketones) and consequently it behaves rene) structure IXa for this substance. Although the compound IXa was unknown, like a carbonyl function in reactions with Grignard spiro-(cyclopropane-1,9’-fluorene)(IXb) has been reagents, lithium aluminum hydride, etc.’l Simde~cribed.5J.~ We therefore allowed fluorenone ilarly, the 9-alkylidenefluorene V reacts with the (I) to react with excess triphenylphosphine- triphenylphosphine-alkylidene I1 to give an intermethylene (IIb). This reaction led to triphenyl- mediate such as VI1 and/or VIII, which then dephosphine, triphenylphosphine oxide and a sub- composes to yield the cyclopropane derivative I X stance which proved to be identical with an authen- and triphenylphosphine. Support for this mechtic sample of the spiro compound IXb. By anal- anism is provided by the fact that 9-n-butylideneogy, the cyclopropane structure IXa assigned to fluorene (Va)4 on being allowed to react with trithe product from fluorenone and triphenylphos- phenylphosphine-n-butylidene (IIa) does indeed phinebutylidene must be correct. lo As expected, smoothly give the spiro compound IXa in addition the ultraviolet spectra of the two spiro compounds to triphenylphosphine. IXa and IXb are almost identical. Acknowledgments.-We are indebted to Dr. The formation of the spiro compounds readily can R. E. Dean of the Coal Tar Research Association, be explained if it is assumed that the normal reac- Leeds, England, for a sample of spiro-(cycloprotion2 between the triphenylphosphine-alkylidene pane-l,g’-fluorene) (IXb). and fluorenone takes place first, leading via the Experimental’ 2 intermediate I11 and/or IV to the 9-alkylidenen-Butyltriphenylphosphonium Bromide .--A solution coilfluorene V and triphenylphosphine oxide. The 5 g. of triphenylphosphine and 16 cc. of n-butyi semicyclic double bond in dibenzfulvenes of type V taining bromide in 5 cc. of benzene was boiled under reflux for 4 hr. is strongly polarized (the polarization, as in VI, be- aiid then cooled. The resulting precipitate was collected
0-0
+ RHz-;Ph3
+
\ Y0 ’ I
0-0 \A v / 03
RHC
\
111 @PPh3
IIa, R = C3H7 h,R = H
J. Va, R = (2x117 I>, R = H
vI
I S a , R = C3H7 b,R = H
” PPh,
1v
‘CH IIa, R = c311? b,R = H
1
VI1
R
‘Cd VI11
1
R
Handbook, 14th International Congress of Pure and Applied Chemist r y , Zurich, July, 1965, p. 66; THIS JOURNAL,in press. (9) H. Wieland a n d 0. Probst, A n n . , 530, 274 (1937). (10) Although two isomers of structure I X a are possible, it was n o t determined whether t h e suhstancc belongs t o the nzeso or the racemic series.
and washed with benzene. Crystallization from chloroform yielded 4.2 g. of the salt with m.p. 242-243”, reported3 m.p. 235’. Anal. Calcd. for CS2Hz4BrP:Br,20.02. Found: Br, 20.20.
