Model Steroid Glycoside Synthesis via a Glycosyl Transfer Mediated by Heterocycloaddition

Full text of DOI:10.1021/jo970287r

4510
J . Org. Chem. 1997, 62, 4510-4512
Sch em e 2a
Mod el Ster oid Glycosid e Syn th esis via a
Glycosyl Tr a n sfer Med ia ted by
Heter ocycloa d d ition
War War Win* and Richard W. Franck*
Department of Chemistry, Hunter College/ CUNY,
695 Park Avenue, New York, New York 10021
Received February 14, 1997
Steroid glycosides have been a subject of study for the
synthetic chemist for many years.1 The construction of
the glycoside-steroid linkage has been successfully ac-
complished by several groups. Every recorded synthesis
used the most common method of glycoside bond forma-
tion, namely electrophilic activation of the anomeric
carbon of the carbohydrate followed by its trapping with
a steroidal alcohol. Our laboratory has been exploring
an alternate approach to glycosyl transfer,2 shown by the
general cycloaddition in Scheme 1.
a
Key: (a) oxidation: Swern 72%, TPAP 87%, J ones 97%; (b)
Sch em e 1
bromination: Br2-HBr 62%, CuBr2 92% (2R/2â 95:5); (c) CaCO3-
DMF, reflux 30 min 71% + 20% ∆-4,5; (d) H2O2/NaOH 16%, Triton
B-TBHP 53%; (e) LAH, ether, reflux 99%; (f) J ones 81%.
Scheme 3 illustrates the cycloaddition chemistry.
Thus, steroid dione 4 was cleanly phthalimidosulfenyl-
ated to afford sulfenyl dione 8, the ultimate precursor to
our heterodiene 3, which was generated and trapped in
situ in the presence of 0.5 equiv of tribenzyl glucal 9. Two
cycloadducts were obtained in identical yield and were
separable by flash chromatography on silica gel. That
both adducts had arisen from heterodiene attack on the
bottom or R face of the glucal was clear from proton NMR
data. In particular, the axial H at C-2′ (δ 3.2), adjacent
to sulfur, is quite diagnostic in all our cycloadducts. In
order to differentiate between the two adducts, an
We were interested to discover if our concept could be
applied to the glycosidation of steroids. Our planning
was based on our earlier demonstration that the struc-
tural element required for a successful heterocycloaddi-
tion was a 1,3-dioxo-2-thiono assembly such 1, which is
symmetrical, and 2, where the carbonyls were differenti-
ated electronically. Thus, our proposed steroid hetero-
diene 3 led to an interesting question: would there be
selectivity between the carbonyls at C-1 and C-3 in the
cycloaddition?
undecoupled
13C NMR spectrum of each material was
used. Thus, the product assigned structure 10 exhibited
a triplet for the resonance at δ 162.1 assigned to the vinyl
carbon at C-3, whereas the material assigned structure
11 revealed a distorted triplet for the CdO at δ 194.7.
Desulfurization of 10 and 11 with Raney nickel in
benzene afforded the novel steroid glycosides 12 and 13
in 57% and 21% yield, respectively.
Discu ssion
Resu lts
The absence of regioselectivity in the cycloaddition step
of diene 3 was surprising at first glance since the 2-D
representation implies that congestion might exist near
the C-1 carbonyl. However, a simple calculation of the
ground-state geometry of heterodiene 3 corrected our
naive first impression. Figure 1 shows that, indeed, the
C-1 carbonyl carbon is somewhat blocked by the angular
methyl. But, the A ring is quite twisted, presumably to
avoid eclipsing of the carbonyls and thiocarbonyl, so that
the oxygen at C-1, which is the terminus of the diene, is
not near any group that might have provided some steric
hindrance to approach of the tribenzyl glucal. Our
cycloaddition method is novel, and it does provide a
2-deoxyglycoside without electrophilic glycosyl transfer
The first task was to prepare cholestane-1,3-dione (4),
which was to be the precursor to heterodiene 3. Dione 4
had been reported by Tamm,3 using the sequence shown
in Scheme 2. We repeated the work with one significant
change. In our hands, the Grieco modification of the
Yang method using t-BuOOH for the epoxidation of the
enone 6 produced the desired 7 in 53% yield, compared
to 16% yield in our hands, when alkaline hydrogen
peroxide was the reagent.4
(1) Randolph, J . T.; Danishefsky, S. J . J . Am. Chem. Soc. 1995, 117,
5693-5700. This recent paper has a good bibliography.
