CA2123684A1 - Method for making steroidal peracyl glycosides - Google Patents

Abstract

Processes for the synthesis of tigogenin beta-O-cellobioside heptaalkanoate which is an intermediate for the known hypocholesterolemic agent tigogenin beta-cellobioside. The process comprises reacting .alpha.-cellobiosyl bromide heptaalkanoate and .beta.-tigogenin in the presence of zinc fluoride or zinc cyanide under conditions capable of forming said tigogenyl .beta.-O-cellobioside heptaalkanoate. The analogous preparations of hecogenin .beta.-O-cellobioside heptaalkanoate 11-ketotigogenin .beta.-O-cellobioside heptaalkanoate, and diosgenin .beta.-O-cellobioside heptaalkanoate are also disclosed. The process provides both high .beta.-anomeric selectivity and high yields.

Description

Wo 93/ 111 50 P~/ US92/0863X
21236~

METHOD FOR MAKING STEROIDAL PERACYL GLYCOSIDES
Backqround of the Invention The present invention relates to processes for the synthesis of steroidal glycosides, and particularly to the preparation of st~roidal peracyl glycosides ~sed as intermedi.~es therein.
Tigogenin beta-O-cellobioside is a known compound having utility in the treatrnent of hypercholesterolemia and atherosclerosis (Malinow, U.S. Patents 4,602,003 and 4,602,005; Malinow e~ ai., Stero;ds, vol. 48, pp. 197-211, 1986). Each patent discloses a different synthesis of this compound from alpha-D-cellobiose octaacetate;
the first via ~he ~Iycosyl bromide heptaacetate which is couplecl with tigo~enin in the presence of silver carbonate, and finally hydrolyzed; and the second via direct stannic chloride catalyzed coupling of the cellobiose octaacetate with tigogenin in methylene chloride, again foliowed by hydrolysis. In Malinow et al., reaction of collobiose octaacetate with titanium tetrabromide gave the cellobiosyl brornide heptaacetate, which was coupled with tigogenin by means of mercuric cyanide, and then hydrolyzed. All of these methods have sorious drawbacks for producing bulk material to be used as a pharmaceutica! drug. A desir~ble goa!, met by the present invention, has been to devise synthetic methods which avoid toxic and/or expensive reagents, and which cleanly produce the desired tigogenin beta-O~cellobioside, avoiding tedious and expensive pur^dication steps.
Schmiclt, Angew~ Chem. lnt. Ed. Engl., vol~ 25, pp~ 212-235 (1~86) has reviewed the synthesis ~and reactions of O-glycosyl trichloroacetimida~es Sormed by the rea~tion of ~.ugars possessîng a 1-hydroxy group (but wîth other hydroxy groups protected, e~g., by benzyl or acet~l) with trichloroacetonitrile in the presen~e o~ a base~ There is preterential formatîon of the alpha-anomer when sodium hydride is used as base, and pr~erentialformationofthebeta-anornerwhenthebaseispotassiumcarbonat~. The alpha anomer of t~rabenzy591ucosyl trichloroacetimi'date when coupled with cholesterol gave anomeric mi)çtures which~varied with~ catalyst (p-toluenesulfonic acid or boron trifluoride etherate) and temperature (~0 to ~+20 C~. On the other hand, both the alpha and beta anomers of tetraacetyigluoosyl analog reporteclly yield exclusively beta-anomPric products.;
Thus, there has been a eontinuing search in this iield of art for irnproved methods of stereocontrolled syntheses of steroidal glycosides.

