Unprecedented FeCl3·6 H2O-promoted skeleton rearrangement of 1-Aryl2,3,4,5-tetrahydro-1H-3-benzazepines: A new strategy for the synthesis of C1 quaternary tetrahydroisoquinolines

Article abstract of DOI:10.1002/chem.200901197

A study was conducted to demonstrate that 1-aryl-tetrahydroisoquinolines were prepared through a skeleton-rearrangement promoted by FeCl ₃·6H₂O. 7,8-dimethoxy-3-methyl-1-(m-tolyl)-2,3,4,5- tetrahydro-1H-benzo[d]azepine was selected a

Full text of DOI:10.1002/chem.200901197

COMMUNICATION
DOI: 10.1002/chem.200901197
Unprecedented FeCl3·6H2O-Promoted Skeleton Rearrangement of 1-Aryl-
2,3,4,5-tetrahydro-1H-3-benzazepines: A New Strategy for the Synthesis of
C1 Quaternary Tetrahydroisoquinolines
Jing Zhang and Ao Zhang*[a]
The tetrahydroisoquinoline (THIQ) bicyclic system is an
important structure unit which occurs in a variety of natural
alkaloids and synthetic analogues (Figure 1).[1] These com-
ate b-phenylethylamide or an aminoacetal, or by condensa-
tion of a b-phenylethylamine with a carbonyl reagent (alde-
hyde, ketone) under acidic conditions. Although some new
modifications have been reported in recent years, these syn-
thetic methods are still limited to the preparation of struc-
tural simple THIQs. A general strategy involving skeleton-
rearrangement of 1-aryl-2,3,4,5-tetrahydro-1H-3-benzaze-
pines has not been discovered, especially for the synthesis of
THIQs containing a quaternary C1 center.
1-Aryl-2,3,4,5-tetrahydro-1H-benzazepines represent an
important class of compounds with dopamine D1 receptor
agonistic or antagonistic activities (Figure 1).[8] For example,
SKF 82526 (Fenoldopam) is a specific dopamine receptor
agonist, and used clinically as an antihypertensive agent.
SCH 23390, another benzazepine analogue, is a typical D1
receptor antagonist, widely used as a tool drug.[8,9] General-
ly, these compounds can be readily prepared by reaction of
an b-phenylethylamine and styrene oxide followed by acid-
assisted cyclization.[10] In the pursuit of novel benzazepine
analogues as D1 receptor ligands,[2,8,10] we recently found
that 1-aryl-2,3,4,5-tetrahydro-1H-benzazepines upon treating
with FeCl3·6H2O/PhNO2 can undergo a unprecedented skel-
eton rearrangement leading to a series of THIQs. Although
oxidative skeleton rearrangement of several specific benza-
zepines has been reported in literatures,[11] our current com-
munication showed for the first time that 1-aryl-tetrahydroi-
soquinolines can be prepared through a skeleton-rearrange-
ment promoted by FeCl3·6H2O. More interestingly, structur-
ally complicated THIQs containing a quaternary C1 center
equipped with a formyl function can be also prepared with
this novel protocol.
Figure 1. Representative structures of THIQs and benzazepines.
pounds often possess significant bioactivities and were re-
ported to be involved in some pathologies of CNS disorders
and cardiovascular diseases.[2–4] Therefore, synthesis of tetra-
hydroisoquinolines (THIQs) has been and is still a valuable
objective for both synthetic and medicinal chemists.[4]
Traditionally, there are three methods that are widely
used to construct THIQs, including Bischler–Napieralski
cyclization/reduction, Pictet–Spengler synthesis and Pomer-
anz–Fritsch synthesis.[5–7] These methods generally involve
the formation of a THIQ ring by cyclization of an appropri-
First, 7,8-dimethoxy-3-methyl-1-(m-tolyl)-2,3,4,5-tetrahy-
dro-1H-benzo[d]azepine (1a) was chosen as the prototypic
mode substrate for reaction condition optimization. The re-
arrangement was conducted in different solvents with
FeCl3·6H2O (2 equiv) as the activator. A significant differ-
ence caused by various solvents was observed (see Support-
ing Information). The best result was obtained with PhNO2
as the solvent, leading to the rearranged product, 1-aryl-tet-
[a] J. Zhang, Prof. Dr. A. Zhang
Synthetic Organic & Medicinal Chemistry Laboratory (SOMCL)
Shanghai Institute of Materia Medica (SIMM)
Chinese Academy of Sciences, 555 Zuchongzhi Road
Shanghai 201203 (P.R. China)
Fax : (+86)21-50806035
E-mail: aozhang@mail.shcnc.ac.cn
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200901197.
