Mild generation of o-quinodimethanes via fluoride induced 1,4-elimination of α-(o-trimethylsilylmethyl)benzylesters: stereoselective synthesis of 19-nor steroids and RU486 precursors

Article abstract of DOI:10.1016/j.tetlet.2006.04.129

19-Nor steroids and RU486 tricyclic synthetic intermediates were stereoselectively prepared by an intramolecular Diels-Alder reaction involving an o-quinodimethane possessing a pro-17 unique chiral stereocenter, substituted by a protected hydroxyl group.

Full text of DOI:10.1016/j.tetlet.2006.04.129

Tetrahedron Letters 47 (2006) 4671–4675
Mild generation of o-quinodimethanes via fluoride induced
1,4-elimination of a-(o-trimethylsilylmethyl)benzylesters:
stereoselective synthesis of 19-nor steroids and RU486 precursors
Marc Port and Robert Lett*
Unite´ Mixte CNRS-ROUSSEL UCLAF (UMR 26) 102, route de Noisy, 93235 Romainville, France
Received 23 March 2006; revised 19 April 2006; accepted 26 April 2006
Available online 22 May 2006
Abstract—19-Nor steroids and RU486 tricyclic synthetic intermediates were stereoselectively prepared by an intramolecular Diels–
Alder reaction involving an o-quinodimethane possessing a pro-17 unique chiral stereocenter, substituted by a protected hydroxyl
group. The o-quinodimethane was generated in mild conditions by fluoride induced 1,4-elimination of a-(o-trimethylsilyl-
methyl)benzylesters and the present methodology allows a flexible access to a,a0-disubstituted o-quinodimethanes, as shown by
the 11b-substituted steroid approach. The IMDA diastereoselections reported herein were highly dependent on the nature of the
hydroxyl protective group and the diastereoselectivities superior to those observed with the thermolysis of the corresponding
benzocyclobutenes.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
from the opening of an epoxide, triggered by fluoride
cleavage of the C–Si bond in diglyme at 27 °C, to obtain
the desired IMDA cycloadduct in 70% yield.5 Hence, for
an asymmetric synthesis of 19-norsteroids and 11b-
substituted steroids, we examined another route involv-
ing a fluoride anion induced 1,4-elimination with an
ester as a leaving group, which might have a greater
flexibility for the substitution of the aromatic part of
the precursor and for the substitution at the a,a0 posi-
tions by an alkyl or an aryl substituent.6 Thus, also
based on the pioneering work of Kametani and Fuku-
moto,7 and starting from a precursor 3 for generating an
o-quinodimethane 2 having a unique chiral stereocenter
at pro-17, our aim was to see if we might obtain effi-
ciently and highly stereoselectively the desired enantio-
pure cycloadduct 1 from a precursor such as 3
(Scheme 1). The required absolute configurations of
the C11, C13, and C14 stereocenters of a steroid might
be obtained from a protected hydroxyl group of S con-
figuration via transition state i, or from the enantiomer
R via the enantiomeric transition state of ii. Hence a pre-
cursor, substituted at the pro-11 position, should thus
afford the 11b-substituted cycloadduct, provided one
of the two transition states, i or ii, is sufficiently favored
with the proper configuration of the unique chiral ste-
reocenter (Scheme 2). In this note, we report our results
concerning the mild and efficient generation of o-quino-
dimethanes from precursors such as 3, easily obtained
Intramolecular cycloadditions of o-quinodimethane
intermediates have been extensively studied by many
groups in total synthesis, and especially in that of
steroids.1–3 Although different routes for generating
in situ, the o-quinodimethane intermediate have shown
quite a potential, it might be worth trying to simplify
the access to its precursor and to make it more flexible
with regard to its structure, especially for the prepara-
tion of a,a0-disubstituted o-quinodimethanes. It might
also be useful to generate the o-quinodimethane in
milder conditions than those required for the thermoly-
sis of benzocyclobutenes or that of cyclic sulfones.