(2) (a) Capozzi, G.; Franck, R. W.; Menichetti, S.; Mattioli, M.;
Nativi, C.; Valle, G. J . Org. Chem. 1995, 60, 6416-6426. (b) Capozzi,
G.; Dios, A.; Franck, R. W.; Geer, A.; Marzabadi, C.; Menichetti, S.;
Nativi, C.; Tamarez, M. Angew. Chem., Int. Ed. Engl. 1996, 35, 777-
779.
(4) (a) Grieco, P. A.; Nishizawa, M.; Oguri, T.; Burke, S. D.;
Marinovic, N. J . Am. Chem. Soc. 1977, 99, 5773-5780. (b) Yang, N.
C.; Finnegan, R. A. J . Am. Chem. Soc. 1958, 80, 5845-5848.
(3) Tamm, C.; Albrecht, R. Helv. Chim. Acta 1960, 43, 768-782
S0022-3263(97)00287-9 CCC: $14.00 © 1997 American Chemical Society

Notes
J . Org. Chem., Vol. 62, No. 13, 1997 4511
Sch em e 3a
F igu r e 1. Chem3D+ representation of an AM1 calculation
(Spartan software) of the A and B rings of 3.
1H NMR (CDCl3, 300 MHz) δ 7.5-8.7 (m, 8H); 13C NMR (CDCl3,
75 MHz) δ 124.3, 132.2, 134.9, 166.5.
To a solution of N,N′-dithiobis(phthalimide) (3.4 g, 9.6 mmol)
in 150 mL of HPLC-grade CH2Cl2 were added dry pyridine (0.2
mL) and sulfuryl chloride (8.3 g, 5.0 mL, 62 mmol) via a dropping
funnel at rt. The yellow reaction mixture was stirred for 2 days,
and all volatile materials were removed under vacuum, affording
the title compound in 90% yield (3.7 g): 1H NMR (CDCl3, 300
MHz) δ 7.8-8.0 (m, 4H); 13C NMR (CDCl3, 75 MHz) δ 124.8,
131.6, 135.7, 165.3.
1,3-Dioxo-2-(th iop h th a lim id o)ch olesta n e (8). A solution
of phthalimidosulfenyl chloride (144 mg, 0.7 mmol) in dry THF
(5 mL) was cooled to -60 °C. Into it was added a solution of
cholestane-1,3-dione (4) (200 mg, 0.5 mmol; see Supporting
Information for details of its preparation) in dry THF (5 mL)
dropwise. After the white suspension was stirred for 45 min at
-60 °C, the reaction mixture was allowed to warm to room
temperature and then filtered. The filtrate was evaporated
under vacuum to give the title compound in 62% yield (180
mg): 1H NMR (CDCl3, 300 MHz) δ 0.67 (s, 3H,), 0.87 (d, 6H, J
) 6.6), 0.89 (m, 6H,), 7.76 (m, 4H), 11.00 (s, 1H).
a
Key: (a) benzene-d6, 2,6-lutidine, rt 21 h; (b) Raney nickel,
toluene.
