Synthesis of pyridine fused polycyclic amines using sequential ring-closing metathesis and radical cyclisation reactions

Article abstract of DOI:10.1016/S0040-4039(03)00389-7

The syntheses of novel pyridine fused polycyclic bridgehead amines are described using sequential ring-closing metathesis (RCM) and 5-exo-trig intramolecular radical cyclisation reactions. Critical to the success of the two sequential steps were the RCM catalyst and/or the nature of the nitrogen atom.

Full text of DOI:10.1016/S0040-4039(03)00389-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 2995–2999
Synthesis of pyridine fused polycyclic amines using sequential
ring-closing metathesis and radical cyclisation reactions
S. Richard Baker, Manuel Cases,* Martine Keenan, Richard A. Lewis and Paul Tan
Eli Lilly & Co. Ltd, Lilly Research Centre, Erl Wood Manor, Windlesham, Surrey, GU20 6PH, UK
Received 19 November 2002; revised 30 January 2003; accepted 7 February 2003
Abstract—The syntheses of novel pyridine fused polycyclic bridgehead amines are described using sequential ring-closing
metathesis (RCM) and 5-exo-trig intramolecular radical cyclisation reactions. Critical to the success of the two sequential steps
were the RCM catalyst and/or the nature of the nitrogen atom. © 2003 Elsevier Science Ltd. All rights reserved.
The selective activation or inhibition of a specific sub-
class of receptor is one of the many challenges facing
the medicinal chemist. One approach is to design com-
pounds, which are conformationally constrained so
they can only adopt a single shape, which exactly fits a
single receptor subtype. These rigid structures can rep-
resent a significant synthetic challenge. Sequential ring-
closing metathesis and radical cyclisation has proved to
be a very powerful tool for the synthesis of polycyclic
systems.1 Radical cyclisations are well known for their
tolerance towards a multitude of functional groups
including basic nitrogens.2 On the other hand, the
tolerance of commercially available metathesis catalysts
towards functional groups and more specifically basic
tertiary amines is less understood as most examples
reported in the literature use a deactivated nitrogen
(amides, carbamates, sulfonamides).3 In this paper we
report a new general procedure for the synthesis of
novel strained bridgehead azabicyclic amines fused to a
pyridine ring. The initial targets were the novel pyridi-
nes 1, 2 and 3 (Scheme 1).
Our general tactic employed a ring-closing metathesis
(RCM) reaction followed by a radical cyclisation to
form the critical bonds (Scheme 2). Retrosynthetically
the process can be thought of as the cleavage of the
dihydrofuran ring to give the key intermediate 4. This
compound then possesses the requisite 3-halo-pyridine
and an appropriately placed double bond, which would
allow, in the forward sense, the creation of the car-
bonꢀcarbon bond by a radical cyclisation or a Heck
reaction. Intermediate 4 could be derived from 5 using
an RCM reaction, which would close the ring and
introduce the double bond in one-step.
We started with the synthesis of the simplest tricyclic
molecule 1 in order to gain experience with RCM
reactions in this context (Scheme 3). The known inter-
mediate 9 was fashioned in three steps from commer-
cially available epoxide 6,4 using allylamine and
di-tert-butyl dicarbonate. The RCM step was achieved
with the commercially available Grubbs’ ruthenium cat-
alyst 8,3,4 to give the secondary alcohol 9. Compound 9
Scheme 1.
* Corresponding author. E-mail: cases manuel@lilly.com
–
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)00389-7

2996
S. R. Baker et al. / Tetrahedron Letters 44 (2003) 2995–2999
Scheme 2.
Scheme 3.
was then O-arylated with 3-bromo-2-fluoropyridine5 to
give the 2-pyridyl ether 10, which was treated under
classic radical cyclisation conditions to yield the desired
tricyclic compound 11.4 An attempt to carry out a
related cyclisation using a Heck reaction failed.6
Finally, compound 11 was either deprotected using
trifluoroacetic acid in dry dichloromethane, to give
amine 12 or reduced with lithium aluminium hydride to
give the N-methyl amine 1. The expected cis ring fusion
of amines 1 and 12 was confirmed by NOE experi-
ments.
dichloroethane, followed by a regioselective opening of
the epoxide 14 with allylamine in wet ethanol/tetra-
hydrofuran. During this step, concurrent substitution
of allylamine at the ester group could not be avoided.
Thus, the resulting mixture 15 was heated in a
microwave oven at 200°C for 30 min, giving the lactam
16. Oxidation of the secondary alcohol in lactam 16,
using Swern conditions,7 gave the ketone 17. This step
was followed by the selective addition of vinyl Grignard
to the ketone 17 in THF furnishing 18, the precursor
for the RCM step. The RCM using Grubbs’ ruthenium
3
catalyst 8 failed to give the desired olefin 19.
Invigorated by this success we decided to apply the
same methodology to the more complex target 2
(Scheme 4). The synthesis of diene 18, the precursor for
the RCM step, commenced with a straightforward
epoxidation of commercially available ethyl 4-pen-
tenoate 13 with meta-chloroperbenzoic acid in 1,2-
A second precursor 23 for the RCM step without a
carbonyl functionality in the ring was then envisaged
(Scheme 5). 3-Hydroxypiperidine 20 was N-alkylated
with allyl bromide in the presence of excess triethyl-
amine in THF to give alcohol 21, which was converted

