NOTES
3202
lsolated in 78% yield (14 mg.) by addition of water to the solution, treatment of the solution with ammonium hydroxide, extraction of the precipitate with ether, and recrystallization from pentane of the solid obtained from the ether extract. The oxime prepared from this sample of phenacylpyridine melted a t 115117”. 2-[p-Hydroxy-D-(o-boronophenyl)vinyl]pyridine was obtained by addition of sodium hydroxide in 50% alcohol-water to the hydrochloride salt (0.2940 g.), evaporation of the solution to dryness, extraction of the resulting solid with chloroform, and evaporation of the extract. For further purification, the sample was dissolved in methanol and the solution was filtered and taken to dryness. The resulting pale yellow solid charred and decomposed without melting when heated in the range of 230’. The analysis of the material dried a t room temperature corresponded to a dihydrate of compound 11.
VOL. 28
Anal. Calcd. for C13HlzBN03.2HzO: C, 56.3; H , 5.82; N, 5.06. Found: C,55.9; H,5.90; N,4.92. After the sample had been heated a t 65” (1 mm.) for 24 hr., the analysis agreed with that for 2- [p-hydroxy-p-(0-boronophenyl)vinyl] pyridine . Anal. Calcd. for C13H12BN03: C, 64.7; H, 5.02; N, 5.81. Found: C, 64.5; H,5.33; N,5.99. A semicarbazone derivative was obtained by warming a solution of 86 mg. of compound 11, 0.2 g. of semicarbazide hydrochloride, and 0.3 g. of sodium acetate in 4 ml. of ethanol and 2 ml. of water for 15 min. When the solution was cooled, 60 mg. of the yellow crystalline semicarbazone derivative of compound I1 was obtained, m.p. 169-172’. The analytical sample was dried a t 65’ (1 mm.) for 8 hr. Anal. Calcd. for Cl4H16BN4O3: C, 56.4; H, 5.07; X , 18.80. Found: C , 56.5; H , 4.88; N, 18.78.
Notes Reduction of 1-Methyl-3-acylindole Derivatives with Lithium Aluminum Hydride’ K. T. POTTS AND D. R. LILJEGREN~ Department of Chemistry, University of Louisville, Louisville, Kentucky, and Department of Organic Chemistry, University of Adelaide, Adelaide, Australia Received Febrwzry 1 , 1963
The application of lithium aluminum hydride to the reduction and cyclization of (2-3‘-indolylethyl)- or (2-3’-indolyl-2-oxoethyl)pyridinum and isoquinolinium sa’.,, iiiis been described in previous papersS3 This note reports some results obtained in the 2nd-X-methyl ~eries.~ The reduction of various 3-acylindoles with lithium aluminum hydride is a well authenticated5hydrogenolysis reaction, the 3-alkylindoles being readily obtained. With ind-K-methyl-3-acylindoles the reduction has been reported to stop a t the intermediate alcohol stage and does not appear t o be analogous to the reduction of a disubstituted vinylogous amide. Thus 1,X-dimethyl3-indoleglyoxamide (I, R = CH3; R’ = H) and lithium aluminum hydride gave the alcohol6 I1 (R = CHI; R’ = H ; R ” = OH). However, conflicting reports have appeared, e.g., the reduction of I (R = R’ = -CH2Ph) with lithium aluminum hydride to give the oxygen-free product’ I1 (R = R’ = -CH2Ph; R ” = H) ; on the other hand, 1-methyl-3-indolylaldehydes have been shown* to undergo reduction to l-methyl-3(1) Regarded as P a r t I V in the series: Synthetic Experiments Related to the Indole Alkaloids. (2) Recipient of a C.S.I.R.O. Senior Postgraduate Studentship, 19611962. ( 3 ) P a r t 111: K. T. Potts and D. R . Liljegren, J . Org. Chem., 28, 3066 (1963). (4) This work was supported in part b y P H S G r a n t H-C475 from the Xational Heart Institute, Public Health Service. ( 5 ) E. Leete a n d L. Marion, Can. J . Chem.. 31, 775 (1953). (6) M. E. Speeter, U. S. Patent 2,815,734: Chem. Absli., 62, 12923f (1958). (7) A . Buzas, C. Hoffman, and G. Regnier, Bull. soc. chim. France. 043 (1950).
