The interactions of bis-phenanthridinium-nucleobase conjugates with nucleotides: adenine-conjugate recognizes UMP in aqueous medium

Article abstract of DOI:10.1016/j.tet.2010.01.063

Among series of novel bis-phenanthridinium-nucleobase conjugates, the adenine derivative revealed high (log Ks=6.9 M^-1) and selective affinity toward complementary nucleotide (UMP), accompanied by specific change in the UV-vis spectrum of phena

Full text of DOI:10.1016/j.tet.2010.01.063

Tetrahedron 66 (2010) 2501–2513
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
The interactions of bis-phenanthridinium–nucleobase conjugates with
nucleotides: adenine-conjugate recognizes UMP in aqueous medium
a
b
b
,
a,
*
*
´
Lidija-Marija Tumir , Marina Grabar , Sanja Tomic , Ivo Piantanida
a
ˇ
Division of Organic Chemistry & Biochemistry, RuCer Boskovi´c Institute, HR 10002 Zagreb, P.O.B. 180, Croatia
Laboratory for Chemical and Biological Crystallography, Division of Physical Chemistry, RuCer Boskovi´c Institute, HR 10002 Zagreb, P.O.B. 180, Croatia
b
ˇ
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 14 August 2009
Received in revised form 9 December 2009
Accepted 18 January 2010
Available online 2 February 2010
Among series of novel bis-phenanthridinium–nucleobase conjugates, the adenine derivative revealed
high (log Ks¼6.9 Mꢀ1) and selective affinity toward complementary nucleotide (UMP), accompanied by
specific change in the UV–vis spectrum of phenanthridine subunits, differing significantly from changes
caused by addition of other nucleotides. High stability and selectivity of adenine-conjugate/UMP non-
covalent complex is, according to the molecular modeling studies, correlated to the number of inter- and
intramolecular aromatic stacking interactions between phenanthridinium subunits, covalently attached
adenine and added UMP, while the selectivity of adenine-conjugate toward UMP in respect to other
nucleotides is most likely the consequence of additional hydrogen bonding between UMP and adenine.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
linkers connecting aromatic subunits.6 Previously prepared bis-
phenanthridinium compounds have shown at the time the high-
Efficient synthetic receptors with the capability of selective
substrate binding in aqueous solution are important for the un-
derstanding of molecular recognition and self-assembly in chem-
ical and biological systems.1 Detection of nucleosides and
nucleotides in aqueous medium has paramount importance as they
form the fundamental units of all the life forms. However, differ-
entiation among naturally occurring nucleobases based on differ-
ent hydrogen bonding patterns within the artificial receptor is
strongly limited due to competitive hydrogen bonding of water;2
therefore among many artificial receptors reported, most of them
lacked base selectivity. Actually, till now there are only a few re-
ceptors able to selectively bind some of nucleobases in water.
Lhomme et al. showed the capacity of aryl-nucleobase conjugates
to recognize certain nucleobases in water,3 while Kimura et al.
demonstrated that zinc(II) complexes of the macrocyclic tetra-
amine 1,4,7,10-tetraazacyclododecane (cyclen) have a unique pro-
pensity to bind with deprotonated imides like thymine and uracil,
by forming non-covalent stable complexes in biologically relevant
conditions.4 Moreover, cyclens appended with aromatic rings such
as acridine and ditopic receptors yielded binding constants for TMP
and UMP up to K¼107 Mꢀ1.5 Furthermore, some cyclo-bis-aromatic
derivatives revealed selectivity toward certain nucleobases or
basepairs due to the selective interactions of nucleobases with the
est affinity toward nucleosides and nucleotides but not selectivity
among studied nucleobases.7,8 Intriguingly, comparison of the
binding constants of monomer (order of magnitude Ks¼102 Mꢀ1
)
with calculated binding constants of bis-phenanthridinium analogs
(order of magnitude Ks¼106 Mꢀ1) revealed that not only simulta-
neous involvement of two monomeric units in complex formation
was present, which should give Ksz104 Mꢀ1, but also their coop-
erativity in binding. The difference between the expected
Ksz104 Mꢀ1 and the obtained Ksz106 Mꢀ1 could be a consequence
of hydrophobic effects (both, entropy- and enthalpy-driven),9
pre-organisation of bis-phenanthridinium analogs suitable for
nucleobase insertion (template effect),10 as well as of the other
interactions yielding significant template effect. Furthermore,
previously reported phenanthridinium–nucleobase conjugates
were not able to differentiate among selected nucleotides in
aqueous medium, most likely due to the strong competition of bulk
water with the expected hydrogen bonds between complementary
nucleotide and nucleobase attached to the intercalator.11–13 How-
ever, the same phenanthridinium–nucleobase conjugates inter-
acted highly selectively with complementary polynucleotide
sequences, most likely due to the polynucleotide hydrophobic
environment, which allowed the formation of specific hydrogen
bonds between nucleobase attached to intercalator and nucleo-
bases of polynucleotide.14,15
The aforementioned results suggested that the nucleobase
positioned within the hydrophobic cavity could recognize a com-
plementary nucleotide by hydrogen bonding. To achieve both high
* Corresponding authors. Tel.: þ385 1 45 71 210; fax: þ385 1 46 80 195.
´
E-mail addresses: pianta@irb.hr, tomic@irb.hr (S. Tomic).
0040-4020/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2010.01.063

2502
L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
Scheme 1. The asymmetric (4) or symmetric (3, 5) alkylation of tosylamino substituents of 2 by mono- and dibromopropane; (a) TsCl/pyridine/40–50 ꢁC; (b1) POCl3/120 ꢁC (b2)
NaOH/H2O; (c) Br(CH2)2CH3 (10 equiv)/K2CO3/DMF/Ar/rt (d1) Br(CH2)2CH3 (1,5 equiv)/K2CO3/DMF/Ar/rt (d2) Br(CH2)3Br (10 equiv)/K2CO3/DMF/Ar/rt.
stability and selectivity we used the bis-phenanthridinium skele-
ton, which, as previously reported,8 forms highly stable complexes
with nucleotides by aromatic stacking interactions, to which we
covalently attached various nucleobases. Linkers between phe-
nanthridinium units and between aromatic units were chosen to
allow insertion of nucleobase between two phenanthridinium
subunits, forming in this way a possible recognition site for
complementary nucleotides within a hydrophobic cleft and addi-
tionally stabilising the targeted basepair by aromatic stacking
interactions.
2. Results and discussion
2.1. Synthesis
Since it was not possible to covalently link nucleobases directly
to the phenanthridine of previously studied bis-phenanthridinium
derivatives,7,16 novel synthetic strategy for building the
bis-phenanthridinium skeleton had to be developed. The general
strategy that was used for the synthesis of the novel
bis-phenanthridinium-nucleobase conjugates 10–12 and reference