S. R. Baker et al. / Tetrahedron Letters 44 (2003) 2995–2999
2997
Scheme 4.
Scheme 5.
to the ketone 22 by Swern oxidation.7 Ketone 22 was
then treated with vinyl Grignard to give the diene 23.
Sadly again, when treated with Grubbs’ catalyst, diene
23 did not give the desired bicyclic compounds, proba-
bly because the Grubbs’ ruthenium catalyst is incom-
patible with basic tertiary amines. We then turned to
another commercially available RCM catalyst, the
Schrock catalyst 26.
obtained in 54% yield when the diene 24 was reacted
with a catalytic amount of molybdenum complex 26 in
dry and degassed dichloromethane at room tempera-
ture. The olefin 25 gave the desired tetracyclic product
2 on treatment with tri-n-butyltin hydride and AIBN in
toluene at 80°C.
The same methodology was applied to the synthesis of
the [3.2.1] ring compound 3 (Scheme 6). However, in
this case the required 3-pyrrolidinone 30 was synthe-
sised by condensation of DL-malic acid 27 and allyl-
amine in refluxing xylene to give the imide 28. This was
followed by lithium aluminium hydride reduction to
Thus, alcohol 23, was deprotonated with potassium
hexamethyldisilazide in THF and condensed with 3-
bromo-2-fluoropyridine5 yielding the diene ether 24.
Gratifyingly, the expected RCM product 25 was

2998
S. R. Baker et al. / Tetrahedron Letters 44 (2003) 2995–2999
Scheme 6.
give alcohol 29.8 Swern oxidation gave the ketone 30,
which on addition of vinylmagnesium bromide gave the
vinyl alcohol 31. Selective 2-substitution of 3-bromo-2-
fluoropyridine with the potassium salt of 31 yielded the
diene 32. RCM reaction of diene 32 with the Schrock
catalyst 26 gave the bicyclic intermediate 33 that was
readily transformed into the target compound 3 after
the 5-exo-trig radical cyclisation step, using the previ-
ously described conditions.
software. Significant changes in bond lengths (>5%)
were observed between the two methods, highlighting
the strain in the ring system: the ab initio results were
used in the NMR-structure confirmation.
As expected a syn approach of the 3-pyridyl radical on
the double bond accounts for the results.
In conclusion we have developed a simple and efficient
route to novel pyridine fused polycyclic amines, based
on a sequential ring-closing metathesis, followed by a
5-exo-trig intramolecular radical cyclisation using tri-n-
butyltin hydride and AIBN.
NMR studies confirmed the stereochemistry of these
novel tetracyclic systems.9 The NOE interactions for the
[3.2.1] ring system showed that the bridgehead proton a
was on the opposite face to proton b on the one carbon
bridge as shown in Figure 1. Distances between key
This work emphasises how critical the choice of RCM
catalysis versus substrate can be to a successful syn-
thetic outcome.
,
protons are given in A (Fig. 1). An accurate model of
the structure was built to aid in the NMR-structure
confirmation. Optimisation of the mono-protonated
species was performed by molecular mechanics using
the MMFF force field, and by ab initio methods using
a 6-31 G* basis set, as implemented in the Spartan
References
1. Paquette, L. A.; Leit, S. M. J. Am. Chem. Soc. 1999, 121,
8126–8127.
2. For a leading review of radical cyclisations, see: Giese, B.;
Kopping, B.; Gobel, T.; Dickhaut, J.; Thoma, G.; Kulicke,
K. J.; Trach, F. Organic Reactions; John Wiley: New
York, 1996; Vol. 48, pp. 301–856.
3. For reviews of RCM, see: (a) Grubbs, R. H.; Chang, S.
Tetrahedron 1998, 54, 4413–4450; (b) Armstrong, S. K. J.
Chem. Soc., Perkin Trans. 1 1998, 371–388; (c) Schuster,
M.; Blechert, S. Angew. Chem., Int. Ed. Engl. 1997, 36,
2036–2056; (d) Fu¨rstner, A. Angew. Chem., Int. Ed. 2000,
39, 3012–3043; (e) Roy, R.; Das, S. K. Chem. Commun.
2000, 519–529.
4. Evans, D. A.; Campos, K. R.; Tedrow, J. S.; Michael, F.
E.; Gagne´, M. R. J. Org. Chem. 1999, 64, 2994–2995.
5. For synthesis, see: (a) Estel, L.; Marsais, F.; Queguiner, G.
J. Org. Chem. 1988, 53, 2740–2744; (b) Marsais, F.;
Pineau, Ph.; Nivolliers, F.; Mallet, M.; Turck, A.; Godard,
A.; Queguiner, G. J. Org. Chem. 1992, 57, 565–573.
Figure 1. The 6-31G* optimised structure of the [3.2.1] ring
,
system, showing the measured distances (A) between the
protons used to assign the NMR spectrum.

