A flexible radical approach to 5-substituted 4,5-dihydro-3H-pyrido[4,3-b] azepin-2-ones. Some mechanistic observations on the radical cyclisation- aromatisation process

Article abstract of DOI:10.1016/j.tetlet.2012.04.020

Variously substituted novel dihydropyridoazepinones have been prepared by an intermolecular radical addition followed by a radical cyclisation on a pyridine ring. The latter process involved the use of a combination of two different peroxides, an experimental contrivance resulting from a careful product analysis and a better understanding of the cyclisation step.

Full text of DOI:10.1016/j.tetlet.2012.04.020

Tetrahedron Letters 53 (2012) 3220–3224
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
A flexible radical approach to 5-substituted 4,5-dihydro-3H-pyrido[4,3-b]azepin-
2-ones. Some mechanistic observations on the radical cyclisation-aromatisation
process
Laurent Petit a, Iuliana Botez b, André Tizot b, Samir Z. Zard a,
⇑
a Laboratoire de Synthèse Organique, UMR 7652 CNRS, Ecole Polytechnique, 91128 Palaiseau, France
b Institut de Recherches Servier (IdRS), 125 chemin de Ronde, 78290 Croissy sur Seine, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Variously substituted novel dihydropyridoazepinones have been prepared by an intermolecular radical
addition followed by a radical cyclisation on a pyridine ring. The latter process involved the use of a com-
bination of two different peroxides, an experimental contrivance resulting from a careful product analysis
and a better understanding of the cyclisation step.
Received 30 January 2012
Revised 23 March 2012
Accepted 4 April 2012
Available online 20 April 2012
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Radical cyclisation
Radical addition
Xanthates
Pyridoazepinones
Peroxides
Pyridine derivatives arguably represent a very important class
of heterocyclic structures. They are used in a wide variety of appli-
cations such as pharmaceuticals, agrochemicals, vitamins, food fla-
vourings, paints, dyes, rubber products, adhesives, etc. Thus, their
preparation has garnered much interest in both academic and
industrial settings.1 The search for innovative and efficient access
to this class of molecules is indeed one of the driving forces in this
field. The degenerative transfer of xanthates and related groups
discovered in our laboratory has proved efficient at accomplishing
notoriously difficult radical transformations such as intermolecular
additions to unactivated alkenes and ring-closure onto aromatic
rings.2 In this context, we have explored the synthesis of heterocy-
clic substructures using either a radical cyclisation onto the rele-
hydropyridoazepine ring (Scheme 1).4 To date, the most useful
method to construct this scaffold relies on the ortho-metalation
of an N-acyl-aminopyridine followed by alkylation with a dihaloal-
kane and in situ cyclisation. However, this method, introduced by
Spivey in 1998 for the synthesis of DMAP analogues,5 is hampered
by the limited access to suitable dihaloalkanes and by the low
functional group tolerance inherent to the use of strong lithiated
bases. We have reported a radical based approach to two examples
of 6,7-dihydro-5H-pyrido[2,3-b]azepin-8(9H)-one (the extra-nu-
clear nitrogen atom is at the 2-position of the pyridine), but further
investigation was required to generalise this strategy and to extend
it to the synthesis of other isomers in this family.2b
We became interested in molecules bearing the nitrogen substi-
tuent at the 4-position of the pyridine ring because these isomers
are virtually unknown in the literature.6 Our synthetic analysis re-
lies on a sequence consisting of an intermolecular radical addition
of a xanthate to an unactivated olefin such as 1 to furnish xanthate
2, followed by the radical cyclisation to give 3 (Scheme 2). The
chlorines at the 2- and 6-positions were chosen to reduce the
nucleophilicity and basicity of the pyridine nitrogen atom and to
serve as further points of diversification.2b A nucleophilic pyridine
slowly decomposes the xanthate group by an ionic process and the
resulting sulphur-containing products can interfere with the radi-
cal sequence.
vant aromatic ring or
a combination of an intermolecular
addition and a radical cyclisation. For example, azaindolines, diaza-
indolines, pyrimidinones, azaindoles, azaoxindoles, pyridones, tet-
rahydronaphthyridines and tetrahydroazaquinolones have all been
assembled using this technology in a straightforward fashion.3 Se-
ven-membered rings fused to a pyridine core, such as pyridoaze-
pines, are very scarce structures in the chemical literature,
reflecting the difficulty of their access. Yet, they can be valuable
targets, as demonstrated by the considerable effort invested by
the process research department at Merck to develop a practical
route to a non-peptidic
avb3 antagonist containing a tetra-
Initial studies revealed that the nitrogen atom at the 4-position
needed to be substituted for the cyclisation to occur; otherwise,
almost complete reduction of the xanthate function in 2 was
⇑
Corresponding author.
