Asymmetric phase-transfer-catalyzed intramolecular N-alkylation of indoles and pyrroles: A combined experimental and theoretical investigation

Article abstract of DOI:10.1002/chem.201000560

Asymmetric phase-transfer catalysis (PTC) has risen to prominence over the last decade as a straightforward synthetic methodology for the preparation of pharmacologically active compounds in enantiomerically pure form. However, the complex interplay of weak nonbonded interactions (between catalyst and substrate) that could account for the stereoselection in these processes is still unclear, with tentative pictorial mechanistic representations usually proposed. Here we present a full account dealing with the enantioselective phase-transfer-catalyzed intramolecular aza-Michael reaction (IMAMR) of indolyl esters, as a valuable synthetic tool to obtain added-value compounds, such as dihydro-pyrazinoindolinones. A combined computational and experimental investigation has been carried out to elucidate the key mechanistic aspects of this process. La catalisi asimmetrica per trasferimento di fase e, a tutt'oggi, una fra i metodi di sintesi stereoselettivi pi efficaci per la preparazione di intermedi chiave dell′indotto farmaceutico. Tuttavia, i complessi network di legami deboli che generalmente si instaurano fra catalizzatore e substrato in queste trasformazioni non consentono di comprendere appieno gli aspetti alla base delle elevate stereoselezioni ottenute. In questo studio, con un approccio combinato sperimentale/computazionale, vengono esplicitati gli aspetti meccanicistici chiave nella sintesi enantioselettiva di composti ad elevato valore aggiunto come i diidro-pirazinoindoloni, tramite addizione tipo aza-Michael intramolecolare per trasferimento di fase (IMAMR). Union is strength: A combined computational and experimental investigation has shed light on a network of weak interactions responsible for the high stereochemical translation in the phase-transfer-catalyzed intramolecular N-alkylation of indoles (see figure).

Full text of DOI:10.1002/chem.201000560

DOI: 10.1002/chem.201000560
Asymmetric Phase-Transfer-Catalyzed Intramolecular N-Alkylation of
Indoles and Pyrroles: A Combined Experimental and Theoretical
Investigation
Marco Bandini,*[a] Andrea Bottoni,[a] Astrid Eichholzer,[a] Gian Pietro Miscione,*[a] and
Marco Stenta[b]
Abstract: Asymmetric phase-transfer
catalysis (PTC) has risen to promi-
tion in these processes is still unclear,
with tentative pictorial mechanistic
representations usually proposed. Here
we present a full account dealing with
the enantioselective phase-transfer-cat-
alyzed intramolecular aza-Michael re-
action (IMAMR) of indolyl esters, as a
valuable synthetic tool to obtain
added-value compounds, such as dihy-
dro-pyrazinoindolinones. A combined
computational and experimental inves-
tigation has been carried out to eluci-
date the key mechanistic aspects of this
process.
nence over the last decade as
a
straightforward synthetic methodology
for the preparation of pharmacological-
ly active compounds in enantiomeri-
cally pure form. However, the complex
interplay of weak nonbonded interac-
tions (between catalyst and substrate)
that could account for the stereoselec-
Keywords: asymmetric catalysis
·
computational chemistry · indoles ·
Michael addition · phase-transfer
catalysis
Introduction
based and metal-free catalysts.[3] Additions to carbonyl com-
pounds, Michael additions to electron-deficient olefins, hy-
ꢀ
The indole nucleus represents a key molecular motif with
widespread occurrence in nature and featuring peculiar
pharmacological and agrochemical activities.[1] Accordingly,
it is not surprising that a great deal of attention has been de-
voted to the use of innovative organic synthetic methodolo-
gies to prepare tailored synthetic indolyl-containing com-
pounds.[2] Within these synthetic approaches, the develop-
ment of enantioselective Friedel–Crafts alkylations (FCAs)
has triggered increasing interest over the last few years,
thanks to the growing availability of efficient chiral metal-
droarylations of unactivated C C multiple bonds, and SN2/
SN1 processes involving epoxides/p-activated alcohols de-
serve particular mention among the most reliable catalytic
stereoselective FC-type processes.[4]
Indole is commonly referred to as an electron-rich (gener-
ous) heteroaromatic system, showing enhanced reactivity
(compared to benzene) in electrophilic aromatic substitu-
tions. It has been involved in nearly 70% of the reported ex-
amples addressing catalytic and stereoselective FCA pro-
cesses with the quasi-exclusive functionalization of the most
nucleophilic C3 position. On the other hand, direct C2-se-
lective catalytic enantioselective alkylations of indoles is still
an open synthetic task (see Figure 1). Commonly, an alter-
native indirect strategy, which consists of a two-step alkyla-
tion/oxidation protocol, involving 4,7-dihydroindoles as Frie-
del–Crafts partners, is used for this purpose.[5]
[a] Dr. M. Bandini, Prof. A. Bottoni, Dr. A. Eichholzer,
Dr. G. P. Miscione
Dipartimento di Chimica “G. Ciamician”
Alma Mater Studiorum—Universitꢀ di Bologna
via Selmi 2, 40126 (Italy)
In contrast, the stereoselective functionalization of indoles
at the N1 position, which would represent a valuable short-
cut to numerous enantiomerically pure natural tricyclic in-
dolyl alkaloids (e.g., Mitomycins[6] and indolocarbazoles[7])
is still mostly unexplored. Only a few examples have been
reported so far: chiral Pd (Trost)[8a,b] and Ir (Hartwig)[8c]
complexes have been found to assist the chemo-, regio-, and
stereoselective N-allylic alkylation of indoles, and very re-
Fax : (+39)51-2099456
E-mail: marco.bandini@unibo.it
gianpietro.miscione@unibo.it
[b] Dr. M. Stenta
Laboratory for Biomolecular Modeling
Ecole Polytechnique Fꢁdꢁrale de Lausanne (EPFL)
1015 Lausanne (CH)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000560.
12462
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12462 – 12473

FULL PAPER
Figure 1. Catalytic enantioselective routes to functionalized indolyl cores.
Regioselective N1-alkyaltion is still a synthetic challenge.
Figure 2. a) General formula structures of 3,4-dihydropyrazino
ACHTUNGTRENNUNG[1,2-
a]indol-1(2H)-ones (1) and 1,2,3,4-tetrahydropyrazino[1,2-a]indoles (2).
A
ACHTUNGTRENNUNG
b) Pictorial representation of the postulated contact ion pair between
chiral quaternary ammonium salts and indolinate intermediate in the
PTC enantioselective intramolecular N-alkylation of indoles.
cently Chen[9a] and Enders[9b] have demonstrated the effi-
ciency of chiral organocatalysis in the stereoselective indole
N1 functionalization through Morita–Baylis–Hillman reac-
tions and domino aza-Michael/aldol condensation reactions,
respectively.
As a part of our continuing effort at developing new cata-
lytic and enantioselective procedures for the “decoration”
of indolyl cores,[2f,10] we have recently addressed the chal-
lenging N1 alkylation reaction based on the use of intramo-
lecular aza-Michael addition.[11] The main goal of the proto-
col was the preparation of active pharmaceutical ingredients
forming process, exploiting the preferential N1-site reactivi-
ty when electron-withdrawing groups (EWGs) are present
at both C2 and C3 indole positions.[9] In particular, based on
our recent findings on the K2CO3-catalyzed synthesis of
functionalized
pyrazino-indolinones
(1
with
R1 =
CH2CO2Et),[15] we envisioned that intramolecular aza-Mi-
chael addition, under phase-transfer-catalysis (PTC)[16] con-
ditions, could ensure the formation of sufficiently tight ion
pairs between the chiral ammonium salt and the indolate in-
termediate (obtained from 3) to provide efficient stereodis-
crimination. A schematic representation of the hypothetical
tight ion pair, with a portion of quaternary ammonium salt
deriving from Cinchona alkaloids, is given in Figure 2b. It is
worth pointing out that, despite the impressive synthetic po-
tential for the preparation of nitrogen-containing heterocy-
cles, the catalytic and enantioselective intramolecular aza-
Michael addition still remains quite an unexplored field.[17]
After a severe optimization phase, we were delighted to
find out that N-benzylcinchonidinium bromide 4a
(10 mol%), carrying a strong EWG (e.g., CF3) in the para
position of the benzyl substituent, proved to be an effective
catalyst for the ring-closing event. The phase-transfer
regime (toluene, KOHaq, ꢀ458C) led to (S)-1a in 93% yield
and 91% enantiomeric excess (Scheme 1).
(APIs), such as functionalized 3,4-dihydropyrazino
ACHTUNGTNER[NUNG 1,2-
a]indol-1(2H)-ones (1), which have recently attracted atten-
ACHTUNGTRENNUNG
tion as specific inhibitors of serotonergic receptors and for
their noncompetitive antihistamine activity (Figure 2a).[12]
Moreover, structurally correlated 1,2,3,4-tetrahydropyrazino-
ACHTUNGTRENNUNG[1,2-a]indoles (2) have risen to prominence as effective me-
dicinal additives for anti-obesity, and in the treatment of
non-insulin-dependent diabetes.[13] Notably, the direct syn-
thetic access to such a class of compounds in enantiomeri-
cally pure form is still a demanding task, and the existing
methodologies typically require crucial preparative steps,
such as the resolution of racemates or the use of chiral aux-
iliaries/chiral pool.[13,14]
Our working hypothesis deals with the direct construction
ꢀ
of 1 through a stereoselective intramolecular C N bond-
The reaction scope was then promptly studied. Wide tol-
erance towards electron-deficient and electron-donating
Abstract in Italian: La catalisi asimmetrica per trasferimento
di fase ꢀ, a tuttꢁoggi, una fra i metodi di sintesi stereoselettivi
piꢂ efficaci per la preparazione di intermedi chiave dell’indot-
to farmaceutico. Tuttavia, i complessi network di legami
deboli che generalmente si instaurano fra catalizzatore e sub-
strato in queste trasformazioni non consentono di compren-
dere appieno gli aspetti alla base delle elevate stereoselezioni
ottenute. In questo studio, con un approccio combinato speri-
mentale/computazionale, vengono esplicitati gli aspetti mec-
canicistici chiave nella sintesi enantioselettiva di composti ad
elevato valore aggiunto come i diidro-pirazinoindoloni, tra-
mite addizione tipo aza-Michael intramolecolare per trasferi-
mento di fase (IMAMR).
Scheme 1. Optimal reaction conditions for the PTC asymmetric synthesis
of pyrazinoindolone 1a.
Chem. Eur. J. 2010, 16, 12462 – 12473
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12463

