Scope and limitations of the Minisci reaction for the synthesis of aza-heterocycles

Article abstract of DOI:10.1016/j.tetlet.2009.09.112

Attempts to prepare several classes of aza-heterocycles by application of the Minisci radical cyclisation reaction are described. Competing β-scission, hydrolytic cleavage and lactonisation reactions were found to be major hurdles to adopting this strategy for the synthesis of such targets.

Full text of DOI:10.1016/j.tetlet.2009.09.112

Tetrahedron Letters 50 (2009) 6772–6774
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Scope and limitations of the Minisci reaction for the synthesis
of aza-heterocycles
b
Ryan N. Burgin a, Simon Jones a, , B. Tarbit
*
a Department of Chemistry, University of Sheffield, Dainton Building, Brook Hill, Sheffield S6 1XD, UK
b Vertellus Specialities UK Ltd, Seal Sands Road, Seal Sands, Middlesbrough TS2 1UB, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 July 2009
Revised 5 September 2009
Accepted 18 September 2009
Available online 24 September 2009
Attempts to prepare several classes of aza-heterocycles by application of the Minisci radical cyclisation
reaction are described. Competing b-scission, hydrolytic cleavage and lactonisation reactions were found
to be major hurdles to adopting this strategy for the synthesis of such targets.
Ó 2009 Elsevier Ltd. All rights reserved.
Aza-heterocycles represent important scaffolds in drug discov-
ery platforms, including the azaindole nucleus which has shown
great potential as a structural subunit and as an indole biostere
in the search for biologically relevant molecules (Fig. 1). Although
significantly less predominant than their indole counterparts, sev-
eral natural products that contain this substructure have also been
isolated, including the variolin family, some of which have been
found to exhibit anti-cancer and anti-viral properties.1 It is not
surprising that numerous methods have been described for their
synthesis, from classical condensation reactions to transition me-
tal-mediated routes, all of which have been extensively reviewed.2
In addition to these methods, a common strategy to form aza-
heterocycles involves the use of radical cyclisation as the key step.
Most of these transformations occur via 4-, 5- or 6-exo-trig/dig
modes from a carbon or nitrogen radical onto an alkene or alkyne
acceptor.3 An alternative approach would be by the direct addition
of a carbon-centred radical to a pyridine nucleus (Fig. 2).
Such a radical could be generated from a carboxylic acid, mak-
ing use of the elegant work of Minisci et al., who demonstrated that
carbon-centred radicals could be formed through the silver-cata-
lysed decomposition of carboxylic acids, and that the radical
formed would add in an intermolecular fashion to a protonated
pyridine at the 2- and 4-position.4 Use of an acid and oxidant, such
as ammonium persulfate, increases the efficiency of the reaction
through activation of the pyridine as an electron acceptor, in addi-
tion to ensuring efficient re-oxidation of the aromatic moiety and
silver salts.5 An advantage of this strategy would be the ability to
couple readily available and structurally diverse bromopyridines
and b-amino acids, ultimately allowing access to substituted
azaindoles.
pling reaction. However, despite numerous attempts at conducting
Ullman-type couplings using conditions reported by Ma and Xia
(10 mol % CuI, K2CO3 in DMF) for the coupling of bromobenzenes
with amino acids,6 the reaction of b-alanine with 3-bromopyridine
did not afford anything other than non-quantitative return of start-
ing material. An alternative approach using 3-aminopyridine as a
nucleophile, and methyl acrylate as a Michael acceptor proceeded
reasonably well under solvent-free conditions in 41% yield
(Scheme 1). Lewis acid catalysis [Yb(OTf)3, Bi(OTf)2, Zn(OTf)2] in
CH3CN afforded much lower yields (<10%). The subsequent hydro-
lysis step proceeded smoothly, but it was not possible to obtain the
carboxylic acid (pKa 4.8) due to the similar pKa values of the acid
(propanoic acid, pKa 4.9) and the conjugate acid of the pyridine
NH2
N
N
OH
N
MeO
OMe
O
N
N
HN
O
N
N
N
NH2
O
Ph
Variolin B
BMS-488043
Figure 1. Representative azaindoles.
radical
H
H
H
cyclization
N
N
OH
N
O
For this strategy, formation of the key carboxylate precursor 1
was initially envisioned using an organometallic-mediated cou-
N
+
N
N
Br
H2N
OH
O
N
* Corresponding author. Tel.: +44 (0) 114 222 9483; fax: +44 (0) 114 222 9346.