E-mail address: zard@poly.polytechnique.fr (S.Z. Zard).
0040-4039/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2012.04.020

L. Petit et al. / Tetrahedron Letters 53 (2012) 3220–3224
3221
OMe
N
N
1) HexLi, TMEDA
Cl
O
O
O
2)
H
N
H
N
O
I
N
N
Cl
N
CO2H
N
Cl
3) reflux
78%, 5.05 kg scale
O
Scheme 1. Merck process route to a tetrahydroazepine.
O
O
O
R1
R1
R1
N
R2
N
N
intermolecular
radical addition
S
OEt
R2
radical cyclisation
S
Cl
N
Cl
Cl
N
Cl
Cl
N
Cl
2
3
1
Scheme 2. Retrosynthetic analysis.
observed. While this failure may be attributed to the slow rotation
around the amide bond and the predominance of the geometrically
non-desired rotamer, this result must be contrasted with the sur-
prisingly successful formation of a six-membered ring starting
with an unsubstituted amide as in the first transformation shown
in Scheme 3.3a
a complex mixture of more polar by-products. These by-products
were identified by HPLC/MS and crude NMR analysis as dimeric
bicyclic compounds present as a mixture of regio and atropoi-
somers. For the sake of clarity only two of these, 5 and 6, are
shown (Scheme 4).
This was the first time such dimers were observed in the course
of a radical cyclisation onto a pyridine ring. The radical cyclisation
process usually entails four main steps. First, the radical is gener-
ated from the corresponding xanthate by the action of the peroxide
under thermal conditions. Second, the key cyclisation event leads
to a delocalised dihydropyridinyl radical. Third, the oxidation of
this radical by the peroxide (used in stoichiometric amounts) by
a one-electron transfer precedes the fourth and final rearomatisa-
tion steps through loss of a proton. In the present case, the oxida-
tion of the dihydropyridinyl radical is clearly a relatively slow
process. The lifetime of this radical intermediate is extended and
its concentration therefore increases to the point where self-
coupling becomes a very significant side reaction. Loss of two mol-
ecules of hydrochloric acid from the resulting dimer furnishes the
observed dimeric products 5 and 6 and their various isomers.
These mechanistic features are summarised in Scheme 5 (only
one isomer is shown).
The methyl group was chosen as the substituent on the
amide nitrogen. The preparation of olefin 1a was executed in
two steps. Reaction between 4-amino-2,6-dichloropyridine and
4-pentenoic acid chloride in the presence of diisopropylethyl-
amine provided the corresponding secondary amide in 97% yield.
Treatment of the latter with methyl iodide in acetone in the
presence of potassium carbonate in a sealed vessel gave olefin
1a in 93% yield. The intermolecular radical addition of the xan-
thate derived from chloroacetone proceeded smoothly under
standard conditions [sub-stoichiometric quantities of dilauroyl
peroxide (DLP) in refluxing ethyl acetate] furnishing xanthate
2a in a good yield (61%). With this new xanthate in hand, the
stage was set for the radical cyclisation. The first attempt using
stoichiometric amounts of DLP in refluxing ethyl acetate resulted
in a disappointingly poor yield (10%) of the desired heterocycle
3a, along with the prematurely reduced compound 4 (15%) and
OEt
O
O
O
S
S
S
HN
N
HN
N
HN
N
EtO
S
CO2Et
CO2Et
CO2Et
Cl
Cl
Cl
Cl
Cl
Cl
(lauroyl peroxide)
DCE, reflux
lauroyl peroxide
PhCl, reflux
51%
82%
O
HN
O
O
OEt
S
OEt
HN
N
HN
N
OEt
OEt
OEt
OEt
lauroyl peroxide
PhCl, reflux
S
S
OEt
OEt
S
Cl
Cl
Cl
Cl
Cl
Cl
N
Cl
Cl
(lauroyl peroxide)
AcOEt, reflux
O
OEt
OEt
74%
HN
N
Scheme 3. Reactivity contrast between the six and the seven-membered ring formation with an unsubstituted amide.

