A diastereoselective switch in the access to isobenzofuran-derived α-selenoesters

Article abstract of DOI:10.1021/ol005560h

(equation presented) The cycloaddition of 1-phenylisobenzofuran (PIBF) with methyl acrylate yields, in a moderate endo/exo ratio, the expected oxa-bridged adduct, which can be deprotonated and condensed on diphenyl diselenide to provide, in a stereoconver

Full text of DOI:10.1021/ol005560h

ORGANIC
LETTERS
2000
Vol. 2, No. 7
923-925
A Diastereoselective Switch in the
Access to Isobenzofuran-Derived
r-Selenoesters
,†
Ce´line Martin,† Patrick Mailliet,‡ and Jacques Maddaluno*
Laboratoire des Fonctions Azote´es et Oxyge´ne´es Complexes de l'IRCOF,
UniVersite´ de Rouen, 76821 Mont St Aignan Cedex, France, and Rhoˆne-Poulenc
Rorer, Centre de Recherche de Vitry-AlfortVille, 13 quai Jules Guesde,
94403 Vitry-sur-Seine Cedex, France
jmaddalu@crihan.fr
Received January 17, 2000
ABSTRACT
The cycloaddition of 1-phenylisobenzofuran (PIBF) with methyl acrylate yields, in a moderate endo/exo ratio, the expected oxa-bridged adduct,
which can be deprotonated and condensed on diphenyl diselenide to provide, in a stereoconvergent step, the “endo” r-selenoester. Its “exo”
epimer is obtained by reacting PIBF and methyl r-phenylselenoacrylate. These adducts can be oxidized to give a common unsaturated bridged
ester that can react with an imminium ylide to provide the expected pyrrolidine stereoselectively.
Farnesyltransferase inhibitors have opened promising new
directions in cancer therapy lately, particularly in the case
of tumors associated to Ras protein mutations.1 The poly-
cyclic inhibitor RPR 115135, which exhibits remarkably low
geranylgeranyltransferase inhibition profiles while keeping
significant in vivo activities,2 constitutes a particularly
inspiring starting point for a focused study on the structure-
activity relationship in the benzo[f]perhydroisoindole (BPHI)
series. The possible importance of both the carbon bridge in
the BPHI structure and the position of the ester group for
the drug-enzyme interaction made analogue 1 an appealing
targets. Unsaturated ester 2 cannot be prepared by a direct
cycloaddition between 1-phenylisobenzofuran 5 and methyl
propiolate which, under classical thermal conditions, instead
leads to a complex mixture of unidentified compounds.3
Thus, an alternative retrosynthetic approach of the polycyclic
skeleton is disclosed in Figure 1. The pyrrolidine ring in 1
can result from a 1,3-dipolar cycloaddition between 2 and
Sakurai’s dipole.4 The synthesis and oxidation of the
corresponding R-selenoesters 3 thus offered an alternative
route to 2. We describe here two methods giving a totally
stereoselective access to both isomers of the key intermediate
3.
† Universite´ de Rouen.
‡ Rhoˆne-Poulenc Rorer.
(1) Bos, J. L. Cancer Res. 1989, 49, 4682.
Selenoesters such as 3 are classically prepared by depro-
tonation and selenation of the corresponding esters. In our
(2) Mailliet, P.; Laoui, A.; Bourzat, J. D.; Capet, M.; Cheve´, M.;
Commerc¸on, A.; Dereu, N.; Lebrun, A.; Martin, J. P.; Peyronel, J. F.;
Salagnad, C.; Thompson, F.; Zucco, M.; Guitton, J. D.; Pantel, G.; Bissery,
M. C.; Brealey, C.; Lavayre, J.; Lelie`vre, Y.; Riou, J. F.; Vrignaud, P.;
Duchesne, M.; Lavelle, F. In Farnesyltransferase and Geranylgeranyl-
transferase Inhibitors in Cancer and CardioVasculartherapy; Sebte, S., Ed.;
Humana Press: Totowa, NJ, in press.
(3) Martin, C.; Mailliet, P.; Maddaluno, J. Full paper in preparation.
(4) (a) Hosomi, A.; Sakata, Y.; Sakurai, H. Chem. Lett. 1984, 1117. (b)
Terao, Y.; Kotaki, H.; Imai, N.; Achiwa, K. Chem. Pharm. Bull. 1985, 33,
2762.
10.1021/ol005560h CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/11/2000

