Rapid entry into heterocycle-fused benzylic azepines and azocines via directed metallation/ring-closing metathesis

Article abstract of DOI:10.1016/j.tetlet.2013.05.142

The efficient synthesis of a range of heterocycle-fused benzylic azepine and azocine derivatives is reported, employing a directed metallation/ruthenium- catalysed ring-closing metathesis approach. A base-mediated tautomerisation approach can be used to access both the azepine and azocine derivatives from the same starting material.

Full text of DOI:10.1016/j.tetlet.2013.05.142

Tetrahedron Letters 54 (2013) 4254–4258
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Rapid entry into heterocycle-fused benzylic azepines and azocines
via directed metallation/ring-closing metathesis
Thomas A. Moss
AstraZeneca Mereside, Alderley Park, Cheshire SK10 4TG, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
The efficient synthesis of a range of heterocycle-fused benzylic azepine and azocine derivatives is
reported, employing a directed metallation/ruthenium-catalysed ring-closing metathesis approach. A
base-mediated tautomerisation approach can be used to access both the azepine and azocine derivatives
from the same starting material.
Received 2 April 2013
Revised 17 May 2013
Accepted 30 May 2013
Available online 10 June 2013
Ó 2013 Elsevier Ltd. All rights reserved.
Keywords:
Metathesis
Azepine
Azocine
Heterocycle
Bicycle
The efficient construction of medium sized nitrogen-containing
rings is an ongoing challenge for synthetic chemists, as the entro-
pic penalty for ring closure is typically higher than for five- and
six-membered systems. Nonetheless, fused nitrogen-containing
seven- (azepine) and eight- (azocine) membered ring systems are
common in Nature, being found in a number of bioactive molecules
and pharmaceuticals. These include natural products such as the
galanthamine family1 and apparicine,2 as well as promising
medicinal chemistry targets3,4 (Fig. 1).
Due to their attractive structures, fused 2-benzazocines have
been investigated previously in the literature, with methods such
as radical chain insertion,5 electrophilic cyclisation6 and Schmidt
rearrangement7 being employed to build the saturated ring. A
powerful method for synthesising medium sized nitrogen-contain-
ing ring systems is the ruthenium-catalysed ring-closing metathe-
sis reaction (RCM),8 which has been widely utilised for the
synthesis of benzo-fused azepines and azocines. However one area
that remains poorly developed is the synthesis of heterocycle-
fused ring systems, in particular those which contain key molecu-
lar binding sites such as pyrimidines, pyridines and azoles. As part
of our ongoing interest into the synthesis of partially saturated
fused ring systems,9 we previously described the efficient
construction of various heterocycle-fused 1-azepine derivatives.10
As detailed in Figure 1, fused 2-azepines and 2-azocines are an
important target in both natural product and medicinal chemistry.
To this point, the majority of attention amongst academic groups
has been directed at the synthesis of 1-benzazepine derivatives
by RCM,11 including some examples of heterocycle-fused 1-ben-
zazepines.12 With this in mind, we embarked upon a project to
investigate the efficient construction of fused 2-azepines and -
azocines. Our initial aim was to build the saturated ring through
a
directed metallation/ring-closing metathesis strategy, as
described previously by Snieckus and co-workers for the synthesis
of benzazepines and benzazocine derivatives13 (Scheme 1).
Although this method was reported to work well for substituted
phenyl derivatives, heteroatom-containing examples were not
explored. We anticipated that moving from simple phenyl deriva-
tives to heterocycles would present a number of challenges, in
particular the sensitivity of the substrates to metallation. Further-
O
N
O
(R)
O
O
N
OH
OH
F
N
H
N
O
N
H
tyrosine kinase inhibitor3
(-) galanthamine1
Ph
O
OtBu
N
NH
O
O
S
N
OH
N
H
F
apparicine2
HIV protease inhibitor 4
Figure 1. Examples of fused azepines and azocines found in nature and in
E-mail address: thomas.moss@astrazeneca.com
medicinal chemistry.