Chem. Eur. J. 2009, 15, 11119 – 11122
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11119

rahydroisoquinoline (2a[12]) in 63% yield. Other substituted
benzenes (PhBr, PhF, PhCN, toluene, xylene), nitromethane,
isopropanol, or frequently used solvents (CH3CN, THF,
DMF) either did not initiate the reaction, or provided the
products in trace yield (<5%). Thus, PhNO2 was chosen as
the best choice of solvent for iron reagent optimization.
Among the metal reagents (activators) we screened,
2 equiv of FeCl3·6H2O generated the best yield (63%, see
Supporting Information). Catalytic or stoichiometric or ex-
cessive amount of the iron activator gave decreased yields.
Interestingly, anhydrous FeCl3 is not beneficial for this rear-
rangement and provided product 2a in only 21% yield.
FeBr3 still enabled this reaction and gave the product in
60% yield, compatible to that from FeCl3·6H2O. Using
other iron or non-iron promoters, the rearrangement did not
occur at all. Oxidative reagents (MnO2) or its combination
with FeCl3·6H2O also gave negative results.
also took place but with lower yields (entries 9, 10). Note-
worthy, N-propargyl- (1k), and N-benzyl- (1l) benzazepines
yielded two pairs of skeleton-rearranged products 2k, 4k,
and 2l,[13] 4l,[14] respectively, in roughly 1:1 ratios (en-
tries 11,12). Further, in the case of N-H norbenzazepine
1m, no THIQ 2m was obtained; instead a dehydrogenated
dihydroisoquinoline 3m was isolated as the major product
(entry 13).
From the results described above, it is apparent that most
of the N-alkyl benzazepines (1a–m) participated the skele-
ton rearrangement and yielded the expected THIQs (2a–l)
in moderate to good yields confirming the generality of this
rearrangement. The somewhat lower yields in most cases
may be ascribed to the difficulty in isolating the products
from the reaction complex formed by the excessive amount
of FeCl3·6H2O. Meanwhile, the influence of different N-sub-
stituents on the yields and products is notable. To broaden
the scope and probe the limitations of this skeleton rear-
rangement, a series of N-acylated benzazepines 5a–g were
prepared and their FeCl3·6H2O-assisted skeleton rearrange-
ment was investigated under the same reaction conditions.
The results were summarized in Table 2.
With the optimal solvent (PhNO2) and activator
(FeCl3·6H2O, 2 equiv), we decided to explore the substrate
scope and generality of this rearrangement. Accordingly,
various benzazepines (1a–m) bearing different N-alkyl sub-
stituents were tested (Table 1). N-methyl, ethyl, n-propyl,
To our surprise, treating benzazepines 5a–g with
FeCl3·6H2O in PhNO2 under the same conditions, optimized
for benzazepines 1a–m, did undergo the expected skeleton
rearrangement, but the structures of the products were not
Table 1. FeCl3·6H2O-promoted skeleton-rearrangement of N-alkylated-
1-arylbenzazepines.[a]
1
the same as we expected. A careful examination of the H
and 13C NMR along with other spectroscopic data (see Sup-
porting Information) indicated that 1-formyl-1-aryl-N-acyl-
tetrahydroisoquinolines 6a–g were formed. All rearrange-
Table 2. FeCl3·6H2O-promoted skeleton-rearrangement of N-acylated-1-
phenyl benzazepines.[a]
Entry
R1
R2
Product
Yield [%][b] Ratio 2/4[c]
1 (1a)
2 (1b)
3 (1c)
4 (1d)
5 (1e)
6 (1 f)
7 (1g)
8 (1h)
9 (1i)
10 (1j)
11 (1k)
12 (1l)
13 (1m)
Me
Et
nPr
3’-Me 2a
3’-Me 2b
3’-Me 2c
3’-Me 2d
2’-Me 2e
63
56
53
54
87
61
69
52
40
38
58
–
–
–
–
–
–
–
–
allyl
allyl
allyl
CPM
CBM
nBu
iBu
propargyl
Bn
H
H
2 f
H
2g
Entry
1
Substrate
R
Product
Yield [%][b]
60
H
H
H
2h
2i[13]
2j
–
–
5a
6a
H
H
2k + 4k
56:44
41:59
–
2
3
5b
5c
6b
6c
28[c]
66
2l[13] + 4l[14] 86
3’Me
3m
53
[a] A mixture of benzazepine 1 (1.0 equiv) and FeCl3·6H2O (2.0 equiv) in
PhNO2 (7 mL) was stirred at 1208C for 10 h, then worked up. [b] The
yields of isolated products. [c] Determined by isolated yields.