The deprotonation step in the R position of such an oxa-
bridged system could also be troublesome. Rodrigo et al.
have indeed shown that the deprotonation of ester 7 by LDA
triggers an instantaneous opening of the bridge, leading to
alcohol 8.11 However, the same authors have shown that if
the deprotonation takes place in the presence of an electro-
phile such as D2O, the R-deuterated bridged esters 9 are
recovered in almost quantitative yields (Scheme 2).
Scheme 2a
Figure 1. Retrosynthetic route to inhibitor analogue 1.
case, the precursor 6 (Scheme 1) has been obtained by a [4
+ 2] cycloaddition between 1-phenylisobenzofuran 5 and
acrylates. PIBF 5 itself was handily prepared in situ by
n-butyllithium-induced δ-elimination5 on phthalane 4.6
a (a) LDA, THF, 0 °C; (b) MeONa, MeOD, 20 °C.
Scheme 1a
In our case, LDA appeared rather inefficient, the starting
esters 6 being recovered unaltered. By contrast, KHMDS
yielded the expected selenoester provided that diphenyl
diselenide was added to the medium before deprotonation
(Scheme 3). It is worth noting that both 6a and 6b lead to
Scheme 3a
a (a) 2 equiv of n-BuLi, Et2O, rt, 15 min then t-BuOH quench;
(b) methyl or tert-butyl acrylate, THF, rfx, 2 h.
The thermal cycloaddition of 5 with methyl acrylate
provided 6 as a single regioisomer but as a 38:62 endo/exo
mixture of stereoisomers.7 The regioselectivity was as
expected given previous results on related systems obtained
by Gilchrist8 or, more recently, by Iwasaki9 and fit our own
AM1 calculations on the HOMO coefficients of 5.3
Because the lack of endoselectivity of this step is possibly
a drawback for the selenation, we tried to further tune a few
structural parameters. However, when replacing the phenyl
group by an o-tolyl group, the selectivity is marginally
affected (44:56 in favor of exo 6c); resorting to tert-butyl
acrylate hardly reverses this ratio (57:43 in favor of endo10
6b this time).
(a) 2 equiv of (PhSe)2 and then 2 equiv of KHMDS, THF, rt.
the same “endo” epimer 11a. However, the kinetics of their
transformation are rather different: after 30 min at room
temperature, the reaction is total with 6a while the conversion
hardly reaches 40% with 6b; the relative ease of access to
the acidic proton in 6a and 6b are probably at the origin of
this phenomenon.
The mixture of 6 isomers can, of course, be treated without
prior separation provided that the reaction time is long
enough. The stereoconvergence of the selenation is likely
due to the intermediate formation of the same (except,
(5) Tobia, D.; Rickborn, B. J. Org. Chem. 1986, 51, 3849.
(6) Plaumann, H. P.; Smith, J. G.; Rodrigo, R. J. Chem. Soc., Chem.
Commun. 1980, 534.
(7) An exo-preference has been reported for similar systems. See for
instance: (a) Nozaki, H.; Yamaguti, T.; Ueda, S. Kondo, K. Tetrahedron
1968, 24, 1445. (b) Hickmott, P. W.; Simpson, R. J. Chem. Soc. (C) 1972,
302. (c) Rodrigo, R.; Knabe, S. M. J. Org. Chem. 1986, 51, 3973.
(8) Faragher, R.; Gilchrist, T. L. J. Chem. Soc., Perkin Trans. 1 1976,
336.
(10) As expected according to Mellor, J. M.; Webb, C. F. J. Chem. Soc.,
Perkin Trans. 2 1974, 17-22.
(11) Keay, B.; Rajapaksa, D.; Rodrigo, R. Can. J. Chem. 1984, 62, 1093.
(9) Sugahara, M.; Moritani, Y.; Terakawa, Y.; Ogiku, T.; Ukita, T.;
Iwasaki, T. Tetrahedron Lett. 1998, 39, 1377.
924
Org. Lett., Vol. 2, No. 7, 2000