L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
2503
Scheme 2. Synthesis of conjugates 6–8. (a) NaH/DMF/Ar/40–50 ꢁC.
compound 9 comprised the asymmetric or symmetric alkylation of
the amino substituents of bis-phenanthridine 2 by mono- and
dibromopropane (Scheme 1), followed by the introduction of
nucleobase at the other end of one or both alkyl linkers (Scheme 2),
and subsequent deprotection of tosylated compounds (Scheme 3).
Compound 1 was prepared starting from N,N0-bis-[(40-amino)-2-
biphenylyl]-octanediamide16 that was tosylated in pyridine. The
bis-phenanthridine 2 was obtained by the Morgan–Walls re-
action17 based on the central pyridine ring formation by intra-
molecular electrophilic cyclisation of the bis-biphenylyl 1 using
POCl3. Then, bis-phenanthridine 2 was alkylated with a large excess
of mono-bromopropane to give symmetric alkylaminobis-
phenanthridine 3. To get the asymmetric product 4, one of two
tosylamino groups of 2 was alkylated in the first reaction step over
7 days in the dark and at room temperature, using a small excess
(1.5 equiv) of 1-bromopropane. Subsequently, a large excess of
potassium carbonate and 1,3-dibromopropane were added drop-
wise in situ, in order to obtain asymmetric compound 4, while
symmetric compound 5 was obtained as a side product after
purification by TLC (Scheme 1).
A pronounced hypochromic effect of 9–12 absorption maxima in
comparison to the monomer compound Ph–C3 (aH, Table 1) is most
likely the consequence of intramolecular aromatic stacking in-
teractions. Furthermore, comparatively weak fluorescence of ref-
erence compound 9 is most likely caused by strong intramolecular
aromatic stacking between phenanthridinium subunits (Table 1),
while significantly stronger fluorescence of bis-phenanthridinium-
nucleobase conjugates 10–12 (in comparison to 9) could be the
result of intramolecular un-stacking of phenanthridinium subunits
caused by at least partial nucleobase insertion.
2.3. Interactions with nucleotides
Interactions of compounds 9–12 with nucleotides in aqueous
medium were studied by UV–vis and fluorimetric titrations. Due to
the low solubility of 9–12, UV–vis titrations were performed using
an immersion probe with 5 cm light path length, which allowed
measurements at concentration range of 10ꢀ6 mol dmꢀ3, thus at
experimental conditions comparable to fluorimetric titrations. It
should be noted that UV–vis spectra were collected in the range
The reaction of bromo-derivatives 4 and 5 with a large excess of
uracil or adenine was performed under argon atmosphere at 40–
50 ꢁC in dry DMF in the presence of NaH, giving compounds 6–8.
Under these conditions the alkylation of uracil selectively occurred
at the N1 position, while adenine was selectively alkylated at the
N9 position (Scheme 2).
Tosyl-groups were removed by heating at 100 ꢁC under acidic
conditions, followed by neutralization using 5 M NaOH aqueous
solution (Scheme 3). Compounds 9–12 were found to be sufficiently
soluble in water under acidic conditions (pH 5).
l
¼260–300 nm, at which both, 9–12 and also nucleotides absorb
light, therefore for the processing of the titration-induced changes
in complete spectral range multivariate analysis program was
necessary (we applied Specfit).19 It should be stressed that at
l>290 nm adenine and uracil (in contrast to guanine and cytosine)
do not have an UV–vis spectra (for UV–vis spectra of nucleotides
see Supplementary data) and therefore titration-induced changes
in this part of 9–12 UV–vis spectra can be attributed only to the
changes in the absorption properties of phenanthridinium chro-
mophore. Titration with AMP and GMP yielded significantly
stronger changes in the UV–vis spectrum (l>290 nm) of the ref-
2.2. Spectroscopy
erence compound 9 in comparison to effects induced by UMP and
CMP, most likely due to the larger aromatic surface of purine
nucleobases in comparison to pyrimidines and consequently more
efficient aromatic stacking interactions.
Intriguingly, titration with UMP induced a significantly stronger
hypochromic effect in the UV–vis spectra (at l>290 nm) of adenine
conjugates 11 and 12, if compared to reference compound 9 and
uracil-conjugate 10 (Fig. 1). Even more interesting is the observa-
tion that titration with AMP induced a clear hyperchromic effect in
The UV–vis spectra of compounds 9–12 are strongly pH
dependent, exhibiting a one-step change at pKaz6, which was
attributed to the protonation of the phenanthridine heterocyclic
nitrogen.11,18 Due to the poor solubility of examined compounds in
neutral and basic conditions, all further measurements were per-
formed at pH¼5.0, with more than 90% of all compounds being in
the protonated (phenanthridinium) form. The absorbance of com-
pounds 9–12 was linearly dependent on the concentration within
the c¼1ꢂ10ꢀ6–4ꢂ10ꢀ5 mol dmꢀ3 range (Table 1), while at higher
concentrations aggregation of chromophores, as well as some
precipitation occurred. The compounds 9–12 exhibited fluores-
cence emission (Table 1) proportional to the concentration of
compound up to c¼5ꢂ10ꢀ6 mol dmꢀ3. Excitation spectra moni-
tored at emission maxima agree well with the corresponding
UV–vis spectra.
the UV–vis spectra (at l>290 nm) of adenine conjugates 11 and 12
(see, for example, Fig. 2), demonstrating that electronic absorption
properties of phenanthridinium chromophores of 11 and 12 are
significantly different upon complexation of UMP and AMP.
Changes in the UV–vis spectra at
l>290 nm of 9–12 upon
titration with GMP and CMP were less informative due to the
partial masking of changes by the intrinsic UV–vis spectra of
nucleotides.

2504
L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
Scheme 3. Deprotection of compounds 3 and 6–8. (a1) H2SO4/CH3CO2H/80–100 ꢁC (a2) NaOH/H2O.
However, fluorimetric titrations (Fig. 3) yielded more pro-
nounced spectroscopic changes than UV–vis titrations and there-
fore binding constants (Ks) and stoichiometries of the complexes
determined upon processing the titration data by Specfit19 program
are more accurate than those calculated from UV–vis titrations.
Nevertheless, both methods yielded quite comparable Ks values
and for all titrations the best fit was obtained for stoichiometry 9–
12/nucleotide¼1:1 (Table 2).
Table 1
Molar extinction coefficients and absorption maxima of 9–12 and monomer com-
pounds cPh–C3, cUr–C3, cAd–C3, hypochromic effects (H)a of 9–12 in respect to
monomer compounds. Fluorescence emission intensities at emission maxima of
compounds 9–12
UV–vis
Fluorescence
lmax/nm
3
(mmolꢀ1 cm2)
9269
aH (%)
lem/nm
Iib/I2 (550 nm)
9
10
11
12
269
275
270
273
277
268
262
85
62
75
55
d
d
d
552
565
560
558
547
d
1
19
14
31
,
d
25,497
16,436
30,526
29,282
9841
The binding constants (Table 2) obtained for reference com-
pound 9 and all studied nucleotides are comparable with those of
previously studied phenanthridinium-based bis-intercalands and
cyclo-bis-intercalands.7 It should be stressed that monomer Ph–C3
binds nucleotides predominantly by aromatic stacking interactions
yielding log Ksz2. Since compound 9 consists of two Ph–C3 sub-
units linked by an inert aliphatic chain, if aromatic stacking
interactions would be dominant in 9/nucleotide complexes, the
values of Ks (9/nucleotide)zKs (Ph–C3/nucleotide),2 which are
actually not the case (Table 2); the obtained values of Ks (2/nu-
cleotide) are more than two orders of magnitude higher, suggesting
the presence of a significant template effect.7 The affinity of bis-
cPh–C3
cUr–C3
cAd–C3
c d
13,733
d
d
a
(Na citrate/HCl buffer, pH¼5.0, I¼0.03 mol dmꢀ3),
H
(hypochromic
effect)¼{[2ꢂ3277nm (Ph–C3)þnꢂ3277nm (Ur–C3 or Ad–C3)ꢀ3277nm (9–12)]/[2ꢂ3277nm
(Ph–C3)þnꢂ3277nm (Ur–C3 or Ad–C3)]}ꢂ100; n¼0 for compound 9, n¼1 for com-
pounds 10, 11; n¼2 for 12.
b
For all compounds c¼2.2ꢂ10ꢀ6 mol dmꢀ3
lexc¼270 nm; relative intensities
,
calculated at
l¼550 nm taking 9 as a reference.
c
Published results.11
Not possible to compare due to different experimental conditions.
d

L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
2505
1.0
0.9
0.8
0.7
0.6
0.5
0.4
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
9
10
11
12
0.0
5.0x10
1.0x10
c (UMP) / mol dm
260
270
280
290
300
/ nm
Figure 1. UV–vis titration of 12 with UMP (c(12)¼2ꢂ10ꢀ6 mol dmꢀ3; c(UMP)¼1.2ꢂ10ꢀ6–12ꢂ10ꢀ6 mol dmꢀ3); Inset: changes in the UV–vis spectra (at
(c¼2ꢂ10ꢀ6 mol dmꢀ3) upon titration with UMP, done at pH¼5.0 (Na cacodylate/HCl buffer, I¼0.05 mol dmꢀ3).
l
¼290 nm) of 9–12
phenanthridinium-nucleobase conjugates 10–12 toward most of
the studied nucleotides is comparable to the reference compound 9
affinities. That is also pointing toward a significant template effect
in respect to previously studied phenanthridinium-nucleobase
conjugates,11,12 as well as phenanthridinium-bis-nucleobase
conjugates.13
Most intriguingly, the adenine-conjugate 11 binds comple-
mentary nucleotide UMP with the binding constant (Ks 11/UMP)
an order of magnitude higher than any of the binding constants
2.4. Discussion of results of spectrophotometric titrations
The UV–vis spectrum of an aromatic moiety usually reveals
a hypochromic effect upon stacking with another aromatic
p–p
system, although the intensity of hypochromic effect is not directly
related to the binding constant. Therefore, the hypochromic effect
of 9–12 compared to monomer Ph–3 can be explained by intra-
molecular aromatic stacking of two phenanthridinium units,
accompanied by additional stacking of covalently linked nucleo-
base (only for 10–12). However, the fact that phenanthridinium
obtained for the reference compound
9 (Ks 9/nucleotide).
Moreover, the affinity of 11 toward UMP is significantly higher
than affinity of 11 toward other nucleotide mono-phosphates
(AMP, GMP, CMP, Table 2, Fig. 3B). Such significantly stronger
affinity points toward additional interactions between 11 and
UMP (not present in the case of other nucleotide mono-
phosphates).
chromophores of adenine conjugates 11 and 12 at l>290 nm
revealed a much stronger hypochromic effect upon UMP titration in
comparison to reference 9 and uracil-conjugate 10 (Fig. 1) sug-
gested more efficient overlapping of aromatic surfaces in the case of
11/UMP and 12/UMP complexes, whereby one of the possible ex-
planations is formation of an adenine–UMP basepair within the
lypophilic pocket between two phenanthridinium subunits.
Moreover, such adenine–UMP basepair interactions within 11/UMP
and 12/UMP complexes could be correlated to the observed op-
posite changes (hyperchromic effect) in the UV–vis spectra of ad-
1,3
1,2
1,1
1,0
0,9
enine conjugates 11 and 12 at l>290 nm upon titration with UMP
and AMP (Fig. 2). Namely, the freedom of orientation of covalently
bound adenine between two phenanthridinium subunits is very
limited and basepair formation with AMP is hard to imagine.
Moreover, the surface of such an adenine–adenine basepair would
exceed the surface of phenanthridinium and therefore could not
effectively yield better overlapping of aromatic surfaces in com-
parison to an uracil–adenine basepair. Therefore, it is most likely
that AMP and covalently bound adenine compete for the binding
sites within 11 and 12, yielding as a final result a hyperchromic
0,8
AMP
UMP
0,7
0,6
0,5
0,4
0,3
0,2
0,1
effect at
l>290 nm (UV–vis range of phenanthridinium
chromophores).
An order of magnitude higher binding constant of 11/UMP
complex in comparison to any other 11/nucleotide complex or 9/
nucleotide complex (Table 2) is also in line with the proposed ad-
enine–UMP basepair formation. Assuming that hydrogen bonding
is contributing to the selectivity of 11 toward UMP, the adenine of
11 should be positioned in a hydrophobic surrounding (e.g., be-
tween phenanthridinium subunits), within which water molecules
are mostly excluded. Otherwise, competition of the extremely high
0,0
2,0x10 4,0x10 6,0x10 8,0x10 1,0x10 1,2x10 1,4x10
c (nucleotide) / mol dm
Figure 2. Changes in the UV–vis spectrum (at
l
>290 nm) of 11 (c¼2ꢂ10ꢀ6 mol dmꢀ3
)
upon titration with UMP (-) and AMP (C), done at pH¼5.0 (Na cacodylate/HCl buffer,
I¼0.05 mol dmꢀ3).