4
5
6
5d
5e
5 f
6d
6e
6 f
70
71
77
allyl, cyclopropylmethyl (CPM), and cyclobutylmethyl
(CBM) substituted benzazepines (1a–h, entries 1–8) allowed
the reaction to occur smoothly and afforded the correspond-
ing THIQs 2a–h in 53–87% yield, suggesting that larger or
slightly steric N-substituents, as well as a C1-aryl with appro-
priate substitution were tolerated. The rearrangement of N-
butyl- (1i), and N-isobutyl- (1j) substituted benzazepines
7
5g
6g
89
[a] A mixture of benzazepines 5 (1.0 equiv) and FeCl3·6H2O (2.0 equiv)
in PhNO2 (7 mL) was stirred at 1208C for 10 h, then worked up. [b] The
yields of isolated products. [c] Product is unstable upon standing.
11120
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11119 – 11122

Quaternary Tetrahydroisoquinolines
COMMUNICATION
ment reactions went through nicely and afforded the prod-
ucts in good yields (60–89%) except N-propionyl-THIQ 6b
which was very unstable and obtained in only 28% yield.
The small cyclic N-substituents, such as cyclopropanecarbon-
yl (5c) and cyclobutanecarbonyl (5d) survived very well
under the reaction conditions and the corresponding alde-
hydes 6c and 6d were obtained in 66 and 70% isolated
yields, respectively. The rearrangement of carbamate 5g
proceeded in highest yield (entry 7, 89%). The structural
feature of 6a–g is very unique and secured by all the spec-
troscopic data, and by the X-ray diffraction of 6c (Figure 2).
To the best of our knowledge, there is no simple routine to
facilitate such highly substituted dihydroisoquinolin-1-carb-
aldehydes.
12 under the general rearrangement conditions. This result
eliminated the possibility that the N-alkyl-THIQs (2) were
obtained through the corresponding aldehydes (6) during
the rearrangement. Therefore, the arrangement of N-alkyl
and N-acyl benzazepines may undergo different pathways.
Although the family of iron reagents (FeCl3, FeCl2,
FeSO4, FeACHTNUGRTNENUG(ClO4)3, FeCAHTUNGTREN(NUGN acac)3) has been a hot topic in the
field of organic reactions and synthesis recently due to its
low cost and readily availability, the exact reaction mecha-
nism of these transformations was barely discussed.[15] Since
FeCl3·6H2O and PhNO2 generally act as oxidative reagents
and are widely used in Skraup or its modified reactions to
form isoquinolines,[16] we shed light on the pathway of
Skraup reaction, and proposed a mechanism to rationalize
our FeCl3·6H2O-promoted skeleton-rearrangement of 1-
aryl-N-alkyl/acyl-benzazepines.
As outlined in Scheme 2, benzazepines 1 and 5 upon
treating with FeCl3·6H2O and PhNO2 would likely undergo
benzyl-oxidation to yield intermediate I. When R group is
alkyl (substrate 1), the intermediate I would go further oxi-
À
dation (II) followed by C1 C2 bond breakage to form spe-
Figure 2. ORTEP diagram of the single-crystal X-ray structure of 6c.
It is quite intriguing that why the corresponding aldehydes
were not obtained when the N-substituents in substrates 1a–
m are alkyl groups. One possible explanation is that the cor-
responding N-alkyl-1-aryl-tetrahydroisoquinolin-1-formalde-
hydes are not stable and readily undergo oxidation to yield
acids which are then decarboxylated to afford THIQ 2. To
test this hypothesis, we initiated an indirect method to syn-
thesize the N-alkyl-1-aryl-tetrahydroisoquinolin-1-formalde-
hydes 10–12, by reducing the aldehydes 6 with LiAlH4, fol-
lowed by Dess–Martin oxidation (Scheme 1). By using this
approach, N-alkyl aldehydes 10–12 were obtained in moder-
ate yields. These aldehydes were found very stable within
weeks at RT and no significant oxidation was observed. Fur-
ther, no desired product 2 was observed when treating 10–
Scheme 2. Plausible mechanism for the rearrangement of 2,3,4,5-tetrahy-
dro-1H-benzazepines 1 and 5 under FeCl3·6H2O in PhNO2.
cies III which is stabilized by both the methoxy and the N-
alkyl groups. Hydrolysis (IV) followed by recyclization
would provide THIQs 2a–l. However, when the R group is
acyl (substrate 5), the intermediate I would go further oxi-
dation followed by deprotonation to yield species V. Intra-
molecular aziridine formation (VI) followed by water-attack
would yield aldehydes 6a–g. Although it can be argued that
during the rearrangement, a primary alcohol like THIQs 7–
9 may be produced which can undergo similar process to
provide the rearranged products 6a–g, however, after treat-
ing alcohol 9 under the same condition (FeCl3·6H2O/
PhNO2) in 12 h, no reaction occurred.
