DECEMBER, 1964
3687
NOTES
A mixture melting point with an authentic sample6 of the diacetate was not depressed. The solvent was removed from the filtrate and the resulting oil was dissolved in water and extracted with ether. The combined extracts were concentrated and gave 0.12 g. (4.7y0) of V, m.p. 144-146'. A mixture melting point with an authentic sample was not depressed. Reaction of 2-Mercaptoacetophenone (VI) with Potassium Ethoxide in the Presence of Sulfur.-To a suspension of 0.9 g. (0.03 g.-atom) of sulfur in 20 ml. of ethanol containing 1.08 g. (0.028 g.-atom) of potassium, was added 0.5 g. (0.0229 mole)
of VI. On addition of the mercaptan, the solution turned red and the suspended sulfur dissolved. After 2 hr. a solid began to precipitate from the reaction mixture. The suspension was allowed to stand a t room temperature for 72 hr. Analysis of a 1-ml. aliquot in the vapor fractometer indicated the presence of 30.2y0 acetophenone in the reaction mixture. Filtration of the suspension gave 1.46 g. (40.6%) of V, m.p. 144-146", m.m.p. 145-146". The filtrate, worked up in the usual manner,6 provided 0.3 g. (15.6%) of hydrogen sulfide (as lead sulfide) and 1.02 g. of a salt which was not identical with XXVIII and not further characterized.
Notes The Formation of N-(P-Anilinoethyl)-2-pyridone from the Action of N-Phenylethanolamine on 2-Bromopyridine RICHARD G. HISKEYAND JEROME HOLLANDER'
Venable Chemical Laboratory, The University of North Carolina, Chapel Hill,North Carolina Received July 10, 1964
As part of another program a sample of 2-(N-P-hydroxyethylani1ino)pyridine (I) was desired for investigation. A well-known route to compounds of this type2a-c involves displacement of a halogen by the appropriate amine. When 2-bromopyridine (11) was heated a t 200-250' with 2 equiv. of K-phenylethanolamine (111), a crystalline solid (analysis showed CIIH14NZO) was obtained in 41.5% yield. The structure of this product was assigned as N-(p-anilinoethyl)-2-pyridone(IV) on the basis of the following transformations. The alkyla-
converted IV to N-vinyl-2-pyridone (V) and N,K-dimethylaniline. Acylation of IV with acetyl chloride in benzene provided the hydrochloride of N-(@-acetanilinoethyl)-2pyridone (VI) which could be converted to N-(P-acetanilinoethyl)-2-pyridone (VII) with either silver oxide or triethylamine. The amide VI1 could also be prepared directly from IV by acylation with acetic anhydride in the presence of base. Hydrolysis of VI1 with either acid or base provided IV. Several mechanistic possibilities can be proposed to explain the rather unusual course of the alkylation reaction. The 2-alkoxypyridines are known3 to rearrange to N-alkyl-2-pyridones when heated. Caldwell and Schweiber4 have reported that a similar rearrangement of VI11 afforded I X and suggested that the transformation O
z
N - o OCH2CHCHzNPhth
I
NaOH H2O
OCOCH3
VI11
/
I1
J
I
O~N ~ N-H C H ~ C H C H ~ O H
I
OH
CHzCHzOH 1
IX
was analogous to the Smiles r e a ~ ~ a n g e m e n tThus . ~ ~ ~0alkylation of I11 followed by rearrangement may have , o occurred. However, the initial formation of a 0CH= CH, CH~CH~NHC~H & ~H ~ C H ~ - N - C ~ H ~ alkylation product is considered unlikely since NI V IV rn phenylethanolaniine normally provides N-alkyl derivacii, tives. Alternatively I may have been produced and VI, hydrochloride rearranged to IV via either intermediate X or XI. VII, free base Compounds similar to X have previously been reported2 in reactions of this type while decomposition of XI tion product, IV, reacted smoothly with dry hydrogen would provide the 2-alkoxypyridine. KO decision chloride and phenyl isothiocyanate to provide the corbetween these possible pathways can be reached from responding amine derivatives. A Hofmann degradation the present data. degradation
I
(1) Abstracted in part from a dissertation by J. Hollander submitted in partial fulfillment of the requirements for the Ph.D. degree t o the University of North Carolina, June, 1959. (2) (a) H. S . Mosher "Heterocyclic Compounds," Vol. I . , R. C. Elderfield, Ed.. John Wiley and Sons, Inc., New York, N. Y . , 1950,p. 397; (b) N. Weiner and I. A. Kaye, J . Org. Chem., 14, 868 (1949); (c) A. P. Gray, D . E. Heitmeier, and E. E . Spinner, J . A m . Chem. Soc., 81,4351 (1959).
(3) K. B. Wiberg, T. M. Shryne, and R. R. Kintner [ibid., 79, 3160 (1957)l have shown that the reaction is probably intermolecular and may involve a radical chain mechanism. (4) W. T. Caldwell and G. C. Schweiber, ibid., 74, 5187 (1952). (5) W. Evans and S. Smiles, J . Chem. S o c . , 181 (1935). (6) J. F. Bunnet and R. Zahler, Quort. Reu.. 69, 273 (1951).
3688
NOTES H+
I1 t 1 1 1 4 1
k
i H+
+N' Br-
Y
y/c 6Hs + HzO
dH2- CH2
X
XI
J. CHTCH~
Experimental7 Preparation of N-(p-Anilinoethyl)-2-pyridone(IV).-A mixture of 47.4 g. (0.3 mole) of 2-bromopyridine and 82.2 g. (0.6 mole) of N-phenylethanolamine were heated a t reflux for 6 hr. The solution was diluted with water and extracted with chloroform; the aqueous layer (pH 6) was made alkaline with 50% sodium hydroxide solution. The brown oil which separated solidified on standing. Two recrystallizations from carbon tetrachloride provided 26.8 g. (41.570) of IV as the one-half hydrate, m.p. 109.5-110". A n a l . 'Calcd. for C I ~ H I ~ N Z O . O . ~ HC, ~ O 69.92; : H , 6.77; N, 12.55. Found: C, 69.97; H , 6.49; N, 12.55. A 2.0-g. sample of 11. was refluxed for 39 hr. in benzene under a Dean-Stark trap. Anhydrous I\' was obtained as white needles, m.p. 108-109". A n a l . Calcd. for CL3Hl4N20: C, 72.87; H , 6.58; N, 13.08. Found: C, 72.82; H , 6.58; N, 13.13. The infrared spectrum of anhydrous IV exhibited peaks at 1653, 3340, 1513, and 1256 cm.-'. The ultraviolet spectrum in ethanol exhibited absorption maxima a t 298 mp ( e 2375), 247 (4575), and 237 (4280). Similar spectra have been observed for other 2-pyrid0nes.~ Treatment of IV with phenyl isothiocyanate and recrystallization of the resulting solid from carbon tetrachloride provided the phenylthiourea derivative of IT', m.p. 144-146'. A n a l . Calcd. for C20HlJaOS: C, 68.74; H , 5.48; N, 12.03. Found: C, 68.61; H , 5.37; N , 12.10. Hofmann Degradation of 1V.-To 1.30 g. (0.006 mole) of IV in 25 ml. of ethanol was added 1.7 g . (0.012 mole) of potassium carbonate and 2.56 g. (0.018 mole) of methyl iodide. The mixture was refluxed for 31 hr. and additional potassium carbonate and methyl iodide were occasionally added. The reaction mixture was filtered and evaporated to provide the yellow quaternary salt. The salt was dissolved in water and treated with 5.0 g. of silver oxide. The solution was stirred for 1 hr., filtered, and slowly distilled. The distillate was extracted with ether to provide 0.6 g. of N,S-dimethylaniline: b . p . 193.9', n M D 1.5583; lit.g b.p. 192.5-193.5, n 2 0 ~1.5582. The infrared spectrum of the compound was identical with that of authentic N,N-dimethylaniline. The distillation residue was extracted with ether. Removal of the solvent provided an oil which crystallized to yield 0.9 g. of V, m.p. 118-121'. Sublimation raised the melting point to 121123", lit.'Om.p. 119-122". The mercury(I1) chloride salt melted a t 193-195", lit.lo m.p. 193". Preparation of N-(p-Acetanilinoethyl)-2-pyridoneHydrochloride (VI).-To 7.7 g. (0.036 mole) of 11' in 600 ml. of benzene was added 3.14 g. (0.04 mole) of acetyl chloride. The addition was carried out a t 5". Crystallization of the precipitate from ethanol-ether provided 8.5 g. (80.6%) of VI, m.p. 118-124". ilnal. Calcd. for ClsH,7C1N202: C, 61.53; H, 5.85; N, 9.57. Found: C, 61.46; H , 5.95; N , 9.66. (7) hlelting points and boiling points are uncorrected. Elemental analysea were by Micro-Tech Laboratories, Skokie, 111. (8) J. A. Rereon and T. Cohen, J . A m . Chem. Soc., 78,416 (1956); F. Ramirrz and A. P. Paul, J . Org. Chem.. 19,183 (1954). (9) E. H. Rodd. "Chemistry of Carbon Compounds," Val. 111, Elsevier Publishing Co.. Amsterdam. 1954,p. 174. (10) S. Okhi. J. Pharm. SOC. J a p a n , 70,101 (1950).
VOL. 29
Titration of VI with standard potassium hydroxide gave d neutralization equivalent of 245 (calcd. for C1~H&lN202,292.8). The pK, value of the hydrochloride salt VI was determined as 2.65. Titration of VI with standard silver nitrate solution, using a potassium chromate indicator, provided a molecular weight of 267. Preparation of N-( 8-Acetanilinoethy1)-2-pyridone (VII).-A solution of 1.0 g. (0.0034 mole) of VI in 30 ml. of a 1 0 : l etheralcohol mixture was treated with 0.34 g. (0.0034 mole) of triethylamine. Filtration of triethylamine hydrochloride followed by evaporation of the solution provided an oil which crystallized. Recrystallization from carbon tetrachloride provided 0.83 g. (95.47,) of VII, m.p. 100-104°. When IV was allowed to react with 25 ml. of acetic anhydride followed by treatment with dilute sodium hydroxide, a brown oil was obtained. The oil was extracted with ether, the extract was evaporated, and the oil was allowed to crystallize. Recrystallization of the yellow solid from carbon tetrachloride provided 1'11, m.p. 104-105', identical in all respects with the sample obtained from VI. Hydrolysis of N-(P-Acetanilinoethyl)-2-pyridone(VII).-A 2.0g. (0.007-mole) sample of VI1 was refluxed with 50 ml. of 257, sodium hydroxide solution for 26 hr. The brown oil was extracted with chloroform and the extracts were evaporated to provide 1.40 g. (93.3y0) of IT'as yellow needles, m.p. 109-111". A mixture melting with a sample of IV prepared from I1 and I11 was not depressed. Acid hydrolysis of VI followed by neutralization provided a i3.3YOyield of Is', m.p. 108-110".
Exchange of Magnesium between the Ethyl Grignard Reagent and Magnesium Bromidela DWAINE0. COWAN, JOHN Hsn,Ib
AND
JOHN D. ROBERTS"
Contribution N o . 3148 from Gates and Crellin Laboratories of Chemistry, California Institute of Technology, Pasadena, California Received August 4, 1964
Much has been said but little settled about the precise nature of the Grignard reagent. The X-ray structure studies of Stucky and Rundle,2the vapor pressure studies of Vreugdenhil and B l ~ m b e r g ,the ~ work of Ashby and B e ~ k e rand , ~ the far-infrared experiments of Salinger and Mosher5 all indicate that the Grignard reagent can exist, a t least in part, as a monomer (RMgX) under a variety of experimental conditions. However, the nuclear magnetic resonance studies of Roos and Zeil,6 the association studies of Slough and UbbeIohde' (concentrat>ed solutions), plus the Grignard reaction studies of several groups (Bikales and Becker,8 Mosher and co-w~rkers,~ House and Traficante, lo Anteunis and D'Hollander," Hanielin,I2 and Dessy and (1) (a) Supported by the National Science Foundation: (b) National Science Foundation Undergraduate Fellow. summer. 1963: (0) to whom correspondence should he addressed a t the Department of Chemistry. The Job-s H pkins University, Baltimore, M d . 21218. (2) G. D. Stucky and R. E. Rundle. J . Am. Chem. Sac.. 86, 1002 (1963). (3) A. D. Vreugdenhil and C. Blomberp. Rec. trao. chim.,82, 453, 461 (1963). (4) E. C. Ashby and W. E. Becker, J. A m . Chem. Sac., 86, 118 (1963). (5) R. M. Salinaer and H. S. Mosher, ibid., 86, 1782 (1964). ( 6 ) H. Roos and W. Zeil, Z . Elektrochem.. 67. 28 (1963). (7) W. Slough and A. R. Ubbelohde. J . Chem. Soc.. 108 (1955). (8) N. M. Bikales and E. I. Rrcker, Can. J . Chem.. 41, 1329 (1963). (9) D. 0.Cowan and H. S. Masher, J . Org. Chem.. 28, 204 (1963); 27, 1 (1962): J. hliller, G. Gregoriou. and H. S. Moshcr, J. A m . Chem. SOC.. 89,3966(1961). (10) H. 0.House and D. D. Traficante. J . Ore. Chem.. 28,355 (1963). (11) hl. J. Anteunis and R. D'Hollander, Tetrahedron Letters, 1275 (1962). (12) R. Hamelin. Bull. m c . chim. Fvance, 915 (1961).
NOTES
DECEMBER, 1964 Salinger13) all indicate that the Grignard reagent in solution either consists in part of a dimeric species (RzT\lg.llgXn)or contains a mixture of dialliylmagnesium and magnesium halide ( R N g X’IgX2). Dessy and co-workers14have studied the exchange of magnesium between Grignard species using isotopic labeling techniques. Magnesium bromide f roni niagnesium-28 and bromine was mixed with diethylniagnesium prepared by the dioxane-precipitation method. Precipitation of the magnesium bromide from the solution with dioxane followed by isotopic analysis showed only 6 to 8% exchange of the labeled magnesium. From this, it was concluded that “the ethyl Grignard reagent is better represented by a coniplex EtNg.MgBrz than by EtMgBr” and that equilibrium 1 in ether is “at
+
2EthlgBr
e EtJlg + RIgBrn or (Et2Mg.hlgBrz)
(1)
most a very unimportant side reaction.” In a later paper15 it was reported that substitution of the stable isotope magnesium-25 for magnesium-28 gave a statistical distribution of magnesium but that this exchange could be caused by an impurity in the magnesium-%. We have carried out a somewhat different type of exchange experiment with high-purity magnesium-25 and observed statistical exchange. Magnesium-25 bromide, prepared from magnesium-25 and ethylene dibromide, was added directly to a Grignard reagent prepared from Dow Grignard magnesium and ethyl bromide in ethyl ether. This procedure precludes questions as to whether a mixture of diethylmagnesium and magnesium bromide is equivalent to a Grignard reagent and furthermore avoids having impurities such as dioxane which might arise in the preparation of the diethylniagnesiuni. After 1.5 hr., the magnesium bromide was precipitated from the solution with dioxane washed wit’h ether and converted to magnesium oxide; the proport’ions of magnesium-24, -23, and -26 were determined by mass spectrographic analysis. The results indicate that statistical equilibrium is achieved among the various forms of magnesium present. Therefore, if the Grignard reagent is to be represented as Etzi\Ig.;\lgBrz, exchange of magnesium-25 occurs not only with the complexed niagnesium bromide, but also with the dialkylmagnesium. The impurities in Dow Grignard grade magnesium and the impurities in the magnesium-25 used are listed in Table I. It must be concluded that under these conditions exchange does in fact take place, either by way of the equilibrium 1 or some equivalent process. After the completion of the above experiments, Dessy and ~ o - w o r k e r spublished ~~ the results of a number of exchange studies that are in general agreement with ours save for those carried out with one maverick variety of pagnesium, Dow atomized shot, which was reported to give no exchange. The unusual character of this sample of the metal in the exchange is not understood; however, as will be reported in more detail elsewhere, 3,3-diniethylbutylmagnesium chloride prepared from Dow at,omized shot has been shown by Dr. George (13) R. E. Dessy and R. M. Salinger. J. Am. Chem. SOC., 83, 3530 (1961). (14) R. E. Dessy, G. S. Handler, J. H . Wotiz, and C. A. Hollingsaorth, ibzd., 79, 3476 (1957). (15) R. E. Dessy and G. 8. Handler, ibid., 80, 5824 (1958). (16) Mms spectrographic analyses were performed by the Mass Spectrometer Laboratory, Oak Ridge National Laboratory. (17) R. E. Dessy, S. E. I. Green, and R. h‘l. Salinger, Tetrahedron Letters, NO. 91,1369 (1461).