WO 9~/1 1 15(~ PC'r/US92/0~63X
21~6~

Sumrnarv ot the Invention This invention is directed to a process for lhe synthesis of tigogenin 13-O, 11-ketotigogenin ~-O, hecogenin B-O, or diosgenin B-O cellobioside heptaalkanoale that provides greater f3-anorneric selectivity and increased ylelds. The process is particularly useful ~or preparing tigogenin B-O-cellobioside heptaalkanoate, which is an interrnediate ~or the known hypocholesterolemic agent tigogenin ~-O-cellobioside. The process cornprises reacting a-cellobio~yl bromide heptaalkanoate and 13-tigogenin, 11-13-ketotigogenin, 13-hecogenin or ~-diosgenin in the presence of zinc fluoride or zinc cyanide under conditions suitabie for 1Orming the tigogenin 13-O-, 11 -ketotigogenin B-O, hecogenin 13-O-, or diosgenin B-O-cellobioside heptaalkanoate.
Other features and advantages will be apparent from the specification and claims .
etailed Description of the In~/entlon Preferably the rnetcl salt used in the stereospecific reaction of a-cellobiosyl 1~ bromide heptaalkanoate and B-tigogenin, 11-keto-B-tigogerlin, 13-hecogenin or fl-diosgenin is zinc 11uoride or zino cyanide. It is ~sperially preferred ~hat the metal salt is zinc fluoride. It is preferred that about 0.5 equivalents to about ~ equivalents and especially preferred that about 1.5 e~uivalents to about 2.25 equivalents metal salt is used.
It may also be preferred lo conduct the :zinc fluoride or zinc cyanide-activatedcoupling in the presence of additional zinc salts such as zinc halides (e.g., 2inc bromide, zinc chloride, zinc iodide) or basic salts of zinc (zinc oxide, zinc hydroxide, zinc hydroxy fluoride, zinc carbonate, etc.) to buffer or to activate the prornoter (i.e., zinc tluoride or zinc cyanide metal sait). Trialkyl tertiary amines ~e.g., diisopropylethyl 26 amine, tri~thy!amine, tri~utylamine), te~raalkyiureas ~e.g., tetramethyi urea, tetraethyl urea) or dia!kylanilines ~e.g., diisopropyl aniline, dibutylanilin~) are alss usefui reaction buffers. The above a~ditives are generaliy used at 10-50% msle equiYalents of the promoters.
Aithough any of the alkanoate ~C,~C4) substituted alpha-cellobiosyl bromides may be used it is preferred th~ acetate (i.e., Cl) is usecl. They may be prepared ~rom conventional starting materials according to methods described in K. Freudenberg and .

WO 93/11150 PCr/US~2/0863~
2123~

W. Nagai, Ann., 494,63 ~1932) (e.g. Example 3). It is preferred that about 0.5 equiva-lents to about 3 equivalents, and especially preferred that about 1 equivalent to about 2 equivalents alkanoate (C1-C4) substituted alpha-cellobiosyl bromides are used.Any reaction inert solvent may be used. As used above and elsewhere herein, 5 the expression "reaction-inert solvent" refers to a solvent which does not reaet or decompose with starting materials, reagents, intermediates or products in a rnanner which adversely affects the yield of the desired product. In general, the solvent can cornprise a single entity, or contain multiple components. Preferably the solvent is a non-protic reac~ion inert solvent and it is especially preferred lhat the solvent is 10 acelonitrile because of the excellent stereoselectivity it provides. Other solvents include rnethylene chloride, ethyl acetate and nitromethane.
It is preferred that ~he reaction is acid catalyzed as 1his can increase the selectivity of the 13~cellobioside product over the a-cellobioside anomeric produet.
Pre~erably rnineral acids are used. Hydrobromic acid has been shown to be particularly 15 effe~ive in increasing the 13-ce!lobioside product yield. Other pre~erred acids include hydrochloric, hydrofluoric and sulfuric acid. It is preferred that about 0.05 equivalents to about 2 equivalents, and especially preferred that about 0.1 equivalents to about 0.5 equivalents acid catalyst is used.
13-Tigogenin's preparation is described by Rubin in U.S. Patents Nos. 2,991,282 20 and 3,303,187, by B. Loken in U.S. Patent No. 3,935,194 and Cag1ioti et aJ., Tetrahedron 19, 1127 t~963). Its structure is depicted below.

' .
H
_H 3 ~.~lC H3 C H 3 - ~ /
~ J
H 0 h H

.

WO 93/11150 PCr/US92/0f~63X

212'~1i8~ ~-13-Hecogenin's preparation is described in a paper on Steroidal Sapogenins by Russell E. Marker et al., in J. Amer. Chem. Soc., 69, 2167-2211 (1947). Its structure is depictsd below.