S. R. Baker et al. / Tetrahedron Letters 44 (2003) 2995–2999
2999
6. (a) Tamaru, Y.; Yamada, Y.; Yoshida, Z. J. Org. Chem.
1978, 43, 3396–3398; (b) Evans, P.; Grigg, R.; Ramzan, M.
I.; Sridharan, V.; York, M. Tetrahedron Lett. 1999, 40,
3021–3024; (c) Grigg, R.; Sridharan, V.; York, M. Tetra-
hedron Lett. 1998, 39, 4139–4142.
7. Mancuso, A. J.; Swern, D. Synthesis 1981, 165–185.
8. Bowers Nemia, M. M.; Lee, J.; Joullie, M. M. Synth.
Commun. 1983, 13, 1117–1123.
J=6.71, 12.21 Hz, 1H, H-6), 3.33 (d, J=10.99 Hz, 1H,
H-3), 3.38 (dt, J=6.71, 12.82 Hz, 1H, H-15), 3.48 (m, 1H,
H-5), 3.64 (d, J=10.99 Hz, 1H, H-3), 3.91 (m, 2H, H-13,
H-15), 4.06 (dt, J=3.66, 12.21 Hz, 1H, H-5), 7.13 (dd,
J=5.49, 7.32 Hz, 1H, H-10), 7.86 (d, J=7.32 Hz, 1H,
H-11), 8.05 (d, J=5.49 Hz, 1H, H-9). ESI-MS C12H14N2O
Mr (calcd) 202.1 Mr (found) 203.1 [M+H+].
9. Spectral data for compounds 1, 2 and 3 are reported
below.
Compound 1 (maleic acid salt) 1H NMR (300 MHz,
DMSO-d6) l 1.60 (broad, 1H), 2.20 (m, 1H), 2.76 (s, 3H),
3.00 (broad, 1H), 3.1 (broad, 1H), 3.30 (m, 1H), 3.45 (m,
1H), 3.7 (broad, 1H), 4.90 (dt, J=3.7, 6.79 Hz, 1H), 6.03
(s, 2H), 6.97 (dd, J=7.15, 5.27 Hz, 1H), 7.72 (broad d,
J=7.16 Hz, 1H), 8.00 (dd, J=4.9, 1.51 Hz, 1H), 9.7
(broad, 1H). ESI-MS C11H14N2O Mr (calcd) 190.1 Mr
1
(found) 191.1 [M+H+]. Compound 2 H NMR (300 MHz,
CD3OD) l 2.02–2.65 (broad, 6H), 3.25–3.43 (broad, 4H),
3.82 (d, J=11.31 Hz, 2H), 3.92 (dd, J=6.4, 11.68 Hz, 1H),
7.35 (dd, J=6.02, 7.16 Hz, 1H), 8.15 (d, J=5.66 Hz, 1H),
8.21 (d, J=7.53 Hz, 1H). ESI-MS C13H16N2O Mr (calcd)
216.1 Mr (found) 217.1 [M+H+]. Compound 3 (Fig. 2): H
NMR (500 MHz, CD3OD) l 2.05 (m, 1H, H-14), 2.41 (m,
1H, H-6), 2.54 (dt, J=6.71, 14.65 Hz, 1H, H-14), 2.68 (dt,
1
Figure 2. Compound 3 numbering used for the 1H NMR
assignment.