Reductive aldol reactions on aromatic heterocycles

Article abstract of DOI:10.1016/S0040-4039(99)02224-8

The Birch reduction of both pyrroles and furans can be quenched with an aldehyde electrophile, thus constituting a reductive aldol reaction. Although the reaction was not stereoselective, the pyrrolines derived from reduction of pyrroles could be oxidised and then reduced with NaBH₄ to provide syn- aldol adducts with high levels of stereoselectivity. The relative stereochemistry of two adducts was proven by X-ray crystallography. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(99)02224-8

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 989–993
Reductive aldol reactions on aromatic heterocycles
a,∗
a
a
a,†
Timothy J. Donohoe, Karl W. Ace, Paul M. Guyo, Madeleine Helliwell and
Jeffrey McKenna b
a
Department of Chemistry, University of Manchester, Oxford Road, Manchester, M13 9PL, UK
b
Rhône-Poulenc Rorer, Dagenham Research Centre, Dagenham, RM10 3RX, UK
Received 27 October 1999; revised 24 November 1999; accepted 29 November 1999
Abstract
The Birch reduction of both pyrroles and furans can be quenched with an aldehyde electrophile, thus constituting
a reductive aldol reaction. Although the reaction was not stereoselective, the pyrrolines derived from reduction
of pyrroles could be oxidised and then reduced with NaBH4 to provide syn-aldol adducts with high levels of
stereoselectivity. The relative stereochemistry of two adducts was proven by X-ray crystallography. © 2000 Elsevier
Science Ltd. All rights reserved.
Keywords: Birch reduction; aldol; pyrrole; furan.
The aldol reaction is one of the most important carbon–carbon bond forming reactions used in
synthesis; one of the principal reasons for its success is that control of both relative and absolute
1
stereochemistry is possible. Normally, the enolate component of an aldol reaction is generated by
reaction of a carbonyl compound with base: however, other methods of enolate generation include
conjugate addition to unsaturated carbonyl compounds and reduction of, for example, α-halo carbonyl
compounds. There is another popular way of generating enolates, which involves reduction of aromatic
2
carboxylic acids with a group I metal in ammonia: however, to the best of our knowledge, reaction of
such enolates with carbonyl electrophiles (in what would then constitute a Birch reductive aldol reaction)
3
is unknown.
As part of a programme designed to study the partial reduction of heterocycles, we recently examined
4
5
the reductive alkylation of electron deficient pyrroles and furans: we believe that addition of two
electrons to such heterocycles generates a dianion and subsequently (after protonation at C-5 by
ammonia) an enolate, Scheme 1. This enolate is a willing partner in many alkylation reactions and in this
paper we wish to report on the use of aldehydes as electrophiles in the Birch reduction. Our study has
concentrated on the development of the reductive aldol reaction of pyrroles, as the products so produced
may have application in the synthesis of natural products such as lactacystin and oxazolamycin.
∗
†
Corresponding author. E-mail: t.j.donohoe@man.ac.uk (T. J. Donohoe)
Author to whom correspondence regarding the X-ray crystal structure should be addressed.
0
040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(99)02224-8
tetl 16152

9
90
Scheme 1.
Compound 1 was prepared in a three high yielding steps from pyrrole-2 carboxylic acid and then
subjected to reduction using lithium in ammonia/THF at −78°C. After quenching excess electrons with
isoprene, the enolate was treated with a series of aldehydes (Scheme 2); in each case the result was a high
yielding and regioselective aldol reaction (Table 1).6
Scheme 2.
Table 1
Each aldol adduct was formed as a 1:1 mixture of diastereoisomers
Presently, there are two limitations to this methodology: (i) the aldol adducts are formed as essentially
1
:1 mixtures of diastereoisomers. It is unclear at this time whether the lack of stereoselectivity in the
aldol reaction is a consequence of formation of a mixture of enolate geometric isomers or whether there
is a lack of stereospecificity during the aldol reaction of either one of these isomers; and (ii) enolisable
i
‡
aldehydes, such as Pr CHO, protonate the enolate and so do not give aldol adducts using this protocol.
In an attempt to address the first issue raised above and control the stereochemistry of the aldol adducts
2–5, we subjected each mixture to an oxidation/reduction sequence. Several oxidants were screened for
the transformation of these hindered secondary alcohols to ketones, and only the Dess–Martin reagent,
activated by water, gave good yields (Scheme 3).7
Ketones 6–9 were then reduced with sodium borohydride and cerium trichloride in isopropanol solvent
(
−78°C→rt) Scheme 4; we were delighted to observe high stereoselectivity for formation of the syn
diastereoisomer. The CeCl additive appeared to speed-up the rate of reduction, thus allowing (more
3
‡
Very reactive electrophiles such as formaldehyde were also not suitable for the reductive aldol reaction, although related
compounds could be accessed by Birch reduction of 1, followed by quenching with SEMCl.