2506
L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
300
250
200
150
100
50
have investigated the possible conformations of such complexes by
molecular modeling studies.
2.5. Molecular modeling
All studied molecules were prepared in both; extended and
maximally folded shape with rings stacked conformations, sol-
vated, energy optimised, and subjected to MD simulations (for
details of MD simulations see section ‘4.3. Molecular modeling’).
During the MD simulations the extended conformations folded and
the stacked ones slightly unfolded. However the majority of mol-
ecules retained their folded (more or less stacking conformation),
with no water molecules accommodated within the two phenan-
thridinium units (Fig. 4). Obtained structures are in accordance
with the pronounced hypochromic effect of 9–12 absorption
maxima (Table 1) in comparison to that of the reference compound
Ph–C3, whereby the strongest hypochromic effect of 9 (if compared
to nucleobase conjugates 10–12) supports the insertion of a nucle-
obase between phenanthridinium subunits (as shown in Fig. 4).
Apparently, the impact of multiple aromatic stacking interactions
on the hypochromicity in the UV–vis spectrum is significant for the
studied bis-phenanthridinium skeleton. The stacking interaction is
the most efficient between two phenanthridinium subunits (9),
insertion of another aromatic moiety decreases the stacking in-
teraction intensity, whereby the effect of uracil, i.e., smaller aro-
matic moiety (10) insertion is in comparison with stacking of the
adenine (11, 12) being less favorable. Although the fluorescence of
small molecules in water is a complex phenomenon and often
cannot be directly correlated to structural properties, it is intriguing
that the intensity of fluorescence emission of phenanthridinium
units of all studied nucleobase conjugates is significantly stronger
than that of compound 9 (Table 1), also supporting intramolecular
interactions of nucleobases with fluorescence emitting
chromophores.
500
520
540
560
580
600
620
640
/ nm
100
80
60
40
20
0
UMP
AMP
GMP
CMP
0,0
Since the conformations presented in Figure 4 resemble to the
molecular shape of a hydrophobic cavity in which there are no
water molecules, we considered them to be excellent starting
points for further modeling studies of the non-covalent complexes
with nucleoside mono-phosphates AMP and UMP.
c (nucleotide) / mol dm-3
Figure 3. (A) Fluorimetric titration of 11 (c¼2ꢂ10ꢀ6 mol dmꢀ3
) with UMP; (B)
percentage of formed 11/nucleotide complex calculated by Specfit.19
In the initial conformation of the 9–AMP complex, used in MD
simulations, adenine was inserted between two phenanthridinium
units in a similar manner as obtained for covalently bound adenine-
conjugate 11 during MD simulations (Fig. 4). Complexes 11–UMP
and 11–AMP were built in a way to enable adenine from 11 and base
from mono-phosphate to form hydrogen bonds (Figs. 6A and 7A).
The complexes were solvated in water and the systems were
geometry optimised and subjected to molecular dynamics
simulations for 8.5 ns. The initial orientations of the bis-
excess of bulk water would not allow formation of hydrogen bonds
between nucleobases, as previously noted for phenanthridinium–
nucleobase conjugates.11,12
Since the aforementioned UV–vis and fluorimetric titrations
cannot directly prove proposed adenine–UMP basepair formation,
and low solubility of 9–12 hampered detailed studies by structur-
ally more specific methods (NMR, crystallographic studies), we
Table 2
Binding constants log Ks calculated from fluorimetric titrations and UV–vis titrations (in brackets) for 9–12/nucleotide complexesa,b
b
9
10
11
12
Ph–C3
AMP
5–6c
5.75ꢃ0.03
6.34ꢃ0.06
5.63ꢃ0.06
1.73ꢃ0.02
(nd)
1.78ꢃ0.03
2.29ꢃ0.02
1.72ꢃ0.09
(nd)
(5.60ꢃ0.11)
(z5–6)c
(z5–6)c
(z5–6)c
ADP
ATP
GMP
>6c
6.19ꢃ0.17
6.14ꢃ0.05
5.97ꢃ0.03
(z5–6)c
6.21ꢃ0.23
6.91ꢃ0.23
6.42ꢃ0.16
6.63ꢃ0.19
>6c
5–6c
5.45ꢃ0.04
5.55ꢃ0.08
d
d
(5.73ꢃ0.15)
CMP
UMP
5–6c
5.69ꢃ0.04
5.24ꢃ0.04
5.48ꢃ0.07
(5–6)c
1.93ꢃ0.08
(5–6)c
5–6c
(nd)
d
d
6.11ꢃ0.04
6.89ꢃ0.11
5.86ꢃ0.09
1.59ꢃ0.09
(nd)
(5–6)c
(5–6)c
(6.23ꢃ0.15)
(5–6)c
a
Titrations done at pH¼5 (Na cacodylate/HCl buffer, I¼0.05 mol dmꢀ3) and log Ks values are given for stoichiometry 9–12/nucleotide¼1:1.
AMP2ꢀ¼adenosine mono-phosphate; GMP2ꢀ¼guanosine mono-phosphate; CMP2ꢀ¼cytidine mono-phosphate; UMP2ꢀ¼uridine mono-phosphate.
Due to the small spectroscopic changes less than 10 data points were collected, allowing only estimation of binding constant.
b
c
d
Small spectroscopic changes of complex compared to ligand and nucleotide resulted in linear change of absorbance, which hampered even estimation of the binding
constant.

L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
2507
Figure 4. Conformations of studied compounds obtained by MD simulations.
phenanthridinium conjugates nucleobase and the nucleosides
did not change significantly during the optimisation and the
Watson–Crick (W–C) type of interaction was retained in 11–UMP
complex (Fig. 6A). However, during the MD simulation the con-
formations of the complexes changed (see, for example, Fig. 5). In
comparison with the initial, optimized structures the simulation
yielded less organized structure of 9–UMP complex. The differ-
ence between the final structures of 9–AMP and 9–UMP com-
plexes (Supplementary data) obtained upon MD simulation,
clearly point out the importance of the size of aromatic part of
nucleobase, whereby only larger purine nucleobase was able to
form stable complex by insertion between the phenanthridinium
subunits.
The 11–UMP complex reorganized into the more compact form
(Fig. 6B) stabilized by two intermolecular stacking inter-
actionsdface to face and face to edge between uracil and two Phen.
Unit and one intramolecular stacking interaction (Phen. Uni-
tdadenine). The hydrophobic pocket outlined by two perpendic-
ularly oriented phenanthridinium units and the alkyl linker
prevented water molecules to compete with uracil from UMP in
forming two intermolecular hydrogen bonds: one with adenine
and the other with phenanthridinium subunit. The potential en-
ergy of the final system is about 7% lower than that of the initial.
Furthermore, stabilization due to solvation effects also increased
during the MD simulation: the non-polar solvent-accessible surface
area decreased for about 22% and the polar solvent-accessible
surface area increased for about 6%.
The 11–AMP complex (Fig. 7A) also reorganized into the more
stable conformation during MD simulations (Fig. 7B). However the
stabilisation due to the solvation effects is insignificant, i.e., the
non-polar solvent-accessible surface area of the complex decreased
by only about 8% and the polar solvent-accessible surface area
decreased by about 4%.
The overall shape of the final 11–AMP complex (Fig. 7B) is less
compact (as can be seen from decrease of solvent-accessible surface
area) and the overlapping of aromatic units is less pronounced in
comparison to the 11–UMP complex (Fig. 6B). The latter property
could be correlated to the opposite changes in UV–vis titration
experiments (Fig. 2); namely three aromatic stacking interactions
between UMP and 11 could yield hypochromic effect with respect
to free 11, while less pronounced aromatic overlapping caused by
addition of AMP to 11 could result in the hyperchromic effect.
3. Conclusions
The bis-phenanthridinium–adenine derivative 11 successfully
Figure 5. Conformation of the 11–UMP complex significantly changed during MD
simulation in water.
combined the high affinity of previously known bis-intercalands7

