Expanding the palette of phenanthridinium cations

Article abstract of DOI:10.1002/chem.201304241

5,6-Disubstituted phenanthridinium cations have a range of redox, fluorescence and biological properties. Some properties rely on phenanthridiniums intercalating into DNA, but the use of these cations as exomarkers for the reactive oxygen species (ROS), superoxide, and as inhibitors of acetylcholine esterase (AChE) do not require intercalation. A versatile modular synthesis of 5,6-disubstituted phenanthridiniums that introduces diversity by Suzuki-Miyaura coupling, imine formation and microwave-assisted cyclisation is presented. Computational modelling at the density functional theory (DFT) level reveals that the novel displacement of the aryl halide by an acyclic N-alkylimine proceeds by an S_NAr mechanism rather than electrocyclisation. It is found that the displacement of halide is concerted and there is no stable Meisenheimer intermediate, provided the calculations consistently use a polarisable solvent model and a diffuse basis set.

Full text of DOI:10.1002/chem.201304241

DOI: 10.1002/chem.201304241
Full Paper
&
Phenanthridinium Cations
Expanding the Palette of Phenanthridinium Cations
Andrew G. Cairns,[a] Hans Martin Senn,*[a] Michael P. Murphy,[b] and Richard C. Hartley*[a]
Abstract: 5,6-Disubstituted phenanthridinium cations have
a range of redox, fluorescence and biological properties.
Some properties rely on phenanthridiniums intercalating
into DNA, but the use of these cations as exomarkers for the
reactive oxygen species (ROS), superoxide, and as inhibitors
of acetylcholine esterase (AChE) do not require intercalation.
A versatile modular synthesis of 5,6-disubstituted phenan-
thridiniums that introduces diversity by Suzuki–Miyaura cou-
pling, imine formation and microwave-assisted cyclisation is
presented. Computational modelling at the density function-
al theory (DFT) level reveals that the novel displacement of
the aryl halide by an acyclic N-alkylimine proceeds by an
SNAr mechanism rather than electrocyclisation. It is found
that the displacement of halide is concerted and there is no
stable Meisenheimer intermediate, provided the calculations
consistently use a polarisable solvent model and a diffuse
basis set.
Introduction
5,6-Disubstituted phenanthridinium salts have found wide utili-
ty as fluorescent dyes for DNA,[1] as sensors[2] and as potential
therapeutics,[3] including for the treatment of trypanosomal in-
fections.[4] The planar phenanthridinium cation can intercalate
into duplex nucleic acids,[1,5,6] particularly if it has 3- and/or 8-
amino groups.[7] The fluorescence of the 3,8-diamino-6-aryl de-
rivatives, such as ethidium bromide (1) and propidium iodide
(2), is hugely enhanced upon intercalation, and both the
amino groups and the 6-aryl substituent contribute significant-
ly to this effect (Figure 1).[8] As a result, in spite of its toxicity,
1 is widely used in molecular biology laboratories to detect
DNA on agarose gels following electrophoresis,[9] and its dis-
placement from DNA is often used to assess the binding of
novel ligands to DNA.[10] Propidium iodide is also a valuable
DNA stain employed in flow cytometry and histochemistry.[11]
More efficient visualisation of RNA can be achieved by linking
a second fluorophore to the 5,6-disubstituted phenanthridini-
um.[12] Incorporation of phenanthridinium cations into DNA has
enabled the study of its redox properties and electron-transfer
processes within its double helix.[13]
Figure 1. Important phenanthridinium cations 1–3 (counterion unimportant)
and phenanthridines 4–6 (substituent numbering and our designation of
rings are shown).
Phenanthridiniums can inhibit enzymes in ways that are not
related to their intercalation into DNA, which would represent
an undesirable off-target effect. The most important example
is the allosteric inhibition of acetylcholine esterase (AChE) by 2
and other 5,6-disubstituted phenanthridiniums.[14,15] This allo-
steric interaction has been exploited to demonstrate the
enzyme-induced assembly of AChE inhibitors using in situ Click
chemistry,[15,16] and as a means of detecting AChE.[17]
5,6-Disubstituted dihydrophenanthridines, which are easily
prepared by reduction of the corresponding phenanthridini-
ums, also have useful properties. They can be oxidised in vivo
to the corresponding phenanthridiniums by reactive oxygen
species (ROS).[2] Such ROS are predominantly produced by re-
duction of oxygen to superoxide in the mitochondria,[18] and
underlie much of the damage in cardiovascular disease, stroke
and neurodegeneration. They are also implicated in the pro-
cess of ageing. The specificity of hydroethidine (4) oxidation
by superoxide to give the hydroxyethidium cation (3) makes
this reaction a relatively robust method for assessment of su-
peroxide production in biological systems,[2,19] and hydropropi-
dine (5) provides a useful cell-impermeant analogue.[20] A mito-
[a] A. G. Cairns, Dr. H. M. Senn, Dr. R. C. Hartley
WestCHEM School of Chemistry
University of Glasgow, Glasgow, G12 8QQ (UK)
E-mail: hans.senn@glasgow.ac.uk
richard.hartley@glasgow.ac.uk
[b] Dr. M. P. Murphy
MRC Mitochondrial Biology Unit
Wellcome Trust/MRC Building, Cambridge, CB2 0XY (UK)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304241.
ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons At-
tribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
Chem. Eur. J. 2014, 20, 3742 – 3751
3742ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
chondria-targeted version, MitoSOX (6), allows specific assess-
ment of this ROS in mitochondria.[21] Unfortunately, the detec-
tion of the hydroxyethidium (3) to infer superoxide concentra-
tions in biological experiments often relies on the enhanced
fluorescence when the cation 3 binds to DNA, and this is com-
plicated by the presence of other ethidium derivatives formed
by different oxidative processes that produce similar emissions.
More accurate quantification requires isolation and HPLC sepa-
ration of the various phenanthridiniums[2] as so-called exo-
markers.[22] Here, the interaction with DNA can be problematic
because it adversely affects extraction of the phenanthridini-
ums from the biological medium and so makes quantification
less reliable. Therefore, structures that allow the independent
modulation of the redox, DNA intercalating and enzyme inter-
action properties of this class of compounds would facilitate
the rational design of biologically useful probe molecules.
Herein, we present a highly versatile route to 5,6-disubstitut-
ed phenanthridiniums so as to allow their properties to be
tuned to a desired function, be it redox reactivity or interaction
with DNA or enzymes. The key step is a novel cyclisation in
which an acyclic N-alkylimine displaces an aryl halide intramo-
lecularly. We use DFT calculations to determine the mechanism
of this reaction, distinguishing between electrocyclisation and
SNAr pathways.
ent, is also generally prepared by N-alkylation. Addition of sta-
bilised carbanions to C-6 can be reversible, allowing intercon-
version of non-fluorescent and fluorescent compounds 8 and
10, respectively,[25,26] a property used by Cronin and co-workers
in their design of pH sensors.[26,27] Nucleophilic additions with
Grignard reagents have worked well in the preparation of simi-
lar dihydrobenzophenanthridines, which are used to make
benzophenanthridinium salts.[28]
There are some promising methods that give direct access
to 5,6-disubstituted dihydrophenanthridines and do not in-
volve the processes shown in Scheme 1. In particular, examples
of CꢀH olefination–cyclisation[29] Dçtz benzannulation,[30] and
a Pictet–Spengler reaction of an N-alkylaniline[31] have been re-
ported. A few benzophenanthridiniums have been prepared
by cyclisation of N-aryl-N-alkyl-acetamides using POCl3,[32] but
the direct synthesis of 5,6-disubstituted phenanthridinium salts
by this method is not know.