Thus, we were particularly interested by the mild conditi-
ons (acetonitrile, 50 °C) developed by Saegusa, for fluo-
ride ion induced 1,4-elimination of o-(a-trimethylsilyl-
alkyl)benzyltrimethylammonium salts.4 Moreover, quite
remarkably in a synthesis of (d,l)-11a-hydroxyestrone,
Magnus generated the required o-xylylene intermediate
*
Corresponding author at present address: ICMMO, UMR CNRS
8182, baˆtiment 410, Universite´ de Paris-Sud, 91405 Orsay Cedex,
France. Tel.: +33 1 69 15 63 03; fax: +33 1 69 15 46 79; e-mail:
robert.lett@icmo.u-psud.fr
Present address: Guerbet Research, BP 57400, 95943 Roissy CDG
Cedex, France.
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.04.129

4672
M. Port, R. Lett / Tetrahedron Letters 47 (2006) 4671–4675
pro-17
pro-14
OR
S
R'
S
CHO
Met OR
OR
OR
MeO
5
OCOR''
R'
H
+
MeO
MeO
MeO
1
2
3
4
SiMe3
SiMe3
R'
R' = H, alkyl, aryl
pro-11
Scheme 1.
OMe
OMe
H
RO
H
H
Me
H
S
H
R'
R'
H
RO
H
Me
H
H
Me
R'
H
H
S
S
H
R'
H
H
OR iii
H
H
H
H
Me
H
i
ii
iv
H
H
OMe
S
OMe
H
H
H
OR
trans syn adduct
trans anti adduct
cis anti adduct
cis syn adduct
Scheme 2.
from 4 and 5 (Scheme 1), and also the diastereoselectiv-
ities obtained in such cycloaddition conditions with a
racemic precursor in preliminary studies.6
tained after flash chromatography in 65% overall yield
from 6, with no intermediate purification. Reaction of
the lithio derivative prepared from iodide 10, via halo-
gen–lithium exchange,8 with aldehyde 45,9 (0.95 equiv)
afforded intermediate 11 in 91% yield, as a 60/40 mix-
ture of two diastereoisomers. Further benzoylation
(93%) and hydrolysis in mild acidic conditions gave 13
as a 60/40 mixture of diastereoisomers which were not
separated (95.5%). It is worth pointing out that the
choice of that pro-17 hydroxyl protective group, in addi-
tion to the advantages of its simplicity and of not intro-
ducing a supplementary stereogenic center, gave far
better results than a THP ether, allowing in the hydro-
lysis step to completely avoid the formation of a five-
membered cyclic ether which was always formed with
the THP ether in different acidic conditions.6
2. Convergent synthesis of the o-quinodimethane
precursors (Scheme 3)
In order to also further examine the influence of the nat-
ure of the hydroxyl protective group on the stereoselec-
tivity of the intramolecular cycloaddition,7 a convergent
synthesis leading to the key-intermediate 13 was
achieved, involving the protection of the pro-17 hydr-
oxyl as a mixed ketal with 2-methoxypropene for the
whole sequence 6 (prepared in three steps from 4-pen-
ten-1-ol, 69% overall yield) to 12. Iodide 10 was thus ob-
OH
O
OMe
O
OMe
O
OMe
d
a
b
c
TBSO
TBSO
HO
MsO
6
7
8
9
O
OH
O
OMe
OMe
O
OMe
OH
OCOPh
OCOPh
e
f
g
I
MeO
MeO
MeO
SiMe3
SiMe3
SiMe3
10
11
12
13
65% overall from 6
91%
93%
95.5%
Scheme 3. Reagents and conditions: (a) POCl3 (1 mol %), 2-methoxypropene (1.5 equiv), CH2Cl2, rt; (b) TBAF 1 M (1.1 equiv), THF, rt; (c) MsCl
(1.13 equiv), NEt3 (1.2 equiv), i-PrNEt2 (0.25 equiv), CH2Cl2, 0 °C!rt; (d) NaI (5 equiv), i-PrNEt2 (0.2 equiv), acetone, rt; (e) 10, t-BuLi (1.9 equiv),
Et2O–pentane, ꢀ78 °C, 3 min, then 4 (0.95 equiv), ꢀ78 °C!rt; (f) PhCOCl (2 equiv), DMAP cat., pyridine, rt and (g) PPTS (0.1 equiv), EtOH, rt.

M. Port, R. Lett / Tetrahedron Letters 47 (2006) 4671–4675
4673
3. In situ o-quinodimethane formation and intramolecular
cycloaddition
(89%), DMF (97.5%), and DMPU (93.5%), at 80 °C.
However conversion was much slower in MeCN with
the stannate as a source of fluoride, than with CsF. Con-
centration of the precursor 3 (R = MEM, R0 = H,
R00 = Me or Ph) was always found to be quite optimal
at 0.05 M to obtain good yields of the cycloadducts
and a higher concentration (0.5 M) led to a significantly
lower yield (R00 = Me, 42.5%), due to intermolecular
reactions and dimerization.
Quite surprisingly, an ester as a leaving group was never
really examined before this work.6 We only found an
isolated previous example of fluoride induced elimina-
tion on o-(acetoxymethyl)benzyltrimethylsilane10 and
another one on 2-(acetoxymethyl)-3-(triisopropylsilyl-
methyl)-benzofuran11 when this study was initiated. With
derivatives of 3 (R = MEM, R0 = H), we first examined
different conditions in order to get an efficient in situ
generation of the o-quinodimethane intermediate and
IMDA cycloadducts formation. Noteworthy, no signifi-
cant difference in reactivity was observed for the two
diastereoisomers, which are epimeric at pro-14, as shown
by following their conversion, consistently with earlier
results of Saegusa in a parent methodology.4 In similar
anhydrous conditions (CsF (10 equiv), MeCN, 80 °C),
for 3 (R = MEM, R0 = H, 0.05 M), the global yields
of cycloadducts were quite comparable (67.5–71%)
with the acetate, benzoate, and p-methoxybenzoate,
but significantly lower with the p-nitrobenzoate (30%).
No significant difference of IMDA stereoselectivity
was observed for all those esters.
With 3 (R = MEM, R0 = H, R00 = Me or Ph, 0.05 M)
and CsF (2 equiv), at 80 °C, polar solvents like DMF,
sulfolane, and DMPU gave faster conversions than
MeCN and good yields of cycloadducts. With respect
to MeCN, at 80 °C, other solvents like THF, DME or
diglyme gave slower reactions and lower yields. On the
other hand, 3 (R = MEM, R0 = H, R00 = Ph) gave with
CsF (2 equiv) in DMSO, at 80 °C, a much faster reac-
tion, but those reaction conditions led to much inter-
molecular reactions at 0.05 M and to a lower yield of
cycloadducts (44%). No solvent effect or influence of
the source of the fluoride was observed on IMDA stereo-
selectivity, at the same temperature.
Benzocyclobutenes were not observed and shown not to
be intermediates. It is worth to point out that, for the
TBS ether 14K, the yield of cycloadducts was signifi-
Concerning the source of the fluoride, TBAF/THF was
not satisfactory, leading to protodesilylation products,
due to the difficulty to get anhydrous TBAF.12 Anhy-
drous CsF, which is easy to obtain (200 °C, 0.8 mmHg)
gave better results than KF in MeCN. However, KF or
CsF (2 equiv) gave comparable yields of cycloadducts in
sulfolane (89.5% and 92%). The stannate [n-Bu4N+,
Ph3SnF2ꢀ],13 which is soluble in many organic solvents
and not hygroscopic, gave good yields of cycloadducts
from 3 (R = MEM, R0 = H, R00 = Ph, 0.05 M) in MeCN
cantly improved with the stannate [n-Bu4N+, Ph2SnF ꢀꢁ
2
with respect to CsF (Table 1). Also noteworthy, alcohol
13 led to much protodesilylation (21%) and the corre-
sponding lithium alkoxide here was unsatisfactory.6
With 14B (0.05 M, toluene, 80 °C, 23 h), potassium tri-
methylsilanolate (1.2 equiv) was found to yield only 15.5%
of cycloadducts and the major product (68%) resulted
from a Brook rearrangement of the pro-11 alcoholate
Table 1. Yields and diastereoselectivity of the Diels–Alder reaction from the o-quinodimethane generated in situ
OR
OR
OR
OR
OR CsF(10 equiv)
OCOPh
H
H
H
H
MeCN
+
+
+
80 °C, 1.5 h
MeO
MeO
MeO
MeO
cis anti adduct
relative ratios of diastereoisomers
MeO
14
(0.05M)
(d,l) 15
(d,l) 16
(d,l) 18
(d,l) 17
SiMe3
trans syn adduct trans anti adduct
cis syn adduct
cycloadducts yield
(from 14)
R = H
R = MEM
R = MOM
R = DMPM
R = THP
R = CMe2(OMe)
R = COCF3
R = p-NO2-C6H4-CO
R = p-MeO-C6H4-CO
R = Piv
13
20.5%
82%
63.5%
58%
65%
96%
15A
36%
58%
58%
63%
55%
46%
55%
48%
41%
37%
26%
27.5%
16A
16B
16C
16D
16E
16F
16G
16H
16I
55%
29%
29%
25%
33%
45%
23%
32%
40%
48%
64%
62%
17A
17B
17C
3%
2%
2%
18A
18B
18C
6%
11%
11%
14B
14C
14D
14E
14F
14G
14H
14I
15B
15C
15D
15E
15F
15G
15H
15I
17D + 18D 12%a
17E
17F
17G
17H
17I
2%
2%
4%
7%
8%
4%
4%
4.5%
18E
18F
18G
18H
18I
9%
7%
80%
18%
13%
11%
10%
5%
23.5%
68.5%
70.5%
69%
14J
14K
15J
15K
16J
16K
17J
17K
18J
18K
R = TBS
80.5%b
5.5%
a Deprotection of the DMPM (3,4-dimethoxybenzyl) protective group was here not achieved, and the ratio of the two cis-fused could not be
determined.
b 14K (0.05 M), [n-Bu4N+, Ph3SnF2ꢀ] (2 equiv), anhyd DMF, 80 °C, 45 h.

4674
M. Port, R. Lett / Tetrahedron Letters 47 (2006) 4671–4675
formed in situ. On the other hand, no o-quinodimethane
was formed from 3 (R = MEM, R0 = H, R00 = Me) in
the conditions reported by Mann (CCl4, SiO2, 80 °C).14
results, with those later obtained,9 clearly show that the
IMDA stereoselectivity is at present difficult to
understand.
4. Pro-17 hydroxyl protective group influence on IMDA
diastereoselectivity (Table 1)
5. Temperature effects on diastereoselectivity
As anticipated for an IMDA, there was only a slight
effect of the temperature on diastereoselectivity and the
conversions were too slow at 25 °C to be of preparative
interest, although the stereoselectivity was then
improved (Supplementary data).6
Using standard conditions, the pro-17 hydroxyl was
protected very efficiently by different groups. For
each IMDA, the stereoselectivities were determined on
15A, 16A, 17A, and 18A, after cleavage of the hydroxyl
protective group on the crude product, by 1H NMR
integration of the distinct angular methyl group for each
diastereomer (200 MHz, CDCl3) and by HPLC
(Lichrosorb Si 60 lm; eluent: 11:89:0.2 EtOAc/hexane/
AcOH; flow rate: 2 mL/min; tR (min): 16A, 13.66; 18A,
14.75; 17A, 16.38; 15A, 18.21). Structures of the cycload-
ducts were unambiguously assigned by identification
of 15A with a reference sample provided by Roussel
Uclaf, and by chemical filiation.15 The results are given
in Table 1 and they were shown to be kinetic results,
since we checked that diastereoselectivity was the same
at different conversions of the precursor and that
isolated cycloadducts remained unchanged in reaction
conditions.6
6. a,a0-disubstituted o-quinodimethanes: a synthetic
approach of RU486 (mifepristone) (Scheme 4)
In preliminary studies concerning a new synthetic
approach of RU486 or more generally of 11b-substituted
steroids, we first examined on racemic intermediates, pre-
pared from (d,l)-6, if we might extend our methodology
to the preparation of a,a0-disubstituted o-quinodime-
thanes and achieve the corresponding IMDA to afford
stereoselectively the desired synthon 26. Thus, the pre-
cursor 25 was obtained as a mixture of diastereoisomers,
starting from p-N,N-dimethylamino-benzaldehyde.6
Further reaction of 25 (0.05 M) with CsF (12 equiv) in
CH3CN, at 80 °C for 2 h gave only 56% of the cyclo-
adducts 26, 27, 28 in a ratio 61.5:31:7.5, and 37% of
o,o0-quinodimethane dimers were isolated. Best results
were obtained by the reaction of 25 (0.05 M) with
As already observed in the thermolysis of benzocyclo-
butenes by Kametani and Fukumoto,7 and now in the
present work,6,15 IMDA diastereoselection is highly
dependent on the nature of the protective group of the
pro-17 hydroxyl as a unique chiral stereocenter of the
generated o-quinodimethane. With respect to the results
obtained by thermolysis at 180 °C of the correspond-
ing benzocyclobutenes, diastereoselectivity is slightly
improved by in situ generation of the o-quinodimethane
at 80 °C from precursors such as 3.6,15 From each
precursor, the trans-hydrindanes are the major cyclo-
adducts and the configuration of the major trans-
hydrindane cycloadduct can be changed depending on
the pro-17 hydroxyl protective group, as already
observed for silyl ethers by Kametani and Fukumoto,7
but also with other protective groups such as here the
pivalate 14J or the t-butyl ether.9 The formation of the
cis anti adduct 17 can be expected as a minor cyclo-
adduct, and rationalized to some extent by considering
the strong steric interactions of the OR group with the
o-quinodimethane subunit involved in transition state
iii (Scheme 2). With respect to the trans syn and trans
anti cycloadducts, the rationalization of the nature of
the major isomer appears to be quite difficult and these
[n-Bu4N+, Ph3SnFꢀ] in anhydrous DMF, at 80 °C for
2
48 h, which afforded in 87% yield the cycloadducts 26,
27, and 28 as a 60.5:32.5:7 mixture, and no dimers were
formed in those conditions. Either with the stannate or
with CsF, no difference of reactivity between the diaste-
reoisomers of 25 was noticed and no traces of the fourth
possible cycloadduct were observed. The structures of
the cycloadducts 26, 27, and 28 were assigned unambig-
1
13
uously by H and C NMR.6
As a conclusion, the present methodology is quite simple
and flexible for the in situ generation of o-quinodi-
methanes from precursors such as 3. Quite remarkably,
the relative rates for the formation of the intermediate o-
quinodimethane and for the intramolecular cycloaddi-
tion remain compatible from 25 to 180 °C to afford good
yields of the cycloadducts, with no dimerization or inter-
molecular reactions in the appropriate reaction condi-
tions reported herein. Precursors such as 3 were also
shown in intermolecular Diels–Alder reactions to give re-
OMEM
N
N
N
OMEM
OMEM
OMEM
OAc
+
+
87%
SiMe3
H
H
H
MeO
25
(d,l)-26
(d,l)-27
(d,l)-28
MeO
MeO
MeO
Ar
Ar = p-N,N-dimethylamino-phenyl
26 : 27 : 28 = 60.5 : 32.5 : 7
Scheme 4. Reagents and conditions: [n-Bu4N+, Ph3SnFꢀ] (2 equiv), anhyd DMF, 80 °C, 48 h.
2

M. Port, R. Lett / Tetrahedron Letters 47 (2006) 4671–4675
4675
3. Michellys, P.-Y.; Pellissier, H.; Santelli, M. Org. Prep.
Proc. Int. 1996, 28, 545–608.
sults quite comparable to those obtained by the method-
ology developed by Ito and Saegusa,4a,c with the same
dienophiles.9 During the completion of this work and
after it, some related results were also reported for fluo-
ride-induced elimination of an ester14,16 or a sulfone.17
4. (a) Ito, Y.; Nakatsuka, M.; Saegusa, T. J. Am. Chem. Soc.
1980, 102, 863–865; (b) Ito, Y.; Nakatsuka, M.; Saegusa,
T. J. Am. Chem. Soc. 1981, 103, 476–477; (c) Ito, Y.;
Nakatsuka, M.; Saegusa, T. J. Am. Chem. Soc. 1982, 104,
7609–7622.
The present method involving o-quinodimethane pre-
cursors such as 3 was also applied to the asymmetric
synthesis of 19-nor steroids such as enantiopure trenbo-
lone acetate, starting from enantiopure (S)-6.6 As shown
also by our synthetic approach of RU486 (mifepristone)
from 3 (R = MEM, R0 = p-NMe2-C6H4-, R00 = Me), the
methodology described herein gives a good access to
a,a0-disubstituted o-quinodimethanes which are difficult
to synthetize by other methods.1–4
5. Djuric, S.; Sarkar, T.; Magnus, P. J. Am. Chem. Soc. 1980,
102, 6885–6886.
6. Port, M. Ph.D. Thesis, Paris VI University, 1995.
7. (a) Kametani, T.; Matsumoto, H.; Honda, T.; Fukumoto,
K. Tetrahedron Lett. 1980, 4847–4848; (b) Kametani, T.;
Matsumoto, H.; Honda, T.; Nagai, M.; Fukumoto, K.
Tetrahedron 1981, 37, 2555–2560; (c) Nemoto, H.;
Matsuhashi, N.; Fukumoto, K. J. Chem. Soc., Chem.
Commun. 1991, 705–706; (d) Nemoto, H.; Matsuhashi, N.;
Satoh, A.; Fukumoto, K. J. Chem. Soc., Perkin Trans. 1
1992, 495–498.
8. (a) Corey, E. J.; Beames, D. J. J. Am. Chem. Soc. 1972, 94,
7210–7211; (b) Seebach, D.; Neumann, H. Chem. Ber.
1974, 107, 847–853.
9. Delpech, B.; Lett, R., unpublished results.
10. Askari, S.; Lee, S.; Perkins, R. R.; Scheffer, J. R. Canad.
J. Chem. 1985, 63, 3526–3529.
Acknowledgments
We thank the staff of the Analytical Department of the
Roussel Uclaf Research Center at Romainville and also
Roussel Uclaf and the CNRS for a Ph.D. Grant to M.
Port, the Direction des Recherches Chimiques of Rous-
sel Uclaf for support of this work.
11. Bedford, S. B.; Begley, M. J.; Cornwall, P.; Knight, D. W.
Synlett 1991, 627–629.
12. (a) Cox, D. P.; Terpinski, J.; Lawrynowicz, W. J. Org.
Chem. 1984, 49, 3216–3219; (b) Clark, J. H. Chem. Rev.
1980, 80, 429–452; Clark, J. H.; Smith, D. K. Tetrahedron
Lett. 1985, 26, 2233–2236.
Supplementary data
13. (a) Gingras, M. Tetrahedron Lett. 1991, 50, 7381–7384; (b)
Sau, A. C.; Carpino, L. A.; Holmes, R. R. J. Organomet.
Chem. 1980, 197, 181–197.
14. Beard, A. R.; Hazell, S. J.; Mann, J.; Palmer, C. J. Chem.
Soc., Perkin Trans. 1 1993, 1235–1238.
Temperature effects (25–180 °C) on IMDA yield and
diastereoselectivity are given for 14B (R = MEM,
R00 = Ph) and 14K (R = TBS, R00 = Ph). Supplementary
data associated with this article can be found, in the
online version, at doi:10.1016/j.tetlet.2006.04.129.
15. Port, M.; Lett, R. Tetrahedron Lett., accompanying
paper, doi: 10.1016/j.tetlet.2006.04.128.
16. (a) Bedford, S. B.; Begley, M. J.; Cornwall, P.; Foreman,
J. P.; Knight, D. W. J. Chem. Soc., Perkin Trans. 1 1996,
2827–2832; (b) Leung, M.; Trahanovsky, W. S. J. Am.
Chem. Soc. 1995, 117, 841–851; (c) Fleming, I.; Morgan, I.
T.; Sarkar, A. K. J. Chem. Soc., Perkin Trans. 1 1998,
2749–2763.
17. (a) Lenihan, B. D.; Shechter, H. Tetrahedron Lett. 1994,
35, 7505–7508; (b) Lenihan, B. D.; Shechter, H. J. Org.
Chem. 1998, 63, 2072–2085; (c) Lenihan, B. D.; Shechter,
H. J. Org. Chem. 1998, 63, 2086–2093.
References and notes
1. (a) Kametani, T.; Nemoto, H.; Ishikawa, H.; Shiroyama,
K.; Fukumoto, K. J. Am. Chem. Soc. 1976, 98, 3378–3379;
(b) Kametani, T.; Nemoto, H. Tetrahedron 1981, 37, 3–16;
(c) Nemoto, H.; Fukumoto, K. Tetrahedron 1998, 54,
5425–5464.
2. (a) Oppolzer, W. Synthesis 1978, 793–802; (b) Oppolzer,
W. Heterocycles 1980, 1615–1630.