2508
L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
Figure 6. The starting conformation of 11–UMP complex with the Watson–Crick type of H-bonds (A) changed to the conformation (B) in which the complex is stabilized by three
intermolecular stacking interactions ( ) and two intermolecular H-bonds ( ).
toward nucleobases with selectivity toward complementary nu-
cleotide (UMP). Molecular modeling studies suggest that selectivity
of 11 toward UMP with respect to other nucleotides is most likely
a consequence of organization of the 11–UMP complex in the
compact form stabilized by efficient intra- and intermolecular
stacking interactions (as shown by hypochromic effect in UV–vis
titration) as well as by intermolecular hydrogen bonds between
uracil and 11.
Other bis-phenanthridinium–nucleobase derivatives (10 and
12) were not able to distinguish between studied nucleotides sig-
nificantly. The MD simulations of uracil-conjugate 10 as well as 9–
UMP complex suggest that uracil due to the small aromatic surface
was not able to form a stable conformation in which it would si-
multaneously form stacking interactions with both phenan-
thridinium subunits and therefore failed to induce formation of the
hydrophobic cavity necessary for hydrogen bonding recognition of
nucleotides. On the other hand, two adenines attached to derivative
12 could compete with any nucleotide added, thus lowering the
binding constant value.
and double stranded DNA/RNA sequences highly promising,
whereby selectivity of 11 toward UMP could be even more pro-
nounced in a case of more hydrophobic poly U. In addition, other
bis-phenanthridinium–nucleobase derivatives could also reveal
selective affinity and/or spectroscopic sensing toward comple-
mentary DNA/RNA sequences. Furthermore, all studied compounds
and especially derivative 9 are expected to show high affinity to-
ward ds-DNA, and consequently pronounced biological activity,
like many other bis-aromatic compounds.20,21
4. Experimental
4.1. General procedures
1H and 13C NMR spectra were recorded on Bruker Avance
DRX 500 operating at 500 MHz. Chemical shifts
(d) are
expressed in ppm, and J values in Hz. Signal multiplicities are
denoted as s (singlet), d (doublet), t (triplet), ps. t. (pseudo
triplet), q (quartet) and m (multiplet). The electronic absorption
spectra of newly prepared compounds were measured on
a Varian Cary 100 Bio spectrometer in quartz cuvettes (1 cm and
Finally, the high affinity of novel compounds 9–12 toward nu-
cleotides makes studies of their interactions with single stranded

L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
2509
Figure 7. The starting conformation of 11–AMP complex (A) changed to the conformation (B), in which covalently attached adenine was displaced by AMP from the cavity between
two Phen. units.
10 cm). UV–vis titration were performed on a Varian Cary 100
Bio spectrometer and also on Varian Cary 50 using immersion
probe with 5 cm light path length. IR spectra were recorded on
a Perkin–Elmer 297 instrument using KBr pellets. Fluorescence
spectra were recorded on Varian Cary Eclipse fluorimeter. Mass
spectra were obtained using Applied Biosystems 4800 Plus
MALDI TOF/TOFÔ Analyzer. Preparative thin layer chromatog-
raphy (TLC) was carried out using Kieselgel HF254 ‘Merck’.
Melting points were determined on Kofler apparatus and are
uncorrected. All products were characterized by NMR, IR, ESI-
MS or HRMS. Hygroscopic character of compounds yielded el-
emental analyses with non-stoichiometric amounts of water-
dhowever, since NMR spectra of final compounds were in
accordance with other, previously prepared close analogs,22
proposed structures are not questionable.
4.2. UV–vis and fluorescence measurements
Nucleotides were purchased from Sigma and Aldrich, and used
without further purification. The measurements were performed in
aqueous buffer solution (pH¼5, I¼0.05 mol dmꢀ3, sodium cacody-
late/HCl buffer). Under the experimental conditions used (con-
centration of compounds 9–12z10ꢀ6 mol dmꢀ3) the absorbance
and fluorescence intensities of 9–12 were proportional to their
concentrations. Spectroscopic titrations were performed at con-
stant ionic strength (buffer, I¼0.05 mol dmꢀ3) by adding portions of
nucleotide solution into solution of the tested compound. Obtained
data were corrected for dilution. UV–vis titrations were performed
using immersion probe with 5 cm light path length, which allowed
measurements at concentration range of 10ꢀ6 mol dmꢀ3, thus at
experimental conditions comparable to fluorimetric titrations. It

2510
L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
should be noted that UV–vis spectra were collected in the range
¼260–300 nm, at which both, 9–12 and also nucleotides absorb
yellow solid precipitated. Recrystallization from methanol gave
white solid 1 (780 mg, 70% yield). Rf (SiO2, 5% MeOH in
l
light, therefore for the processing of the titration-induced changes
in complete spectral range multivariate analysis program was
necessary. In fluorimetric titrations, excitation wavelengths at
lmax¼320 nm were used in order to avoid absorption of excitation
light by added nucleotides and changes in emission at maxima
were monitored. The binding constants and stoichiometries of
complexes of 9–12 with nucleotides were calculated for the con-
centration range corresponding to ca. 20–80% complexation by
non-linear least-square fitting program SPECFIT.19
CH2Cl2)¼0.31; mp 110–112 ꢁC; 1H NMR (DMSO-d6)
d: 1.17 (br s, CH2,
4H), 1.44 (br s, CH2, 4H), 2.10 (t, CH2, 4H, J¼6.8 Hz), 2.32 (s, Ts–CH3,
6H), 7.11 (d, Ts, 4H, J¼8.6 Hz), 7.20–7.22 (m, Ar-H, 7H), 7.27–7.29 (m,
Ar-H, 2H), 7.33–7.34 (m, Ar-H, 5H), 7.39–7.41 (m, Ar-H, 2H), 7.68–
7.69 (m, Ar-H, 4H), 9.11 (s, NH–CO, 2H), 10.36 (s, NH–Ts, 2H); 13C
NMR (DMSO-d6)
125.95, 126.85, 127.35, 127.63, 129.65, 129.84, 130.13, 134.49, 134.99,
136.15, 137.08, 137.16, 143.37, 171.61; IR (KBr) : 3464, 3246, 2924,
d: 21.09 (Ts–CH3), 25.07, 28.57, 35.74, 119.42,
n
2853, 2366, 2345, 1647, 1524, 1508, 1458, 1445, 1385, 1339, 1325,
1227, 1157, 1092, 924, 841, 814, 764, 658, 573, 546 cmꢀ1. Anal. Calcd
for C46H46N4O6S2 (Mr 815.03 g molꢀ1): C 67.72, H 5.96, N 6.87%.
Found: C 67.36, H 5.48, N 6.70%.
4.3. Molecular modeling
Molecules were built using the module ‘Builder’ within the