We now present a general modular approach to 5,6-disubsti-
tuted phenanthridiniums that does not involve N-alkylation.
We reasoned that the central ring of the phenanthridinium
cation could be assembled in a convergent way from an ortho-
fluoroarylboronate (11), a primary amine (12) and an aromatic
ketone (13) by Suzuki–Miyaura cross-coupling,[33] imine forma-
tion and intramolecular nucleophilic aromatic substitution
(Scheme 2). Aryl fluorides seemed the most suitable precursors
as they react more quickly than other aryl halides in SNAr reac-
tions.[34]
Results and Discussion
N-Alkylphenanthridinium cations (7) and N-alkyl-dihydrophen-
anthridines (8) bearing an alkyl or aryl substituent at C-6 can
be interconverted by simple redox reactions. Two methods
have almost invariably been used to construct these structures
(retrosynthetic Scheme 1). The first is N-alkylation of the phen-
Scheme 2. Retrosynthetic analysis of our convergent approach to 5,6-disub-
stituted phenanthridinium salts 7 beginning from three components 11–13.
Many ortho-fluoroarylboron compounds are commercially
available and they are also typically straightforward to assem-
ble. For example, aryltrifluoroborate 17 was easily prepared
from 4-tert-butyl-bromobenzene (14; Scheme 3).
Nitration of aryl bromide 14 and reduction gave aniline de-
rivative 15. Diazotisation in DCM and solvent exchange gave
a toluene suspension of the aryltetrafluoroborate without iso-
Scheme 1. Retrosynthetic analysis of standard approaches to 5,6-disubstitut-
ed phenanthridinium salts 7 and phenanthridines 8.
anthridine 9, an often inefficient reaction suffering from the
poor solubility of phenanthridines in many organic solvents
and requiring powerful alkylating agents, such as alkyl tri-
flates.[8,13,17] The 6-substituted phenanthridine precursors
needed for N-alkylation have been accessed by a wide range
of methods.[23] The second approach involves nucleophilic
attack on a phenanthridinium salt (10) with a carbanion.[24] The
phenanthridinium precursor 10, which lacks the C-6 substitu-
Scheme 3. Synthesis of an ortho-fluoroborate 17.
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3743ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
lating the potentially explosive intermediate. This underwent
the Balz–Schiemann reaction[35] to give aryl fluoride 16 upon
heating. Bromine–lithium exchange and borylation was fol-
lowed by conversion to the aryltrifluoroborate 17,[36] which is
easy to isolate by crystallisation. Arylboron compounds 17–19
were coupled with iodoarenes 20 to give biaryl compounds 21
using the air- and moisture-stable palladium N-heterocyclic
carbene catalyst PEPPSI-iPr,[37] either under thermal or micro-
wave conditions (Scheme 4 and Figure 2). Microwave condi-
Scheme 4. Suzuki cross-coupling to form ketones 21. See Figure 2 for yields
and conditions.
tions were preferred for the more electron-poor arylboron
compounds because side-products occurred under thermal
conditions. Differences in yields mostly reflected the varying ef-
ficiency of the purifications involved. The dinitro compound
21g was obtained by nitration of biaryl 21e.
Treating 2-arylbenzophenone (21a) with hexylamine and ti-
tanium tetrachloride produced imine 22a, which was isolated
by chromatography in modest yield and was used to investi-
gate the conditions for the proposed cyclisation (Scheme 5).
Interestingly, the CH2N exhibited diastereotopicity indicating
that the biaryl unit is not planar and there is restricted rotation
about the CꢀC bond that links the two rings. Heating to 2008C
in nitrobenzene led to the formation of a 6-phenylphenanthri-
dine, presumably by way of phenanthridinium formation and
fluoride-induced elimination. Transferring the imine 22a to
a short-path distillation (Kugelrohr) apparatus in chloroform,
distilling off the chloroform so that the neat sample could
then be heated to 1008C for 1 h gave complete conversion
from the imine 22a. This resulted in a mixture of compounds
in which the A-ring fluoride had been substituted. Treatment
with TFA then gave the desired phenanthridinium salt 23a in
79% overall yield from the imine 22a. Similar use of acid to
prepare phenanthridinium salts from pseudo-bases (the neutral
adduct of hydroxide attack on the N-alkylimminium group) is
well-known.[38] Although this solventless procedure for the
preparation of N-alkylphenanthridinium salt 23a from imine
22a was high yielding, it led to charring for other imines that
solidified upon removal of the chloroform. A better procedure
was to heat a solution of the imine in THF in the microwave at
1008C for 2 h. The reaction was faster in acetonitrile, but this is
Figure 2. Ketones 21 were prepared from: [a] aryltrifluoroborate 17, [b] the
corresponding aryboronic acids 19, or [c] arylboronate ester 18 (yields and
conditions in parenthesis). [d] Aryl bromide rather than aryl iodide was used
as reactant.
Scheme 5. Method development for construction of the B-ring.
more expensive and gave minor impurities so it was only used
for less reactive imines.
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3744ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 6. Exemplification of N-alkylphenanthridinium synthesis. See
Figure 3 for yields and conditions.
We then explored the range of substrates that could be
used in an optimised three-step procedure in which the crude
imine is heated in the microwave after an aqueous work-up,
and the product mixture is treated with ethereal HCl to form
the closed phenanthridinium salts 23–25, which in many cases
were isolated by simple precipitation (Scheme 6 and Figure 3).
8-Nitrophenanthridiniums 23a–c were produced in good
yields under these conditions. An electron-withdrawing group
is not required for reaction and the simple 5,6-disubstituted
phenanthridinium 23d could be prepared in good yield using
acetonitrile as solvent and a slightly higher temperature. How-
ever, placing an electron-donating methoxy group para to the
fluoride significantly impeded reaction even when the 8-nitro
group was present, so that phenanthridinium 23e was pro-
duced in lower yield under all conditions (the best yield was
obtained under the original solventless conditions with ex-
tended heating). Introduction of a second electron-withdraw-
ing nitro group gave rise to a dramatic acceleration and the
second step was unnecessary for the preparation of 3,8-dinitro-
phenanthridinium (23 f) in very high yield, because the cyclisa-
tion occurred spontaneously during removal of the toluene on
the rotary evaporator with modest heating. Unsurprisingly, the
synthesis of phenanthridinium 23g, which has an electron-do-
nating methoxy group para to the fluoride, required heating.
