Chemo-, Regio-, and Diastereoselective Hydrogenation of Oxopromegestone into Trimegestone over Supported Platinoids: Effects of the Transition Metal, Support Nature, Tin Additives, and Modifiers

Article abstract of DOI:10.1006/jcat.1999.2760

The selective hydrogenation of oxopromegestone (17α-methyl-17β-(1,2-dioxopropyl)-estra-4,9-dien-3-one), 1, into Trimegestone (17α-methyl-17β-(2(S)-hydroxy-1-oxopropyl)-estra-4,9-dien-3-one), 2, was carried out on various monometallic catalysts. The order of activity, Pd〉Rh〉Pt〉Ir〉Os, was almost opposite that of the chemoselectivity (selective hydrogenation of the carbonyl versus the internal olefinic double bond), Os〉Ir〉Pt=Pd〉Rh. No metal gave any required diastereoselectivity (selective formation of the 21(S)-OH alcohol). However, silica-supported rhodium, palladium, and platinum exhibited 100% diastereomer excess (d.e.) for the hydrogenation of the C₃ carbonyl group. Pt/SiO₂ catalyst exhibited a low chemoselectivity (24%), but its TON activity was relatively high (42×10^-3 s^-1). Pt_s(Sn)_n/SiO₂ catalysts, prepared by the interaction of Sn(CH₃)₄ with reduced Pt/SiO₂ under H₂ at room temperature, exhibited a significant increase of chemoselectivity compared to Pt/SiO₂, but still a low diastereoselectivity to the desired 21(S)-OH unsaturated ketoalcohol, 2. In parallel, the d.e. for the hydrogenation of the C₃ carbonyl group decreased from 100% to 34%. The addition of (-)cinchonidine and hydrocinchonidine on the platinum-tin catalysts sharply increased the d.e. in favor of the 21(S)-OH unsaturated ketoalcohol to 70% without changing the chemoselectivity.

Full text of DOI:10.1006/jcat.1999.2760

Journal of Catalysis 190, 364–373 (2000)
doi:10.1006/jcat.1999.2760, available online at http://www.idealibrary.com on
Chemo-, Regio-, and Diastereoselective Hydrogenation of
Oxopromegestone into Trimegestone over Supported
Platinoids: Effects of the Transition Metal, Support
Nature, Tin Additives, and Modifiers
Yurii A. Ryndin,† Catherine C. Santini, ,1 Denis Prat,‡ and Jean Marie Basset
Laboratoire de Chimie Organome´tallique de Surface, UMR CPE-CNRS 9986, 43 Bd du 11 Novembre 1918, 69616 Villeurbanne Cedex, France;
†Boreskov Institute of Catalysis, Russian Academy of Sciences, Siberian Branch, Prosp. Akad. Lavrentieva 5, 630090 Novosibirsk,
Russian Federation; and ‡Hoechst-Marion-Roussel, 102 route de Noisy, 93235 Romainville Cedex, France
Received July 30, 1999; revised November 12, 1999; accepted November 12, 1999
(HMR) with almost 100% chemo-, regio-, and diastereos-
electivity from oxopromegestone (17 -methyl-17 -(1,2-di-
oxopropyl)-estra-4,9-dien-3-one), 1 (1, 2). The main draw-
back of this bioreduction is the high dilution which requires
the use of rather huge reactors. This difficulty opens the
field for alternative processes. We report here the prepa-
ration and use of supported mono- or bimetallic catalysts
in the selective hydrogenation of oxopromegestone, 1, into
Trimegestone, 2 (Scheme 1).
The selective hydrogenation of oxopromegestone (17 -methyl-
17 -(1,2-dioxopropyl)-estra-4,9-dien-3-one), 1, into Trimegest-
one (17 -methyl-17 -(2(S)-hydroxy-1-oxopropyl)-estra-4,9-dien-
3-one), 2, was carried out on various monometallic catalysts. The
order of activity, Pd > Rh > Pt > Ir > Os, was almost opposite that
of the chemoselectivity (selective hydrogenation of the carbonyl ver-
sus the internal olefinic double bond), Os > Ir > Pt = Pd > Rh. No
metal gave any required diastereoselectivity (selective formation
of the 21(S)-OH alcohol). However, silica-supported rhodium, pal-
ladium, and platinum exhibited 100% diastereomer excess (d.e.)
for the hydrogenation of the C3 carbonyl group. Pt/SiO2 catalyst
exhibited a low chemoselectivity (24%), but its TON activity was
relatively high (42 10 3 s 1). Pts{Sn}n/SiO2 catalysts, prepared by
the interaction of Sn(CH3)4 with reduced Pt/SiO2 under H2 at room
temperature, exhibited a significant increase of chemoselectivity
compared to Pt/SiO2, but still a low diastereoselectivity to the de-
sired 21(S)-OH unsaturated ketoalcohol, 2. In parallel, the d.e. for
the hydrogenation of the C3 carbonyl group decreased from 100%
to 34%. The addition of ( )cinchonidine and hydrocinchonidine
on the platinum–tin catalysts sharply increased the d.e. in favor of
the 21(S)-OH unsaturated ketoalcohol to 70% without changing the
c
To achieve this aim it was necessary to solve simultane-
ously three problems of chemo-, regio-, and diastereoselec-
tivity in the hydrogenation of 1 by supported metals cata-
lysts. Selective hydrogenation of
, unsaturated ketones
by supported group VIII metals into the corresponding un-
saturated alcohols was considered in a recent review (3).
There isno example in the literature ofselective hydrogena-
tion of an unsaturated polyketone into the corresponding
unsaturated ketoalcohols. We report here the first example
of such a selective reduction.