0040-4039/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2013.05.142

T. A. Moss / Tetrahedron Letters 54 (2013) 4254–4258
4255
O
O
DMG
DMG
DMG
O
RCM
R
R
NaH, DMF
90%
R
n-BuLi, THF
N
N
N
N
Het
45%
S
Cl
DMG = NHBoc, CONHMe, SO2NHEt
4
11
Scheme 1. Directed metallation/RCM strategy described by Snieckus.
Scheme 3. Tautomerisation of C-allyl substrates.
consistent with over-reduction. Alternatively, attempted metalla-
tion/allylation of the benzylic CH2N(Boc)allyl pyridine resulted
largely in alkylation on the allylic methylene. The desired product
was accessed eventually by reductive amination of 4-iodopyridine
6a with allylamine followed by N-Boc protection to give 6b, then a
magnesium-halogen exchange with iPrMgCl17 and quenching with
allyl bromide to give 6 (Scheme 4). As with other examples, the 4-
allyl moiety could be tautomerised to the styrene 7 by stirring with
NaH in DMF.
more, we aimed to synthesise RCM products containing handles for
further elaboration (e.g., halogens for Pd-catalysed cross-
couplings), which could potentially pose problems with the
competitive metal–halogen exchange pathway. With these factors
in mind, we began our studies into the synthesis of the RCM
precursors. Allyl amides and allyl sulfonamides were intially
chosen as the metallation directing groups, as these could easily
be accessed in near quantitative yields from the carboxylic acid
and sulfonyl chlorides, respectively. In the case of carboxylic acid
derivatives, formation of the acid chloride with oxalyl chloride
prior to treatment with allylamine proved to be cleaner and quick-
er than direct amide coupling methods. The C-allyl group could
then be inserted by ortho-metallation with n-BuLi, followed by
quenching with allyl bromide. In most cases, it was necessary to
perform a transmetallation with copper(I) bromide prior to the
addition of the electrophile in order to attenuate the high basicity
of the lithium species. The more sensitive pyridine substrates
turned out not to be stable to n-BuLi, however, metallation could
be readily performed using milder TMPMgClÁLiCl (TMP = 2,2,6,6-
tetramethylpiperidide).14 This had the added benefit that carbon–
halogen bond insertion did not occur with this base, allowing aryl
bromide substrates to be accessed. Once C-allylation had been
performed, the RCM reaction could be carried out directly, or the
amide could be N-protected under standard conditions in cases
where the free NH turned out to be detrimental to RCM (Scheme 2).
Pleasingly, this approach worked well for a range of heteroaryl
substrates, including pyridines, thiophenes, pyrroles and pyrazoles,
giving the azocine RCM precursors in good overall yields. Further-
more, by exploiting the relative acidity of the C-allyl group, several
substrates were successfully tautomerised into the styrenyl
compounds by treatment with NaH (for pyridines) or n-BuLi. This
approach is particularly useful as it allows both the azocine and
azepine RCM products to be accessed from a single starting
material and provides a useful alternative to the traditional
formylation/Wittig15 and iodination/Stille16 pathways to these
substrates (Scheme 3).
With the precursors in hand, we began our studies into the RCM
reaction. Table 1. Amido-pyridine 1 reacted sluggishly when trea-
ted with 5 mol% of Grubbs (II) catalyst in CH2Cl2, giving only traces
of the RCM product 17 after heating to reflux for 4 h. Further heat-
ing to 80 °C in toluene gave mostly the competitive cross-metath-
esis product rather than the RCM product. Pleasingly, protection of
the NH with either Boc or methyl did allow the RCM reaction to oc-
cur. N-Boc precursor 2 gave the azocine 17 in reasonable yield after
Boc group cleavage (TFA in CH2Cl2), although significant amounts
of the cross-metathesis product were still seen. The formation of
this unwanted side-product could be suppressed to a degree by
increasing the reaction dilution from 0.1 to 0.02 M (47% vs 60%
yield of 17), however further dilution was not attempted as it
would limit the practicality of the reaction. Interestingly, N-Me
precursor 3 reacted significantly more smoothly in the RCM reac-
tion than the N-Boc compound 2, giving methylated azocine 18
in an 81% yield, with only a small amount of the cross-metathesis
product observed in the reaction mixture. Tautomerised pyridine 4
reacted more quickly in the RCM reaction than the allyl compound
giving azepine 19 in a 63% yield after N-Boc cleavage. This is not
entirely surprising as seven-membered ring formation is generally
more favourable than eight-membered systems. When a methyl
group was introduced onto the C-allyl (compound 5), only cross-
metathesis via the N-allyl was observed, highlighting the sensitiv-
ity of these substrates to steric contraints. The absence of the
amide carbonyl in precursors 6 and 7 resulted in much smoother
RCM reactions, giving azocine 21 and azepine 22, respectively, in
excellent yields (93% and 85%). Gratifyingly, sulfonamides 8 and
9 not only reacted smoothly in the RCM reaction without the need
for N-protection, but they also tolerated a methylallyl moiety,
giving methylated azocine 24 in an excellent 89% yield. The reac-
tion was then extended to include five-membered heterocycles.