Spiro-(2,3-di-n-propylcyclopropane-l,9’-fluorene)(IXa). (a) From Fluorenone (I).-A 1 Nethereal solution of butpllithium (9 cc.) was added to a suspension of 4 g . (10 millimoles) of n-butyltriphenylphosphonium bromide in 50 cc. of ether with swirling under nitrogen. The resultiiig deep red mixture was shaken in nitrogen for 2 hr. A solutioii of 0.6 g. (3.3 millimoles) of fluorenone in 15 cc. of ether was added, whereby a bulky white precipitate was formed ;vhich quickly turned brown-black. The mixture was shaken ffir 4 hr. and then allowed to stand overilight. Ether was distilletl off at the same time as tetrahydrofuran was added uiitil tlie distillation temperature reached 60”. The mixture was boiled under reflux for 6 hr., cooled, dilutcd with water aiid extracted with ether. The ether extract was washcd with dilute hydrochloric acid, sodium bicarbonate solution atid then was dried and evaporated. Chroniatography on 50 g. of alumina and elution with pentane gave 0.46 g. (.50y0)of the spiro compound I X a which after crystallization from ethanol showed m.p. 44-45’; 227, 269, 292 and 303 mp (log c 4.33, 4.25, 3.99 aiid 4.06, respectively). Anal. Calcd. for C21HZ4: C, 81.25; H, 8.75. Fuutid: C, 91.31, 91.45; H , 8.75, 8.71. Further elution of the chromatogram with petrcilcriin ctlier yielded 0.48 g. of triphenylphosphine and clutioii with ether gave 0.84 g. of triphenylphosphine oxide. Both these substances were identified by comparison with authctitic samples. (b) From 9-n-Butylidenefluorene (Va).-The reaction between 9-n-butylidenefluorene (Va)l (0.73 g., 3.3 millimoles) and triphenyl phosphine-n-butylidene was carried out exactly as described above, on the sa” scale. Chromatography on 50 g. of alumina and elution with pentane yielded 0.48 g. (52%) of the spiro compound I X a with m.p. 44-45’. Identity with a sample prepared by method a was established by mixture m.p. determination and infrared comparison. Further elution with petroleum ether gave 0.47 g . of triphenylphosphine. Elution with ether gave only 60 mg. of triphenylphosphine oxide. Oxidation of Spiro-( 2,3-di-n-propylcyclopropane-l,9’fluorene) ( m a ) to Fluorenone (I).-A solution containing (11) Cf. E. D. Bergmann, in “Progress in Organic Chemistry,” Editor, J. W. Cook, Butterworths Scientific Publications, London, 1955, Vol. 3, pp. 107-109. (12) Melting points are uncorrected. All chromatograms were made with Alcoa activated alumina, grade F-20 (Aluminum Co. of America, Pittsburgh, Pa.). Ultraviolet spectra were measured in 95% ethanol solution on a Unicam model S.P. 500 spectrophotometer. Analyses were carried o u t in our microanalytical laboratory under the direction uf hIr Erich bleier.
200 ing. of the spiro conipound 1Xa and 500 iiig. of cliruriiic :wid iii 5 cc. of glacial acetic acid was boiled under reflux for 2 hr. The mixture was cooled, diluted with water and the iicutral product was isolated with ether ill t h e usual way. Crystallization of the product from ethaiiol gave 65 mg. of fluorenone, n1.p. 80-8Z0, undepressed on admixture with ;iii authentic sample (m.p. 82-83'). Spiro-(cyclopropane-l,g'-fluorene)(IXb). --.a solutioii of t 1-ipIrenylpliosphinemethylene was prepared by adding 2 1 cc. of a 1 ethereal solution of butyllithium to a suspension o f 8.!1 g. ( 2 5 millimoles) of methyltriphenylphosphoniuni l)rornitle13in 100 cc. of ether with swirling under nitrogeii. 'l'lic mixture was shaken for 2 hr., fluorenone (1.44 g.) iii 2,; cc. of ether was added and tlic rc:iction was then carrictl
exactly as described a l o v c . l'lic proiluct ili~solvetli l l benzene was filtered through 100 g . of alumina to reiiiuvc triphenylpliosphine oxide. The eluate was evaporated to dryness, dissolved in pentane and clirolliatcigraphecl O I I 100 g. of alumina. Elution with pentane and crystallizatioil from methanol yielded 302 mg. of t h r spiro conipiiuild 1x11, 1n.p. 50-55'. -\ sample, further purified b>-crystallizatii~i from methanol and high-vacuum sublimation, showc(1 m . p . 68-70"; A,, 226, 268, 292 and 303 r n p (log e 4.;3tJ, 4.20, 3.92 arid 4.02, respectively). There \\-as rio tielire.-siori iii m . p . u i i irdniixture with an authentic spcciriicii (m.p. 70-7lo)6 atiii tlie infrared spectra xvcre ideiitical. Further elution of the coluinn xitli pctriilcun1 ctlier (l>.lJ. (JUL
I
[ C O N T K I I I L v l I U N FKO.11 XUYES CI11;MICAL L A 1 3 O R A T U K Y , ~ S I V h K 5 rY 1 Ill*' ~l,l,l.YUlS
Polyphosphoric Acid as a Reagent in Organic Chemistry. IX.' Cyclization to Diaminoacridines 2 131, 11. I