eventually, for the E(O)/Z(O) ratio) potassium enolate 10,
which reacts with the electrophile on its only accessible
convex face. Thus, the selenoester is finally obtained as a
single diastereomer 11a (bearing the ester group endo). This
compound appeared relatively unstable on silica gel and is
therefore difficult to purify under standard chromatographic
conditions. It can, however, be used as a crude mixture in
the following oxidation step. This instability could be related
to the steric constraints introduced by the selenium atom;
adduct 6b was also observed to be much more difficult to
purify by chromatography than its analogue 6a.
Scheme 5a
Resorting to R-phenylselenyl methyl acrylate 12 offered
an attractive shortcut in our retrosynthetic route since its
cycloaddition with PIBF 5a would yield selenoester 11
directly. The starting selenoacrylate 12 has been synthesized
according to the procedure described by Piettre and col-
leagues,12 which consists of the reversible addition of PhSeCl
to methyl acrylate in the presence of ZnCl2. Upon basic
quenching (NEt3) only the captodative gem-disubstituted
regioisomer is obtained in 82% yield.
a (a) H2O2, CH2Cl2, -40 °C; (b) 0.1 equiv of CF3COOH, CH2Cl2,
The [3 + 2] cycloaddition between crude 13 and the dipole
derived from amine 14 has been achieved under acidic
conditions (Scheme 5).4b The pyrrolidine 15 was obtained
as a single diastereomer, possessing an endo ester append-
age.14 This selectivity makes sense in regards to the folded
structure of 13, which presents a single convex face
accessible to the dipole (Figure 2). This final adduct was
purified by chromatography and recovered in 14% overall
yield from 4a (four steps).
rt.
The cycloaddition was performed between 5a, generated
in situ as described above, and 12 in toluene at 80 °C for 2
h (Scheme 4). A single adduct 11b was recovered. The NMR
Scheme 4a
a (a) Tol, 80 °C, 2 h.
Figure 2. Approach of the ylide along the convex face of 13.
showed that the regioselectivity was unchanged while the
exo selectivity became total this time. This dramatic prefer-
ence can be tentatively attributed to π-π interactions
between the phenyl group borne by the selenium in 12, and
the conjugated cyclohexadienic system in 5a. Because isomer
11b is also very sensitive to silica gel, probably for steric
reasons comparable to those described above, its chromato-
graphic yield plummets to impractical values.
The next step toward 2 was the oxidation of 11 into the
corresponding epimeric selenoxides, which is expected to
undergo an instantaneous â-elimination of selenenic acid
since the selenium atom is syn to a proton in both 11a and
11b. Thus, hydrogen peroxide has been reacted separately
with the two epimers of 11 at -40 °C, giving in both cases
the expected bridged ethylenic ester 13 (Scheme 5). Albeit
the NMR spectra of the reaction mixture leaves no doubt to
the structure of the product, this compound is as sensitive
to silica gel as its precursors, preventing its chromatographic
purification and the evaluation of the yield. A facile retro-
Diels-Alder reaction, previously observed for a carbon-
bridged precursor of RPR 115135,13 probably explains this
instability.
In conclusion, this preliminary study on the synthetic
routes to pyrrolidine 15 provides evidence for a diastereo-
meric switch that allows for a selective access to both
epimers of R-selenoester 11. These two isomers give the
same 1,4-epoxy-1,4-dihydronaphthalene 13 after H2O2 oxida-
tion; their different topologies could possibly be exploited
through radical chemistry. This selectivity, which seems
mainly governed by the marked concavity of the oxa-bridged
skeleton, could also be extended to other electrophiles. All
details on the developments brought about by the synthesis
of these and others farnesyltransferase inhibitors analogues
will be reported in a full paper.3
Acknowledgment. C.M. thanks the Conseil Re´gional de
Haute-Normandie and Rhoˆne-Poulenc Rorer for a Ph.D.
grant. Dr. C. Harrison is acknowledged for proofreading this
manuscript.
OL005560H
(14) 1H NMR (200 MHz, CDCl3) δ (ppm) 2.27 (1H, d, J ) 9.6), 2.35
(1H, t, J ) 8.1), 3.20 (1H, t, J ) 7.9), 3.34 (1H, d, J ) 7.9), 3.40 (3H, s),
3.48 (1H, d, J ) 9.5), 3.56 (2H, s), 5.22 (1H, s), 6.94-7.77 (14H, m); 13
C
NMR (50 MHz, CDCl3) δ (ppm) 52.02, 54.82, 58.28, 59.71, 59.91, 66.27,
81.06, 92.94, 119.29, 120.43, 126.57, 126.84, 127.06, 127.22, 127.87,
128.15, 128.38, 136.00, 138.63, 144.87, 145.83, 173.83; CIMS (i-BuH) m/z
412 (M + 1, 100); exact mass calcd for C27H26O3N m/z 412.1913, found
412.1913.
(12) Piettre, S.; Janousek, Z.; Merenyi, R.; Viehe, H. G. Tetrahedron
1985, 41, 2527.
(13) Mailliet, P. Unpublished results.
Org. Lett., Vol. 2, No. 7, 2000
925