M. Bandini, G. P. Miscione et al.
substituents on the indolyl ring was recorded, thus providing
direct access to a library of indolinones 1 in high yields and
with excellent enantiomeric excesses. A representative col-
lection of tricyclic compounds (1b–g) is reported in
Scheme 2.
scribing intramolecular processes have been reported so far.
Prompted by such a task, we present here a full account on
the PTC asymmetric aza-Michael alkylation of indoles. In
particular, we carry out a combined experimental and theo-
retical investigation of the reaction reported in Scheme 1, in
which we selectively introduce specific structural variations
in the substrate and catalyst, with the aim of gaining an in-
sight into the mechanistic aspects of this process.
Results and Discussion
The hypothetical mechanism of the PTC intramolecular N1-
indole alkylation can be thought to involve three main
events. Initially, the indolyl ester 3 is deprotonated by OHꢀ
ions provided by the external inorganic species (presumably
at the interfacial region), and the resulting reactive indoli-
nate intermediate is intercepted by the chiral quaternary
ammonium salt. The well-constructed structure of the Cin-
chona alkaloid provides the electronic and steric constraints
that favor a proper folding of the acyclic precursor 3 to
reach a structural arrangement in which the nucleophilic
(indole nitrogen atom) and electrophilic (b-carbon atom of
the ester moiety) centers are spatially close. It is plausible to
believe that a network of electronic interactions between
the substrate and the peripheral catalyst substituents work
simultaneously to stabilize the ring-closing transition state
(TS) with synchronized control of the stereochemistry of the
process. Finally, the negatively charged enolate intermediate
is rapidly neutralized by the proton originating from the re-
aromatization of the indole nucleus, leading to the final
cyclic product 1, which is released in solution.
Scheme 2. Enantiomerically enriched dihydro-pyrazinoindolinones syn-
thesized through asymmetric PTC (PMP: para-methoxyphenyl; *: after
recrystallization).
It has been frequently observed that Cinchona mediated,
asymmetric phase-transfer catalysis deals with chemical and
optical outcomes which are strongly related to the structural
decoration of the chiral ammonium salts at the C9 position,
at the C6’ position of the quinoline ring, and the benzyl
framework (Figure 3). This aspect has generally caused frus-
The crucial phase for the actual overall transfer of chirali-
ty from the catalyst to the final pyrazinoindolone should be
represented by the formation of conformationally rigid and
stereochemically defined 3a–4a ion pair resulting from a
network of complementary attractive interactions between
catalyst and substrate. Within this conceptual model, the
ester moiety (A), the indole ring (B), and the amide group
(C) should represent crucial binding sites of the substrate
(see Figure 3). On the other hand, it is conceivable that the
Cinchona derivative preferentially interacts with the precur-
sor 3a through the hydroxyl group (a potential source of hy-
drogen bonds) at the C9 position (D), the positively charged
quinuclidine nitrogen atom (E), the benzyl group bonded to
nitrogen (F), and the quinoline ring (G).
With this model in mind, we have carefully investigated
the potential-energy surface (PES) for the complexes that
originate from the interaction and binding of the deproton-
ated model substrate 3c with the catalyst 4a (Scheme 1).
The key factors that differentiate the possible resulting ad-
ducts are the weak “complementary” interactions that
engage the various groups of substrate (A, B, and C) and
catalyst (D, F, and G) as previously shown in Figure 3. Most
of these attracting “contacts” are parallel or T-shaped p-
stacking interactions involving the various p systems of both
molecules. An additional important interaction should be
Figure 3. Crucial binding sites of substrate and Cinchona derivative
acting as a catalyst.
trating and time-consuming trial-and-error approaches for
discovering optimal conditions, because of the large number
of structural variables to be taken into account. Within this
context, a reliable model providing a rationale of the net-
work of weak, nonbonded catalyst–substrate interactions re-
sponsible for the enantiodiscriminating event would dramat-
ically enhance the synthetic feasibility of such a class of pro-
cesses in the total synthesis of natural products and large-
scale productions by shortening the optimization phase.
Despite the extent of chiral Cinchona ammonium salts in
asymmetric phase-transfer catalysis, and the enormous ef-
forts devoted to elucidating its mechanism of action (mainly
concerning alkylation processes),[18] very few examples de-
12464
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12462 – 12473

Intramolecular N-Alkylation of Indoles and Pyrroles
FULL PAPER
the effective hydrogen bond that can be established between
the free hydroxyl group (D) bonded at the C9 position of
the catalyst and the substrate carbonyl groups of the ester
(C) or the amide group (A).
It is important to stress some structural features of M1-S
and M1-R. In both structures, the two atoms involved in the
ꢀ
formation of the new C N bond (e.g., the formally negative
nitrogen atom N2 and the electrophilic C2 carbon atom of
the a,b-unsaturated ester group) are facing each other, and,
thus, they already point in a suitable direction for the nucle-
ophilic attack to take place. Isomers M1-S and M1-R differ
mainly in the orientation of the ester moiety. Interestingly,
the two species are conformational isomers which can inter-
Among the various catalyst–substrate adducts located on
the PES, one special type of structural arrangement has
been found to be more stable than the others. This particu-
lar arrangement characterizes the two conformationally
rigid isomers M1-S and M1-R, which represent the starting
complexes for the two reaction channels leading to the (S)-
1c and (R)-1c products, respectively. A schematic represen-
tation of the two structures is given in Figure 4, in which a
set of important geometrical parameters is also reported.
One point concerning these drawings should be stressed
here. Because, in addition to a conventional three-dimen-
sional picture (bottom part of Figure 4) we have also adopt-
ed a two-dimensional representation, in this latter case sev-
eral atomic distances are not realistic and appear much
longer (or shorter) than in the real molecule. A comparison
to the more realistic three-dimensional picture reported in
the same figure should help to comprehend the relative po-
sitions of the various groups. A further point that must be
mentioned concerns the numbering of the various atoms: in
the following discussion we have replaced the separated no-
tations for substrate and catalyst (previously used) with a
global notation (reported in the figure) for the whole sub-
strate–catalyst adduct. Furthermore, for the sake of simplici-
ty, we have indicated only the atoms explicitly considered in
the discussion.
ꢀ
convert by a simple rotation around the C1 C2 bond, thus
exposing one or other of the two ester pro-chiral faces to
the attack of the nucleophilic nitrogen N2.
We consider first in detail the structural features of the
more stable isomer M1-S, affording the (S)-1c enantiomeric
ꢀ
product (C2 N2 distance for the new forming bond=
3.021 ꢃ). The main force driving the substrate and catalyst
together is electrostatic in nature. Two opposite charges are
located on the former (negative) and latter (positive) spe-
cies. Even if the two charges are formally located on N2 and
N3, the computational results clearly show that they are
strongly delocalized. The real (computed) net charge on N2
is ꢀ0.58, but significant negative charges are also present on
O1 and O3: ꢀ0.61 and ꢀ0.63, respectively. Furthermore, the
actual computed net charge on N3 is +0.57, but other atoms
of the catalyst belonging to the quinoline ring bear a signifi-
cant positive charge: +0.86 on C8 and +0.70 on C7. Addi-
tional attractive interactions between substrate and catalyst
are evident in M1-S. First of all, as previously suggested, a
strong hydrogen bond involves H1 (D hydroxyl group
Figure 4. A schematic representation of M1-S and M1-R in two and three dimensions. Bond lengths are in ꢃ. When not connecting atoms, dashed lines
connect the ring centroids.
Chem. Eur. J. 2010, 16, 12462 – 12473
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12465