However, unlike the methoxyphenanthridinium 23e without
a 3-nitro group, reaction proceeded in good yield under the
standard conditions in THF. The acetophenone derivative 21h
also cyclised spontaneously and was isolated in quantitative
yield as the enamine 23h following an alkaline wash. The 2,8-
dinitrophenanthridinium 23i and 4-aza-8-nitrophenanthridini-
um 23j were also produced in high yield. Both formed more
easily than phenanthridiniums 23a–c bearing only one nitro
group, even though conversion of ketone 21i to phenanthridi-
nium 23i involved displacement of a chloride rather than
a fluoride.
Figure 3. Phenanthridinium salts 23–25 prepared from ketones 21 by imine
formation and microwave irradiation according to Scheme 6 (yields and con-
ditions in parenthesis) and phenanthridinium salt 26c prepared by reduction
of nitro compound 23c. [a] Prepared under thermal conditions [b] Second
step unnecessary as spontaneous cyclisation occurred.
such it is a useful tag in materials chemistry[40] and because it
combines with alkynes under bioorthogonal conditions has
found wide application in chemical biology.[39,41] Indeed, a li-
brary of 5,6-disubstituted phenanthridiniums bearing azido
tags was used for the in situ generation of AChE inhibitors.[16]
Therefore, we demonstrated that the azido tag was straightfor-
wardly introduced by our method. Thus, combining 6-azido-
hex-1-ylamine with ketone 21d gave phenanthridinium 25d in
good overall yield.
A 4-methyl group greatly impedes reaction, presumably be-
cause of steric interactions, so that extensive microwave heat-
ing in acetonitrile was required to prepare 2-methyl-8-nitro-
phenanthridinium 23k. Steric interactions were further investi-
gated with the use of a more sterically hindered amine, cyclo-
hexylmethylamine. The phenanthridinium 24a was isolated in
good yield, but unlike phenanthridinium 23a required the
higher temperature conditions in acetonitrile.
An azido group provides a way of introducing diversity
through Click chemistry[39–41] and Staudinger ligation;[39] as
The bromo group in phenanthridinium 23c also gives a po-
tential site for the introduction of diversity through cross-cou-
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3745ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
pling reactions, and so was chosen to exemplify access to the
important 8-amino compounds. Reduction by iron in AcOH/
EtOH at 408C gave phenanthridinium 26c in good yield.
Computational mechanistic study
The nucleophilic substitution of aryl halides by pyridine deriva-
tives is well known,[42,43] particularly for the preparation of
Zincke salts[42] (2,4-dinitrophenyl derivatives that have a variety
of uses). However, similar reactions by cyclic N-alkylimines are
extremely rare[44] and such reactions by acyclic N-alkylimines
almost unprecedented (we found one example[45]). The reac-
tion is mechanistically interesting because cyclisation of, for ex-
ample, imine 22d to give the corresponding Meisenheimer
complex 27d could occur either by electrocyclisation involving
the C=N p bond,[46] or by the traditional SNAr mechanism
through nucleophilic attack on the aromatic ring by the nitro-
gen lone pair (Scheme 7). The putative Meisenheimer complex
Scheme 8. Cyclisation reaction with transition state and intermediates as
identified computationally. The Meisenheimer complex II is not a stable min-
imum in MeCN solution.
Scheme 7. Two putative mechanisms for the formation of phenanthridinium
salt 28d from N-alkylimine 22d differing in the mode of cyclisation to give
Meisenheimer complex 27d.
27d would then fragment to give the phenanthridinium fluo-
ride 28d, which would be a precursor to the chloride salt 23d.
Since the line of attack in the first step is different for the two
mechanisms, we reasoned that they could be distinguished by
calculating the structure of the transition state.
We, therefore, embarked on a computational study at the
DFT level (M06-2X/def2-TZVP+) to elucidate the mechanism of
the cyclisation reaction, including structures and energies of
possible intermediates and transition states. Staying close to
the experimentally studied compounds, we considered the
imines Ia–h as reactants (Scheme 8). The series covers a range
of substitution patterns, which will allow us to analyse the in-
fluence of substituent effects on the reaction. All computation-
al results refer to reaction conditions of T=383 K and P=
320 kPa, corresponding to the conditions used in the cyclisa-
tion of 22d to 23d. Unless noted otherwise, all calculations
used a polarisable continuum solvent model of acetonitrile,
which was calibrated for the reaction conditions. For compari-
son, we also considered the reaction of Ia in vacuum.
Figure 4. Optimised structure of the transition state TSa with NBOs involved
in important donor–acceptor interactions. a) Donation of imine lone pair nN
into aryl p*(C4a=C4); b) Donation of nN into the CꢀF antibonding orbital,
s*(C4aꢀF).
state (Figure 4). As substituent and environment effects on the
structure of TS are minor, we are using TSa as the type speci-
men in the following discussion. The sp2 lone pair of the at-
The cyclisation transition state, TS, could readily be located
for all cases. Its structure is characteristic of an SNAr transition
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3746ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tacking imine nitrogen (N5) points straight at the halogen-
bearing aryl carbon (C4a), at an angle of 1258 to the aryl plane.
The coordination geometry at C4a is clearly pyramidal, with
the CꢀF bond out of the aryl plane by 368. The CꢀF bond is
lengthened from 1.35 ꢁ in the reactant Ia to 1.40 ꢁ in TSa. The
C=C bonds adjacent to C4a are also lengthened slightly (from
1.39 to 1.42 ꢁ), while the imine C=N bond remains practically
unchanged (1.27 to 1.28 ꢁ).
direct conjugation to C4a. However, it is important to note the
still significant stabilisation by the strong inductive effect of
a meta nitro group (TSc, R2 =NO2).[48] In contrast, a nitro group
on the C-ring provides less stabilisation even when it is formal-
ly in conjugation with the same carbon atom (TSb, R3 =NO2),
because the A- and C-rings are not co-planar (the biphenyl tor-
sion is ꢀ548 in the reactant Ia and reduces to ꢀ378 in TSa).
For the cases considered, substituent effects on the barrier
height are additive: the barriers for the disubstituted TSf and
TSg are within 2 kJmolꢀ1 of the values predicted from the in-
crements derived from the monosubstituted cases.
The characterisation of the cyclisation reaction as a nucleo-
philic attack, rather than an electrocyclic process, was further
corroborated through a natural bond orbital (NBO) analysis.[47]
In essence, NBOs are constructed such as to correspond opti-
mally to a Lewis picture of localised (anti-)bonding orbitals and
lone pairs (in contrast to the often highly delocalised canonical
molecular orbitals).
The reactivity towards nucleophilic attack is significantly re-
duced in the chloro compound (TSh) compared to the fluoro
congener (TSa), again as expected for an SNAr reaction. The
more electronegative the leaving group, the more positive is
the partial charge on the attacked carbon and the lower in
energy is the acceptor s*CX orbital, which translates into higher
reactivity. In Ia and Ih, the Hirshfeld charges on the halogen
are ꢀ0.094 e (F) and ꢀ0.075 e (Cl), while the charges on the
carbon are 0.080 and 0.031 e, respectively. Similarly,
DEdeloc(nN!s*CCl)=ꢀ60 kJmolꢀ1 in TSh, compared to
ꢀ126 kJmolꢀ1 in TSa, indicating the reduced stabilisation of
the transition state by the less electronegative halogen.