program InsightII,23 and using the option ‘Modify Torsion’ the
stacking conformation was prepared for each of the molecule.
The crystal structure of AMP was separated from crystal structure of
complex with PDB-id code 1Z6S. UMP was constructed using the
crystal structure of AMP by replacing A with U. The replacement
was done using the module ‘Biopolymer’ within the program
InsightII. The AMBER ff03 force field of Duan et al.24 and the general
AMBER force field GAFF were used to obtain parameters for the bis-
phenanthridinium conjugates, nucleoside mono-phosphates, and
water molecules. The tLeap module of AMBER 9 was used to obtain
topology and coordinate files for molecules and complexes. Each
molecule was placed in the center of an octahedron that was filled
with TIP3P type water molecules; the water buffer of 8 Å was used.
Besides water molecules, Clꢀ ions were added to neutralize the
system when necessary Geometry optimization and molecular
dynamics (MD) simulations were accomplished using the AMBER 9
program package.25 The simulation was accomplished using Peri-
odic Boundary Conditions (PBC). The Particle Mesh Ewald (PME)
method was used for calculation of electrostatic interactions. In the
direct space the pairwise interactions were calculated within the
cutoff-distance of 11 Å. Before molecular dynamics (MD) simula-
tions, the system was optimized using steepest descent and con-
jugate gradient methods, 1500 steps of each. After energy
minimization, the system was equilibrated during 10 ps. During
equilibration, the temperature was linearly increasing from 0 to
300 K and the volume was held constant. The equilibrated system
was then subjected to at least 8.5 ns (UMP-11 and AMP-11, 13.5 ns
AMP-9) of productive unconstrained molecular dynamics simula-
tion at constant temperature and volume (300 K). The time step
during the simulation was 1 fs and temperature was held constant
4.4.2. 1,6-Bis-(8-tosylaminophenanthridine-6-yl)-hexane
(2). Compound 2 was obtained by suspending N,N0-bis-[(40-tosy-
lamino)-2-biphenylyl]-suberamide 1 (2 g, 2.45 mmol) in 8 ml POCl3
and heating the reaction mixture at 100–110 ꢁC during 3 h. Mixture
was allowed to cool and poured into ice, and afterward was made
alkaline (pH¼8–9) by addition of 3 M NaOH water solution. Yellow
solid precipitated and was filtered and washed with water to give
pale yellow powder (1.8 g, 94% yield); Rf (SiO2, 10% MeOH in
CH2Cl2)¼0.52; mp 269–271 ꢁC; 1H NMR (DMSO-d6)
d: 1.52 (br s,
CH2, 4H), 1.78 (br s, CH2, 4H), 2.17 (s, Ts–CH3, 6H), 3.18 (t, CH2, 4H,
J¼7.7 Hz), 7.22 (d, Ts, 4H, J¼8.2 Hz), 7.59–7.66 (m, Phen-2, Phen-3,
Phen-4, 6H), 7.68 (d, Ts, 4H), 7.93–7.95 (m, Phen-7, Phen-9, 4H),
8.58 (dd, Phen-1, 2H, J1–3¼0.9 Hz, J1–2¼8.2 Hz), 8.70 (d, Phen-10,
2H, J9–10¼9.0 Hz), 10.68 (s, NH–Ts, 2H) ppm; 13C NMR (DMSO-d6)
d: 20.80 (Ts–CH3), 28.46, 29.16, 35.48, 114.67, 122.15, 122.90, 123.68,
124.34, 125.09, 126.63, 126.71, 128.21, 129.08, 129.65, 129.97, 136.63,
142.58, 143.29, 160.85, 174.64 ppm; IR (KBr) : 3275, 3067, 2934,
2858, 2363, 2345, 1618, 1576, 1535, 1491, 1448, 1389, 1348, 1242,
1161, 1092, 953, 895, 814, 762, 669, 575, 544, 473, 459 cmꢀ1
n
;
(MALDI/TOF-HRMS) m/z: 779.2695 (calcd for C46H42N4O4S2:
779.2720).
4.4.3. 1,6-Bis-[8-(propyltosylamino)phenanthridine-6-yl]-hexane
(3). 1-Bromopropane (234 ml; 257 mmol; 20 equiv) and K2CO3
(266 mg; 1.93 mmol, 20 equiv) were suspended in dry DMF (10 ml).
To this suspension, a solution of 1,6-bis-(8-tosylaminophenan-
thridine-6-yl)-hexane (2) (100 mg; 0.128 mmol) in dry DMF (5 ml)
was added dropwise during 10 min and the reaction mixture was
stirred for 4 days under argon atmosphere at room temperature.
Water (70 ml) and CH2Cl2 (70 ml) were added to this suspension,
the water layer was washed twice with CH2Cl2, organic extracts
were dried over Na2SO4, and evaporated, yielding a brown oil. The
oily residue was triturated with water to give a light brown pre-
cipitate that was filtered (71 mg, 64%), washed with water and
dried, and used without further purification. Pure compound 2 was
obtained by TLC (SiO2, 2% MeOH in CH2Cl2, Rf¼0.54) as a white
solid, additionally recrystallized from MeOH; mp 187–189 ꢁC; 1H
using Langevin dynamics with a collision frequency of 1 psꢀ1
.
The trajectories were visualized using the VMD 1.8.6 program.
The RMSDs (root mean square deviations) between the initial
conformation and those obtained during the MD simulation were
calculated for each complex. The trajectories were divided into
several stages (consisting of subsequent conformations with simi-
lar RMSD), and for each of this stage the average conformation was
determined. The average conformations, as well as the final one,
were energy minimized using the same procedure as for the initial
one. The obtained conformations were visually compared using the
InsightII software.
NMR (CDCl3)
d: 0.93 (t, CH3, 6H, J¼7.3 Hz), 1.46–1.53 (m, CH2–
hexylene chain, CH2–propyl chain, 8H), 1.85 (br s, CH2–hexylene
chain, 4H), 2.41 (s, Ts–CH3, 6H); 3.21 (t, CH2–hexylene chain, 4H,
J¼7.6 Hz), 3.65 (t, NCH2, 4H, J¼7.0 Hz), 7.23 (d, Ts, 4H, J¼8.1 Hz), 7.45
(d, Ts, 4H), 7.54 (dd, Phen-9, 2H, J7–9¼1.9 Hz, J9–10¼8.7 Hz), 7.64 (t,
Phen-2, 2H), 7.73 (t, Phen-3, 2H), 7.87 (d, Phen-7, 2H), 8.12 (d, Phen-
4, 2H, J3–4¼8.0 Hz), 8.50 (d, Phen-1, 2H, J1–2¼8.0 Hz), 8.59 (d, Phen-
4.4. Synthesis of compounds
4.4.1 . N,N0-Bis-[(40-tosylamino)-2-biphenylyl]-octanediamide (1). A
solution of tosyl-chloride (1.5 g, 6.71 mmol) in 15 ml of pyridine
was added dropwise during 1 h to the ice-cold solution of N,N0-bis-
[(40-amino)-2-biphenylyl]-octanediamide16 (690 mg, 1.3 mmol) in
pyridine (15 ml). After addition was completed, the reaction mix-
ture was heated at 50–60 ꢁC during 4 h. Subsequently, the reaction
mixture was allowed to cool and then poured into water. A light
10, 2H) ppm; 13C NMR (CDCl3)
d: 9.93 (CH3), 20.47 (Ts–CH3), 20.53
(CH2–propyl chain), 28.06 (CH2–hexylene chain), 28.58 (CH2–hex-
ylene chain), 34.50 (CH2–hexylene chain), 51.18 (NCH2), 120.90
(Phen-10), 121.99, 123.39 (Phen-1), 125.05 (Phen-7), 125.50, 126.59
(Ts), 127.05, 127.87 (Phen-3), 128.28, 128.40 (Ts), 128.63 (Phen-4),
129.81, 134.01, 142.46 ppm; IR (KBr)
n: 3425, 3065, 2961, 2932,
2874, 2858, 2363, 2345, 1599, 1572, 1528, 1479, 1458, 1344, 1238,

