Metal-salen-base-pair complexes inside DNA: Complexation overrides sequence information

Article abstract of DOI:10.1002/chem.200600558

Two isomeric salicylic alde-hyde nucleobases have been prepared and incorporated into various DNA duplexes. Reaction with ethylenediamine leads to formation of the well-known salen ligand inside the DNA double helix. Addition of transition-metal ions such as Cu²⁺, Mn²⁺, Ni ²⁺, Fe²⁺, or VO²⁺ results in the formation of metal-salen-base-pair complexes, which were studied by using UV and circular dichroism (CD) spectroscopy. HPLC and ESI mass spectrometric measurements reveal an unusually high stability of the DNA-metal system. These metal-salen complexes act as interstrand cross-links and thereby lead to a strong stabilization of the DNA duplexes, as studied by thermal de- and renaturing experiments. Complex formation is strong enough to override sequence information even when the preorganization of the ligand precursors is unfavorable and the DNA duplex is distorted by the metal complexation. Furthermore, melting-point studies show that the salen complex derived from ligand 2 fits better into the DNA duplex, in accordance with results obtained from the crystal structure of the corresponding copper-salen complex 8.

Full text of DOI:10.1002/chem.200600558

DOI: 10.1002/chem.200600558
Metal–Salen-Base-Pair Complexes Inside DNA: Complexation Overrides
Sequence Information
Guido H. Clever, Yvonne Sçltl, Heather Burks, Werner Spahl, and Thomas Carell*[a]
Abstract: Two isomeric salicylic alde-
hyde nucleobases have been prepared
and incorporated into various DNA
duplexes. Reaction with ethylenedi-
HPLC and ESI mass spectrometric
measurements reveal an unusually high
stability of the DNA–metal system.
These metal–salen complexes act as in-
terstrand cross-links and thereby lead
to a strong stabilization of the DNA
duplexes, as studied by thermal de- and
renaturing experiments. Complex for-
mation is strong enough to override se-
quence information even when the pre-
organization of the ligand precursors is
unfavorable and the DNA duplex is
distorted by the metal complexation.
Furthermore, melting-point studies
show that the salen complex derived
from ligand 2 fits better into the DNA
duplex, in accordance with results ob-
tained from the crystal structure of the
corresponding copper–salen complex 8.
A
known salen ligand inside the DNA
double helix. Addition of transition-
metal ions such as Cu2+, Mn2+, Ni2+
,
Fe2+, or VO2+ results in the formation
of metal–salen-base-pair complexes,
which were studied by using UV and
circular dichroism (CD) spectroscopy.
Keywords: circular dichroism
DNA structures · DNA · nucleo-
bases · transition metals
·
AHCTREUNG
Introduction
by nature to form the DNA duplex, are replaced by metal
coordination interactions.[5] This concept may in the future
allow a switching of the DNA single-strand interaction due
to the presence or absence of specific metals. The motiva-
tion to incorporate metals into a DNA double helix in a de-
fined fashion reaches far beyond the development of new
base-pairing modes, however, as metal ions tightly coordi-
nated inside the chiral double helix provide the possibility
to generate new catalysts for enantioselective transforma-
tions. Here, the chiral environment established by the DNA
is, in principle, amenable to optimization using evolutionary
techniques.[6]
A second line of interest behind the incorporation of
metal–base-pairs into DNA is the desire to program the
construction of metal arrays. Solid-phase DNA synthesis
allows the synthesis of highly modified DNA in which sever-
al metals may be stacked on top of each other. This provides
new perspectives for the nanotechnological exploitation of
the DNA structure as a molecular wire or as an electronic
switch.[7] From an inorganic chemistsꢀ point of view, access
to a variable set of multidentate ligands is of great interest
to study metal interactions in homo- or heterometallic coor-
dination compounds.
The molecular structure of DNA has fascinated scientists
since the report of its elucidation by Watson and Crick in
1953.[1] The ascent of molecular genetics is directly based on
this discovery and has led to revolutionary changes in the
science of biology. Modern techniques of molecular biology,
like cloning, the polymerase chain reaction (PCR), and
DNA sequencing, would be unimaginable without knowl-
edge of the structure of DNA and the mechanisms of semi-
conservative replication.
In contemporary chemistry, the synthesis of novel DNA-
like structures is of tremendous interest. Recent examples
include the synthesis of new base-pairing systems based on
altered sugar components such as TNA,[2] pRNA, and ho-
moDNA,[3] or the preparation of DNA-like materials with
extended nucleobases.[4]
A more recent development is the design of metal–base-
pairs in which the hydrogen bonds, which are typically used
[a]Dipl.-Chem. G. H. Clever, Dipl.-Chem. Y. Sçltl, H. Burks,
Dr. W. Spahl, Prof. Dr. T. Carell
Department of Chemistry and Biochemistry
Ludwig Maximilians University Munich
Butenandtstrasse 5–13, Haus F, 81377 Munich (Germany)
Fax : (+49)089-2180-77756
We recently reported a metal–salen-base-pair complex in
DNA that accepts a variety of metal ions. This metal–base-
pair complex yielded the highest ever reported duplex stabi-
lization in terms of melting-point increase.[8] In contrast to
E-mail: Thomas.Carell@cup.uni-muenchen.de
8708
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8708 – 8718