3222
L. Petit et al. / Tetrahedron Letters 53 (2012) 3220–3224
O
Me
N
N
Cl
Cl
O
S
OEt
O
61%
Me
Me
O
DLP
S
O
O
O
Me
Me
N
N
Me
N
Me
H
DLP, AcOEt
S
OEt
Me
reflux
Cl
O
N
Cl
Cl
N
Cl
S
O
Cl
N
Cl
3a
4
2a
O
O
Me
Me
Cl
N
Me
Cl
N
N
Me
N
N
N
Cl
Cl
O
N
Me
Me
O
O
5
6
Me
N
O
Me
O
Scheme 4. Products from a first cyclisation attempt.
O
O
O
O
Me
Me
Me
N
Me
N
N
R
N
N
H
H
oxidation
reversible
cyclisation
R
R
R
DLP
Cl
Cl
Cl
N
Cl
Cl
N
Cl
Cl
N
Cl
H
-
DLP
3
dimerisation
reduction
O
O
O
O
O
Me
Me
Me
Me
Me
N
R
N
N
N
N
H
Cl
N Cl
N
N
N
Cl
N
Cl
R
R
R
R
Cl
Cl
Cl
Cl
-2 HCl
4
5
Scheme 5. Mechanistic outcomes.
The lifetime of a peroxide is exponentially dependent on the
temperature. Thus, the half-life of lauroyl peroxide (DLP) is
60 min at 85 °C. It is extended to 10 hours at 65 °C. Therefore, by
decreasing the temperature of the reaction mixture to 73–75 °C,
the quantity of DLP available for the oxidation process is greater.
Improved yields of 3a varying from 20% to 25% were attained using
this straightforward modification. However, these yields were still
far from being useful from a preparative standpoint. We therefore
decided to incorporate a co-oxidant to accelerate the formation of
the dihydropyridinyl cation and shut down more effectively the
unwanted path to the dimers. Dibenzoylperoxide (DBP) has a
half-life of 60 min at 95 °C and is known to be a better oxidant than
lauroyl peroxide.7 By keeping the lower temperature of 73–75 °C,
only the lauroyl peroxide will thermally decompose and start the
radical process, whereas the benzoyl peroxide will be present in
a sufficiently large concentration to accomplish the desired oxida-
tion of the dihydropyridinyl radical intermediate. Indeed, we found
that using 1.1 equiv of DBP and adding 20 mol % of DLP every
60 min at 73–75 °C limited strongly the formation of the dimers
and provided the desired dihydropyridoazepinone 3a in 60% yield.8
It should be noted that it is really the synergistic effect of both per-
oxides that allows the cyclisation to proceed, as a blank experi-
ment with benzoyl peroxide alone was not successful. Even if
this yield is rather moderate, it represents a six-fold increase in
comparison with the initial yield observed under non-optimised
conditions. It must also be remembered that cyclisations onto aro-
matic structures leading to the direct formation of fused seven-
membered rings are inherently difficult and almost impossible to
accomplish with classical methods. The optimised conditions were
then applied to the preparation of a series of novel pyridoazepi-
nones starting from compound 1a and the results are displayed
in Table 1.
The reaction conditions proved to be compatible with a variety
of functionalities and allowed the straigthforward synthesis of a
series of original heterocycles. Ketones, esters, malonates and even
sensitive a-chloroketones were well tolerated. The use of xanthate
derived from an acetophenone represents however a limitation of
the method. The radical generated from xanthate 2j can cyclise