Scheme 1. Preparation of 1-aryl-1-formyl-N-alkyltetrahydroisoquinolines
10–12. a) LiAlH4, THF, reflux; b) Dess–Martin periodinane, CH2Cl2,
258C.
Chem. Eur. J. 2009, 15, 11119 – 11122
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11121

A. Zhang and J. Zhang
d) Y. Yamanaka, M. J. Walsh, V. E. Davis, Nature 1970, 227, 1143–
1144.
[2] A. Zhang, Y. Zhang, A. R. Branfman, R. J. Baldessarini, J. L. Neu-
meyer, J. Med. Chem. 2007, 50, 171–181.
The existence of the C7-MeO group in the benzazepine
backbone is critical to stabilize both the two key intermedi-
ates III and V in this mechanism. This can be confirmed by
the fact that no skeleton rearrangement or other reactions
occurred when substrates 1 or 5 lacking the two methoxy
substituents were subjected to the same reaction conditions.
To gain more insight to the mechanism, racemic amide 5c
was separated on chiral column (OD), and its (+)-enantio-
mer was subjected to the rearrangement reaction. However,
the aldehyde 6c was obtained as a racemate (see Supporting
Information). This result supports our proposed mechanism
in which an intermediate like III or V may be formed.
In summary, a skeleton-rearrangement of 1-arylbenzaze-
pines was developed with the assistance of 2 equiv of
FeCl3·6H2O. The N-substituents played a dramatic influence
on the structures of the products. In the case of N-alkyl-
benzazepines, 1-aryl-tetrahydroisoquinolines (THIQs) were
obtained, whereas N-acylbenzazepine substrates yielded a
series of novel 1-aryl-1-formyl-tetrahydroisoquinolines
(THIQs). The N-acyl-1-aryl-1-formyl-THIQs can serve as
the intermediates to prepare corresponding N-alkyl-1-aryl-1-
formyl-THIQs. These findings not only added an additional
example of organic transformations aided by iron reagent,
but also paved a novel avenue to access THIQs, especially
those with a fully functionalized quaternary C1 carbon
center. Further, the unique structural features of 1-formyl-1-
aryl-THIQs represent a series of novel compounds worthy
for biological screening, and the 1-formyl group also pro-
vides a valuable functionality warranting for further func-
tional transformations and for the synthesis studies on 1,2-
conjucted natural products as depicted in Figure 1.
[3] R. Gitto, R. Caruso, B. Pagano, L. D. Luca, R. Citraro, E. Russo,
G. D. Sarro, A. Chimirri, J. Med. Chem. 2006, 49, 5618–5622.
[4] a) P. Siengalewicz, U. Rinner, J. Mulzer, Chem. Soc. Rev. 2008, 37,
2676–2690; b) M. Chrzanowska, M. D. Rozwadowska, Chem. Rev.
2004, 104, 3341–3370; c) M. D. Rozwadowska, Heterocycles 1994,
39, 903–931.
[5] a) A. Bischler, Chem. Ber. 1893, 26, 1891–1903; b) G. Fodor, J. Gal,
B. A. Philips, Angew. Chem. 1972, 84, 947–949; Angew. Chem. Int.
Ed. Engl. 1972, 11, 919–920; c) R. Fu, X. Xu, Q. Dang, X. Bai, J.
Org. Chem. 2005, 70, 10810–10816.
[6] a) A. Pictet, T. Spengler, Chem. Ber. 1911, 44, 2030–2036; b) E. D.
Cox, J. M. Cook, Chem. Rev. 1995, 95, 1797–1842; c) M. J. Vande-
n Eynden, J. P. Stambuli, Org. Lett. 2008, 10, 5289–5291; d) J.
Seayad, A. M. Seayad, B. List, J. Am. Chem. Soc. 2006, 128, 1086–
1087.
