3223
EDWARD S. ROTHMAN AND MONROE E. WALL [CONTRIBUTION FROM THE
IrOL
79
EASTERN REGIONAL RESEARCH LABORATORY1]
Steroidal Sapogenins. XLII. Partial Synthesis of 11-Ketodiosgenin? BY EDWARD S. ROTHMAN AND MONROE E. WALL RECEIVED DECEMBER 13, 1956 The steroidal sapogenin gentrogenin (33-hydroxy-22a,25D-spirost-5-en-12-one, I), considered as a suitable cortisone precursor substance, was transformed to 11-ketodiosgenln (3B-hydroxy-22a,25D-spirost-5-en-ll-one, VIa) and t o l l a - h y lTa). The route employed involved the tetrabromination of I t o 36droxydiosgenin (22a,25D-spirost-5-en-3~,lla-diol, acetoxy-Sa,6~,11~,23-tetrabromo-22a,2.jD-spirostan-12-one(11). This tetrabromide was selectively debrominated t o restore the olefinic bond by means of iodide ion t o form 3~-acetoxy-lla,23-dibromo-22a,2~D-spirost-5-en-l2-one (111). Hydrolysis of I11 formed, after acetylation, 23-bromo-3P,12B-diacetoxy-22a,25D-spirost-5-en-ll-one ( I V b ) which was de(IVc). Selective deacetoxylbrominated wTith zinc dust in acetic acid t o form 3B,12~-diacetoxy-22a,25D-spirost-5-en-ll-one ation of IVc with calcium, sodium or lithium in liquid ammonia gave rise t o 11-ketodiosgenin (VIa) or t o lla-hydroxydiosgenin, depending on reaction conditions utilized. Small amounts of 3-desoxy sapogenins also were detected as by-products.
In the course of examination of natural steroidal materials of botanic origin as possible source materials for steroid hormones, we have been studying the 12-ketosteroids hecogenin, gentrogenin3 and ~orrellogenin.~When we read of the discovery by Chapman, Elks and Wyman4 of the conversion of hecogenin to 11-ketotigogenin by the route of selective deacetoxylation of 3/3,12P-diacetoxy-5a,22a,25D-spirostan-ll-one by means of calcium in liquid ammonia, i t seemed that such a method might be advantageously applied to gentrogenin to prepare the unknown C-11 oxygenated derivatives of diosgenin. Since 11-ketotigogenin is a commercially utilized intermediate for cortisone production, it was reasonable to expect that its 5,6-dehydro derivative, 11-ketodiosgenin, might be similarly used. The present communication describes the preparation of 11-ketodiosgenin and 11a-hydroxydiosgenin. For this work we had available a ketonic sapogenin fraction obtained from Dioscorea spiculi$ora5 containing a mixture of the 25D- and 25L-diastereoisomeric sapogenins gentrogenin and correllogenin, the 25D-isomer greatly predominating in the mixture. Fortunately, separation of isomers is unnecessary for practical work since conversion to derivatives in the C-21 series expels the single differing asymmetric center and leads to a single series of compounds. The constants reported in this paper are for sterically pure 25D-gentrogenin derivatives, although we have carried out the same transformations on the natural mixtures as well. Bromination of I6 gave a 90-94% yield of crystalline 3P-acetoxy-5a,G@, 11a,23-tetrabromo-22a,(1) A laboratory of t h e Eastern Utilization Research and Development Division, Agricultural Research Service, United States Uepartment of Agriculture. Article not copyrighted. (2) Paper XI.1, H. E. Kenney and AI. E. Wall, J . Org. Chem, 22, 468 (1957). (3) For a preliminary description of these new sapogenins see H. A. Walens, S. Serota and M. E. Wall, THISJOURNAL, 77, 5196 (1955). A more complete description h a s been submitted t o J . Org. Chem. (4) J . H. Chapman, J. Elks and L. J . Wyman, Chemistry 6'Indrrst r y , 603 (1955). (5) This material was collected in Chiapas, Mexico, by Drs. H. S. Gentry