Figure 4 shows the imine lone pair together with the two
principal empty orbitals it is interacting with. The lone pair do-
nates substantial electron density (0.4 e) into the aryl antibond-
ing orbital p*(C4a=C4) and the CꢀF antibonding orbital
s*(C4aꢀF). Both interactions are energetically highly favoura-
ble, DEdeloc =ꢀ519 and ꢀ126 kJmolꢀ1, respectively; the former
exceeds the next smaller donor–acceptor interaction more
than threefold. In contrast, the imine C=N p and p* orbitals
have no significant interaction with any orbital on the halo-
gen-bearing A-ring. The incipient CꢀN bond therefore arises
principally from the interaction of the imine lone pair with the
aryl p* adjacent to the carbon, which is the signature of a clas-
sic SNAr attack.
The effect of solvation on the barrier is remarkably small:
TSa is only 9 kJmolꢀ1 lower than TSavac. The vacuum transition
state is slightly later, judging from the shorter NꢀC distance
(1.81 vs. 1.85 ꢁ); however, there is hardly any difference in the
degree of activation of the CꢀF bond (1.39 vs. 1.40 ꢁ). This is
another clear indication that the transition state corresponds
to the formation of the CꢀN bond, while the CꢀX bond, al-
though weakened, is essentially maintained. The transition
state can therefore be described as Meisenheimer-like in terms
of its geometric and electronic structure.
The free-energy barrier for the formation of the CꢀN bond is
103 kJmolꢀ1 in the parent TSa (Table 1). The barriers show the
Table 1. Gibbs free energies[a] relative to I along the reaction sequence
shown in Scheme 8. Also given are the lengths of the forming (N5ꢀC4a)
and breaking (C4aꢀX) bonds in the transition state.
Although we have not collected any experimental kinetic
data in this study, we may still use the yields and reaction con-
ditions as indicators for the required activation energies. Com-
paring the unsubstituted 23d (corresponding to the cyclisation
product of Ia) with 23b (Ib), 23e (Ig), and 23 f (If; Figure 3),
we find that the yields and required conditions agree with the
trends in the computed free-energy barriers.
DG [kJmolꢀ1
]
d° [ꢁ]
TS
103
98
93
72
115
90
112
120
112
III
III+ +Xꢀ
ꢀ81
IV
N5ꢀC4a
C4aꢀX
1.40
1.39
1.39
1.37
1.41
1.39
1.40
1.87
1.39
a
b
c
d
e
f
ꢀ65
ꢀ57
ꢀ60
ꢀ52
ꢀ70
ꢀ51
ꢀ61
ꢀ126
80[d]
ꢀ79
ꢀ84
ꢀ93
ꢀ89
ꢀ78
ꢀ88
1.85
1.88
1.87
1.96
1.84
1.88
1.87
1.88
1.81
ꢀ74
ꢀ72
ꢀ71
Energy-minimising the SNAr transition states towards prod-
ucts, we expected to obtain the Meisenheimer complexes II.
However, we found that IIa–h are not stable minima in MeCN
solution. Rather, the CꢀX bond is spontaneously cleaved to
form the ion pairs III, in which the halide anion is located at
a distance of approximately 3 ꢁ directly “beneath” the nitrogen
N5 of the planar phenanthridinium cation. We are therefore
dealing with a concerted, one-step SNAr reaction, in which the
formation of the Meisenheimer s-complex constitutes the rate-
limiting barrier, while its decomposition is barrier-free.
ꢀ90
ꢀ59
g
h
avac
ꢀ77
ꢀ76
[b]
ꢀ133
440
[c]
ꢀ80
[a] Calculated at M06-2X/def2-TZVP+ level in polarisable continuum
MeCN (er =22.5) for T=383 K, P=320 kPa. [b] The 6-chlorophenanthridine
IVh is not a stable minimum, but dissociates into the ion pair IIIh during
structure optimisation. [c] Calculated in vacuum. [d] Energy of IIavac. In
vacuum, the ion pair IIIavac is not stable, but collapses to IIavac
.
Energetically, these reactions are exergonic by 50–
70 kJmolꢀ1 (Table 1). The driving force is mainly enthalpic, due
to the aromatisation of the B-ring and the strong solvation of
the anion. For X=Cl (IIIh), the reaction is even more exergonic,
despite the weaker solvation of Clꢀ compared to Fꢀ, because
of the weaker CꢀCl bond.[49] For all cases in solution, the com-
plete separation of the ion pair into free ions III+ +Xꢀ is fur-
substituent effects expected for a nucleophilic mechanism,
which is promoted by stabilising the developing partial nega-
tive charge on the attacked moiety: electron-withdrawing sub-
stituents lower the barrier (TSb, c, d, f), while donors increase
it (TSe). The stabilising effect of the nitro substituent is stron-
gest when it is para to the halogen (TSd, R1 =NO2) because of
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3747ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ther exergonic, which is driven entropically; enthalpically, the
ion pair is favoured over separated ions (see the Supporting In-
formation). The relative stability of the phenanthridinium cat-
ions is governed by substituent effects, which act opposite to
the situation in electron-rich TS: acceptor substituents destabi-
lise the cation (IIIb, c, d, f), whereas donor substituents stabi-
lise it (IIIe).
Instead of remaining as separated ions, the halide may reat-
tach to the phenanthridinium to form the neutral 6-halophe-
nanthridines IV. The thermodynamic viability of this step de-
pends on the substituents; for X=F, the reaction free energies
relative to separated ions are between ꢀ23 kJmolꢀ1 (IVf) and
+12 kJmolꢀ1 (IVe). On enthalpic grounds, IV is always more
stable than the ion pair or separated ions (see the Supporting
Information). The relative stability of the 6-fluorophenanthri-
dines IV follows the same pattern as in the TS, opposite to the
substituent effects in the cations. In contrast to the fluoro
compounds IVa–g, IVh with X=Cl is not a stable energy mini-
mum, but spontaneously dissociates into the ion pair IIIh. This
is in pleasing agreement with experiment, in which the original
fluoride is exchanged for chloride during the work-up (step iii
in Scheme 6) and phenanthridinium chlorides are obtained ex-
clusively.
Both cases are special in that the departing fluoride can bind
to the nucleophile (hydrogen bonding in the former, formation
of a fluorosilane in the latter) and is therefore greatly stabi-
lised, despite both reactions being in vacuum. Two recent
studies have highlighted that the potential-energy surface in
the vicinity of the Meisenheimer complex can be very flat, so
that even small changes in the computational method can de-
termine its nature. For instance, in the reaction of pentafluoro-
pyridine with NH3, the Meisenheimer complex is only prevent-
ed from decomposing spontaneously by a vanishingly low bar-
rier (0.1 kJmolꢀ1
)
at the B3LYP/6-31+G(d,p)/PCM(water)
level,[51] but the barrier increases to a significant 19 kJmolꢀ1
when the M06-2X functional is used. For the substitution of
a phenylamidate by an amine in polarisable continuum MeOH,
the existence of a stable Meisenheimer intermediate was simi-
larly found to depend on the choice of exchange-correlation
functional.[55]
Together with the present results, these observations point
to the key factor that determines the nature of the Meisen-
heimer complex in SNAr reactions: it is the balance between
the bond strength of the leaving group X and the stabilisation
of free Xꢀ. For X=F in particular, the strong arylꢀF bond,
which in the Meisenheimer complex is already close to
a weaker C(sp3)ꢀF bond, must be compensated for by a stabili-
sation of the incipient Fꢀ that acts early on along the reaction
coordinate. Unless stabilisation is provided as soon as the CꢀF
bond begins to break and significant negative charge to build
up on the fluoride, the Meisenheimer complex will become
a minimum, with a barrier to decomposition.