2. EXPERIMENTAL
chemoselectivity.
2000 Academic Press
Key Words: oxopromegestone; Trimegestone; diastereoselective
hydrogenation; chemoselective hydrogenation; platinum–tin cata-
lyst; hydro( )cinchonidine modifier.
2.1. Origin of the Organic Compounds
Oxopromegestone and isomerically pure hydrogenated
products 2–7 (Scheme 2) were provided by HMR.
( )Cinchonidine (CD), (+)cinchonine (CN), hydro( )
cinchonidine (HCD), quinuclidine, quinoline, quinidine
(QD), quinine (QN), and tetramethylstannane purchased
from Aldrich were used without further purification. Ethyl
acetate was purified by distillation as described in the liter-
ature (4).
1. INTRODUCTION
Trimegestone, 2 (17 -methyl-17 -(2(S)-hydroxy-1-oxo-
propyl)-estra-4,9-dien-3-one), is a new progestomimetic
molecule developed for the treatment of postmenopausal
diseases. On an industrial scale, 2 is obtained by a biore-
duction process developed by Hoechst Marion Roussel
2.2. Catalyst Preparation and Characterisation
2.2.1. Monometallic catalysts. The commercial Pd/C
(5.0% ) catalyst was purchased from Aldrich. Degussa
Aerosil 200 silica with a surface area of about 200 m2/g was
1 To whom correspondence should be addressed. Fax: 0033472431795.
E-mail: santini@mulliken.cpe.fr.
364
0021-9517/00 $35.00
c
Copyright
2000 by Academic Press
All rights of reproduction in any form reserved.

HYDROGENATION OF OXOPROMEGESTONE INTO TRIMEGESTONE
365
SCHEME 1
used as the support for Rh, Os, Ir, and Pt. It was calcined tube. The desired amount of tetramethylstannane and 10 ml
under an air flow at 500 C before impregnation. Al2O3, of n-heptane were added. The reaction medium was stirred
with a surface area of 100 m2/g, ZrO2 with a surface area at room temperature for 22 h under hydrogen. The com-
of 49 m2/g, and TiO2, P25TM, with a surface area of 50 m2/g plete hydrogenolysis reaction was proved by GC analysis
were obtained from Degussa.
of evolved methane (capillary column, KCl/Al2O3, 130 C).
The Os/SiO2 (1.0% ) catalyst was prepared by treat- After several washings with n-heptane, the solid was kept
ment of the dehydroxylated silica with Os3(CO)12 in under dynamic vacuum and 10 4 Torr (1 Torr = 1.33 N/m2)
dichloromethane solution under an argon atmosphere (5). at 50 C for 1 h to eliminate all physisorbed tetramethyl-
The Rh/SiO2 (1.4% ) and Pt/SiO2 (1.3% ) catalysts were stannane. The catalyst was then covered with 10 ml of ethyl
obtained by ionic exchange from [Rh(NH3)5]+(Cl ) and acetate and this suspension was introduced into the auto-
[Pt(NH3)4]2+(OH )2 respectively, by surface protons of sil- clave under argon.
ica (6, 7). The Ir/SiO2 (1.5% ) catalyst was obtained by im-
Elemental analysis performed on the bimetallic catalyst
pregnation of silica with Ir(acac)3 in toluene. The Pt/ZrO2 before and after the catalytic run showed no decrease in the
(2.0% ) catalyst was obtained by impregnation of zirco- Sn% and gave the number, n, of fixed tin atoms by surface
nia with Pt(acac)2. The Rh/Al2O3 (1.8% ) and Rh/TiO2 platinum atoms, n = Sn/ Pts. These catalysts are described
(1.4% ) catalysts were obtained by reduction with H2 of as Pts{Sn}n/SiO2.
(M–O)Rh( 3-C3H5)2 (M = Al, Ti) (8).
All the supported metallic precursors described above
2.3. Catalytic Tests of Oxopromegestone Hydrogenation
were washed, dried, and decomposed by calcination at
The hydrogenation was carried out either under 1 atm of
400 C under a stream of nitrogen/oxygen mixture (5/1).
All monometallic catalysts except Pd/C were reduced un-
der flowing H2 at 350 C overnight before the hydrogenation
reaction took place.
The dispersion of the supported metals, expressed as the
fraction of surface atoms (Ms/Mt), was calculated from H2
then introduced into the Schlenk tube or autoclave under
chemisorption data, obtained in a volumetric apparatus,
Pd (9, 10), Pt (11–14), and Rh (6). According to transmis-
sion electron microscopy the average surface metal parti-
cle size in Rh/SiO2 and Ir/SiO2 catalysts was 1 and 3 nm,
4.6 mm; eluent: acetonitrile–water 70/30 (v/v), flow rate:
respectively; this calculation was in good agreement with
H2 in a Schlenk tube, or at 80 atm (1 atm = 0.1 MPa) in a
Parr 4560 minireactor, fitted with a mechanical stirred. In
a usual run, 0.4 g of oxopromegestone, 1, were dissolved
into 50 ml of ethyl acetate under argon. The thin powder of
reduced monometallic or dried platinum–tin catalyst was
an argon flow.
Monitoring of the reaction and analysis were carried out
by HPLC {column: symmetry C18, 5 m, l = 25 cm; d =
1 ml/mn; detection: UV 210 nm}.
chemisorption data.
The selectivities were calculated at 50% conversion, a
value at which the overreduction products could be ne-
glected (i.e., products in which more than one keto or olefin
group is reduced). The numbers in Tables 1–5 are the per-
2.2.2. Bimetallic catalyst Pts{Sn}n/SiO2. The prepara-
tion of the bimetallic catalysts has already been described
(7). Pt/SiO2 was reduced at 350 C under hydrogen for 4 h.