Amido-thiophenes 10 and 11 worked reasonably well in the
RCM, as did amido-furan 13. As with the pyridine examples, the
sulfonamide analogues reacted considerably more smoothly than
the amide series, with thiophene 12, pyrazole 14 and pyrrole 15
giving the sulfonamide azocines in high yields (70–84%).
Sulfonamido-azaindole 16 appeared to react somewhat slowly in
the RCM, requiring prolonged heating for 16 h to achieve a 25%
conversion into tricyclic azocine 31.
We also wanted to evaluate benzylic examples (i.e., with the
amide reduced), as such structures appear in several natural
products (e.g., galanthamine, apparacine, see Fig. 1). Unfortunately,
direct reduction of the amide did not prove to be a reliable reac-
tion. Heating pyridine 1 with LiAlH4 in THF was required to achieve
the initial reduction of the amide, with the second reduction of the
amino-alcohol intermediate being somewhat sluggish. Prolonged
heating or addition of further reductant appeared to give products
O
H
O
O
H
a
c
N
H
N
H
OH
Het
Het
Het
Het
Het
Het
I
I
O
O
O
O
O
O
CHO
Cl
S
H
a, b
S
S
H
c
N
Boc
b
c
N
H
N
H
Cl
N
Boc
N
N
Cl
6b
N
Cl
6
6a
Scheme 2. Initial route to heterocycle-fused azocines and azepines. Reagents and
conditions: (a) (COCl)2, CH2Cl2, DMF (cat.), then allylamine, Et3N; (b) allylamine,
Et3N, CH2Cl2; (c) n-BuLi, CuBr, allyl bromide, THF.
Scheme 4. Reagents and conditions: (a) allylamine, MeOH, rt, then NaBH4, 77%; (b)
Boc2O, MeOH, rt, 91%; (c) PrMgCl, THF, À30 °C–rt, then allyl bromide, 70%.
i

4256
T. A. Moss / Tetrahedron Letters 54 (2013) 4254–4258
Table 1
Scope of the RCM reactiona
X
NH
i) Base tautomerisation
ii) Grubbs (II) 5 mol%
X
R
X
NH
N
Grubbs (II) 5 mol%
Het
Het
Het
CH2Cl2, 45 °C
0.02 M
CH2Cl2, 45 °C
0.02 M
n
n
n
1 16
-
Entry
1
Starting material
Time (h)
Product
Yieldb (%)
trace
O
O
O
O
NH
NH
N
N
N
NH
N
4
Cl
Cl
Cl
Cl
Cl
Cl
17
17
1
O
O
Boc
N
N
2c
4 Then TFA/CH2Cl2, 1 h
60
81
2
N
N
N
3
1.5
Cl
Cl
18
19
3
O
O
O
Boc
NH
NH
N
N
N
N
N
4c
1.5 then TFA/CH2Cl2, 1 h
6 then TFA/CH2Cl2, 1 h
63
0d
4
5
O
Boc
N
5
Cl
N
Cl
N
20
21
Cl
Boc
Cl
N
Boc
N
N
6
7
8
1
1
2
93
85
90
6
7
Cl
Cl
Boc
Boc
O
N
N
N
22
O
O
O
O
S
S
NH
N
NH
N
Cl
Cl
Cl
Cl
Br
Br
23
Br
8
9
O
O
O
S
S
NH
N
O
NH
N
9
1.5
89
60
Br
24
H
O
Boc
N
N
10c
4 then TFA/CH2Cl2, 1 h
25
S
S
10

T. A. Moss / Tetrahedron Letters 54 (2013) 4254–4258
4257
Table 1 (continued)
Entry
Starting material
Time (h)
Product
Yieldb (%)
O
O
Boc
NH
N
11c
12
3 then TFA/CH2Cl2, 1 h
67
78
79
70
S
S
26
11
O
O
O
S
H
N
O
S
S
NH
2
S
27
12
13
14
O
H
N
O
Boc
N
13c
14
4 then TFA/CH2Cl2, 1 h
28
29
O
O
O
O
S
H
N
O
O
S
NH
3
N
N
N
N
O
S
O
S
H
N
O
O
NH
15
1
84
N
N
15
30
SO2Ph
SO2Ph
O
S
O
S
H
N
O
N
O
N
N
NH
16
16
(25)e
N
31
16
a
Reactions were performed on 1.0 mmol scale with 50 mL of CH2Cl2 (0.02 M) at 45 °C.