3689 TABLE I IMPURITIES I N MAGNESIUM
Element
Dow Grignard grade,a %
MgZs,*7%
0.02 A1 0,005 Ca ... 0.01 0.01 cu 250 125 15.6 125 62.5 3.90 >250
Spectrum: Ai: characteristic absorption for anhydride group a t 5.45 and 5.68 M for carbon-oxygen stretching. B. With Water.-N-Chloroformyl-X-phenylglycine (2.14 9.) in 50 ml. of dioxane was added to 150 ml. of water a t 5' and stirred. A fine, white precipitate formed immediately. After 5 min., the mixture wae extracted with 500 ml. of ethyl acetate. The extract was concentrated in vacuo to yield 3-phenyl-2,5oxazolidinedione (0.65 g., 3770), m.p. 137-139". Anal. Calcd. for CgH?N03: C, 61.01; H, 3.96; N, 7.90. Found: C, 60.73; H , 3.80; N, 8.09; C1, 0.4. Infrared spectrum was the same as that given in A. Further crops of the anhydride contained increasing amounts of the chloroformyl compound. Prolongation of the reaction time should improve the yield. C . With Heat.-A small amount of S-ch1oroformyl-Nphenylglycine was heated a t 100" in an Abderhalden dryer a t atmospheric pressure for 4.5 hr. The weight of the product recovered was 96% of that expected from a conversion to 3phenyl-2,5-oxazolidinedione,m.p. 135-137". An infrared spectrum of this material further confirmed the identification. Cyclization takes place during the melting point determination, since melting points of about 140" for the K-chloroformyl derivative and the anhydride must result from the cyclization of the former to the latter during the heating process in the melting point bath. Potassium 6-(2-Anilinoacetamido)penicillinate. A.-N-Chloroformyl-N-phenylglycine (2.14 g., 0.01 mole) in 50 ml. of dioxane was added to a stirred solution prepared by adding 6aminopenicillanic acid (1.08 g., 0.005 mole) to 100 ml. of water, adjusting p H to 6.0 with dilute sodium hydroxide, diluting to 150 ml., and cooling to 5 ' . After 20 min. a t this temperature, the pH was adjusted to 7.0, and the solution was bioassayed and freeze dried to give 3.67 g. of product (384 y of ampicillin equivalent/mg.) . l l The crude product was purified by dissolving 2.00 g. in 6 ml. of water containing 1 g. of ammonium sulfate, adjusting the pH to 2.5-3.0 with 870 phosphoric acid, extracting the free acid with two 20-ml. portions of amyl acetate, and adding 3.2 ml. of 2 M potassium acetate in 90% isopropyl alcohol. Storage a t - 10" precipitated the purified potassium salt, which was collected, washed with 90% isopropyl alcohol, and dried in vacuo. The yield was 0.25 g. (825 y of ampicillin equivalent/mg.). Anal. Calcd. for CleH,,KN,OS: C, 49.63; H, 4.68; K, 10.10; N , 10.85; S, 8.28. Found: C, 49.16; H , 4.71; K , 9.84; N , 10.09; S, 8.5. Spectrum: : :A: 3.01 (N-H), 5.62 (p-lactam C=O), 5.97 (amide C=O), 6.25 and 7.18 p (-Cot-). By calculation from the bioassay of the reaction mixture and the activity of the pure potassium salt, the penicillin yield was 97%. Substantiation was by paper chromatographic data. Bioautography of the reaction mixture chromatographed in the system butyl alcohol-sec-butyl alcohol-acetone-water (12 : 12 : 10:9, v./v.) by descending chromatography with Whatman KO. 1 (11) R y disk bioassay against Staphylococcus aureus ATCC 6538P in terms of an ampicillin [6-(~-2-amino-2-phenylacetamido)penicillanic acid] standard.
DECEMBER, 1964
NOTES
paper impregnated with 0.1 M pH 6 potassium phosphate buffer showed only one round zone of activity against both Staphylococcus aureus and Escherichia coli plates at Rr 0.65-0.68. The pure salt in tube serial dilution inhibited Bacillus subtzlis ATCC 6633 and Staphylococcus aureus ATCC 6538P a t 0.122 y/ml. When it was tested against various microorganisms by agar serial dilution, the results shown in Table I were obtained. B.-The penicillin could also be made in 380/, yield by shaking equivalent amounts of N-chloroformyl-N-phenylglycine and sodium 6-aminopenicillinate in ethyl acetate for 5 min. C.-3-Phenyl-2,5-oxazolidinedione (3.54 g., 0.02 mole) in 100 ml. of dioxane was added to a stirred solution of sodium 6aminopenicillinate (2.38 g., 0.01 mole) in 300 ml. of water a t 5 " . After 2 hr. a t 5' the reaction was stopped, and the crude penicillin was isolated as in A. The yield was 97%, Rr 0.68.
The Adduct of Triphenylphosphine and Maleic Anhydride C. OSUCH,J. E. FRANZ, A N D F. B. ZIENTY Research Department, Organic Chemicals Division, Monsanto Company, St. Louis, Missouri 68177 Received June 8, 1964
The reports of Chopard and Hudson' prompt us to report work which confirms their results. The reaction of triphenylphosphine with maleic anhydride in an inert solvent gave an adduct which, on recrystallization from benzene, agreed in nielting point and analysis with earlier results.2 Since the infrared spectrum did not conform to that expected for the proposed2 structure I, other possibilities were considered (11-IV). Aksnes has proposed3 I11 to be the correct formulation based on infrared evidence alone.
+
372
Ph3PHBr I I
+
+
R0zC--CH2-CH-PPh3
ROzCCH=CHCOzR
Br
1
COzR
4
base
R02CCHzC=PPha
1
COzR Va, R = CzHs b, R = CH3
Since I11 was the only tenable structure remaining, the adduct was related to a known substance by the following route. The reaction of triphenylphosphine hydrobromide with diethyl fumarate followed by treatment with base has been reported' to give Va. When dimethyl fumarate was used, Vb was obtained in 69% yield. Treatment of the adduct I11 with methanol8 followed by diazomethane also gave Vb in 82.5% yield. Furthermore, Vb has been previously prepared by an alternate We have also found that chloromaleic anhydride reacts with triphenylphosphine to give an adduct having the same characteristic infrared spectrum as the maleic anhydride adduct. Citraconic anhydride did not react in this fashion but gave, instead, an unidentified, red solid which cont'ained only one carbonyl band in t'he infrared spectrum. Experimental
Triphenylphosphine-Maleic Anhydride Adduct .-Equimolar amounts of triphenylphosphine and maleic anhydride solutions in benzene were mixed with stirring. The precipitated product was washed and dried to give a crude, orange, amorphous-looking powder in 92.8% yield.. An analytical sample was obtained from benzene, m.p. 162.5-163.A" dec., lit.*m.p. 160" dec. Anal. Calcd. for CzzHL703P: C, 73.33; H, 4.75; P , 8.60; equiv. wt., 360 (monobasic), 180 (dibasic). Found: C, 73.58; H, 4.78; P , 8.39; equiv. wt., 366. The infrared spectrum (mull in mineral oil) showed two strong carbonyl bands a t 1682 and 1787 em.-'. In KBr these bands are reported3 to appear at 1702 and 1805 em.-'.
TriphenylphosphinecarbomethoxymethylcarbomethoxymethylI
I11
I1
IV
The adduct, showed the following properties. Titration of a dioxane solution with aqueous base indicated a monobasic acid.4 Decomposition a t t'he melting point resulted in formation of triphenylphosphine. The P31n.ni.r. spect'runi showed a peak at - 13 p.p.m. relative to phosphoric acid. Compounds containing five groups attached to a phosphorus atom have been reported5 to show a large positive shift. The proton magnetic resonance showed two nonvinylic protons6 and 15 aromatic hydrogen atoms. F. Hudson and P. A. Chopard, Helu. Chim. Acta, 46, 2178 (1963); (bl P. A. Chopard and R. F. Hudson, Z. Naturforsch., lSb, 509 (1) (a) R.
(1963). (2) A . Schonberg and A. F. A. Ismail. J . Chem. Soc.. 1374 (1940). ( 3 ) G . Aksnes, Acta Chem. Scand., 15, 692 (1961). (4) On the basis of 111, this can be rationalized b y assuming t h a t the half-acid sodium salt (PhaP=C-CH?CO*H Na or PhaP=C(CO%H)+
I
cos CHe-COz-Na +) can tautomerize to a stable zwitterion (PhaP +-CH(COQ-)-CHz-C02-Na+) which may exist as either a n open chain as shown or as a cyclic lactone-type structure containing phosphorus in the ring. (5) R. A. 1 ' . Jones and A. R. Katritzky, Angew. Chem., Inlern. Ed., 1, 32 (1962).
(6) The nonaromatic protons appear a s a single peak. Closer investigation disclosed that a 0.Ebc.p.s. coupling was present. This is in essential agreement with the report of Hudson and Chopard.1'
ene (Vb).-Methanol (2 nil.) was added to 208 mg. of adduct I11 and the solution refluxed 15 min. After cooling, an ether solution of diazomethane was added until no further signs of reaction occurred and the mixture had a definite yellow color. The solvent was then removed under reduced pressure, and the resulting oil was triturated in ether to form a tan solid. One recrystallization from chloroform-n-hexane gave 193 mg. of product, m.p. 156-158" (lit.9 m.p. 157-158') alone and 163165" when mixed with an authentic sample prepared as follows. To 721 mg. of dimethyl fumarate in 10 ml. of acetonitrile was added 1.716 g. of triphenylphosphine hydrobromide (prepared by bubbling HBr into an ether solution of triphenylphosphine), and the mixture refluxed 30 min. The mixture was diluted with water and extracted with ether; the aqueous layer was neutral___
(7) H. Hoffmann, Chem. Ber.. 94, 1331 (1961).
(8) The resulting half-ester was not purified, but the crude material had the reported'" infrared spectrum. The proton n.m.r. spectrum of this crude material did not show the expected CH-CH, splitting required for the structure proposed by Hudson and Chopard.' Instead, a doublet ( J = 15 c.p.8.) arising from the CHz group and split by P H coupling was obaerved a t 2.86 p.p.m. ( T M S = 0 ) . The methyl group appeared a t 3.3 p.p.m. as a sharp singlet. The H atom presumed to be on oxygen did not give a discernible peak. (9) €1. J. Bestmann and H. Schulz, Chem. Ber., 95, 2921 (1962). (10) The proton n.m.r. spectrum of Vb showed a doublet ( J = 17 c.p.8.) a t 2.93 (2H). a sharp peak a t 3.43 ( 3 H ) . a broad hand a t 3.28 p.p.m. ( 3 H ) , and the expected aromatic hand (15H). We have no explanation for the anomalous behavior of one of the methyl groups of this diester nor for the absence of P H splitting in the anhydride I11 despite the apparent presence of such splitting in the mono- and diester. T h a t the observed splitting in Vb ( J = 17 c.p.8.) is due to P H coupling and not nonequivalent H's was confirmed by the fact t h a t a t 100 Mc. the coupling was unchanged ( J = 17.3 c.p.8.).
NOTES
3722
ized with 2 N NaOH. The resulting solid was filtered, washed, and dried to give 1.405 g., 69.270, of crude product. An analytical sample was obtained from chloroform-n-hexane, m.p. 165.5167' Anal. Calcd. for Cz4Hz3O4P: C, 70.93; H , 5.70; P, 7.62; mol. wt., 406. Found: C, 70.95; H, 5.91; P , 7.69; mol. wt., 379. .The infrared spectra of the esters prepared by the two methods were superimposable and showed two strong carbonyl bands at 1725 and 1610 cm.-l. Reaction of Other Anhydrides with Tripheny1phosphine.Ethereal solutions of equimolar amounts of chloromaleic anhydride and triphenylphosphine were mixed to give a purple precipitate in 94% yield. The infrared spectrum of the crude product showed absorption bands a t 1775 and 1705 cm.? (in mineral oil). Efforts to obtain pure material by recrystallization have been unsuccessful. When ether solutions of citraconic anhydride (1.12 g.) and &phenylphosphine (261 g.) were mixed, no visible reaction occurred even after refluxing and then allowing the mixture to stand 4 hr. The solvent was removed under vacuum and the tan crystalline mass was allowed to stand overnight. The resulting mixture was deep red and, after washing with ether, gave 1.53 g. of a red, crystalline solid, m.p. 100-103.5" dec. The infrared spectrum (mineral oil) showed a strong band a t 1770 cm.-l. Unreacted triphenylphosphine (2.05 g.) was recovered from the ether washings.
VOL. 29
The four possible dibromobenzofurazan oxides were synthesized and compared with these two products. The synthesis of IIa, IIb, and IIc was accomplished by ring closure in refluxing toluene of the corresponding azidodibronionitrobenzenes (IIIa, IIIb, and IIIc) which were prepared by diazotization3 of the corresponding dibromo-o-nitroanilines and treatment of the diazonium salts with sodium azide. 5,6-Dibromobenzofuraza11 oxide was prepared from 4,5-dibronio-2-nitroa1dine~by oxidation with alkaline hypochlorite.6 The products melting at 148-149' and 92.3-93' were found to be identical with synthetic 4,s- and 4,6-dibromobenzofurazan oxide (IIa and IIb), respectively, by mixture melting point and comparison of their infrared spectra. The four dibroniobenzofurazan oxides were converted to the corresponding furazans by reduction with hydroxylamine in alkaline solution followed by steam distillation. The effect of the nucleophile and the solvent on the isomer distribution produced in the dehydrobromination of I is currently under investigRtion. Experimental'
The Structure of Two New Dibromobenzofurazan Oxides from the Dehydrobromination of Tetrabroniotetrahydrobenzofurazan Oxide WILLIAMM O J E ~ Department of Plant Pathology, University of California, Riverside, California Received June 19, 1964
During an investigation of substituted benzofurazan oxides for fungicidal activity, it was found that treatment of tetrabromotetrahydrobenzofurazan oxide (I) with several nucleophiles yielded two dibromobenzofurazan oxides which had not been described previously.
Br
x 3
I
I1
xs
111
a, XI = Br, XZ = XB = H b, XZ = Br, XI = XB = H c, X3 = Br, XI = XZ = H
Treatment of I with aqueous alkali has been shown to produce a dibroniobenzofurazan oxide melting a t 132°.2a The structure was originally postulated as I I b but was later established as 1Ic by reduction to the known 3,6-dibrom0-0-phenylenediamine.~~ I n the present study, when I, m.p. 169-170', was treated with pyridine or potassium acetate in glacial acetic acid, a niixture of two new dibromohenzofurazan oxides was formed. One of these had n1.p. 148-149' and the other had m.p. 92.5-93'. (1) Deceased Aug. 12, 1964. '. A. M. Edwards, and E. R. Steiner. J . Chem. 12) (a, D. L. Hammick. W Soc.. 3308 (1931): (b) J. H. Boyer. U. Toggweiler, and G. A. Stoner, J . A m . Chem. SOC., 79, 1748 (1957).
Preparation of Compounds. Dibromo-o-nitroanilines.-5,6Dibromo-2-nitroaniline was prepared in a four-step reaction sequence: o-bromoaniline -* 2-bromo-6-nitroaniline8~8 + 2,3-dibromonitrobenzene~~10 -* 2,3-dibromoaniline11 -* 5,B-dibromo-2-nitroaniline . I z The last step in the reaction sequence was conducted as follows. 2,3-Dibromoaniline was acetylated to give 2,3-dibromoacetanilide, and 7.0 g. (0.024 mole) of the latter was slowly added with vigorous agitation to 15 ml. of fuming nitric acid (sp. gr. 1.51) a t -5". After 1 hr. atO" theorange solution was added dropwise to an icewater mixture to produce a yellow precipitate, which was filtered, washed with water, and recrystallized from 95' ethanol to give 3.27 g. (40%) of 5,6-dibromo-2-nitroacetanilidelz as fine white needles, m.p. 235-237" dec. A suspension of 4.69 g. (0.014 mole) of the latter in a mixture of 50 ml. of 95% ethanol plus 12 ml. of concentrated hydrochloric acid was refluxed for 48 hr. The suspension was diluted with hot 957; ethanol (ca. 30 ml.) to produce a solution which on cooling deposited 3.88 g. (94Y0) of 5,6-dibromo-2-nitroaniline as yellow needles, m.p. 150.5-151°, lit.12m.p.149'. 4,5-Dibromo-2-nitroaniline was prepared in 75% yield via the ethyl nitrate nitration of 3,4-dibromoacetanilide to 4,5-dibromo2-nitroacetanilide,4 followed by acid hydrolysis as above, forming orange needles, m.p. 203.5-204.5O, lit.4m.p. 204-205'. 4,6-Dibromo-2-nitroaniline was prepared in 84y0 yield by the bromination of o-nitroaniline in glacial acetic acid, forming orange needles, m.p. 127-128" (from 95% ethanol), lit.'$ m . p 127-128". 3,6-Dibromo-2-nitroaniline was prepared by Austin's procedure,14 heating 2,3-dinitro-l,4-dibromobenzenein a sealed tube with ammonia for 3 hr. at 100" and, much more conveniently, by passing ammonia through a suspension of the compound in ethanol. Thus, anhydrous ammonia was bubbled through a suspension of 2.0 g. (6.1 mmoles) of 2,3-dinitro-1,4-dibromoben-
(3) H. H. Hodgson and J. Walker, J. Chem. Soc.. 1620 (1933). (4) F. H. Case, J. Ore. Chem., 16, 941 (1951). (5) A. G. Green and E. M . Rowe, J . Chem. Soc.. 101, 2452 (1912). (6) T. Zincke and P. Schwars, Ann.. 307, 28 (1899). ( 7 ) Melting points are uncorrected and were determined on a ThomasHoover melting point apparatus, spectra u.ere taken with a Perkin-Elmer Model 137-B infrared spectrophotometer, and analyses were by Elek Microanalytical Lttboratories. Torrance, Calif. (8) C . S. Gibson and J . D. 4 . Johnson, J . Chem. Soc.. 3092 (1928). (9) H. Franzen and E. Engel. J . prakt. Chem.. 102, 156 (1921). (10) A. F. Hollaman, Rec. trau. chim., Z7, 156 (1908). (11) G. Korner and A. Contardi, ktti accad. naz. Lincei Mem. Classe sci. fis. mat. e nat., [51 1 6 , 526 (1906). (12) G . Korner and A . Cuntardi. ibid.. [5] 16, 580 (1906). (13) C. L. Jackson and E. W.Russe. A m . Chem. J . , 36, 148 (1906). (14) P. T. Austin. B e r . , 9 , 621 (1876).