L. Petit et al. / Tetrahedron Letters 53 (2012) 3220–3224
3223
Table 1
Intermolecular additions and subsequent cyclisations
Xanthate
Adduct (isolated yield)
Bicycle (isolated yield)
O
O
O
O
EtO
EtO
EtO
EtO
S
S
S
Me
Me
Me
Me
N
Me
N
N
Me
S
S
S
S
S
OEt
Me
S
Cl
Cl
Cl
Cl
Cl
N
Cl
O
2a (61%)
3a (60%)
O
O
O
O
O
O
CO2Et
CO2Et
Me
N
N
N
S
OEt
CO2Et
S
Cl
Cl
N
Cl
O
2b (49%)
3b (45%)
O
O
O
OPh
OPh
Me
N
N
N
S
OEt
OPh
S
Cl
Cl
N
Cl
O
2c (70%)
EtO2C
3c (56%)
CO2Et
O
O
S
Me
OEt
Me
Me
N
N
N
CO2Et
CO2Et
O
OEt
S
OEt
S
Cl
N
Cl
Cl
Cl
Cl
3d (53%)
O
2d (69%)
O
O
O
EtO
S
O t
-
Bu
N
O-tBu
Me
N
S
S
S
OEt
t
CO2 Bu
S
N
Cl
Cl
N
Cl
2e (71%)
3e (44%)
O
O
O
O
EtO
EtO
S
S
Me
Me
N
N
S
OEt
Decomposition
S
Cl
Cl
Cl
2f (52%)
O
O
O
Cl
Cl
Me
N
N
N
S
S
S
OEt
Cl
S
Cl
Cl
N
Cl
O
2g (81%)
3g (48%)
OBz
BzO
CN
O
O
Me
EtO
S
CN
Me
N
N
N
OBz
CN
S
OEt
S
Cl
N
Cl
Cl
Cl
3h (41%)
2h (79%)
(continued on next page)

3224
L. Petit et al. / Tetrahedron Letters 53 (2012) 3220–3224
Table 1 (continued)
Xanthate
Adduct (isolated yield)
Bicycle (isolated yield)
O
O
O
EtO
S
Me
N
N
S
S
OEt
1:1:1:1 inseparable mixture of starting material,
reduced product, tetralone and dihydropyridoazepinone
F
F
S
Cl
Cl
2j (47%)
3. (a) Bacqué, E.; El Qacemi, M.; Zard, S. Z. Heterocycles 2012, 84, 291; (b) Bacqué,
E.; El Qacemi, M.; Zard, S. Z. Org. Lett. 2004, 6, 3671. See also: (c) Laot, Y.; Petit, L.;
Tran, N. D. M.; Zard, S. Z. Aust. J. Chem. 2011, 64, 416; (d) Laot, Y.; Petit, L.; Zard, S.
Z. Org. Lett. 2010, 12, 3426; (e) Laot, Y.; Petit, L.; Zard, S. Z. Chem. Commun. 2010,
46, 5784; (f) El Qacemi, M.; Ricard, L.; Zard, S. Z. Chem. Commun. 2006, 4422.
4. Keen, S. P.; Cowden, C. J.; Bishop, B. C.; Brands, K. M. J.; Davies, A. J.; Dolling, U.
H.; Lieberman, D. R.; Stewart, G. W. J. Org. Chem. 2005, 70, 1771.
5. Spivey, A. C.; Fekner, T.; Adams, H. Tetrahedron Lett. 1998, 39, 8919.
6. (a) Moehrle, H.; Dwuletzki, H. Chem. Ber. 1986, 119, 3600; (b) Stukenbrock, H.;
Mussmann, R.; Geese, M.; Ferandin, Y.; Lozach, O.; Lemcke, T.; Simone Kegel, S.;
Lomow, A.; Burk, U.; Dohrmann, C.; Meijer, L.; Austen, M.; Kunick, C. J. Med.
Chem. 2008, 51, 2196; (c) Egert-Schmidt, A.-M.; Dreher, J.; Dunkel, U.; Kohfeld,
S.; Preu, L.; Weber, H.; Ehlert, J. E.; Mutschler, B.; Totzke, F.; Schächtele, C.;
Kubbutat, M. H. G.; Baumann, K.; Kunick, C. J. Med. Chem. 2010, 53, 2433.
7. Fraind, A.; Turncliff, R.; Fox, T.; Sodano, J.; Ryzhkov, L. R. J. Phys. Org. Chem. 2011,
24, 809. and references cited therein.
8. Typical cyclisation procedure: A three-necked round bottomed flask equipped
with a water condenser and a thermometer was charged with the appropriate
xanthate 2 (1.0 mmol). EtOAc (15 mL) was added and the solution was refluxed
for 10 min under a nitrogen atmosphere. The temperature of the solution was
then decreased to 73–75 °C and dibenzoylperoxide (1.1 mmol) was added in one
portion. Lauroyl peroxide was then added (20 mol %) and additional portions
(20 mol %) were added every 60 min until total consumption of the xanthate.
