Pyrrole versus quinoline formation in the palladium catalyzed reaction of 2-alkynyl-3-bromothiophenes and 2-alkynyl-3-bromofurans with anilines. A combined experimental and computational study

Article abstract of DOI:10.1039/c2ob26689j

Benzofuroquinolines were prepared by a new type of Pd catalyzed annulation reaction. In the first step, 2-alkynyl-3-bromobenzofurans were prepared by Sonogashira reactions of 2,3-dibromobenzofuran. Their Pd catalyzed reaction with electron-rich anilines afforded benzofuroquinolines by a domino C-N coupling/annulation process. This reaction proceeds as a C,N-cyclization via the nitrogen atom and the ortho-carbon of the aniline. Similarly, furoquinolines were prepared from 2,3-dibromofuran. In contrast, benzofuropyrroles and furopyrroles were formed by N,N-cyclization when electron-poor anilines were used. Earlier, we reported results related to the thiophene and benzothiophene series. Quinolines were formed from 2,3-dibromobenzothiophene when electron rich anilines were used. In contrast, pyrroles were obtained in the case of electron-poor anilines. On the other hand, pyrroles were generally obtained, not depending on the type of aniline, when 2,3-dibromothiophene was employed as the starting material. In the present article, a detailed DFT study related to the mechanism (quinoline versus pyrrole formation) is reported which provides a rationalization of the selectivities observed for the furan, benzofuran, thiophene and benzothiophene series and for the different selectivities observed for electron-rich and -poor anilines.

Full text of DOI:10.1039/c2ob26689j

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
PAPER
Pyrrole versus quinoline formation in the palladium
catalyzed reaction of 2-alkynyl-3-bromothiophenes and
2-alkynyl-3-bromofurans with anilines. A combined
experimental and computational study†
Cite this: Org. Biomol. Chem., 2012, 10,
9464
Ghazwan Ali Salman,a Riffat Un Nisa,b Viktor O. Iaroshenko,a,c Jamshed Iqbal,d
Khurshid Ayubb and Peter Langer*a,e
Benzofuroquinolines were prepared by a new type of Pd catalyzed annulation reaction. In the first step,
2-alkynyl-3-bromobenzofurans were prepared by Sonogashira reactions of 2,3-dibromobenzofuran. Their
Pd catalyzed reaction with electron-rich anilines afforded benzofuroquinolines by a domino C–N coup-
ling/annulation process. This reaction proceeds as a C,N-cyclization via the nitrogen atom and the ortho-
carbon of the aniline. Similarly, furoquinolines were prepared from 2,3-dibromofuran. In contrast, benzo-
furopyrroles and furopyrroles were formed by N,N-cyclization when electron-poor anilines were used.
Earlier, we reported results related to the thiophene and benzothiophene series. Quinolines were formed
from 2,3-dibromobenzothiophene when electron rich anilines were used. In contrast, pyrroles were
obtained in the case of electron-poor anilines. On the other hand, pyrroles were generally obtained, not
depending on the type of aniline, when 2,3-dibromothiophene was employed as the starting material. In
the present article, a detailed DFT study related to the mechanism (quinoline versus pyrrole formation) is
reported which provides a rationalization of the selectivities observed for the furan, benzofuran, thio-
phene and benzothiophene series and for the different selectivities observed for electron-rich and -poor
anilines.
Received 27th August 2012,
Accepted 24th October 2012
DOI: 10.1039/c2ob26689j
www.rsc.org/obc
this class of molecule, due to their antimuscarinic, antibacter-
ial, antiviral, antiplasmodial, and antihyperglycemic
Introduction
Pyridoindoles (carbolines)1 are of considerable pharmacologi- activities.5–7 Benzofuro[3,2-b]quinolones, which can be
cal relevance. Their benzo-annulated derivatives are, in con- regarded as oxygen analogues of these alkaloids, are equally
trast to simple derivatives, relatively rare in nature. Examples important because of their broad spectrum of biological
include the indoloquinoline alkaloids quindoline (A)2 and activity against bacteria and protozoa, as well as anticancer
cryptolepine (B)3,4 which were isolated in 1977 and 1929, and cytotoxic activity. For example, compound C has been
respectively, from the West African plant Cryptolepis sanguino- reported as a potent antituberculosis agent, while compound
lenta (Periplocaceae) (Fig. 1). There is considerable interest in D exhibits cytotoxic activity.8,9
Many natural products containing a furoquinoline moiety
are known to exhibit a wide range of biological activities, such
as antiasthmatic,10 anti-inflammatory11 and antiproliferative
activities.12 Among the methods developed to date for the syn-
thesis of furoquinoline derivatives is the Lewis acid catalysed
imino Diels–Alder reaction between N-benzylideneanilines and
nucleophilic olefins.13 Due to their biological relevance, a
number of syntheses of furoquinolines and benzofuroquino-
lines have been reported. Avetisyan et al.14 reported a multi-
step synthesis of furo[3,2-c]quinolones in 20% overall yield
based on the allylation of 4-hydroxy-2-methylquinoline with
2,3-dichloropropene and subsequent Claisen rearrangement.
aInstitut für Chemie, Universität Rostock, Albert-Einstein-Str. 3a, 18059 Rostock,
Germany. E-mail: peter.langer@uni-rostock.de; Fax: +49 (0)381 4986412
bDepartment of Chemistry, COMSATS Institute of Information Technology,
Abbottabad, 22060, Pakistan
cNational Taras Shevchenko University, 62 Volodymyrska st., Kyiv-33, 01033,
Ukraine
dDepartment of Pharmacy, COMSATS Institute of Information Technology,
Abbottabad, 22060, Pakistan
eLeibniz-Institut für Katalyse e. V. an der Universität Rostock, Albert-Einstein-Str.
29a, 18059 Rostock, Germany
†Electronic supplementary information (ESI)
available. See DOI:
10.1039/c2ob26689j
9464 | Org. Biomol. Chem., 2012, 10, 9464–9473
This journal is © The Royal Society of Chemistry 2012

Organic & Biomolecular Chemistry
Paper
View Article Online
alkynes has been employed for the synthesis of indoles, benzo-
furans, benzopyrans, isocoumarins, isoquinolines, and poly-
cyclic aromatic hydrocarbons.26–29
Recently, we reported the synthesis of benzothienoquino-
lines by a new type of Pd catalyzed annulation reaction. In the
first step, 2-alkynyl-3-bromobenzothiophenes were prepared by
the Sonogashira reaction of 2,3-dibromobenzothiophene.30
The Pd catalyzed reaction of the 2-alkynyl-3-bromobenzothio-
phenes with anilines afforded benzothienoquinolines by a
domino C–N coupling/annulation process. This reaction pro-
ceeds as a C,N-cyclization via the nitrogen atom and the ortho-
carbon of the aniline. In contrast, thienopyrroles were formed
by a domino C–N coupling/hydroamination reaction (N,N-cycli-
zation) when the reactions were carried out starting with thio-
phenes instead of benzothiophenes. The latter type of reaction
is more common and has been earlier studied for other ring
systems. Herein, we report, for the first time, the application of
our methodology to the synthesis of furoquinolines and benzo-
furoquinolines starting with 2,3-dibromofuran and 2,3-dibro-
mobenzofuran, respectively. While pyrroles are formed for the
thiophene series and quinolines are formed in the case of the
benzothiophene series, quinolines are generally formed for
both the furan and benzofuran series. The type of aniline (elec-
tron rich or poor) also has an influence on the mode of cycliza-
tion. In addition, we report a detailed DFT study to shed light
on the mechanism.
Fig. 1 Structures of alkaloids quindoline (A) and cryptolepine (B) and of the
bioactive benzofuroquinolines C and D.
Avetisyan et al.15 later developed a more convenient approach
which involves bromination of 3-(2-chloro-2-propenyl)-4-
hydroxy-2-methyl-quinolone to give 2-bromomethyl-2-chloro-4-
methyl-2,3-dihydrofuro[3,2-c]quinoline and subsequent reac-
tion with various nucleophiles to afford the corresponding
2-substituted 4-methylfuro[3,2-c]quinoline in 83–98% yield.
Kawase et al.16,17 reported the synthesis of benzofuroquino-
lines by the reaction of 4-hydroxy-3-(2-methoxyphenyl)-1,2-
dihydroquinolin-2-one, prepared from methyl anthranilate and
2-methoxyphenylacetyl chloride, with pyridine hydrochloride.
Yang et al. reported the synthesis of an 11-chlorobenzofuro-
[2,3-b]quinolone by reflux of anthranilic acid with 2-cumara-
none in POCl3. Other methods represent modifications of the
Pfitzinger synthesis which relies on the base mediated cyclo-
condensation of isatin with ketones.18 Zhu et al. reported a
multistep synthesis of benzofuroquinolines from substituted
anthranilic acids, acyl chlorides and phenols.19 All these synth-
eses of benzofuroquinolines proceed under harsh conditions
or require several synthetic steps. Therefore, the development
of alternative synthetic strategies is of considerable interest.
In recent years, transition metal-catalyzed syntheses of
heterocycles have been developed which nicely complement
classic synthetic approaches because they proceed under mild
conditions and show a high degree of chemoselectivity and
functional group tolerance.20,21 A variety of palladium-cata-
lyzed syntheses of indoles and carbazoles have been reported.
For example, Ackermann et al. reported the synthesis of
indoles and various other ring systems by domino N-arylation/
hydroamination reactions.22 They also developed an efficient
approach to indoles by Pd- or Cu-catalyzed domino C–N coup-
ling/hydroamination reactions of ortho-alkynylated aryl
halides.23 Buchwald et al. reported the synthesis of pyrroles
and pyrazoles by a related strategy.24,25 The palladium-cata-
lyzed annulation of alkynes has recently been proven to be a
powerful method for the construction of a variety of carbo-
and heterocycles. For example, the annulation of internal
Results and discussion
Synthetic studies
The bromination of benzofuran (1) afforded the known 2,3-
dibromobenzofuran (2) in high yield. The Sonogashira reac-
tion of 2 with alkynes 3a–c (1.0 equiv.) afforded the 2-alkynyl-
3-bromobenzofurans 4a–c with very good site-selectivity
(Scheme 1, Table 1). The first attack occurred at the more elec-
tron-poor position.
Scheme 1 Synthesis of 4a–c. Conditions: (i) Br2 (2.0 equiv.), KOAc (2.0 equiv.),
CH2Cl2, reflux, 5 h; (ii) 2a–c (1.0 equiv.), Pd(PPh3)2Cl2 (5 mol%), CuI (10 mol%),
diethylamine (1.5 equiv.), THF, 20 °C, 24 h.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 9464–9473 | 9465

Paper
Organic & Biomolecular Chemistry
View Article Online
The Pd catalyzed reaction of 2-alkynyl-3-bromobenzofurans quinolines 6a–t in good yields (Scheme 2, Table 2). The best
4a–c with various anilines 5a–k afforded the benzofuro[3,2-b]- results were obtained when, similar to the previously reported
synthesis of benzothieno[3,2-b]quinolines,30 Pd(OAc)2 (10 mol%),
P(tBu)3·HBF4 (20 mol%) and KOtBu (2 equiv.) were employed.
Table 1 Synthesis of 2-alkynyl-3-bromobenzofurans 4a–c
The yields of the products were higher for anilines containing
3,4
R
% (4)a
electron-donating substituents (e.g., products 6b,c,e,l,m,p) as
compared to those containing electron-withdrawing substitu-
ents (e.g., products 6i,j). This result suggests that some side
reactions may occur for such substrates. To study this obser-
vation in more detail, we carried out the reaction of 2-alkynyl-
3-bromobenzofuran 4c with 4-chloroaniline (5l), 3-nitroaniline
(5m) and 4-cyanoaniline (5n). In the case of 5l, quinoline 6s
was isolated in only 40% yield. As a by-product, pyrrole 7a was
isolated in 30% yield. In the case of 5m and 5n, only quino-
lines 7b and 7c were isolated, respectively. This result indicates
that pyrroles are formed (by N,N-cyclization) when electron-
poor anilines, containing π-acceptor substituents, are
employed. In contrast, quinolines are obtained for all the
other anilines (C,N-cyclization). A similar result was earlier
observed for the corresponding sulfur analogues.30 Surpris-
ingly, only quinolines 6i and 6j were obtained when (electron
poor) 4-fluoro- and 4-trifluoromethylaniline were employed,
respectively. In fact, tlc and GC-MS analysis of the crude
product mixture indicated the formation of small amounts of
pyrroles which, however, could not be isolated in pure form,
due to practical problems during the chromatographic
purification.
a
b
c
H
90
82
75
Me
OCH3
a Yields of isolated products.
To study the mechanism of the cyclization reaction, we
tried to isolate an intermediate formed after initial C–N coup-
ling. The reaction of 4c with 5l, carried out at 40 °C (4 h)
instead of 110 °C (12 h), afforded product A1 (Scheme 3).
Scheme 2 Synthesis of 6a–u and 7a–c. Conditions: (i) 5a–n (1.3 equiv.), Pd-
(OAc)2 (10 mol%), P(tBu)3·HBF4 (20 mol%), KOtBu (2 equiv.), 110 °C, 12 h.
Table 2 Synthesis of 6a–u and 7a–c
4
5
6
7
R1
R2
R3
R4
R5
% (6)a
% (7)a
a
a
a
a
a
a
a
a
a
c
b
b
b
b
b
c
c
c
c
c
c
c
c
a
b
c
d
e
f
a
b
c
d
e
f
g
h
i
H
H
CH3
H
H
H
H
OCH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
60
65
68
60
70
61
67
65
60
55
64
70
68
63
61
67
72
68
66
70
40
0
0
H
H
OCH3
OC2H5
C2H5
OCH3
CH(CH3)2
H
H
0
H
H
H
H
0
H
H
0
H
OCH3
H
H
0
H
H
0
g
h
i
OCH3
CH3
H
H
H
0
H
CH3
H
0
H
F
H
0
j
j
k
l
H
H
CF3
OCH3
CH3
CH3
CH3
CH3
CH3
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
0
k
b
c
f
H
OCH3
H
H
0
H
OCH3
OC2H5
CH(CH3)2
CH3
0
m
n
o
p
q
r
H
H
0
H
H
0
h
b
d
g
a
f
l
m
n
CH3
H
H
0
H
OCH3
C2H5
H
0
H
H
0
OCH3
H
H
0
s
H
CH3
0
t
H
H
CH(CH3)2
Cl
0
u
a
b
c
H
H
30
75
88
H
NO2
H
H
CN
H
0
a Yields of isolated products.
9466 | Org. Biomol. Chem., 2012, 10, 9464–9473
This journal is © The Royal Society of Chemistry 2012

Organic & Biomolecular Chemistry
Paper
View Article Online
Scheme 3 Isolation of intermediate A1. Conditions: (i) 5l (1.3 equiv.), Pd(OAc)2
(10 mol%), P(t-Bu)3·HBF4 (20 mol%), KOtBu (2 equiv.), toluene, 40 °C, 4 h.
Scheme 5 Synthesis of 10a–f and 11. Conditions: (i) 5b,c,h,o,p (1.3 equiv.), Pd-
(OAc)2 (10 mol%), P(t-Bu)3·HBF4 (20 mol%), KOtBu (2.0 equiv.), toluene, 110 °C,
12 h.
Table 4 Synthesis of 10a–f and 11
5
10
11
R1
R2
R3
% (10)a
% (11)a
b
b
c
h
h
o
p
a
b
c
d
e
f
a
b
c
d
e
f
H
OCH3
OCH3
OC2H5
CH3
CH3
OCH3
OCH3
H
OCH3
OCH3
H
70
65
64
60
55
72
0
0
0
H
H
0
Scheme 4 Synthesis of 9a–c. Conditions: (i) 3a–c (1.0 equiv.), Pd(PPh3)2Cl2
CH3
CH3
OCH3
H
0
(5 mol%), CuI (10 mol%), triethylamine (1.2 equiv.), THF, 20 °C, 48 h.
CH3
0
OCH3
NO2
0
g
g
65
Table 3 Regioselective alkynylation of 2,3-dibromofuran
a Yields of isolated products.
3,9
R
% (9)a
a
b
c
H
CH3
OCH3
82
75
85
symmetry constraints by the hybrid B3LYP method. The B3LYP
method, which consists of the three-parameter hybrid func-
tional of Becke32 in conjunction with the correlation func-
tional of Lee, Yang, and Parr,33 provides a nice balance
between cost and accuracy. For geometries optimization, the
6-31G*34 basis set was used for C, H, N, O and S atoms and the
LANL2DZ pseudopotential35 for Pd. Each optimized structure
was confirmed by frequency analysis at the same level as a true
minimum (no imaginary frequency) or as a transition state
(with one imaginary frequency). Imaginary frequencies of tran-
sition states were also evaluated to confirm that their associ-
ated eigenvector corresponds to the motion along the reaction
coordinates. The reported energies for all structures are in
a Yields of isolated products.
Simple stirring of a toluene solution of A1 at 110 °C for 12 h,
in the absence of a catalyst, failed to give any cyclization
product. In contrast, a cyclization could be successfully
induced in high yield when the reaction was carried out in the
presence of Pd(OAc)2 (10 mol%), P(tBu)3·HBF4 (20 mol%) and
KOtBu (2 equiv.). This result shows that, similar to the reac-
tions of 2-alkynyl-3-bromobenzothiophenes,30 the cyclization
step of the one-pot process leading to products 6 is a palla-
dium-catalyzed process and not a thermal electrocyclization.
The palladium catalyzed reaction of 2,3-dibromofuran (8)
with alkynes 3a–c afforded the 2-alkynyl-3-bromofurans 9a–c
in very good yields (Scheme 4, Table 3).
The palladium catalyzed reaction of 2-alkynyl-3-bromofur-
ans 9a–c with electron-rich anilines 5b,c,h,o afforded furo-
[3,2-b]quinolines 10a–f. In contrast, furo[3,2-b]pyrrole 11 was
isolated when aniline 5p, containing an electron-withdrawing
nitro group, was employed (Scheme 5, Table 4). These results
are in line with the discussion above for the different chemical
behaviour of electron-rich and -poor anilines.
kcal mol−1
.
Density functional theory (DFT) calculations have been per-
formed to gain mechanistic insight into the reaction to explain
the regioselectivity difference observed for pyrrole vs. quino-
line formation. We have recently reported30 that a thiophene
intermediate (AT) is cyclized to a pyrrole, whereas a benzothio-
phene (ABT) favours the formation of a quinoline product. In
this study, we describe that both furans (AF) and benzofurans
(ABF) generally favour the formation of quinolines. For all sub-
strates, the use of electron poor anilines results in the for-
mation of pyrroles. Intermediate type A is generated by a
Buchwald–Hartwig reaction and the product subsequently
undergoes a cyclization to deliver either a pyrrole or a quino-
line skeleton.
Computational studies
THERMAL NON-CATALYZED CYCLIZATION. First of all, transition
states for a non-catalyzed thermal cyclization reaction of
METHOD. All calculations were performed using Gaussian
09.31 Geometries of the structures were optimized without any
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 9464–9473 | 9467

Paper
Organic & Biomolecular Chemistry
View Article Online
Scheme 6 Number scheme for discussion.
Scheme 8 Energy profile for cyclization of anionic ET into pyrrole FT and quino-
line precursor GT. All energies are in kcal mol−1 relative to ET and include
unscaled zero point correction.
kinetically favourable over quinoline formation by 6-exo cycli-
zation (through DT) via the ortho carbon (C3) (26.7 kcal mol−1
vs. 33.1 kcal mol−1). A similar selectivity in favour of pyrrole
over quinoline formation (by 4.3 kcal mol−1) was also observed
Scheme 7 Energy profile for cyclization of anionic BT to give pyrrole CT and
quinoline precursor DT. All energies are in kcal mol−1 relative to BT and include
unscaled zero point correction. All bond distances are in Angstroms.
for the benzothiophene intermediate BBT (25.2 vs. 29.5 kcal mol−1
,
see ESI†). Although the energy available at 110 °C may
be sufficient to surpass the barrier, the experimental regio-
selectivities cannot be explained by this model. Pyrrole for-
mation is highly favourable over its competitive quinoline
formation, regardless the skeleton. This is probably due to the
fact that the electron density is more located at the aniline
nitrogen than at the ortho carbon C3. Due to the inability of
the model to explain the experimental regioselectivity, no
further exploration of furo and benzofuro systems has been
carried out.
intermediate ABT to give 12 were explored, but no transition
state for quinoline formation could be located. This is consist-
ent with the experimental observation that intermediate ABT
failed to cyclize by simple heating, supporting the fact that it
is not purely a thermal cyclization. A plausible explanation for
this failure is the lack of alkyne activation combined with low
nucleophilicity of carbon C3 of aniline (see Scheme 6 for
numbering).
Aniline intermediate AT has an acidic proton which can be
abstracted by a base (KOtBu in this case) to generate an
anionic intermediate BT with enhanced nucleophilicity of both
aniline nitrogen N1 and of ortho carbon C3 for pyrrole and qui-
nolone formation, respectively (Scheme 7). However, nucleo-
philic attack of aniline nitrogen N1 to alkyne carbon C4 via
5-endo cyclization to form pyrrole through CT was found to be
PD(II)-CATALYZED MECHANISM. Cyclization reactions in the
absence of Pd failed to explain the observed regioselectivity,
therefore, activation of the alkyne by the palladium species
was taken into account (Scheme 8). For this purpose, Pd(0)
and Pd(II) may be considered, however, the latter is a better
electrophilic activator of alkynes for nucleophilic attack. With
Pd(II), both electroneutral aniline (with an electroneutral Pd(II)
9468 | Org. Biomol. Chem., 2012, 10, 9464–9473
This journal is © The Royal Society of Chemistry 2012

Organic & Biomolecular Chemistry
Paper
View Article Online
species) and deprotonated aniline (with an anionic Pd(II)
species) were computationally considered for the cyclisation.
Firstly, the reaction of the anionic aniline derivative with the
alkyne coordinated to Pd(II) was studied with regard to its cycli-
zation selectivity. For this purpose, hydroxy ligands were
chosen and this choice was based on the supposition that the
electron density on Pd with two OH groups will be comparable
to two relatively electron deficient OAc groups and an electron
rich phosphine. This approach turned out to be successful
(vide infra) and allowed us to carry out the calculations with
reasonable calculation time.
Coordination of the alkyne to palladium(II) in the anionic
intermediate BT to generate ET is a thermodynamically favour-
able process by 29.1 kcal mol−1. The high coordination energy
of the complex stems from participation of the negative charge
at nitrogen. The negative charge in ET resides mostly at N1 and
C6. The latter along with alkyne carbons behaves like an allylic
anion (4e donor), therefore, the coordination geometry of pal-
ladium is more close to a square plane. The C6–Pd and C4–Pd
bond distances are 2.32 Å, whereas the C5–Pd bond distance is
2.18 Å. Moreover, the O–Pd–O bond angle is 85 degrees.
Nucleophilic attack of N1 at C4, to give the pyrrole-like inter-
mediate FT, is a highly favourable process both kinetically
(11.22 kcal mol−1) and thermodynamically (over 30 kcal
mol−1). In the transition state (TS-Py-ET), geometry on palla-
dium changes from square plane to T-shaped with two oxygen
ligands located trans to each other (O–Pd–O bond angle is
167.5). The palladium is more bonded to C5 (Pd–C5 bond dis-
tance is 2.03 Å). Moreover, the N1–C5 bond distance is 2.68 Å
Scheme 9 Energy profile for cyclization of anionic EBT into pyrrole FBT and qui-
noline precursor GBT. All energies are in kcal mol−1 relative to EBT and include
unscaled zero point correction.
Finally, we turned our attention to the effect of the elec-
which is reduced to 1.33 Å in the case of FT. The competitive tronic character of the aniline on the mode of cyclization. We
transition state (TS-Qu-ET) for the formation of quinoline GT is computationally studied as model reactions the use of an elec-
1.9 kcal mol−1 higher in energy (Scheme 11).
tron-rich and of an electron-poor aniline. Our computational
We next focused on the cyclization of EBT, because it should results strongly agree with the experimental observation that
favour the formation of the 6-exo cyclization product GBT and, electron withdrawing groups favour formation of a pyrrole,
indeed, this is predicted computationally. The transition state while electron donating substituents favour formation of a qui-
for cyclization to give the quinoline lies 1.85 kcal mol−1 lower noline. For the nitro derivative, the kinetic barrier for pyrrole
in energy than the corresponding TS for pyrrole (Scheme 9). In formation is 6.03 kcal mol−1, compared to 12.54 kcal mol−1 for
TS-Qu-EBT, palladium is also T-shaped with two oxygen ligands quinoline formation. For the methoxy derivative, the kinetic
located trans to each other (O–Pd–O angle is 167°). The C3–C4 demand for quinoline formation is 13.64 kcal mol−1 as com-
bond distance is 1.98 Å. In order to rationalize the difference pared to 12.86 kcal mol−1 for pyrrole formation.
in the selectivity of cyclization, we analyzed Mullikan charges
Our findings strongly suggest that the cyclization proceeds
in BT and BBT. A slight difference was observed in Mullikan through anionic intermediates with Pd(II) catalysis. However,
charges on C-3 (−0.184 for BBT vs. −0.17 for BT), however, it is to consider also the participation of other possible intermedi-
unclear whether this difference in Mullikan charges is ates in the reaction mechanism, the reaction was also studied
sufficient to cause the change in cyclization mode.
for Pd(II) complexes of neutral aniline derivatives as shown in
We studied computationally the cyclization of Pd(II)-com- Table 6. For the cyclization of thiophene-based intermediates,
plexes EF and EBF derived from furan and benzofuran, respect- HT is kinetically favoured in the direction of 5-endo (TS-Py-HT)
ively. We found that the formation of quinoline products is over 6-exo (TS-Qu-HT) by 19 kcal mol−1. Similar selectivities for
indeed favourable over the formation of pyrrole products for pyrrole formation were observed in the case of benzothiophene
both the furan and benzofuran series (Table 5). Transition and furan intermediates HBT and HF, respectively. Kinetic bar-
state TS-Qu-EF is 0.3 kcal mol−1 lower in energy than TS-Py-EF, riers are higher, but may still be accessible under the reaction
whereas this difference rises to 4.45 kcal mol−1 for the benzo- conditions, however, the experimentally observed regioselectiv-
furan system (TS-Qu-EBF is 4.45 kcal mol−1 lower in energy ities cannot be explained by this model. In all cases, pyrrole
than TS-Py-EBF). In both systems, furan and thiophene, fusion formation is favourable compared to quinoline formation.
of a benzene ring kinetically favours the formation of the
quinoline by about 3.4 kcal mol−1
PALLADIUM(0) CATALYZED MECHANISM. In a final study, it was ana-
lyzed computationally whether Pd(0), instead of Pd(II), can
.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 9464–9473 | 9469

Paper
Organic & Biomolecular Chemistry
View Article Online
Table 5 Comparison of computationally calculated and experimentally observed regioselectivities for Pd(II) catalysis and anionic anilines
Entry
Transition state
X
Y
R1
R2
TS-Pyr-
TS-Qu-
Product (experiment)
1
2
3
4
5
6
ET
S
H
Benzo
H
Benzo
H
H
H
H
11.22
11.09
14.17
17.16
6.03
13.1
Pyrrole
EBT
S
H
H
9.24
Quinoline
Quinoline
Quinoline
Pyrrole
EF
EBF
EF-NO2
EF-OCH3
O
O
O
O
H
H
NO2
OCH3
H
13.87
11.50
12.54
12.86
H
H
CH3
13.64
Quinoline
Table 6 Comparison of computationally calculated and experimentally observed regioselectivities for Pd(II) catalysis and electroneutral anilines
Entry
Name
X
Y
R1
R2
TS-Pyr-
TS-Qu-
Exp.
1
2
3
HT
HBT
HF
S
H
Benzo
H
H
H
H
H
H
H
14.5
11.64
18.5
33.5
24.95
38.34
Pyrrole
Quinoline
Quinoline
S
O
catalyze the cyclization of intermediates B to explain the exper- generate an intermediate OT. This process is found to be
imentally observed regioselectivities. Complexation of BT with favourable both kinetically (ΔG# = 5.38 kcal mol−1) and ther-
Pd(0) generated KT, which shows a different coordination of modynamically (24.9 kcal mol−1). The protonated ligand trans-
the alkyne to Pd as compared to complex ET. In this case, it is fers its additional proton to the neighbouring carbon via
an eta-2 complexation and Pd–C4 and Pd–C5 bonds are of transition state TS-OT which lies just 0.5 kcal mol−1 above
equal length (2.02 Å), however, the L–Pd–L bond angle is com- from its precursor OT. The final product PT with perfect quino-
paratively large (105 deg). Cyclization of complex KT favours line geometry is thermodynamically more stable by 20 kcal
the formation of pyrrole LT over quinoline MT by 6.7 kcal mol−1
.
mol−1. The thermodynamic downhill process from GT to PT
The energy of activation is considerably higher for the (45 kcal mol−1) easily outweighs the 30 kcal mol−1 stability of
Pd(0) catalyzed reaction. Moreover, this model does not pyrrole FT (Schemes 8 and 10). A main reason for the higher
predict the right regioselectivity and is, therefore, ruled out thermodynamic stability of the quinoline product over pyrrole
(see ESI†).
is the higher number of fused rings and their aromaticities.
PROTON SHIFT AND HYDROLYSIS. Our computational studies of the For example, FT has a thiophene ring fused with a pyrrole
transition states, which include deprotonated anilines coordi- ring, while in PT a thiophene ring is fused to a quinoline
nated to Pd(II), are in agreement with the observed regioselec- moiety. The aromaticity of six membered nitrogen-
tivities. These computations refer, of course, to an irreversible, containing heterocycles, such as pyridine, is almost compar-
kinetically controlled reaction. However, in all cases, the able to that of benzene,36 while it is considerably higher than
pyrrole products are thermodynamically more stable than the that of pyrrole.
quinoline precursors (e.g., GT vs. FT). Moreover, the kinetic
Based on our computational studies, the most probable
differences are not extraordinarily large which means that the mechanism is shown in Scheme 11. The mechanism is
quinoline precursor should ultimately transform into a pyrrole depicted for an example of the thiophene series, but it is iden-
product via retro-cyclization, unless there is a thermodynamic tical to the benzothiophene, furan, and benzofuran series.
sink for the quinoline precursor. To investigate the post-cycli-
zation pathway, we have performed calculations on the quino-
line precursor GT which can deliver the quinoline product
either by an intermolecular deprotonation–protonation
Conclusions
sequence or by intramolecular proton shift. In our efforts to
study proton abstraction by a base, we have found that tert-
butanol (generated in the first step of proton abstraction) can
easily deliver a proton H-3 to the oxygen based ligand on Pd to
We have reported a new synthesis of benzofuroquinolines by a
new type of Pd catalyzed annulation reaction. In the first step,
2-alkynyl-3-bromobenzofurans were prepared by the Sonoga-
shira reaction of 2,3-dibromobenzofuran. The Pd catalyzed
9470 | Org. Biomol. Chem., 2012, 10, 9464–9473
This journal is © The Royal Society of Chemistry 2012

Organic & Biomolecular Chemistry
Paper
View Article Online
Scheme 10 Transformation of GT to PT. All free energies are in kcal mol−1 rela-
tive to GT. Unnecessary hydrogen atoms are removed for clarity.
reaction of the 2-alkynyl-3-bromobenzofurans with electron-
rich anilines afforded benzofuroquinolines by a domino C–N
coupling/annulation process. This reaction proceeds as a C,N-
cyclization via the nitrogen atom and the ortho-carbon of the
aniline. Similarly, furoquinolines were prepared from 2,3-
dibromofuran. In contrast, benzofuropyrroles and furopyrroles
were formed by an N,N-cyclization when electron-poor anilines
were used. In our previous studies,30 we reported that quino-
lines were formed from 2-alkynyl-3-bromobenzothiophenes
when electron-rich anilines were used. In contrast, pyrroles
were formed in the case of electron-poor anilines. On the other
hand, pyrroles were generally obtained, no matter which type
of aniline was employed, when 2-alkynyl-3-bromothiophenes
were used as the starting material. In the case of electron-rich
anilines, quinolines are formed from 2-alkynyl-3-bromofurans,
-benzofurans and benzothiophenes, while pyrroles are formed
in the case of 2-alkynyl-3-bromothiophenes. For all hetero-
cycles, the formation of pyrroles is observed when electron-
poor anilines are employed.
We have reported a detailed DFT study related to the mech-
anism (quinoline versus pyrrole formation) which provides a
rationalization of the selectivities observed. In the case of elec-
tron-rich anilines, the nucleophilicity of the phenyl group is
higher and quinoline formation is favoured. On the other
hand, the product distribution seems to also depend on the
Scheme 11 Possible mechanism for the formation of pyrroles and quinolines.
Experimental section
General procedure A for the synthesis of 6, 7, 10, and 11
degree of aromaticity of the heterocyclic core structure. Quino- In a pressure tube (glass bomb) a suspension of Pd(OAc)2
lines are formed for less aromatic heterocycles, while pyrroles
are formed for more aromatic systems. The aromaticities of
(10 mol%), Pd(t-Bu)3·HBF4 (20 mol%), KO-t-Bu (2 equiv.) in
toluene (5 mL) was purged with argon and stirred at 20 °C to
benzothiophene and benzofuran are lower than those of thio- give a brownish clear solution. To the stirred solution was
phene and furan, respectively. On the other hand, the aromati-
city of thiophene is considerably higher than the aromaticity
added 4a–c (1.0 equiv.), or 9a–c and aniline 5a–p (1.3 equiv.).
The reaction mixture was heated at 110 °C for 12 h. The solu-
of furan. While these differences are in line with the observed tion was cooled to 20 °C, poured into H2O and CH2Cl2 (25 mL
selectivities, a simple explanation for this effect cannot be
given.
each), and the organic and the aqueous layers were separated.
The latter was extracted with CH2Cl2 (3 × 25 mL). The
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 9464–9473 | 9471

Paper
Organic & Biomolecular Chemistry
View Article Online
combined organic layers were dried (Na2SO4), concentrated in
vacuo, and the residue was purified by chromatography (flash
silica gel, heptane–EtOAc) to give 6 and/or 7, 10 and/or 11.
References
1 According to the IUPAC nomenclature, a benzo-η-carboline
(here) is a 10H-indolo[3,2-b]quinoline.
2 D. Dwuma-Badu, J. S. K. Ayin, N. I. Y. Fiagbe, J. E. Knapp,
P. L. Shiff and D. J. J. Slatkin, Pharm. Sci., 1978, 67, 433.
3 E. Clinquart, Bull. Acad. R. Med. Belg., 1929, 9, 627.
4 A. N. Tackie, M. H. M. Sharaf, P. L. Schiff, Jr., G. L. Boye,
R. C. Crouch and G. E. Martin, J. Heterocycl. Chem., 1991,
28, 1429.
5 H. W. Rauwald, M. Kober, E. Mutschler and G. Lambrecht,
Planta Med., 1992, 58, 486.
6 K. Cimanga, T. de Bruyne, A. Lasure, B. van Poel, L. Pieters,
M. Claeys, D. Vanden Berghe, K. Kambu, L. Tona and
A. J. Vlietinck, Planta Med., 1996, 62, 22.
7 K. Cimanga, T. de Bruyne, L. Pieters, A. J. Vlietinck and
C. A. Turger, J. Nat. Prod., 1997, 60, 688.
8 C. L. Yang, C. H. Tseng, Y. L. Chen, C. M. Lu, C. L. Kao,
M. H. Wu and C. C. Tzeng, Eur. J. Med. Chem., 2010, 45,
602.
9 Y. L. Chen, C. H. Chung, I.-L. Chen, P. H. Chen and
H. Y. Jeng, Bioorg. Med. Chem., 2002, 10, 2705.
10 D. Edmont, R. Rocher, C. Plisson and J. Chenault, Bioorg.
Med. Chem. Lett., 2000, 10, 1831.
11 D. Narsinh and S. Anamik, Indian J. Pharm. Sci., 2001, 63,
211.
11-Benzyl-2-methylbenzofuro[3,2-b]quinoline (6a)
Following general procedure A, 6a was prepared from 3-bromo-
2-(phenylethynyl)benzofuran (4a) (150 mg, 0.5 mmol),
Pd(OAC)2 (11 mg, 10 mol%), P(t-Bu)3·HBF4 (29 mg, 20 mol%),
4-methylaniline (5a) (70 mg, 0.65 mmol), toluene (5 mL), KO-t-
Bu (112 mg, 1.0 mmol), as a brown solid (98 mg, 60%); mp
1
170–172 °C. H NMR (300 MHz, CDCl3): δ = 2.45 (s, 3H, CH3),
4.66 (s, 2H, CH2), 7.06–7.22 (m, 5H, ArH), 7.34–7.39 (m, 1H,
ArH), 7.44 (dd, 1H, J = 1.8, 8.8 Hz, ArH), 7.51–7.54 (m, 2H,
ArH), 7.78 (brs, 1H, ArH), 8.12 (d, 1H, J = 8.8 Hz, ArH),), 8.30
(d, 1H, J = 7.7 Hz, ArH). 13C NMR (62.9 MHz, CDCl3): δ = 22.1
(CH3), 31.1 (CH2), 122.2, 122.9, 123.5 (CH), 125.1 (C), 126.4
(CH), 126.6 (C), 128.4, 128.6, 129.5, 130.0, 130.5 (CH), 136.0,
138.6, 145.0, 145.7, 147.0, 159.2 (C). IR (KBr): ν = 3056, 3023,
2946, 2914, 2857, 2731, 1621 (w), 1594 (m), 1575, 1565, 1551
(w), 1502, 1491 (m), 1469 (w), 1443 (m), 1426, 1402, 1376,
1364, (w), 1337 (m), 1286, 1262, 1231 (w), 1184 (s), 1179, 1169,
1146, 1123, (w), 1109, 1068, 1043 (m), 1024, 1015, 1001, 958,
927, 903, 872, 858 (w), 820 (s), 783, 763 (m), 748, 732, 702 (s),
691, 672 (w), 644 (m), 620 (w), 589, 571, 554, cm−1. GC-MS (EI,
70 eV): m/z (%) = 323 ([M]+, 100), 322 (22), 308 (13), 307 (13),
246 (13). HRMS (EI, 70 eV): calcd for C23H17NO [M]+:
323.13047; found: 323.130011.
12 A. Shi, T. A. Nguyen, S. K. Battina, S. Rana, D. J. Takemoto,
P. K. Chiang and D. H. Hua, Bioorg. Med. Chem. Lett., 2008,
18, 3364.
13 H. Mahesh, C. Reddy, K. V. Reddy, P. V. K. S. Raju and
V. V. Reddy, Synth. Commun., 2004, 34, 4089.
14 A. A. Avetisyan, I. L. Aleksanyan and A. A. Pivazyan,
Russ. J. Org. Chem., 2006, 42, 739.
2-Ethyl-11-(4-methoxybenzyl)benzofuro[3,2-b]quinoline (6q)
Following general procedure A, 6q was prepared from 3-bromo-
2-((4-methoxyphenyl)ethynyl)benzofuran
0.5 mmol), Pd(OAC)2 (11 mg, 10 mol%), P(t-Bu)3·HBF4 (29 mg,
20 mol%), 4-ethylaniline (5d) (78 mg, 0.65 mmol), toluene 16 Y. Kawase, S. Yamaguchi, O. Maeda, A. Hayashi, K. Tabata
(5 mL), KO-t-Bu (112 mg, 1.0 mmol), as a brown solid (120 mg, and M. Kondo, J. Heterocycl. Chem., 1979, 16, 487.
72%); mp 160–162 °C. H NMR (250 MHz, CDCl3): δ = 1.23 (t, 17 Y. Kawase, S. Yamaguchi, M. Morita and T. Uesugi, Bull.
3H, J = 7.6 Hz, CH3), 2.76 (q, 2H, J = 7.6 Hz, CH2), 3.63 (s, 3H, Chem. Soc. Jpn., 1980, 53, 1057.
OCH3), 4.60 (s, 2H, CH2), 6.70 (d, 2H, J = 8.7 Hz, ArH), 7.13 (d, 18 L. W. Deady, A. J. Kaye, G. J. Finlay, B. C. Baguley and
2H, J = 8.7 Hz, ArH), 7.33–7.39 (m, 1H, ArH), 7.47 (dd, 1H, J = W. A. Denny, J. Med. Chem., 1997, 40, 2040.
(4c)
(150
mg, 15 A. A. Avetisyan, I. L. Aleksanyan and K. S. Sargsyan,
Russ. J. Org. Chem., 2007, 43, 426.
1
1.8, 8.7 Hz, ArH), 7.51–7.54 (m, 2H, ArH), 7.82 (brs, 1H, ArH), 19 X. Y. Zhu, L. G. Mardenborough, S. Li, A. Khan, W. Zhang,
8.14 (d, 1H, J = 8.7 Hz, ArH), 8.30 (d, 1H, J = 7.8 Hz, ArH). 13C
NMR (62.9 MHz, CDCl3): δ = 15.4 (CH3), 29.2, 30.3 (2CH2), 55.2
P. Fan, M. Jacob, L. Walker and S. Y. Ablordeppey, Bioorg.
Med. Chem., 2007, 15, 686, and references cited therein.
(OCH3), 112.2, 114.1, 121.7, 122.2, 123.4 (CH), 123.5, 125.7, 20 (a) M. S. Butler, J. Nat. Prod., 2004, 67, 2141; (b) J. A. Joule
126.6 (C), 128.9, 129.4, 129.7, 130.5 (CH), 130.7, 142.0, 145.3,
and K. Mills, Heterocyclic Chemistry, Blackwell Science,
145.8, 146.9, 158.2, 159.2 (C). IR (KBr): ν = 3055, 3024, 2956,
Oxford, 2000.
2923, 2866, 1621 (w), 1607 (m), 1575, 1564, 1555 (w), 1507 (s), 21 For general reviews on the metal-catalyzed synthesis of het-
1470 (w), 1445 (m), 1403, 1366 (w), 1340, 1331 (m), 1285, 1264
(w), 1245 (s), 1230, 1179, 1170, 1148, 1128 (w), 1110 (m), 1068,
1056 (w), 1042 (m), 1025, 1015, 1001, 993, 981, 953, 937, 927,
908, 899 (w), 878 (m), 856 (w), 830 (s), 784, 760, 748 (m), 733,
725, 700 (s), 672, 646, 620, 605, 587, 556 (m), cm−1. GC-MS (EI,
70 eV): m/z (%) = 367 ([M]+, 100), 352 (15), 338 (17), 322 (06),
erocycles, see: (a) I. Nakamura and Y. Yamamoto, Chem.
Rev., 2004, 104, 2127; (b) T. L. Gilchrist, J. Chem. Soc.,
Perkin Trans. 1, 1999, 2849; (c) J. F. Hartwig, Synlett, 2006,
1283; (d) L. Jiang and S. L. Buchwald, in Metal-Catalyzed
Cross-Coupling Reactions, ed. A. de Meijere and F. Dieder-
ich, Wiley-VCH, Weinheim, 2nd edn, 2004, p. 699.
307 (10), 294 (08), 259 (06). HRMS (EI, 70 eV): calcd for 22 (a) L. Ackermann, S. Barfüßer and H. K. Potukuchia, Adv.
C25H21NO2 [M]+: 367.15668; found: 367.156544.
Synth. Catal., 2009, 351, 1064; (b) L. Ackermann,
9472 | Org. Biomol. Chem., 2012, 10, 9464–9473
This journal is © The Royal Society of Chemistry 2012

Organic & Biomolecular Chemistry
Paper
View Article Online
R. Sandmann, A. Villar and L. T. Kaspar, Tetrahedron, 2008, 31 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
64, 769.
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers,
K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand,
K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene,
J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas,
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaus-
sian 09, Rev. C.01, Gaussian, Inc., Wallingford CT, 2009.
32 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
23 L. Ackermann, Org. Lett., 2005, 7, 439.
24 (a) R. Martin, M. Rodriguez Rivero and S. L. Buchwald,
Angew. Chem., 2006, 118, 7237; R. Martin, M. Rodriguez
Rivero and S. L. Buchwald, Angew. Chem., Int. Ed., 2006, 45,
7079; (b) R. Martin, C. H. Larsen, A. Cuenca and
S. L. Buchwald, Org. Lett., 2007, 9, 3379.
25 For related transformations, see: (a) Z.-Y. Tang and
Q.-S. Hu, Adv. Synth. Catal., 2006, 348, 846; (b) Y. Yu,
G. A. Stephenson and D. Mitchell, Tetrahedron Lett., 2006,
47, 3811; (c) K. Hiroya, S. Itoh and T. Sakamoto, J. Org.
Chem., 2004, 69, 1126; (d) L. Xu, I. R. Lewis, S. K. Davidsen
and J. B. Summers, Tetrahedron Lett., 1998, 39, 5159.
26 (a) R. C. Larock and E. K. Yum, J. Am. Chem. Soc., 1991,
113, 6689; (b) R. C. Larock, E. K. Yum and M. D. Refvik,
J. Org. Chem., 1998, 63, 7652.
27 R. C. Larock, E. K. Yum, M. J. Doty and K. K. Sham, J. Org.
Chem., 1995, 60, 3270.
28 K. R. Roesch and R. C. Larock, J. Org. Chem., 1998, 63, 33 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785.
5306.
34 P. C. Hariharan and J. A. Pople, Theor. Chim. Acta, 1973, 28,
29 (a) R. C. Larock, M. J. Doty, Q. Tian and J. M. Zenner,
213.
J. Org. Chem., 1997, 62, 7536; (b) R. C. Larock and Q. Tian, 35 (a) P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 270;
J. Org. Chem., 1998, 63, 2002. (b) P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 299.
30 G. A. Salman, M. Hussain, V. O. Iaroshenko, A. Villinger 36 Y. Wang, J. I.-C. Wu, Q. Li and P. v. R. Schleyer, Org. Lett.,
and P. Langer, Adv. Synth. Catal., 2011, 353, 331.
2010, 12, 4824.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 9464–9473 | 9473