M. Bandini, G. P. Miscione et al.
O2H1) and O1 (carbonyl oxygen), as evidenced by the short
Now the following question arises: why is the ester chain
bent in M1-R even if this arrangement results in an overall
destabilization? An explanation can be found by analyzing
the structure of the aminopropene unit N1-C1-C2-C3. It is
known that in propene the eclipsed conformation is more
stable than the staggered one. This can easily be explained
by the change of orbital interactions between the two out-
of-plane methyl hydrogen atoms and the double-bond p
system on passing from the staggered to the eclipsed confor-
mer.[19] In agreement with this structural preference, in both
M1-S and M1-R adducts the aminopropene unit adopts an
ꢀ
H1···O1 distance (1.682 ꢃ) and the lengthening of the C5
O1 carbonyl bond (1.247 ꢃ). The crucial role of the hydrox-
yl group of 4a in tightening the catalyst/indolinate ion pair
has also been demonstrated experimentally. For instance,
ꢀ
the model reaction, in which C OH was replaced by C-O-
allyl in 4a, afforded 1a in good yield (85%) but in nearly
racemic form (ee=5%). Two p-stacking interactions (both
engaging the catalyst electron-poor benzyl–CF3 ring) con-
tribute to bind the substrate and catalyst in a rigid confor-
mation. One of the two interactions is p T-shaped and in-
volves the catalyst benzyl–CF3 group and the substrate
indole ring. The second interaction is a weaker parallel p-
stacking interaction occurring between the catalyst benzyl–
CF3 group and the substrate ester p system, which lie on
parallel planes. As a matter of fact both interactions are nei-
ther perfect T-shaped nor parallel p-stacking interactions. A
slightly oblique relative assembly of the interacting groups is
evident from our computations, and this allows the benzyl–
CF3 ring to interact with both partners. A series of test cal-
culations on a simpler model system (see Figure S1 of the
Supporting Information) has shown that a perfect T-shaped
interaction between the benzyl–CF3 ring and the indole sys-
tems would be highly favored. However, such a perfect T-
shaped arrangement is not actually observed in the real
system due to steric hindrance, and because it would not
allow the additional interaction between the benzyl–CF3
ring and the ester. Additional calculations have demonstrat-
ed that a different orientation of the benzyl group bonded
to N1, and resulting in a p-stacking interaction with quino-
line, is less stable than the one actually observed, for which
the N1-benzyl ring rotates and moves far away from the
quinoline group. It is evident that, when quinoline and the
N1-benzyl are spatially close, the presence of the benzyl
ring causes a steric perturbation on the conjugation of the
planar N2-C6-C5-N1-C1 system. Thus, a conformation in
which this conjugation is preserved is favored at the expense
of the loss of the benzyl–quinoline interaction.
ꢀ
eclipsed conformation (i.e. the C1 N1 bond is eclipsing the
C2=C3 double bond). However, in M1-R this moves the
ester chain far away from the benzyl–CF3 ring, thus prevent-
ing any stabilizing p-stacking interaction. Thus, the more
stable eclipsed conformation of the aminopropenic unit con-
tributes in turn to the lower stability of M1-R with respect
ꢀ
ꢀ
to M1-S. The relative position of the C1 N1 and C2 C3
bonds is evident in the three-dimensional representation of
Figure 4. However, for a better understanding of the amino-
propene fragment structural features, a more detailed repre-
sentation is given in Figure S2 of the Supporting Informa-
tion.
A conformational transition state (rotation around the
ꢀ
C1 C2 bond) connecting the two isomers M1-S and M1-R
has been located (see Figure S3 in the Supporting Informa-
tion). The two barriers (2.27 and 1.35 kcalmolꢀ1 considering
M1-S and M1-R, respectively, as the starting point) are con-
sistent with the rotational barrier values usually found for
single bonds. These values show that the interconversion be-
tween the two minima is a rather easy process.
Other possible conformers that can be obtained from M1-
S and M1-R have been examined. In particular, rotation
ꢀ
around the C3 C4 bond has been considered in both spe-
cies. Specifically, we have located a conformer (obtained
from M1-R) in which the ester chain is twisted towards the
benzyl–CF3 fragment. However, in all cases examined here,
the new structures are higher in energy with respect to M1-S
and M1-R. Thus, all the possible reaction paths stemming
from these additional initial adducts are quite unlikely, and
will not be discussed in the following.
The conformer M1-R lies 0.92 kcalmolꢀ1 higher in energy
than M1-S. Analysis of the various interactions can provide
a rationale for this energy gap. Firstly, the H1···O1 hydrogen
bond is weaker in M1-R than in M1-S (the H1···O1 distan-
ces are 1.719 and 1.682 ꢃ, respectively). Also, the p-stacking
interaction between the benzyl–CF3 group and the indole
rings becomes weaker. This is evidenced by the lengthening
of the distance between H3 and the centroid of the six-
membered ring of the indole fragment (2.640 and 2.536 ꢃ in
M1-R and M1-S, respectively). Finally, in M1-R, the interac-
tion between the substrate ester group and the catalyst
benzyl–CF3 fragment is absent, because the two groups are
now too distant. These structural features are a direct conse-
quence of the different orientation of the ester moiety in
the two isomers. Whereas in M1-S the C1-C2-C3-C4 plane is
almost perfectly orthogonal to the C1-N2-C5-C6-indole
system plane (the dihedral angle N1-C1-C2-C3 is 28), in M1-
R, the ester is bent away from the benzyl–CF3 ring (the di-
hedral angle N1-C1-C2-C3 becomes 1178).
ꢀ
A transition state describing the C2 N2 bond formation
and the consequent ring closure has been found along both
reaction paths (originating from M1-S and M1-R). The two
transition states are TS1-S and TS1-R, and lead to the (S)-
1c and (R)-1c products, respectively (Figure 5). Transition
state TS1-S is 5.73 kcalmolꢀ1 higher than the preceding com-
plex M1-S (intrinsic activation barrier). The formation of
ꢀ
the new C2 N2 bond (bond length=2.003 ꢃ) following the
nucleophilic attack causes a shifting of the electronic density
along the C2-C3-C4-O3 framework and
a consequent
change in character of the three bonds. In particular, the
ꢀ
ꢀ
single-bond character of C2 C3 (1.373 ꢃ) and C4 O3
ꢀ
(1.229 ꢃ) increases, whereas C3 C4 (1.421 ꢃ) is turning
into a double bond. The strength of the important intermo-
lecular O1···H1 hydrogen bond does not change too much,
the O1···H1 distance being only 0.019 ꢃ longer with respect
12466
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12462 – 12473

