Epimerization of C-22 in (25R)- and (25S)-sapogenins

Article abstract of DOI:10.1016/j.steroids.2014.10.004

Most of the naturally occurring steroidal sapogenins (C-23 non-substituted frameworks), possess an R configuration at the spiro C-22 center. Their C-22 epimers have become important targets in biological research. This paper describes a procedure to obtain 22S-spirostans from 22R-sapogenins and pseudosapogenin skeletons, without affecting the chirality at either C-25 or C-20. An optimal way to synthesize the pair of C-22 stereoisomers of 23-acetyldiosgenin is also reported. The latter was obtained from a 22,26-epoxycholestane or from 23-acetylfurostene compounds.

Full text of DOI:10.1016/j.steroids.2014.10.004

Steroids 93 (2015) 60–67
Contents lists available at ScienceDirect
Steroids
journal homepage: www.elsevier.com/locate/steroids
Epimerization of C-22 in (25R)- and (25S)-sapogenins
Omar Viñas-Bravo a, Penélope Merino-Montiel b, Anabel Romero-López b, Sara Montiel-Smith b,
,
⇑
Socorro Meza-Reyes b, Francisco J. Meléndez b, Jesús Sandoval-Ramírez b,
⇑
a Instituto de Química Aplicada, Universidad del Papaloapan, Circuito Central # 200, Colonia Parque Industrial, Tuxtepec, Oax. 68301, Mexico
b Facultad de Ciencias Químicas, Benemérita Universidad Autónoma de Puebla, Ciudad Universitaria, Puebla, Pue. 72570, Mexico
a r t i c l e i n f o
a b s t r a c t
Article history:
Most of the naturally occurring steroidal sapogenins (C-23 non-substituted frameworks), possess an R
configuration at the spiro C-22 center. Their C-22 epimers have become important targets in biological
research. This paper describes a procedure to obtain 22S-spirostans from 22R-sapogenins and pseudos-
apogenin skeletons, without affecting the chirality at either C-25 or C-20. An optimal way to synthesize
the pair of C-22 stereoisomers of 23-acetyldiosgenin is also reported. The latter was obtained from a
22,26-epoxycholestane or from 23-acetylfurostene compounds.
Received 18 February 2014
Received in revised form 1 October 2014
Accepted 14 October 2014
Available online 20 November 2014
Keywords:
Ó 2014 Elsevier Inc. All rights reserved.
Boron trifluoride etherate
22-epi-Sapogenins
Pseudosapogenins
23-Acetyldiosgenin
1. Introduction
temperatures (steam bath), and the limited characterization (mp
and microanalysis) for the reported 22-epi-sarsasapogenin product
Large amounts of steroidal saponins are extracted from natural
sources containing a spirostane aglycone and these are widely
employed in traditional medicine [1]. Variation in the stereochem-
istry of the spiroketal C-22 carbon atom is relatively rare in sapo-
nins, which usually differ only in the configuration of C-25. In most
cases the stereochemistry of spirostans at C-22 is R (in C-23 non-
substituted frameworks) [2]. Kudo et al. [3] provided the first reli-
able characterization of 22S sapogenins as minor compounds
accompanying the natural spirostanol glycosides. Other examples
of 22-epimeric saponins have been isolated as 1 [4], 2 [5], 3 [6],
4 [7] and 5 [8] (Fig. 1) from different natural sources and exhibit
important biological activities such as inhibitory activity against
cAMP phosphodiesterase, and HeLa cells growth and cytotoxicity
against six human cancer cell lines (HepG2, Hep3B, A549, Ca9-
22, MDA-MB-231, and MCF-7).
In nature, the occurrence of steroidal sapogenins (non-substi-
tuted at C-23) with an S configuration at C-22 is limited. However,
their significant biological activities and the scarcity of methods for
their preparation have motivated a search for improved synthetic
routes to these 22-epimeric spirostans. The first isomerization of
the spiroketal carbon was reported by Marker [9] in 1939 for sar-
sasapogenin (6, Fig. 2). His methodology involved strong acidic
conditions (aq HCl/EtOH), a long reaction time (4 days), and high
(7, Fig. 2) may not exclude other chemical changes, such as the
configuration at C-20 or C-25. Different epimerization was
observed by Wall [10a] who reported that treatment of 6 in the
same acid medium for 24 h at room temperature gave 20R-sarsas-
apogenin [10b]. During the past seventy years, this isomerization
methodology has been employed with some minor modifications
[11].
Recently, a procedure was reported for the asymmetric intra-
molecular spiroketalization of alcohol-enol frameworks, catalyzed
by chiral BINOL confined phosphoric acids (see Fig. 3) to give high
enantiomeric or diastereoisomeric ratios of (5,7)-, (6,6)-, (5,6)-, and
(5,5)-spiroketals [12]. Those designed chiral BINOL Brønsted acids
are sterically hindered and contain bifunctional sites producing a
chiral microenvironment. When any of those enantiomeric chiral
Brønsted acids were added to the enol ether present in pseud-
odiosgenin (8, Fig. 3), a mixture of 4 diastereoisomers (20R,22R;
20R,22S; 20S,22R, and 20S,22S) of diosgenin (9) was obtained in
a 20:1:1:1 diastereomeric ratio (88–89% transformation yield).
Among those stereoisomers, the non-natural 20R,22R diastereo-
mer (20-epi-diosgenin) was the major compound; this probably
results from the inability of the confined BINOL acid to efficiently
accommodate such large substrates [12]. This result is important
because it shows that some reagents can epimerize the chiral cen-
ter at C-20 more easily than at C-22; the 22S epimer of diosgenin
10 (the 20S, 22S compound) was obtained in a very low proportion.
In this paper we report a simple, efficient method to obtain 22-
epimeric spirostanic sapogenins from acetates of sarsasapogenin
⇑
Corresponding authors. Tel.: +52 222 2295500x7382; fax: +52 222 2295584.
E-mail addresses: maria.montiel@correo.buap.mx (S. Montiel-Smith), jesus.
sandoval@correo.buap.mx (J. Sandoval-Ramírez).
http://dx.doi.org/10.1016/j.steroids.2014.10.004
0039-128X/Ó 2014 Elsevier Inc. All rights reserved.

O. Viñas-Bravo et al. / Steroids 93 (2015) 60–67
61
(6, 22R,25S), diosgenin (9, 22R,25R) and yamogenin (11, 22R,25S,
Fig. 4), with no modifications at the stereogenic center C-20.
not separated, and the proportion was determined by 1H NMR as
1:1.4.
2.3. General procedure for the formation of 22-epi-sapogenin acetates
7a and 10a from their pseudosapogenins, employing a Brønsted acid
2. Experimental
2.1. General methods
To a solution of the pseudosarsasapogenin 3b,26-diacetate (12a,
1.05 mmol) or pseudodiosgenin 3b,26-diacetate (8a, 1.05 mmol) in
CHCl3 (20 mL), p-TsOHÁH2O (1.05 mmol) was added. The reaction
mixture was held at 50 °C for 30 min. Longer periods of time did
not change the product ratio. The mixture was poured over brine
(50 mL) containing smashed ice and was vigorously shaken. The
product was extracted with CH2Cl2 (50 mL) and washed with a sat-
urated solution of NaHCO3 (2 Â 50 mL) and finally water (50 mL).
The resulting solution was dried over Na2SO4 and concentrated
to dryness under vacuum. The residue was purified by column
chromatography (eluent: hexane–AcOEt 8:2) to give the corre-
sponding C-22 epimers 7a/6a and pseudosarsapogenin diacetate
(12a) in a 50:30:20 ratio or 10a/9a and pseudodiosgenin diacetate
(8a) in a 10:70:20 ratio. Spectroscopy data are in agreement with
the precedent procedure.
NMR spectra 1D and 2D 1H and 13C (DEPT, COSY, NOESY, HSQC,
HMBC) were recorded on
a Varian Mercury spectrometer
(400 MHz for 1H, or Bruker spectrometer 500 MHz for 1H). Chemi-
cal shifts are stated in ppm (d), and are referred to the residual 1H
signal (d = 7.27) or to the central 13C triplet signal (d = 77.0), for
CDCl3. Coupling constants (J) are quoted in Hz. IR spectra were
acquired on a Nicolet Magna FT-IR 750 spectrophotometer using
KBr pellets. Optical rotations were determined on a Perkin Elmer
241 polarimeter, at 24 °C, at the D line, using chloroform solutions.
Melting points were measured with a Mel-Temp apparatus and
were not corrected. Analytical TLC was performed on silica gel
ALUGRAMÒSIL G/UV254 plates and chromatographic columns con-
tained silica gel Davisil™ grade 633 (200–425 mesh).
2.3.1. (22S,25S)-5b-Spirostan-3b-yl acetate (22-epi-sarsasapogenin
acetate, 7a)
2.2. General procedure for the epimerization of C-22 of sarsasapogenin
acetate (6a), diosgenin acetate (9a), and yamogenin (11) using a
Lewis acid
Mp 144–146 °C [
C@O ester), 1261 (m
a
]
À52.2° (c 1.0, CHCl3); IR (cmÀ1): 1732
D
(m
asC–O). 1H NMR (CDCl3) d: 5.05 (1H, br s,
H-3), 4.38 (1H, ddd, J16,17 = 6.1 Hz, J15 ,16 = 7.3 Hz, J15b,16 = 8.3 Hz,
a
To a solution of sarsasapogenin acetate (6a, 1.0 g, 2.18 mmol) in
CHCl3 (20 mL), BF3ÁOEt2 (1.0 mL, 7.96 mmol) was added. The reac-
tion mixture was stirred for 30 min, at 50 °C. Longer periods of
time did not change the products ratio. The reaction was poured
into brine (50 mL) containing smashed ice and was vigorously sha-
ken. Then the mixture was extracted with CH2Cl2 (1 Â 50 mL). The
organic phase was washed with brine (1 Â 50 mL), a saturated
solution of NaHCO3 (2 Â 50 mL) and water (1 Â 50 mL), then dried
over Na2SO4 and concentrated to dryness under vacuum. The resi-
due (950 mg) was purified by column chromatography using silica
gel and eluted with an 8:2 hexane/AcOEt solution. The isolated
products were 22-epi-sarsasapogenin acetate (7a) (700 mg, 70%),
pseudosarsasapogenin 3b-acetate (12b) (40 mg, 5%), and starting
material 6a (200 mg, 20%).
H-16), 3.46 (1H, ddd, J26eq,26ax = 10.3 Hz, J25eq,26eq = 4.4 Hz,
J24eq,26eq = 1.8 Hz, H-26eq), 3.36 (1H, dd, J25ax,26ax = 10.3 Hz,
J26eq,26ax = 10.3 Hz, H-26ax), 2.02 (3H, s, CH3COO-3), 0.98 (3H, s,
CH3-19), 0.96 (3H, d, J20,21 = 6.9 Hz, CH3-21), 0.79 (3H, d,
J25,27 = 7.2 Hz, CH3-27), 0.76 (3H, s, CH3-18). 13C NMR (CDCl3) d:
30.7 (C-1), 25.0 (C-2), 70.6 (C-3), 30.6 (C-4), 37.3 (C-5), 26.5
(C-6), 26.5 (C-7), 35.3 (C-8), 39,8 (C-9), 35.0 (C-10), 20.9 (C-11),
40.2 (C-12), 40.7 (C-13), 56.3 (C-14), 31.8 (C-15), 80.8 (C-16),
62.2 (C-17), 16.5 (C-18), 23.9 (C-19), 41.6 (C-20), 14.6 (C-21),
109.1 (C-22), 31.4 (C-23), 28.8 (C-24), 30.3 (C-25), 66.8 (C-26),
17.2 (C-27), 21.6 (CH3COO-3), 170.4 (CH3COO-3).