FULL PAPER
all other known metal–base-pair complexes, the salen con-
cept combines a covalent cross-link of both DNA single
strands with the metal complexation event (Figure 1), and
we showed that the assembly of the salen complex inside
DNA proceeds cooperatively—the diamine is needed to
form the ligand, while the coordinated metal prevents hy-
drolysis of the formed imines.
resonance) mass spectrometry experiments provide detailed
information about the integrity and stability of the metal–
base-pair-containing DNA. We observe that the formation
of the metal–salen cross-link overrides geometrical con-
straints preset by the nucleoside geometry or the DNA se-
quence context in all cases. Formation of the complex
always occurs, even when the double helical structure is
strongly distorted or when one
native base pair has to be
broken.
Results and Discussion
Synthesis of nucleoside 1 and
its incorporation into desoxy-
oligonucleotides: The salicylic
aldehyde nucleoside 1 was pre-
pared according to the previ-
ously reported procedure for
preparation of its isomer
(Scheme 1).[8]
This synthesis requires pro-
tection of both the carbonyl
and hydroxyl groups of the sal-
2
Figure 1. Assembly of the metal–salen-base-pair complex in the DNA double helix.[8]
Here we report a detailed investigation of the formed
metal–base-pair complex in DNA, including thermal de-
and renaturing studies, UV and circular dichroism (CD)
spectroscopy measurements, and high-resolution mass spec-
trometry. The influence of the nucleoside geometry and var-
iation of the sequence context on the complex formation
and duplex stability are also examined. The precursors for
the assembly of the metal–salen-base-pair complexes are the
salicylic aldehyde nucleobases 1 and 2 (Figure 2). A compar-
icylic aldehyde; this was performed bearing in mind that the
protecting groups must be compatible with the organometal-
lic C-glycosidation and subsequent solid-phase DNA synthe-
sis. Furthermore, both protecting groups have to be cleaved
under mild conditions after DNA synthesis to avoid degra-
dation of the oligonucleotide. Commercially available 5-bro-
mosalicylic aldehyde (3) was first protected as a 1,3-dioxane
and was then treated with TIPS-Cl to yield the protected
ligand 4. Compound 4 was treated with two equivalents of
tert-butyllithium to achieve a bromine–lithium exchange.
The transmetalation of the organolithium compound to cop-
per(I) is the most critical step in the C-glycosidation proce-
dure. The reaction was found to work best when the lithiat-
ed ligand was transferred to a suspension of CuBr·SMe2 in
freshly dried diethyl ether at ꢀ788C. Careful warming of
this mixture to ꢀ308C until all solids had dissolved to give a
clear yellow solution yielded the corresponding organocop-
per species, which turned out to be very unstable. It was
therefore treated immediately with a solution of 5 in dry di-
chloromethane at ꢀ788C to give compound 6. For the de-
sired b-anomer, the configuration at C-1’ was assigned by
evaluation of the nuclear Overhauser effect between the hy-
drogen atoms C1’-H, C2’-H, and C3’-H. The through-space
coupling between C1’-H and a-C2’-H as well as between b-
C2’-H and C3’-H lead to the appearance of cross-peaks in
the NOESY NMR spectrum of nucleoside b-6. Attempts to
increase the overall yield of the b-nucleoside by transform-
ing a fully deprotected a-nucleoside into its b-anomer by
acid treatment, as described for unsubstituted aryl-C-nucleo-
sides,[9] resulted in decomposition of compound a-6. Depro-
tection of the sugar hydroxyl groups by transesterification in
methanol[10] was followed by 5’-DMT protection and genera-
Figure 2. Salicylic aldehyde nucleosides 1 and 2 used to prepare the salen
complexes in the DNA strands described in this study.
ison of oligonucleotides containing either ligand 1 or its
isomer 2 by melting-point experiments shows that nucleo-
side 2 leads to a higher duplex stabilization upon complex
formation. Furthermore, the superposition of the crystal
structure of the copper–salen base pair 8 (derived from nu-
cleoside 2) with a natural AT Watson–Crick base pair shows
a surprisingly large geometrical match. We present spectro-
scopic data supporting the assembly process of the metal–
salen-base-pair complex inside the DNA duplex and we ad-
dress the question of chirality transfer from the double helix
to the metal complex. Finally, ESI-ICR (ICR=ion cyclotron
Chem. Eur. J. 2006, 12, 8708 – 8718
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8709