L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
2511
1213, 1167, 1090, 1074, 1020, 964, 872, 812, 766, 725, 708, 667, 642,
582, 550 cmꢀ1. Anal. Calcd for C52H54N4O4S2 (Mr¼863.16): C 72.36,
H 6.31, N 6.49%. Found: C 72.05, H 6.22, N 6.54%; (MALDI/TOF-
HRMS) m/z: 863.3644 (calcd for C52H54N4O4S2: 836.3659).
phenanthridine-6-yl]-6-[8-(propyltosyl)aminophenanthridine-6-
yl]-hexane 4 (90 mg; 0.095 mmol) in dry DMF (10 ml) was added
dropwise and the reaction mixture was stirred during 48 h under
argon atmosphere at 50 ꢁC. Then, water (70 ml) and CH2Cl2 (70 ml)
were carefully added to this suspension. The water layer was washed
twice with CH2Cl2, organic extracts were dried over Na2SO4 and
evaporated, yielding oily residue that was triturated with water to
give 95 mg of white precipitate. The precipitate was filtered, washed
with water and dried; and then purified by TLC (SiO2, 10% MeOH in
CH2Cl2, Rf¼0.55). Compound 6 was obtained as white solid (35 mg,
4.4.4. 1-[8-(3-Bromopropyltosyl)aminophenanthridine-6-yl]-6-[8-
(propyltosyl)aminophenanthridine-6-yl]-hexane (4). 1-Bromopro-
pane (56 ml; 0.617 mmol; 1.6 equiv) and K2CO3 (133 mg;
0.964 mmol, 2.5 equiv) were suspended in dry DMF (10 ml). To this
suspension, a solution of 1,6-bis-(8-tosylaminophenantridine-6-
yl)-hexane (2) (300 mg; 0.386 mmol) in dry DMF (5 ml) was added
dropwise over 10 min and the reaction mixture was stirred for 7
days under argon atmosphere at room temperature. Then, 1,3-
37% yield); mp 200–203 ꢁC; 1H NMR (CDCl3)
d: 0.90 (t, CH3, 3H,
J¼7.8 Hz), 1.52 (m, CH2–hexylene chain, CH2–propyl chain, 6H), 1.84
(m, CH2–hexylene chain, CH2–propylene chain, 6H), 2.39 (s, Ts–CH3,
6H), 3.19–3.24 (m, CH2–hexylene chain, 4H) 3.62 (t, NTsCH2–propyl
chain, 2H, J¼6.7 Hz), 3,72 (t, NTsCH2–propylene chain, 2H, J¼6.3 Hz),
3.85 (t, uracil-NCH2–propylene chain, 2H, J¼6.70 Hz), 5.66 (d, uracil-
5, 1H, J5–6¼7.8 Hz), 7.21 (d, Ts, 4H, J¼8.1 Hz), 7.31 (d, uracil-6, 1H),
7.36–7.46 (m, Ts, Phen-9, 5H), 7.51 (dd, Phen-containing-base-9, 1H,
J7–9¼1.9 Hz, J9–10¼8.8 Hz), 7.59–7.74 (m, Phen-2, Phen-3, 4H), 7.83 (d,
Phen-containing-base-7, 1H), 7.94 (d, Phen-7, 1H), 8.10 (d, Phen-4,
2H, J3–4¼7.3 Hz), 8,47 (d, Phen-1, 2H, J1–2¼8.1 Hz), 8.54 (d, Phen-10,
1H, J9–10¼8.9 Hz), 8.58 (d, Phen-containing-base-10, 1H), 8.78 (s, U-
dibromopropane (525 ml, 5.14 mmol, 13 equiv) and K2CO3 (533 mg;
3.86 mmol, 10 equiv) were added to the reaction mixture, and
stirred for 2 days under argon atmosphere at room temperature.
Water (70 ml) and CH2Cl2 (70 ml) were added to this suspension,
the water layer was washed twice with CH2Cl2, organic extracts
were dried over Na2SO4, and evaporated, yielding brown oil. The
oily residue was triturated with water to give 826 mg of light brown
precipitate that was filtered (160 mg, 44%), washed with water and
dried. Pure compound 4 was obtained by TLC (SiO2, 2% MeOH in
CH2Cl2, Rf¼0.54) as white solid (160 mg, 44%); mp 198–200 ꢁC; 1H
NH, 1H) ppm; 13C NMR (CDCl3)
d: 11.23 (CH3), 21.73 (Ts–CH3), 21.75
NMR (CDCl3)
d: 0.92 (t, CH3, 3H, J¼7.4 Hz) 1.44–1.53 (m, CH2–hex-
(Ts–CH3), 21.81 (CH2–propyl chain), 27.41 (CH2–propylene chain),
29.21 (CH2–hexylene chain), 29.76 (CH2–hexylene chain), 36.05
(CH2–hexylene chain), 46.51 (uracil–NCH2–propylene chain), 47.69
(NTsCH2–propylene chain), 52.35 (NTsCH2–propyl chain), 102.27
(uracil-5), 114.76, 121.30, 122.22 (Phen-1), 122.23 (Phen-1), 123.1,
123.29, 123.7 (Phen-10), 124.1 (Phen-10), 125.69, 126.11, 126.85,
126.99, 127.84, 127.95, 129.44, 129.72, 129.87, 132.35, 132.70, 134.27,
135.27, 137.45, 143.8, 144.35, 145.24 (uracil-6), 150.74, 161.75,
ylene chain, CH2–propyl chain, 6H), 1.84 (br s, CH2–hexylene chain,
4H), 2.08 (m, CH2–propylene chain, 2H), 2.40 (s, Ts–CH3, 6H), 3.22
(t, CH2–hexylene chain, 4H, J¼6.2 Hz), 3.45 (t, CH2Br, 2H, J¼6.4 Hz),
3.64 (t, NCH2–propyl chain, 2H, J¼7.0 Hz), 3.83 (t, NCH2–propylene
chain, 2H, J¼6.6 Hz), 7.22 (d, Ts, 4H, J¼8.0 Hz), 7.43 (m, Ts, 4H), 7.51–
7.54 (m, Phen-9, 2H), 7.63 (m, Phen-2, 2H), 7.73 (t, Phen-3, 2H),
7.85–7.89 (m, Phen-7, 2H), 8.12 (d, Phen-4, 2H, J3–4¼7.8 Hz), 8.49 (d,
Phen-1, 2H, J1–2¼7.6 Hz), 8.56–8.60 (m, Phen-10, 2H) ppm; 13C NMR
163.51 ppm; IR (KBr)
1541, 1508, 1458, 1385, 1340, 1159, 1092, 812, 766, 723, 669, 584,
546 cmꢀ1
(MALDI/TOF-HRMS) m/z: 973.3742 (calcd for C56
n:3462, 2928, 2853, 2361, 2343, 1686, 1647,
(CDCl3) d: 11.03 (CH3), 21.52 (Ts–CH3), 21.62, 29.19, 29.25, 29.57,
29.85, 31.71, 49.26, 52.24, 76.79, 77.00, 77.21, 122.03, 123.02, 123.12,
123.54, 123.77, 125.30, 125.36, 125.96, 126.06, 126.25, 126.36,
126.90, 127.65, 127.72, 129.25, 129.51, 129.61, 132.23, 132.37, 134.53,
;
H56N6O6S2: 973.3776).
135.06, 138.17, 143.60, 143.92, 161.72, 161.82 ppm; IR (KBr)
n
: 3452,
4.4.7. 1-[8-(9-(Aden-1-yl)propyltosyl)aminophenanthridine-6-yl]-6-
[8-(propyltosyl)aminophenanthridine-6-il]hexane (7). Compound 7
was obtained as described for 6; 1-[8-(3-bromopropyltosyl)ami-
nophenanthridine-6-yl]-6-[8-(propyltosyl)aminophenanthridine-
6-yl]-hexane 4 (130 mg; 0.138 mmol), adenine (187 mg; 1.38 mmol,
10 equiv), and NaH (55 mg, 60% w/w, 1.38 mmol, 10 equiv) in dry
DMF (10þ10 ml) gave white powder 7 (50 mg, 36% yield), Rf (SiO2,
2926, 2854, 2363, 2345, 1684, 1647, 1541, 1508, 1340, 1163, 1090,
964, 812, 766, 669, 582, 548, 473 cmꢀ1; (MALDI/TOF-HRMS) m/z:
941.2756 (calcd for C52H53BrN4O4S2: 941.2764).
4.4.5. 1,6-Bis-[8-(3-bromopropyltosylamino)phenanthridine-6-yl]-
hexane (5). Compound (5) was obtained as a side product during
preparation of 4 as a white powder (30 mg, 8% yield), Rf (SiO2, 2%
10% MeOH in CH2Cl2)¼0.48; mp 184–186 ꢁC; 1H NMR (CDCl3)
d:
MeOH in CH2Cl2)¼0.25; mp 205–209 ꢁC; 1H NMR (CDCl3)
d
: 1.54 (br
0.88 (t, CH3, 3H, J¼7.3 Hz), 1.45 (ps q CH2–propyl chain, 2H), 1.51 (br
s, CH2–hexylene chain, 4H), 1.83 (br s, CH2–hexylene chain, 4H),
2.05 (br s, CH2–propylene chain, 2H), 2.36 (s, Ts–CH3, 3H), 2.37 (s,
Ts–CH3, 3H), 3.19 (m, CH2–hexylene chain, 4H), 3.61 (t, NTsCH2–
propyl chain, 2H, J¼7.0 Hz), 3.68 (t, NTsCH2–propylene chain, 2H,
J¼5.9 Hz), 4.30 (t, adenine–NCH2–propylene chain, 2H, J¼6.3 Hz),
5.80 (br s, adenine–NH2, 2H), 7.16–7.19 (m, Ts, 4H), 7.34 (d, Ts, 2H,
J¼7.9 Hz), 7.42 (d, Ts-Phen-containing-base-, 2H, J¼8.0 Hz), 7.45 (d,
Phen-containing-base-9, 1H, J9–10¼8.6 Hz), 7.48 (d, Phen-9, 1H, J9–
s, Phen-CH2, 4H), 1.85 (br s, CH2–hexylene chain, 4H), 2.07 (t, CH2,
2H, J¼6.4 Hz), 2.41 (s, Ts–CH3, 6H), 3.24 (br s, CH2–hexylene chain,
4H), 3.45 (t, Br–CH2, 4H, J¼6.9 Hz) 3.84 (t, NCH2, 4H, J¼6.9 Hz), 7.22
(d, Ts, 4H, J¼8.0 Hz), 7.42 (d, Ts, 4H), 7.53 (d, Phen-9, 2H,
J9–10¼8.8 Hz), 7.64 (m, Phen-2, 2H), 7.73 (m, Phen-3, 2H), 7.88
(s, Phen-7, 2H), 8.15 (br s, Phen-4, 2H), 8.49 (d, Phen-1, 2H, J1–
¼8.2 Hz), 8.58 (d, Phen-10, 2H) ppm; 13C NMR (CDCl3)
d: 21.53,
2
29.12, 29.59, 29.83, 31.76, 48.21, 49.29, 122.03, 123.01, 123.74,
125.45, 126.00, 126.77, 126.80, 127.74, 129.16, 129.60, 130.47, 132.33,
¼8.7 Hz), 7.57 (ps. t., Phen-containing-base 2, 1H), 7.60 (ps. t.,
10
134.67, 138.15, 143.89, 161.65 ppm; IR (KBr)
2926, 2854, 2365, 2345, 1717, 1653, 1541, 1458, 1346, 1242, 1163,
1092, 949, 812, 764, 725, 708, 667, 582, 548, 419, 397 cmꢀ1
n
: 3447, 3065, 3032,
Phen-2, 1H), 7.65 (ps. t., Phen-containing-base-3, 1H), 7.70 (ps. t.,
Phen-3, 1H), 7.82 (s, Phen-containing-base-7, 1H), 7.86 (s, adenine-
8, 1H), 7.89 (s, Phen-7, 1H), 8.06 (d, Phen-containing-base-4, 1H, J3–
;
(MALDI/TOF-HRMS) m/z: 1019.1891 (calcd for C52H52Br2N4O4S2:
1019.1869).
¼8.0 Hz), 8.08 (d, Phen-4, 1H, J3–4¼8.0 Hz), 8.21 (s, adenine-2, 1H),
4
8.42 (d, Phen-containing-base-1, 1H, J1–2¼8.0 Hz), 8.46 (d, Phen-1,
1H, J1–2¼8.1 Hz), 8.50 (d, Phen-containing-base-10, 1H), 8.55 (d,
4.4.6. 1-[8-(3-(Urac-1-il)propyltosyl)aminophenanthridine-6-yl]-6-[8-
(propyltosyl)aminophenanthridine-6-yl]hexane (6). Uracil (107 mg;
0.955 mmol, 10 equiv) that was previously dried, and NaH (38 mg,
60% w/w, 0.955 mmol, 10 equiv) were suspended in dry DMF (5 ml)
and stirred during 1 h in argon atmosphere at room temperature.
To this suspension, a solution of 1-[8-(3-bromopropyltosyl)amino-
Phen-10, 1H) ppm; 13C NMR (CDCl3)
d: 11.25 (CH3), 21.72 (Ts–CH3),
21.89 (CH2–propyl chain), 28.57 (CH2–propylene chain), 29.17
(CH2–hexylene chain), 29.25 (CH2–hexylene chain), 29.80 (CH2–
hexylene chain), 29.85 (CH2–hexylene chain), 36.16 (CH2–hexylene
chain), 36.23 (CH2–hexylene chain), 41.20 (adenine–NCH2–pro-
pylene chain), 47.94 (NTsCH2–propylene chain), 52.52 (NTsCH2–

