The partial reduction of heterocycles: An alternative to the birch reduction

Article abstract of DOI:10.1016/S0040-4039(99)02315-1

The partial reduction of a series of heterocycles in THF is described. By adding amine 3 as a proton source and naphthalene as an electron carrier, we were able to produce reductively alkylated pyrrolines and dihydrofurans in moderate to good yields. This reaction does not require liquid ammonia as a solvent, which may have interesting ramifications for large-scale synthesis. Moreover, we were also able to quench the reduction reactions with an acid chloride so performing a reductive acylation reaction. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(99)02315-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 1331–1334
The partial reduction of heterocycles: an alternative to the Birch
reduction
Timothy J. Donohoe,a,∗ Rakesh R. Harji a and Rick P. C. Cousins b
aDepartment of Chemistry, University of Manchester, Oxford Road, Manchester, M13 9PL, UK
bGlaxoWellcome Research and Development, Medicines Research Centre, Gunnels Wood Road, Stevenage, SG1 2NY, UK
Received 15 November 1999; revised 10 December 1999; accepted 12 December 1999
Abstract
The partial reduction of a series of heterocycles in THF is described. By adding amine 3 as a proton source and
naphthalene as an electron carrier, we were able to produce reductively alkylated pyrrolines and dihydrofurans in
moderate to good yields. This reaction does not require liquid ammonia as a solvent, which may have interesting
ramifications for large-scale synthesis. Moreover, we were also able to quench the reduction reactions with an acid
chloride so performing a reductive acylation reaction. © 2000 Elsevier Science Ltd. All rights reserved.
In the preceding paper we described the double reductive alkylation of 2,5-diester 1 under Birch
type conditions (Scheme 1).1 Later, we discovered that the same reduction could be performed in THF
(no ammonia) with lithium metal and catalytic amounts of naphthalene used as an electron shuttle.1
We feel that removing the need for ammonia solvent will make the partial reduction reaction more
practicable, especially on a large scale, and may also allow us to quench reduction reactions with sensitive
electrophiles that would otherwise react with ammonia itself.
Scheme 1.
The reduction of other electron-deficient heterocycles using this method would prove a potentially
useful extension to this methodology. However, there are some problems that need to be addressed
first. Reduction of 1 with naphthalene is straightforward because the pyrrole acquires two electrons to
form a stable dianion that is subsequently alkylated twice. Most of the pyrroles and furans that we have
reduced previously have only one ester activating group and form relatively unstable dianions that are
∗
Corresponding author. E-mail: t.j.donohoe@man.ac.uk (T. J. Donohoe)
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(99)02315-1
tetl 16243

1332
basic enough to deprotonate ammonia (protonation of this dianion forms an enolate which can then be
alkylated). Thus, a proton source is needed if the reaction is to become widely applicable.
Previously we have added an excess of (MeOCH2CH2)2NH 3 to the Birch reduction of heterocycles
because it improves the yields.3 We suppose that the amine would either protonate the heterocyclic
dianion itself, or react favourably with LiNH2 to form (MeOCH2CH2)2NLi in situ. The advantage of
having this anion present, rather than LiNH2, lies in its reduced nucleophilicity (presumably the anion
is chelated) and basicity: this means that the anion is not so aggressive towards both starting material
and product and the yields are increased. One essential property of this amine is its inability to protonate
the enolate formed after protonation of the dianion: this means that we can add excess 3 with impunity
and makes its role distinctly different from conventional proton sources used in the Birch reduction (e.g.
t-BuOH).4
Clearly, the properties of this amine make it an ideal substitute for ammonia. Therefore, we took the
electron deficient furan 4 and reduced it with lithium in THF with naphthalene (5 equiv.) and amine 3 (5
equiv.). Sonication during formation of lithium naphthalenide was found to be beneficial to the yield.5
After 2 h at −78°C, the reaction was quenched with two electrophiles (Scheme 2). We were delighted to
find that the reductive alkylation reaction proceeded well and gave products identical to those observed
in liquid ammonia. However, the yields of reduced products were not quite as high as that observed in a
bona fide Birch reduction.6
Scheme 2.
We then set about reducing 3-substituted and 3,4-disubstituted pyrroles under identical conditions to
assess the scope of the reduction method (Scheme 3). In both cases, reductive alkylation products were
isolated in reasonable yields.
Scheme 3. Compound 8 was formed as a 1:1 mixture of diastereoisomers
Pyrrole 9 is one of the most useful heterocyclic systems we can partially reduce and a study was
therefore undertaken (Scheme 4). In each case, the reduction reaction worked well to produce substituted
pyrrolines. Formation of 15 via a reductive aldol reaction is noteworthy.7
The only example in which the product was at variance with that observed under Birch conditions
in ammonia involved butyliodide as an electrophile. In this case, the protonated material 14 (which is a
protected form of dehydroproline) was isolated in 57% yield; this result implies that the enolate generated
under these conditions is not as nucleophilic as that formed in the Birch reduction.