Intramolecular N-Alkylation of Indoles and Pyrroles
FULL PAPER
Figure 5. A schematic representation of the two transition states TS1-S and TS1-R and the corresponding product adducts M2-S and M2-R. Bond lengths
are in ꢃ. When not connecting atoms, dashed lines connect the ring centroids.
to M1-S. Because in this transition structure the ester chain
is twisting toward the indole moiety, it moves nearer to the
benzyl–CF3 fragment, and the relative p-stacking interac-
tions become stronger. This is evidenced by the variation of
the distances between the centroid of the benzene ring and
the two atoms C4 and O4 on passing from M1-S to TS1-S.
These parameters decrease from 4.733 and 5.058 ꢃ to 4.332
and 4.203 ꢃ, respectively.
that characterizes TS1-S into a less stabilizing parallel p in-
teraction in TS1-R (compare TS1-S and TS1-R in Figure 5).
As mentioned above, calculations on simpler model systems,
for which only the p systems of the indole and benzyl–CF3
groups are considered, clearly show that a T-shaped p inter-
action is more stable than a parallel one, and, when steric
hindrance can be neglected, this represents the preferred ar-
rangement. In summary, our analysis indicates that the
higher energy of TS1-R with respect to TS1-S is caused by:
1) a weaker O1···H1 interaction; 2) the unfavorable confor-
mation of the amino-propene unit; 3) the lack of p-stacking
interactions involving the ester moiety; and 4) the different
and less favorable relative orientation of the indole and
benzyl–CF3 groups.
TS1-S and TS1-R lead to the product adducts M2-S and
M2-R, respectively (bottom sections of Figure 5a and b).
The two structures, characterized by the six-membered cycle
C1-N1-C5-C6-N2-C2, are epimers, which differ in their con-
figuration at the C2 center. In both cases, the cycle adopts a
half-chair-like conformation, where C1, N1, C5, C6, and N2
are coplanar and C2 is bent out of the plane. In M2-S the
unsaturated chain is pseudo-axial, whereas it becomes
pseudo-equatorial in M2-R. Even if the negative charge is
The computed activation energy for TS1-R (along the al-
ternative (R)-1c pathway) is 6.90 kcalmolꢀ1, and, important-
ly, TS1-R is 2.09 kcalmolꢀ1 higher in energy than TS1-S. The
ꢀ
length of the new forming bond C2 N2 (2.008 ꢃ) varies
only slightly with respect to TS1-S. In contrast, the intermo-
lecular H1···O1 distance becomes longer (1.780 ꢃ) than in
TS1-S, indicating a weakening of this relevant interaction.
Two important structural differences are evident in the
comparison between TS1-S and TS1-R: these are the inner
conformation of the ester chain and the relative position of
the indole and benzyl–CF3 aromatic systems. In TS1-R, the
direction of the ester chain allows N2 and C2 to reach suita-
ble relative positions for the formation of the new bond, but
forces the aminopropene framework (N1-C1-C2-C3) to
adopt a less stable staggered conformation. As a conse-
quence, the ester moiety cannot be properly oriented for an
effective p-stacking interaction with the benzyl–CF3 frag-
ment. At the same time, to avoid destabilizing steric interac-
tions, the benzyl–CF3 ring must rotate, and its plane be-
comes almost parallel to the indole plane. This structural
change transforms the indole-benzyl T-shaped p interaction
ꢀ
ꢀ
ꢀ
now formally located on O3, the C2 C3, C3 C4 and C4 O3
ꢀ
bond lengths clearly indicate a strong delocalization: C3 C4
is longer than a double bond (1.376 ꢃ in M2-S and 1.374 ꢃ
ꢀ
in M2-R), whereas C2 C3 (1.474 ꢃ in M2-S and 1.465 ꢃ in
ꢀ
M2-R) and C4 O4 (1.249 ꢃ in M2-S and 1.269 ꢃ in M2-R)
are shorter than normal single bonds. In M2-S the ester
Chem. Eur. J. 2010, 16, 12462 – 12473
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12467

M. Bandini, G. P. Miscione et al.
chain is slightly twisted towards the benzyl–CF3 fragment.
This arrangement gives rise only to a weak p-stacking inter-
action, which is not possible in M2-R, where (as noted in
TS1-R) the benzyl and indole planes are parallel and the
ester moiety is not involved in any p-stacking interaction.
The energy profile of the reaction, which is exothermic
(the product M2-S is 7.09 kcalmolꢀ1 lower than M1-S) is re-
ported in Figure 6. It is interesting to note that the relative
stability of the two product adducts is reversed with respect
to that of the corresponding reagents and transition states,
M2-R being 2.38 kcalmolꢀ1 more stable than M2-S. This
finding cannot be ascribed to the O1···H1 hydrogen bond,
because the O1···H1 distance increases on going from transi-
tion states to products, but remains significantly shorter in
M2-S (1.769 ꢃ) than in M2-R (2.020 ꢃ). It is plausible that
an important contribution to the greater stability of M2-R is
represented by the pseudo-equatorial orientation of the
ester substituent in the cycle.
tween the two products is determined by the energy differ-
ence between the corresponding transition states TS1-R and
TS1-S. This difference is equal to 2.09 kcalmolꢀ1, and at
ꢀ458C this value corresponds to a ratio M2-S:M2-R of ap-
proximately 99:1. This would lead to a 98% ee, which is in
good agreement with the experimentally observed value of
90%.
To reveal the effects of the catalyst we have carried out
additional computations on the sole substrate (uncatalyzed
reaction). The results can be summarized as follows: a) the
two starting isomers are very close in energy, their energy
difference being only 0.31 kcalmolꢀ1 (with the R isomer fa-
vored); b) the two transition states become almost degener-
ate because their energy gap drops from 2.09 kcalmolꢀ1 to
0.03 kcalmolꢀ1. Thus, it is evident that the catalyst inclusion
(and its interaction with the substrate) is essential for differ-
entiating the two diastereomeric pathways.
The role of the catalyst–substrate interactions has been
further investigated by means of new calculations on “mu-
tated” complexes. To this purpose we have considered the
model substrate 3h (Scheme 3) in which the amide benzyl
Scheme 3. Proving the effect of the amide substituents on the course of
the asymmetric cyclization.
fragment has been replaced by a methyl group. We have
found that, after optimization, the energy difference be-
tween the two starting complexes M3-S and M3-R (Fig-
ure S4 in the Supporting Information) is 0.79 kcalmolꢀ1 with
M3-S favored (thus, very close to the value of 0.92 kcal
molꢀ1 found for M1-S and M1-R). Our computations show
that the O1···H1 hydrogen bond becomes weaker on passing
from M3-S to M3-R (the H1···O1 distance is 0.130 ꢃ longer
in the latter case). However, in M3-R, as the cumbersome
benzyl group is lacking, the quinoline can approach the
ester group to give rise to a T-shaped p-stacking interaction.
The structural features indicate that this interaction is
weaker in M3-S and, thus, can partly balance the reduced
strength of the O1···H1 hydrogen bond. The computations
carried out on the transition states TS2-R and TS2-S along
the two reaction pathways originating from M3-R and M3-S,
respectively, show that their energy difference (2.10 kcal
molꢀ1) is nearly identical to the value found for TS1-R and
TS1-S, and the S pathway is still favored. This finding is in
agreement with the evidence that the benzyl group did not
Figure 6. Computed reaction energy profile.
We now compare the computed energy profile to the ex-
perimental evidence. Because the reaction is carried out
under kinetic conditions (at a temperature of ꢀ458C), to
evaluate the ratio between the two possible products (and
ultimately the reaction ee) we do not have to consider their
relative stability, which would erroneously indicate the R
isomer as the favored product. As the barrier for the confor-
mational interconversion between the two starting com-
plexes is significantly lower (2.31 kcalmolꢀ1 for the M1-S!
M1-R conversion) than that required by the ring-closing
process (5.73 kcalmolꢀ1 for the M1-S!TS1-S transforma-
tion), following the Curtin–Hammet principle, the ratio be-
12468
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12462 – 12473