2512
L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
propyl chain), 122.20 (Phen-containing-base-1), 122.25 (Phen-1),
123.15, 123.28, 123.67 (Phen-containing-base-10), 124.04 (Phen-
10), 125.75, 125.80, 126.16 (Phen-containing-base-7), 126.61 (Phen-
7), 126.81 (Phen-containing-base-2), 126.92 (Phen-2), 127.93 (Ts),
127.99 (Ts), 129.16 (Phen-containing-base-3), 129.36 (Phen-3),
129.71 (Ts), 129.80 (adenine-8), 129.84 (Ts), 129.91 (Phen-contain-
ing-base-4), 130.04 (Phen-4), 130.47 (Phen-containing-base-9),
130.54 (Phen-9), 132.31, 132.65, 134.62, 135.50, 137.74, 138.28,
143.76, 144.16, 144.24, 144.29, 152.98 (adenine-2), 155.64, 161.69,
hexylene chain, CH2–propylene chain, 6H), 3.12 (m, NCH2, 2H), 3.,21
(m, CH2–hexylene chain, 4H), 3.79 (t, uracil-NCH2–propylene chain,
2H, J¼6.3 Hz), 5.54 (d, uracil-5, 1H, J5–6¼7.9 Hz), 6.27 (NH, br s, 2H),
7.14 (dd, Phen-9, 2H,), 7.25 (s, Phen-7, 1H), 7.28 (s, Phen-7, 1H),7.50
(m, Phen-2, Phen-3, 4H), 7.62 (d, uracil-6, 1H), 7.87 (d, Phen-4, 2H),
8.51 (m, Phen-1, Phen-10), 11.25 (s, uracil-NH, 1H) ppm; 13C NMR
(DMSO-d6) d: 11.23 (CH3), 21.81 (CH2–propyl chain), 27.41 (CH2–
propylene chain), 29.21 (Phen-CH2), 29.76 (Phen-CH2), 36.05
(Phen-CH2), 46.51 (uracil-NCH2–propylene chain). 47.69 (NHCH2–
propylene chain), 52.35 (NHCH2–propyl chain), 102.27 (uracil-5); IR
161.88 ppm; IR (KBr) n: 3448, 2959, 2932, 2856, 2361, 2343, 1653,
1541,1508,1458,1340,1157,1090,1072, 951, 812, 762, 723, 706, 665,
582, 548 cmꢀ1; (MALDI/TOF-HRMS) m/z: 996.4083 (calcd for
C57H57N9O4S2: 996.4047).
(KBr)
1508, 1458, 1387, 1340, 1259, 1232, 1200, 1136, 1034, 997, 949, 864,
824, 760, 721, 669, 617, 548 cmꢀ1
(MALDI/TOF-HRMS) m/z:
n: 3398, 3057, 2926, 2853, 2363, 2345, 1684, 1655, 1618, 1541,
;
665.3580 (calcd for C42H44N6O2: 665.3599).
4.4.8. 1,6-Bis-[8-(3-(aden-9-yl)propyltosylamino)phenanthridine-6-
yl]-hexane (8). Compound (8) was obtained as described for 6; 1,6-
bis-[8-(3-bromopropyltosylamino)phenanthridine-6-yl]-hexane 5
(25 mg; 0.024 mmol), adenine (40 mg; 0.29 mmol, 10 equiv) and
NaH (12 mg, 60% w.w., 0.29 mmol, 10 equiv) in dry DMF (5þ5 ml)
gave white powder 8 (20 mg, 70% yield), Rf (SiO2, 10% MeOH in
4.4.11. 1-[8-(9-(Aden-1-yl)propyl)aminophenanthridine-6-yl]-6-[8-
(propyl)aminophenanthridine-6-yl]hexane (11). Compound (11)
was obtained as described for 9; 1-[8-(9-(aden-1-yl)propyltosyl)-
aminophenanthridine-6-yl]-6-[8-(propyltosyl)aminophenanthridine-
6-yl]hexane 7 (45 mg, 0.045 mmol) in concd H2SO4 (1 ml) and concd
acetic acid (2 ml) gave yellow powder 11 (15 mg, 50% yield); Rf (SiO2,
CH2Cl2)¼0.28; mp 151–155 ꢁC; 1H NMR (CDCl3)
d: 1.53 (br s, CH2–
hexylene chain, 4H), 1.84 (br s, CH2–hexylene chain, 4H), 2.06 (t,
CH2–propylene chain, 4H, J¼6.2 Hz), 2.38 (s, Ts–CH3, 6H), 3.22 (t,
Phen-CH2–hexylene chain, 4H, J¼7.7), 3.63 (t, NTsCH2–propylene
chain, 4H, J¼6.0 Hz), 4,33 (t, adenine–NCH2–propylene chain, 4H,
J¼6.3 Hz), 6.41 (br s, adenine–NH2, 4H), 7.19 (d, Ts, 4H, J¼8.1 Hz),
7.36 (d, Ts, 2H, J¼8.2 Hz), 7.43 (dd, Phen-9, 2H, J7–9¼2.0 Hz, J9–
10% MeOH in CH2Cl2)¼0.32; mp 119–122 ꢁC; 1H NMR (DMSO-d6)
d:
0.90 (t, CH3, 3H, J¼7.4 Hz), 1.51–1.61 (m, CH2–hexylene chain, CH2–
propyl chain, 6H), 1.88 (br s, Phen-CH2, 4H), 2.13 (m, CH2–propylene
chain, 2H), 3.08 (m, NHCH2–propyl chain, 2H), 3.16–3.21(m, Phen-
CH2, NHCH2–propylene chain, 6H), 4.27 (t, adenine–NCH2–propylene
chain, 2H, J¼6.7 Hz), 6.26 (br s, NH, 1H), 6.39 (br s, NH, 1H), 7.10 (s,
Phen-7, 2H), 7.21–7.29 (m, adenine–NH2, Phen-9, Phen-7, 5H), 7.48–
7.52 (m, Phen-2, Phen-3, 4H), 7.86 (m, Phen-4, H), 8.14 (br s, adenine-
2, adenine-8, 2H), 8.46–8.52 (m, Phen-1, Phen-10, 4H) ppm; 13C
¼8.8 Hz), 7.61 (t, Phen-2, 2H), 7.70 (t, Phen-3, 2H), 7.89 (s, Phen-7,
10
adenine-8, 4H), 8.09 (dd, Phen-4, 2H, J2–4¼1.0 Hz, J3–4¼8.1 Hz), 8.18
(s, adenine-2, 2H), 8.45 (d, Phen-1, 2H, J1–2¼7.5 Hz), 8.54 (d, Phen-
10, 2H) ppm; 13C NMR (CDCl3)
d: 21.58 (Ts–CH3), 28.14, 29.03, 29.52,
NMR (DMSO-d6) d: 11.80 (CH3), 21.83, 28.07, 28.21, 28.92, 29.19,
29.70, 35.93. 41.04, 47.46, 122.04, 122.87, 123.85, 125.51, 126.34,
126.76, 127.72, 129.21, 129.66. 129.80, 132.38, 134.09, 137.29, 143.90,
144.12, 152.44, 155.51, 161.62 ppm; IR (KBr) n: 3421, 2922, 2851,
29.29, 41.21, 44.70, 103.02, 103.32, 114.63, 119.04, 119.65, 120.59,
121.29, 121.36, 122.63, 122.93, 123.74, 123.86, 124.14, 124.22,
126.20. 126.24, 126.31, 126.73, 126.75, 129.05, 129.08, 129.62,
141.50, 141.58, 148.36. 148.70, 149.79, 152.55, 156.15, 160.59,
2363, 2345, 1647, 1597, 1574, 1475, 1420, 1385, 1340, 1304, 1244,
1159, 1109, 1088, 991, 935, 872, 814, 764, 725, 698, 667, 582,
160.70 ppm; IR (KBr) n: 3447, 2928, 2853, 2361, 2343, 1869, 1772,
544 cmꢀ1
;
(MALDI/TOF-HRMS) m/z: 1129.4405 (calcd for
1734, 1647, 1618, 1541, 1508, 1458, 1387, 1339, 1315, 1259, 1232,
1200, 822, 760, 669, 650, 519 cmꢀ1; (MALDI/TOF-HRMS) m/z:
688.3886 (calcd for C43H45N9: 688.3871).
C62H60N14O4S2: 1129.4435).
4.4.9. 1,6-Bis-[8-(propylamino)phenanthridine-6-yl]-hexane
(9). Compound (9) was obtained by heating solution of 1,6-bis-[8-
(propyltosylamino)phenanthridine-6-yl]-hexane 3 (27 mg, 0.032 mmol)
in a mixture of concd H2SO4 (1 ml) and concd acetic acid (2 ml)
under reflux at 80–100 ꢁC for 2 h. The reaction mixture was cooled,
poured on ice, and made alkaline (pH¼8–9) by addition of 2 M
NaOH. The obtained yellow-brown solid was precipitated, filtered,
and washed with lots of water to afford pure compound 9 (5 mg,
28% yield); mp 221–224 ꢁC; Rf (SiO2,10% MeOH in CH2Cl2)¼0.49; 1H
4.4.12. 1,6-Bis-[8-(3-(aden-9-yl)propylamino)phenanthridine-6-yl]-
hexane (12). Compound (12) was obtained as described for 9;
1,6-bis-[8-(3-(aden-9-yl)propyltosylamino)phenanthridine-6-yl]-
hexane 8 (30 mg, 0.027 mmol) in concd H2SO4 (1 ml) and concd
acetic acid (2 ml) gave yellow powder 12 (17 mg, 77% yield); Rf
(SiO2, 10% MeOH in CH2Cl2)¼0.45 mp 147–149 ꢁC; 1H NMR
(DMSO-d6) d: 11.53 (br s, CH2–hexylene chain, 4H), 1.85 (br s, CH2–
hexylene chain, 4H); 2.26 (m, CH2–propylene chain, 4H), 3.15 (m,
CH2–hexylene chain, NCH2, 8H), 4.26 (t, adenine–NCH2–propylene
chain, 4H, J¼6.7 Hz), 6.38 (br s, NH, 2H), 6.98 (s, Phen-7, 2H), 7.22
(m, adenine–NH2, Phen-9, 6H), 7.48–7.51 (m, Phen-2, Phen-3,
4H), 7.86 (m, Phen-4, H), 8.12 (s, adenine, 2H), 8.13 (s, adenine,
2H), 8.44–8.51 (m, Phen-1, Phen-10, 4H) ppm; 13C NMR (DMSO-
NMR (DMSO-d6)
d
: 0.94 (t, CH3, 6H, J¼7.5 Hz), 1.57–1.64 (m, CH2–
hexylene, CH2–propyl chain, 8H),1.89 (br s, CH2–hexylene, 4H), 3.12
(br s, CH2–hexylene, 4H), 6.28 (br s, NH, 2H), 7.12 (s, Phen-7, 2H),
7.27 (d, Phen-9, 2H, J9–10¼7.8 Hz), 7.46–7.53 (m, Phen-2, Phen-3,
4H), 7.88 (m, Phen-4, 2H), 8.49 (m, Phen-1, Phen-10, 4H) ppm; IR
(KBr)
1508, 1458, 1387, 1340, 1315, 1256, 1232, 1205, 1140, 824, 762, 669,
598, 517 cmꢀ1
(MALDI/TOF-HRMS) m/z: 555.3493 (calcd for
n
: 3447, 3246, 2961, 2926, 2854, 2361, 2334, 1653, 1618, 1541,
d6) d: 27.91, 28.7, 29.08, 35.11, 38.28, 41.01, 119.44, 121.33,
123.71, 123.95, 126.1, 126.53, 128.92, 130.79, 133.99, 138.37,
141.14, 141.4, 142.18, 148.18, 149.57, 155.99, 158.99, 160.45 ppm;
;
C38H42N4: 555.3482).
IR (KBr) n: 3337, 3200, 2924, 2852, 1640, 1619, 1575, 1541, 1479,
1462, 1420, 1395, 1335, 1308, 1240, 1210, 1178, 830, 800, 762,
730, 660 cmꢀ1; (MALDI/TOF-HRMS) m/z: 821.4279 (calcd for
C48H48N14: 821.4259).
4.4.10. 1-[8-(3-(Urac-1-il)propyl)aminophenanthridine-6-yl]-6-[8-
(propyl) aminophenanthridine-6-yl]hexane (10). Compound (10)
was obtained as described for 9; 1-[8-(3-(urac-1-il)propyltosyl)a-
minophenanthridine-6-yl]-6-[8-(propyltosyl)aminophenanthridine-
6-yl]hexane 6 (40 mg, 0.04 mmol) in concd H2SO4 (1 ml) gave yellow
powder 10 (15 mg, 53% yield); mp 108–110 ꢁC; Rf (SiO2, 10% MeOH
Acknowledgements
We thank the Ministry of Science, Education and Sport of the
Republic of Croatia for financial support of this study (grants no.
098-0982914-2918 and 098-1191344-2860).
in CH2Cl2)¼0.39; 1H NMR (DMSO-d6)
: 0.94 (t, CH3, 3H, J¼7.4 Hz),
d
1.59 (m, CH2–hexylene chain, CH2–propyl chain, 6H), 1.90 (m, CH2–