DECEMBER, 1964
NOTES
3723
solution gave 3.88 g. (69%) which waa chromatographed in nzene1611ein 80 ml. of refluxing absolute ethanol for 3 hr. After 48 hexane on silica gel. Evaporation of the eluate and crystallizahr. at 25' the orange solution was refluxed for an additional 3 hr. tion of the residue from n-hexane gave 5,6-dibromobenzofurazan while ammonia was passed through. Evaporation gave an orange as long colorless needles, m.p. 87-87.5". solid which was chromatographed in benzene on silica gel. A yellow band eluted rapidly and the eluate on evaporation yielded Anal. Calcd. for C6H2Br2N20:C, 25.93; H , 0.73; N , 10.08. 1.6 g. which on recrystallization from aqueous ethanol gave 1.49 g. Found: C, 26.27; H , 1.04; N, 9.78. (82%) of 3,6-dibromo-2-nitroaniline as yellow needles, m.p. 4,5-Dibromobenzofurazan was obtained in 43% yield from 73.5-74.5', lit.I4 m.p. 75". IIa, forming colorless needles from 95% ethanol, m.p. 123-124". Azidodibromonitrobenzenes.-The Hodgson and Walker proAnal. Calcd. for CeH2BrzNzO: C, 25.93; H , 0.73; N, 10.08. cedure3 for the diazotization of nitroamines was utilized for the Found: C, 26.02; H , 0.95; N,9.83. conversion of 5,6-, 4,6-, and 3,6-dibromo-2-nitroaniline to the cor4,6-Dibromobenzofurazan waa obtained in 55% yield from responding diazonium salts. Treatment of the diazonium salts IIb, forming long colorless needles from 9570 ethanol, m.p. 71.5with aqueous sodium azide led to the azidodibromonitrobenzenes. Thus, for the preparation of l-azido-5,6-dibromo-2-nitrobenzene 72". A n d . Calcd. for C6H2Br2N20:C, 25.93; H , 0.73; N, 10.08. (IIIa), 1.16 g. (0.017 mole) of finely pulverized sodium nitrite Found: C, 26.23; H , 0.93; N,9.98. was gradually added to 8 ml. of concentrated sulfuric acid with stirring a t 0". The suspension was heated to 50' and held a t this 4,7-Dibromobenzofurazan was obtained in 69% yield from temperature until the sodium nitrite dissolved. The solution was IIc, forming colorlessneedles from 95% ethanol, m.p. 112-112.5", cooled and a finely divided suspension of 3.88 g. (0.013 mole) of 1it.I m.p. 113". 5,6-dibromo-2-nitroaniline in 40 ml. of glacial acetic acid was Conversion of Tetrabromotetrahydrobenzofurazan Oxide (I) added gradually with stirring while the temperature was mainto Dibromobenzofurazan Oxide. A. With Pyridine.-A solutained below 15'. After the diazotization was complete, the pale tion of 4.56 g. (0.01 mole) of I, m.p. 169-17Oo,l in 25 ml. of yellow suspension wa8 added dropwise to a solution of 1.16 g. pyridine was allowed to stand a t 25' for 2 hr. The suspension (0.018 mole) of sodium azide in 20 ml. of water with vigorous stirwas poured into cold water and the resulting yellow precipitate ring a t 10". The resulting precipitate was filtered, washed with was filtered, washed with water, and dried. The yield was 2.62 water, and recrystallized from 9570 ethanol. The yield of pale g. (89% based on conversion to dibromobenzofurazan oxide). yellow, feathery needles, m.p. 71.5-72.5', was 3.73 g. (8870). Recrystallization twice from n-hexane and once from 95% ethanol A n d . Calcd. for C6H2Br2N402: C, 22.38; H , 0.63; Br, gave 0.70 g. of yellow needles, m.p. 148-149". This material 49.65. Found: C, 22.61; H , 0.84; Br, 50.07. was identical (mixture melting point and infrared spectrum) with 1-Azido-4,6-dibromo-2-nitrobenzene (IIIb) was obtained in 4,5-dibromobenzofurazan oxide obtained by the decomposition 767' yield in the same manner from 4,6-dibromo-2-nitroaniline, of IIIa as previously described. forming pale yellow needles from 957' ethanol, m.p. 52-53'. The hexane filtrates were combined and evaporated to a small C, 22.38; H, 0.63; N, 17.40. Anal. Calcd. for C6H2Br2N4O2: volume to yield 0.96 g., m.p. 70-89'. Recrystallization from Found: C,22.65; H,0.84; K , 17.23. aqueous dioxane and then from 9570 ethanol gave 0.55 g. of l-Azido-3,6-dibromo-2-nitrobenzene(IIIc) was obtained in yellow plates, m.p. 92.593'. This material waa identical (mix8670 yield from 3,6-dibromo-2-nitroaniline, forming colorless ture melting point and infrared spectrum) with 4,6-dibromobenzoneedles from n-hexane, m.p. 65.5-66.5". furazan oxide obtained by the decomposition of IIIb, previously Anal. Calcd. for c6H2BrZN4O2:C, 22.38; H , 0.63; Br, described. 49.65. Found: C, 22.31; H, 0.71; Br, 49.84. B. With Potassium Acetate in Acetic Acid.-A mixture of Dibromobenzofurazan Oxides.-5,6-Dibromobenzofurazan ox22.8 g. (0.05 mole) of I and 22 g. of potassium acetate in 200 ml. ide was prepared in 7670 yield by the sodium hypochlorite oxiof glacial acetic acid was refluxed for 24 hr. The suspension was dations of 4,5-dibromo-2-nitroaniline, forming yellow needles, poured into water and the resulting solid was filtered, washed m.p. 128-128.5' (from 95Y0 ethanol). with water, and recrystallized twice from 9570 ethanol and once Anal. Calcd. for C6H2Br2N202: C, 24.52; H, 0.69; N, 9.53. from acetic acid. The yield of high-melting isomer, m.p. 148Found: C, 24.73; H,0.75; K,9.56. 149', waa 4.4 g. The remaining dibromobenzofurazan oxides were prepared The acetic acid filtrate was evaporated to a small volume giving by the thermal decomposition of the corresponding o-nitroazides. 2.2 g., m.p. 77-80', which after recrystallization from methanol For example, a solution of 1.O g. of l-azido-5,6-dibromo-2-nitroweighed 1.4 g. Two recrystallizations from n-hexane gave an benzene ( I I I a ) in 25 ml. of anhydrous toluene was refluxed for 29 additional 0.17 g. of the high-melting isomer. The n-hexane hr. Evaporation gave a tan solid. Chromatography in benzenefiltrates were combined and evaporated to a small volume; the n-hexane ( 1 : l v./v.) on silica gel gave a yellow band, which resulting solid was recrystallized from aqueous dioxane and then eluted readily to yield 4,5-dibromobenzofurazan oxide (IIa), 95% ethanol. The yield of the low-melting isomer, m.p. 92.5which crystallized from 95% ethanol as pale yellow, feathery 93', wai 0.59 g. needles, m.p. 148-149". The yield was 0.80 g. (88%). Anal. Calcd. for C6H2Br2N2O2: C, 24.52; H, 0.69; N, 9.53. Found: C,24.57; H,0.81; N,9.90. 4,6-Dibromobenzofurazan oxide ( I I b ) was obtained in 82y0 Methyl P-Deoxy-4-rnercapto-~-ribofuranoside~ yield in the same manner from IIIb, forming canary yellow platelets from 95Y0 ethanol, m.p. 92.593". ROYL. WHISTLER, W. E . DICK,T. R. INGLE, Anal. Calcd. for C6H2Br2Nd&: C, 24.52; H , 0.69; N, 9.53. R. M. ROWELL, AND B. URBAS Found: C, 24.67; H, 0.79; N , 9.65. 4,7-Dibromobenzofurazan oxide (IIc) was obtained in 71% Department of Biochemistry, yield from IIIc, forming yellow prisms from 95% ethanol, m.p. Purdue University, Lafayetle, Indiana 133-133.5.', lit.* m.p. 132". Dibromobenzofurazans.-The dibromobenzofurazans were Received May 19, 1964 prepared by reduction of the furazan oxides with hydroxylamine followed by steam distillation of the alkaline solution of the dioximes.' For example, a solution 5.88 g. (0.02 mole) of 5,6-diThis 1abor.ttory and others have been engaged in bromobenzofurazan oxide in 100 ml. of hot 95% ethanol was producing modified sugars wherein the ring oxygen is cooled rapidly in an ice bath to produce a finely divided suspenreplaced by another atom such as sulfur, selenium, or sion. A solution of 2.2 g. (0.032 mole) of hydroxylamine hydrochloride in 12 ml. of water was added followed by 2570 aqueous nitrogen. Xlost work has dealt with the placement of potassium hydroxide with stirring and cooling until nitrogen sulfur in the sugar ring. Many of the thiosukars thus evolution ceased. Steam distillation of the deep red alkaline far produced are of biological interest, such as the ana(15) C. L. Jackson and D. F. Calhane, Am. Chem. J.,as, 451 (1902). (16) C. J. Sunde, G. Johnson, and C. F. Kade. J. Org. Chem., 4, 548 (1939).
(1) Journal Paper No. 2285 of the Purdue Agricultural Experiment Station, Lafayette, Ind. Presented in part a t the 1 4 7 t h National Meeting of the American Chemical Society, Philadelphia Pa. April 1964.
3724
NOTES
VOL. 29
of components on the paper was determined by spraying guide logs of D-glucose,2~ - f r u c t o s e~-xylose,~-' ,~ L-arabinose,8 strips cut from the edges and center of each paper. The bands ~ - r i b o s e , ~and . ~ 2-deoxy-~-ribose.~In all of these were excised and eluted with several portions of solvent. Purity sugars the sulfur atom is a part of a six-membered pyraof crystalline products was determined by thin layer chromanose ring. It is now of interest to prepare 4-deoxy-4tography on silica gel G1* coated microscope slides irrigated with mercapto-D-ribose since this sugar would be useful in ( E ) chloroform-acetone (12: 1 v./v.) or ( F ) butanol-ethanolwater ( 3 : l : l v./v.). Location of components was obtained by making nucleosides. A report of 4-deoxy-4-mercaptospraying with 570 sulfuric acid in ethanol and charring. Sugar L-ribose has been givenlo recently. flow rates are reported relative to that of D-ribose ( R , values). The starting material for the synthesis of the D Molecular weights were measured in a Mechrolab vapor phase analog is a-L-lyxose" which initially is converted to osmometer with water as solvent. methyl 2,3-0-isopropylidene-4-0-(p-tolylsulfonyl)-a-~- Methyl 2,3-O-Isopropylidene-4-0-(p-tolylsulfonyl)-~-~-lyxopyranoside (I).-This compound was made by the procedure of Kect lyxopyranoside (I) by the method of Kent and Ward12 D (c 2.81, and Ward.12 It had m.p. 105-106", [ ( Y ] ~ ~ +16.7" and also t o the methyl 2,3-0-isopropylidene-4-0ethan 01). (met hy lsulfony 1)-a-L-lyxopyranoside (11). .4nal. Calcd. for ClsHzz07S: C, 53.6; H , 6.19; S, 8.95. Displacements of the tosyloxy from I and the mesylFound: C, 53.3; H, 6.14; S, 9.2. Methyl 2,3-O-Isopropylidene-4-0-( methylsulfony1)-a-~-lyxooxy from I1 by thioacetate anion in N,N-dimethylpyranoside (II).-Methylsulfonyl chloride (5 ml., 0.063 mole) formamide (DMF) were compared. The ratio of sugar was dissolved in 30 ml. of dry pyridine and the solution was derivative to thioacetate was 1 :4 in both instances and cooled to -5 to - 10". To this was added, dropwise with stirreactions were conducted at an oil-bath temperature a t ring, a solution of methyl 2,3-0-isopropylidene-~-~-lyxopyrano154-156'. After 2 hr., compound I gave methyl sidelg (10.2 g., 0.05 mole) in 20 ml. of dry pyridine. The mixture was stirred at -5 to - 10" for 4 hr., then 25 ml. of water 4 -deoxy-2,3-0-isopropylidene-4 -thioacetyl -6 -D -ribopy was added slowly. The mixture was extracted four times with ranoside (111) in 50y0 yield while even after 3 hr. of 50-ml. portions of chloroform and the combined extracts were reaction time compound I1 gave only 30% of compound washed with water and dried over sodium sulfate. The dry 111. These results are as expected and confirm the chloroform solution was concentrated to a thick sirup which was dissolved in methanol and concentrated to dryness. The comgeneral observation that the tosyloxy group is a better pound was crystallized from hot ethanol to give 13 g. (92%) of leaving group than mesyloxy. Compound I11 showed compound 11: m.p. 92-93', [ c Y ] ~ ~ -14.7" D (c 1.3, ethanol). the characteristic a b ~ o r p t i o n 'for ~ thiolacetate at 230Anal. Calcd. for ClaH1807S: C, 42.5; H, 6.4; S, 11.4. 240 mp. Found: C, 42.3; H , 6.5; S, 11.2. Methanolysis of 111 gave sirupy methyl 4-deoxy-4Methyl 4-Deoxy-2,3-O-isopropylidene-4-thioacetyl-p-~-ribopyranoside (III).-Compound I1 (5.7 g., 0.02 mole) and 9.2 g., mercapto-D-ribofuranoside (IV) which was purified by 0.08 mole, of recrystallized potassium thioacetate were dissolved paper chromatography. The product showed no thiol in 70 ml. of dry DMF and the mixture was heated in an oil bath activity in the ultraviolet or with sodium nitroprusside a t 154-156" for 3 hr. The reaction mixture was poured with (SNP)l4,I5 or 2,3,5-triphenyl-2H-tetraxoliumchloride stirring into 300 ml. of xylene and the precipitated salts were filtered and washed with xylene. The combined solutions were (TTC) l6 reagents, indicating preferential ring formaevaporated under reduced pressure to dryness and the residue tion on sulfur. extracted twice with 100-ml. portions of heptane. The combined Deacetylation of I11 in methanolic sodium methoxide heptane extracts were concentrated to a volume of 15 ml. and produced the free mercapto derivative which was concooled. The crystals which formed were separated by filtration and suspended in 50 ml. of petroleum ether (b.p. 30-37"). verted to the disulfide by oxygen and iodine oxidation in The crystals which did not dissolve were filtered to give 0.5 g. refluxing methanol, The disulfide on treatment with unreacted compound I1 which waa used for further disAmberlite I R 120(H) gave bis(methy1 4-deoxy-P-~- of placements. The mother liquor, when concentrated to 10 ml., ribopyranoside) 4,4'-disulfide. deposited crystals. The crystalline product was recrystallized Experimental Analytical Methods.-Chromatographic separations of sugar derivatives were made a t 25" on Whatman Xo. 3 MM filter paper developed in irrigants (A) 1-butanol-ethanol-water (40: 11 : 19 v./v.) or (B) ethyl acetate:pyridine:water (25: 1 :25 v./v.). Spray indicators employed were ( C ) potassium permanganate periodate" and ( D ) silver nitratesodium hydroxide.lS Location (2) M. S. Feather and R. L. Whistler, Tetrahedron Letters, No. 16, 667 (1962). (3) M. S. Feather and R. L. Whistler, J . Or#. Chem., 48, 1567 (1963). (4) D. L. Ingles and R. L. Whistler, ibid., 21, 3896 (1962). (5) J. C. P. Schwarz and C. P. Yule, Proc. Chem. Soc., 417 (1961). ( 6 ) T. J. Adleyand L. N . Owen, ibid., 418 (1961). ( 7 ) R. L. Whistler, M. S.Feather, and D. L. Ingles, J . Am. Chem. Soc., 84,122 (1962). (8) R. L. Whistler and R. M. Rowell. J . O w . Chem.. 49, 1259 (1964). (9) C. J. Clayton and N. A. Hughes, Chem. Ind. (London), 1795 (1962). (10) E. J. Reist, D. E. Gueffroy, and L. Goodman, J . Am. Chem. Soc., 85, 3715 (1963). (11) R. L. Whistler and J. N. BeMiller, "Methods in Carbohydrate Chemistry," Vol. 1, R. L. Whistler and M. L. Wolfrom, Ed., Academic Press I n c . , New York, N. Y., 1962,p. 79. (12) P. W. Kent and P. F. V. Ward, J . Chem. SOC.,416 (1953). (13) H.P. Koch, ibid.. 387 (1949). (14) D. A . H. Morner, Z . physzol. Chem., 48, 594 (1899). ( 1 5 ) W. I. Patterson, W. B. Geiger, L. R. Mizell, and &' Harris, I. J. Research N a t l . Bur. Std., 2 1 , 89 (1941). D. P. Procter. and J. S. Harrison, Nature, 166, (16) W. E. Trevelyan, 444 (1950). (17) R. C . Lemieux and H. F. Bauer, A n a l . Chem., %6, 920 (1954).
several times from heptane giving a total yield of 1.42 g. (30%) of material: m.p. 91-92', [ C Y ] ~ ~-26.8" D (c 0.75, ethanol). The mixture melting point of I1 and I11 was 68". Anal. Calcd. for CllHlaOsS: C, 50.4; H , 6.9; S, 12.2. Found: C, 50.6; H, 7.2; S,11.9. This compound showed characteristic absorption for thiolacetate a t 230-240 m r and also gave a strong positive test nith TTC and SSP. Compound I11 was also prepared from compound I. Compound I (5.5 g., 0.154 mole) and 8 g., 0.07 mole of potassium thioacetate were dissolved in 50 ml. of dry DMF and the solution was heated in an oil bath a t 154-156" for 2 hr. It was worked up as described above with 200 ml. of xylene. Initial crystals from heptane had m.p. 76-85'. These were recrystallized several times from petroleum ether (30-37"). The final product had the same constants as given above. The yield was 2.0 g. (49.6%). Methyl 4-Deoxy-4-mercapto-~-ribofuranoside (IV).-compound I11 (1.0 g., 0.0038 mole) was dissolved in 10 ml. of 0.7 A' methanolic hydrogen chloride and the solution was refluxed for 24 hr. Thiol activity was determined periodically by testing an aliquot of the reaction mixture with TTC. N o thiol test was found after 24 hr. The reaction mixture was cooled and was neutralized by passing it through a column of Amberlite IR-45 cationic resin. The effiuent was concentrated under reduced pressure to a thin sirup (0.6 g.). The glycoside was purified by paper chromatography with irrigant A. Elution of a compound with R, 2.03 ( R , 2.1 in irrigant B) with water followed by (18) Brinkmann Instruments Inc., Great Neck, L. I., N. Y. (19) J. P. Verheljden and P. J. Stoffyn, Tetrahedron, 1, 253 (1957).