and D. S. Correll. Horticultural Crops Research Branch, Agricultural Research Service, Beltsville, Md. (6) T h e conditions of t h e bromination were essentially those used by J. W. Cornforth, J . hl Osbond and G. H. Phillips, J. Ciiem. Soc., 907 (1954), b u t several differences from results with hecogenin acetate are worth mentioning. We obtained a 3770 yield of crystalline llm,23dibromohecogenin acetate, melting point 188-190°, rather t h a n t h e 5470 yield reported by Cornforth, et al. (See also G. P. Mueller, L. L. Norton, R. E. Stobaugh, L. Tsai and R. S . Winniford, THISJ O U R N A L . 75, 41192 jll)53), for yield and melting point data.) T h e mother
25D-spirostan-12-one (11). This crystalline tetrabromide on treatment with zinc in refluxing acetic acid for 1.5 hr. regenerated starting material I essentially completely (95%). However, selective regeneration of the 5,G-olefinic bond with iodide gave only a 48-G3% yield of the desired dibromide 3B-acetoxy-lla,23-dibrorno-22a,25D -spirost-5-en12-one (III).' The mother liquors were not useful for reconversion to starting material. The speed of the debromination was surprisingly rapid. The reaction in acetone solution was complete in less than six minutes a t room temperature. The hydrolysis of 11a,23-dibrornohecogenin acetate had been studied by Mueller, et a1.,6 and by Rosenfeld and Gallagher.* We applied the conditions of the latter authors to 11a,23-dibromogentrogenin acetate (111) and obtained, after reacetylation and debromination, the desired product 3~,12/3-diacetoxy-22a,25D-spirost-5-en-1l-one (IV c ) . ~The 12P-hydroxyl group adjacent to the 11-ketone group does not reliably acetylate with pyridine and acetic anhydride on the steam-bath so that we used perchloric acid catalysis.1° When we attempted to carry out the hydrolysis reactions in smaller volumes by using refluxing ethanol in 2-hr. reactions or when we used water-miscible co-solvents (e.g., tetrahydrofuran), the yields were sharply reduced. A carboxylic acid by-product was obtained in low yields from the hydrolysis reactions. Wendler, Hirschmann, Slates and Walker, working with 11,12-dimesylates, obtained by alkaline treatment a C-ring carboxylic norsteroid.12 We suspect a liquors from t h e crystallization were deep green, viscid residues which could be freed of bromine by treatment with zinc and refluxing acetic acid b u t which yielded only half t h e calculated content of crystallizable hecogenin acetate. (7) Attempts t o improve the yield of I11 by treating IT with zincdust in benzene or tetrahydrofuran were not successful since t h e l l a bromine a t o m was more susceptible t o reductive attack than t h e 5,Bbromine a t o m pair. T h e direction of attack could be indirectly observed by infrared spectral means since t h e l l a - b r o m o ketone absorbs at 1735 cm.-l, while the free 11-ketone absorbs in the usual region near 1709 cm. -1. (8) R . S. Rosenfeld and T. F. Gallagher, THIS J O U R N A L , 77, 4367 (1955). (9) We confirm t h e 9270 yield for 20-hour, room temperature hydrolysis of lla,?3-dibromo hecogenin acetate reported by Rosenfeld and Gallagher; however, identical reaction conditions applied t o the closely related I11 gave IVc in somewhat lower yields averaging 50 t o S5%.
(10) S e e in this connection t h e observations of Mueller, rl 01.. ref. 6. (11) N. L. Wendler, R . F. Hirschmann, H. I.. Slates and R . W Walker, Chemistry b I n d n s t u y , 901 (1954). (12) T h e carboxylate was obtained by air-oxidation of a n intermedia t e aldehyde.