2.3.2. (22S,25R)-Spirost-5-en-3b-yl acetate (22-epi-diosgenin acetate,
10a)
Using diosgenin acetate (9a, 1.0 g) as starting material, the iso-
lated products were 22-epi-diosgenin acetate (10a, 170 mg, 17%),
pseudodiosgenin 3b-acetate (8b, 50 mg, 6%), and starting material
(720 mg, 72%). For yamogenin (11) as starting material, the prod-
uct mixture of yamogenin (11) and 22-epi-yamogenin (13) was
Colorless oil; [
1242 (
a]
D À18.4° (c 1.1); IR (cmÀ1): 1728 (
mC@O ester),
m
asC–O). 1H NMR (CDCl3) d: 5.36 (1H, m, H-6), 4.59 (1H, m,
H-3), 4.40 (1H, ddd, J16,17 = 6.5 Hz, J15 ,16 = J15b,16 = 7.5 Hz, H-16),
a
3.94 (1H, dd, J26eq,26ax = 11.0 Hz, J26ax,25eq = 2.7 Hz, H-26ax), 3.29
R4
25
O
26
22
20
O
R1O
R2
R3
R1
R2
R3
R4
β−Η
Δ5
H
H
1 20S,22S,25R
2 20S,22S,25R
3 20S,22S,25R
β-
D
-Galp-(1 2)-β-
D
-Xylp-(1 3)-β-
D
-Glcp-(1 4)-β-D-Galp-
H
H
H
H
α-
L
-Rhap-(1 4)-β-
D
-Glcp-(1 4)-α- -Rhap-(1 2)-β-
L
D-Glcp-
α−Η
β-
D
-Glcp-(1 2)-α-
L
-Rhap-(1 4)-β-D-Glcp-
4 20S,22R,25S
,
H
α−Η
Δ5
β-D
-Xylp-(1 4)-β-
D
-Quip OH
5 20R,22S,25R
Rhap-(1 2)-Glup-
Η
H
Fig. 1. Structures of natural spiroketalic saponins possessing unusual configuration at C-22.

62
O. Viñas-Bravo et al. / Steroids 93 (2015) 60–67
(S)
O
(S)
(S)
(S)
O
(R)
(S)
O
O
RO
RO
7: R= H
7a: R=Ac
H
6: R= H
6a: R=Ac
H
Fig. 2. Structures of sarsasapogenin (6) and 22-epi-sarsasapogenin (7).
(R)
R1O
O
(R)
O
(S)
(S)
(R)
O
(S)
O
O
10: R=H
10a: R=Ac
8: R=R1=H
8a: R=R1=Ac
9: R=H
9a: R=Ac
RO
RO
RO
8b: R=Ac, R1=H
Derivative of the confined chiral BINOL
imidophosphoric acid
Fig. 3. Structures obtained from the reaction of pseudodiosgenin (8), catalyzed by chiral BINOL confined imidophosphoric acids [12].
a catalytic quantity of DMAP (10 mg) was added and the solution
was stirred for 15 min. Subsequently, the crude reaction was
washed with water (20 mL) and extracted with CH2Cl2
O
(S)
(S)
(R)
O
(2 Â 100 mL). The organic phase was washed with saturated
NaHCO3 (2 Â 100 mL), brine (30 mL), dried over Na2SO4 and con-
centrated under reduced pressure to dryness. The diasteroisomeric
mixture was separated by column chromatography (hexane/AcOEt
95:5) yielding the 22S,23R compound 15a (639 mg, 59%) and its
diastereoisomer, the 22R,23S spirostan 16a (422 mg, 39%).
RO
11: R=H
11a: R=Ac
Fig. 4. Structure of yamogenin (11).
2.4.1. (22S,23R)-23-Acetyldiosgenin acetate (15a)
(1H, ddd, J26eq,25eq = 2.7 Hz, J26eq,24eq = 1.0 Hz, J26eq,26ax = 11.0 Hz, H-
26eq), 2.02 (3H, s, CH3COO-3), 1.07 (3H, d, J25,27 = 7.2 Hz, CH3-27),
1.03 (3H, s, CH3-19), 1.00 (3H, d, J20,21 = 6.7 Hz, CH3-21), 0.79 (3H, s,
CH3-18). 13C NMR (CDCl3) d: 37.0 (C-1), 27.7 (C-2), 73.8 (C-3), 38.1
(C-4), 139.6 (C-5), 122.3 (C-6), 32.0 (C-7), 31.4 (C-8), 49.9 (C-9),
36.7 (C-10), 20.9 (C-11), 39.7 (C-12), 42.2 (C-13), 56.4 (C-14),
31.7 (C-15), 80.8 (C-16), 61.9 (C-17), 16.2 (C-18), 19.3 (C-19),
40.2 (C-20), 14.3 (C-21), 109.7 (C-22), 25.9 (C-23), 25.7 (C-24),
27.0 (C-25), 65.1 (C-26), 16.1 (C-27), 21.4 (CH3COO-3), 170.4 (CH3-
COO-3).
IR (cmÀ1): 2920 (
m
C–H); 1732 (
C@C), 1033 (
C–O). 1H NMR (CDCl3) d: 5.32 (1H,
d, J6,7 = 4.8 Hz, H-6), 4.57 (1H, m, H-3), 4.40 (1H, ddd, J16,17
mC@O ester); 1678 (mC@O
ketone), 1636 (
m
m
= J16,15 = 16.0 Hz, J16,15b = 7.6 Hz, H-16), 3.49 (1H, m, H-26ax),
a
3.41 (1H, dd, J25,26eq = 5.2 Hz, J26eq,26ax = 11.0 Hz, H-26eq), 2.64
(1H, dd, J23,24ax = 12.5 Hz, J23,24eq = 4.4 Hz, H-23), 2.30 (2H, m, H-
4), 2.15 (1H, m, H-20), 1.96 (1H, m, H-8), 1.71 (1H, m, H-14),
1.50 (2H, m, H-11), 1.01 (3H, s, CH3-19), 0.98 (3H, d,
J20,21 = 6.8 Hz, CH3-21), 0.92 (1H, m, H-9), 0.86 (3H, d,
J25,27 = 6.4 Hz, CH3-27), 0.66 (3H, s, CH3-18). 13C NMR (CDCl3) d:
37.0 (C-1), 27.8 (C-2), 73.8 (C-3), 38.1 (C-4), 139.6 (C-5), 122.1
(C-6), 32.1 (C-7), 31.3 (C-8), 49.9 (C-9), 36.8 (C-10), 20.9 (C-11),
40.6 (C-12), 39.8 (C-13), 56.3 (C-14), 32.2 (C-15), 81.1 (C-16),
61.2 (C-17), 16.2 (C-18), 19.4 (C-19), 38.5 (C-20), 14.3 (C-21),
108.9 (C-22), 55.0 (C-23), 31.6 (C-24), 29.9 (C-25), 66.1 (C-26),
17.0 (C-27), 170.3 (CH3COO-3), 21.5 (CH3COO-3), 209.9 (CH3COO-
23), 28.7 (CH3COO-23).