NOTES
DECEMBER, 1964 concentration of the extract, produced a sirup (IV) which gave [ a l z 6-118' ~ ( c 1.58, methanol). Anal. Calcd. for CoHlzOlS: S, 17.8; OCH3, 17.3; mol. wt., 180. Found: S, 17.5; OCH3, 17.6; mol. wt., 178. Periodate oxidationz0 showed 2 molar equiv. of periodate consumed, 0.2 molar equiv. of total acids produced, but with no formic acid produced.' The excess periodate consumed was probably due to the oxidation of the sulfur to a sulfone or sulfoxide.?l Compound I V showed no thiolacetate adsorption or free mercapto groups with T T C and SKP. Hydrolysis of IV in 0.5 'V hydrochloric acid resulted in a change of the specific optical rotation from -118' to + 3 8 O in 30 min. a t 75". This is suggestive that the glycoside is predominantly in the p-D-configuration. After hydrolysis was complete, the product was isolated by passing the solution through a column of Amberlite IR-45(OH) resin and concentrating to a sirup. This material, 4-deoxy-4mercapto-D-ribofuranose, gave a positive test for reducing groups and had R, 1.36 in irrigant A and 1.5 in irrigant B. Compound Is' gave a crystalline tri-p-nitrobenzoatelO: m.p. 193-194", [ a ] " D +85.5" ( c 0.38, chloroform). Bis(methy1 4-deoxy-p-~-ribopyranoside)4,4'-Disulfide (V) .Compound 111 (1.0 g.) was dissolved in 20 ml. of 2 N methanolic sodium methoxide and allowed to stand a t 25" for 16 hr. The solution was passed through a column of Amberlite IR-l20(H) resin for neutralization and removal of the isopropylidene group. To the effluent was added a few crystals of iodine and the mixture was refluxed for 3 hr. with oxygen bubbling through it. The cooled solution was concentrated to a thick sirup which was dissolved in 10 ml. of water and extracted twice with 25-ml. portions of chloroform to remove the excess iodine. The aqueous solution was concentrated under reduced pressure to dryness and the residue was crystallized from hot ethanol to give m.p. 152", [ c Y ] ? ~ D-229" ( c 0.43 water), yield 0.41 g. (60%). Titration of the product with 0.1 N iodine solutionzzshowed no thiol activity. Reaction with T T C and S N P gave no color test until after reduction of the disulfide bond with lithium aluminum hydridez3 in diethyl ether. The R, for V in irrigant A was 1.90, in irrigant B, 2.3. Anal. Calcd. CIZH,ZO&: S , 17.8. Found: S, 17.5.
Acknowledgment.-The authors gratefully acknowledge Grant No. AM 06782-01 from the Department of Health, Education, and Welfare which helped support this work. (20) R. D. Guthrie, "Methods in Carbohydrate Chemistry," Vol. 1,
R. L. Whistler and M. L. Wolfrom, Ed., Academic Press Inc., New York, S . Y., 1962, p. 432. (21) S. J. Leonard and C. R. Johnson, J . Org. Chem., 17,282 (1962). (22) R. M. Evans and L. N. Owen, J . Chem. Soc., 244 (1949). (23) R . C. Arnold, A. P. Lien, and R. M. A l a , J . A m . Chem. Soc., 73, 731 (1950).
Sodium-Liquid Ammonia Debenzylations in Nucleoside Synthesis' ELMER J. REIST,VICTORJ. BARTUSKA, A N D LEONGOODMAN Stanford Research Institute, M a l o Park, Calijornia Received July 24,1964
Recent publications2 concerning spongoadenosine
[ (9-p-D-arabinofuranosyl)adenine,I111, a nucleoside
first synthesized in these laboratories, have disclosed some interesting biological activities for the compound. (1) This work was carried out under the auspices of the Cancer Chemotherapy National Service Center, Xational Cancer Institute, National Institutes of Health, U. S . Public Health Service, Contract No. PH43-64500. The opinions expressed are those of the authors and are not necessarily those of the Cancer Chemotherapy Xational Service Center. (2) (a) M. Hubert-Habart and S. S. Cohen, Biochim. Biophys. Acta, 69, 468 (1962); (b) G. A. LePage and I. G. Junga, Cancer Res., 3 8 , 739 (1963); (cj J. J. Brink and G. A. LePage, ibid., 14, 312 (1964). (3) (a) W. W. Lee, A . Benitea, L. Goodman, and B. R. Baker, J . A m . Chem. Soc., 83, 2648 (1960); (b) E. J. Reist, A . Benitez, L. Goodman, B. R. Raker, and W. W. Lee, J . Org. Chem., 37, 3274 (1962).
3725 2"
WCl
ROCHi 0
+
OR 11, R = CeHaCH2
OR
I, R=CeH&Hz
111, R = H
Accordingly the synthesis of I11 in quantity has become important both for further biological evaluation and for conversion to potentially useful analogs. The description by Glaudemans and Fletcher4 of the condensation of 2,3,5-tri-O-benzyl-~-arabinofuranosyl chloride (I) with 6-benzamido purine to give, after deblocking, the desired b-nucleoside (111)offered a practical, direct route to 111. However, the final step in the synthesis, the catalytic hydrogenolysis of the intermediate 11, was convenient only when small quantities of I1 were employed. An alternative technique for removal of the benzyl blocking group of I1 was sought and the use of sodium in liquid ammonia was investigated. The numerous examples of debenzylation of S- and N-benzyl groups with sodium in ammonia made this a logical choice; surprisingly, however, there is virtually no mention in the literature of the use of this method for cleaving 0benzyl group^.^ Recent work in this laboratory6 described the smooth removal of both the 0- and S-benzyl group of 6-aniino-3-0-benzyl-5-S-benzyl-6-deoxy-l,2-0isopropylidene-5-thio-~-idofuranose through the action of sodium in liquid ammonia. We wish to draw attention to this method of 0-debenzylation because of its convenience and its applicability in situations where previously described methods of 0-debenzylation are inappropriate.' It was possible to effect the conversion of I1 to I11 in high yield using sodium in liquid ammonia; this modification of the Glaudemans-Fletcher procedure4is especially convenient for large-scale synthesis of 111. The stability of the adenine ring to the action of sodium in ammonia is noteworthy; there are numerous references to the reduction of nitrogen-containing heterocycles by this reducing agent.9 Experimental To a stirred suspension of 3.75 g. (6.98 mmoles) of 9-(2',3',5'tri-0-benzyl-p-o-arabinofuranosy1)adenine(11)4 in 160 ml. of liquid ammonia was added a total of 600 mg. (26 mg.-atoms) of sodium in portions over 10-12 min. by which time the characteristic deep blue color persisted. At this point, the blue color (4) C. P. J. Glaudemans and H. G. Fletcher, Jr., ibid., 28, 3004 (1963). ( 5 ) H. Smith, "The Chemistry of Sonaqueous Ionizing Solrents." Vol.
I , P a r t 2, "Organic Reactions in Liquid Ammonia." G. Jander. H. Spandau, and C. C. Addison, Ed., Interscience Publishers, Inc.. New York, N. Y.. 1963. Reductive fission with metal ammonia reagenta is reviewed (p. 156). The author mentions unpublished work by W. Grassman, E . Wunsch. and G. Fries which is quoted by W. Grassman and E. Wunsch [Fortschr. Chem. Org. Naturstofe, 18, 487 (1956)], in which they report the regeneration of L-tyrosine and serine from their respective 0-benzyl ethers using sodium and ammonia. (6) L. Goodman and J. E. Chriatensen, J . 078. Chem., 39, 1787 (1964). (7) I n addition t o the method of catalytic debenzylation already mentioned, it has been demonstrated8 t h a t carbohydrate benzyl ethers arc readily acetolyzed. Acetolysis conditions are totally incompatible with nucleoside stability, however. (8) R. Allerton and H. G. Fletcher, Jr.. J . A m . Chem. Soc., 7 6 , 1757 (1954). (9) Ref. 5, p. 276.
NOTES
3726
was discharged by the careful addition of ammonium chloride and the reaction was evaporated to dryness under a stream of nitrogen. The solid residue was triturated with 50 ml. of benzene (to remove bibenzyl), then was dissolved in 40 ml. of water. The aqueous solution was treated with Norit, then acidified with acetic acid to cause the precipitation of product (TIT). Filtration gave 1.53 g. (82%) of colorless product, n1.p. 255-256" dec. (uncor.), which was identical in all respects with authentic spongoadenosine (III).3 I n a larger run, 13 g. of I1 gave 6.1 g. (947,) of product 111, m.p. 254-256' dec.'O (IO) Private communication from Dr. R. R. Engle, of Riker Laboratories, Inc., Northridge, Calif.
Pinacol and Pinacolone Derivatives of Some Acylferrocenes LAWRENCE R. MOFFETT, JR. Thiokol Chemical Corporation, Huntsville Division, Huntsville. Alabama Received June 11, 1964
Several examples of the bimolecular reduction of ferrocenyl ketones have been reported in the literature. 1 Weliky and Gould, using methylmagnesium bromide and cobaltous chloride, obtained the pinacol derivative of benzoylferrocene and effected its conversion to the pinacolone. From the Clemniensen reduction of acetylferrocene, Pauson and Watts isolated an 87% yield of ethylferrocene and a 3y0 yield of the pinacolone derivative. We wish to report bimolecular reduction of several alkylferrocenyl ketones to the corresponding vicinal diols and subsequent conversion of the diols to the rearranged diferrocenyl ketones. Yields of the diferrocenyl pinacols were 31 to 60% and of the diferrocenyl pinacolones 42 to 60%. Acetylferrocene failed to undergo the two-step reaction described herein to the pinacolone derivative. However, n-propionyl-, n-butyryl-, and n-valerylferrocene all underwent bimolecular reduction in the presence of amalgamated magnesium to give the corresponding pinacol. With the exception of the bimolecular reduction product of n-propionylferrocene, which formed as a monohydrate, the diols were unstable and tended to undergo air oxidation with regeneration of the original ketones. Rearrangement of the diols could lead to two possible structures for the pinacolones. All of the pinacolones OH Fc-C-
' A
-Fc
R
R
OH I
Fc
O
I
+Fc-C-C-Fc I I
II
+ Fc-C--C-R I
R
R
0
II
is directly attached to the carbon atom of the carbonyl group showed carbonyl absorption a t 1675 cni.-' (carbon tetrachloride). No absorption bands at 1675 cm.-' were observed in any spectra of the pinacolones. The formation of I1 would be expected with the anticipated migration of the aromatic group rather than the alkyl group during the pinacol-pinacolone rearrangement. N o indications were found of the presence of substances having structure I in any instance. Colunin chromatography of the pinacolones on Alcoa F-20 activated alumina using 20% benzene in petroleum ether (60-90') as the eluting agent always gave rise to only one band. The infrared data plus the facts that only one component could be isolated from the pinacolones through column chromatography and that all the pinacolones melted over a narrow temperature range, even before chromatography, tend to indicate that the rearranged ketones are of structure I1 and are not contaminated with significant amounts of products of structure I. The pinacolones failed to form either 2,4-dinitrophenylhydrazones or oximes and could not be reduced by either the Clemmensen or Wolff-Kishner reactions. However, confirmatory evidence for the presence of a carbonyl group, in addition to the infrared data, was obtained through the ultraviolet spectra of the pinacolones which showed maximal absorption at 270 mp. Experimental3 Acylferrocenes.-The liquid ketones were synthesized by the procedure of Voge14; most refractive indices have not been reported hitherto. n-Propionylferrocene was obtained as a red oil, b.p. 95-97" (0.1 mm.), n Z 51.6195 ~ (lit.5 m.p. 38-39"). Anal. Calcd. for C13H14Fe0: C, 64.49; H , 5.83. Found: 64.29; H, 6.11. n-Butyrylferrocene upon distillation had b.p. 120-122" (0.25 mm.), nZ5D1.6079 [lit.Bb.p. 144-146" (1.5 mm.), n% 1.60731. n-Valerylferrocene upon distillation had b.p. 119-121" (0.08 mm.), 12251) 1.5943 [lit.?b.D. llC-120" (air-bath temperature) (0.01 mm.), m.p. 36-38']. Anal. Calcd. for C16HIPFeO: C. 66.68: H . 6.71. Found: .i ." C, 66.82; H, 7.01. Preparation of Diferrocenyldiok-The diferrocenyldiols prepared through bimolecular reduction of the acylferrocenes are listed in Table I. Preparation of 4,5-diferrocenyl-4,5-octanediol ( R = n-CZH7, Table I ) exemplifies the general procedure. A solution of n-butyrylferrocene (41.8 g., 0.16 mole) and 2.5 g. of mercuric chloride in 75 ml. of dry tetrahydrofuran was added in a slow but steady stream to a boiling mixture of magnesium (2.2 g., 0.09 g.-atom) in 30 ml. of dry benzene. The reaction mixture was heated under reflux with stirring for 36 hr., with protection from atmospheric moisture. The cooled mixture was treated with 15 ml. of water and heated a t reflux for 1 hr. Solid was removed by filtration and was extracted with several portions of boiling benzene. The combined filtrates were concentrated in uacuo, and the solid residue was washed with petroleum ether (6(t90°) until the washings were only faintly yellow. The analytical sample was obtained as brown plates from benzene-
I
R
I Fc
I
VOL. 29
I1
ferrocenyl R = C2Hs-, n-CZH,-, n-CaHs=
were shown to have structure I1 by virtue of strong carbonyl absorption a t 1710 cm. - l . All alkylferrocenyl ketones used in this study in which the aromatic moiety (1) (a) N. Weliky and E. S. Gould, J. A m . Chem. SOC.,79, 2742 (1957); (b) P. L. Pauson and W. E. Watts, J. Chem. SOC.,3880 (1962).
(2) C. K. Ingold. "Structure and Mechanism in Organic Chemistry," Cornel1 University Press, Ithaca, N. Y., 1953, p. 477. (3) All melting points are uncorrected. Infrared apectra were determined with a Perkin-Elmer ,Model 21 spectrophotometer. The author wishes t o thank Messrs. J. E. Sharpe and R. L. Hill for the microanalyses and Mr. R. A. Jewel1 for the infrared spectra. (4) M. Vogcl, M. Rauach, and H. Rosenburg, J . O w . Chem., I'd, 1016 (1957). (5) K. L. Rinehart, Jr.. R. J. Curby, Jr., and P. E. Sokol. J. A m . Chem. Soc.,79, 3420 (1957). (6) E. L. DeYoung, J . Org. Chem., S6, 1312 (1961). (7) K. Schloegl and II. Pelousek, A n n . , 661, 1 (1962).
NOTES
DECEMBER, 1964
3727
TABLEI DIFERROCENYLDIOLS OH
FcFc
L A
- -Fc
I
I
R
R
=
ferrocenyl
___--__ R
Yield, yo
CzHs 31 n-C3Hr 55b n-C4H960* a Recrystallized product.
M.p., aC.a
113-115 126-128 94-97
OH
C
61.93 65.39 66.44
_-___--
Calcd., %------H
6.40 6.66 7.06
Fe
C
22.15 21.72 20.60
61.86 65.27 66.25
Found, yo------H
6.43 6.69 7.11
Fe
22 30 21.93 20.85
* Crude yield. TABLE I1 DIFERROCENYL KETONES
I
Fc Fc = ferrocenyl
_______ Calcd., %------n
Yield, %
CzHsn-CaHr n-C4H9--
47 60 42
M.p., O C .