[7] a) M. Boudou, D. Enders, J. Org. Chem. 2005, 70, 9486–9494;
b) E. V. Brown, J. Org. Chem. 1977, 42, 3208–3209.
[8] a) A. Zhang, J. L. Neumeyer, R. J. Baldessarini, Chem. Rev. 2007,
107, 274–302; b) J. Zhang, B. Xiong, X. Zheng, A. Zhang, Med. Res.
Rev. 2009, 29, 272–294.
[9] J. Weinstock, J. P. Hieble, J. W. Wilson, Drugs Future 1985, 10, 645–
697.
[10] J. Zhang, X. Chen, L. Yu, X. Zhen, A. Zhang, Bioorg. Med. Chem.
2008, 16, 9425–9431.
[11] a) H. Mçhrle, C. Rohn, G. Westle, Pharmazie 2006, 61, 391–399;
b) E. Napolitano, R. Fiaschi, V. Scartoni, A. Marsili, J. Chem. Soc.
Perkin Trans. 1 1986, 781–783; c) P. L. Barili, R. Fiaschi; E. Napoli-
tano, L. Pistelli, V. Scarteni, A. Marsili, J. Chem. Soc. Perkin Trans.
1 1981, 1654–1657; E. Napolitano, L. Pistelli, V. Scarteni, A. Marsili,
J. Chem. Soc. Perkin Trans. 1 1981, 1654–1657; d) G. Blaskꢁ, N.
Murugesan, A. J. Freyer, D. J. Gula, B. Sener, M. Shamma, Tetrahe-
dron Lett. 1981, 22, 3139–3142.
[12] P. Ozic, H. S. Gunes, M. Cizmecioglu, V. Pabuccuoglu, B. Gozler, B.
Ozcelik, U. Abbasoglu, Marmara Univ. Eczacilik Derg. 1997, 13,
13–26.
[13] a) R. K. Tiwari, D. Singh, J. Singh, A. K. Chhillar, R. Chandra, A. K.
Verma, Eur. J. Med. Chem. 2006, 41, 40–49; b) J. Wu, F. Wang, Y.
Ma, X. Cui, L. Cun, J. Zhu, J. Deng, B. Yu, Chem. Commun. 2006,
1766–1768.
[14] a) K. Mohri, K. Suzuki, M. Usui, K. Isobe, Y. Tsuda, Chem. Pharm.
Bull. 1995, 43, 159–161; b) A. Kubo, N. Saito, N. Kawakami, Y. Mat-
suyama, T. Miwa, Synthesis 1987, 824–827.
[15] For recent reviews on iron-catalyzed reactions, see: a) S. Enthaler,
K. Junge, M. Beller, Angew. Chem. 2008, 120, 3363–3367; Angew.
Chem. Int. Ed. 2008, 47, 3317–3321; b) B. D. Sherry, A. Fꢂrstner,
Acc. Chem. Res. 2008, 41, 1500–1511; c) A. Correa, O. G. Manche-
Ço, C. Bolm, Chem. Soc. Rev. 2008, 37, 1108–1117; d) C. Bolm, J.
Legros, J. L. Paih, L. Zani, Chem. Rev. 2004, 104, 6217–6254.
[16] a) Z. H. Skraup, Monatsh. Chem. 1881, 2, 139–170; b) R. H. Reitse-
ma, Chem. Rev. 1948, 43, 43–68; c) R. H. Manske, Chem. Rev. 1942,
30, 145–158; d) S. E. Denmark, S. Venkatraman, J. Org. Chem.
2006, 71, 1668–1676.
Acknowledgements
The authors are grateful to the Chinese National Science Foundation
(grant 30772625), Shanghai Commission of Science
(07pj14104), the National Department of Science
(2009ZX09103-062) as well as SIMM (SIMM0809KF-02) for financial
support. We also thank Professor Liming Zhang from University of
Nevada, and Professor Guigen Li from Texas Tech University for valua-
ble discussion on the mechanism.
&
&
Technology
Technology
Keywords: benzazepines · iron · quaternary carbon centers ·
rearrangement · tetrahydroisoquinolines
[1] a) K. W. Bentley, Nat. Prod. Rep. 2006, 23, 444–463; b) J. D. Scott,
R. M. Williams, Chem. Rev. 2002, 102, 1669–1730; c) K. W. Bentley,
The Isoquinoline Alkaloids, Harwood Academic, Amsterdam, 1998;
Received: May 4, 2009
Revised: June 25, 2009
Published online: September 18, 2009
11122
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11119 – 11122