At room temperature, the hydrogen flow was replaced by centages of each product in the crude reaction mixture cal-
argon, and then the catalyst was introduced into a Schlenk culated by HPLC assay with respect to a pure sample; e.g.,

366
RYNDIN ET AL.
SCHEME 2
in Table 1, the Os-catalysed reduction gave at 50% conver- ketones, and even fully reduced into the saturated alcohols
sion 12% Trimegestone, 2, 12% 21(R)-OH, 3, 4% 20-OH, 4 (Scheme 2).
and 5, and 66% 3-OH, 6 and 7. The missing 6% is compat-
The initial unsaturated ketoalcohols, which can be pro-
ible with the overall experimental error and not identified, duced by route 2, are the three pairs of diastereomers:
fullyreduced products. The Rh-catalysed reaction gave only 21(S)-OH, 2 (Trimegestone), and 21(R)-OH, 3, 20(S)-
12% alcohols; the main pathway was in this case the hydro- OH and 20(R)-OH, 4 and 5, 3 -OH, 6, and 3 -OH, 7
genation of the olefinic double bonds.
(Scheme 2).
In this work chemoselectivity refers to the reduction of
keto groups vs all possible reduced products, regioselec-
tivity to the reduction of the ketone at C21 vs all possible
reduced products, and diastereoselectivity to the Si/Re re-
3. RESULTS AND DISCUSSION
Hydrogenation of the initial unsaturated triketone, 1, can
take place at five different sites, two olefins and three ke-
tones at C3 , C20, and C21 and in each case from two sides,
duction at C21 and
/ reduction at C3.
2
If S(i) is the selectivity in the compound i, i.e., the ratio
between the amount of desired product and the amount of
/
( = cis to methyl group at C17) for bonds located on
substrate consumed, the chemoselectivity (SC O), regiose-
==
the steroid skeleton, or Si/Re for the two ketones C20 and
C21 on the side chain. If the carbon–carbon double bonds
are reduced first (Scheme 2, route 1), partially saturated
triketones are formed. If the ketone groups are reduced
first (Scheme 2, route 2), the unsaturated hydroxyketones,
2–7, are obtained. Then, both kinds of compounds can be
further reduced into partially or fully saturated hydroxy-
lectivity (RS), and diastereomer excess (d.e.) will be defined
by the following equations:
(SC=O ) = S(2) + S(3) + S(4) + S(5) + S(6) + S(7); [1]
RS = S(2) + S(3);
d.e. C21 = 100[S(2) S(3)]/[S(2) + S(3)];
d.e. C3 = 100[S(7) S(6)]/[S(7) + S(6)].
[2]
[3]
[4]
2 Ketone at C is more reactive than the other keto groups at any
3
position on steroid (15).

HYDROGENATION OF OXOPROMEGESTONE INTO TRIMEGESTONE
367
TABLE 1
Monometallic Catalysts: Influence of the Metal Nature on the Catalytic Properties in the Hydrogenation
of Oxopromegestone at Room Temperature, S(i) = Selectivity, mol% for Compound i
M,
wt%
P,
atm
Dispertion,
Ms/Mt
TON,
103, s
d.e.
C21
d.e.
C3
1
Cat.
S(2)
S(3)
S(4, 5)
S(6)
S(7)
SC
=
O
Pda
Rh
Pt
Ir
Os
5.0
1.6
1.3
1.5
1.0
1
1
1
80
80
0.25
0.80
0.55
0.30
0.90
252
58
42
1.2
0.2
0
0
<1
4
0
0
3
4
12
0
0
0
4
4
0
0
0
13
22
26
12
21
25
44
26
12
24
50
94
—
—
?
0
0
100
100
100
32
12
33
a Pd on carbon, all others metals are supported on silica.
3.1. Oxopromegestone Hydrogenation on Different Metals
(i) The carbonyl group at C3 of oxopromegestone, 1, and
==
==
the C4 C5 and C10 C9 double bonds of the steroid skele-
ton are conjugated, leading to a flat part of the molecule, 1.
The 100% diastereoselectivity of supported Rh, Pd, and Pt
with respect to 3 -OH, 7, may be explained by the chirality
of oxopromegestone and the preferable -side multicenter
chemisorption of this flat part on the several neighbouring
surface atomsofthese metals. In thiscase, asproposed in the
diastereoselective hydrogenation of a steroidic -diketone
(19), the order of activities and selectivities are also gov-
erned by the geometry of the steroid.
The activities and the selectivities of different metals in
the hydrogenation of oxopromegestone, 1, are reported in
Table 1. Regarding the activity, the most active metals were
the supported Rh and Pd catalysts, the less active being Ir
and Os. The difference of TON between the most active
Pd/C catalyst and the least active Os/SiO2 catalyst is more
than 3 orders of magnitude. The most active metals were
tested at 1 atm. The order of TON activity of supported
platinum metals is Pd > Rh > Pt > Ir > Os.