E-mail address: simon.jones@sheffield.ac.uk (S. Jones).
Figure 2. Retrosynthetic strategy.
0040-4039/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.09.112

R. N. Burgin et al. / Tetrahedron Letters 50 (2009) 6772–6774
6773
access to 1,5-naphthyridines (Scheme 4), perhaps the least studied
of the naphthyridine family, but still having an important role in
medicinal chemistry applications.8
H
N
O
NH2
No solvent
O
OMe
+
80 oC, 48 h
41%
N
N
Amide 7 was prepared in good yield (69%) by condensation of 3-
aminopyridine and succinic anhydride, and using this substrate
under the optimised Minisci reactions yielded 3-aminopyridine 2
in 60% yield. Two control experiments, the first using TFA and
water, and second omitting the TFA from the Minisci conditions,
led to the formation of 3-aminopyridine 2, and in the latter case
also returned succinic acid. Thus, although a radical-mediated
cleavage cannot be ruled out, it is highly likely in this case that
the Achilles heel of this strategy is the labile nature of this partic-
ular amide bond. To remove the problematic amide functionality, a
number of reducing agents were applied to the key substrate 7,
with H2/PtO2 returning starting material, whilst LiAlH4, LiBH4
and Red-Al at different temperatures and in various solvents
yielded none of the desired product. Related transformations have
been described in the literature, although not for pyridine-contain-
ing substrates.9
OMe
H
KOH, MeOH/H2O
N
O
K
rt, 24 h
100%
O
N
1
Scheme 1. Synthesis of key carboxylate intermediate 1.
(3-aminopyridine, pKa 6.0) moieties. The carboxylate 1 was iso-
lated directly by removal of solvent in quantitative yield.
The Minisci reaction was performed on carboxylate 1 by applica-
tion of optimised literature conditions,7 and upon work-up, the only
product isolated in low yield was 3-aminopyridine 2 (Scheme 2).
A degradation pathway to account for this observation can be
envisioned where the radical generated undergoes irreversible b-
scission, which upon work-up would yield 3-aminopyridine and
ethene. The low yield of the 3-aminopyridine can be accounted
for by the poor mass return resulting from the water solubility of
both this compound and the starting material 1. This hypothesis
was supported by attaching a bromine water trap to the outlet of
the reaction condenser, which decolourised as the reaction pro-
ceeded and ethene gas was generated. A control reaction involving
intermolecular Minisci reaction of potassium 3-phenylpropionate
3 and pyridine 4 led to the coupled products 5 and 6 in yields of
50% and 25%, respectively, with no decolouration of the bromine
water (Scheme 3).
Heating the potassium salt 1 at reflux in the presence of TFA and
H2O without a radical initiator for 6 h at 70 °C, followed by quench-
ing with KOH, returned starting material 1 in quantitative yield,
ruling out hydrolytic cleavage. Thus, although the key radical
was formed, b-scission competes at a faster rate than the desired
cyclisation. It is therefore unlikely that aza-heterocycles could be
formed by this strategy if the radical is in the b-position relative
to a good leaving group, in this case 3-aminopyridine. In order to
circumvent b-scission, a homologue was considered, thus allowing
A new route was devised that avoided the problems of the
amide functionality. Methanolysis of
c-butyrolactone 8 under
acidic conditions, followed by immediate oxidation of the crude
reaction mixture with PCC afforded, after distillation, the aldehyde
9 in 39% yield (Scheme 5).