1333
Scheme 4. Compound 15 was formed as a 1:1 mixture of diastereoisomers
Remarkably, we were able to quench reduction reactions with benzoyl chloride to form β-ketoesters
(Scheme 5); we have previously found that this type of electrophilic quench does not work at all in liquid
ammonia, presumably because the acid chloride reacts with ammonia and the HCl by-product protonates
the enolate.8
Scheme 5.
The mechanism of this reduction has not been studied in detail. However, we presume that the (green)
radical anion derived from naphthalene (Li Np) is formed in solution (Fig. 1). Two equivalents of this
reagent should form the corresponding dianion from the relevant heterocycle (only small amounts need
be generated at any given time). Presumably, only this dianion is basic enough to deprotonate amine 3 to
yield an enolate A and the amide anion 18. With the exception of reaction with butyl iodide, this enolate
behaves as if it were generated in liquid ammonia and reacts at the α-position with a variety of different
electrophiles.
Fig. 1.
Attempts to optimise this reaction are underway and we have made the following observations
regarding the reaction of of 9→10 (Scheme 4).8
(i) The product 10 can be isolated in 57% yield using sodium, rather than lithium, metal.
(ii) Addition of the pyrrole substrate over 2 h, using a syringe pump, increased the yield of 10 to 81%
(however, not all heterocycles examined gave this increase in yield and further investigation is called
for).
(iii) The reaction was successful in diethylether (71% yield) using lithium metal and a syringe pump.

1334
To conclude, we have developed an interesting method for the partial reduction of heterocycles that
does not require ammonia as solvent. Replacing ammonia with THF and amine 3 means that a wider
range of electrophiles can be tolerated in the reduction which obviously increases its scope. Efforts are
currently underway to explore this reaction further.
Representative experimental procedure: A mixture of lithium metal (28 mg, 3.9 mmol) and naphtha-
lene (506 mg, 3.9 mmol) in THF (25 mL) was placed in a sonic bath at room temperature for 1.5 h under
an atmosphere of argon. The dark green solution was then cooled to −78°C and a premixed solution of
pyrrole 9 (200 mg, 0.8 mmol) and amine 3 (0.59 mL, 3.9 mmol) was added. After 1 h, benzoyl chloride
(1 mL, 8.6 mmol) was added and the reaction stirred at −78°C for 10 min before being quenched with aq.
NH4Cl. The crude mixture was evaporated to dryness and purified by chromatography on silica (eluting
with EtOAc: petrol, 1:10) to afford 17 as a pale yellow oil (315 mg, 59%): 1H NMR (200 MHz, CDCl3)
δ 7.90–7.75 (m, 2H, ArH), 7.60–7.30 (m, 3H, ArH), 6.24–6.03 (m, 2H, 2×CH), 5.20–5.03 (m, 1H,
HC(CH3)2), 4.58–4.30 (m, 2H, NCH2), 1.50–1.00 (m, 15H, C(CH3)3, HC(CH3)2).
Acknowledgements
We would like to thank GlaxoWellcome (R.R.H.) and the EPSRC for financial support. Karl W. Ace
and James G. Logan are thanked for performing additional experiments.
References
1. Donohoe, T. J.; Harji, R. R.; Cousins, R. P. C. Tetrahedron Lett. 2000, 41, 1327. For a brief description of other alternatives
to the Birch reduction, see: Rabideau, P. W.; Marcinow, Z. Organic Reactions 1992, 42, 1.
2. (a) Donohoe, T. J.; Guyo, P. M. J. Org. Chem. 1996, 61, 7664; (b) Donohoe, T. J.; Guyo, P. M.; Beddoes, R. L.; Helliwell,
M. J. Chem. Soc., Perkin Trans. 1 1998, 667; (c) Donohoe,T. J.; Harji, R. R.; Cousins, R. P. Chem. Commun. 1999, 141.
3. Donohoe, T. J.; Guyo, P. M.; Harji, R. R.; Cousins, R. P. C. Tetrahedron Lett. 1998, 39, 3075.
4. For reviews of the Birch reduction, see: Schultz, A. G. Chem. Commun. 1999, 1263. Rabideau, P. W.; Marcinow, Z. Organic
Reactions 1992, 42, 1; Mander, L. N. In Comprehensive Organic Synthesis; Trost, B. M.; Fleming, I., Eds.; Pergamon Press:
New York, 1991; Vol. 8.
5. Azuma, T.; Yanagida, S.; Sakurai, H.; Sasa, S.; Yoshino, K. Synth. Commun. 1982, 12, 137; see also: Handbook of Reagents
for Organic Synthesis: Oxidising and Reducing Agents; Burke, S. D.; Danheiser, R. L. Eds.; John Wiley: New York, 1999.
6. All new compounds were fully characterised (1H , 13C NMR, mass spectra, IR and HRMS or microanalysis) and the data of
known compounds compared to that in the literature: 7 Ref. 3; 8 Ref. 2c; 10 Ref. 2b.
7. Donohoe, T. J.; Ace, K. W.; Guyo, P. M.; Helliwell, M.; McKenna, J. Tetrahedron Lett. 2000, 41, 989.
8. Donohoe, T. J. Harji, R. R., unpublished work.