Regioisom er ic Cyclic Oxa th iin es (10 a n d 11). Into a
suspension of 1,3-dioxo-2-(thiophthalimido)cholestane (8) (335
mg, 0.6 mmol) and tri-O-benzylglucal (9) (125 mg, 0.3 mmol) in
benzene-d6 (5 mL) was added 0.3 mL of 2,6-lutidine at rt under
Ar. After being stirred for 21 h, the reaction mixture was diluted
with CH2Cl2, washed first with aqueous saturated sodium
bicarbonate solution and second with brine, dried over Na2SO4,
and evaporated in vacuo to give 480 mg of crude product. The
components were separated by column chromatography (silica
gel, petroleum ether -10% ethyl acetate) to afford regioisomeric
cyclic oxathiines 10 (Rf ) 0.76) and 11(Rf ) 0.24) each in 33%
yield (80 mg). Data for 10: [R]25 ) +119.5 (c ) 8.5, CHCl3); 1H
NMR (CDCl3, 300 MHz) δ 0.65 (s, 3H), 0.84 (d, 6H, J ) 6.6),
0.89 (d, 3H, J ) 6.5), 1.05 (s, 3H), 3.20 (dd, 1H, J ) 10.7, 2.9),
3.57-3.96 (m, 3H), 4.52-4.83 (m, 8H), 5.63 (d, 1H, J ) 2.8),
7.33 (m, 15H); 13C NMR (CDCl3, 75 MHz) δ 11.6, 12.6, 18.8, 22.8,
23.0, 23.7, 24.1, 24.5, 28.1, 28.2, 28.3, 29.9, 30.7, 34.0, 36.1, 36.4,
37.0, 39.8, 40.3, 41.5, 42.1, 42.7, 47.1, 48.2, 56.6, 68.3, 73.4, 73.8,
75.7, 78.5, 96.5, 102.8, 128.3, 128.4, 128.5, 128.7, 138.0, 162.1,
202.0. Anal. Calcd for C48H70 O6S: C 76.56, H 8.33, S 3.79.
Found: C 76.22, H 8.55, S 3.72. Data for 11: [R]25 ) +112.4 (c
) 9.25, CHCl3); 1H NMR (CDCl3, 300 MHz) δ 0.67 (s, 3H), 0.86
(d, 6H, J ) 6.6), 1.15 (s, 3H), 3.25 (dd, 1H, J ) 10.8, 2.8), 3.71-
3.90 (m, 3H), 4.56-4.92 (m, 8H), 5.57 (d, 1H, J ) 2.8, 1-H), 7.15-
7.20 (m, 15H); 13C NMR (CDCl3, 75 MHz) δ 12.1, 12.2, 19.0, 22.5,
22.6, 23.5, 24.2, 25.5, 27.6, 27.9, 29.7, 30.1, 35.5, 35.7, 37.2, 39.8,
40.4, 40.6, 42.1, 42.3, 43.4, 44.9, 50.9, 56.2, 56.4, 68.2, 74.4, 74.6,
75.8, 76.2, 78.8, 100.5, 128.1, 128.2, 128.3, 128.4, 137.8, 137.9,
175.0, 194.7.
chemistry. However, it is not a practical alternative to
the classical methods for steroid glycoside synthesis
because of the several steps needed to prepare the
steroidal heterodiene precursor.
Exp er im en ta l Section
Pulsed Fourier transform 300 MHz 1H and 75 MHz 13C
spectra were obtained with deuterated chloroform (99.8%, 0.03%
v/v TMS, Sigma-Aldrich). Chemical shifts are in δ or ppm units
downfield from Me4Si as internal reference. Coupling constants
are reported in hertz (J values). The assignment of the 13C was
confirmed by single-frequency off-resonance decoupled and
proton-coupled spectra. TLC analyses were done on silica gel
60 F254 plates available from EM Science and visualized by
dipping them in a cerium sulfate or polymolybdic acid solution.
All regular and flash column chromatography separations were
performed using 230-400 mesh, 60 Å silica gel, available from
EM Science. Optical rotations were recorded on an automatic
polarimeter using a 1 dm cell at the reported temperatures and
concentrations.
P h th a lim id osu lfen yl Ch lor id e. Into an ice-chilled suspen-
sion of potassium phthalimide (73 g, 0.4 mol) in dry CH2Cl2 (400
mL) was added a solution of sulfur monochloride (26 g, 0.2 mol,
15.4 mL) in dry CH2Cl2 (15 mL) via a dropping funnel over a
period of 10 min. The reaction mixture was stirred mechani-
cally. After being stirred for 15 min at 0 °C, it was allowed to
warm to rt for 18 h. The insoluble material was filtered. After
the solvent was removed under reduced pressure, 22.8 g (33%,
mp 235-236 °C) of N,N′-dithiobis(phthalimide) was obtained:
3-(3′,4′,6′-Tr iben zyl-2′-d eoxy-r-glu cop yr a n osyl)ch olest-
2-en -1-on e (12). A teaspoonful of Raney nickel was washed first
with absolute ethanol (4 × 5 mL) and then with toluene (4 × 5

4512 J . Org. Chem., Vol. 62, No. 13, 1997
Notes
mL). Into an ice-cooled suspension of Raney nickel in toluene
(1 mL) was added cyclic oxathiine 18 (40 mg, 0.05 mmol) in
HPLC-grade benzene (1 mL) under Ar. After being stirred at 0
°C for 2 h, the mixture was allowed to stir at rt for 29 h. The
suspension was filtered through a Celite pad, which was then
washed with ethyl acetate, and the filtrates were evaporated in
vacuo and purified by flash chromatography over silica gel,
eluting with petroleum ether -20% ethyl acetate, to give 23 mg
(57% yield) of glycoside 12: 1H NMR (CDCl3, 300 MHz) δ 0.67
(s, 3H), 0.87 (d, 6H, J ) 6.6), 0.89 (d, 3H, J ) 6.5), 1.00 (s, 3H),
2.33 (m, 2H), 2.52 (m, 1H), 3.63-3.98 (m, 3H), 4.47 (d, 2H, J )
12.3), 4.55 (d, 2H, J ) 10.5), 4.67 (d, 2H, J ) 10.5), 4.90 (d, 2H,
J ) 10.8), 5.48 (s, 1H), 5.54 (m, 1H, 1-H), 7.33 (m, 15H); 13C
NMR (CDCl3, 75 MHz) d 11.4, 12.6, 18.8, 22.8, 23.0, 23.8, 24.1,
24.5, 28.2, 28.3, 30.9, 33.3, 35.1, 36.1, 36.4, 37.0, 39.7, 40.4, 41.6,
42.7, 46.2, 48.1, 56.6, 68.5, 72.2, 72.5, 73.8, 75.3, 95.8, 104.7,
127.8, 127.9, 128.2, 128.6, 138.1, 138.5, 138.6, 170.2, 206.4.