110-111.5 108.5-109.5 62-64
____--_
C
H
Fe
c
Found, %------H
Fe
66.70 67.76 68.72
6.03 6.50 6.92
23.86 22.51 21.31
66.67 67.65 68.69
6.25 6.65 7.11
23.80 22,55 21.33
petroleum ether (60-90"). The infrared spectrum of the diol (tetrachloroethane solution) showed strong hydroxyl absorption a t 3400 cm.-'. After several days' storage a t room temperature, a sample of the solid slowly turned to a reddish brown oil which which was identified through its infrared spectrum as n-butyrylferrocene. Preparation of Diferrocenyl Ketones.-The rearranged ketones prepared through the dehydration of the diols are listed in Table 11. Preparation of 5,5-diferrocenyl-4-octanone ( R = n-C3H7,Table 11) illustrates the general procedure. A solution of 4,5-diferrocenyl-4,5-octanediol(10.3g., 0.02mole) in 50 ml. of chlorobenzene was added to a solution of 20 ml. of concentrated sulfuric acid in 80 ml. of water. The resulting two-phase system was heated under reflux with vigorous stirring for 6 hr. The organic layer waa separated, washed to neutrality with water, and connected in vacuo. The residual oil was triturated with methanol and the mixture was allowed to stand a t -25" for several hours until solidification was complete. The yellow solid waa recrystallized from methanol-benzene as short golden needles. The infrared spectrum of the compound (carbon tetrachloride) showed in addition to the carbonyl band a t 1710 cm.-I, and other bands, the two absorption bands a t 1105 and 995 cm.? characteristic of mono-ring substitution of ferrocene.8 Reduction of acetylferrocenes (22 g., 0.1 mole) with magnesium (1.44g., 0.069 g.-atom) and 1.6 g. of mercuric chloride (50 ml. of tetrahydrofuran and 15 ml. of benzene) as described above gave 18.1 g. of solid with m.p. 146-175". Even upon temporary storage in vacuo it began to decompose to a dark liquid. I t was therefore immediately dissolved in 125 ml. of chlorobenzene and added to a solution of 48 ml. of concentrated sulfuric acid in 192 ml. of water. The two-phase system was treated as described above to yield 3.4 g . of solid, m.p. ca. 240" dec. The infrared spectrum of the material showed no carbonyl absorption bands a t either 1685 or 1715 cm.?. Anal. Calcd. for CpaHzdFezO: C, 65.49; H, 5.50; Fe, 25.38. Found: C, 65.73; H , 5.72; Fe, 23.00.
(8) M. Rosenblum and R. B. Woodward, J . A m . Chem. Soc., 80, 5443 (1958). (9) P. J. Graham, et al., ibid., 79, 3416 (1957):
The Structure of Alantolactone JAMES A. MARSHALL AND NOALCOHEN Department of Chemistry, Xorthwestern University, Evanston, Illinois Received July 1, 1964
The sesquiterpene lactones alantolactone (I), isoalantolactone (II), and dihydroisoalantolactone (111) were related to eudalene by Ruzicka and co-workers.' Subsequently, Tsuda, et C L ~ . proved ,~ the correct placement of the lactone oxygen and assigned structure IV to alantolactone and structure I1 to isoalantolactone. Asselineau and Bory3 questioned the validity of IV on the grounds that dihydroisoalantolactone (111) and dihydroalantolactone (V)* do not give the same chloride with hydrochloric acid.5 The French workers saponified and oxidized dihydroalantolactone to a keto acid which was isomerized by alkali to a conjugated ketone VI 240 nib ( E 6000)].3 This evidence supports formulation I for alantolactone. Recently, Nakazawa6 obtained identical bromides by hydrobromination of dihydroisoalantolactone and di(1) L. Ruzicka, P. Pieth, T. Reichstein. and L. Ehmann. Helu. Chzm. Acta, 16, 268 (1933). These workers favored structure I V (lactone a t C-6). T h e stereochemistry of these compounds has been revieued by W. Cocker and T. B. H. McMurrv Tetrahedron. 8 . 181 (19130\. (2) K. Tsuds, K. Tansbe, I. Iwai, and K. Funakoshi. J. A m . Chem. Soe., 79. 5721 (1957). ( 3 ) C. Asselineau and S.Bory, Compt. rend., 246, 1874 (1958). (4) T h e configuration a t C-11 has recently been proved by W. Cocker and M. A. Nisbet, J . Chem. Soc., 534 (1963). (5) K. F. W. Hansen IChem. Eer., 64, 1904 (1931)l obtained twoisamptic hydrochlorides from dihydroisolantolactone. J. Bredt and W. Pcsth [ A n n . , 286, 349 (1895) I prepared the hydrochloride of dihydroalantolactone. (6) 5.Nakazawa, J. A m . Chem. Soc., 8 2 , 2229 (1960).
3728
NOTES
VOL. 29
7cps
C - 4 CH, (1.11 ppm)
=
hydroalantolactone under comparable conditions. Furthermore, Ukita and Yakazawa? found that ozonation of dihydroalantolactone yielded a lactone keto acid which gave an iodoform test. Seemingly, these results are compatible only with formulation IV for alantolactone.
I
I1
I11
IV
V
VI
Our interest in this structural dilemma arose from projected synthetic approaches to the alantolactone fanlily of sesquiterpenes. I n view of the differing synthetic problenis posed by I and IV we decided to (7) T. U k i t a and S.Sakazawa. J . Am. Chem. Sac., 83. 2224 (1960).
seek evidence which would allow a clear choice between the two. For this task, nuclear magnetic resonance spectroscopy served admirably to define alantolactone as I and dihydroalantolactone as V. The proton resonance peaks (Fig. 1) which are crucial to the choice of I in preference to IV occur at gCC1 TM$ = 5.14 (H-6 doublet; J 6 , 7 = 4 c.P.s.) and 1.11 p.p.m. (C-4 CH, doublet; J = 7 c.P.s.). Additional structural details in support of I are furnished by peaks a t 4.73 (H-8 quintet; J8,9 = J8,9! = 3 c.P.s., J8,1 = 7 c.P.s.) and 3.53 p.p.m. (H-7 multiplet; J7,e = ! 4 c.P.s., J7,8 = 7 c.P.s., J7,13 = 2 C.P.S.).~The coupling constants between H-8 and H-9 (dihedral angle 0' = 60°), H-8 and H-9' (e = 60°), and H-7 and H-8 (e = 15') are in good agreement with the predicted value^.^ Inspection of Dreiding inodels reveals that the longrange coupling between H-7 and H-13 in alantolactone (I) should be about 2 c.P.s., (observed J = 2 and 1.8 c.P.s.) while the analogous J,,13 in isoalantolactone (11) should be less than 2 C.P.S. (observed J = 0.8 and 0.6 c.p.s.).'O Thus the spectrum of alantolactone (Fig. 1) confirms not only the placement of the carbocyclic double bond at C-5 but the stereochemical details of I as well. As expected, the inultiplet a t 3.5 p.p.m. due to the C-7 proton of alantolactone is shifted down(8) T h e internal consistency this first-order interpretation. (9) H. Conroy, "Advances Raphael, E. C. Taylor, and H. New York, N. Y . ,1960, p. 311. (10) T. A. Wittstruck, S. K. Soc.. 85, 1699 (1963).
of the aplitting patterns appears t o justify
in Organic Chemistry," Vol. 2 , R . A. Wynberg, Ed., Interscience Publishers, Inc., Malhotra, and H. J. Ringold, J . Am. Chem.
DECEMBER, 1964
KOTES SCHEME I
CH,
' T C H 3 OH
-
RTo
CH3
+
NaOH I CH3CHO L C H I 3
field in the spectrum of dihydroalantolactone but the doublets at 5.1 and 1.1 p.p.m., attributed to the C-6 proton and the C-4 methyl group, remain. It is noteworthy that neither alantolactone (I) nor isoalantolactone (11) show appreciable Hl3/HI3r spin-spin coupling. l1 We repeated the low-temperature hydrobromination of dihydroisoalantolactone (111) and obtained Nakazawa's6 120-121 O bromo derivative which, on the basis of its n.m.r. spectrum, is 4-broniotetrahydroalantolactone. Since the same bromo compound is obtained by hydrobrominatiori of dihydroalantolactone (V),6 it is clear that this reaction is accompanied by double-bond (or carbonium ion) rearrangement. We were unable to prepare the 189-191" bromo derivative which is reported6 for both dihydroalantolactone (V) and dihydroisoalantolactone (111). All attempts produced colored oils which evolved hydrogen bromide on exposure to air and during attempted crystallization at room temperature or below. Similar behavior was noted for the 120-121" bromo derivative. Hydrochlorination of dihydroisoalantolactone (111)s was effected smoothly and afforded a mixture of isomeric 4-chlorotetrahydroalantolactones in good yield. The chloro compounds were easily handled and showed no tendency to dehydrochlorinate. Therefore, structural conclusions based on these derivatives3 must be considered more compelling than conclusions based on the hydrohromides.6 The lactone keto acid obtained by ozonation of dihydroalantolactone (V)' must now be formulated as VII. A reasonable pathway for the formation of iodoform from this compound is pictured (R = 1,3dimethyl-2-ketocyclohexyl) in Scheme I.
3729
Alantolactone (I)gave ;::6 6.06 (H-13 doublet, J = 2 c.P.s.), 5.52 (H-13' doublet, J = 1.8 c.P.s.), 5.14 (H-6 doublet, J = 4 c.P.s.), 4.73 (H-8 multiplet), 3.53 (H-7 multiplet), 1.18 (C-10 CH,), and 1.11 (C-4 CHI doublet, J = 7 c.P.s.) p.p.m.; m.p. 78.5-80", lit3 m.p. 78-79'. Isoalantolactone (11)gave 6:g; 6.02 (H-13 doublet, J = 0.8 c.p.s.),5.55(H-l3'doublet, J = 0.6c.p.s.),4.76,4.55(C=CH2), 4.45 (H-8), and 0.82 (C-10 CH3) p.p.m.; m.p. 112-113", lit.3 m.p. 111-113". Dihydroisoalantolactone (111) gave;;:8 4.78, 4.51 (C=CH2j 4.3-4.6 (H-8), 1.17 (C-11 CHI doublet, J = 7 c.P.s.), and 0.82 (C-10 CH3) p.p.m.; m.p. 172-173", lit.' m.p. 171-172". This material was prepared by hydrogenation of isoalantolactone over reduced platinum oxide in methanol until 1 mole equiv. was taken UP. Dihydroalantolactone (V) gave ;::6 5.18 (H-6 doublet, J = 3 c.P.s.), 4.5-4.8 (H-8), 1.22 (C-10 CH,), and 1.17 ((3-4 CHI and C-11 CHI doublet, J = 8 c.p.s.)p.p.m.; m.p. 132-132.5", lit.' m.p. 133.5134". This material was prepared in the manner described for dihydroisoalantolactone. 4-Bromotetrahydroalatolactone gave ;::6 4.2-4.5 (H-8), 1.78 (C-4 CH,), 1.18 (C-10 CH,), and 1.18 (C-11 CHI doublet, J = 7 c.P.s.) p.p.m.; m.p. 119-121", lit.em.p. 12C-121". This material was obtained in only 10% yield by hydrobromination of dihydroisoalantolactone a t 0" according to the published method .e Hydrobromination of dihydroisoalantolactone at room temperature by the published procedure6 afforded a brown oil which fumed in air. Crystallization of the oil could not be induced and dissolution in ether and hexane was attended by fuming and separation of a red-brown oil. 4-Chlorotetrahydroalantolactonegave ;::8 4.6-4.2 (H-8), 1.52 (C-4 CH3), 1.20 (C-11 CHI doublet, J = 7 c.P.s.), and 1.00 (C-10 CH3) p.p,m.; m.p. 142-145", lit.6 m.p. 145'. The procedure of Hansen5was followed using 0.50 g. of dihydroisoalantolactone. The crude hydrochloride, m.p. 104-137", was obtained in 977, yield. Several recrystallizations from ethanol afforded O.IOg., m.p. 142-145". The mother liquors were evaporated and the residue was recrystallized from ethanol. A second recrystallization afforded 0.05 g., m.p. 127-136". This material is a mixture of 4-chloro isomers since the n.m.r. spectrum exhibited a new peak a t 0.81 p.p.m. ((2-10 CH3) in addition to intense peaks a t 1.52 (C-4 CH3j and 1.00 p.p.m. (C-10 CH3) present in the 145" isomer. The chloro derivatives (in distinct contrast t o the bromo derivatives) showed no tendency to fume. Solution in boiling 95y0 ethanol could be effected with no sign of decomposition.
Acknowledgment.-We thank the U. S. Public Health Service for support of this work through a research grant (AI-04965, National Institute of Allergy and Infectious Diseases) and a fellowship (5-Fl-Gl1-19,839 to N. Cohen). The A-60 spectrometer was purchased with funds furnished Northwestern University by the National Science Foundation. We are indebted to Professor R. E. Ireland for his kind assistance in obtaining spectra of dihydroalantolactone and dihydroisoalan tolactone.
Experimental12 Alantolactone was purchased from Chemicals Procurement Laboratories, Inc., College Point, N . Y . The commercial material was about 607, isoalantolactone which was largely removed by crystallization from aqueous methanol. The remaining alantolactone was purified to the reported melting point3 by numerous recrystallizations, first from hexane and finally ethanol a t -20'. This material retained 5 1 0 % of isoalantolactone (detected by the n.m.r. spectrum and estimated by integration) which could not be removed by further crystallization.' (11) This result might be expected if the C-13 H-C-H bond angle assumes a value near 125' (ref. 9, p. 310). (12) Melting points were taken on a Fisher-Johns hot stage and are corrected. N.m.r. spectra were determined n,ith a Varisn A-60 spectrometer.
Preparation of Triaminoguanidine LAWRENCE E. BENJAMIN
Contract Department, Central Research Division, American Cyanamid Company, Stamford, Connecticut Received July S I , 1964
During the course of a research program involving the use of triaminoguanidinium salts, a convenient method for the preparation of triaminoguanidine was discovered. The synthesis of this compound in its
3730
NOTES
free-base form has not been reported in the literature.' A simple reaction of triaminoguanidinium chloride with liquid ammonia results in the formation of ammonium chloride and a precipitate of triaminoguanidine. Al[(NHzNH)&]+Cl-
liq. NHa + NHI -+
NNHz /I
NHzKHCNHNHz4
+ NHaCl
though triaminoguanidine is a much stronger base than ammonia, its very low solubility in liquid ammonia,2 together with the use of a large excess of ammonia, causes the reaction to go to completion. Triaminoguanidine is a white crystalline solid which is stable when stored in vacuo or in an inert atmosphere. In air it undergoes slow decomposition of an unknown nature, becoming pink and ultimately deep purple in color. Perhaps owing to greater purity, some samples have initially been more resistant to this decomposition. A relatively small amount of decomposition, not detected by a change in infrared spectrum, imparts rather intense color to the material. Samples which become colored after exposure to air usually revert to a nearly colorless state again when resealed and allowed to stand. However, the color reappears very rapidly when the container is again opened. Triaminoguanidine is extremely soluble in water, insoluble or sparingly soluble in common organic solvents. I n water it hydrolyzes to carbohydrazide and hydrazine (half-life a t 25" about 14 h ~ - . ) . Titration ~ of its aqueous solution gives the neutralization curve characteristic of a strong base. Experimental' The apparatus used consisted of a 125-ml., three-neck, roundbottom flask with a sealed-in sintered-glass disk and draw-off tube in the bottom, equipped with a drying tube and two gas outlet adapters. Seven grams (0.05 mole) of triaminoguanidinium chloride was placed in the flask, the flask was purged with nitrogen, and about 30 ml. of anhydrous liquid ammonia was introduced. The mixture was stirred with a magnetic stirrer for about 10 min. and then the ammonium chloride solution was removed by applying vacuum to the draw-off tube. During the filtration, nitrogen was passed through the flask. The solid remaining in the flask was treated twice more with liquid ammonia in this manner, then nitrogen was drawn through the white crystalline product until it warmed to room temperature. It was dried in uacuo over Pz06a t 40°, yield 4.7 g. (90%). On a FisherJohns block the compound turned red and melted with gas evolution a t about 100". In an evacuated capillary, melting with decomposition began a t 141". A n a l . Calcd. for CH8Ne: C, 11.54; H , 7.74; N , 80.72; equiv. wt., 104.1. Found: C, 11.61; H , 7.52; N, 80.76; equiv. wt., 103.2 (titration in acetic acid-acetonitrile with perchloric acid). The infrared spectrum (Nujol mineral oil and halocarbon mulls) showed absorption maxima a t 3310 (sh), 3275 (9) (NHz asymmetric stretch), 3165 (8) (NHz symmetric stretch), 2860 (vw), 1677 (ms) (C=N stretch), 1'643 (8) (NHz deformation), 1632 (sh), 1485 (s), 1443 (m), 1355 (w), 1343 (w), 1316 (w), 1177 (1) Early in 1959, Dr. V. P. Wystrach and Mrs. J. H. Smalley at these laboratories prepared aqueous solutions of triaminoguanidine by passing solutions of t h e hydrochloride through a strongly basic anion-exchange resin. Isolation of the free base from such solutions was reported t h a t same year in t h e classified literature by Mrs. P. D. Oja and Mr. G. E. Hartzell of t h e Dow Chemical Co. This was accomplished by low-temperature concentration and precipitation. (2) This is in contrast to related strong organic bases such as guanidine, guanylurea, and biguanide. which are reported to be soluble in liquid ammonia: cf. I+'. H. Hill, U. S. Patent 2.274.412 (Feb. 24. 1942). (3) 0 . S . Sprague and E. A . Takacs. unpuhlished work. (4) Microanalyses were perforinnil b y Mr. J. H. Deoqarine and Miss J. Schuler. titrations by Dr. C. A . Streuli nod Mr. S. Sandler. Infrared spectral data was determined by Mr. N . B. Colthup.
VOL.29
(ms), 1130 (ms), 1002 (m), 954 (8, broad), 868 (m, broad), 753 (ms, broad), 699 (m) cm.-'. Purification of impure material has not been easily accomplished because of the sensitivity of the base to air and moisture, and its low solubility in solvents other than water. The bulk of the colored decomposition products can be removed by washing with anhydrous methanol. Recrystallization then can be carried out by stirring the solid in dimethylformamide (ca. 25 ml./g.) a t 80" and adding water until it dissolves (ca. 6 ml./g.). Subsequent cooling of the solution to - 10' usually yields nearly colorless crystals.
Acknowledgment.-This research was supported by the Advanced Research Projects Agency, Propellant Chemistry Office, and was monitored by the Bureau of Naval Weapons, RMlLIP, under Contract No. NOrd 18728.
Hydrogenation i n the Pyridine Series. 11. Catalytic Reduction of 2-Monoalkyl- and 2-Dialkylaminopyridines
M.FREIFELDER, R. W. MATTOON, AND Y. H. No Organic Chemistry and Physical Chemistry Departments, Research Division, Abbott Laboratories, North Chicago, Zllinois Received April 16, 1964
Numerous 2-substituted pyridines have been converted to the corresponding piperidines, but in the catalytic hydrogenation of 2-aminopyridine only 2 molar equiv. are absorbed, yielding 2-iminopiperidine. Further uptake gives only hydrogeno1ysis.l I n a study of catalytic debenzylations, Birkhofer did not obtain 2-aminopyridine from 2-benzylaminopyridine. He described the resultant product as 2-benzylamino-3,4,5,6tetrahydropyridine. I t was anticipated that 2-diethylaminopyridine (I), incapable of tautomerizing, could be reduced to the corresponding piperidine. However, under the most favorable reaction conditions -hydrogenation in glacial acetic acid in 'the presence of a high ratio of rhodium-onalumina catalyst -only 2 molar equiv. were absorbed. The reduction product was shown to be 2-diethylamino3,4,5,6-tetrahydropyridine (11) by the absence of vinyl proton absorption in the n.m.r. spectrum3 and by infrared examination: X ~ ~ 6.28 ~ ' p* , strong (C=N), no bands for NH or pyridine ring. 2-Dimethylamino- and 2-methylaminopyridine were hydrogenated to determine whether smaller substituents would lead to piperidines. I n each reduction only 2 equiv. were absorbed giving the corresponding tetrahydropyridines IV and VI (VIa). The reduction product from V can exist as an endo or exo double-bonded cyclic amidine. The results of
(1) T. B. Grave, J . Am. Chem. Soc., 46, 1468 (1924). (2) L. Birkhofer, Ber., 76, 429 (1942). (3) The n.m.r. spectra were run by Mr. R. Kriese on a Varian A-60 spectrometer at 60 Mc./sec. with tetramethylsilane as internal standard and, unless stated, with deuteriochloroform as solvent
NOTES
DECEMBER, 1964
3731
TABLE I
Yield,a
KO.
R'
R2
%
B.p., "C.
mm.
n16D
124-126 108
47 38
1.4884
95-97 84-87
30 22
1.4952 1.4951
11
C2Hs CzH5
71.8
IV
CHI
60
Hydrochloride m.p., OC.
-HH.%Calcd. Found
70.07
69.91
11.76 11.76 18.13 18.07
C13H24Nz08 57.33
57.33
CeHisNz 103b
CHI
%-
--4, %Calcd. Found
Formula
8.88
-N, Calcd.
Found
8 . 8 3 10.28 10.36
11.18 11.07 22.92 22.92 66.62 66.59 CTHiLVz 9.28 9.22 17.22 17.04 51.68 51.48 C7HjsClNz 10.78 10.89 24.88 24.63 64.23 63.45' VI H CHI 60 104-105 14-16 e CGHizNz 8.81 8.67 18.84 18.82 162d CeH13ClNz 48.47 48.29 a The yields could be increased by continuous extraction. b Succinate salt. Oxygen was also run. Calcd., 23.50. Found, 23.89. The hydrochloride salts were extremely hygroscopic. They had to be weighed in a dry atmosphere to obtain good analyses. e Material solidified. f The base absorbed carbon dioxide rapidly. Analysis had to be carried out immediately after distillation. Carbon value of the same material analyzed 2 hr. later dropped to 61 %. 124-126d
bon dioxide. Unless it was analyzed immediately after distillation, carbon values were always low. 2-Diethylamino-3,4,5,6-tetrahydropyridine(11).-A solution of 22.5 g. (0.15 mole) of I in 100 ml. of glacial acetic acid was hydrogenated under 3 atm. in the presence of 9.0 g. of 5y0 rhodium on alumina.8 Uptake did not proceed beyond 2 equiv. (2 hr.). The solution, after removal of catalyst, was concentrated under reduced pressure to a thick mass. Water was added, and the solution was made strongly basic with 507' sodium hydroxide solution. The mixture was cooled and thoroughly extracted with ether, and the ether extract was dried over anhydrous magnesium sulfate. After removal of the drying agent, the solution was concentrated, and the residue was distilled. Compounds IV and VI were similarly prepared from I11 and V (see Table I for physical constants). When the reduction of I was run in the presence of 0.45 g. of platinum oxide8uptake was slow. When 1.2 g. was used the rate Experimental increased. I n each case only 2 equiv. were absorbed. The infrared spectra of TI, IV, and VI showed the presence of a The reaction of 2-bromopyridine and diethylamine to form I is C=N band a t 6.15-6.28 p and the absence of pyridine ring. In typical of the method used to prepare I11 and V. Each of the addition, there was an NH band a t 2.9 p for VI.1o starting substituted aminopyridines is known. However, the N.m.r. Spectra of the 3,4,5,6-Tetrahydropyridines (6) .-2-Didifficulty in separating any contaminating 2-bromopyridine from ethylamino derivative 11: methyl protons of the ethyl group, the products made elemental analysis necessary so that only pure triplet centered a t 1.07 ( J = 7 c.P.s.), integration, 6; methylene products would be used for subsequent ring reduction. protons, quartet, 3.11, 3.23, 3.35, 3.47 ( J = 7 c.p.5.); C-3 pro2-Diethylaminopyridine (I).-A solution of 79 g. (0.5 mole) of tons,11 triplet with fine splitting, centered a t 2.22, integration, 2; 2-bromopyridine and 146 g. (2.0 moles) of diethylamine in 350 C-4 and C-5 protons, multiple peaks with fine splitting, 1.33ml. of ethyl alcohol was heated and shaken in a 1-1. rocker-type 2.00, integration, 4; C-6 protons,ll triplet centered a t 3.51 ( J = bomb for 20 hr. a t 175°.a After removal of the material from the 5 c.P.s.), partially obscured by one of the methylene peaks a t 3.47 bomb, the mixture was concentrated. The residue wm treated (this and the methylene quartet integrate to 6). with anhydrous ether. After diethylamine hydrobromide was 2-Dimethylmino compound IV: methyl protons, singlet, 2.87, filtered from the solution, the solvent was distilled and the residue integration, 6; C-3 protons," triplet centered a t 2.23 ( J = 6 fractionated. The yield of the fraction collected a t 116-119" c.P.s.), integration, 2; C-4 and C-5 protons, multiple peaks, ~ was 72y0. The described boiling (32 mm.), n Z 51.5357-1.5360, 1.50-2.03, integration, 4; C-6 protons,11 triplet centered a t 3.53 point is 208-214', a t atmospheric pressure.6 (J = 5 c.P.s.), integration, 2. 2-Dimethylaminopyridine (111), b.p. 105-107" (40-41 mm.),? Compound VI (VIa) in chloroformlz: methyl protons, sharp n% 1.5547-1.5551, was obtained in 81-84y0 yield. singlet, 2.79; NH, 3.97 (disappears on addition of DzO); C-3, 2-Methylaminopyridine (V), b.p. 94-96' (15 mm.),* n Z 5 ~ C-4, and C-5 protons, multiple split peaks 1.37-1.97 and a 1.5716, was obtained in 81.5% yield. I t has an affinity for carpossible triplet centered a t 2.13, integration, 618; C-6 protons,li triplet centered a t 3.50 ( J = 5 c.p.5.). I n dimethyl sulfoxide and in water the separation between the (4) Basic amines of t h e type CHaNHR d o not always show a doublet for C-3 and the C-4 and -5 signals was better defined. The N-methyl the methyl protons. "High Resolution N . M . R . Catalog," Vol. 2 , N. S.
n.m.r. spectra of the base in chloroform, dimethyl sulfoxide, and water4 showed the ?ri-methyl protons as a sharp singlet. The signal in the spectrum of the hydrochloride salt in water or of a solution in water and acid to pH 0.8 was still a singlets6 No real evidence is seen for either structure. However, a comparison of the n.m.r. spectra of I1 and IV, each with a fixed endocyclic bond, with that of VI (VIa) shows the C-6 protons of the three compounds as essentially symmetric triplets ( J = 5 c.P.s.) with almost identical chemical shifts. This suggests that the latter compound exists primarily in the endo form.
Bhacca, D. P. Hollis, L. F. Johnson, and E. A. Pier, Varian Associates, Palo Alto, Calif.: example 652 shows splitting while example 489 does not. Some unpublished spectra from this laboratory of similar type compounds show singlets in, some instances, doublets in others. I n t h e spectrum of V, the signal is a'doublet ( J = 5 c.p.8.). ( 5 ) L. M . Jackman, "Application of Nuclear Magnetic Resonance in Organic Chemistry," Pergamon Press, Inc., New York, N. Y., 1959; section 5.2 points out t h a t methylamine hydrochloride does n o t show splitting because of rapid exchange of t h e amino proton with water. (6) British P a t e n t 265,167 (1928); the reaction was carried out with 2chloropyridine a t 225' for 8 hr. When we used a n 8-hr. period a t 150' with 2-bromopyridine, a mixture of I and bromopyridine was obtained. The longer reaction period and higher temperature were used t o ensure complete conversion. (7) A. E . Chichibabin and I. L. Knunianz [ B e y . , 61, 427 (1928)l give 88' (15 mm.). (8) F. W. Bergstrom, H. G. Sturz, a n d H . W. Tracy [. Org. I.Chem., 11, 289 (1946)] give 100-102° (18mm.).
(9) Available from Engelhard Industries, Newark, N. J. (IO) Infrared spectra run b y A. Kammer and W. Washburn of this laboratory. (11) Adjacency t o t h e double bond should move t h e C-3 signal further downfield t h a n t h a t of t h e C-4 and C-5 protons. T h e C-6 protons should be seen farthest downfield because of proximity t o the heterocyclic nitrogen atom. Each should be a triplet split b y the C-4 and C-5 protons, respectively. (12) When t h e spectrum was run in CDCla, no signal for N H was seen, b u t a very strong chloroform peak appeared a t 7.47. Evidence of exchange was confirmed by infared when the solution showed almost total elimination of N H a t 2.9 p and N D appeared a t 3.9 a. All other points in the n.m.r. spectra in CDCls and CHCls were identical. (13) T h e triplet is not too well defined. There is a definite separation between i t and the multiplet. The portion further don.nfield which we assign t o the C-3 protons (see ref. 11) integrates t o 2, while t h e multiplet integrates t o 4.
NOTES
3732
VOL. 29
protons in each solvent were represented by a single peak. The C-0 protons in each instance appeared as a triplet ( J = 5 c.P.s.) .I4 Attempted Hydrolysis of VI (VIa).-The material was heated for about 4 hr. with 187, hydrochloric acid, cooled, made basic, and extracted with ether. Similarly, hydrolysis with a 2OY0 sodium hydroxide solution was carried out. The cooled solution was extracted with ether. The two extracts were dried and concentrated. The residues were examined. Their n.m.r. spectra were unchanged. However, when hydrolysis was carried out in 20% sodium hydroxide for 18-20 hr., a sample of the solution submitted for n.m.r. showed complete removal of the methyl protons.
Acknowledgment.-The authors are indebt,ed to Dr. John Tadanier of the Organic Research Department of this laboratory and to Professor Peter Beak of the Universit'y of Illinois a t Urbana for the discussions of the i1.ni.r.data while trying to develop the structure of \'I (VIa). (14) T h e spectrum of VI (VIa) in dimethyl sulfoxide t o which a few drops
of concentrated hydrochloric acid were added showed the N-methyl protons as a doublet, 2.83 and 2.92 (J = 5 c.P.s.), and the C-6 protons were seen a s an unsharp group with separate peaks (J = 2 c.p.8.) suggestive of two overlapping triplets probably caused b y protonation of the heterocyclic nitrogen. T h e spectrum in concentrated hydrochloric acid also showed a doublet, 2.89 and 2.97 (J = 5 c.P.s.), for the N-methyl protons and split peaks, 3.45 a n d 3.49 ( J = 2 c.P.s.), amid a n unsharp group for the C-6 proton signal. These two spectra only establish t h a t the protonated form of VI (VIS) exists as a resonance stabilized hybrid.
Microbial Hydroxylation of 5,6-Dihydrosolasodine YOSHIOSATO,JAMES A. WATERS,AND HIDEHIKOK A N E K O ~ .Vational Institute of Arthritis and Metabolic Diseases, AVationalInstitutes of Health, Bethesda 14, Maryland Received August 18, 1964
The hydroxylation of diosgenin, the well-known steroidal sapogenin12 with the fungus, Helicostylum p i ~ i f o r m e (ATCC, 8992), leads to the formation of 7P,11a-dihydroxy- ( lO-l5% yield) and 11a-hydroxy7-oxodicsgenin (5-10% yield) . 3 However, when the same fungus was incubated with the 5,6-dihydro derivat'ive of diosgenin, L e . , tigogenin, no detectable amount of hydroxylation was o b ~ e r v e d . ~ I n view of t'his somewhat surprising effect of the saturated 5,B-dihydro st'eroid in suppressing hydroxylation, it became of some interest to study the behavior of 5,6-dihydro~olasodine~(solasodan-3/3-01) since its precursor solasodine has previously been shown to hydroxylate readily to form 9a- (ca. 35%), lla- (ca. 1%), and 76-hydroxysolasodine (ca. 1%). 5 The hydroxylation of 5,6-dihydrosolasodine (I), contrary to our expectation, was not altered to any appreciable degree in comparison with solasodine; the corresponding Sa-hydroxy- (11) and 7p-hydroxy-5,6dihydrosolasodine (111) were obtained. The yields were roughly comparable with t,hat obtained in the
c
1.
IT, R = H
IU,R = H
?Lo, R = A c
ma, R = A c
m
hydroxylation of solasodine. The diol IT formed the expected 0,K-diacetylhydroxy derivative, IIa, and 111, the O,O,N-triacetate IIIa, upon acetylation. The identity of diol I1 was determined by comparison with a sample of 9a-hydroxy-5,6-dihydrosolasodine prepared by the catalytic reduction of Sa-hydroxysolasodine.5 The location of the hydroxyl function in the latter has been authenticated. The structure of the second diol I11 was likewise ascertained by comparison with a specimen of 7p-hydroxy-5,6-dihydrosolasodine obtained from the catalytic reduction of 7P-hydroxyso1asodine.j Molecular rotation data was also in agreement for a 7P-configuration. The seemingly contradictory data observed in the hydroxylation of tigogenin (5,6-dihydrodiosgenin) and 5,6-dihydrosolasodine attest to the necessity for more fundamental knowledge concerning the mechanism of microbiological hydroxylation. Experimental6 Microbiological Hydroxylation of 5,6-Dihydrosolasodine.Erlenmeyer flasks (500-ml.) each containing 200 ml. of corn steep medium,6 were inoculated with newly formed spores of the fungus H . piriforme and agitated on a platform shaker at 29-30' for 67 hr. A solution of 25 mg. of dihydrosolasodine (m.p. 205-208') in 2.0 ml. of ethanol was added to each flask. The flasks were incubated a t 29-30' for 100 hr. The mycelium from the combined flasks was removed by filtration through a thin layer of Celite, and then washed with ethanol. The filtrate was made basic with ammonium hydroxide and extracted with chloroform. Thus in a typical run 1.5 g. of dihydrosolasodine yielded 2.0 g. of light brown, resinous residue, which was shown to consist principally of a lipid material and of three steroidal components by thin layer chromatography (silica gel G, n-heptane-ethyl acetate-triethylamine, 2 :4 : 4). The extract was chromatographed on 50 g. of neutral alumina (grade 11, Woelm) with the following eluents: absolute ether, 0.5, 1.0, and 2.0y0 methanol in ether, finally chloroform. Each fraction was tested by t.1.c. The 0.5% methanol in ether eluate gave 50 mg. of the starting material. The 27, methanol in ether eluate yielded 542 mg. of crude crystalline material which through repeated fractional crystallization from chloroform-ether and from methanol furnished 350 mg. of rhombic prisms: m.p. 221~ f 2' ( c 0.86, CHCl,), A",? 2.78 and 2.91 p 223", [ a ] z o-64.3 (OH and NH). It was identical in properties (melting point, mixture melting point, and infrared spectrum) with an authentic specimen of Sa-hydroxydihydrosolasodine(11).
( 1 ) Visiting Scientist (1963-19641, National Institutes of Health.
( 2 ) R . E. Marker. T. Tsukamoto, and D. L. Turner, J . A m . Chem. Soc., 63,2525 (1940).
(3) S.Hayakawa a n d Y .Sato. J . Org. Chem., 38, 2742 (1963). (4) L . H . IJriggs. R . P. Newbold. and N . E. Stace, J . Chem. SOC., 3 (1942). ( 5 ) Y .Sato and S.Hayakawa, J . Org. Chem.. 18, 2739 (1963).