3229
PARTIAL SYNTHESIS OF 1~-KETODIOSGEXIN
June 20, 1957
BLr
A
c
RO
O
W Br Br
@R = H
,
OR = A c '
R ' = Br R'= H
-
IY C
XI-b
RO
R=Ac
RO
Yb R = A c
+
H2
The immiscible toluene layer could be separated from the liquid ammonia metal solution phase in a separatory funnel provided with a fritted glass disk. We have also used bromobenzene to quench the reaction as suggested by Chapman, Elks and Wyman.14 Reduction of IVc gave a 90% yield of products consisting chiefly of 11-ketone VIa but containing about 10-20Y0 of lla-diol Va. The conversion of gentrogenin t o 11-ketodiosgenin is effected in about 35% over-all yield. A further observation in the calcium deacetoxylation experiments is the formation by the reduction system of small amounts of 3-desoxysapogenins by apparently non-specific deacetoxylation. Isolated and identified were 3-desoxy-11-ketodiosgenin (22a,25D-spirost-5-en-ll-one, VII) and 3-desoxy-12p-hydroxydiosgenin (22aI25D-spirost-5-en-l20-01, VIII) . The latter compound was characterized by oxidation to the 12-ketone, 3-desoxygentrogenin (22a,25D-spirost-5-en-12-one). The 126-configuration is tentatively suggested since the 12p-configuration is the equatorial one usually favored by metal reduction systems. The C-11 ketones can be distinguished from the c-12 ketones not only by reduced chemical reactivity of the former but also by infrared differences. Thus, the 11-ketones show a band a t 1704-1710 cm.-' (VIb a t 1706 crn.-I)I5 while the 12-ketones show the band a t 1706-1712 cm.-I (I a t 1713 cm.-I).16 I n addition the 11(13) F. Sondheimer, 0. Mancera, G.Rosenkranz and C. Djerassi, ketones show a band16 a t 970 cm.-I not given by THIS JOURNAL, 76, 1282 (1953), using calcium, sodium or lithium in liquid ammonia to reduce 3~-propionoxy-5a,22a,25D-spirost-8(9)-en~ the 12-ketones. Such a band occurs also for exrelationship between these two products but have as yet no experimental verification. We carried out our preliminary C-12 deacetoxylation experiments of conversion of IVc to 11ketodiosgenin and to 11a-hydroxydiosgenin on the basis of the communication of Chapman, Elks and W ~ m a n . These ~ authors described the parallel experiments in the hecogenin series but in their preliminary disclosure gave no experimental details. In view of the fact that even closely related sapogenins often have inherent peculiarities that affect direction of reaction and yield it may not be amiss to note here some observations on the metalammonia reduction systems as applied to the compound IVc. The selective deacetoxylations were carried out in liquid ammonia using calcium or sodium or lithium as the reducing element. The primary product of the reduction was 11-ketodiosgenin, 3phydroxy-22a,25D-spirost-5-en-ll-one(VIa), which may further react with protonic materials (e.g., water, alcohol) to form 11a-hydroxydiosgenin, 22a,25D-spirost-5-en-3Pl11a-diol (Va), in a secondary r e a ~ t i 0 n . I ~When water is used to quench the reaction and decompose excess metal reagent, a greater or lesser amount of lla-01 is formed. Henry A. Walens of this Laboratory devised an experiment avoiding this quenching operation.
11-one reported formation of the saturated ketone in the absence of ethanol and the l l a - o l in the presence of ethanol. S. Bernstein, R. Littell and J. H. Williams, i b i d . , p. 1481, used lithium-ammoniaethanol to reduce a saturated 11-ketone to an lla-01 in a similar manner.
(14) Private communication. (15) R. N. Jones and F. Herling, J . Ore. Chem.. 19, 1256 (1954). (16) J. Schmidlin and A. Wettstein, Hels. Chim. Acta, 36, 1245 (1953); cf. Fig. 1 of article for spectra.