The question of whether the Meisenheimer s-complex is
a stable minimum or a transition state along the reaction coor-
dinate of the SNAr reaction, that is, whether the reaction is
stepwise or concerted, has been addressed in a number of pre-
vious computational studies (Figure 5).[50–56] The answer de-
In general, solvation is the single most important factor sta-
bilising the departing (anionic) leaving group. Accordingly, in
the absence of solvent, we find that the Meisenheimer com-
plex, IIavac, is stable (Table 1), as separating charges is highly
disfavoured in vacuum, which agrees with previous stud-
ies.[50,52] While an explicit representation of the solvent com-
bined with extensive configurational sampling (e.g., a Monte
Carlo QM/MM approach[53]) may be required to model reliably
the reaction in polar-protic solvents, we believe that a much
simpler polarisable continuum model is adequate for dipolar-
aprotic solvents, like MeCN. However, it is important that the
model accurately reproduces the dominating solvation of the
free anion. We have therefore taken great care to calibrate our
solvation model to the best available experimental solvation
data for Fꢀ and Clꢀ. Moreover, it is crucial that structures are
fully optimised in solution as the potential-energy surfaces
differ qualitatively between vacuum and solution. Simply cor-
recting the energies of vacuum structures for solvation effects
will necessarily miss the possibility of spontaneous decomposi-
tion of the s-complex.
Figure 5. The Meisenheimer complex (MC) can be either a transition state
(solid black line) or an intermediate (dashed grey line) along the SNAr reac-
tion coordinate.
pends on the nature of the nucleophile (neutral/anionic, nucle-
ophilicity), the leaving group (bond strength, stability of free
leaving group), the arene core (substitution pattern), and the
reaction medium (gas phase, solvent polarity). For reactions of
fluoroarenes (mostly fluorobenzenes) with various nucleo-
philes, the s-complex was in most cases found to be an inter-
mediate,[50–53] only two studies reported it to be a transition
state: in the reaction of para-NO2C6H4F with NH3 or NH2Me[50]
and in the reaction of C6H5F with Me2NSiMe3 or Me2PSiMe3.[54]
The second requirement to model the reaction reliably is an
adequate description of the electronic structure. Crucially, the
localised negative charge of the free anion must be described
sufficiently well otherwise its stability will be artificially under-
estimated. This is not possible without a reasonably large basis
set that includes diffuse functions. We initially ran exploratory
calculations with the def2-SVP basis set, which predicted
stable Meisenheimer intermediates for all cases with X=F; it
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3748ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Variations to reaction times, solvent and temperature are shown in
Figure 3. The characterisation data and procedures for all com-
pounds are provided in the Supporting Information. Assignment of
the 1H NMR spectra of phenanthridiniums 23–25 was achieved
with the assistance of COSY, HMQC and NOESY and agree well
with those described by Luedtke et al.[5]
was only for X=Cl that we obtained the ion pair (see the Sup-
porting Information). The def2-TZVP+ basis set, which includes
diffuse functions on all atoms, provided a qualitatively different
picture in which the s-complexes are not stable. Again, struc-
tures need to be optimised at the higher level; single-point
corrections will fail to capture the effect. Where both basis sets
yield qualitatively the same structure (e.g., for the transition
states TS), energies differ insignificantly (<5 kJmolꢀ1).
Computational methods
All calculations were done with the Gaussian 09 program.[57] We
chose the M06-2X[58] exchange-correlation functional for its good
performance in organic thermochemistry and reaction barriers and
its ability to describe accurately dispersive interactions.[59] Produc-
tion calculations used the def2-TZVP+ basis set, which we derived
from def2-TZVP[60] by adding one diffuse function per valence an-
gular momentum. Gibbs energies were calculated using the stan-
dard rigid rotor/harmonic oscillator approximation for T=383 K,
P=320 kPa. The SMD solvation model was used as implemented
in Gaussian, which combines the IEF-PCM polarisable continuum
model for electrostatic solvation with the SMD non-electrostatic
terms. We specified non-default values for the following solvation
parameters: dielectric constant of MeCN, er(383 K)=22.5; Rsolv(F)=
1.63 ꢁ, Rsolv(Cl)=2.48 ꢁ. Structures were fully optimised in solvent.
Minima and transition states were verified through the correct
number of imaginary frequencies. The IRC was followed for ten
steps to either side of transition states, followed by optimisation to
the adjacent minimum. NBO analyses were done using the
NBO 6.0 program.[61] Additional details, in particular on the calibra-
tion of the solvation model, can be found in the Supporting Infor-
mation.
Conclusion
In summary, we present a versatile, efficient and convergent
route to 5,6-disubstituted phenanthridiniums using three easily
accessible components: arylborons, ortho-iodoaryl ketones and
primary amines. The key step in the synthesis is the novel mi-
crowave-assisted intramolecular displacement of an aryl halide
by an acyclic imine. We demonstrate through DFT calculations
that the reaction proceeds by an SNAr rather than an electrocy-
clic mechanism, and that it is concerted and does not involve
a stable Meisenheimer complex. Of significant importance to
the computational work is the proper modelling of the solvat-
ed free anion. The concerted mechanism for the SNAr that
comprises our key step is very unusual: the vast majority of ex-
amples of this type of reaction appear to involve Meisenheimer
complexes as stable intermediates. The use of our synthetic
method provides access to a wide range of structural ana-
logues and should enable the development of phenanthridini-
um probes to respond to redox, DNA intercalation and
enzyme reactivity independently.
Acknowledgements
We thank the BBSRC for funding through BB/D526310/1 and
BB/I012826/1, and the EPSRC National Mass Spectrometry Fa-
cility in Swansea for MS.
Experimental Section
General procedure for imine formation-cyclisation using
conversion of ketone 21j into 23j as an example
Keywords: density functional calculations
·
nucleophilic
substitution · phenanthridinium · solvation · synthetic methods
A solution of titanium tetrachloride in dry toluene (1m, 2.40 mL,
2.40 mmol, 1.49 equiv) was added over 5 min to a stirred solution
of benzophenone 21j (520 mg, 1.61 mmol, 1.00 equiv) and dry
hexylamine (1.50 mL, 11.4 mmol, 7.05 equiv) in dry toluene (5.0 mL)
at 08C. The reaction was allowed to warm to RT and stirring was
continued for 20 h. The mixture was poured into H2O, then filtered
through Celite eluting with DCM. HCl(aq.) (1m) and more DCM were
added to the filtrate, the mixture was shaken and the layers were
separated. The aqueous layer was re-extracted (DCMꢂ2). The com-
bined organics were washed with 1m HCl(aq.) then NaHCO3(aq.) and
dried (MgSO4) then filtered and the solvent was removed under re-
duced pressure. From the crude yield of 631 mg of brown solid
a portion (100 mg, 0.25 mmol, 1.00 equiv) was transferred into a mi-
crowave tube (10 mL) containing dry THF (5.0 mL). The tube was
filled with argon and sealed, and then heated in a microwave to
1008C for 30 min. It was then cooled and transferred to a round-
bottomed flask in dry CHCl3. Then the solvent was removed under
reduced pressure. The residue was dissolved in dry Et2O and then
precipitated with 2m ethereal HCl (1.00 mL, 2.00 mmol, 8.13 equiv).