Anal. Calcd for C54H72O6: C, 79.37; H, 8.88. Found: C, 79.22;
H, 8.72.
(21% yield) of the title compound: 1H NMR (CDCl3, 300 MHz)
δ 0.65, (s, 3H), 0.86 (d, 6H, J ) 6.6), 0.88 (d, 3H, J ) 6.5), 1.14
(s, 3H), 3.60-3.96 (m, 4H), 4.47 (d, 2H, J ) 12.3), 4.55 (d, 2H,
J ) 10.8), 4.88 (d, 2H, J ) 10.5), 5.56 (m, 1H), 5.58 (s, 1H), 7.38
(m, 15H); 13C NMR (CDCl3, 75 MHz) δ 11.9, 12.5, 12.7, 14.3,
18.8, 22.8, 23.0, 24.0, 24.4, 24.5, 26.0, 27.8, 28.1, 28.2, 28.3, 29.9,
31.0, 31.2, 35.2, 35.4, 35.98, 36.0, 36.3, 36.4, 37.7, 39.71, 39.74,
40.1, 40.7, 41.2, 42.7, 43.0, 44.1, 46.5, 46.8, 49.9, 56.5, 56.7, 57.2,
68.5, 70.3, 72.4, 73.2, 73.8, 75.5, 77.5, 78.4, 96.2, 105.0, 127.5,
127.9, 128.0, 128.1, 128.3, 128.6, 138.2, 138.4, 138.7, 184.7, 199.4.
Ack n ow led gm en t. The authors are indebted to
Professor G. Capozzi (University of Florence) for sharing
the results of his parallel heterocycloaddition project.
This research was supported by NIH Grants Nos.
GM51216 and RR03037 (RCMI). The AM1 calculation
was done on an IBM RISC computer acquired through
HEAT funding from the State of New York.
1-(3′,4′,6′-Tr iben zyl-2′-d eoxy-r-glu cop yr a n osyl)ch olest-
2-en -3-on e (13). A teaspoonful of Raney nickel was washed first
with absolute ethanol (4 × 5 mL) and then with toluene (4 × 5
mL). Into an ice-cooled suspension of Raney nickel in toluene
(1 mL) was added cyclic oxathiine 19 (70 mg, 0.08 mmol) in
HPLC-grade benzene (1.5 mL) under Ar. After being stirred at
0 °C for 2 h, the mixture was allowed to stir at rt for 18 h. The
suspension was filtered through a Celite pad, which was washed
repeatedly with ethyl acetate, and the filtrate was evaporated
in vacuo and purified by flash chromatography over silica gel,
eluting with petroleum ether -20% ethyl acetate, to give 14 mg
Su p p or tin g In for m a tion Ava ila ble: The complete ex-
perimental procedures for the preparation of steroid materials
4, 6, and 7 plus the intermediates not shown in Scheme 2 along
1
with copies of 13C and H NMR spectra for every compound in
the schemes (28 pages). This material is contained in libraries
on microfiche, immediately follows this article in the microfilm
version of the journal, and can be ordered from the ACS; see
any current masthead page for ordering information.
J O970287R