L.-M. Tumir et al. / Tetrahedron 66 (2010) 2501–2513
2513
10. Schneider, H.-J.; Yatsimirsky, A. In Principles and Methods in Supramolecular
Chemistry; Wiley-VCH: Weinheim, 2000.
Supplementary data
ˇ
11. Tumir, L.-M.; Piantanida, I.; Novak, P.; Zinic´, M. J. Phys. Org. Chem. 2002, 15, 599.
The UV–vis spectra of studied nucleotides, additional molecular
modeling results associated with this article can be found in online
version at doi:10.1016/j.tet.2010.01.063.
ˇ
´
´
´
´
12. Tumir, L.-M.; Piantanida, I.; Juranovic Cindric, I.; Hrenar, T.; Meic, Z.; Zinic, M.
J. Phys. Org. Chem. 2003, 16, 891.
ˇ
13. Tumir, L.-M.; Piantanida, I.; Zinic´, M.; Juranovic´ Cindric´, I.; Meic´, Z.; Kralj, M.;
´
Tomic, S. Eur. J. Med. Chem. 2006, 41, 1153.
14. Juranovic´, I.; Meic´, Z.; Piantanida, I.; Tumir, L.-M.; Zinic´, M. Chem. Commun. 2002, 1432.
ˇ
ˇ
´
´
´
´
15. Tumir, L.-M.; Piantanida, I.; Juranovic, I.; Meic, Z.; Tomic, S.; Zinic, M. Chem.
Commun. 2005, 2561.
16. Cudic´, P.; Zinic´, M.; Skaric´, V.; Kiralj, R.; Kojic´-Prodic´, B.; Vigneron, J.-P.; Lehn, J.-
M. Croat. Chem. Acta 1996, 69, 569.
References and notes
ꢀ
ˇ
ˇ
1. (a) Lehn, J. M. Supramolecular Chemistry. Concepts and Perspectives; Wiley-VCH:
Weinheim, 1995; (b) Atwood, J. L. Comprehensive and Supramolecular Chemistry;
Pergamon: Oxford, 1996.
2. (a) De Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley, A. J. M.; McCoy,
C. P.; Rademacher, J. T.; Rice, T. E. Chem. Rev. 1997, 97, 1515; (b) Sessler, J. L.; Kral,
V.; Shishkanova, T. V.; Gale, P. A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4848.
3. Constant, J.-F.; Fahy, J.; Lhomme, J. Tetrahedron Lett. 1987, 28, 1777.
4. Shionoya, M.; Ikeda, T.; Kimura, E.; Shiro, M. J Am Chem Soc. 1994, 116, 3848.
5. Aoki, S.; Kimura, E. J. Am. Chem. Soc. 2000, 122, 4542.
17. Morgan, G. T.; Walls, L. P. J. Chem. Soc. 1931, 2447.
18. Jones, R. L.; Wilson, W. D. Biopolymers 1981, 20, 141.
19. Specfit Global Analysis Program Gampp, H.; Maeder, M.; Meyer, C. J.; Zu-
berbuehler, A. D. Talanta 1985, 32, 257; Maeder, M.; Zuberbuehler, A. D. Anal.
Chem. 1990, 62, 2220.
20. Wakelin, L. P. G. Med. Res. Rev. 1986, 6, 275.
21. Demeunynck, M.; Bailly, C.; Wilson, W. D. In DNA and RNA binders; Wiley-VCH:
Weinheim, 2002.
6. (a) Teulade-Fichou, M.-P.; Vigneron, J.-P.; Lehn, J.-M. Supramol. Chem. 1995, 5,
139; (b) Baudoin, O.; Gonnet, F.; Teulade-Fichou, M.-P.; Vigneron, J.-P.; Tabet, J.-
C.; Lehn, J.-M. Chem.dEur. J 1999, 5, 2762.
ꢀ
ˇ
´
´
ˇ ´
22. Cudic, P.; Zinic, M.; Tomisic, V.; Simeon, V.; Vigneron, J.-P.; Lehn, J.-M. J. Chem.
Soc., Chem. Commun. 1995, 1073.
23. INSIGHTII. Accelerys; 2001–2008; http://accelrys.com/services/training/life-
science/insight-migration.html San Diego.
24. Duan, Y.; Wu, C.; Chowdhury, S.; Lee, M. C.; Xiong, G.; Zhang, W.; Yang, R.;
Cipelak, P.; Luo, R.; Lee, T. J. Comput. Chem. 2003, 24, 1999.
25. http://amber.scripps.edu/.
ꢀ
ˇ
7. Piantanida, I.; Palm, B. S.; Cudic´, P.; Zinic´, M.; Schneider, H.-J. Tetrahedron Lett.
2001, 42, 6779.
ꢀ
ˇ
ˇ ´
8. Cudic´, P.; Zinic´, M.; Tomisic, V.; Simeon, V.; Vigneron, J.-P.; Lehn, J.-M. J. Chem.
Soc., Chem. Commun. 1995, 1073.
9. Meyer, E. A.; Castellano, R. K.; Diederich, F. Angew. Chem., Int. Ed 2003, 42, 1210.