Isolated yield after silica chromatography.
N-Boc cleavage was performed withTFA/CH2Cl2.
b
c
d
e
Cross-methathesis product isolated as the only product.
Conversion according to the 1H NMR spectrum, product was not isolated.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.tetlet.2013.05.
142.
Boc
N
Boc
N
Cl
Cl
H2 / Pd (cat.)
N
N
EtOAc, RT, 1 h
91%
References and notes
Scheme 5. Selective reduction of the alkene moiety.
1. (a) Chen, G.-Q.; Xie, J.-H.; Bao, D.-H.; Liu, S.; Zhou, Q.-L. Org. Lett. 2012, 14, 2714;
(b) Peng, C.; Xu, B.; Zhang, L.-F.; Ding, M.; Han, X.-J.; Li, J.; Zhang, G.-B.; Tu, Y.-
Q.; Fan, C.-A. Angew. Chem., Int. Ed. 2011, 50, 8161.
2. (a) Kettle, J. G.; Roberts, D.; Joule, J. A. Heterocycles 2010, 82, 349; (b) Bennasar,
M.-L.; Zulaica, E.; Sole, D.; Roca, T.; Garcia-Diaz, D.; Alonso, S. J. Org. Chem. 2009,
74, 8359.
3. Peng, C. T.; Yi, D. S.; Feng, J.; Jian, H. F.; Jiang, L. Y.; Xiao, L.; Jiang, H. P.; Ya, L. L.;
Zhang, L.; Hu, B.; Zhou, Y.; Fang, Q. L.; Bei, B. F.; Li, G. L.; Ai, S. G.; Gao, H. S.; Wei,
H. S.; Xian, T. M. J. Med. Chem. 2010, 53, 8140.
4. Ganguly, A. K.; Alluri, S. S.; Caroccia, D.; Biswas, D.; Wang, C.-H.; Kang, E.;
Zhang, Y.; McPhial, A. T.; Carroll, S. S.; Burlein, C.; Munshi, V.; Orth, P.;
Strickland, C. J. Med. Chem. 2011, 54, 7176.
5. Benati, L.; Nanni, B.; Sangiorgi, C.; Spagnolo, P. J. Org. Chem. 1999, 64, 7836.
6. Homsi, F.; Rousseau, G. J. Org. Chem. 1998, 63, 5255.
7. Ortega, R.; Ravina, E.; Masaguer, C. F.; Areias, F.; Brea, J.; Loza, M. I.; Lopez, L.;
Selent, J.; Pastor, M.; Sanz, F. Bioorg. Med. Chem. Lett. 2009, 19, 1773.
8. For a review, see: Chattopadhyay, S. K.; Karmakar, S.; Biswas, T.; Majumdar, K.
C.; Rahaman, H.; Roy, B. Tetrahedron 2007, 63, 3919. and references therein.