- 3 ~ \C H 3 ~ ~H3 H H

~' 11-Keto-B-tigogenin switches the carbonyl group from th~ 12 position to the 11 position of the structure ~epicted above. 11~Keto-B-tigogerlin is prepared ~rom hecogenin by lhe following procedure. According to the proc0dure of Conforth, et al., ). Chem. Soc., 19~4, 907), hecogenin is acetylated, bromina~ed, treated with sodium hydroxi~e and reduced with zinc to give the 12-hydroxy-11-keto analog. Then 12-hyclroxy~ keto ana!og is acetylated and reduced with calcium and amrnonia ts give 1 1-ketotigogenin.
Diosgenin's:preparation is described in "Diosgenin and C)ther Steroidal Dr~g Precursors" by Asolkar, L.V.,~ Chadha, Y.R., and Rawat, P.S" Council of Sci~ntific and Industrlàl Research, New D~lhi, India, 1~83 pages, 1979 and also in T. Kawasaki et al., C~hem, Pharm. Bull., Japan~10 698 t1962). Its ~tructure is depicted below.

.
:

~ .

Wo 93t 1 1 15(1 PCl / US92/1)~63X

~123(i8~

~ H 3 J. ~ C H 3 HO R

Pr~ferably about 1 equivalent to about 2 equivalents of the steroid is used. It is esp~clally pref~rred that about 1 equivalent to about 1 5 equivalerlts of the steroid is . used.
Any envirorlment or conditions (e.g., lernperature, time, solvent) suitable for (i.e, capable of) forming the desir~d tigogenin, 11-ketotigogenin, hecogenin- or diosgenin-10 beta-O-cellobloside heptaalkanoate may ~e used However, it is preferred that the reaction occurs at a temperature of about 20C to about 100C and preferably ~omabout ~OQC t~ about 65C. Below:about 2ûC the reaction can b~ slow and above about 100~ undesir~d side rea~tions (e.g. anorneri~ation~ can occur This reaction conveniently ca~ried out at ambient pressure how ver, pressures from about 0.5 to 15 about 3 atmospheres may be~ used.
Pre~erably ths steroid, metal salt and solvent are heated to retlux and sufficient solvent is a~eotropically:distilled~:to remove substantially all the water. Then the ce!l~biosyl bromide heptaacetate is~ added to the above mixture and heated ~or about ~:0.5 to::about 6.0 hours,~ typioally under nitrogen The desired cornpounds are then 20 isola~ed by conventional methsds :: For example, the ~Iycosides may be precipitated from the erud2 ~iltered reaction mlxture (e.g. acetonitrile ~product solution) by the addition of about 25% to 75% water : ~ :

:

WO 93/11150 Pcr/l-ls92/ox63x 2~.2~j8 9i -6-and the remainder alcohol (e g. methanol). Precipitation of the product ~rom aqueous rnethanol/acetonitrile requires less processing than an extractive isolation, and provides a product of greater purity.
The steroidal peracyl glycosides rnay be deacetylated by conventional methods 5 such as treatment with triethylarnine in methanol, basic anion exchange resins or 5 sodiurn methoxide in rrethanol or methanol/THF solvents (e.g. Example 2 below). For exarnple~ the deacetylated pro~uct may be prepared by refluxing in methanol/THF using a non-cataiytic amount of sodium methoxide followed by conventional work-up. Theexcess methoxide is used to decompose the fluoro sugar, if any ~3-cellobiosyl fluoride heptaacetate is present, otherwise the deacetylation would be catalylic in sodium 10 methoxide. The tigogenyl-~-O-cellobioside or analogs are then isolated by conventional methods such as ~iltration.
Although the above process is designed to synthesize steroidal glycosides of the 13 con~iguration, the more therrnodynamically stable a-anorners are accessible by acid-catalyzed isomeriza~ion of the 13-glycosides. For example1 tigogenyl a-O-1~ cellobioside heptaalkanoate can be prepared from tigogenyl. 13-0-cellobiosidehept~alkanoate by heating the 13-glycoside in a methylene chloride solution containing hydrogen bromide.
The influence of r~action stoichiometry, lemperature, solvents, molecular sieves, and vestigial hydrogen bromide on the stereoselectivity and yield o~ tigogenyl 13-O;cello-20 bioside heptaac~etate (using the process o~ Example 1) are summarized in Table 1.