Intramolecular N-Alkylation of Indoles and Pyrroles
FULL PAPER
concur substantially with the catalyst–model substrate aggre-
gation, and, consequently, to the differentiation of the two
reaction channels.
puted for M1-S (89.60 and 91.29 kcalmolꢀ1, respectively).
The two corresponding transition states, TS3-S and TS3-R
(also depicted in Figure S5), are almost degenerate in
energy, TS3-R being only 0.16 kcalmolꢀ1 lower than TS3-S.
This computational outcome is in agreement with the mod-
erate stereoselection (ee=59%, see Scheme 4) observed ex-
perimentally. The reduced energy gap between the two tran-
sition states is a direct consequence of the replacement of
an amide group by an ester group. Because the conjugation
in the N2-C6-C5-O-C1 system becomes poorer, the oxygen
is now placed out of the N2-C6-C5 plane. Especially in TS3-
R, this allows the ester chain to adopt a different conforma-
tion that minimizes the steric repulsion with the benzyl–CF3
ring, which, in turn, can establish a stronger key p-stacking
interaction with the indole fragment with respect to the cor-
responding structures with the amide group. This structural
change in the substrate also determines a lengthening of the
O1···H1 hydrogen bond with respect to the reference transi-
tion states TS1-S and TS1-R (from 1.719 ꢃ to 1.754 ꢃ in the
S structures and from 1.780 ꢃ to 1.920 ꢃ in the R struc-
tures). This indicates that in TS3-R it is more convenient for
the system to weaken the hydrogen bond strength in order
to maximize the above-mentioned p-stacking interactions.
In contrast, the p-stacking interactions between the benzyl–
CF3 group and the terminal ester moiety are not significant-
ly affected: the distance between C4 and the centroid of
benzyl–CF3 is 4.332 ꢃ in TS1-S and 4.278 ꢃ in TS3-S.
To validate this theoretical model, the compound (E)-
ethyl-4-(N-methyl-1H-indole-2-carboxamido)but-2-enoate
3h was synthesized and subjected to IMAMR under optimal
conditions. However, the poor solubility of 3h prevented
the catalytic process from taking place (Scheme 3). There-
fore, acyclic precursors bearing benzyl groups with electron-
donating and electron-withdrawing substituents (e.g. 3i,j)
were readily prepared and allowed to cyclize in the presence
of 4a (10 mol%) in toluene at ꢀ458C. The corresponding
pyrazinoindolones 1i,j were isolated with similar chemical
and optical yields (ee=84–90%). These results, together
with the case reported above for 1g (R=PMP, Scheme 2,
ee=82%) supports the previously described computational
outcomes underlying marginal importance of the amide sub-
stituent (Scheme 3) on the overall chemical course.
In contrast, the amide moiety in 3 exerts a significant in-
fluence on the overall catalytic reaction. This has been theo-
retically and experimentally verified by taking into account
the indolyl-ester 3k (Scheme 4), which was directly obtained
In addition to the effect of the weaker O1···H1 hydrogen
bond described previously, it is worth outlining the less pro-
nounced propensity of the ester moiety (with respect to its
amidic counterpart) to fold the acyclic precursor 3k to reach
the optimal geometry for the ring-closing process. This
factor certainly contributes to the lower reactivity of 3k
with respect to 3i and 3j. In this case, an appreciable isolat-
ed yield (75%) (in addition to the moderate ee (59%) pre-
viously mentioned) was obtained using a stoichiometric
amount of tBuOK as a base in toluene at room temperature
(Scheme 4). Attempts to improve the level of stereodiscrimi-
nation by lowering the reaction temperature failed, causing
an unacceptable drop in the reaction rate with no significant
variations on the final optical yield.
Additional computations have been performed on modi-
fied model systems in which we have replaced the methyl
group on the ester chain with a tert-butyl group character-
ized by greater steric hindrance. The presence of the more
cumbersome ester substituent is responsible for a steric per-
turbation that destabilizes the S structures more, both in the
two starting adducts M5-S and M5-R (see Figure S6 in the
Supporting Information) and in the two transition structures.
As a matter of fact, the energy differences between M5-S
and M5-R (0.79 kcalmolꢀ1) and between the two related
transition states TS4-S and TS4-R (1.67 kcalmolꢀ1) decrease.
The bulky tert-butyl group introduces a more evident steric
perturbation with the benzyl–CF3 ring in the S structures,
and also causes a weakening of the parallel p-stacking inter-
action between the catalyst benzyl–CF3 group and the sub-
strate ester p system, which is present in the S structures
Scheme 4. Asymmetric phase-transfer-catalyzed aza-Michael addition of
indolyl-ester 3k.
in unoptimized 25% yield through condensation of indole-
2-carboxylic acid and ethyl 4-bromocrotonate under basic
conditions (see Supporting Information). In this case, the re-
placement of the nitrogen N1 with the oxygen O5 is expect-
ed to cause a decrease of the electron density on the sub-
strate carbonyl oxygen O1. To validate the effect of this sub-
strate “mutation” we have optimized the structures of the
new adducts M4-S and M4-R obtained by replacing the sub-
strate N1-benzyl unit with an oxygen atom. The energy dif-
ference between the two complexes M4-S and M4-R (see
Figure S5 of the Supporting Information) is 1.77 kcalmolꢀ1
with the S isomer favored. It is interesting to notice the sig-
nificant lengthening of the O1···H1 distance (about 0.15 ꢃ)
with respect to M3-S and M3-R (from 1.666 to 1.704 ꢃ and
from 1.796 to 1.852 ꢃ, respectively). This finding clearly
demonstrates that the O1···H1 hydrogen bond is significant-
ly stronger when an amidic carbonyl is involved. The de-
crease of the energetic contribution of the O1···H1 interac-
tion is also in agreement with the stabilization energy com-
puted for M4-S with respect to the isolated catalyst and sub-
strate, which is about 2 kcalmolꢀ1 lower than the value com-
Chem. Eur. J. 2010, 16, 12462 – 12473
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12469

M. Bandini, G. P. Miscione et al.
Table 1. Asymmetric PTC aza-Michael addition of pyrrolyl ester 3l and
3m.[a]
only. The two intrinsic activation barriers are rather similar
(7.62 and 8.32 kcalmolꢀ1 for the S and R path, respectively),
and are in both cases larger than in the reference model
system with the ester methyl group. These computational
data are in rather good agreement with the experimentally
observed decrease of the reaction yield and of the enantio-
meric excess (yield: 90%!61%, ee: 90%!53%). Besides
the importance of the O1···H1 hydrogen bond, the trend of
the previously described p-stacking interactions between the
indolyl ring and the benzyl substituent of the quinuclidine
nitrogen atom seems to match the reported experimental
data satisfactorily. In particular, it has been found that:
1) EWGs favor the discriminating T-shaped p interactions
(e.g., p-F, p-NO2 and p-benzyl–CF3 substituents lead to com-
parable results); and 2) a para substitution pattern of the
benzyl unit is strictly required, as the introduction of groups
in the meta and/or ortho positions (e.g., F, NO2, 9-anthryl)
causes an important decrease in the enantiomeric excess.
In principle, the T-shaped interactions should become
much less important if we reduce the extension of the inter-
acting p-electron clouds. This would be the case of pyrrole-
analogous 3l and 3m, which, after a survey of chiral alka-
loid-derived quaternary ammonium salts 4b–d (Scheme 5),
Entry
3
4
Conditions[b]
Yield [%][c]
ee [%][d]
1
2
3
4
5
6
7
8
3l
3l
3l
3l
3l
3l
3l
3m
4a
4b
4b
4c
4c
4d
4e
4a
A
A
B
B
A
B
82
–
54
–
88
79
75
85
80[e]
77
18
30
29
13
31
56
A
A
[a] All the reactions were carried out in reagent grade toluene without
any moisture restriction. [b] A: KOHaq (25%, 0.5 equiv), ꢀ458C, 16 h; B:
KOHaq (25%, 0.5 equiv), RT, 16 h. [c] Yields after flash chromatography.
[d] Determined by HPLC on a chiral stationary phase. [e] After 48 h.
In particular, M6-S is only 0.14 kcalmolꢀ1 more stable than
M6-R, and the two corresponding transition states also are
very close in energy, TS5-R being only 0.35 kcalmolꢀ1 more
stable than TS5-S. It is therefore evident that the interac-
tions involving the phenyl moiety of the indole ring and the
benzyl–CF3 group are crucial for determining the stereo-
chemical outcome of the reaction. Even in this case, the
computational evidence is in good agreement with the ex-
perimental outcome, showing only moderate stereoselection
(ee=54–56%), and indicates that the indole system is essen-
tial in determining the enantioselective character of the pro-
cess.
led to the synthetically useful 1,2,3,4-tetrahydropyrroloACTHNUTRGNEU[GN 1,2-
a]pyrazinones 1l and 1m in high yield but lower enantio-
meric excess up to 56% (Table 1).
Conclusion
In this paper a full account addressing the stereoselective
N1-alkylation of indoles through intramolecular phase-trans-
fer catalysis has been presented. In particular, combined
computational (DFT) and experimental investigations have
been carried out to understand how chiral Cinchona quater-
nary ammonium salts exert perfect stereodiscrimination in
the ring-closing event. The computational results have dem-
onstrated that a complex network of single- and multi-point
nonbonded catalyst/substrate interactions is responsible for
the observed stereoselection.
We have considered as a reference model system the ad-
ducts formed by the substrate 3c and the catalyst ammoni-
um salt 4a. Because the reaction is carried out under kinetic
conditions (at a temperature of ꢀ458C) and the two starting
adducts M1-S and M1-R can interconvert easily, the stereo-
selective character of the reaction and the ee are determined
by the energy difference (2.09 kcalmolꢀ1) between the two
diastereomeric transition states TS1-R and TS1-S leading to
the two products (R)-1c and (S)-1c, respectively.
Scheme 5. Set of chiral ammonium quaternary salts 4b–e used in the
PTC intramolecular N-alkylation of pyrrolyl derivatives.
These points have been validated by further computa-
tions, showing that in both adducts M6-S and M6-R (emulat-
ing the pyrrole-analogous 3l and 3m) no interactions be-
tween the pyrrolyl group and the benzyl–CF3 fragment are
evident (see Figure 7). The p systems of the two groups are
in this case too far away to give rise to an effective interac-
tion. On the other hand, to force them to stay closer would
cause a weakening of the strong H1···O1 hydrogen bond,
which, on the contrary, becomes stronger in M6-S (1.643 ꢃ)
and M6-R (1.685 ꢃ) with respect to M1-S (1.682 ꢃ) and
M1-R (1.719 ꢃ), respectively.
Our computations show a significant decrease of the
energy difference between the two diastereomeric pathways.
12470
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12462 – 12473

Intramolecular N-Alkylation of Indoles and Pyrroles
FULL PAPER
Figure 7. A schematic representation of M6-S, M6-R, and the two corresponding transition states, TS4-S and TS4-R. Bond lengths are in ꢃ. When not
connecting atoms, dashed lines connect the ring centroids.
Experimental Section
The key structural features that are responsible for the
General methods: 1H NMR spectra were recorded on Varian 200
(200 MHz) and Varian 300 (300 MHz) spectrometers. Chemical shifts are
reported in ppm from TMS with the solvent resonance as the internal
standard (deuterochloroform: d=7.27 ppm). Data are reported as fol-
lows: chemical shift, multiplicity (s=singlet, d=duplet, pd=pseudo
duplet, t=triplet, q=quartet, br=broad, brs=broad singlet, m=multip-
let), coupling constants (Hz). 13C NMR spectra were recorded on Varian
200 (50 MHz) or Varian 300 (75 MHz) spectrometers with complete
proton decoupling. Chemical shifts are reported in ppm from TMS with
the solvent as the internal standard (deuterochloroform: d=77.0 ppm).
GC–MS spectra were taken by EI ionization at 70 eV on a Hewlett–Pack-
ard 5971 with GC injection. LC-electrospray ionization mass spectra
were obtained using an Agilent Technologies MSD1100 single-quadru-
pole mass spectrometer. Chromatographic purification was performed
with 240–400 mesh silica gel. Precoated preparative TLC plates were pur-
chased from FLUKA (silica gel 60 F254). Elemental analyses were car-
ried out by using an EACE 1110 CHNOS analyzer. Analytical high per-
formance liquid chromatography (HPLC) was performed on a liquid
chromatograph equipped with a variable wavelength UV detector (deute-
rium lamp 190–600 nm), using a Daicel ChiracelTM OD or AD column
(0.46 cm I.D.ꢄ25 cm) (Daicel Inc. HPLC grade isopropanol and hexane
were used as the eluting solvents). Separation through preparative HPLC
was performed on an Agilent Technologies MSD 1100 liquid chromato-
graph using a Zorbax Eclipse SDB-C18 column (21.2ꢄ150 mm) with ace-
tonitrile and milliQ-H2O as eluents. Optical rotations were determined in
a 1 mL cell with a path length of 10 mm (NaD line). Melting points were
determined with Bibby Stuart Scientific Melting Point Apparatus SMP3
and are not corrected. Compound 3l was obtained as previously de-
scribed.[15] Because the catalyst–substrate adducts examined here involve
different aromatic groups, it is reasonable to believe that the interactions
involving their p systems should play a key role in the reaction. Conse-
quently, it is mandatory to choose an appropriate method for their de-
TS1-R/TS1-S energy difference (favoring TS1-S) have been
identified. These are: 1) the hydrogen bond O1···H1 be-
tween substrate and catalyst; 2) the conformation of the
substrate amino-propene unit; and 3) a complex interplay of
the p-stacking interactions involving different groups of sub-
strate and catalyst. In particular, we have found that the p
interaction between the indole and the benzyl–CF3 groups is
crucial.
The key-role played by this interaction has been demon-
strated clearly by the computations carried out on the
model systems emulating the pyrrole-analogous 3l and 3m.
In these cases the lack of interactions between the indole
(replaced by the pyrrolyl group) and the benzyl–CF3 frag-
ment determines a significant decrease in the energy differ-
ence between the two diastereomeric pathways.
The mechanistic picture based on the above-mentioned
nonbonded interaction has been fully corroborated by con-
trol experiments on tailored acyclic precursors. These results
also nicely account for the moderate stereochemical levels
obtained with pyrrolyl derivatives 3l and 3m (ee up to
56%) for which the crucial p–p interaction between the
larger indolyl core and the CF3-substitued benzyl group of
the catalyst is missing. Efforts toward the exploitation of
these findings in the rational design of new chiral PT-cata-
lysts for intramolecular aza-Michael of pyrrolyl derivatives
are in progress.
Chem. Eur. J. 2010, 16, 12462 – 12473
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12471

M. Bandini, G. P. Miscione et al.
scription. Unfortunately, this class of interactions cannot be correctly
treated by the most popular DFT functionals (for instance, B3LYP),
which are inaccurate for interactions where medium-range correlation ef-
fects are dominant.[21] Even if p–p interactions can be satisfactorily de-
scribed by the MP2 method, this approach is unfeasible for the large
model system used here, because it would require too much CPU time
for practical and extensive usage. However, a new hybrid functional (de-
noted as MPWB1K) has recently been proposed by Trulhar.[22] This func-
tional has been demonstrated to be capable of treating medium-range
correlation effects and to provide a good estimate of the p–p interactions
and reaction energetics[23,24] using reasonable amounts of CPU time.
Thus, all DFT computations reported in the present paper were carried
out with the Gaussian 03 series of programs[25] using the MPWB1K[22]
functional and the DZVP basis set.[26] The DZVP basis is a local spin
density (LSD)-optimized basis set of double-zeta quality, which includes
polarization functions and is suitable for describing weak hydrogen
bonds and p interactions such as those occurring in the system investigat-
ed here. The transition vector of the various transition states was ana-
lyzed by means of frequency computations.
3J
1H), 4.13–4.26 (m, 2H), 4.81 (s, 2H), 5.06 (br, 1H), 7.21–7.31 (m, 2H),
7.46 (s, 2H), 7.51 (s, 1H), 7.80 ppm (d, 3J(H,H)=8.1 Hz, 1H); 13C NMR
(50 MHz, [D]CHCl3): d=14.1, 35.8, 47.3, 61.5, 69.7, 110.1, 111.1, 121.8,
122.8, 123.5, 126.5, 127.2, 135.8, 159.3, 170.3 ppm; ESI-MS: m/z (%): 274
[M+1]+.
(H,H)=12.0 Hz, 1H), 3.02 (dd, 3J(H,H)=16.8 Hz, 3J
ACHTUNGTRENUNGN ACHTUNRTGENNGNU ACHTNUGTRENUN(GN H,H)=9.3 Hz,
AHCTUNGTRENNUNG
Typical procedure for the PTC intramolecular aza-Michael addition (1l
and 1m): Method A: A sample vial was charged with reagent grade tolu-
ene (6 mL), the indolyl-ester 3l or 3m (50 mmol) and catalyst 4a (3 mg,
5 mmol). An aqueous solution of KOH (25%, 6 mL, 0.5 equiv) was added
to the mixture by syringe and immediately cooled to ꢀ458C. The reac-
tion was stirred at the same temperature for 16 h, then the solvent was
evaporated under reduced pressure. The crude product was purified di-
rectly through a pad of silica. Method B: same conditions but at RT.
Data for 1l:[15] Yield: 82%; flash chromatography (c-Hex/AcOEt=8:2);
ee=54%; [a]D =ꢀ20.1 (c=1.3 in chloroform); HPLC analysis: OD
column (210 nm), RT, method: n-Hex/IPA=90:10, flow 1.0 mLminꢀ1
,
t(major) =21.3 min, t(minor) =25.6 min.
Data for 1m:[15] Yield: 77%; flash chromatography (c-Hex/AcOEt=
85:15); ee=56%; [a]D =+4.6 (c=1.5 in chloroform); HPLC analysis:
OD column (214 nm), RT, method: n-Hex/IPA=90:10, flow
1.0 mLminꢀ1, t(minor) =14.9 min, t(major) =19.4 min.
Typical procedure for the PTC intramolecular aza-Michael addition (1i
and 1j): A sample vial was charged with reagent grade toluene (6 mL),
the indolyl-ester 3 (50 mmol) and catalyst 4a (3 mg, 5 mmol). An aqueous
solution of KOH (25%, 6 mL, 0.5 equiv) was added to this mixture by sy-
ringe, and immediately cooled to ꢀ458C. The reaction was stirred at the
same temperature for 16 h, then the solvent was evaporated under re-
duced pressure. The crude product was purified directly through a pad of
silica.
Acknowledgements
(S)-Ethyl-2-[2-(4-methoxybenzyl)-1-oxo-1,2,3,4-tetrahydropyrazinoACTHNUTRGNE[UNG 1,2-
a]indol-4-yl]acetate (1i): Yellow solid; flash chromatography (c-Hex/
AcOEt=8:2); m.p. 78.5–80.88C; yield: 97%; ee=90%; [a]D =+4.6 (c=
0.9 in chloroform); HPLC analysis: AD column (214 nm), 408C, method:
Acknowledgement is made to the MIUR (Rome), FIRB Project (Proget-
tazione, preparazione e valutazione biologica e farmacologica di nuove
molecole organiche quali potenziali farmaci innovativi), Universitꢀ di Bo-
logna and Fondazione del Monte di Bologna e Ravenna.
n-Hex/IPA=70:30, flow 1.0 mLminꢀ1
,
t
(S) =15.9 min,
G
t
(R) =23.0 min;
1H NMR (300 MHz, [D]CHCl3): d=1.19 (t, 3J
3
2.53 (m, 2H), 3.59 (d, J
G
3H), 4.12 (d, 3J
(H,H)=14.1 Hz, 1H), 6.87 (d, 3J
7.29–7.32 (m, 5H), 7.71 ppm (d, 3J
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
[1] a) D. J. Faulkner, Nat. Prod. Rep. 2002, 19, 1–48; b) A. Kleeman, J.
Engel, B. Kutscher, D. Reichert, Pharmaceutical Substances, 4th ed.,
Thieme, New York, 2001; c) J. Sundberg, Pyrroles and Their Benzo-
derivatives: Synthesis and Applications in Comprehensive Heterocy-
clic Chemistry, Vol. 4 (Eds.: A. R. Katritzky, C. W. Rees), Pergamon,
Oxford, 1984, pp. 313–376; d) A. Joule, Indole and its Derivatives in
Science of Synthesis: Houben-Weyl Methods of Molecular Transfor-
mations, Vol. 10 (Ed.: E. J. Thomas), Thieme, Stuttgart, 2000; Chap-
ter 10.13.
A
ACHTUNGTRENNUNG
(75 MHz, [D]CHCl3): d=13.9, 35.8, 47.2, 48.1, 48.7, 55.2, 61.0, 107.0,
109.5, 114.2, 120.9, 122.9, 124.7, 127.7, 128.3, 128.6, 130.1, 135.3, 159.3,
159.6, 170.4 ppm; ESI-MS: m/z (%): 393 [M+1]+.
(S)-Ethyl-2-[2-(4-fluorobenzyl)-1-oxo-1,2,3,4-tetrahydropyrazinoACTHNUTRGNE[UNG 1,2-
a]indol-4-yl]acetate (1j): Pale white solid; flash chromatography (c-Hex/
AcOEt=8:2); m.p. 88.3–91.78C; yield: 85%; ee=84%; [a]D =+2.4 (c=
1.7 in chloroform); HPLC analysis: AD column (214 nm), 408C, method:
[2] a) G. W. Gribble, J. Chem. Soc. Perkin Trans. 1 2000, 1045–1075;
b) R. Dalpozzo, G. Bartoli, Curr. Org. Synth. 2005, 2, 163–178; c) S.
Cacchi, G. Fabrizi, Chem. Rev. 2005, 105, 2873–2920; d) G. W. Grib-
ble, M. G. Saumier, E. T. Pelkey, T. L. S. Kishbaugh, Y. Liu, J. Jiang,
H. A. Trujillo, D. J. Keavy, D. A. Davis, S. C. Conway, F. L. Switzer,
S. Roy, R. A. Silva, J. A. Obaza-Nutatis, M. P. Sibi, N. V. Moskalev,
T. C. Barden, L. Chang, W. M. Habeski, B. Pelcman, W. R. Spon-
holtz III, R. W. Chau, B. D. Allison, S. D. Garaas, M. S. Sinha, M. A.
McGowan, M. R. Reese, K. S. Harpp, Curr. Org. Chem. 2005, 9,
1493–1519; e) S. Agarwal, S. Cꢅmmerer, S. Filali, W. Froçhner, J.
Knçll, M. P. Krahl, K. R. Reddy, H.-J. Knçlker, Curr. Org. Chem.
2005, 9, 1601–1614; f) M. Bandini, A. Melloni, S. Tommasi, A.
Umani-Ronchi, Synlett 2005, 1199–1222; g) G. R. Humphrey, J. T.
Kuethe, Chem. Rev. 2006, 106, 2875–2911; h) J. Campo, M. Garcꢆa-
Valverde, S. Marcaccini, M. J. Rojo, T. Torroba, Org. Biomol. Chem.
2006, 4, 757–765.
[3] a) K. A. Jørgensen, Synthesis 2003, 1117–1125; b) M. Bandini, A.
Melloni, A. Umani-Ronchi, Angew. Chem. 2004, 116, 560–566;
Angew. Chem. Int. Ed. 2004, 43, 550–556; c) M. Bandini, E. Emer,
S. Tommasi, A. Umani-Ronchi, Eur. J. Org. Chem. 2006, 3527–
3544; d) Y.-F. Sheng, A.-J. Zhang, X.-J. Zheng, S.-L. You, Chin. Org.
Chem. 2008, 28, 605–616; e) B. Poulsen, K. A. Jørgensen, Chem.
Rev. 2008, 108, 2903–2915; f) S.-L. You, C. Quan, Z. Mi, Chem. Soc.
Rev. 2009, 38, 2190–2201; g) M. Rueping, B. J. Nachtsheim, Beilstein
J. Org. Chem. 2010, 6, DOI: 10.3762/bjoc.6.6.
n-Hex/IPA=70:30, flow 1.0 mLminꢀ1
1H NMR (200 MHz, [D]CHCl3): d=1.14 (t, 3J
(d, 3J(H,H)=7.4 Hz, 2H), 3.59 (d, 3J
(H,H)=13.0 Hz, 1H), 3.90–4.07 (m,
3H), 4.87 (d, 3J(H,H)=14.6 Hz, 1H), 4.34 (d, 3J
(H,H)=14.6, 1H), 4.87–
4.96 (m, 1H), 5.17 (d, 3J(H,H)=14.6 Hz, 1H), 7.01–7.09 (t, 3J
(H,H)=
8.0 Hz, 2H), 7.17 (dt, 3J(H,H)=8.0, 3J
(H,H)=1.8 Hz, 1H), 7.32–7.39 (m,
5H), 7.70 ppm (d, 3J(H,H)=8.0 Hz, 1H); 13C NMR (75 MHz, [D]CHCl3)
d=14.2, 36.0, 47.5, 48.5, 49.0, 61.3, 107.5, 109.8, 116.0 (d, J=21.8 Hz),
121.3, 123.2, 125.1, 128.0, 128.3, 130.8 (d, J=7.8 Hz), 134.2 (d, J=
206.3 Hz), 159.9, 162.7 (d, J=244.5 Hz), 170.6 ppm; ESI-MS: m/z (%):
381 [M+1]+.
,
t
(S) =17.3 min,
t(R) =28.8 min;
AHCTUNGTRENNUNG
U
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
R
ACHTUNGTRENNUNG
G
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
Typical procedure for the PTC intramolecular aza-Michael addition (1k):
Indolyl ester 3k (20 mg, 0.07 mmol) was dissolved in toluene (4 mL) in a
10 mL round-bottomed flask equipped with a magnetic stirrer. After the
addition of the catalyst 3a (4 mg, 10 mol%), solid tBuOK (8.2 mg,
0.07 mmol) was added. After 1 h stirring at room temperature the start-
ing material was consumed (judged by TLC), and the reaction mixture
brought to dryness under vacuum and purified by flash chromatography.
(ꢀ)-Ethyl-2-(1-oxo-3,4-dihydro-1H-[1,4]oxazino
ACHTUNGTNER[NUNG 4,3-a]indol-4-yl)acetate
(1k): Pale yellow viscous oil; flash chromatography (c-Hex/AcOEt=
8:2); yield: 75%; ee=59%; [a]D =ꢀ12.5 (c=0.4 in chloroform); HPLC
analysis: AD column (214 nm), RT, method: n-Hex/IPA=80:20, flow
1.0 mLminꢀ1
,
t
(minor) =10.9 min,
t
(major) =13.3 min; 1H NMR (300 MHz,
(H,H)=7.2 Hz, 3H), 2.81 (dd, 3J
(H,H)=4.8 Hz,
3
[D]CHCl3): d=1.27 (t, J
A
ACHTUNGTRENNUNG
12472
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12462 – 12473