Reductive amination of aldehyde 9 with 3-aminopyridine 2
using NaBH3CN in glacial acetic acid in the presence of activated
molecular sieves gave the reduced product 10 in 66% yield. Hydro-
lysis was performed as previously using KOH in MeOH/H2O,
providing the potassium salt 11 in quantitative yield. This was
subjected to the standard Minisci conditions (Scheme 6), but the
expected cyclised product was not seen, the 1H NMR spectrum of
the crude reaction mixture revealing a nearly equal mixture of 3-
aminopyridine 2 and
c
-butyrolactone 8.
O
H
CH3CN
rt, 3 h
NH2
+
N
O
O
69%
O
N
N
HO
O
7
1 eq. (NH4)2S2O8
0.1 eq. AgNO3 (10 mol%)
0.2 eq. TFA, H2O
H
1 eq. (NH)4S2O8
H
N
N
O
70 oC, 6 h
60%
0.1 eq. AgNO3
NH2
N
O
K
N
0.2 eq. TFA, H2O
70 oC, 6 h
O
N
N
1
2
NH2
20%
N
N
2
1,5-naphthyridine
H
N
H2C CH2
Scheme 4. Attempted synthesis of 1,5-naphthyridine.
β-scission
N
Scheme 2. Attempted Minisci cyclisation.
O
O
H2SO4 (cat.)
PCC
HO
O
OMe
CH2Cl2, 3 h
MeOH, Δ, 5 h
O
8
39% over 2
steps
K
O
1 eq. (NH)4S2O8
0.1 eq. AgNO3
3
+
+
H
N
O
0.2 eq. TFA, H2O
70 oC, 6 h
NaBH3CN
N
AcOH, MeOH
NH2
+
N
N
4Å MS, rt, 24 h
N
N
O
OMe
4
5
6
O
OMe
66%
50%
25%
9
2
10
Scheme 3. Control intermolecular Minisci reaction.
Scheme 5. Synthesis of a 1,5-naphthyridine intermediate.

6774
R. N. Burgin et al. / Tetrahedron Letters 50 (2009) 6772–6774
H
H
N
H
H
N
N
N
KOH, MeOH/H2O
rt, 24 h
KOH, MeOH/H2O
rt, 24 h
N
N
N
N
O
MeO
O
OMe
O
O
K
100%
100%
12
13
N
11
10
O
K
O
O
O
1 eq. (NH)4S2O8
0.1 eq. AgNO3
1 eq. (NH)4S2O8
0.1 eq. AgNO3
NH2
O
O
O
N
0.2 eq. TFA, H2O
70 oC, 6 h
N
14
Minor
component
0.2 eq. TFA, H2O
70 oC, 6 h
8
40%
2
45%
Major
component
H2
N
Scheme 8. Attempted Minisci cyclisation of a 1,6-naphthyridine precursor.
N
by an independent route failed. However, it is clear that the pre-
dominant pathway in this reaction is likely to once again be a rad-
ical-mediated cyclisation.
O
O
Scheme 6. Attempted Minisci cyclisation.
Through the attempts made towards the synthesis of azain-
doles, 1,5-naphthyridines and the homologue towards a 6,7-fused
system, it has become clear that a route based on the Minisci reac-
tion is not feasible for any of these systems. Though disappointing,
key points about the nature of the mechanism have been investi-
gated and elucidated in one case. Importantly, although the 3-ami-
nopyridine moiety provides a useful scaffold, it creates significant
problems arising from its ability to act as a leaving group, leading
to b-scission or by facilitating competing ionic or radical pathways
(5-exo-tet vs Minisci reaction). Additionally, the study demon-
strates that for the synthesis of fused aza-heterocycles using the
Minisci reaction, it is essential that the second ring system com-
prises a carbon atom at the benzylic position.
O
O
PCC, CH2Cl2
OMe
rt, 3 h
H2SO4 (cat.)
MeOH, Δ, 24 h
98%
HO
O
65%
H
N
NaBH3CN
AcOH, MeOH
O
NH2
+
N
MeO
4Å MS, rt, 24 h
N
O
OMe
59%
O
12
Scheme 7. Preparation of a 1,6-naphthyridine intermediate.