2.4. Procedure for the formation of (22S,23R)- and (22R,23S)-23-
acetyldiosgenin acetate (15a and 16a) from (25R)-23-acetyl-22,26-
epoxycholesta-5,22-diene-3b,16b-diyl diacetate (14)
An ethanolic 20% KOH solution (10 mL) containing 14 (1.0 g,
2.18 mmol) was stirred at 70 °C for 30 min. Once the reaction
cooled to room temperature, dilute aqueous HCl was added until
the mixture became acidic. The organic compounds were extracted
with CH2Cl2 (2 Â 100 mL). The organic layer was washed with a
saturated solution of NaHCO3 (30 mL) then with water
(2 Â 100 mL), dried over Na2SO4 and concentrated under reduced
pressure to dryness. To facilitate the chromatographic separation,
the residue was dissolved in acetic anhydride (2 mL, 21.90 mmol),
2.4.2. (22R,23S)-23-Acetyldiosgenin acetate (16a)
IR (cmÀ1): 2937 (
m
C–H); 1735 (
mC@O ester); 1700 (mC@O ketone),
1238, 1036 (
m
C–O ester). 1H NMR (CDCl3) d: 5.34 (1H, d, J6,7 = 4.8 Hz,
H-6), 4.57 (1H, m, H-3), 4.38 (1H, ddd, J16,17 = J16,15 = 16.0 Hz,
J16,15b = 7.6 Hz, H-16), 4.11 (1H, dd, J25,26ax = 13.2 Hz, J26eq,26ax = 2.9
Hz, H-26ax), 3.34 (1H, br d, J26eq,25 = 11 Hz, H-26eq), 2.77 (1H, dd,
a

O. Viñas-Bravo et al. / Steroids 93 (2015) 60–67
63
J23,24ax = 13.2 Hz, J23,24eq = 4.0 Hz, H-23), 2.32 (4H, m, H-4, H-20, H-
24), 1.54 (1H, m, H-24), 1.36 (2H, m, H-15) 1.12 (3H, d,
J20,21 = 7.3 Hz, CH3-21), 1.09 (3H, d, J25,27 = 7.3 Hz, CH3-27), 1.02
(3H, s, CH3-19), 1.00 (3H, s, CH3-18), 0.56 (2H, m, H-14, H-9), 2.13
(3H, s, CH3-232), 2.02 (3H, s, CH3-32). 13C NMR (CDCl3) d: 37.0 (C-1),
27.8 (C-2), 73.8 (C-3), 38.2 (C-4), 139.5 (C-5), 122.1 (C-6), 32.0 (C-7),
30.7 (C-8), 50.0 (C-9), 36.7 (C-10), 20.6 (C-11), 40.5 (C-12), 41.7 (C-
13), 56.7 (C-14), 34.1 (C-15), 84.1 (C-16), 59.1 (C-17), 15.7 (C-18),
19.4 (C-19), 42.7 (C-20), 19.5 (C-21), 106.5 (C-22), 51.5 (C-23), 29.9
(C-24), 26.9 (C-25), 66.0 (C-26), 16.7 (C-27), 170.2 (CH3COO-3), 21.5
(CH3COO-3), 211.0 (CH3COO-23), 28.6 (CH3COO-23).
compound, characterized as its C-22 epimer (7a, Scheme 2). After
10 min, the spectrum showed about 30% of the epimeric com-
pound and equilibrium was established 20 min later. This experi-
ment demonstrated that the new product, the 22 epimer is not
merely an artifact. The 22-epi-sarsasapogenin (7a) is slightly less
polar by TLC than compound 6a, and could be easily separated
by column chromatography in high yield (70%). The yield of the
22-epi derivative 7a was not improved using higher temperatures
or longer periods of time; instead some decomposition was
observed.
A plausible reaction mechanism is shown in Scheme 2. The F
ring opening is promoted by the action of BF3ÁOEt2, with the subse-
quent formation of two dispositions of the oxacarbenium ion (iv-a
and iv-b) or the carbenium ion (iv-c and iv-d). If the side chain
maintains the original spatial disposition (iv-a or iv-c), a subse-
quent nucleophilic attack by the C-26 oxygen would occur on the
3. Results and discussion
The rarity of C-22 epimeric spirostans in nature [3–11] has
increased interest in their synthesis, with a focus on finding new
biologically active sapogenins.
a
face (Si-face) generating 6a; on the other hand, if the side chain
is conformationally rearranged (iv-b or iv-d) the cyclization would
take place through the more crowded b face (Re-face) affording the
22-epimeric derivative 7a. Under reaction conditions, using BF3-
ÁOEt2, we previously observed no epimerization of spirostans at
C-20 [13a]; such epimerizations normally occur at higher temper-
atures and longer periods of time [9].
During the epimerization at C-22 of 6a, promoted by BF3ÁOEt2, a
small amount of pseudosarsasapogenin 3b-acetate (12b, Scheme 1),
was produced and identified by 1H NMR and by comparison with
an authentic sample.
An alternative methodology to obtain 7a was developed start-
ing from the pseudosarsasapogenin diacetate (12a) (Scheme 3),
under reflux in CHCl3 with an equimolar amount of p-TsOHÁH2O.