hydroxymethyl indoles, in agreement with the former reaction. In attempts to effect the reductive cyclization of 1-[2-(1’-methyl-3’-indolyl)-2-oxoethy1]pyridinium derivatives, no clear-cut results could be obtained and it was decided to investigate the reduction of l-methyl3-acetylindole.
- mcH-c R’ I
R,:c-ocJ--)Q
\ /
I
CH, I
E
\
/
R
I I
CH3
I1
Immediately after isolation in the usual way, the product from the lithium aluminum hydride reduction showed intense hydroxyl absorption in its infrared spectrum, indicating the presence of a predominant amount of structure I11 (R = CH3). However, no derivative of the alcoholic function could be obtained and, on standing, the crude product developed an odor of acetaldehyde. This was of interest in view of the reported* decomposition of 1-methyl-3-hydroxymethylindole (111, R = H) to formaldehyde and the diindolylmethane IV (R = H). However, this decomposition pathway was not followed in the case of our 1-(1’-
methyl-3’-indolyl)ethanol. On distillation or on boiling with water it underwent ready dehydration to l-methyl3-vinylindole, which immediately polymerized t o poly(1-methyl-3-vinylindole). There was no evidence of the formation of an appreciable amount of 1,l-di(1’methyl-3’-indolyl)ethane (IV, R = CH3), authentic (8) E. Leete. J . A m . Chem. Soc., 81, 6023 (1959).
NOVEMBER, 1963
NOTES
samples of which were prepared by (1) methylation of l,l-di(3’-indolyl)ethane,and (2) from 1-methylindole, paraldehyde, and zinc chloride. The poly (l-methyl-3viny1)indole structure was assigned on the basis of the following evidence. Analytical data established the empirical formula (CllHIlN) supported by complete absence of hydroxyl absorption in the infrared spectrum. The ultraviolet spectrum showed, in addition to the peak characteristic of 1-methylindole, a broad shoulder a t 2670 A. ; 1-methyl-3-vinylindoles h a y been foundg to exhibit maxima in the region of 2650 A. The isolation of a small amount of acetaldehyde can be most simply attributed to a retroaldol type reaction, probably catalyzed by the alkali of the glass though we were unable to establish the conditions under which this was the major decomposition pathway. The conjugate base of the indole system must assist in the hydrogenolysis of the coordinated oxygen atom of the 3-ketone function and, as no conjugate base can be formed in the case of the 1-methylindole products, the reduction was next carried out in a system in which the carbon-oxygen bond was weakened by coordination with a stronger Lewis acid such as aluminum chloride. 1-Methyl-3-acetylindole was accordingly reduced with lithium aluminum hydride-aluminum chloride mixture and, as expected, 1-methyl-3-ethylindole was isolated in excellent yield. Parallel results were also obtained with similar arylmethyl ketones such as 3,4-dimethoxyacetophenone. Reduction with lithium aluminum hydride gave 1-(3,4dimethoxyphenyl)ethanol, which readily lost water on distillation forming 3,4-dimethoxystyrene, characterized as the dibromide. With lithium aluminum hydride and aluminum chloride, the course of the reaction was a1tered to yield 3,4-dimethoxye t h y lbenzene. This is analogous to the hydrogenolysis1° of benzophenone to diphenylmethane, and acetophenone t o ethylbenzene. With this more definite knowledge of the behavior of 1-methyl-3-acetylindole toward lithium aluminum hydride, a series of reductions of the salt 2-[2-(1’methyl-3’-indolyl)-2-oxoethy1]isoyuinoliniumiodide (V) was carried out. Despite repeated attempts using lithum aluminum hydride alone or niixed with aluminum chloride, no pure products could be isolated when tetrahydrofuran was used as the solvent. Spectral evidence indicated that in the majority of fractions a t Isast partial reduction had occurred and it is likely that