Intramolecular N-Alkylation of Indoles and Pyrroles
FULL PAPER
[4] Catalytic Asymmetric Friedel–Crafts Alkylations (Eds.: M. Bandini,
A. Umani-Ronchi), Wiley-VCH, Weinheim, 2009.
VCH, Weinheim, 2005; e) T. Ooi, K. Maruoka, Aldrichimica Acta
2007, 40, 77–86; f) J. L. Vicario, D. Badia, L. Carrillo, Synthesis
2007, 2065–2092, and references therein; g) H. Pellissier, Tetrahe-
dron 2007, 63, 9267–9331; h) K. Maruoka, Asymmetric Phase Trans-
fer Catalysis, Wiley-VCH, Weinheim, 2008; i) K. Maruoka, Org. Pro-
cess Res. Dev. 2008, 12, 679–697; <lit l>K. Maruoka, Synlett 2009,
652–652.
[5] a) D. A. Evans, K. R. Fandrick, Org. Lett. 2006, 8, 2249–2252;
b) D. A. Evans, K. R. Fandrick, H.-J. Song, K. A. Scheidt, R. Xu, J.
Am. Chem. Soc. 2007, 129, 10029–10041; c) G. Blay, I. Fernꢇndez,
J. R. Pedro, C. Vila, Tetrahedron Lett. 2007, 48, 6731–6734; d) G.
Blay, I. Fernꢇndez, J. R. Pedro, C. Vila, Org. Lett. 2007, 9, 2601–
2604; e) Q. Kang, X.-J. Zheng, S.-L. You, Chem. Eur. J. 2008, 14,
3539–3542; f) M. Zeng, Q. Kang, Q. Kang, Q.-L. He, S.-L. You,
Adv. Synth. Catal. 2008, 350, 2169–2173; g) Y.-F. Sheng, G.-Q. Li, Q.
Kang, A.-J. Zhang, S.-L. You, Chem. Eur. J. 2009, 15, 3351–3354.
[6] a) J. J. Vepsꢅlꢅinen, S. Auriola, M. Tukiainen, N. Ropponen,
Callaway, J. Planta Med. 2005, 71, 1049–1057; b) A. L. Williams,
J. M. Srinivasan, J. N. Johnston, Org. Lett. 2006, 8, 6047–6049; c) S.
Grꢈschow, L. C. Chang, Y. Mao, D. H. Sherman, J. Am. Chem. Soc.
2007, 129, 6470–6476, and references therein; d) M. B. Johansen,
M. A. Kerr, Org. Lett. 2008, 10, 3497–3500; e) S. Chamberland, S.
Grueschow, D. H. Sherman, R. M. Williams, Org. Lett. 2009, 11,
791–794.
[7] a) D. Moffat, C.J. Nichols, D. A. Riley, N.S. Simpkins, Org. Biomol.
Chem. 2005, 3, 2953–2975; b) E. Wincent, H. Shirani, J. Bergman,
U. Rannug, T. Janosik, Bioorg. Med. Chem. 2009, 17, 1648–1653.
[8] a) B. M. Trost, M. J. Krische, V. Berl, E. M. Grenzer, Org. Lett. 2002,
4, 2005–2008; b) For an analogous approach on pyrrole derivatives,
see: B. M. Trost, G. Dong, Chem. Eur. J. 2009, 15, 6910–6919;
c) L. M. Stanley, J. F. Hartwig, Angew. Chem. 2009, 121, 7981–7984;
Angew. Chem. Int. Ed. 2009, 48, 7841–7844.
[17] For representative examples see: a) K. Tajasu, S. Maiti, M. Ihara,
Heterocycles 2003, 59, 51–55; b) S. Fustero, D. Jimꢁnez, J. Moscar-
dꢉ, S. Catalꢇn, C. del Pozo, Org. Lett. 2007, 9, 5283–5286; c) S. K.
Bur, A. Pawda, Adv. Heterocycl. Chem. 2007, 94, 1–105; d) E. C.
Carlson, L. K. Rathbone, H. Yang, N. D. Collet, R. G. Carter, J. Org.
Chem. 2008, 73, 5155–5158; e) S. Fustero, J. Moscardꢉ, D. Jimꢁnez,
M. D. Pꢁrez-Carrion, M. Sanchez-Rosellꢉ, C. del Pozo, Chem. Eur.
J. 2008, 14, 9868–9872; f) Asymmetric Synthesis of Nitrogen Hetero-
cycles, (Ed.: J. Royer), Wiley-VCH, Weinheim, 2009, and references
therein.
[18] For representative examples: a) U.-H. Dolling, P. Davis, E. J. J. Gra-
bowski, J. Am. Chem. Soc. 1984, 106, 446–447; b) D. L. Hughes, U.-
H. Dolling, K. M. Ryan, E. F. Schoenewaldt, E. J. J. Grabowski, J.
Org. Chem. 1987, 52, 4745–4752; c) E. J. Corey, F. Xu, M. C. Noe, J.
Am. Chem. Soc. 1997, 119, 12414–12415; d) E. J. Corey, Y. Bo, J.
Busch-Petersen, J. Am. Chem. Soc. 1998, 120, 13000–13001; e) T.
Ooi, M. Kameda, M. Taniguchi, K. Maruoka, J. Am. Chem. Soc.
2004, 126, 9685–9694; f) A. Berkessel, M. Guixꢀ, F. Schmidt, J. M.
Neudçrfl, J. Lex, Chem. Eur. J. 2007, 13, 4483–4498.
[19] F. Bernardi, A. Bottoni, G. Tonachini, J. Chem. Soc. Perkin Trans. 2
1980, 467–470.
[9] a) H.-L. Cui, X. Feng, J. Peng, J. Lei, K. Jiang, Y.-C. Chen, Angew.
Chem. 2009, 121, 5847–5850; Angew. Chem. Int. Ed. 2009, 48, 5737–
5740; b) D. Enders, C. Wang, G. Raabe, Synthesis 2009, 4119–4124.
[10] a) M. Bandini, A. Eichholzer, Angew. Chem. 2009, 121, 9786–9824;
Angew. Chem. Int. Ed. 2009, 48, 9608–9644, and references therein;
b) M. Bandini, A. Eichholzer, Angew. Chem. 2009, 121, 9697–9701;
Angew. Chem. Int. Ed. 2009, 48, 9533–9537.
[11] M. Bandini, A. Eichholzer, M. Tragni, A. Umani-Ronchi, Angew.
Chem. 2008, 120, 3282–3285; Angew. Chem. Int. Ed. 2008, 47, 3238–
3241.
[20] Obtained with an analogous 4a ammonium salt in which the CF3
group at the benzylic pendant was replaced by a NO2 group.
[21] Y. Zhao, D. G. Truhlar, Acc. Chem. Res. 2008, 41, 157–167.
[22] Y. Zhao, D. G. Truhlar, J. Phys. Chem. A 2004, 108, 6908–6918.
[23] Y. Zhao, D. G. Truhlar, J. Phys. Chem. A 2005, 109, 5656–5667.
[24] Y. Zhao, O. Tishchenko, D. G. Truhlar, J. Phys. Chem. B 2005, 109,
19046–19051.
[25] Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
[12] a) F. Cafieri, E. Fattorusso, O. Tagliatela-Scafati, J. Nat. Prod. 1998,
61, 122–125, and references therein; b) K. G. Poullennec, D. Romo,
J. Am. Chem. Soc. 2003, 125, 6344–6345; c) N. Nettekoven, J.-M.
Plancher, H. Richter, O. Roche, S. Taylor, PCT Int. Appl. WO US
2007/0135416A1, 2007.
[13] a) J. M. Bentley, P. Hebeisen, M. Muller, H. Richter, S. Roever, P.
Mattei, S. Taylor, PCT Int. Appl. WO 2001-EP8520 20010724, 2002;
b) R. K. Tiwari, A. K. Verma, A. K. Chhillar, D. Singh, J. Singh,
V. K. Sankar, V. Yadav, G. L. Sharma, R. Chandra, Bioorg. Med.
Chem. 2006, 14, 2747–2752.
[14] a) J. Patel, N. Pelloux-LIon, F. Minassian, Y. VallIe, J. Org. Chem.
2005, 70, 9081–9084; b) J. Patel, N. Pelloux-LIon, F. Minassian, Y.
VallIe, Tetrahedron Lett. 2006, 47, 5561–5563.
[15] M. Bandini, A. Eichholzer, M. Monari, F. Piccinelli, A. Umani-
Ronchi, Eur. J. Org. Chem. 2007, 2917–2920.
ACHTUNGTRENeNUNG ry, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
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, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
[16] For general reviews see: a) A. Nelson, Angew. Chem. 1999, 111,
1685–1687; Angew. Chem. Int. Ed. 1999, 38, 1583–1585; b) B. Lygo,
B. I. Andrews, Acc. Chem. Res. 2004, 37, 518–525; c) L.-W. Xu, C.-
G. Xia, Eur. J. Org. Chem. 2005, 633–639, and references therein;
d) A. Berkessel, H. Grçger in Asymmetric Organocatalysis, Wiley-
[26] N. Godbout, D. R. Salahub, J. Andzelm, E. Wimmer, Can. J. Chem.
1992, 70, 560–571.
Received: March 3, 2010
Published online: September 13, 2010
Chem. Eur. J. 2010, 16, 12462 – 12473
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12473