Acknowledgements
Several possibilities exist for the formation of the lactone. Ring
closure may occur by an ionic 5-exo-tet mechanism before the car-
boxyl radical is formed, or by an analogous radical mechanism,
which is presumably faster than the rate of any decarboxylation
process. When the potassium salt was heated at 70 °C in the pres-
ence of TFA and H2O for the same length of time followed by basic
work-up (KOH), the 1H NMR spectrum of the crude reaction prod-
uct revealed mainly starting material as well as a small amount
(ꢀ20%) of potassium 4-hydroxybutyrate resulting from hydrolytic
cleavage of the substrate. This indicates that although an ionic cyc-
lisation occurs under the reaction conditions, the radical-based
pathway occurs at a significantly faster rate.
We thank the EPSRC for funding and Vertellus Specialties for
use of their resources and additional financial support.
References and notes
1. Perry, N. B.; Ettouati, L.; Litaudon, M.; Blunt, J. W.; Munro, M. H. G.; Parkin, S.;
Hope, H. Tetrahedron 1994, 50, 3987–3992.
2. (a) Popowycz, F.; Mérour, J.-Y.; Joseph, B. Tetrahedron 2007, 63, 8689–8707; (b)
Song, J. J.; Reeves, J. T.; Gallou, F.; Tan, Z.; Yee, N. K.; Senanayake, C. H. Chem. Soc.
Rev. 2007, 36, 1120–1132.
3. Giese, B. Radicals in Organic Synthesis: Formation of Carbon–Carbon Bonds;
Pergamon: Oxford, 1986. Chapter 4.
Since the problem with this reaction was the fast five-mem-
bered ring closure, formation of a six-membered ring would be
expected to be slower, and hence may allow for the decarboxyl-
ation-Minisci step to occur before any competing ring closure.
The synthesis of the Minisci precursor 12 for this new system
was analogous to that of substrate 10 and was easily accessed in
moderate to good yields (Scheme 7).
With the ester 12 in hand, hydrolysis to the carboxylate 13 was
carried out in KOH in MeOH/H2O in quantitative yield (Scheme 8).
Reaction of carboxylate 13 under the standard Minisci condi-
tions led to a complex mixture, predominantly containing d-val-
4. Minisci, F.; Galli, R.; Cecere, M.; Malatesta, V.; Caronna, T. Tetrahedron Lett. 1968,
9, 5609–5612.
5. Minisci, F.; Bernardi, R.; Bertini, F.; Galli, R.; Perchinummo, M. Tetrahedron 1971,
27, 3575–3579.
6. Ma, D.; Xia, C. Org. Lett. 2001, 3, 2583–2586.
7. Extensive optimisation reactions were performed on the intermolecular Minisci
reaction depicted in Scheme 3. Critical to success was to use fresh supplies of
ammonium persulfate, the shelf-life of which is only guaranteed for one year.
Additionally, the potassium salt of 3-phenylpropionic acid 3 was employed in an
identical manner as the more commonly used carboxylic acid. This was more
useful in our approach due to the problems associated with isolating
a
homogenous sample of the carboxylic acid because of the similar pKa values
of the pyridine and carboxylic acid functional groups.
8. For example, see: Gellibert, F.; Woolven, J.; Fouchet, M.-H.; Mathews, N.;
Goodland, H.; Lovegrove, V.; Laroze, A.; Nguyen, V.-L.; Sautet, S.; Wang, R.;
Janson, C.; Smith, W.; Krysa, G.; Boullay, V.; de Gouville, A.-C.; Huet, S.; Hartley,
D. J. Med. Chem. 2004, 47, 4494–4506.
9. Conrow, R. E.; Delgado, P.; Dean, W. D.; Callen, G. R.; Plummer, S. V. Tetrahedron
Lett. 2008, 49, 2348–2350.
erolactone. Repeated column chromatography resulted in
a
fraction with a pyridine-containing component that was tenta-
tively assigned as the lactam 14 by analogy with the 1H NMR spec-
trum of N-acetyl 3-aminopyridine. Attempts to prepare lactam 14