the large number of fractions formed is due to the many ways in which a highly reactive intermediate such as V I can undergo further reaction. The use of dioxane as solvent gave an intermediate reaction complex that was almost insoluble so that a t no stage was the reaction mixture homogeneous. It is interesting that the main product isolated from this system was 2-[a-(1’methyl - 3’- indolyl) - 2 - oxoethyl] -1,2,3,4-tetrahydroisoquinoline (VII), besides unidentified oily fractions similar to those already mentioned. The identity of this product was indicated by analytical and spectral data. The infrared spectrum showed main absorption bands a t 3100, 3020, 2911, 2805, and 1620 cm.-’, and the ultraviolet spectrum showed the presence of a 1-methyl-3-acylindole chromophore1’ (Ama, 248, 306 mp). The n.m.r. spectrum1*was not complex, showing main peaks a t 7.1 7 (K-CHJ, 6.26 and 6.23 7 (-CH2peaks), and aromatic protons. These results indicated that the 3-acylindole group had not undergone hydrogenolysis and that the pyridine ring system had been reduced to the tetrahydro stage. Confirmation was obtained by the synthesis of VI1 from the chloride of the salt V by reduction with Adam’s catalyst in acetic acid solution. The failure of the 3-acylindole function to undergo hydrogenolysis in this reaction can be attributed to a large extent to the heterogeneous reaction medium. The results obtained with these 1-methylindole compounds indicate the desirability of working with the unniethylated series. The cyclized products may then be readily converted into the methyl-substituted series with sodamide and methyl iodide in liquid ammonia. l 3 l 4
VI1
v
(9) E. Leetc, . I . A m . C l i r m . Soc., 88, 6838 (l9GO). (10) R. F. Nystron and C. R. A. Borger, iDid., 80, 2896 (1958).
3203
Experimental15 Reduction of 1-Methyl-3-acetylindole with Lithium Aluminum Hydride. Poly( I-methyl-3-vinylindole) .-l-Methyl-3-acetylindole16 (6.0 g., 0.03 mole) and a solution of lithium aluminum hydride (4.0 g., 0.11 mole) in dry tetrahydrofuran (300 ml.) were heated together under reflux for 4.5 hr. After decomposition of the reaction complex with water, the reaction mixture was extracted with ether and the ether solution dried (sodium sulfate). Removal of the solvent left a pale yellow oil ( 6 g , ) , m i 3400 em.-’ (broad, very intense). Extraction of the crude product, whirh had an odor of acetaldehyde, with cold water and treatment of the aqueous extract with Brady’N reagent gave a yellow precipitate of acetaldehyde 2,4-dinitro~henylhydr22zone, m.p. 168’, alone and when mixed with an authentic tlample.1’ The infrared spectra of the two samples were superimposable. The residue from the water extraction was divided into two portions. The first was boiled with water for 24 hr. and on cooling a brown solid, m.p. 70-80’ was deposited. It showed no hydroxyl group absorption and was purified by dissolution in benzene and filtration through a coliimn of neutral ttlumiria (11) !J. A. Ballantine, C. B. Barrett, R. J. S. Beer, B. G. Boggiano, S. Eardley. B. E. Jennings, and A . Robertson, J . Chem. Soc., 2227 (19.57). (12) The spectra were recorded from a Varian V-4302 dual purpose 60Mc. n.m.r. spectrometer and chemical shift values are reported in T units using tetramethylsilane (I 10) as internal standard. We are indebted to Dr. T. Spotswood for his assistance in determining this spectrum. (13) Irierite.15 The mixture was To