The solvent was removed under reduced pressure and then the
solid triturated with pet. ether and dried under reduced pressure
to give 4-aza-5-(hex-1’-yl)-6-phenyl-8-nitrophenanthridinium chlo-
ride 23j as a white amorphous solid (106 mg, 98%).
[1] Review: H. Ihmels, D. Otto, Top. Curr. Chem. 2005, 258, 161–204.
[2] Review: J. Zielonka, B. Kalyanaraman, Free Radical Biol. Med. 2010, 48,
983–1001.
[3] Recent examples include: a) N. Jain, S. Francis, S. H. Friedman, Bioorg.
Med. Chem. Lett. 2012, 22, 4844–4848; b) M. R. Stojkovic, S. Marczi, L.
Glavas-Obrovac, I. Piantanida, Eur. J. Med. Chem. 2010, 45, 3281–3292;
c) N. N. Shaw, H. Xi, D. P. Arya, Bioorg. Med. Chem. Lett. 2008, 18, 4142–
4145; d) S. Rangarajan, S. H. Friedman, Bioorg. Med. Chem. Lett. 2007,
17, 2267–2273; e) A. D. C. Parenty, L. V. Smith, K. M. Guthrie, D. L. Long,
J. Plumb, R. Brown, L. Cronin, J. Med. Chem. 2005, 48, 4504–4506;
f) M. L. Bolognesi, V. Andrisano, M. Bartolini, R. Banzi, C. Melchiorre, J.
Med. Chem. 2005, 48, 24–27.
[4] Review: M. Wainwright, Biotech. Histochem. 2010, 85, 341–354.
[5] N. W. Luedtke, Q. Liu, Y. Tor, Chem. Eur. J. 2005, 11, 495–508.
[6] T. Kubar, M. Hanus, F. Ryjacek, P. Hobza, Chem. Eur. J. 2006, 12, 280–
290.
[7] a) N. W. Luedtke, Q. Liu, Y. Tor, Bioorg. Med. Chem. 2003, 11, 5235–5247;
b) L. P. G. Wakelin, M. J. Waring, Mol. Pharmacol. 1974, 10, 544–561;
c) W. Leupin, J. Feigon, W. A. Denny, D. R. Kearns, Biophys. Chem. 1985,
22, 299–305.
[8] C. Prunkl, M. Pichlmaier, R. Winter, V. Kharlanov, W. Rettig, H. A. Wagen-
knecht, Chem. Eur. J. 2010, 16, 3392–3402.
[9] Q. Huang, W. L. Fu, Clin. Chem. Lab. Med. 2005, 43, 841–842.
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3749ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[10] Recent examples include: a) W. B. Hu, C. Blecking, M. Kralj, L. Suman, I.
Piantanida, T. Schrader, Chem. Eur. J. 2012, 18, 3589–3597; b) F. Dimiza,
F. Perdih, V. Tangoulis, I. Turel, D. P. Kessissoglou, G. Psomas, J. Inorg.
Biochem. 2011, 105, 476–489; c) Z. Liu, A. Habtemariam, A. M. Pizarro,
S. A. Fletcher, A. Kisova, O. Vrana, L. Salassa, P. C. A. Bruijnincx, G. J.
Clarkson, V. Brabec, P. J. Sadler, J. Med. Chem. 2011, 54, 3011–3026;
d) S. P. Jones, G. M. Pavan, A. Danani, S. Pricl, D. K. Smith, Chem. Eur. J.
2010, 16, 4519–4532; e) M. A. Kostiainen, J. Kotimaa, M. L. Laukkanen,
G. M. Pavan, Chem. Eur. J. 2010, 16, 6912–6918; f) E. B. Veale, T. Gunn-
laugsson, J. Org. Chem. 2010, 75, 5513–5525.
Y. F. Song, G. Cooke, L. Cronin, J. Am. Chem. Soc. 2008, 130, 13059–
13065.
[28] For example: T. Nakanishi, A. Masuda, M. Suwa, Y. Akiyama, N. Hoshino-
Abe, M. Suzuki, Bioorg. Med. Chem. Lett. 2000, 10, 2321–2323.
[29] Y.-Y. Liu, R.-J. Song, C.-Y. Wu, L.-B. Gong, M. Hu, Z.-Q. Wang, Y.-X. Xie, J.-
H. Li, Adv. Synth. Catal. 2012, 354, 347–353.
[30] J. Barluenga, M. Fananas-Mastral, F. Aznar, C. Valdes, Angew. Chem.
2008, 120, 6696–6699; Angew. Chem. Int. Ed. 2008, 47, 6594–6597.
[31] K. Augustine, A. Bombrun, P. Alagarsamy, A. Jothi, Tetrahedron Lett.
2012, 53, 6280–6287.
[11] Reviews: a) M. Zille, T. D. Farr, I. Przesdzing, J. Muller, C. Sommer, U. Dir-
nagl, A. Wunder, J. Cereb. Blood Flow Metab. 2012, 32, 213–231; b) Z.
Darzynkiewicz, S. Bruno, G. Delbino, W. Gorczyca, M. A. Hotz, P. Lassota,
F. Traganos, Cytometry 1992, 13, 795–808.
[32] a) N. Sotomayor, E. Dominguez, E. Lete, J. Org. Chem. 1996, 61, 4062–
4072; b) H. Ishii, Y. Ichikawa, E. Kawanabe, M. Ishikawa, T. Ishikawa, K.
Kuretani, M. Inomata, A. Hoshi, Chem. Pharm. Bull. 1985, 33, 4139–
4151.
[12] N. A. O’Connor, N. Stevens, D. Samaroo, M. R. Solomon, A. A. Marti, J.
Dyer, H. Vishwasrao, D. L. Akins, E. R. Kandel, N. J. Turro, Chem.
Commun. 2009, 2640–2642; M. van der Wiel, J. Cheng, R. Koukiekolo,
R. K. Lyn, N. Stevens, N. O’Connor, N. J. Turro, J. P. Pezacki, J. Am. Chem.
Soc. 2009, 131, 9872–9873; N. Stevens, N. O’Connor, H. Vishwasrao, D.
Samaroo, E. R. Kandel, D. L. Akins, C. M. Drain, N. J. Turro, J. Am. Chem.
Soc. 2008, 130, 7182–7183.
[33] a) A. Suzuki, Angew. Chem. 2011, 123, 6854–6869; Angew. Chem. Int. Ed.