(6) Melting points were taken on the Kofler block and are uncorrected. Microanalyses were performed b y the Microanalytical Services Unit of this laboratory under the direction of Dr. W. C . Alford. T h e infrared spectra were taken on the Model 21 Perkin-Elmer infrared spectrometer b y Mr. H. K . Miller a n d Mrs. A . H. Wright of this laboratory.
NOTES
DECEMBER, 1964 Anal. Calcd. for CZ7H45N03: C, 75.13; H, 10.51. Found: C, 74.88; H , 10.50. The acetate IIa, prepared in the usual manner (acetic anhydride-pyridine, 1-hr . refluxing) and purified through chromatography (alumina), afforded white prisms: m.p. 197-199"; [ a l Z o-26.7 ~ f 1.0' (c 1.01, ethanol); ::A: 2.77(OH), 5.75and 8.07 (3-Ac0), and 6.02 p (N-Ac). Anal. Calcd. for C 3 1 H d 0 5 : C, 72.19; H, 9.58. Found: C , 72.36; H , 9.64. The mother liquor, after removal of 11, yielded 68 mg. of a substance whose RI value (t.1.c.) was slightly lower than that of 9ahydroxydihydrosolasodine and which gave a green coloration on spraying with sulfuric acid (507,). I1 gave a violet coloration with sulfuric acid. The impure crystals of I11 were repea,tedly crystallized from methanol-ether until homogeniety was achieved as shown by t.1.c. Seedles, m.p. 216-219", [ a I Z o-37.6 ~ f 1" (c 1.0, CHCl,), were obtained. I11 analyzed for a monohydroxydihydrosolasodine and agreed in properties (melting point, mixture melting point, and infrared spectrum) with a sample of 7Phydroxydihydrosolasodine (111, solasodane-3p,7p-diol) prepared by the reduction of 7~-hydro~ysolasodine.~ Molecular rotation difference (AMD= -MDof I11 - MD of I ) of 102 also agrees well for a 7P-hydroxy-5a-steroid listed as +110.? Anal. Calcd. for C2,Hd5N03: C, 75.13; H , 10.51. Found: C, 75.26; H , 10.73. The acetate IIIa failed to crystallize but t.1.c. and g.1.c. indicated it to be a homogenous product. I t analyzed for a O,O,Ntriacetyl derivative: [ a I z o D 4-7" (c 0.578, CHC13); Ai:: 5.76 and 8.05 (OAc), and 6.02 p (N-Ac). Anal. Calcd. for C33H5&OB: C, 71.06; H, 9.22. Found: C, 70.84; H, 9.52. Eiydrogenation of Pa-Hydroxyso1asodine.-A solution of 43 mg. of Sa-hydroxysolasodine5 in 5 ml. of glacial acetic acid was reduced over 40 mg. of palladium-charcoal (10%) catalyst until slightly more than 1 mmole of hydrogen was absorbed. The product, when chromatographed on alumina (Woelm, grade 11) and eluted with 0.5 2.5% methanol in ether, afforded 5.2 mg. of crystalline substance of m.p. 220-223' from methanol-ether. The properties of this compound were in agreement (melting point mixture melting point, and infrared spectrum) with those obtained from the microbial hydroxylation of 5,6-dihydrosolasodine. Hydrogenation of 70-Hydroxysolasodine .-To a solution of 30 mg. of 7p-hydroxy~olasodine~ in 6 ml. of ethyl acetate was added 50 mg. of 10% palladium-charcoal and the mixture was hydrogenated until 1 mole equiv. of hydrogen was absorbed. Thin layer chromatography and infrared spectrum indicated the product to be a mixture of the hydroxy- and ketodihydrosolasodine. The presence of the latter was suspected as due to impure starting material. The crystalline residue was therefore dissolved in 5 ml. of pyridine and reduced with 30 mg. of sodium borohydride. The residue was then submitted to alumina chromatography ( Woelm, grade I ) . The fractions eluted with chloroform yielded needles of m.p. 216-218" from acetone and were identical (melting point and infrared spectrum) with the product from the microbiological hydroxylation.
-
(7) L. F. Fieser and M. Fieser, "Steroids," Reinhold Publishing Corp., New York, N. Y., 1959, p. 179.
The Reaction of Organic Azides with Benzyne G. A. REYNOLDS
3733
bond, should react readily with azides. This was confirmed several years ago when phenyl azide was treated with benzyne and 1-phenylbenzotriazole was obtained in small yield. The difficulty and danger involved in the preparation of benzyneZdiscouraged further investigation of this reaction.3 The recent publication of a convenient synthesis for benzyne4 led to a renewed interest in the problem. Benzyne has been found to react readily with aromatic, aliphatic, and certain heterocyclic azides. The over-all reaction, including the preparation of the benzyne, is shown in the following reaction sequence. The
uNH2 - aNz'i- C,H90NO
coz-
COZH
RI
L
I
experimental procedure consisted in the slow addition of an acetone solution of anthranilic acid to a refluxing chloroform solution of butyl nitrite and the organic azide, thus eliminating the isolation of the explosive diazobenzoate. The substituted benzotriazoles that were prepared by the procedure are listed in Table 1. It is evident that substituents and unsaturation on the aromatic azide do not affect the reaction adversely. It is thought that higher yields of product would be obtained if a metering pump had been used to obtain slower addition of the acetone solution of anthranilic acid. The higher yields of product that were obtained from the aliphatic azides are probably due to better recovery of the product from the reaction mixture. The results obtained with heterocyclic azides were more complex. The azides 11,111,and IV were treated with benzyne by the procedure just described, but the azides were recovered unchanged. It was thought that
I1
I11
lv
these azides might not be typical because of interaction between the azide group and the heterocyclic nitrogen atom. The reaction was repeated with 9-azidoacridine (V) which should not be subject to this type of interaction, and 1-(9-acridyl)benzotriazole (VI) was obtained in 47y0 yield.
Research Laboratories, Eastman Kodak Company, Rochester, New York 14660 Received August 18, 1964
Organic azides react rather sluggishly with olefins and acetylenes to yield triazolines and triazoles. l Since the illcorporation of the double bond in strained cyclic coinpounds greatly accelerates the reaction rate, it was thought that benzyne, which contains a strained triple ( I ) J. H. Boyer and F. C. Canter, Chem. Rev., 64, 1 (1954).
V
VI
(2) M. Stiles and R. G . Miller, J . Am. Chem. Soc., 81, 3802 11960). ( 3 ) At about this time, Wittig reported t h a t 1-phenylbenzotrlaeole was formed from phenyl azide and benzyne which was prepared by a different procedure: G. Wittig and R. W . Hoffman, Angew. Chem.. 73,435 (1961). (4) L. Friedman and F. M. Logullo, J . Am. Chem. Sac., 84, 1549 (1963).
NOTES
3734
VOL. 29
TABLE I SUBSTITUTED BENZOTRIAZOLES
B M.p. or b.p. (mm.), "C.
R
0 CH,CO
0
CHO*
-
m z T . > C = C H O CSHW
C,B
%-
%-
Yield, %
-Calcd., C
H
N
-Found, C
H
N
89-90
52
73.8
4.6
21.6
74.0
4.5
21.5
160-161
50
71.0
4.7
17.8
71.2
4.5
17.8
156-157
50
70.0
4.0
18.8
69.7
4.0
18.6
180-181
55
67.9
4.4
17.6
67.5
4.3
17.9
135-136 ( 0 . 2 ) 127-128 ( 0 . 2 )
70 68
71.0 68.5
8.4 7.4
20.7 24.0
71.0 68.3
8.6 7.5
21.0 23.8
250 dec.
47
77.0
4.1
18.9
76.6
4.2
18.8
Experimental General Procedure.-A solution of 7 g. (0.051 mole) of anthranilic acid in 60 ml. of acetone was added over a 2-hr. period to a refluxing mixture of 0.05 mole of theazide and 6 g . (0.058 mole) of butyl nitrite in 200 ml. of dichloromethane. The volatile components were stripped from the reaction mixture with an aspirator to leave a brown gum. In the examples with aliphatic azides, the residues were distilled. The residue obtained from the reaction with phenyl azide and with V were dissolved in benzene, chromatographed through Florid, and, after evaporation of the benzene, the residues were recrystallized from petroleum ether (b.p. 35-60') and ethanol, respectively. The brown gums obtained in the other examples listed on Table I were all recrystallized from benzene. Azides 11, 111, and I V were treated in the same manner as phenyl azide, but evaporation of the solution which had been passed through the chromatography column yielded only the starting azide.
proton pea :s a t 260 C.P.S.(broad, -CONH SHCH3)an( 434 C.P.S.(-CONHNHCH,) which were exchangeable with DzO. The reaction of 1 with phenylacetic anhydride gave an oil which was characterized as 1methyl-1-phenylacetylhydrazine (3) by conversion to the p-nitrobenzalhydrazone (4). In each of these reactions a minor amount of the other isomer, 3 from the ester and 2 from the anhydride, was detected in the product mixture by thin ,layer chromatography. The bis(phenylacety1) derivative 5 was obtained from either 1 or 3 by treatment with phenylacetyl chloride; an intermediate monoacyl product could not be isolated. CBH~CH~CONHNHCH~
t-
2
CHsNHNHz
----f
CBH~CHZCON-NHz
1
3 LHa
I
Phenylacetyl Derivatives of Methylhydrazinel WILLIAMJ. THEUER AND JAMES A. MOORE
C~H~CH~CONHNCOCHZC CBHbCHzCON-N=CHCsHaNOz BH~
Departmat of Chemistry, University of Delaware, Kewark, Delaware 19711 Received August 14, 1964
The only systematic studies of the acylation of alkylhydrazines have been reported by Hinman and Fulton,2 who found that there action of methylhydrazine with esters gave mainly 1-acyl-2-methylhydraziries and with anhydrides, 1-acyl-1-methylhydrazines. Most of the products were characterized by paper chromatography. In connection with a projected pyridazone synthesis, we had occasion to examine the reactions of methylhydrazine (1) with various derivatives of phenylacetic acid, and have confirmed the earlier conclusions on the course of these acylations. On refluxing 1 with ethyl phenylacetate, crystalline 2-methyl-1-phenylacetylhydrazine (2) was obtained in 76% yield; the n.m.r. spectrum3 contained two one(1) Supported by Grant DA-CmL-18-108-61-G-24 from the Army Chemical Corps. (2) R . L. Hinman and D. Fulton, J. Am. Chem. Soc., 80, 1895 (1958).
I
I
CHa
4
Experimental4 1-Methyl-2-phenylacetylhydrazine (2).-A solution of 0.87 g. (18.8 mmoles) of methylhydrazine and 3.1 g. (18.8 mmoles) of ethyl phenylacetate in 1.6 ml. of ethanol was refluxed for 12 hr. On cooling, 1.59 g. (7670) of a fluffy crystalline precipitate, m.p. 113-116', separated. Recrystallization from ethanolether gave white plates of 2 : m.p. 126-127'; XKBr 3.02 and 6.13 H ; n.m.r., 149 (singlet, 3), 208 (s, 2), 260 (broad, l ) , and 434 c.p.5. (multiplet, 6 ) . Anal. Calcd. for C,HIZNZO: C, 65.83; H , 7.37. Found: C, 65.76; H , 7.42. The ethanolic mother liquor from 2 showed three spots on a t.1.c. plate, corresponding, in order of decreasing RI value, to ethyl phenylacetate, 3 and 2 . 1-Methyl-1-phenylacetylhydrazine (3).-To a solution of 887 mg. (19.3 mmoles) of methylhydrazine in 1.0 ml. of benzene was (3) N.m.r. spectra were obtained in CDCla solutions (tetramethylsilane) with a Varian A-60 instrument. (4) Melting points were observed on a Fisher-Johns block with a calibrated thermometer. Infrared spectra were obtained with a Perkin-Elmer Infracord. Thin layer chromatography was carried out on silica gel C in chloroform-methanol (25: 2); the spots were visualised with iodine vapor.
NOTES
DECEMBER, 1964 added, with stirring, 2.45 g. (9.63 mmoles) of phenylacetic anhydride5 in 5.0 ml. of benzene. The temperature rose slightly during the addition and the solution was then refluxed for I hr., cooled, and extracted with 5% aqueous KOH. The benzene layer was washed with water, dried, and evaporated, leaving 537 mg. (34%) of a viscous yellow oil which solidified upon chilling: A",: 2.94 and 6.08 p . A t.1.c. plate showed two spots, the larger one due to 3 having the higher Ri value, and the smaller having the same Rr as 2 . I-Methyl-1-phenylacetyl-p-nitrobenzalhydrazone ( 4 ).-To a solution of 78 mg. (0.48 mmole) of l-methyl-l-phenylacetylhydrazine in 0.3 ml. of ethanol was added a solution of 76 mg. (0.50 mmole) of p-nitrobenzaldehyde in 0.5 ml. of ethanol. The solution was refluxed for 10 min. on a steam bath. Upon cwoling a precipitate separated which was filtered and washed with ethanol, yielding 96 mg. (68T0) of 4, m.p. 171-172". Recrystallization from acetone-ethanol gave pale yellow needles: m.p. 172-173'; ::A: 5.94, 6.33, and 6.63 p . A n a l . Calcd. for C16H15N303 (297.30): C, 64.63; H , 5.09; N , 14.14. Found: C, 64.61; H , 5.09; S , 14.17. 1-Methyl-1,Z-diphenylacetylhydrazine(5).-To a stirred solution of 866 mg. (18.8 mmoles) of methylhydrazine in 1.6 ml. of water a t 10' was added 2.90 g. (18.8 mmoles) of phenylacetyl chloride followed by a solution of 1.29 g. of &COB in 1.5 ml. of water. After 1 hr. the mixture was extracted three times with ethyl acetate and the ethyl acetate solution was dried and evaporated to a pale yellow oil, which crystallized after standing for 3 days; 1.35 g. (494",), of white crystals, m.p. 75-83', was obtained. Recrystallization from ethanol-ether gave white crystals: m.p. 85-86'; ::A: 3.02 and 5.94 p ; n.m.r., 182 (s, 3) 20s (s, 2), 212 (s, 21, 437 (multiplet, lo), and 467 c.p.5. ( 1 , exvhangeable with D20 in presence of acid). A n a l . Calcd. for CliHi6X202: C, 72.32; H , 6.43. Found: C, 72.46; H, 6.60. Treatment of crude 3 with phenylacetyl chloride under similar conditions gave 5 in 79% yield. (5) I. M. Heilbron, D. H. Hey, and B. Lythgoe, J . Chem. Soc., 297 (1936).
The Structure of Alloocimene Dioxide' WILLIAM C. DOYLE,JR., J. N . ROCKWBLL, E. EARLROYALS, A N D J. H. STCMP, JR.
Research Department, Newport Division, Heyden Sewport Chemical Corporation, Pensacola, Florida
3735
E
Drani~hnikov,~ who further showed that the nionoineric product reacted as a diepoxide, giving a crystalline H ) ~a, dihydroxy diiuethyl ether, tetrol, C I ~ H ~ ~ ( O and CloH16(OH),(OCH&(V), by reaction with water and methanol, respectively, in the presence of catalytic amounts of acid. Later, each of these authors proposed a different structure for the monomeric dioxide, with Desalbres4 CHI
CH3
-
I -
H&-&-CH-CH=CH-C-CH-CH;j -OH -0CHs
Received April 18, 1963
v
CH3
In 1950, Desalbres2described a product, C10H1602, obtained from the spontaneous depolymerization of the insoluble guniiiiy product resulting from autoxidation of alloociiiierie (I). This work was repeated in 1953 by CH3
CH3
I
I
HaC-C=CH-CH=CH-C=CH-CH3
I CH3
CH3
H3C-~-CH-CH=CH-~-CH-CHa
>'
I1
CHI
Y CHI
H3C-&-CH-CH-CH-h=CH-CHs
y y
I11 CH3 CHa H3C-(!-CH=CH-CH=C-CH-CH3I
I
I
0
0
IT
H2C-A--CH-cH2--CH2
-OH -0CHn CHI
I
-cH--C--CH3 I1
favoring the 2,3,6,7-diepoxide I1 based on conversion to a known mixture of CIOmethyl ketones (VI), while Dranishnikovj felt that the 2,3,4,5-diepoxide structure (111) was necessary to explain the results of his oxidation experiments. Previous work in our laboratory6 suggested that the niononier as normally prepared might contain significant amounts of the cyclic peroxide IV. I n view of these coriflictirig results, we have repeated the work of Desalbres, et al., arid in addition have obtained further evidence in support of structure 11, (1) Presented a t the 145th National Meeting of the American Chemical Society, New York, N. Y . ,Sept., 1963. (2) L. Desalbres, Bull. soc. chim. France, 1245 (1950). (3) G. L . Dranishnikov, I z v . Akad. N a u h S S S R , Otd. K h i m . N a u k , 470 (1953). (4) Yves-Rene Nares, L. Desalbres, and P . Ardizio, Bull. soc. chim. France, 1768 (1956). (5) G . L . Dranishnikov, I r v . Vysshikh Cchebn. Zauendenii, Lesn. Zh., 8 , 127 (1960). (6) J. H. Stump. Jr., and 0. G . Wilson, U. S. Patent 3,038,906 (June 12; 1962).
NOTES
3736
am
&El
Fig. 2.-N.m.r.
zio in
li
i.0
1
o m
spectrum of T'II a t 63 Mc./sec. in carbon tetrachloride relative to tetramethylsilane.