Under these conditions, the C-26 acetate was hydrolyzed to the
alcohol at C-26 (compound 12b). The protonation of the double
bond occurs at C-20 on the Re-face, thus directing the C-20 methyl
When a steroidal sapogenin is treated under suitable acidic con-
ditions, the F ring opens to a cation (resonance structures ii and iii),
followed by the fast equilibrium processes shown in Scheme 1
[13]. The alcohol group at C-26 in pseudosapogenins (i.e. 8b,
11b) can be immediately trapped to avoid the regeneration of
the spiroketal framework. For example, in the well known method-
ology developed by Marker [14], acetic anhydride generates the C-
26 acetates (12a and 8a).
The opened cation ii or iii normally reverts, via cation i, to the
natural 22R-spiroketalic framework through an attack of the 26-
OH over the Si-face [15]. However, formation of the 22S epimer
might be achieved by energetics that favor Re-face attack. From
these considerations and the previous experience of our research
group using Lewis acids for the functionalization of steroidal sapo-
genins [13a,16], we envisioned the use of BF3ÁOEt2 to achieve an
equilibrium of the opening-cyclization reaction of steroidal sapo-
genins and raising the temperature, from room temperature to
50 °C, to promote a different disposition of the side chain in order
to obtain the Re-face attack.
group to the steroidal
a-side, presumably to avoid steric conflict
with the bulky C-13 methyl group. In summary, two cationic inter-
mediates could be produced (v-a/b and v-c/d); a subsequent attack
of the C-26 oxygen at C-22 might occur on either the
a or the b
faces, generating the epimers 6a and 7a; for substrate 6a, forma-
tion of 7a was decidedly favored. Under all our reaction conditions
no epimerization occurred at C-20; the S configuration at C-20 was
retained.
In an exploratory reaction, a mixture of sarsasapogenin acetate
(6a) in CH2Cl2 and BF3ÁOEt2 was stirred at room temperature and
followed by TLC during different periods of time (up to 5 h). After
the work-up and neutralization, the 1H NMR spectrum showed
only signals corresponding to 6a, but when the reactions were
run during longer times at reflux temperature, the reaction became
brown and TLC exhibited many spots caused by decomposition of
the starting material. When the reaction was followed directly by
NMR, using CDCl3 as a solvent, at 50 °C, the spectrum showed a
gradual transformation of sarsasapogenin acetate (6a) into a new
The selective epimerization at C-22 was confirmed by 1H NMR
data, mainly by analysis of the C-26 proton couplings. For the start-
ing sarsasapogenin derivative 6a, having a 25S configuration (H-25
equatorial), the axial H-26 shows a dd signal due to the geminal
coupling with the equatorial H-26 (2J26ax,26eq = 11.0 Hz) and a
small axial–equatorial coupling with H-25 (3J26ax,25eq = 2.8 Hz).
c
A
O
O
H
+
O
O
O
H+
O
a:
=
Hβ 25
,
12
8
R
S
c
A
O
R
Δ5
=
c
a:
A
O
R
25R
,
i
a:
6
a:
9
=
=
R
R
R
Hβ 25
,
S
c
A
2O
Δ5
25R
,
HO
HO
HO
H
H
+
O
O
+
O
:
-
12b 5 Hβ 25
S
,
Δ5
25R
,
ii
:
8b
iii
Scheme 1. Acid catalyzed isomerization of steroidal sapogenins to pseudosapogenins.

64
O. Viñas-Bravo et al. / Steroids 93 (2015) 60–67
BF3
α
O
F3B
O
β
O
O
O
O
BF3
β
v-a
i
v-
i
b
α
O
BF3
O
α
O
a
a
7
6
β
O
O
O
BF3
v-c
i
v-
d
i
Scheme 2. Plausible reaction mechanism for the epimerization of C-22 by a Lewis acid.
β
c
A
O
OH
O
O
H
α
O
v-a
/b
/d
OH
β
H3O
a an a
6
d 7
O
8b
c
OH
A
O
H
a
8
α
v-c
Scheme 3. Cyclization of a pseudosapogenin framework by a protic acid medium.
Epimerization to the 22S epimer, results in a new chair conforma-
tion in which the C-25 methyl group is equatorial, leaving H-25 in
an axial position. The 1H NMR data of the 22-epi-derivative 7a
showed an expected change in the chemical shift and the coupling
pattern of the diasterotopic protons on C-26: the axial H-26
showed a geminal coupling (2J26ax,26eq = 10.3 Hz) and a vicinal cou-
pling with H-25, located in the axial position (3J26ax,25ax = 10.3 Hz).
These data are very similar to those reported for steroidal sapoge-
nins having the 25R configuration [17].
appears at 3.94 ppm while de same proton in 22-epi-yamogenin
(13) is observed at 3.36 ppm.
We also investigated the epimerization of 25R sapogenins.
There are very few reports of epimerizing the C-22 stereochemistry
of 25R-sapogenins [20]. The same experimental epimerization pro-
cedure described above was applied to diosgenin acetate 9a and
pseudodiosgenin diacetate 8a (Scheme 4). When BF3ÁOEt2 was
used, the 22-epi-diosgenin derivative 10 was isolated in only
17%. This low conversion might be rationalized by steric effects
and unfavorable energetics due to the axial position of the C-25
methyl in the 22S epimers having a 25R configuration.
The configuration at C-20 was corroborated by 13C NMR accord-
ing to literature data [18]. A change in the stereochemistry of C-20
from 20S to 20R results in a deshielding effect of ꢀ5 ppm for C-20
and a shielding effect of ꢀ3 ppm for C-21. The 13C NMR data of the
derivative 7a concerning C-20 (d 41.6 ppm) and C-21 (d 14.6 ppm)
showed similar chemical shifts in comparison with the starting
material [19] (d 40.1 and 14.4 ppm respectively). NOESY correla-
tions for derivative 7a (Fig. 5) indicated that H-20 and Me-18 are
both on the same b-steroidal face together with H-8 and H-15b
Also, the C-21 protons and H-14 show a strong correlation with
In the 1H NMR spectrum of 22-epi-diosgenin acetate 10a the dd
signal for axial H-26 has a big coupling constant (J26ax,26eq = 11.0 Hz)
and a small one (J26ax,25eq = 2.7 Hz), in contrast with the same proton
of derivative 9a that shows two big coupling constants
(J26ax,26eq = 11.0 Hz, J26ax,25ax = 11.0 Hz). Such observation demon-
strates the change in the orientation of the methyl group from equa-
torial to axial, confirming the change in the configuration on C-22 to
R. When C-20 changes to the S configuration, its 13C NMR chemical
shift is around 46.6 ppm [11]. The chemical shift of C-20 for com-
pound 10a (42.1 ppm) allowed us to deduce the retention in the ste-
reochemistry of C-20 as R.
H-17 located on the
configuration.
a-steroidal face (Fig. 5), supporting the 20S
Another sapogenin of the 25S series, yamogenin (11) was sub-
jected to the same epimerization conditions described above to
give 11 and 22-epi-yamogenin (13) in a 1:1.4 ratio, as judged by
1H NMR. Because the TLC RF values for 11 and 13 were virtually
the same, no attempt was made to separate them (Fig. 6). In the
1H NMR spectrum of yamogenin (11), the signal of H-26axial
Table 1 shows selected 1H NMR chemical shifts of sarsasapoge-
nin acetate (6a), diosgenin acetate (9a) and its corresponding epi-
mers in C-22 (7a and 10a). We point out the similarity in chemical
shifts of the methyl C-21 and C-27 and the diastereotopic protons
H-26 of the 22-epi-diosgenin acetate (10a) and those in sarsasapo-
genin acetate (6a); both have the C-27 in an axial orientation, and
the same pattern is observed between diosgenin acetate (9a) and
22-epi-sarsasapogenin acetate (7a) [11].
In Table 2, selected chemical shifts of 13C NMR of compounds
6a, 7a, 9a and 10a are presented. The observed signals from C-20
to C-27 of the compounds with C-27 in an axial orientation (6a
and 10a) are quite similar; these data suggested that only epimer-
ization at C-22 had occurred; in the case of derivatives with an
equatorial C-27 (7a and 9a) a similar pattern is observed.
In order to know the behavior of the 22-epi-sapogenins under
the action of the BF3ÁOEt2, 7a or 10a were treated under the same
O
H
CH3
H
H
CH3
O
H
Fig. 5. NOESY correlations of the (22S)-sarsasapogenin acetate (7a).

O. Viñas-Bravo et al. / Steroids 93 (2015) 60–67
65
O
O
O
O
.
t
BF3 OE
2
HO
HO
11
13
First fraction epimerization yamogenin
100
90
80
70
60
50
40
30
20
10
0
500
450
400
350
300
250
200
150
100
50
-yamogenin
H-26 axial in 22-epi
H-26 axial in
yamogenin
4.1
4.0
3.9
3.8
3.7
3.6
3.5
3.4
3.3
3.2
f1 (ppm)
0
5.4
5.2
5.0
4.8
4.6
4.4
4.2
4.0
3.8
3.6
3.4
3.2
3.0
f1 (ppm)
2.8
2.6
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
Fig. 6. Formation of 22-epi-yamogenin (13) from yamogenin (11). 1H NMR spectrum of the reaction mixture.
O
BF
O
3
E
t
C
2
O
H
5
0
2 C
°
l
C
2
,
30
m
O
i
n
c
O
A
a
9
O
or
a
9
+
a
8
+
c
OA
c
O
A
O
O
2
H
,
H
O
s
a
10
l 3
T
-
C
n
i
p
H
m
C
30
,
C
º
0
c
O
a
8
A
5
Scheme 4. Formation of the 22-epi-diosgenin acetate (10a).
conditions described above, resulting in the 22-epi-sarsapogenin
(7a) being transformed into sarsapogenin (6a) in a 10:1 ratio and
22-epi-diosgenin (10a) producing diosgenin (9a) in a 1:3 ratio.
Inspired by Chakravarty et al. [21] and Nohara et al. [22], who
studied the effect of different orientations of the substituents at
C-23 and C-25 of 23-hydroxyspirostanes as well as to change the
stereochemistry at C-22, we decided to study the influence of the
acetyl group at C-23 in the ratio of 22-diastereosomers on 23-
acetyldiosgenin.
Table 1
Selected 1H chemical shifts d (ppm) of 6a, 7a, 9a and 10a in CDCl3.
6a
7a
9a
10a
H-19
H-21
H-26a
H-26e
H-27
H-16
0.98
0.99
3.94
3.29
1.07
4.40
0.98
0.96
3.36
3.46
0.79
4.38
1.04
0.98
3.37
3.47
0.79
4.40
1.03
1.00
3.94
3.29
1.07
4.40
Our research group previously reported the transformation of
22,26-epoxycholestane 14 into 23-acetyldiosgenin 15 [23] using
a basic medium. According to the reaction mechanism (Scheme 5)
the formation of the spiroketal framework involves a nucleophilic
attack over the C-22 (sp2 carbon) of the intermediate vi. The alkox-
Table 2
Selected 13C chemical shifts d (ppm) of 6a, 7a, 9a and 10a in CDCl3.
6
7a
80.80
9a
80.73
10a
ide at C-16 then discriminates between the
D
22 faces. If the attack
C-16
C-17
C-20
C-21
C-22
C-23
C-24
C-25
C-26
C-27
80.94
62.00
42.19
14.41
80.83
61.86
42.06
14.31
109.68
25.90
25.72
27.02
65.05
16.13
occurs at C-22 by the Si face the 22S derivative is formed. In fact,
when the reaction was carried out at room temperature the only
reaction product was 15 [23,24]. Nevertheless, when 14 was trea-
ted with an ethanolic solution of KOH at 70 °C, a mixture of 22S
(15) and 22R (16) compounds was obtained in a 3:2 ratio
(Scheme 5). In both diastereoisomers the acetyl group is placed
in the equatorial position but only in 15 the C-25 methyl group
and the acetyl group are in equatorial positions.
62.21
41.60
14.59
109.05
31.41
28.83
30.32
66.77
17.22
62.02
41.62
14.64
109.14
31.43
28.85
30.34
66.80
17.25
109.54
25.79
24.99
27.08
65.05
16.11

66
O. Viñas-Bravo et al. / Steroids 93 (2015) 60–67
O
Ac
O
O
O
15: R=H
15a: R=Ac
RO
+
O
α
OH
O
O
β
Ac
O
O
O
O
14
vi
AcO
16: R=H
16a: R=Ac
RO
Scheme 5. Plausible mechanism to obtain 23-acetyl compounds 15 and 16.
At high temperature the attack at C-22 occurred on both the Si
4. Conclusion
and the Re faces, resulting in the formation of both diastereoiso-
mers. To explain the relative higher yield of the 22R epimer, we
might consider that the enol intermediate has a twisted conforma-
tion and the protonation of the enol at C-23 may produce the ther-
modynamic ketone, in which the C-23 acetyl substituent is
equatorial. The mixture of compounds 15 and 16 was easily sepa-
rated by chromatography after acetylation of the 3b-alcohols (15a
and 16a).
In conclusion, we have accomplished the epimerization at C-22
of the 25S-sapogenins sarsasapogenin and yamogenin into their
22-epimers, in high yield under mild conditions, realizing a selec-
tive E-ring opening of the spirostan framework. In the 22S-sapoge-
nins the methyl group at C-25 is oriented equatorially. Low yields
were obtained in attempts to epimerize C-22 for the 25R-sapoge-
nin diosgenin, perhaps due to the steric hindrance produced by
the methyl group at C-25 which changes from an equatorial posi-
tion to an axial one. In all cases, the configuration of C-20 was not
disturbed. The isomerization at C-22 of 23-acetyldiosgenin gave
increased formation of the 22R-23-acetyldiosgenin in which the
methyl group at C-25 is placed axially while maintaining the 23-
acetyl group in an equatorial disposition.
In the 1H NMR spectrum of 16a a deshielding of the signal cor-
responding to H-26ax (0.7 ppm) and a shielding (0.15 ppm) of the
H-26eq signal, compared to the same protons of the derivative 15a,
is observed. Furthermore, for compound 16a small coupling con-
stants were observed between H-26ax and H-25 (J26ax,25 = 2.9 Hz)
and between H-26eq and H-25 (J26eq,25 ꢀ 0 Hz), which suggest an
equatorial orientation for H-25. In both cases, the acetyl group
on C-23 is placed in an equatorial orientation, as deduced by the
coupling constants between H-23 and H-24ax (12.5 Hz and
13.2 Hz respectively). Those coupling constants are typical for an
axial-axial interaction; even though the diastereoisomer 16a is still
the minor product (39%), the yield of this unnatural C-22 epimer is
remarkable in comparison with the yield of 10a (17%). This result
suggests a stabilizing effect due to the acetyl group on C-23.
The 1H NMR data for compounds 15a and 16a revealed informa-
tive changes in chemical shifts and coupling patterns in the E and F
rings (H-20, H-18, H-21 and H-27; see Table 3) as a result of the
change of configuration in carbons 22 and 23. The similarity of dis-
placement and coupling constants of H-23 (2.64, J23,24ax = 12.5 Hz,
J23,24eq = 4.4 Hz, and 2.77, J23,24ax = 13.2 Hz, J23,24eq = 4.0 Hz) con-
firmed the equatorial position of the acetyl group.
Acknowledgments
The authors thank CONACYT-México for scholarships to ARL,
the Retention grant 190909 to PMM and financial support (grant
176858 to JSR). Also, we thank VIEP-BUAP for financial support
(grants to SMS and FJM).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.steroids.2014.10.
004.
References
[1] (a) Sparg SG, Light ME, van Staden J. Biological activities and distribution of
plant saponins. J Ethnopharmacol 2004;94:219–43;
(b) Sautour M, Mitaine-Offer AC, Lacaille-Dubois MA. The Dioscorea genus: a
Table 3
Selected NMR 1H chemical shifts d (ppm) of 15a and 16a in
review of bioactive steroid saponins.
references cited therein;
J Nat Med 2007;61:91–101. and
CDCl3.
(c) Mimaki Y, Yokosuka A, Kuroda M, Sashida Y. Cytotoxic activities and
structure–cytotoxic relationships of steroidal saponins. Biol Pharm Bull
2001;24:1286–9;
(d) Perron F, Albizati KF. Chemistry of spiroketals. Chem Rev
1989;89:1617–61.
15a
16a
H-3
H-6
4.57
5.32
1.71
4.40
0.66
1.01
2.15
0.98
2.64
0.86
4.57
5.34
0.56
4.38
1.00
1.02
2.32
1.12
2.77
1.09
H-14
H-16
H-18
H-19
H-20
H-21
H-23ax
H-27
[2] Agrawal PK. Determining ring-F configuration in spirostane-type steroidal
sapogenins by 1H NMR. Indian J Chem 2005;44B:1092–4.
[3] Kudo K, Miyahara K, Marubayashi N, Kawasaki T. Characterization of a minor
compound, which accompanies the usual 22a(R)-O,25b(S)-spirostanol
glycoside, as novel type of 22b(S)-O,25b(S) analog. Chem Pharm Bull
1984;32:4229–32.
[4] Inoue T, Mimaki Y, Sashida Y, Nikaido T, Ohmoto T. Steroidal saponins from
Dichelostemma multiflorum tubers and their inhibitory activity on cyclic-AMP
phosphodiesterase. Phytochemistry 1995;39:1103–10.

O. Viñas-Bravo et al. / Steroids 93 (2015) 60–67
67
[5] Ohtsuki T, Sato M, Koyano T, Kowithayakorn T, Kawahara N, Godad Y, et al.
Steroidal saponins from Calamus insignis, and their cell growth and cell cycle
inhibitory activities. Bioorg Med Chem 2006;14:659–65.
Novel transformation of the side chain in sapogenins. Tetrahedron Lett
1999;40:5143–6;
(b) Sandoval-Ramírez J, Meza-Reyes S, Del Río RE, Hernández-Linarez G,
[6] Hayes PY, Jahidin A, Lehmann R, Penman K, Kitching W, De Voss JJ. Steroidal
saponins from the roots of Asparagus racemosus. Phytochemistry
2008;69:796–804.