establish whether one-electron reduction by a filtered through Celite,lB and the filtrate evaporated to a colorless metal ion on ergosteryl acetate epidioxide (IV) would sirup which failed to crystallize. The product gave a positive ninhydrin reaction, migrated as a single zone ( R r = 0.75) on follow a similar reaction course and generate a steroidal thin layer vhromatograms with 8 : 1 : I benzene-methanol-pyrit-alkoxy radical, the reduction of IV was studied. dine as developer, and did not reduce Fehling solution. Treatment of epidioxide IV with chromous chloride3 Conversion of the I)roduct to the hydrobromide salt with an in ethanolic hydrochloric acid resulted in rapid reducequivalent of hydrogen bromide in methanol, followed by evapo2.97 s ( S H ) , 5.73 vs ration, gave a hygroscopic sirup, A::', tion. Chromatography of the materials formed yielded ( O A v ) , 6.12 w, 6.63 p w (XH3+). The product was not obtained ergosteryl acetate (V), a dimeric substance, CBoH~004 crystalline. (VI), and the hydroxy acetate (VII), all formed in equal Methyl 2-Acetamido-3,4,6-tri-0-acetyl-2-deoxy-~-~-galactoyields of about 30%. pyranoside (IV).--The sirully hydrobromide product from the The structures are assigned as follows. The dimer preceding preparation was dissolved in a cold mixture of dry pyridine ( 5 nil.) and acetic anhydride ( 2 . 5 ml.), and left for 3 hr. V I showed no selective ultraviolet absorption. Saponia t room temperature. The mixture was poured into water, fication of the dimer diacetate yielded a diol which difand the product was extracted with chloroform. The extract fered from the well known bisergostatrienol (IX),4the was washed with water, and the find traces of pyridine were product of photodimerization of ergosterol. The 1i.m.r. removed by shaking the extract with aqueous c~adniiiimchloride solution. The cadmium c*hloride-pyridine cmiplex was filtered, spectrum of the dimer VI indicated vinyl proton absorp(13) .\ product of I'fanstielrl Laboratories. Waukegan, Ill. T h e aiitliors tliank I h . I > . G. 1)olierty. of Oak Ridge National Laboratory. Oak Ridce, T e n n . . for a gift of tliis inaterial. (14) Iionover sodium, followed by preparative gas cahromatogra(2) I99.5% pure as indicated by g.1.p.c. were obtained from distillations through a 1000 X 13 mm. column packed with glass helices and equipped with a total reflux head. Preparations of N-alkylpropargylamines from N-(2-chloroally1)alkylamines were carried out in the same manner as preparations of N-alkylallenimines from N-(2-bromoallyl)alkylamines except reaction times were 10 hr. or longer and the sodium amide-N-( 2-ehloroally1)alkylamine mole ratio was 2.1. Initial sodium amide concentrations were 0.8 M . Boiling points, refractive indices, and elemental analyses of new compounds are given in Table I. Analytical Method.-Solutions containing 19-78% N-ethylallenimine (IIIa) by weight were prepared using IIIa and Nethylpropargylamine (IVa) that were >99.57, pure. Each of these solutions was diluted with ether to give a series of solutions containing 64.2-99.17, ether by weight, and g.1.p. chromatograms were taken of 5 pl. samples of all solutions a t 120" with a helium flow rate of 90 ml./min. The retention times (a)in sec. measured from the air peak were 50, 215, and 330 for ether, IIIa, and IVa, respectively. A linear plot with a slope of unity was obtained when the weight fraction of IIIa was plotted against Hrrra/(Hrrt~ Hwa) where HIIT,and Hlva are the peak heights of the bands due to IIIa and IVa, respectively. Data obtained for all solutions (0-99.17, ether by weight) fitted the plot, and no value of H I I I ~ / ( H I I IHrva) ~ was obtained that was different from the weight-fraction of IIIa by more than 27,. The accuracy of the analytical method was not affected by the addition to the solutions of N-(2-bromoallyl)ethylamine, t, = 1550 sec., or N-(2-~hloroallyl)ethylamine, tr = 845 sec., or by variation of the sample size from 2 t o 10 p1." For solutions containing 90.899.17, ether by weight, the sum of the weight fractions of IIIa Hrva and IVa was equal to (2.4 f 0.2) ( H I I I ~ H1va)/(Hiir. I f e ) ,where H e is the normalized height of the ether band. Description of a typical small scale reaction and analysis follows. A slurry of potassium amide was prepared by the addition of 2.38 g. of potassium to 100 ml. of ammonia and 12 mg. of anhydrous ferric chloride contained in a 250-ml. flask equipped with a sealed mechanical stirrer and Dry Ice condenser protected with a tube containing soda lime. N-(2-Bromoallyl)ethylamine ( I a , 8.2 g.) was added to the stirred slurry from a syringe in 2 min. After 4.5 hr., the Dry Ice had been allowed to evaporate, and the condenser was charged with an ice-salt mixture. Ether (50 ml.) and water (50 ml.) were added cautiously in that order. The ammonia was allowed to evaporate, the phases were separated, and the aqueous solution was extracted with 30 ml. of ether. The ether extracts were combined and dried with sodium
+
+
+
+
+
J. V. Rraun, M. Klihn. and J. Weismantel, A n n . , 449, 249 (19261. A standard curve for analysis of mixture of N-t-butylallenimine t i = 465 sec.) and N-t-butylpropargylamine (IVb, t i = 640 sec.) was also prepared from data obtained in a einiilar manner. The plot of weight fraction of IIIb u s . HIIIb/(HxiIb Hrvo) was also linear with a Slope of unity.
+
NOTES
SOVEMBER, 1963
+
3243
+
hydroxide. The average values of (Hrrrs H I V & ) / ( H I I X ~ dimethylnial~naldehyde~ have been reported, they were Hrva H e ) and Hiria/(Hrrra Hrv.) in 3 g.1.p. chromatograms prepared indirectly and, until r e ~ e n t l y could ,~ not be of the ether solution, which weighed 54 g., were (1.84 & 0.08) converted to the free dialdehydes. This paper reports 10-2 and 0.77 f 0.01, respectively, indicating that the conversion the first synthesis of negatively disubstituted malonwas 58% and that the product was 77% IIIa and 23% IT’a. aldehydes, dibromomalonaldehyde (11), and bromoMiscellaneous Reactions of N-(2-Ha!oa!!yl)ethylamines with Various Bases.-Treatment of 8.2 g. of S-(2-bromoally1)ethylchloromalonaldehyde (111). amine ( I a ) with 2 g . of sodium amide in 100 ml. of ether a t 25’ for Both of these compounds were prepared by the 5 hr. and room temperature for 24 hr., gave a 29cjC conversion to action of the appropriate halogen on the sodium salt of IT’a. No IIIa was detected by means of g.1.p.c. A 6.0-g. bromomalonaldehyde, under scrupulously anhydrous sample of N-(2-c.hloroallyl)ethylamine ( H a ) was treated in the condition^.^ same manner. Xo indication of the presence of IIIa or IT’a in the ether solution was obtained by g.1.p.c. Treatment of 36 g. of Ia with 300 ml. of N-methylmagnesium CHO bromide in ether a t reflux for 4 hr. and a t room temperature for + Xa Br+X X16 hr., gave a 10% yield of a mixture of 249’‘ IIIa and 76% IYa. Br%-z CHO Formation of IIIa and IYa was confirmed by means of infrared spectroscopy, and 29 g. (80%) of I a was recovered. N-(2Ch1oroallyl)isopropylamine (3,5 9.) was treated in the same Dibromomalonaldehyde was obtained in better manner. Traces of the corresponding I11 and I V (-90% IT-), yields and was studied in more detail. I t is readily estimated as