9. Moss, T. A.; Hayter, B. R.; Hollingsworth, I. A.; Nowak, T. Synlett 2012, 23, 2408.
10. Moss, T. A. Tetrahedron Lett. 2013, 54, 993.
11. For recent examples see: (a) Ramachary, D. B.; Narayana, V. V. Eur. J. Org. Chem.
2011, 3514; (b) Ghosh, D.; Thander, L.; Ghosh, S. K.; Chattopadhyay, S. K. Synlett
2008, 3011; (c) Kotha, S.; Shah, V. R. Eur. J. Org. Chem. 2008, 6, 1054.
12. For examples of pyrimido-fused azepines by RCM, see: (a) Kaïm, L. E.; Grimaud,
L.; Oble, J. J. Org. Chem. 2007, 72, 5835; (b) Majumdar, K. C.; Mondal, S.; Ghosh,
D. Synthesis 2010, 1176.
With the azocine and azepine products in hand, we decided to
demonstrate that the alkene formed in the product could be
reduced selectively to the fully saturated product in the presence
of aryl halides. Pleasingly, standard hydrogenation of 21 with Pd/
C and H2 in EtOAc at room temperature gave the saturated system
32 in 91% yield after just one hour, with no observable reduction of
the aryl chloride (Scheme 5).
In conclusion, the ruthenium-catalysed ring-closing metathesis
reaction is shown to be an extremely useful transformation in the
synthesis of heterocyclic-fused 2-azocine derivatives.18 Further-
more, we have demonstrated that through a base-induced tauto-
merisation, the 2-azepine products can be accessed from the
same starting materials. The products contain multiple handles
for further elaboration, and as such, constitute exceptionally versa-
tile synthetic building blocks. This transformation should be of
interest to the synthetic community in general, and particularly
in medicinal chemistry campaigns.

4258
T. A. Moss / Tetrahedron Letters 54 (2013) 4254–4258
13. (a) Lane, C.; Snieckus, V. Synlett 2000, 1294; (b) Stefinovic, M.; Snieckus, V. J.
Org. Chem. 1998, 63, 2808.
14. For a review, see: Haag, B.; Mosrin, M.; Ila, H.; Malakhov, V.; Knochel, P. Angew.
Chem., Int. Ed. 2011, 50, 9794.
15. Rios, R.; Jimeno, C.; Carroll, P. J.; Walsh, P. J. J. Am. Chem. Soc. 2002, 124, 10272.
16. Nora de Souza, M. V.; Dodd, R. H. Heterocycles 1998, 47, 811.
17. Abarbri, M.; Thibonnet, J.; Bérillon, L.; Dehmel, F.; Rottländer, M.; Knochel, P. J.
Org. Chem. 2000, 65, 4618.
18. Representative procedure for the synthesis of 23: Grubbs(II) catalyst (42.4 mg,
0.05 mmol, 5 mol %) was added to a solution of 8 (351 mg, 1.00 mmol) in
anhydrous CH2Cl2 (50 mL) at room temperature. The reaction was warmed to
reflux for 2 h. After cooling to room temperature, the volatiles were removed
under vacuum. [Note: for Boc amide examples, the N-Boc group was cleaved
with 1:1 CH2Cl2/TFA (5 mL) and the residue made basic with aq NaHCO3
solution.] The crude product was purified by silica gel chromatography
[hepane/EtOAc, 9:1–1:1 to give (Z)-7-bromo-8-chloro-3,6-dihydro-2H-
pyrido[4,3-g][1,2]thiazocine 1,1-dioxide 23 as a colourless solid (291.3 mg,
90%). Mp 230–231 °C; mmax (cmÀ1) 3350 (br), 2995 (w), 1653 (m); 1540 (m),
1362 (s), 1275 (m), 1175 (s), 1080 (m); 1H NMR [400 MHz, CDCl3] 3.75–4.05
(m, 4H, 2Â CH2), 5.88 (dt, 1H, J = 6.8, 10.0, CH), 6.04 (m, 1 H, CH), 8.15 (br s, 1H,
NH), 8,74 (s, 1H, CH Ar); 13C NMR [100 MHz, CDCl3] 32.4, 38.6, 123.3, 129.8,
130.7, 138.1, 146.8, 149.0, 154.0; m/z (ES) 323 ([M+H]+, 100); HRMS (EI) m/z
[M+H]+ calcd for C9H9O2N2BrClS: 322.92512; found; 322.92416.