2,3,6,7-diepoxy-2,6-diiiiethyl-4-octene, which for convenience we designate alloociniene dioxide (AOD). Redistilled AOD, prepared by a recently described niethod,6 gave two incompletely resolved peaks when examined by gas chromatography. Careful fractionatioii of the m'xture a t 20-30-plate efficiency yielded 5% each of the t o components, IIa and IIb, each in about 95% purity. The n.1ii.r. spectra of IIa and I I b (Fig. 1) each have three well-defined groups of signals centered near 1.2, 3.0, and 5.6 p.p.ni. [relative to Si(CHa),]in the ratio of 12 : 2 : 2. This distribution of methyl, oxirane, and vinyl protons is in agreement with structure I1 and would seem to exclude both 111and IV.
L
IIa
IIb
and may therefore
be ascribed to proton 1. The other, at 3.18 in IIa and 3.12 in IIb, due to proton 2, is spin coupled ( J = 4-6 c.P.s.) with the adjacent vinyl proton 3 in each isomer. In the case of I I b each meiiiber of the resulting doublet is further split ( J = 2 c.P.s.). One explanation of this secondary splitting is that the IIa and I I b differ in their geometry around the central double bond as shown and that the trans structure permits a long-range coupling of protons 2 and 4. It may be argued that the structures for IIa and I I b niight be reversed and that the observed small coupling is due to a H-C/
by a method in which a strongly acidic resin catalyst was substituted for the sulfuric acid of Dranishnikov.* The differences in the infrared spectra of IIa and I I b are slight and involve absorption bands of weak to medium intensity. A band at 990 c n r ' in IIb, not present in IIa, agrees well with expected trans olefin absorption. Bands which disappear 011 reaction with methanol to give V and may be presumed to be due to the intact oxirane ring are a t 1270 and 805 cni.-' in IIa and a t 1340, 840, and 780 c m - ' in IIb. The two oxirane rings in I1 have reactivities sufficiently different that stepwise ring opening by aqueous diniethylamine is possible and by proper choice of reaction conditions either the mono VI1 or the bis VI11 dimethylamino derivative may be prepared in good yield. CH3 CHI H3C-C-CH-CH=CH-d-CH-CH3 I
I1
I
/
\/ 0
HO N(CH3)z 1'11
CH3 I
H3C-k-CH-CH=CH-C-CH-CHa I
HO
I
k(CH3)2
-OH
-N( CH3h VI11
The structure shown for VI1 is assigned on the basis of the following interpretation of its n.m.r. spectrum (Fig. 2). In comparison with Fig. 1, the signal due to proton 2 (3.12-3.18 p.p.m.) in I1 has shifted to higher field a t ca. 2.6 p.p.ni. This is the direction and magnitude of the shift to be expected' for a proton cy to a dialkylamirio group us. an oxirane proton. Had this proton become cy to a hydroxyl group, as required by opening of the ring in the opposite direction, its signal should be found at considerably lower field. Experimentals8s
Examination of the oxirane proton region (ca. 3.0 p.p.ni.) discloses two distinct single-proton signals. One, a t 2.92 in IIa and 2.81 in IIb, is a 1: 3 : 3 : 1 quartet typical of the grouping H-C-CHB
VOL. 29.
I \c=c,/H I
relationship between protons 2 and 4. I n any case, the difference between the two isomers niust involve the environment near proton 2 and an olefinic cis-trans relationship appears to best fulfill this requirement. To aid in the interpretation of the infrared spectra of IIa and IIb, the dimethoxy compound V was prepared
Gas Chromatography of Alloocimene Dioxide.-Material prepared by a recently described method6 was examined by gas chromatography and found to consist of 15y0 (area) of a number of low-boiling components, the remainder giving two partially separated peaks of about equal areas. In the experiments described below, redistilled material, from which the low-boiling components had been removed, was used. Alloocimene Dioxide Isomers (IIa and IIb) .-Alloocimene dioxide6 (4 l., n z o1.46995, ~ dm 0.9448) was distilled through a 1.5 in. X 4 ft. column, containing stainless steel protruded packing, a t 5-mm. pressure and a reflux ratio of 4: 1. Each fraction (200 ml.) was analyzed by gas chromatography, the final fraction (no. 18), b.p. 79" (5 mm.), 7 2 2 0 ~ 1.4638", d2J 0.9539, being 95% IIbl 3050 (w), 3000, 2970, 2920 (a), 2880 (m), 1810 (w), 1710 (w), 1665 (w), 1460 (a), 1425 (m), 1380, 1375 (a), 1320 (m), 1270 (m), 1260 (m), 1255 (m), 1205 (w), 1160 (m), 1120 (m), 1075 (m), 1040-1030 (w), 990 (shoulder), 968 (a), 890 (m), 865 (a), 820 (m), 805 (w), 760 (m), 725 (w), 705 (m), 680 (m), 630 (w) cm.?. Anal. Calcd. for CloH,,02: C, 71.42; H, 9.52. Found: C, 71.67; H , 9.69. Redistillation of fractions 5-8 through a 1.5 in. X 2 ft. column, with the same type packing, a t 5 mm. gave a 200-ml. center ~ d* 0.9519, 95% I I a by g.1.p.c.r cut, b.p. 76-77', 7 ~ 2 01.4616', i::: 3500 (w), 3050-2960-2930-2880 (a), 1710 (w),1630, 1675 ( 7 ) L. M. Jackman, "Nuclear Magnetic Resonance Spectroscopy;': Pergamon Press Inc., New York, N. Y., 1960, pp. 55-56. (8) Gas chromatography was carried o u t using a Burrell Kromo-Tog Model K2 with a 2.5-m. polyethylene glycol on Chrornosorb W column and helium as carrier gas. (9) Infrared spectra were determined with a Perkin-Elmer Model 237 double-grating spectrophotometer.
DECEMBER, 1964 (w), 1450 (s), 1425 (m), 1370 (s), 1340-1320 (m), 1275 (w), 1255 (m),llt50(m),1125 (m),l075(m),1050 (w), 1040 ( w ) , 9 7 5 (s),890 (m), 870 (9); 840 (m), 820 (m), 780 (wj, 760 (w), 705 (w), 680 (m) cm.-l. Anal. Calcd. for CIOHI6O2:C, 71.42; H , 9.52. Found: C, 71.58; H , 9.61. Oxirane Determination.lo-To 0.3116 g. of IIa was added 25.0 mi. of 0.2 S dry hydrogen chloride in ether. After 2 hr., 50 nil. of neutral ethanol was added and the excess hydrogen chloride was titrated with 0.100 S aqueous potassium hydroxide using a Beckman Model B pH meter. A sharp break occurred in the pH us. titrant volume curve a t p H 3.2 t o 8.2 (midpoint 6.2) corresponding to 2.48 mequiv. of oxirane (1.39 oxirane rings/C1oHieOz). Further titration of the sample gave a second pH break a t 8 to 10 and corresponded to titration of the total amount of hydrogen chloride introduced. A similar titration of IIb showed 1.26 oxirane rings. Reaction with Methanol.-To a stirred mixture of 288 g. (9.0 moles) of methanol and 8.5 g. of Amberlyst-15 acidic resin catalyst (Rohm and Haas) was added 190 g. (1.13 moles) of distilled alloocimene dioxide (mixed isomers) during 35 min. a t 2530" with ice bath cooling. After an additional 45 min. a t room temperature, the catalyst was filtered off, and washed with two 50ml. portions of methanol, and the combined filtrate was vacuum stripped to remove the solvent. The residue was pot distilled from 10 g. of potassium hydroxide, giving the dimethoxy product V as a single 186-g. cut, b.p. 137-141" a t 5 mm., 77% pure by g.c. Redistillation gave 987, pure V, ~ Z O D1.4667", d20 1.0215 showing fiy:;: 3450 (s), 3000-2950-2900 (s), 2800 (m) 2750 (m), 14501460 (m, doublet, ) 1380 (m), 1290 (w), 1230 (wj, 1165 (m), 1105 (s), 1095 (vs) 1040 (w), 990 (m), 952 (mj, 920 (w), 875 (w), 828 (wj, 750 (w) cm.?. A n a l . Calcd. for Cl2HZ4Oa:C, 62.07; H , 10.34. Found: C, 61.92; H , 10.50. Reaction with Dimethvlamine. Monodimethvlamino Derivative VI1.-To 630 g. (3.75 moles) of alloocimene dioxide (mixed isomers) was added with stirring 539 g. (3.0 moles) of 257, aqueous dimethylamine over a 2.75-hr. period a t 45O, slight intermittent heating being required to maintain this temperature. The mixture was stirred an additional 2 hr. a t 45O, with increasingly more heat being required. The organic layer (739 9.) was separated and distilled a t 3 mm. through a 1.5 X 24 in. column containing stainless steel protruded packing. After a small forerun consisting mainly of water, 274 g. of recovered alloocimene dioxide, b.p. 60-85', was collected. This was followed by a 276-g. product, cut, b.p. 112-123" (61%, yield based on unrecovered AOD). The product cut was pot distilled; a center cut, b.p. 114115' a t 5 mm., nzo~1.4674, d20 0.9520, taken for analysis, showed 3400 (s), 2950-2900-2850-2800-2750 (s), 1650 (w), 1450 (s), 1365 (s), 1320 (m), 1250 (m), 1205 (w), 1160 (m), 1110 (m), 1060 (m), 1020 (m), 970 (s), 925 (w), 910 (w), 868 (m), 820 (w), 750 (w), 700 (w), 680 (w) cm.?. A n a l . Calcd. for CjzH~aN02: C, 67.60; H, 10.79; N, 6.57. Found: C,67.04; H , 10.77; N,5.87. Bisdimethylamino Derivative VII1.-A mixture of 312 g. (1.85 moles) of alloocimene dioxide and 2000 g. (11.1 moles) of 25% aqueous dimethylamine was stirred at room temperature. During the first 2 hr. the temperature gradually rose to 45" and the originally two-phase system became homogeneous. After 3 hr. more, during which time the solution cooled back to 30°, the bulk of the water and excess amine was removed by distillation a t atmospheric pressure to a pot temperature of 130" and the remainder stripped away under vacuum. Pot distillation of the residue gave a single product cut (389 g., 82% yield), b.p. 138144' a t 5 mm. Redistillation gave a n analytical sample, b.p. 150-153' a t 8 mm., n% 1.4804", d20 0.9653, which showed C:: 3400 (s), 2980-2950-2850-2810-2780 (s), 1650 (w), 1450 (s), 1360 (s), 1300 (w), 1250 (w), 1170 (m), 1150 (m), 1100 (w), 1045 (s), 1015 (m), 990 (m), 970-960 (m), 910 (w), 835 (m), 715 (w) em.-'. A n a l . Calcd. for C14HaOS202: C, 65.11; H, 11.62; N, 10.85. Found: C,65.16; H, 11.53; N, 10.87.
:::Y
Acknowledgment.-The authors wish to express sincere appreciation to Dr. Leon Mandell, Emory University, for his aid in the det,ermination and interpretation of n.m.r. spectra. (10) D. Swern, T. W. Findley, G. N. Billen, a n d J. T. Scanlan, Anal.
Chem., 19, 414 (1947).
3737
NOTES Unsymmetrical Disulfides from an Amino Bunte Salt DANIELL. KLAYMAN, JERRY D. WHITE, ASD THOMAS R. SWEENEY
Walter Reed A r m y Institute of Research, Department of Medicinal Chemistry, Washington, D . C . 20012 Received J u l y 31, 196'4
The reaction of Bunte salts (alkyl thiosulfates) with mercaptans has been used with varying success in the synthesis of unsynimetrical disulfides. The reac~ of niercaption apparently proceeds via an S N attack tide ion on the sulfur atom attached to the alkyl group in the Bunte salt.5 There is no instance reported in the R S$
+ K'S-bS03-
e RSSR' + S03-*
literature, however, in which a Bunte salt bearing an amino group was used for the synthesis of a mixed disulfide. It was of interest, therefore, to see if 2-aminoethanethiosulfuric acid could be used to prepare an unsymmetrical disulfide containing the cysteamine (mercaptoethylamine) moiety. Using the method of 2-aminoethanethiosulfuric acid was conibined with 1 equiv. of a mercaptan in both aqueous and aqueous ethanolic media. A stream of nitrogen was bubbled into the solution to remove the sulfur dioxide anticipated as a by-product. The reacHzXCHzCHzSSOSH
+ RSH + HzXCH2CHzSSR
+ SO, + Hz0
tion failed to proceed a t room temperature. On prolonged heating a very slight evolution of sulfur dioxide was detected, but nearly all the starting Bunte salt was recovered unchanged on work-up of the reaction niixture. The reaction of sodium 2-aminoethanethiosulfate with a sodium mercaptide in water by the method of Footner and Smiles' proceeded with a slow uptake of the thiol. The use of methanol as the solvent for the above reaction markedly increased the rate of the reaction. A prompt precipitation of sodium sulfite occurred and the mercaptan was consumed in less than 5 min. The insolubility of sodium sulfite in methanol presumably served to drive the reaction to completion. Performance of the reaction a t Oo, rather than at room or elevated temperatures, improved yields. Reactions run a t -10" or below were too slow to be useful. The mixed disulfides were isolated and purified as hydrochloride salts (Table I). RS- + HzSCHzCHzSSO3- e HzXCH2CHzSSR + SO3-2 Yields of unsynimetrical disulfides were rarely greater than 60% due to the formation of considerable quantities of syninietrical disulfides. The ease with which mixed disulfides disproporionate into symmetrical disulfides has been observed by Schoberl and Rauer. H. n. Footner and 9. Smiles. J . Chem. S o c . , 197, 2887 (1925). A . Schoberl and G. Rauer, Angem. Chrm., 6 9 , 4 i 8 (1057). J. M.Swan, Nature, 180, 643 (1957). R. Rlilligan and J. M. Swan, J . Chem. Soc., 6008 (1963). 0. Foss. "Organic Sulfur Compounds," Val. I , N. Kharasch, E d . , Pergamon Press, Inc., New York, N. Y., 1961, p. 83. (1) (2) (3) (4) (5)
3738
NOTES
VOL. 29
TABLEI RSSCHzCHzNHz HCI M.p.;
Yield,
R
OC.
%
n-Butyl n-Hexyl n-Octyl n-Decyl Phenyl Benzyl
96-97 112-115 112-114 111-112 137-138 150-151
40.5 42.4 49.0 68.4 63.2 50.9
C
Ethanol Carbon tetrachloride 2-Propanol-acetonitrile Ethanol Acetonitrile 2-Propanol
35.71 41.80 46.57 50.90 43.32 45.84
They found that in alkaline or strongly acidic solutions, a t elevated temperatures especially, mixed disulfides disproportionate rapidly. Because of the possibility 2HzKCHzCH2SSR L_ RSSR
Calcd., yoH N
I -
Recryst. aolvent
7 99 8.77 9.38 9.87 5.46 5.98
6.94 6.10 5.43 4.90 6.32 5.94
Experimental6 2-Aminoethyl Alkyl (or Aryl) Disulfide Hydrochlorides.-To a solution of 3.0 g. (0.075 mole) of sodium hydroxide in 60 ml. of methanol, through which nitrogen was bubbled, was added 7.08 g. (0.045 mole) of 2-aminoethanethiosulfuric acid.' When solution was complete, the flask was immersed in an icewater bath and the temperature was lowered to ca. 0". The mercaptan (0.03 mole), previously distilled under nitrogen, was added to the solution, causing a slightly exothermic reaction and an almost immediate precipitation of sodium sulfite. After the addition of the mercaptan, aliquots of the mixture were centrifuged and the supernatant liquid was tested for the presence of unreacted mercaptan with sodium nitroprusside. When the test became weakly positive or negative, the time varying from 2 to 5 min. after the addition of the mercaptan, the reaction mixture was suction filtered through Whatman No. 50 filter paper. The cloudy filtrate, kept a t O - l O o , was neutralized with ethanolic hydrogen chloride. The solid which formed, most of which was cystamine dihydrochloride, m.p. 210-215" dec., was removed by filtration and the filtrate was evaporated to dryness under reduced pressure at 50' using a rotary evaporator. The white, waxy residue was recrystallized several times to obtain an analytical sample.
Acknowledgment.-We wish to thank Dr. David P. Jacobus for helpful discussions concerning this work. (6) Microanalyses were performed by Mr. Joseph Alicino, hletuchen, X . J. Melting points were determined on a Fisher-Johns melting point
apparatus and are uncorrected. (7) H. Bretschneider. Monolsh., 81, 372 (1950).
S
C
31.78 27,90 24.87 22,43 28,92 27.20
35 41 46 50 42 45
66 68 83 52 84 82
8 9 9 9 5 6
30 01 55 81 47 21
6 5 5 5 6 6
-
89 57 42 41 44 07
S
31.85 27.84 24.75 22.43 29.26 27.21
Preparation of N-Perchlorylpiperidine DAVIDM. GARDNER, ROBERTHELITZER, AND CHARLES J. MACHLEY
+ HzNCHzCH&SCHzCHz?JHz
that the cystamine [bis(Zaminoethyl) disulfide] dihydrochloride isolated from each reaction mixture may have originated, not only by disproportionation of the mixed disulfide, but by the alkaline decomposition of 2-aniinoethanethiosulfuric acid, the stability of the latter compound was determined under experimental conditions. The amino Bunte salt (1 equiv.) was stirred in a nitrogen atmosphere for 5 hr. with 1.75 equiv. of sodium hydroxide in methanol a t 0". Only a trace of sodium sulfite (