.

j , .. . WO93/11150 P~1`/~ ;92/0~63X
~123~

Table 1 Zinc Fluoride or Cyanide Activated Glycosi~ic Couplings with Tigogenin Equivalents Solvenl Sieves T~ remp. ~G~coside Activator Tiqoqenin ~leld 2.00 ZnF2~4.00~ 1.0 CH3CN No 2.5 hrs/65C 79%
1.50 ZnF2(3.00) 1.0 CH3CN No 2.5 hrs/65C 68%
0.50 ZnF~(1.00) 1.0 CH3CN No 3.0 hrs/65C 32%
O.S0 ZnF~(1.00) 1.0 CH3CN No 1.5 hrs/65C 30%
Hbr (0.38) 2.00 ZnF2(4.00) 1.0 CH2CI2 No 3.0 hrs/4~C 10%
25%~anome~
2.00 ZnF2(4.00) 1.0 Toluene 4A 20 hrs/65C 41%
2.00 ZnF2(4.00) 1.0 CH2CI2/CH3(:;N No 1.5 hrs/65C 64%
(2/1 3) 1.00 . ZnF2(2.00) 2.0 CH~CN No 1.0 hrs/65C 30%
Hbr(0.75) 0.50 ZnF2(0.~0j 1.0 CH3CN No 22 hrs/50C 38%
2.00 ZnF2(4.00) 1.0 CH3CN No 1.75 hrs/80C 76%
0.~0 ZnF2~0.50) 1.0 CH3CN No 1.75 hrs/B0C 53%
1.25 ZriF2(Z.2s) 1.0 CH~CN No 1.75 hrs/80C ~1%
2.00 Zn~CN)2(4.00) 1.0 CH3CN No 2.0 hrs/65C ~3%
2.10 ~n(C~N)2(5-60) ~ 1-0 ~H3CN No 3.0 hr~/65C 55%
1.50 2n(~N)2(4,00) 1-0 CH3CN No 2.5 hrs/6~C 4~%

The zinG fluoride-acthtatecl gly~oside c~upling was repeated with hecogenin and diosgenin in analogous processes to the B-tigogenin glycosidic coupling o~ Example 1 below. The results with these other sterols are summarized in Table 2.

~`

WO 93/1 1 150 PC'r/US92/OX63X

~ ~ 236 '~ ~ -8-Tabl2 2 Zinc Fluoride-Mediated Glycosidic Couplings with Hecogenin or Diosgenin Euuivalents Solvent Sieves TimerremP. Yleld Glv~ ~ Activator Sterol Gh,~e Cholesterol 2.00 2.25 (1.0) CH3CN No 2.25 hrs/65C 63%
Hecogenin 2.00 2,25 (1.0) CH3CN No 3.0 hrs/65C 72%
Diosgenin 2.00 2.25 (1.0) CH3CN No 2.5 hrs/65C 65%

This invention makes a significant advance in the ~ield of steroidal glycosides by providing efficient methods of preparing steroidal peracyl glycosides. The deacetylated end products are useful as antihyper~holesterolemic agents.
6 It should be understood that the invention is not limited to the particular embodiments shown and described herein, but that various changes and rnodifications may be made without departing from the spirit and scope of this novel concept asde~ined by the following claims.
Example_1 Tiqo~enyl 13-O-Cellobios~acetate To a dry ~lask equipped with a mechanical stirrer, thermometer, and distillation head were added 13-tigogbnin (4.1 6g; 0.01 mole), anhydrous zinc fluoride (4.1 3g; 0.04 mole), and 160 ml of dry acetonitriie. The slurry was hea~ed to reflux (8~C) and 9û ml of distillates was removed overhead while 60 ml of fresh, dry acetonitrile was added to the slurry. ~he mixture was cooled to 2~C and then a sarnp!e was removed ~or a KarlFisher determin~t3On (K.F.= 0.0~% H20).
.
a-Cellobiosyl bromide heptaacetate (14.00g; 0.02 mole) was added to the ~lask, and th~n the slirred slurry was heated to 65 C under a nitrogen atmosphere. The mixture vvas mairltained at 65~C for 2.5 hours when lthin-layer chrom~tography1 (tlc) showed that the reaction was complete. The reaction was oooled to ambient temperature and 130 mi of methyl ne chloride was add~d. The lhin slurry was filtered through Celite and the filtrate (300 ml) was washed with a sa~urated sodium bicarbonate solution (70 ml) Wo 93/1 l 150 PCT/US92/OX63X
2123~34 g followed by an aqueous wash (70 ml). After lhe organic layer was dried over anhydrous sodium sulfate (20 grams~ and filtered; the solution was then concentrated to 50 ml via a distillation at atmospheric pressure. Two hundred milliliters of 2B-ethanol was added to the warm concentrat~ and the turbid solution was concentrated to 5 approxirnately 50 ml. The thin slurry was cooled to 25C and then it was granulated for 90 minutes at room t~mperature. The crude product was filtered, the cake waswashed with 25 rnl of 2B-~thanol, and then ~ried at 40C in V~cuo for 17 hours to give 10.8 grams o~ white crystalline solids (m.p. = 22~-231 C).
The solids were dissolved in 25 ml of methylene chloride and then 75 rnl ot 2B-ethanol was added. The thin slurry was heated to reflux (760 mm) and 35.ml of distillate was removed overhead. The resulting slurry was cooled to room temperature and then was granulated ~or 90 minutes. The B-glycoside was filtered, and then dried at 40C in vacuo ~or 18 hours to give 9.65 grams of a while crystalline solid (m.p. =
229-234C). Thin-layer chromatography' and high pressure liquid chromatography2 ~hplc) show that the product contains 77% (w/w) tigogenyl 13-O-cellobioside heptaacet~te and 15% (w/w) a-cellobiosyl fluoride heptaacetate. The a-celiobiosyl fluoride heptaacetate is most easily removed trom the product during the deacetylation stepi Example 2 Tiaogenvl 13-0-cellobioside Crude tigogenyl R-O-cellobioside heptaacetate (50.0g; ~.048 moles~ was dissolvedin 250 ml of tetrahydro~uran and 250 ml of methanol while maintained under a nitrogen atmosphere. The hazy solution was ~iltered through a b~d of Ce,ite and then a solution of ~odium rnethoxide (0.46g; 0.008 moles3 in methanol ~10 ml) was added to the filtrate.
. 2~ The solution was hea~ed to reflux (60C) and mairltained at reflux for 1.25 hours generating a thick white slurry. A reaction aliquot was removed and analyzed by thin-I ay~r chrornatography which indicated that the reaction was complete. The slurry was con~entrated by removing 200 ml of distillate and then 200 ml of water was added to the refluxing slurry. Another 20û ml of distillate was removed, and additional water (200 ~0 ml) was added. The slurry was cooled to ambient ternperature and filtered. The product cake was washed with water (50 ml) and then pulled dry on the fiiter. The wat~r-wet~cake was heated to reflux (65C:: in 600 mls of THF and 92 mls of waler).
DARC0 G-60 (1.53 grams~ was added to the solutionl stirred for 15 minutes, and then WO 93/11150 Pcr/~ls92/o~63x 2 1 2 3 t; ~ 4 -10-the mixture was iiltered through Celite The solution was concen~rated by removing 460 ml of distillates and 460 ml of methanol was then added. The methanol addition and concentration sequence was repeated twice again removing an addition 800 mls o~
distillate and 800 mls of ~resh methanol was added. The resultiny slurry was cooled to 20C and then granulated ~or one hour. The product was filtered, rinse with f!esh methanol (50 ml), and then the wet cake was reslurried in 300 mls of fresh methanol (24C). The product was filtered and then dried at 40C in vaCuo overnight. Tigogenyl 13-O-cellobioside (24.4g; 0.036 moles) was isolated in 74% overall yield. Spectral and physical properties were identical to an authentic sample.
Example 3 a-D-Cellobiosvl 8romide Heptaacetate a-D-Cellobiosyl bromide heptaacetate was prepared ~orm a-D-cellobiose octaacatate and hydrogen bromide in glacial acetic acid using a modified procedure of Freudenberg a~d Nagari1.
A 20% (w/w) h~drogen bromide solution (178.7g; 0.44 mole of HBr) in glacial acetic acid was prepared by bubbling gaseous hydrogen bromide into glacial acetic acid until a derlsity of 1.212 was obtained. In a separate dry reactor maintained under a nitrogen atmosphere, o-D-cellobiose octaacetate (50.0g; 0.074 moles) was dissolved in 408 ml of rnethylene chioride. The HBr/HOAC solution was added to the disaccharide sol~ltion to give a yellow solution. Afterthe solution was stirred for ~ hours at ~ 17-255, a small aliquot o~ solution was removed for a reaction completion assay. Once thin-layer chromatography2 indicated th~t the reaction was complete, the solution was cooled to 10C and 0.5 liters o~water was added. The mixture was stirred ~or 10 rninutes, the stirring was stopped, z~nd th~ iayers~ were allowed ts separate. The rnethylene chloride layer was decanted and then washed with 7.50/D W/W sodium bicarbonate solution (0.5 liters~ ~ollowed by water io.5 liters). Finally, the methylene chloride sol~tion was dried over 8 grams of anhydrous magnesium sulfate and then ~iltered. The MgSO4~hydrate ~: :

:: : . _ K. Freudenberg and W. Nag~arij Ann., 494, 63 (1932).
2Merck Pre-Coated TLC Silica Gel 60F-264 Plates using a toluene/acetic acid (4:1 ) eluant.
Plates were sprayed with 10% (w/w) H2SO4 in water and heated ~or charring, after the plates were cleveloped.

.

.~ Wo 93/11150 PC~/lJS92/0863~
2123~

cake was washed with fifty milliliters of fresh methylene chloride, and the filtrate and wash were combined. The methylene chloride solution was concentrated to approxi-mately 0.15 liters by an atmospheric distillation and then cooled to ambient temperature. Diisopropyl ether (0.6 liters) was slowly added over 15 minutes with stirring to generate a thick slurry. The product was granulated for 1 hour at 25C, filtered, and then dried in v~cuo at 40C for 4.5 hours. a-Cellobiosyl bromide heptaacetate (47.6g; 92% yield) was obtained as a white crystalline solid (m.p. = 192-194C) whose 'H NMR spectrum (Cl:)C13) was consistent with its structure.
Example 4 1 O 11 -Ketotiqoqenvl-~-O-(:~ellobioside Heptaacetate To an appropriately equipped one liter, 3-necked round bottom flask were added acetonitrile (305 mls), 11-ketotigogenin (5.00g; 0.011 moles), and rhombohedral,crystalline zinc fluoride (1,65g; 0.016 rnoles). The slurry was heated to reflux (80C) and then 100 ml of distillate was removed overhead. The slurry was coolecl to room temperature, and then 15.39 grams (0.022 moles) of a-ce!lobiosyl bromide heptaacetate was added. The reaction mixture was reheated to ~0-65~ and then maintain~d at 60-, ~ , 65C: for 2 hours. A reaction sample was removed ~or a reaction completion assay.
Thin-layer chromatography assay (EtOAc/hexanes 1.5:1 ) showed the complete disappearanc0 ot the glycosyl bromide starting material so the reaction was cooled to 25C and 1~2 ml of ~methylene chloride was added. After stirring for 10 minutes~ the mixture was tiltered through~ Celite and the filter cake was washed with 25 ml of CH2CI2.
Th~ oombine~ r~action ~lltrate and wash were washed with water ~81 mls), saturated sodium bi~arbonate solution~(75 mls),~and water (137 ml). The organic layer was ~inally dried over 11 grams of anhydrous magnesiurn sulfate. The MgSO4 was tiltered and 26 ~ washed with 16 mls~of ~fresh ÇH2CI2. The filtra~e ~nd wash were combined and then con~entrated at r~duced pressure to one fourth its original volume (300 mls~. 2B-khanoi (250 ml) was added and the resulting solution was concentra~ed to one-half volume (170~mls). The slurry was cooled to 20-25C and then granulated ~or 1 hour.
`The whit~ waxy solids wer~ ~lltered, washed with ~resh 2B-ethanol (50 mls), and then dried in vacuo at 40~:: overnight. 11-ketotigogenyl-B-O-cellobioside heptaacetate ~9.7 grams~ m.p. = 205-219C) was iso!ated in 84% overall yield. Chromatographic and spectral ~haracterization were identical to an authentic sample ot 11 -ketotigogenyl-~-O-cellobioside heptaacetate. Crude 11-ketotigogenyl-B-O-cellobioside could also be :

:

WO 93/11150 P~/l)S92/08638 21 ~ 3 t1~ ~

isolated in 64% yield by the ~ollowing aqueous isolation sequence: The crude reaction mixture was dilutecl with additional fresh acetonitrile (50 ml) and then ~iltered through Celite. Methanol (290 mls) was added to the filtrate and the resulting solution was heated at reflux (65C) lor one howr. One hundred millili~ers of deionized water was 5 slowly added to the refluxing solution to give a ha~y mixture. Alter 20 minutes at rellux (72C), the mixture was slowly cooled to room temperature and then granulated at 23-25C for 1 hour. The crude product was ~iltered and washed with water. The filter cake was suspended in 2B-ethanol (75 ml) and the rnixture was heated to re~lux. The mixture was then cooled to ambient temperature, liltered, and the solids were dried in 10 vacuo at 40C overnight. 1 1 -Ketotigogenyl-e-O cellobioside heptaacetate was isolated in 64% overall yield.
Examp!e 5 1 1-Ketotiqoqer~YI~
11-Ketotigogenyl-B-O-cellobiosid~ heptaacetate (9.7g; 9.2 mmoles) was suspended 15 in 50 mls o~ methanol and SO rnls of tetrahydrofuran. The system was purged with nitrogen and then a solution o~ sodium rnethoxide (O.lOg; 1.9 mmoles~ in rnethanol (1 ml) was added. The solution was heatcd to reflux (61 C) and then maintained at reflu tor 1 hour. Thin-layer chromatography (CH2CI2/methanol 4:1) of the resulting slurry showed that the reaction was complete. The tetrahydroluran was removed ~rom the 20 reaction by an atmospheric distillation and evsntually displacement with methanol (220 mls). A total of 180 ml o~ distillate was collected. The slurry was cooled to 20-25C
and then granulated overnight. The product was ffltered, wash~d with methanol (2 x 20 mls) and then the wet cake was reslurried (2C hours) in 100 ml o~ deionized water.
Alter fllltratio~ and drying in vacuo at 25~ overnight, 4.7 grams of 11^ketoti~ogenyl~
25 O-cellobioside was isolated in 65.5% overall yield. The product was homogenous by tlc and analytical characterizations was consistent wnh the product's structure.1~ ExamPlQ6 Ln Situ Hydrobromic Acld Generation e1-Cellobiosyl bromid~ heptaacetate (5.00g; 7.15 mmole) and 50 ml ol acetonitrile 30 were added to a 75 ml round bottom flask which was equipped with a mechanicalstirrer, thermometer, and vacuum distill~tion head. The system was purged wilh nitrogen and then the pressure was reduced to approximately 425 mm Hg. The solution was heated to 56-58C:; and approxim~tely 13 ml o~ distillate was removed WO 93~ 0 PCI /US92/û863X
~123~

overhead. The solution was cooled to room temperature and the vacuum was slowly released. Methylene chloride (37 rnl) was added to the glycosyl bromide solution and then the solution was extracted with water (2 x 25 ml). The aqueous phases were combined and then titrated to a phenolphthalein endpoint using 0.100~ N NaOH
sohJtion.
Tho millequivalents of HBr acid contained in the a-cellobiosyl bromide heptaacetate solution before and after the ~eotropic ~trip are reported below. The azeotropic strip increas~ the tetr~table acid approximately 8-10 fold depending upon the water content.

TITRATED ACIDS
Initial a-Cellobiosyl Bromide 2.1 milliequivalents Heptaacetate Solution , , _ _ . . .~
Azeotroped a-Cello~iosyl ~romide 16.9 milliequivalents Heptaa etate Solution _ -_ Wh~n th~ azeotropic distillation was used in Example 1 to dry the glycosyl brornide solution and to increase its acid content, the reaction time was decreased to 1.0 hour at 65C. In addition, high yields and high 13-anomeric selectivity were maintained.
Fxample 7 Aqueous Isolation of Tiqoqenyl ,, C~
A zinc ~luoride-mediated glycosidic coupling of B-tigogenin (0.915 mole~ with a-cellobiosyl bromid~ heptaacetate (1.830 mole3 in ac~tonitrile was conducted according to the procedure of Example 1. Once the reaction was complete, the crude reac~ion mixture was worked-up by an aqueous method to give high-quality tigoyenyl 13-O-cellobiosicle heptaacetate by the following sequence.
The crude reaction mixture was filtered through Celite to afford approxirnately 20 liters of a golden colored ~iltrate. The filtrate was heated (55-60~C~ and concentrated at reduced pressure to about 10 liters. The concentrated solution was cooled to 50C
and 6.0 Iiters of methanol was added. Subsequently, 7.5 liters of deionized water was slowly added over 30 minutes. A solid precipitated from solution once about 2 liters of water was charged. The mixture was heated to reflux ~73C) and then maintained WO g3/ 111 50 PCr/ U S9 2/O~s~ ? ~

2 1 2 3 ~ 4 at reflux for 2 hours. The slurry was cooled to 25~C and granulated overnight. The crude product was filtered, washed with methanol (2 x 1.5 liters), and then dried in v~cuo at 40C. The crude solid (1.01 kg) was 9~.5% pure by a hplc assay. In addition, the crude product contained only 1.3% of tigogenyl a-O-cellobioside heptaacetate and no B-cellobiosyl fluoride heptaacetate.
Crude tigogenyl 13-O-cellobioside heptaacetate (1.0 kg) was slurried in 10.2 liters of 2B ethanol ùnder nitrogen and lhen the mixture was heated to reflux (78C). Af~er the slurry was at re~lux tor 1.5 hours, the mixture was cool~d to 25C and then granulated for 12 hours. The product was filtered, washed with tresh ethanol (2 x 300 ml), and finally dried in vacuo at 4~ C overnight. Tigogenyl (3-O-cellobioside heptaacetate (0.89 kg) was isolated in 74% overall yield frorn tigogenin. The white orystalline product was 98.90io pure by high pressure liquid chromatography and contained only 0.5% (w/w) o~ the isomeric a-anomer~
.

~ ~ : Y

,

Claims (12)

-15-

1. A process for the synthesis of tigogenin-, 11-ketotigogenin-, hecogenin- or diosgenin-.beta.-O-cellobioside heptaalkanoate comprising:
reacting .alpha.-cellobiosyl bromide heptaalkanoate wherein R is C1-C4 and a steroid selected from the group consisting of .beta.-tigogenin, 11-keto-.beta.-tigogenin, .beta.-hecogenin and .beta.-diosgenin, in the presence of zinc fluoride or zinc cyanide under conditions capable of forming said tigogenin-, 11-ketotigogenin-, hecogenin- or diosgenin-.beta.-O-cellobioside heptaalkanoate.

2. The process as recited in claim 1 wherein the metal salt is zinc fluoride, R is methyl and the reaction occurs in a non-protic reaction inert solvent.

3. The process as recited in claim 2 wherein the steroid is .beta.-tigogenin or 11-keto-.beta.-tigogenin.

4. The process as recited in claim 3 wherein said reaction is acid catalyzed.

5. The process as recited in claim 4 wherein the acid catalyst is hydrobromic orhydrofluoric acid.

6. The process as recited in claim 5 wherein the reaction occurs in the additional presence of zinc bromide, zinc chloride, zinc iodide, zinc hydroxy fluoride, zinc oxide, zinc carbonate, zinc hydroxide, a trialkyl tertiary amine, a tetraalkyl urea or a N,N-dialkylaniline.

7. The process as recited in claim 6 wherein said reaction occurs at about 20°C
to about 100°C, about 1 to about 2 equivalents 11-keto-.beta.-tigogenin is used, about 0.5 to about 4 equivalents zinc fluoride is used, and about 0.5 to about 3 equivalents a-cellobiosyl bromide heptaalkanoate is used.

8. The process as recited in claim 7 wherein the solvent is acetonitrile and about 0.05 to about 2 equivalents of hydrobromic acid is used.

9. The process as recited in claim 2 wherein the 1-0-steroidal peracyl-.beta.-glycosides are precipitated from acetonitrile by the addition of about 25% to 75% water and the remainder alcohol.

10. The process as recited in claim 2 additionally comprising the step of deacetylating the tigogenyl .beta.-O-Cellobioside heptaalkanoate to form tigogenyl .beta.-O-cellobioside.

11. The process as recited in claim 10 wherein the deacetylation occurs by treatment with sodium methoxide in methanol.

12. The process as recited in claim 2 wherein the steroid is 11-ketotigogenin.