The mixture was then cooled to room temperature and the solvent was removed
under reduced pressure. The residue was purified by flash chromatography on
silica gel to yield the corresponding dihydropyridoazepinone. An example is the
synthesis of 6,8-dichloro-1-methyl-5-(3-oxo-4-phenoxybutyl)-4,5-dihydro-1H-
pyrido[4,3-b]azepin-2(3H)-one (3c): Following the general procedure the
radical cyclisation was carried out starting with xanthate 2c (300 mg,
0.538 mmol) in EtOAc (8 mL). Dibenzoyl peroxide (75% purity, the remainder
is water; 195 mg, 1.1 equiv.) was used and the reaction needed 100 mol % of DLP
to go to completion. Purification by flash chromatography on silica gel
(petroleum ether/EtOAc 1:1) yielded bicycle 3c (123 mg, 56%). 1H NMR
(400 MHz, CDCl3): dH 8.03 (d, J = 7.3 Hz, 2H), 7.59 (t, J = 7.5 Hz, 1H), 7.43 (t,
J = 7.7 Hz, 2H), 7.05 (s, 1H), 4.73 (s, 2H), 3.71 (m, 1H), 3.30 (s, 3H), 2.47–2.35 (m,
3H), 2.32–2.25 (m, 2H), 2.17–2.14 (m, 1H), 2.02–1.91 (m, 2H); 13C NMR
(100 MHz, CDCl3): dc 202.6 (C), 172.1 (C), 165.7 (C), 155.1 (C), 151.5 (C), 149.1
(C), 133.4 (CH), 129.8 (2 CH), 129.0 (C), 128.5 (2 CH), 117.5 (CH), 68.2 (CH2), 37.2
onto the phenyl ring to generate the corresponding tetralone or
onto the pyridine ring and no chemoselectivity was observed. Sur-
prisingly, the cyclopropane moiety was readily introduced by the
intermolecular radical addition (compound 2f), but did not survive
the cyclisation reaction conditions. The reasons underlying this
behaviour are not clear at the moment.
In summary, we have developed a new radical based method for
the preparation of 5-substituted 4,5-dihydro-3H-pyrido[4,3-b]aze-
pin-2-ones. This sequence complements nicely the rare existing ap-
proaches and allows access to a family of structures not readily
available hitherto. We also demonstrated the large scope of this
procedure by applying it to the synthesis of a series of novel pyr-
idoazepinone derivatives. This process allows the introduction, at
the 5-position of the pyridoazepinone, of a variety of functional
groups, and these can be further elaborated into unique, more com-
plex structures. From a more fundamental perspective, this work
has provided us with a better appreciation of some subtle aspects
of the ring-closure mechanism and perhaps a means to improve
the efficacy of radical cyclisations onto aromatic rings by the use
of a combination of cheap, commercially available peroxides.
References and notes
1. For reviews, see: (a) Henry, G. D. Tetrahedron 2004, 60, 6043; (b) Popowycz, F.;
Routier, S.; Joseph, B.; Mérour, J.-Y. Tetrahedron 2007, 63, 1031; (c) Popowycz, F.;
Mérour, J.-Y.; Joseph, B. Tetrahedron 2007, 63, 8689; (d) Song, J. J.; Reeves, J. T.;
Gallou, F.; Tan, Z.; Yee, N. K.; Senanayake, C. H. Chem. Soc. Rev. 2007, 36, 1120; (e)
Schirok, H. J. Org. Chem. 2006, 71, 5538; (f) Humphrey, G. R.; Kuethe, J. T. Chem.
Rev. 2006, 106, 2875; (g) Cacchi, S.; Fabrizi, G. Chem. Rev. 2005, 105, 2873.
2. (a) Quiclet-Sire, B.; Zard, S. Z. Pure Appl. Chem. 2011, 83, 519; (b) Quiclet-Sire, B.;
Zard, Z. S. Chem. Eur. J. 2006, 12, 6002; (c) Quiclet-Sire, B.; Zard, Z. S. Top. Curr.
Chem. 2006, 264, 201; (d) Zard, S. Z. In Radicals in Organic Synthesis; Renaud, P.,
Sibi, M. P., Eds.; Wiley-VCH: Weinheim, 2001; Vol. 1, p 90; (e) Quiclet-Sire, B.;
Zard, S. Z. Phosphorus, Sulfur Silicon 1999, 153–154, 137; (f) Quiclet-Sire, B.; Zard,
S. Z. J. Chin. Chem. Soc. 1999, 46, 139; (g) Zard, S. Z. Angew. Chem., Int. Ed. 1997, 36,
672.
(CH), 36.8 (CH2), 35.1 (CH3), 32.8 (2 CH2), 27.2 (CH2); IR (CCl4): mmax 3155, 2935,
1730, 1688, 1567, 1272. HRMS (EI+): Calcd for C20H20O3N2Cl2: 406.0851. Found
406.0860.