2011, 50, 6722–6737; b) C. Amatore, A. Jutand, G. Le Duc, Chem. Eur. J.
2012, 18, 6616–6625.
[34] H. Amii, K. Uneyama, Chem. Rev. 2009, 109, 2119–2183.
[35] Recent example: M. Dçbele, S. Vanderheiden, N. Jung, S. Brꢅse, Angew.
Chem. 2010, 122, 6122–6125; Angew. Chem. Int. Ed. 2010, 49, 5986–
5988.
[13] a) N. Amann, R. Huber, H. A. Wagenknecht, Angew. Chem. 2004, 116,
1881–1883; Angew. Chem. Int. Ed. 2004, 43, 1845–1847; b) R. Huber, N.
Amann, H. A. Wagenknecht, J. Org. Chem. 2004, 69, 744–751.
[14] a) Y. Bourne, P. Taylor, Z. Radic, P. Marchot, EMBO J. 2003, 22, 1–12; b) Z.
Radic, P. Taylor, J. Biol. Chem. 2001, 276, 4622–4633.
[36] For a review, see: G. A. Molander, B. Canturk, Angew. Chem. 2009, 121,
9404–9425; Angew. Chem. Int. Ed. 2009, 48, 9240–9261.
[37] C. J. O’Brien, E. A. B. Kantchev, C. Valente, N. Hadei, G. A. Chass, A.
Lough, A. C. Hopkinson, M. G. Organ, Chem. Eur. J. 2006, 12, 4743–
4748.
[15] a) Y. Bourne, Z. Radic, P. Taylor, P. Marchot, J. Am. Chem. Soc. 2010, 132,
18292–18300; b) Y. Bourne, H. C. Kolb, Z. Radic, K. B. Sharpless, P. Taylor,
P. Marchot, Proc. Natl. Acad. Sci. USA 2004, 101, 1449–145.
[38] a) H. Ghaneolhosseini, W. Tjarks, S. Sjoberg, Tetrahedron 1997, 53,
17519–17526; b) T. I. Watkins, J. Chem. Soc. 1952, 3059–3064; c) L. P.
Walls, J. Chem. Soc. 1945, 294–300.
[16] a) R. Manetsch, A. Krasinski, Z. Radic, J. Raushel, P. Taylor, K. B. Sharpless,
H. C. Kolb, J. Am. Chem. Soc. 2004, 126, 12809–12818; b) W. G. Lewis,
L. G. Green, F. Grynszpan, Z. Radic, P. R. Carlier, P. Taylor, M. G. Finn, K. B.
Sharpless, Angew. Chem. 2002, 114, 1095–1099; Angew. Chem. Int. Ed.
2002, 41, 1053–1057.
[17] M. R. Lee, I. Shin, Angew. Chem. 2005, 117, 2941–2944; Angew. Chem.
Int. Ed. 2005, 44, 2881–2884.
[18] R. A. J. Smith, R. C. Hartley, M. P. Murphy, Antioxid. Redox Signaling 2011,
15, 3021–3038.
[19] K. Kundu, S. F. Knight, S. Lee, W. R. Taylor, N. Murthy, Angew. Chem.
2010, 122, 6270–6274; Angew. Chem. Int. Ed. 2010, 49, 6134–6138.
[20] R. Michalski, J. Zielonka, M. Hardy, J. Joseph, B. Kalyanaraman, Free Radi-
cal Biol. Med. 2013, 54, 135–147.
[39] E. M. Sletten, C. R. Bertozzi, Acc. Chem. Res. 2011, 44, 666–676.
[40] B. S. Sumerlin, A. P. Vogt, Macromolecules 2010, 43, 1–13.
[41] E. Lallana, R. Riguera, E. Fernandez-Megia, Angew. Chem. 2011, 123,
8956–8966; Angew. Chem. Int. Ed. 2011, 50, 8794–8804.
[42] Recent examples of Zincke salts include: a) G. Schmidt, C. Timm, A.
Grube, C. A. Volk, M. Koeck, Chem. Eur. J. 2012, 18, 8180–8189; b) K.
Hata, Y. Segawa, K. Itami, Chem. Commun. 2012, 48, 6642–6644; c) I. Ya-
maguchi, T. Nakahara, J. Polym. Sci. Part A 2012, 50, 3340–3349; d) B. J.
Coe, S. P. Foxon, E. C. Harper, M. Helliwell, J. Raftery, C. A. Swanson, B. S.
Brunschwig, K. Clays, E. Franz, J. Garꢆn, J. Orduna, P. N. Horton, M. B.
Hursthouse, J. Am. Chem. Soc. 2010, 132, 1706–1723; e) T. D. Michels,
M. J. Kier, A. M. Kearney, C. D. Vanderwal, Org. Lett. 2010, 12, 3093–
3095.
[21] K. M. Robinson, M. S. Janes, M. Pehar, J. S. Monette, M. F. Ross, T. M.
Hagen, M. P. Murphy, J. S. Beckman, Proc. Natl. Acad. Sci. USA 2006, 103,
15038–15043.
[22] A. Logan, H. M. Cochemꢃ, P. B. L. Pun, N. Apostolova, R. A. J. Smith, L.
Larsen, D. S. Larsen, A. M. James, I. M. Fearnley, S. Rogatti, T. A. Prime,
P. G. Finichiu, A. Dare, E. T. Chouchani, V. R. Pell, C. Methner, C. Quin, S. J.
McQuaker, T. Krieg, R. C. Hartley, M. P. Murphy, Biophys. Biochim. Acta
2013, 1840, 923–930.
[23] a) B. Zhang, C. Mꢄck-Lichtenfeld, C. G. Daniliuc, A. Studer, Angew. Chem.
2013, 125, 10992–10995; Angew. Chem. Int. Ed. 2013, 52, 10792–
10795; b) Y. Cheng, H. Jiang, Y. Zhang, S. Yu, Org. Lett. 2013, 15, 5520–
5523; c) I. Deb, N. Yoshikai, Org. Lett. 2013, 15, 4254–4257; d) R. T.
McBurney, J. C. Walton, J. Am. Chem. Soc. 2013, 135, 7349–7354; e) D.
Intrieri, M. Mariani, A. Caselli, F. Ragaini, E. Gallo, Chem. Eur. J. 2012, 18,
10487–10490; f) A. Korotvicka, I. Cisarova, J. Roithova, M. Kotora, Chem.
Eur. J. 2012, 18, 4200–4207; g) Z. J. Liang, L. Ju, Y. J. Xie, L. H. Huang,
Y. H. Zhang, Chem. Eur. J. 2012, 18, 15816–15821; h) D. Takeda, K.
Hirano, T. Satoh, M. Miura, Heterocycles 2012, 86, 487–496; i) M. Tobisu,
K. Koh, T. Furukawa, N. Chatani, Angew. Chem. 2012, 124, 11525–11528;
Angew. Chem. Int. Ed. 2012, 51, 11363–11366; j) M. Blanchot, D. A. Can-
dito, F. Larnaud, M. Lautens, Org. Lett. 2011, 13, 1486–1489.