Regarding the chemoselectivity, the monometallic cata-
lysts could be divided into two groups: (i) With the Pt,
Pd, and Rh catalysts, the hydrogenation of the carbon–
carbon double bonds was favoured (route 1), affording
mainly the 19,14 triketones and/or full reduction products
(ii) According to theoretical calculations (20), activities
and selectivities can be explained by the value of the band
width of d bands of metal catalysts. The larger this band
width, the stronger the repulsive interaction of the metal
==
with the C C bond and thus the lower the probability of
==
(Scheme 2). As no sample of these products of C C reduc-
chemisorption. The bandwidth of d bands of the metals
tion was available, their actual structure is unknown and
their amount was deduced from the value of the chemos-
increases in the series Pd < Pt < Ir Os, which accounts for
=
the observed selectivities (21). In this case, the nature of
the catalyst seems to be the crucial factor in the observed
activity and selectivities.
electivity (SC O). With these catalysts, the 3 -OH, 7, was
==
the only known alcohol formed. Besides 7, a small quantity
of Trimegestone, 2, was produced over Pt/SiO2, preventing
d.e. C21 calculation. (ii) With Ir and Os catalysts, the reduc-
According to this hypothesis, with Os and Ir, the
==
tion of keto groups was favoured (route 2). In particular, chemisorption via the C C bond is expected to be weak.
Os exhibited a total chemoselectivity, but a low regio- and Consequently, the substrate was chemisorpted by keto
stereoselectivity in C21 (d.e. = 0% ) and in C3 (d.e. = 33% ) groups. So the less sterically hindered ketones are reduced
reduction. With these catalysts, the diastereoselectivity in preferentially C3 (66% ) > C21 (24% ) > C20 (4% ) with a low
C3 did not change with conversion.
stereoselectivity.
The order of chemoselectivity (SC O) of supported
The effect of the support nature is given in Table 2. The
==
metals was almost opposite to their order of activities: best activityand diastereoselectivityat C3 were obtained for
Os > Ir > Pt Pd > Rh. This result is close to the literature supported Rh and Pt metals on silica. With TiO2- and ZrO2-
=
data for the hydrogenation of
, unsaturated aldehydes supported metals, low activity and low diastereoselectivity
and ketones into the corresponding unsaturated alcohols at C3 were observed. Morever, at a high reduction tempera-
(3). The trends are the following: Os and Ir are selective, ture (HRT) of Pt/ ZrO2 these effects were increased. In con-
Rh and Pd unselective, and Ru and Pt moderately selective trast, the chemoselectivity is increased (SC O = 51% ). All
==
(16–18).
these results can be explained by strong metal-supported
interaction (SMSI) which is the decoration of the met-
als surface by carrier suboxides (22). These ZrOx species
decrease the multiplicity of active sites (the number of
neighbouring surfaces atoms) and the probability of the
The order of activity and selectivities can be explained
(i) by a steric control governed by the geometry of the
steroid and (ii) by the intrinsic properties of the cata-
lyst.

368
RYNDIN ET AL.
TABLE 2
Effect of the Support on Catalytic Properties of Rh and Pt at P(H2) = 1 atm and Room Temperature,
S(i) = Selectivity, mol% for Compound i
Dispertion,
Ms/Mt
TON,
103, s
1
Cat.
M, wt%
S(2)
S(3)
S(4, 5)
S(6)
S(7)
d.e. C3
Rh/Al2O3
Rh/SiO2
Rh/TiO2
Pt/SiO2
Pt/Al2O3
Pt/ZrO2
Pt/ZrO2
1.8
1.6
1.4
1.3
0.5
2.0
2.0
0.9
0.8
0.9
0.55
0.9
0.2
0.2
21
58
23
42
9
0
0
0
<1
0
2
0
0
0
3
0
0
0
0
0
0
2
2
0
0
8
0
0
3
7
30
12
10
21
7
100
100
11
100
100
62
12
2
3
10
13
22
10
52
a Catalyst reduced at 800 C.
-side multicenter chemisorption of oxopromegestone. Se- tion at C20 and C3 was almost constant (4 and 5 and 6 and
quentially, the activity and 3 -OH diastereoselectivity is 7), the presence of Sn adatoms favours the hydrogenation
decreased.
of 1 into 2 and 3 and the regioselectivity (RS) reaches 60% .
At the same time, the stereoselectivity at C3 decreased and
remained very low at C21; the variation of the latter is com-
parable to the accuracy of HPLC analysis (Table 3). Note
that the catalyst Pts{Sn}0.5/SiO2 represents the best compro-
3.2. Effect of Sn on Pt/SiO2 Properties
The selectivity and activity of several metals such as Ni,
Ru, Rh, Pd, Ir, and Pt can be strongly modified by interac-
tion with organometallic compounds such as R4Sn (23). In
particular, the presence of Sn alkyl fragmentsor Sn adatoms
on the surface of Rh completely reversed the chemoselec-
tivity of this metal in the hydrogenation of , unsaturated
aldehydes (24) and of aromatic ketones (25).
Various Pts{Sn}n/SiO2 catalysts were prepared by selec-
tive hydrogenolysis of Sn(CH3)4 in n-heptane solution with
a reduced Pt/SiO2 catalyst. The TON activity was calcu-
lated in the hypothesis that the platinum dispersion does
not change after tin modification at room temperature. The
activities and selectivities of these Pts{Sn}n/SiO2 catalysts of
different compositions are reported in Table 3.
mise between chemoselectivity (SC O = 89% ), regioselec-
==
tivity (RS = 53% ), stereoselectivity (d.e. C21 = 14% ), and
activity.
All these results were obtained at 50% conversion. More
precise data could be obtained by following the kinetics of
the reaction and the evolution of conversion and selectiv-
ity with time. Typical data are given in Fig. 1. These kinetic
data can tell us what the primary and secondary products
are. With Pts{Sn}n/SiO2 (n = 0.35–0.63), the diastereoselec-
tivity at C3 and at C21 did not vary with time and conversion,
but the chemoselectivity decreased after 40% conversion
(Fig. 1). This decrease of chemoselectivity is due to the sec-
ondary reduction of unsaturated hydroxyketones, as pro-
posed in Scheme 2. The fact that the diastereoselectivity at
C3 and at C21 did not change with conversion means that
there was no difference in the rate of further reduction of
the two diastereomers, 2 and 3, in contrast with results ob-
tained with butane-2,3-dione (26).