9
91
Scheme 3.
selective) reaction at a lower temperature. Isopropanol tended to give slightly more stereoselective results
than methanol solvent, and we presume that this is a consequence of a more bulky (and therefore more
discriminating) alkoxyborohydride reducing agent. Compound 8 resisted attempts at reduction using
NaBH and use of more powerful reducing agents gave multi-component mixtures: presumably the
4
−1
ketone carbonyl in 8 is conjugated with the pyrrole nitrogen (ν C_O 1657 cm ) and is therefore
relatively unreactive.
Scheme 4. The diastereoselectivity of the reaction was assessed by GC and in comparison with a 1:1 mixture of isomers from
the Birch reduction. Compound 6 was reduced satisfactorily without cerium trichloride additive. Compound syn-5 was formed
with approximately 10% of a by-product which may be the 1,4 reduced adduct
We should like to note that the syn selectivity observed during reduction is consistent with the
Felkin–Anh model so long as we assign the following sizes to groups attached to C-2 of the heterocycle,
i
NBoc=large, COOPr =medium, and C_C=small (A, Scheme 4: this ordering seems reasonable on both
8
steric and electronic grounds).
The relative stereochemistry of the aldol adducts was difficult to ascertain by NMR spectroscopy and
was secured by X-ray crystallographic analysis of syn-2 (Fig. 1).
The relative stereochemistry of both syn-3 and syn-5 was assigned by analogy and in the knowledge
that there were several features in the spectroscopic data of the syn diastereoisomers that helped to
reinforce our assignment (for example, each syn isomer had a low field, D O exchangeable, doublet
2
1
in the H NMR spectrum). This data allowed us to assign one of the diastereoisomers of 4 as syn: this
compound was separated and the structure proven by X-ray crystallography (Fig. 1).
In addition, we also attempted the reductive aldol reaction on furan 10 and obtained a good yield of
product 11 (formed as a 1:1 mixture of diastereoisomers) (Scheme 5). Unfortunately, the ketone derived

9
92
Fig. 1.
from oxidation of 11 was not reduced stereoselectively by the borohydride/CeCl mixture; clearly, the
3
Boc group on the nitrogen of 3–5 is essential in attaining a good selectivity.
Scheme 5.
To conclude, we have shown that aldehydes may be used as electrophilic partners for the Birch reac-
tion. Although the adducts were formed as 1:1 mixtures of diastereoisomers, the syn-aldol compounds
could be obtained readily via a two step oxidation/reduction procedure.
Representative experimental procedure: A solution of 1 (269 mg, 1.1 mmol), in THF (10 mL) was
added dropwise to a blue solution of ammonia (50 mL), THF (10 mL) and lithium (19 mg, 2.7 mmol)
at −78°C under nitrogen. After 15 min, isoprene (3 drops) was added followed by trans-hexenal (412
µL) and the reaction stirred at −78°C for 2 h. A saturated solution of NH Cl (10 mL) was added and
4
the mixture warmed to room temperature, under a stream of nitrogen, to remove the ammonia. Brine
(
50 mL) was added and the mixture extracted with ethyl acetate (4×50 mL), dried (Na SO ), filtered
2
4
and concentrated in vacuo. Chromatography on silica (eluting with ethyl acetate:petrol, 1:9) gave 5 as a
colourless oil (293 mg, 78%).
Acknowledgements
We would like to thank Rhône-Poulenc Rorer and the EPSRC (KWA), the University of Manchester
(PMG) and Zeneca Pharmaceuticals (Strategic Research Fund) for financial support.
References
1
2
3
. For a review of the aldol reaction, see: Heathcock, C. H. In Comprehensive Organic Synthesis; Trost, B. M.; Fleming, I.,
Eds.; Pergamon: New York, 1991; Vol. 2, p. 133 and p. 181.
. For reviews of the Birch reduction, see: Rabideau, P. W.; Marcinow, Z. Organic Reactions 1992, 42, 1; Mander, L. N. In
Comprehensive Organic Synthesis; Trost, B. M.; Fleming, I., Eds.; Pergamon: New York, 1991; Vol. 8.
. For the SmI
2
promoted aldol/alkylation reaction of thiophenes, see: Yang, S.-M.; Fang, J.-M. Tetrahedron Lett. 1997, 38,
1
589; for early work on the reductive aldol reaction of enones, see: Stork, G.; d’Angelo, D. J. Am. Chem. Soc. 1974, 96, 7114.

9
93
4
. Donohoe, T. J.; Guyo, P. M. J. Org. Chem. 1996, 61, 7664; Donohoe, T. J.; Guyo, P. M.; Beddoes, R. L.; Helliwell, M. J.
Chem. Soc. Perkin Trans. 1 1998, 667; Donohoe, T. J.; Guyo, P. M.; Harji, R. R.; Cousins, R. P. C. Tetrahedron Lett. 1998,
3
9, 3075.
5
6
7
8
. Donohoe, T. J.; Helliwell, M.; Stevenson, C. A.; Ladduwahetty, T. Tetrahedron Lett. 1998, 39, 3071.
. All new compounds were fully characterised ( H, C NMR, mass spec, IR and HRMS or microanalysis).
. Meyer, S. D.; Schreiber, S. L. J. Org. Chem. 1994, 59, 7549.
. For a discussion of the Felkin–Anh model, see: Eliel, E. L. In Stereochemistry of Organic Compounds; Wilen, S. H.; Mander,
L. N., Eds.; Wiley Interscience: New York, 1994.
1
13