[24] For example: F. Diaba, I. Lewis, M. Grignon-Dubois, S. Navarre, J. Org.
Chem. 1996, 61, 4830–4832.
[43] Recent examples of cyclisations include: a) U. F. J. Mayer, E. Murphy,
M. F. Haddow, M. Green, R. W. Alder, D. F. Wass, Chem. Eur. J. 2013, 19,
4287–4299; b) M. E. Zhidkov, V. A. Kaminskii, Tetrahedron Lett. 2013, 54,
3530–3532; c) A. Jankowiak, E. Obijalska, P. Kaszynski, A. Pieczonka,
V. G. Young, Jr., Tetrahedron 2011, 67, 3317–3327; d) M. E. Zhidkov, O. V.
Baranova, N. N. Balaneva, S. N. Fedorov, O. S. Radchenko, S. V. Dubovit-
skii, Tetrahedron Lett. 2007, 48, 7998–8000; e) M. D. Garcꢆa, A. J. Wilson,
D. P. G. Emmerson, P. R. Jenkins, Chem. Commun. 2006, 2586–2588.
[44] a) C. Marivingt-Mounir, C. Norez, R. Dꢃrand, L. Bulteau-Pignoux, D.
Nguyen-Huy, B. Viossat, G. Morgant, F. Becq, J.-M. Vierfond, Y. Mettey, J.
Med. Chem. 2004, 47, 962–972; b) C. Marivingt-Mounir, D. Sarrouilhe, Y.
Mettey, J.-M. Vierfond, Pharm. Pharmacol. Commun. 1998, 4, 23–25.
[45] H. M. Wolff, K. Hartke, Arch. Pharm. (Weinheim, Ger.) 1980, 313, 266–
279.
[46] Y. Ishihara, S. Azuma, T. Choshi, K. Kohno, K. Ono, H. Tsutsumi, T. Ishizu,
S. Hibino, Tetrahedron 2011, 67, 1320–1333.
[47] a) F. Weinhold, C. R. Landis, Valency and Bonding, Cambridge University
Press, Cambridge, 2005; b) E. D. Glendening, C. R. Landis, F. Weinhold,
Wiley Interdiscip. Rev. Comput. Mol. Sci. 2012, 2, 1–42; c) F. Weinhold,
C. R. Landis, Discovering Chemistry with Natural Bond Orbitals, Wiley,
Hoboken, 2012.
[48] O. Exner, T. M. Krygowski, Chem. Soc. Rev. 1996, 25, 71–75.
[49] S. J. Blanksby, G. B. Ellison, Acc. Chem. Res. 2003, 36, 255–263.
[50] a) K. Tanaka, M. Deguchi, S. Iwata, J. Chem. Res. Synop. 1999, 528–529;
b) K. Tanaka, M. Deguchi, S. Iwata, J. Chem. Res. Miniprint 1999, 2279–
2294.
[25] J.-J. Chen, K.-T. Li, D.-Y. Yang, Org. Lett. 2011, 13, 1658–1661.
[26] P. J. Kitson, A. D. C. Parenty, C. J. Richmond, D.-L. Long, L. Cronin, Chem.
Commun. 2009, 4067–4069.
[27] The same group used reversible intramolecular addition of a nitrogen
nucleophile in redox and pH switches: C. J. Richmond, A. D. C. Parenty,
[51] M. Liljenberg, T. Brinck, T. Rein, M. Svensson, Beilstein J. Org. Chem.
2013, 9, 791–799.
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3750ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[52] a) M. N. Glukhovtsev, R. D. Bach, S. Laiter, J. Org. Chem. 1997, 62, 4036–
4046; b) I. Fernꢇndez, G. Frenking, E. Uggerud, J. Org. Chem. 2010, 75,
2971–2980; c) K. B. Jose, J. Cyriac, J. T. Moolayil, V. S. Sebastian, M.
George, J. Phys. Org. Chem. 2011, 24, 714–719; d) M. Liljenberg, T.
Brinck, B. Herschend, T. Rein, G. Rockwell, M. Svensson, Tetrahedron Lett.
2011, 52, 3150–3153; e) M. Liljenberg, T. Brinck, B. Herschend, T. Rein, S.
Tomasi, M. Svensson, J. Org. Chem. 2012, 77, 3262–3269.
[53] O. Acevedo, W. L. Jorgensen, Org. Lett. 2004, 6, 2881–2884.
[54] L. I. Goryunov, J. Grobe, D. Le Van, V. D. Shteingarts, R. Mews, E. Lork, E.-
U. Wꢄrthwein, Eur. J. Org. Chem. 2010, 1111–1123.
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. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, N. J. 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, ꢊ. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, D. J. Fox, Gaussian, Inc., Wallingford, CT, 2009.
[55] N. Chꢃron, D. Jacquemin, P. Fleurat-Lessard, Phys. Chem. Chem. Phys.
2012, 14, 7170–7175.
[58] Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241.
[59] a) L. Goerigk, S. Grimme, J. Chem. Phys. 2010, 132, 184103; b) Y. Zhao,
D. G. Truhlar, J. Chem. Theory Comput. 2011, 7, 669–676.
[60] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297–3305.
[61] a) E. D. Glendening, J. K. Badenhoop, A. E. Reed, J. E. Carpenter, J. A.
Bohmann, C. M. Morales, C. R. Landis, F. Weinhold, NBO, V. 6.0, Madison,
WI, Theoretical Chemistry Institute, University of Wisconsin, 2013:
http://nbo6.chem.wisc.edu/; b) E. D. Glendening, C. R. Landis, F. Wein-
hold, J. Comput. Chem. 2013, 34, 1429–1437; c) E. D. Glendening, C. R.
Landis, F. Weinhold, J. Comput. Chem. 2013, 34, 2134.
[56] a) T. Giroldo, L. A. Xavier, J. M. Riveros, Angew. Chem. 2004, 116, 3672–
3674; Angew. Chem. Int. Ed. 2004, 43, 3588–3590; b) H. Sun, S. G. Di-
Magno, Angew. Chem. 2006, 118, 2786–2791; Angew. Chem. Int. Ed.
2006, 45, 2720–2725; c) B. Chiavarino, M. E. Crestoni, S. Fornarini, F. La-
nucara, J. Lemaire, P. Maꢈtre, Angew. Chem. 2007, 119, 2041–2044;
Angew. Chem. Int. Ed. 2007, 46, 1995–1998; d) J. R. Pliego, Jr., D. Pilꢉ-
Veloso, Phys. Chem. Chem. Phys. 2008, 10, 1118–1124; e) B. Chiavarino,
M. E. Crestoni, S. Fornarini, F. Lanucara, J. Lemaire, P. Maꢈtre, D. Scuderi,
Chem. Eur. J. 2009, 15, 8185–8195; f) S.-W. Park, S.-Y. Lee, Bull. Korean
Chem. Soc. 2010, 31, 2571–2574.
[57] Gaussian 09, RevisionC.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, 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,
Received: October 30, 2013
Revised: January 3, 2014
Published online on February 19, 2014
Chem. Eur. J. 2014, 20, 3742 – 3751
www.chemeurj.org
3751ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim