Mechanistic studies of the rearrangements of steroidal 16,17-ketols and syntheses of 20→16-cis-γ-carbolactones

Article abstract of DOI:10.1016/S0968-0896(99)00042-5

Utilization of 17-keto-androstanes as starting materials for the synthesis of α- or β-oriented steroidal 20→16-γ-carbolactones has been explored following two different strategies. A highly efficient, stereospecific protocol has been developed for the β-oriented cis-γ-lactone. A different approach, involving prior attachment of a 3-carbon side chain on C-17 of a 17-oxo-16β-acetoxy-androstane led to the epimeric, α-oriented lactone. The mechanism of the rearrangement of epimeric 16β- or 16α-hydroxy-17-keto-androstanes to 17β-hydroxy-16-keto-androstanes was studied by ¹³C NMR spectroscopy. The former occurs through a 1,2-sigmatropic H-shift, while the latter is likely to take place by simple enolization-reprotonation. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(99)00042-5

Bioorganic & Medicinal Chemistry 7 (1999) 943±947
Mechanistic Studies of the Rearrangements of Steroidal
16,17-Ketols and Syntheses of 20!16-cis-ꢀ-Carbolactones
Andrea C. Bruttomesso, Darõo Dollery and Eduardo G. Gros*
Â
Departamento de Quõmica Organica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabellon 2,
Â
Â
Ciudad Universitaria, 1428 Buenos Aires, Argentina
Received 28 September 1998; accepted 20 January 1999
`The teacher who walks in the shadow of the temple, among his followers, gives not of his wisdom but rather of his faith and his lovingness.'
Dedicated with respect and admiration to the memory of Professor Sir Derek Barton.
AbstractÐUtilization of 17-keto-androstanes as starting materials for the synthesis of a- or b-oriented steroidal 20!16-g-carbo-
lactones has been explored following two di€erent strategies. A highly ecient, stereospeci®c protocol has been developed for the b-
oriented cis-g-lactone. A di€erent approach, involving prior attachment of a 3-carbon side chain on C-17 of a 17-oxo-16b-acetoxy-
androstane led to the epimeric, a-oriented lactone. The mechanism of the rearrangement of epimeric 16b- or 16a-hydroxy-17-keto-
androstanes to 17b-hydroxy-16-keto-androstanes was studied by 13C NMR spectroscopy. The former occurs through a 1,2-sigmatropic
H-shift, while the latter is likely to take place by simple enolization±reprotonation. # 1999 Elsevier Science Ltd. All rights reserved.
Introduction
In addition, products containing the 20!16-cis-g-lac-
tone moiety (IVb-c) have more recently been isolated
from di€erent natural sources.6
Our interest in steroidal cis-20!16-g-carbolactones ori-
ginated in the course of studies on the pathway by
which tomato plants convert the steroidal alkaloid
tomatine (Ia) or its aglycone tomatidine (Ib) into allo-
pregnenolone (II, Scheme 1).1 The isolation of 23x-hy-
droxy-tomatidine (III) suggested the possibility of C-23
being the site where initial enzymatic hydroxylation
takes place, which would be followed by oxidative
cleavage leading to putative intermediate lactone IVa
(tigogeninlactone).2 This might su€er further oxidation
with loss of carbon dioxide to a€ord allopregnenolone
II. In order to determine the fate of the atoms in the
piperidine ring we required a specimen of tomatidine
isotopically labeled within this ring. We felt that a
lactone like IVa would be an appropriate synthetic
precursor towards this goal, as addition of an organo-
metallic reagent to the lactone ring, followed by func-
tional group manipulations, might provide the desired
labeled target.
Initial synthetic e€orts
Our initial synthetic approach to this g-lactone system is
shown in Scheme 2. It involved the addition of a
Reformatsky reagent (from ethyl 2-bromopropionate)
on a 16b-acetoxy-17-keto-androstane (V). We obtained
a mixture of four di€erent b-hydroxy esters (Scheme 2).
These were separated by HPLC, and characterized
spectroscopically as compounds VIa±b and VIIa±b.7
We rationalized the formation of products alkylated on
C-16 by calling for an isomerization of the 16b-hy-
droxy-17-keto androstane system to a 17b-hydroxy-16-
keto, which would su€er addition of the Reformatsky
reagent and yield esters VIIa±b after acetylation of
secondary hydroxyl groups (Scheme 3). In support of
this hypothesis, the transposition of carbonyl groups in
steroids, from C-17 to C-16 as well as other positions
such as C-3 to C-2, had been reported.8
Tigogeninlactone (IVa) had been used as a key inter-
mediate in the preparation of a variety of sapogenins3
and alkaloids.4 Sir Derek Barton himself had reported a
photochemical preparation of this type of compound.5
From the synthetic standpoint, we solved this problem
by switching to milder reagents. Lithium ester enolates
at low temperatures a€orded only products of addition
on C-17.7 However, we were curious as to the di€er-
ences with earlier results reported in the literature. The
same reaction carried out on the 16a-epimer (VIII) had
previously been reported to a€ord a single product.9
Key words: Natural products; steroids and sterols; NMR; stereo-
speci®city; biosynthesis.
y*Corresponding author. E-mail: gros@quimor.qo.fcen.uba.ar
Present address: 3-Dimensional Pharmaceuticals, 665 Stockton
Drive, Suite 104, Exton, PA 19341, USA.
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00042-5

944
A. C. Bruttomesso et al. / Bioorg. Med. Chem. 7 (1999) 943±947
Scheme 1. Putative degradative pathway of tomatine.
Scheme 2. Reformatsky reaction on V.
Scheme 3. Rationale for the formation of VIIa±b.
Indeed, we veri®ed that under the same reaction
conditions the 16a-epimer yields exclusively (HPLC
analysis) a single product of addition on C-17.10 Simi-
larly, Reformatsky reactions of ethyl bromoacetate on
both 16a- and 16b-acetoxy-17-keto-androstanes were
reported to yield the expected b-hydroxy esters in good
yields.11
transformations. The ®rst mechanism would involve
enolization to an ene-diol (IX), followed by reprotona-
tion on C-17 from the less-hindered a-face, to produce
epimer X, which would presumably be the thermo-
dynamically more-stable epimer. The second one would
be a sigmatropic 1,2-intramolecular H-shift from C-16
to C-17 (Scheme 4).
13C-NMR spectroscopy experiments were designed to
explore these hypotheses.12 First, we prepared a solu-
tion of compound V in CD3OD/C6D6. Upon addition
of KOD in D2O, we veri®ed the complete transforma-
tion to X after a few minutes. At no point in time were
we able to detect the resonance for C-17. This fact was
indicating the presence of a deuterium atom on C-17,
and thus con®rmed the possibility of using the presence
Mechanistic studies on the interconversion of steroidal
16,17-ketols
We speculated that the di€erent behavior of these C-16
epimers towards Reformatsky reaction conditions was
due to di€erences in their rate of conversion to (the
same) 17b-hydroxy-16-keto-androstane (X). In principle,
we considered two di€erent ionic pathways for these

A. C. Bruttomesso et al. / Bioorg. Med. Chem. 7 (1999) 943±947
945
Scheme 4. Postulated pathways for ketols interconversion.
(or absence) of a resonance for C-17 as a diagnostic tool
for the incorporation of a deuterium atom (rather
than hydrogen) on a given carbon atom of the steroid
molecule.13
reaction product (stabilized by intramolecular hydrogen
bonding). However, for the 16b-hydroxy epimer proto-
nated intermediate (B), the parallel orbital alignment
between the C16±H16a s-orbital and the empty C-17
p-orbital allows a fast 1,2-migration that leads to the
17b-hydroxy-16-keto-epimer. This migration is not
preferred for the 16a-epimer, as the orbital alignment
upon protonation of the carbonyl oxygen would not be
favorable.
More interesting results were obtained by treatment of
the CD3OD/C6D6 solution of V with catalytic amounts
of D2SO4. The conversion of the starting material was
much slower than under basic conditions, and indeed
we detected the presence of rearranged product X
showing the C-17 resonance for a few hours. As time
went by, this resonance decreased in intensity and
eventually disappeared. Since solvents and catalyst are
fully deuterated there is no source of H atoms. There-
fore, we interpret these results as showing that under
acidic catalysis the conversion of V to X involves a 1,2-
migration of H16a to C-17. This is followed by a slower
enolization±reprotonation (which introduces a deuter-
ium on C-17 of X). Similar conclusions were reached by
Numazawa et al. in experiments using deuterated sub-
strates and mass spectrometry analysis.14
Molecular mechanics calculations on the epimeric
starting materials and reaction intermediates A and B
have provided support for this explanation.15 As shown
in Figure 1, the calculated dihedral angles formed by H-
16b/C-16/C-17/O-17 for intermediate A is ca. 106ꢀ,
while the calculated dihedral angles formed by H-16a/
C-16/C-17/O-17 for intermediate B is ca. 93ꢀ. Thus, this
di€erence could be signi®cant, and intermediate B could
provide substantially better orbital overlapping, which
is the driving force for the intramolecular migration
process.
For the 16a-epimer VIII, experiments using stoichio-
metric amounts of KOD as a promoter were also too
fast to allow for the detection of any intermediate, and
only fully-deuterated reaction product X was seen.
However, under acidic conditions (D2SO4) an interest-
ing result was obtained, as no formation of rearranged
product was seen and instead, a new product was
formed. This was isolated by reverse-phase HPLC, and
characterized as hemiketal XI (Scheme 5).12b
Synthesis completion
We achieved addition speci®cally on C-17 of V by using
the lithium enolate of t-butyl propionate at low tem-
peratures (Scheme 7). After exchange of the acid-labile
t-butyl group for a methyl group, methyl ester XIV was
dehydrated by treatment with catalytic PTSA in tolu-
ene. This yielded a mixture of lactone XV and ester
XVI, easily separable by ¯ash chromatography. It is
noteworthy that in forming lactone XV, inversion of
stereochemistry at C-16 accompanies dehydration, as
Thus, both 16-epimers show indeed a di€erent behavior
under acidic conditions (Scheme 6). In both cases reac-
tivity is triggered by protonation of the oxygen on the
C-17 carbonyl group, resulting in reaction intermediates
which are diastereomeric at C-16. In the case of the 16a-
hydroxy epimer (A), the intermediate's most favorable
pathway is the formal addition of CD3OD to the car-
bonyl group, resulting in the hemiketal being the only
1
determined by H NMR nOe experiments. Hydrogena-
tion of lactone XV occurred smoothly, a€ording the a-
oriented g-lactone XVII, which was epimerized under
basic conditions to the more stable lactone XVIII.14 On
the contrary, catalytic hydrogenation on conjugated
ester XVI was extremely slow, and double bond satura-
tion was observed only after hydrogenolysis of the 16b-
acetoxy group. This presumably relieves steric hin-
drance around the tetra-substituted double bond, and
saturated ester XIX could be obtained.
In order to obtain the desired b-oriented g-lactone the
synthetic strategy was modi®ed as shown in Scheme 8.16
Wittig reaction on 3b-(dimethyl-t-butylsilyloxy)-17-oxo-
5a-androstane XX yielded stereo-speci®cally ole®n XXI.
Scheme 5. Fate of VIII under acidic conditions.

946
A. C. Bruttomesso et al. / Bioorg. Med. Chem. 7 (1999) 943±947
Scheme 6. Di€erent fates for 16-epimeric ketols.
Figure 1. Calculated geometries for intermediates A and B.
Scheme 7. Synthesis of a-oriented 20!16-cis-g-carbolactone.
Scheme 8. Synthesis of b-oriented 20!16-cis-g-carbolactones.

A. C. Bruttomesso et al. / Bioorg. Med. Chem. 7 (1999) 943±947
947
Allylic hydroxylation with TBHP in the presence of
References and Notes
catalytic amounts of SeO2, followed by Swern oxidation
produced in very good yield the conjugated ketone
XXIII. Michael addition of NaCN in a THF±EtOH±
H2O mixture a€orded nitrile XXIV as a single diaster-
eoisomer. Stereospeci®c reduction of the ketone XXIV
with lithium tri-t-butoxy aluminumhydride a€orded the
16b-hydroxy-derivative XXV. As expected, the bulky
hydride approaches the carbonyl group from the less-
hindered a-face of the steroid nucleus, producing in
excellent yield and selectivity the needed b-orientation
for the hydroxy group on C-16. Alkaline hydrolysis of
nitrile XXV, including an acidic work-up, produced
lactone IVa, where lactonization as well as deprotection
of the 3b-OTBDMS group took place during acidic
work-up. This lactone resulted identical with authentic
specimens obtained by degradation of tigogenin and
tomatidine.
1. Heftmann, E.; Schwimmer, S. Phytochem. 1972, 11, 2783.
2. Ripperger, H.; Schreiber, K. The Alkaloids; Academic
Press: New York, 1981, Vol XIX, pp 81.
3. Mazur, Y.; Danieli, N.; Sondheimer, F. J. Am. Chem. Soc.
1960, 82, 5889.
4. Fieser, L. F.; Fieser, M. Steroids; Reinhold: New York,
1959, p 810±846.
5. Barton, D. H. R.; Beckwith, A. L.; Goosen, A. J. Chem.
Soc. 1965, 181.
6. Vespertilin (Ib): Gonzalez, A. G.; Francisco, C. G.; Bar-
Â
reira, R. F.; Lopez, E. S. A. Quõm. 1971, 67, 433. Solanolide
(Ic): Chackravarty, A. K.; Das, B.; Pakrashi, S. C. Phytochem.
1982, 21, 2083.
7. Doller, D.; Gros, E. G. Synth. Commun. 1990, 20, 3115.
8. Pelc, B. Steroids; Johns, W. F.; Ed.; Butterworths Uni-
versity Park Press: London, 1973, p 92.
9. Mazur, Y.; Danieli, N.; Sondheimer, F. J. Am. Chem. Soc.
1960, 82, 5889.
10. NMR analysis of the reaction product showed attack
from the b-face. Reference 9 reported the opposite stereo-
chemistry for the attack by the Reformatsky reagent.
11. Ola, K.; Hara, S. J. Org. Chem. 1978, 43, 4408.
12. (a) Doller, D.; Gros, E. G. Magnetic Resonance in Chem-
istry 1988, 26, 529. (b) Idem, Helvetica Chimica Acta 1989, 31,
353.
13. Abraham, R. J.; Loftus, P. Proton and Carbon-13 NMR
Spectroscopy; Heyden and Son: Philadelphia, 1981, pp 72.
14. Numazawa, M.; Nagaoka, M.; Mutsumi, A. Chem.
Pharm. Bull. 1987, 35, 4763.
In summary, we developed a facile, high yielding pro-
tocol for the preparation of steroidal b-oriented 20!16-
cis-g-carbolactones. The di€erent reactivity of epimeric
16-hydroxy (or acetoxy)-17-keto-androstanes has been
explained based on the di€erent pathways they have
available to rearrange to the more-stable 17b-hydroxy-
16-keto isomer. This type of ketol isomerizations have
been reported previously, and di€erent reaction
mechanisms were postulated, including enolization and
hydration±dehydration.17 The generality of the synthetic
protocol presented is being tested with vespertilin (IVb)
and solanolide (IVc).
15. Molecular mechanics calculations (MM2) were carried out
using Chem3D Pro version 4.0 (1997), Cambridge Soft Cor-
poration, Cambridge, MA.
16. Bruttomesso, A. C.; Doller, D.; Gros, E. G. Synth. Com-
mun. 1998, 28, 4043.
Acknowledgements
17. Numazawa, M.; Osawa, Y. J. Am. Chem. Soc. 1980, 102,
5402. Collins, C. J.; Easthman, J. F. The Chemistry of the
Carbonyl Compounds; Patai, S.; Ed.; John Wiley: New York,
1966, Chapter 15, pp 761.
We thank Universidad de Buenos Aires and Consejo
Nacional de Investigaciones Cientõ®cas y Tecnicas
(CONICET, Argentina) for partial ®nancial support.