product formed a mass of felted micro-needles melting suddenly at 199' with browning and effervescence. T h e an3lytical sample was recrystallized from methanol to give silky filaments, m.p. 199O, [ a ] z E ~-84.5'. .4nal. Calcd. for GgHaaOSBrd: C , 44.30; €1, 4.87; H r , 40.66. Found: C, 44.16; H,5.04; Br,40.65. Selective Debromination of I1 to 3p-Acetoxy-llu,23-dibromo-22a,25D-spirost-5-en-l2-one (111) .-A sample of 11, 1 1 g., in 250 ml. of ethanol was refluxed for 2 hr. with 14 g. of .sodium iodide. T h e solids went into solution during t h e course of t h e reaction. T h e mixture was cooled, tlilutetl with water and ether was added. A dilute solution of sodium thiosulfate was added in portions t o just decolorize tnoleculnr iodine, avoiding excess reagent. The ether layer was washed with 1% sodium hydroxide (to remove t r x c s of carboxylic material), with water, with saturatctl s d i n c solution and after drying v i t h sodium sulfate, was evaporated under reduced pressure at 30" t o a pale-orange frtitli. Stirring with 35 ml. of absolute ethanol caused solution r ~ f t h e froth followed by precipitation of Tvhite crystalline 111, 63(;h . z 3 T h e aualytical sample ~ r a srecrystallized from ethanol-methylene chloride t o fine hexagonal blades blackening a t 152' and melting with effervescence froni I(i0 170°, ;ru]*jn - 9 1 . 9 O . Anal. Calcd. for CZgHaaOjBrz: C, 55.60; 11, 6.11; T3r, 25.51. Fouiitl: C , 54.14; H, 6.18; n r , 24.99. T h e compound is not stable in chloroform or methylenc chloride solutions for prolonged periods of time. Aimlytically pure samples let stand overnight in chloroform solution developed deep pink coloration. 23-Bromo-3~,12~-diacetoxy-22a,25D-spirost-5-en-ll-one (IVb).-A sample of 111, ,5 g., suspended in 4 . 5 1. of 0.3 S ethanolic potassium hydroxide was stirred a t room temperature for 20 hr. T h e mixture was treated with 250 nil. Experimental of 6 E hydrochloric acid and was evaporated t o 400 ml. unRotations were measured in chloroform in a 2 dm. tube, der reduced pressure. T h e residue was diluted with water and extracted with ether. T h e ether was washed with Zr;, c = 25 mg./1.5 ml. Melting points were determined on the sodium hylroxide t o remove a solid carboxylic acid'l and was Kofler block but are otherwise uncorrected. 3p-Acetoxy-5a,6p, 1la,23-tetrabromo-22a,25D-spirostan- dried with saturated brine and with sodium sulfate. T h e 12-one (II).-Gentrogenin acetate ( I ) , 7.0 g. (0.01486 solid froth obtained by reduced pressure evaporation of t h e mole), in 76 ml. of ordinary C.P. chloroform was treated at dried ether extract was acetylated in 50 ml. of acetic acid anti 15 ml. of acetic anhydride in t h e presence of 1 ml. of 2.5 AV room temperature with a solution of 0.01486 mole of bromine in 59 ml. of carbon tetrachloride in the course of 5 minutes perchloric acid in acetic acid during 1.5 hr. a t room temper:\time. T h e initial uptake was slow, b u t t h e rate of uptake ture. The product was obtained by pouring the acetylation accelerated with time. T h e bromine was added in batches mixture into an excess of cold water and filtering off the rather than dropwise. T h e colorless solution was further flocculent precipitate. Crystallization of t h e air-dried product from ether gal-e 3.5 g. of IVb, m . p . 21?-216°. T h e treated with 1.9 ml. of liquid bromine dissolved in 15 ml. of analytical sample, [ m ] " j n -88.7", crystallized from ether, chloroform. Various colorations, sometimes yellow-orange, melted from 214-216' without decomposition, although t h e pink or green developed during this second bromination, again carried out in only 5 minutes time despite difficulties colorless melt turned slightly b r m m soon after melting was completed. in visual observation of the bromine uptake. T h e solution Anal. Calcd. for C31H410iBr: C, 61.28; H , 7.13; Br, was stirred, still a t room temperature, for a n additional 10 13.1,j. Found: C, 61.30; H, 7.02; Br, 13.25. minutes, and then all solvents Ivere removed a t 30' using water aspiration. T h e residue wa5 a pale green, glassy 3p,12p-Diacetoxy-22a,25D-spirost-5-en-ll-one (IVc) .froth which on dissolution in 25 ml. of methylene chloridez2 T h e product from t h e preceding preparation, diminished b y gave a dark green solution. Addition of ethanol incrementhe removal of 121 Ing. for analytical and reference purtally caused color change t o a pale pink. Absolute ethanol poses, was refluxed 6 hr. with zinc dust in acetic acid and the was then added as rapidly a s possible without causing floccu- product isolated by dilution with water and extractioll with lation until a volume of 2.50 ml. of clear, nearly colorless ether and washing in t h e usual manner. B sample (after removal of zinc halide) gave a positive Reilstein flame test solution was obtained. Stirring with scratching caused sudden crystallization of 11.1 g. of snow-white product (de- even after multiple crystallizations. T h e materials were composing at 193' with effervescence and reddening). recombined and retreated for a n additional 4 h r . without Recrystallization was effected by dissolution in a minimal change. Apparently the residual halogen content was volume of hot methylene chloride, dilution with ether a n d vanishingly small since the infrared spectrum of t h e product boiling off t h e methylene chloride azeotropically. T h e was typical of 25D-sapogenins and showed no band a t 739 crn-1. T h e product, 2.13 g., obtained by crystallizing the (17) C. R . Eddy, 11.E . Wall and 11. I(.Scott, A n a l . C h e n i . , 2 5 , 266 evaporated ether extract from ethanol, formed feathery, (1963). branched masses, m.p. 195-218'. T h e analytical sample, (18) 51. E. Tvall, J. J. Willaman, T. Perlstein, D. S. Correll and [ a ]2 5 -loo", ~ recrystallized from ethanol, began t o undergo 11. S. Gentry, J . ,477~.Pitarm. A 5 5 0 C , in press (3957). transition over 202' t o elongated needles, but melting oc(19) D. H. 11. Barton and W. Klyne, Cltenzistry Er I n d u s t r y , 755 curred a t 221-225' before the transition was completed. (1948). Anal. Calcd. for C:ilH1407:C, 70.43: H , 8.39. Fountl: (20) From K. Tsuda and R. Hayatsu, THIS J O U R N A L , 77, 0582 C, 70.71; H , 8.53. (1955). A. Reduction of 3p,12p-Diacetoxy-22a,25D-spirost-5-en(21) From C. Djerassi. E. Batres, M. Velasco and G . Rosenkranz, 11-one (IVc) with Calcium and Ammonia to Give a Mixture ibid., 74, 1712 (1952). of Products.--A sample of IVc, 0.55 g., in 9 ml. of tetrahy(22) If isolation of t h e tetrabromide is not required, t h e methylene drofuran was added t o a solution of 0.67 g . of calcium in 200 chloride should not be added, b u t t h e product should be triturated ml. of liquid ammonia. T h e reactants were stirred 2 hr. with ethanol t o loosen any crystalline residues from the walls of the
ample in the spectrum of tigogenin, diosgenin and 3,.J-cyclo-22a,25D-spirostane, but compounds of these types are easily distinguished by other criteria. The relative intensities of the ketone and acetate bands also vary erratically. I n hecogenin acetate the acetate band is slightly stronger than the ketone band. 16,1i In gentrogenin and correllogenin, the ketone band13 is the stronger, while the 11-ketones have a ketone band definitely weaker than the acetate band. In the course of the identification and structurc assignments of the C-11 ketonic products, we observed a substantial deviation of the sapogenin 11ketone molecular rotation contribution from the nican value of +79" calculated by Barton and Klyne.I9 d check of literature values of 11-ketone molecular rotation contributions in the sterol series gave a value2"of about +72" in good agrecnient with Barton and Klyne's value. However, the 1I -ketone contribution in 11-ketotigogenin?' is +150" and in 11-ketotigogenin acetateI6 is lG9'. The +207" molecular rotation coiitribution in 11-ketodiosgenin is even inore positive. Similar deviation in a positive direction from the calculated1g molecular rotation contribution of the 11a-cil also occurs. The calculated value is - 29" ; the observed value13is 3.5".
+
+
flask t o pervent scorching during subsequent heating. Debrominative regeneration of the 4s-bond with sodium iodide d o e s not reqixire th.1t the tetralnnmi