[7] Lu Y, Luo J, Kong L. Steroidal alkaloid saponins and steroidal saponins from
Solanum surattense. Phytochemistry 2011;72:668–73.
[8] Yen CT, Lee CL, Chang FR, Hwang TL, Yen HF, Chen CJ, et al. Indiosides G-K:
Steroidal glycosides with cytotoxic and anti-inflammatory activities from
Solanum violaceum. J Nat Prod 2012;75:636–43.
[9] Marker RE, Rohrmann E. Sterols. LIII. The structure of the side chain of
sarsasapogenin. J Am Chem Soc 1939;61:846–51.
Suárez-Rojas A, Rincón S. Regioselective cleavage of rings
sarsasapogenin. Steroids 2003;68:199–204;
(c) Fernández-Herrera MA, Sandoval-Ramírez J, Meza-Reyes S, Montiel-Smith
S. Side-chain opening of steroidal sapogenins to form 22-oxocholestanic
skeletons an approach to analogues of the aglycone of the potent anticancer
agent OSW-1. J Mex Chem Soc 2009;53:126–30;
E and F in
(d) Fernández-Herrera MA, López-Muñoz H, Hernández-Vázquez JMV, López-
Dávila M, Escobar-Sánchez ML, Sánchez-Sánchez L, et al. Synthesis of 26-
hydroxy-22-oxocholestanic frameworks from diosgenin and hecogenin and
their in vitro antiproliferative and apoptotic activity on human cervical cancer
CaSki cells. Bioorg Med Chem 2010;18:2474–84.
[10] (a) Wall ME, Serota S. Steroidal sapogenins XX. Configuration of spiroketal side
chain at carbon 22. J Am Chem Soc 1954;76:2850–1
[17] Han XW, Yu H, Liu XM, Bao X, Yu B, Li C, et al. Complete 1H and 13C NMR
assignments of diosgenyl saponins. Magn Reson Chem 1999;37:140–4.
[18] Morzycki JW, Paradowska K, Dabrowska-Balcerzak K, Jastrzebska I,
Siergiejczyk L, Wawer I. 13C NMR study of spirostanes and furostanes in
solution and solid state. J Mol Struct 2005;744–7:447–55.
(b) The used structures for sarsasapogenin and smilagenin, in Refs. [10a,11],
and that of smilagenin in J Am Chem Soc 1940;62:1162-3 are wrong at the C22
stereocenter, guiding to false conclusions.
[11] Tobari A, Teshima M, Koyanagi J, Kawase M, Miyamae H, Yoza K, et al.
Spirostanols obtained by cyclization of pseudosaponin derivatives and
comparison of anti-platelet agglutination activities of spirostanol glycosides.
Eur J Med Chem 2000;35:511–27.
[19] Viñas-Bravo O, Hernández-Linares G, Mata-Esma M, Martínez-Pascual R,
Montiel-Smith S, Meza-Reyes S, et al. 1H and 13C NMR of synthetic steroid
sapogenins. Part II. C-23 Substituted derivatives of (25S)-spirostanes. Arkivoc
2003;xi:163–71.
ˇ
[12] Coric´ I, List B. Asymmetric spiroacetalization catalyzed by confined Brønsted
acids. Nature 2012;483:315–9.
[20] Quan HJ, Koyanagi J, Ohmori K, Uesato S, Tsuchido T, Saito S. Preparations of
[13] (a) Viñas-Bravo O, Martínez-Pascual R, Vega-Báez JL, Gómez-Calvario V,
Sandoval-Ramírez J, Montiel-Smith S, et al. Rapid conversion of spirostans
into furostans skeletons at room temperature. Steroids 2012;77:59–66;
(b) Jastrzebska I, Dobrogowska A, Lutostanska E, Morzycki JW. On reactions of
spirostane sapogenins with benzeneseleninic anhydride. Tetrahedron
2010;66:5024–9;
heterospirostanols and their pharmacological activities. Eur
2002;37:659–69.
J Med Chem
[21] Chakravarty AK, Pakrashi SC, Uzawa J. 13C Nuclear Magnetic Resonance spectra
of 23-hydroxy spirostane sapogenins of Solanum hispidum. Can
1981;59:1328–30.
J Chem
[22] Iida Y, Yanai Y, Ono M, Ikeda T, Nohara T. Three unusual 22-b-O-23-hydroxy-
(5 )-spirostanol glycosides from the fruits of Solanum torvum. Chem Pharm
(c) Lee JS, Fuchs PL. Efficient protocol for ring opening of spiroketals using
trifluoroacetyl trifluoromethanesulfonate (TFAT). Org Lett 2003;5:3619–22;
(d) Gould DH, Staeudle H, Hershberg EB. Catalytic isomerization of spirostans
to furostenols. J Am Chem Soc 1952;74:3685–8.
a
Bull 2005;53:1122–5. and references cited therein.
[23] Meza-Reyes S, Sandoval-Ramirez J, Montiel-Smith S, Hernández-Linares G,
Viñas-Bravo O, Martínez-Pascual R, et al. b-Alkoxy-a,b-unsaturated ketone
[14] Marker RE, Sterols Rohrmann E. Conversion of sarsasapogenin to
systems in steroidal frameworks, and their conversion to 23,24-bisnorcholane
lactones. Arkivoc 2005;vi:307–20.
[24] Hernández-Linares MG, Sandoval-Ramírez J, Meza-Reyes S, Montiel-Smith S,
Guerrero-Luna G. Stereospecific synthesis of new steroidal isoxazoles in dry
media. Steroids 2011;76:1521–6.
pregnanediol-3(
[15] Dewick PM. Medicinal natural products
a
),20(
a
). J Am Chem Soc 1939;61:3592–3.
biosynthetic approach. 2nd
a
ed. Chichester: John Wiley & Sons, Ltd.; 2002. p. 238.
[16] (a) Sandoval-Ramírez J, Castro-Méndez A, Meza-Reyes S, Reyes-Vázquez F,
Santillán R, Farfán N. Preparation of 22,26-epoxycholest-22-ene steroids.