Formation of Pt–Sn bonds induced a decrease in the cata-
lytic activity of platinum, but increased its chemoselectivity.
This is a general observation made with several catalytic re-
actions carried out with Pts{Sn}n catalysts (23, 25). Tin ad-
ditives reduced the TON activity of Pt/SiO2 by 1 order of
magnitude, but increased its chemoselectivity (SC O) from
24% to 60–90% . Besides, as the percentage of hydrogena-
==
The ratio of tin introduced by the surface atom of Pt var-
ied from 0 to 0.63. At these values we have already shown
TABLE 3
Influence of the Sn/Pts Atomic Ratio (n) on the Catalytic Properties of Pts{Sn0}n/SiO2 Catalysts in the Hydrogenation
of Oxopromegestone at 25 C and P(H2) = 1 atm, S(i) = Selectivity, mol% for Compound i
1
{Sn}n/Pts, n
TON, 103, s
S(2)
S(3)
S(4, 5)
S(6)
S(7)
d.e. C3
SC
=
RS
d.e. C21
O
0
42.0
6.8
6
4.2
2.3
1.2
<1
16
21
30
28
28
3
14
19
23
28
33
0
8
10
11
10
11
0
5
6
6
7
7
21
22
17
19
15
14
100
63
62
52
36
34
24
65
73
89
88
93
<4
30
40
53
56
61
?
8
5
13
0
8
0.35
0.44
0.5
0.56
0.63

HYDROGENATION OF OXOPROMEGESTONE INTO TRIMEGESTONE
369
with Sn/Pts = 0.25. In the selective reduction of aromatic
ketones, the hydrogenation of the aromatic ring could be
suppressed, when Rhs{Sn}0.3/SiO2 was used as the catalyst
instead of Rhs/SiO2 (25). So it is very reasonable to con-
clude that Sn can acts as a site blocker in the adsorption
==
of the conjugated C C bonds. The hydrogenation of a car-
1
bonyl function via a
mode is less sterically demanding
than that of an olefinic compound.
Consequently, (i) the low diastereoselectivity at C3 can
be explained, again, in this case by a less strong chemisorp-
tion of the side, like in the case of Pt/ZrO2, Rh/TiO2, and
Os/SiO2, and (ii) the hydrogenation of a carbonyl function
1
via an
mode, less sterically demanding than that of an
olefinic compound, is favoured. But the high C3 regioselec-
tivity is not observed, as in the case of Ir and Os catalysts.
FIG. 1. Variation of the chemoselectivity and diastereoselectivity at
C3 and at C21 with the conversion in the hydrogenation of oxopromege-
stone at 25 C and P(H2) = 1 atm on catalyst Pts{Sn}0.5/SiO2.
==
This means that the C O reduction is controlled not only
by the sterical factor but also by the different chemical na-
ture of the three carbonyl groups and different electronic
properties of modified Pt.
that all tin–carbon bonds of SnMe4 were hydrogenolysed
(27). So, the surface of platinum is modified by zerovalent
tin adatoms (28, 29) which are likely located on specific
crystallographic sites at the surface of the metallic particle.
As already reported in the literature (3), the chemoselec-
tivity of Pt in the hydrogenation of unsaturated aldehydes
into unsaturated alcohol is improved by modification with
tin adatoms. The enhancement of chemoselectivity could be
due to two kinds of electronic modifications of the surface:
3.3. Influence of Modifiers on the Properties
of Pts{Sn}0.5/SiO2
As both the chemoselectivity and regioselectivity could
be improved, but not sufficiently, by the use of Pts{Sn}0.5
/
SiO2 catalyst, variousfactorswhich could enhance the stere-
oselectivity at C21 were studied. With supported Pt/Al2O3,
enantioselective hydrogenation of -diketones is possi-
ble with chiral modifiers such as ( )cinchonidine and
hydro( )cinchonidine (26, 35). As oxopromegestone is it-
self a chiral substrate, a chiral modifier could enhance
the diastereoselectivity of the reduction, a concept already
known as double asymmetric induction (36). We first used
the alkaloid ( )cinchonidine (CD) for the modification of
the Pts{Sn}0.5/SiO2 catalyst.
(i) The oxophilic tin could favour the coordination of the
1
keto group in an
mode (Scheme 3). (ii) The formation
of Pt–Sn bonds change the charge on Pt and thus induce a
==
weaker adsorption of the C C bond on platinum (less back
donation) (20, 30, 32). For instance, the heat of adsorption
of the ethylene decreases from 23.3 kcal/mol with Pt/SiO2
to 16.6 kcal/mol with PtSn/SiO2 (31).
Another explanation can be proposed by considering
that tin adatoms are inactive. Their presence on the sur-
face could just prevent the adsorption ofthe multifunctional
molecule to Pt via its -electrons system, as proposed for Pt
modified with bismuth adatoms (33, 34). Bismuth has a very
small electronic effect on Pt and with a ratio of Bi/Pts = 0.25,
ethylene and benzene adsorption and hydrogenation were
suppressed, which require respectively an ensemble of four
and six free adjacent Pt atoms. The same was observed
3.3.1. ( )Cinchonidine(CD). In the presence of CD
and independent of the reaction conditions, the TON ac-
tivity was lower than that in the absence of CD (Fig. 2).
This result is opposite to the strong increase in the reaction
rate usually observed with cinchona alkaloid modified sup-
ported platinum (37). Selectivity to 6 and 7 was decreased.
The diastereomer excess at C21 and the chemoselectivity
were improved when the CD/Pts ratio wasincreased (Fig. 2).
This effect was enhanced at high H2 pressure and low reac-
tion temperature (Table 4).
As with the unmodified Pts{Sn}0.5/SiO2 catalyst, in the
presence of CD the diastereoselectivity at C21 did not de-
pend on the conversion, and the chemoselectivitydecreased
with conversion (Fig. 3). Also in this case, there was no
difference in the rate of further reduction of the different
diastereomers, 2 and 3.
At room temperature, in the presence of cinchonidine,
the highest diastereomeric excess (d.e.) of 63% for the
21(S)-OH, 2, was obtained with the Pts{Sn}0.5/SiO2 catalyst
(Table 4).
SCHEME 3

370
RYNDIN ET AL.
FIG. 2. Influence of the ratio modifier/Pts on chemoselectivity and
FIG. 3. Dependence of the chemoselectivity and diastereoselectiv-
ity at C21 from the conversion in the oxopromegestone, 1, hydrogena-
tion at 25 C, P(H2) = 80 atm with catalyst Pts{Sn}0.5/SiO2 modified by
diastereoselectivity at C21, and activity in the oxopromegestone, 1, hydro-
genation at room temperature, P(H2) = 80 atm on catalyst Pts{Sn}0.5/SiO2
modified by ( )cinchonidine (CD).
(
)cinchonidine (CD), CD/Pts = 4.6, and CD/(1) = 0.028.
3.3.2. Influence of other modifiers. The effects of the provement of activity and diastereoselectivity with hy-
nature of other modifiers, such as (+)-cinchonine, hydro( ) drocinchonidine instead of ( )cinchonidine is generally
cinchonidine, quinuclidine, quinoline, quinidine, and qui- reported (38). When (+)cinchonine is used instead of
nine on the catalyticpropertiesofPts{Sn}0.5/SiO2 catalystsin ( )cinchonidine, the diastereoisomeric excess at C21 was
the oxopromegestone hydrogenation were studied as well lower, but still higher, than the one obtained with no modi-
(Scheme 4).
fier (respectively, 32% , 63% , and 14% ). Quinidine and qui-
Compared to the Pts{Sn}0.5/SiO2, these modifiers im- nine gave roughly the same chemo-, regio-, and stereose-
proved simultaneously, but at different levels, the chemo-, lectivities. These modifiers double the d.e. of the reduction
regio-, and diastereoselectivity at C21 (Table 5).
at C21, but inhibit the reaction strongly. This inhibition has
The TON activity decreases in the order no ligand > already been reported in the enantioselective hydrogena-
hydro( )cinchonidine, quinuclidine > quinoline > ( )cin- tion of an -ketoacid, though to a lower extent (39). It is
chonidine, cinchonine quinine, quinidine. In any case, no clear that the methoxy group of these modifiers plays a ma-
ligand acceleration was observed. Clearly, the observed ac- jor role, either by simple coordination to the catalyst or
tivity cannot be related to the basicity of the modifier. because this group is hydrogenolysed with the formation of
The diastereoselectivity of the hydrogenation at C21 was phenol (17).
improved by additional ligands, even nonchiral: no lig-
Many studies indicate that the three crucial structural
elements of cinchona alkaloids are (1) the quinoline ring,
and quinoline < quinine, quinidine, (+)cinchonine
quinuclidine < ( )cinchonidine, hydrocinchonidine. Im- acting as an anchoring part, (2) the stereogenic region, which
TABLE 4
Influence of CD/Pts Ratio and of Catalytic Conditions on Catalytic Properties Pts{Sn}0.5/SiO2 in the Hydrogenation
of Oxopromegestone, 1, CD = ( )Cinchonidine, CD/(1) = 0.057 for P(H2) = 1 atm and 0.028 for P(H2) = 80 atm, S(i) =
Selectivity, mol% for Compound i
P,
atm
T,
C
TON,
103, s
d.e.
C21
1
CD/Pts
S(2)
S(3)
S(4, 5)
S(6, 7)
SC
=
RS
O
0
1
1
1
80
80
80
80
80
22
26
22
22
23
22
0
4.2
2.3
1.1
20.6
11.8
7.9
30
27
31
12
28
49
31
39
23
13
15
9
10
11
9
11
10
12
6
13
12
10
10
25
>20
6
42
21
7
89
>70
64
69
72
79
68
62
53
40
46
21
38
60
40
46
13
35
35
14
47
63
55
70
0.5
4.6
0
0.5
4.6
0.5
4.6
3.8
3.0
18
6
0
7

HYDROGENATION OF OXOPROMEGESTONE INTO TRIMEGESTONE
371
SCHEME 4
determines the chirality of the product when the substrate dine gave in majority the 21(S)-OH alcohol, 2, and with a
is prochiral, and (3) the quinuclidine moiety, which when better selectivity than the nonmodified catalyst.
protonated, is directly linked to the substrate by a hydrogen
Finally, in the presence of acetic acid the hydrogenation
bond (38, 40). First our results clearly show that tin adatoms was nonstereoselective (d.e. C21 = 0). Generally, the high-
prevent adsorption via the -bonding system, so the quino- est optical yields reported so far were achieved in acetic
line ring cannot be bound to the catalyst by this way. As a acid which protonates the quiniclidine-N of cinchonidine
matter of fact, the effects of quinoline on the reactivity and and so favours the interaction between the modifier and re-
the chemoselectivity are the same as those of the nonaro- actant by hydrogen bonding (38). Our results indicate that
matic quinuclidine (30–50% inhibition, high chemoselec- the quiniclidine moiety is not directly bound to the sub-
strate during the hydrogenation. More likely, quinuclidine
is absorbed on the catalyst and hinders in some way the ap-
proach of the substrate, thus slowing down the reaction and
enhancing the steric control, which favours the formation
of the 21(S)-OH alcohol, 2.
tivity). Second, in our case, the chirality of the modifier
plays a secondary role in the diastereoselectivity; the main
factor is the chirality of the substrate itself. Each pair of
pseudo-enantiomers (cinchonidine, cinchonine), (quinucli-
dine, quinidine) and the nonchiral quinoline and quinucli-
TABLE 5
Effects of the Nature of the Modifier (M) on the Catalytic Properties of Pts{Sn}0.5/SiO2 Catalysts in the Hydrogenation
of Oxopromegestone, 1, at P(H2) = 80 atm, T = 25 C, S(i) = Selectivity, mol% for Compound i
Modifier
M/Pts = 4.6
TON,
103, s
d.e.
C21
1
S(2)
S(3)
S(4, 5)
S(6, 7)
pK
(SC
=
)
RS
O
0
20.6
15.1
12
46
9
10
6
18
42
6
69
80
21
56
14
64
Hydrocinchonidine
5.8
10.0
5.8
10.0
5.8
10.0
5.1
9.7
(
)Cinchonidinea
7.9
6.8
1.0
0.8
49
25
43
40
11
13
22
22
12
11
14
13
7
16
15
10
79
65
94
85
60
38
65
62
63
32
32
29
(+)Cinchonine
Quinine
Quinidine
5.4
10.0
4.9
11
Quinoline
Quinuclidine
10.5
13.7
13
3
8.7
29
45
24
39
55
19
16
24
7
12
20
12
10
21
38
12
23
6
98
93
83
62
91
48
61
48
46
64
21
48
0
70
72
(
(
)Cinchonidine + acetic acid
)Cinchonidineb
Hydrocinchonidineb
9
6
a CD/(1) = 0.028.
b T = 0 C.

372
RYNDIN ET AL.
The selectivity could probably be improved by changing neous catalysis and that the quinuclidine structure has a
the experimental conditions. This is especially true if the peculiar effect in the stereoselectivity. These two fields are
activity is not too low. So with hydrocinchonidine at 0 C, currently under development.
oxopromegestone was hydrogenated with 91% chemose-
lectivity, 64% regioselectivity, and 72% d.e. at C21.
REFERENCES
CONCLUSION
1. Crocq, V., Masson, C., Winter, J., Richard, C., Lemaitre, G., Lenay, J.,
Vivat, M., Buendia, J., and Prat, D., Org. Process Res. Dev. 1, 2 (1997).
2. Buendia, J., Crocq, V., Masson, C., Prat, D., and Vivat, M., Patent EP
0574,317 (Roussel Uclaf) 1993.
The selective hydrogenation of oxopromegestone (17 -
methyl-17 -(1,2-dioxopropyl)-estra-4,9-dien-3-one), 1, in-
to Trimegestone, (17 -methyl-17 -(2(S)-hydroxy-1-oxo-
propyl)-estra-4,9-dien-3-one), 2, requires catalysts with a
high degree of chemoselectivity (selective hydrogenation
of the carbonyl versus the internal olefinic double bond),
regioselectivity (hydrogenation of the exocyclic C21 car-
bonyl group), and stereoselectivity (selective obtention of
the 21(S)-OH alcohol).
The order of activity for the monometallic catalysts
was Pd > Rh > Pt > Ir > Os; the order of chemoselectivity
was almost the opposite: Os > Ir > Pt = Pd > Rh. No metal
gave any required diastereoselectivity. However, silica-
supported Rh, Pd, and Pt exhibited 100% d.e. for the hy-
drogenation of the C3 carbonyl group.
3. Ponec, V., Appl. Catal. A 149, 27 (1997).
4. Perrin, D. D., Armarego, W. F. L. F., and Perrin, D. R., “Purification
of Laboratory Chemicals.” Pergamon Press, Oxford, 1986.
5. Basset, J. M., and Choplin, A., J. Mol. Catal. 21, 95 (1983).
6. Candy, J. P., Ferretti, O. A., Mabilon, G., Bournonville, J. P.,
El Mansour, A., Basset, J. M., and Martino, G., J. Catal. 112, 210 (1988).
7. Humblot, F., Didillon, B., Lepeltier, F., Candy, J.-P., Corker, J., Clause,
O., Bayard, F., and Basset, J.-M., J. Am. Chem. Soc. 120, 137 (1998).
8. Dufour, P., Santini, C. C., Houtman, C., and Basset, J.-M., J. Mol.
Catal. 66, L23 (1991).
9. Aben, P. C., J. Catal. 10, 224 (1968).
10. Sermon, P. A., J. Catal. 24, 460 (1972).
11. Bond, G. C., and Wells, P. B., Appl. Catal. 18, 225 (1985).
12. Candy, J. P., Fouilloux, P., and Renouprez, A. J., J. Chem. Soc. Fara-
day 1 76, 616 (1980).
13. Kunimori, K., Uchijima, T., Yamada, M., Matsumoto, H., Hattori, T.,
and Murakami, Y., Appl. Catal. 4, 67 (1982).
Pt/SiO2 exhibited a low chemoselectivity (24% ), but its
TON was relatively high (42 103 s 1). Pts{Sn}n/SiO2 cata-
lysts, prepared by the interaction of Sn(CH3)4 with reduced
Pt/SiO2 under H2 at room temperature, exhibited a signifi-
cant increase of chemoselectivity compared to Pt/SiO2, but
14. Martin, G. A., Rev. Phys. Appl. 16, 181 (1981).
15. Fried, J., and Edwards, J. A., “Organic Reactions in Steroid Chem-
istry.” Van Nostrand Reinhold Co., New York, 1972.
16. Rylander, P. N., “Catalytic Hydrogenation over Platinum Metals.”
Academic Press, New York, 1967.
still a low diastereoselectivity to the desired 21(S)-OH un- 17. Rylander, P. N., “Catalytic Hydrogenation in Organic Synthesis.”
Academic Press, New York, 1979.
18. Augustine, R. L., “Heterogeneous Catalysis for the Synthetic
Chemist.” Marcel Dekker, New York, 1995.
19. Gertosio, V., Santini, C. C., Basset, J.-M., Bayard, F., Buendia, J., and
saturated ketoalcohol, 2. In parallel, the diastereoselectiv-
ity at the C3 carbonyl group decreased from 100% to 34% .
The addition of ( )cinchonidine and hydrocinchonidine on
the platinum–tin catalysts sharply increased the diastereos-
Vivat, M., J. Mol. Catal. A 142, 141 (1999).
electivity at C21 in favor of the 21(S)-OH unsaturated ke- 20. Delbecq, F., and Sautet, P., J. Catal. 164, 152 (1996).
21. Gallezot, P., in “Handbook of Heterogeneous Catalysis” (G. Ertl,
toalcohol from 14% to 70% without changing the chemos-
electivity.
The experimental conditions to achieve the best chemo-,
H. Kno¨zinger, and J. Weitkamp, Eds.), Vol. 5, p. 2209. Wiley-VCH,
Weinheim, 1997.
22. Haller, G. I., and Resasco, D. E., Adv. Catal. 36, 173 (1989).
regio-, and diastereoselectivities were found to be the
following: H2 pressure of 80 atm, temperature of 0 C, ratio
Sn/surface platinum of 0.5, and ratio cinchonidine/surface
platinum of 4.6. In these conditions, oxopromegestone was
hydrogenated into Trimegestone with 91% chemoselectiv-
ity, 64% regioselectivity, and 72% diastereoselectivity at
C21.
In the case of Trimegestone, these very promising re-
sults are still below those of the enzymatic catalysis. But
it has been demonstrated that the limiting step which is
the overreduction in hydrogenation of polyfunctional sub-
23. Basset, J.-M., Candy, J.-P., and Santini, C. C., in “Transition Metals
for Organic Synthesis” (M. Beller and C. Bolm, Eds.), Vol. 2, p. 387.
J. Wiley, Weinheim, 1998.
24. Didillon, B., Candy, J. P., El Mansour, A., Houtman, C., and
Basset, J. M., J. Mol. Catal. 74, 43 (1992).
25. Humblot, F., Cordonnier, M. A., Santini, C. C., Didillon, B., Candy, J. P.,
and Basset, J. M., in “Studies in Surface Sciences and Catalysis”
(H.-U. Blaser, A. Baiker, and R. Prins, Eds.), Vol. 108, p. 289.
Elsevier, Amsterdam, 1997.
26. Studer, M., Okafor, V., and Blaser, H.-U., J. Chem. Soc. Chem. Com-
mun. 1053 (1998).
27. Lefebvre, F., Candy, J.-P., Santini, C. C., and Basset, J.-M., Top. Catal.
4, 211 (1997).
strates by a heterogeneous catalyst could be partially sup- 28. Cordonnier, M. A., Ph.D. thesis, Universite´ Claude Bernard, Lyon I:
Lyon, 1994.
pressed by the use of a chemoselective catalyst such as
29. Lamy-Pitara, E., and Barbier, J., Appl. Catal. A 149, 49 (1997).
30. Rodriguez, J. A., Chatuverdi, S., Jirsak, T., and Hrbek, J., J. Chem.
Pts{Sn}0.5/SiO2. The significant influence on the stereodif-
ferentiation of several modifiers is still not well understood.
Phys. 109, 4052 (1998).
Nevertheless, we have shown that a double asymmetric in-
duction with a chiral ligand can be obtained in heteroge-
31. Passos, F. B., Schmal, M., and Vannica, M. A., J. Catal. 160, 118
(1996).

HYDROGENATION OF OXOPROMEGESTONE INTO TRIMEGESTONE
373
32. Delbecq, F., and Sautet, P., J. Catal. 152, 217 (1995).
33. Windham, R. G., and Koel, B. E., Langmuir 4, 1113 (1998).
lysis” (G. Ertl, H. Kno¨zinger, and J. Weitkamp, Eds.), Vol. 5, p. 2422.
Wiley-VCH, Weinheim, 1997.
34. Campbell, J. M., Seimanides, S., and Campbell, C. T., J. Phys. Chem. 38. Baiker, A., J. Mol. Catal. 115, 473 (1997).
93, 2815 (1989).
39. Blaser, H.-U., and Jalett, H. P., in “Studies in Surface Sciences
35. Vermeer, W. A. H., Fulford, A., Johnston, P., and Wells, P. B., J. Chem.
Soc. Chem. Commun. 1053 (1993).
36. Eliel, E. L., and Wilen, S. H., “Stereochemistry of Organic Com-
pounds.” J. Wiley, New York, 1994.
and Catalysis” (M. Guisnet, J. Barbier, J. Barrault, C. Bouchoule,
D. Duprez, G. Pe´rot, and C. Montassier, Eds.), Vol. 78, p. 139.
Elsevier, Amsterdam, 1993.
40. Margitfalvi, J. L., Marti, P., Baiker, A., Botz, L., and Sticher, O., Catal.
Lett. 10, 281 (1990).
37. Baiker, A., and Blaser, H. U., in “Handbook of Heterogeneous Cata-