T. Carell et al.
which give broad and multiple peaks in the HPL chromato-
gram. It was subsequently found that addition of 20% acetic
acid prior to the chromatographic separation process quanti-
tatively cleaved the imines to the colorless aldehydes, which
simplified the chromatographic purification. All oligonu-
cleotides were finally desalted on C18 columns and the DNA
was stored in water at ꢀ208C. A list of the prepared DNA
strands used for this study can be found in Figure 4.
Scheme 1. Synthesis of phosphoramidite 7 and its incorporation into vari-
ous DNA strands. a) 1,3-propanediol, HC(OEt)3, N(nBu)4Br3, RT, 24 h,
96%; b) TIPS-OTf, NEt(iPr)2, CH2Cl2, RT, 12 h, 95%; c) 2 equiv tBuLi,
A
ACHTREUNG
ACHTREUNG
Et2O, ꢀ788C, 2 h; d) CuBr·SMe2, ꢀ788C to ꢀ308C, 20 min; e) 5, CH2Cl2,
12 h, ꢀ788C to RT, 10% (only b-anomer isolated); f) K2CO3, CH3OH,
RT, 2 h, 43%; g) DMT-Cl, pyridine, 3 h, 55%; h) CED-Cl, NEt(iPr)2,
A
THF, RT, 2 h, 92%; i) automated DNA synthesis; j) deprotection of alde-
hydes with water-containing 2% CHCl2COOH in CH2Cl2; k) cleavage
from solid support and cleavage of TIPS with NH3(aq)/EtOH (3:1); l) hy-
bridization of matching sequences under stringent conditions (Table 1);
m) addition of excess ethylenediamine (en) and 1 equiv of metal ions
(Table 1). CED-Cl=(2-cyanoethyl)-N,N-diisopropylchlorophosphorami-
dite, DMT=4,4’-dimethoxytrityl, Tf=trifluoromethanesulfonyl, TIPS=
triisopropylsilyl, Tol=toluyl.
Figure 3. a) Top and side views of the crystal structure of compound 8;
b) Superposition of 8 (colored) with an AT base pair (black).
tion of the phosphoramidite 7. DNA synthesis was per-
formed on an ¾kta Oligopilot 10 synthesizer using a stand-
ard desoxyoligonucleotide protocol. The use of controlled-
pore-size glass beads (CPG) as solid support material was
found to give the desired oligonucleotides in excellent yields
and with high purity. Treatment of the support-bound oligo-
nucleotides with water containing 2% dichloroacetic acid in
dichloromethane allowed removal of the acetal protecting
groups without causing measurable depurination. The TIPS
protecting groups were then removed smoothly and the
DNA strands cleaved from the
Crystal structure of the metal–salen-base-pair complex: Re-
moval of all protecting groups from the b-nucleoside yielded
compound 2, which was subsequently treated with half an
equivalent of ethylenediamine (en) in methanol to give the
salen ligand. Treatment of the chelate ligand with one equiv-
alent of [Cu(acac)2]in methanol yielded the copper–salen
A
complex 8 as a purple solution, from which dichroic green-
purple crystals were grown by slowly cooling this solution
from 658C down to room temperature (Scheme 2).[11]
solid support by treatment
with saturated aqueous ammo-
nia/ethanol (3:1) at room tem-
perature overnight. In the am-
monia-containing solution, the
unprotected aldehydes are in
equilibrium with their corre-
sponding yellow aldimines,
Scheme 2. Synthesis of the copper–salen complex 8.
8710
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8708 – 8718

Metal–Salen-Base-Pair Complexes Inside DNA
FULL PAPER
Table 1. Melting-point experiments with the oligonucleotides 9–14.
The crystal structure of complex 8 is depicted in Fig-
Strand(s)[a]
Additive(s)
TM [8C]
ure 3a. The metal–base-pair complex displays a tetrahe
G
U
Entry
[b]
distorted square-planar coordination geometry of the
copper center similar to that reported for related copper–
1
2
3
4
5
6
7
8
9-L1-a/b
9-L1-a/b
9-L1-a/b
9-L1-a/b
9-L2-a/b
9-L2-a/b
9-L2-a/b
9-L2-a/b
9-L2-a/b
9-L2-a/b
10-L2-a/b
10-L2-a/b
10-L2-a/b
10-L2-a/b
11-L2-a/b
11-L2-a/b
11-L2-a/b
11-L2-a/b
13-L2-a/b
13-L2-a/b
13-L2-a/b
13-L2-a/b
14-L2
35.7
40.5
36.8
71.6
39.9
45.0
54.9
82.4
[b]
100 mm en
4 mm Cu2+[b]
100 mm en
100 mm en
100 mm en
100 mm en
100 mm en
100 mm en
100 mm en
100 mm en
100 mm en
100 mm en
4 mm Cu2+[b]
[b]
[b]
4 mm Cu2+[b]
4 mm Cu2+[b]
[c]
9
40.7
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
4 mm Mn2+[c]
68.8[d]
32.0
[b]
4 mm Cu2+[b]
66.8[e]
33.4
[c]
6 mm Mn2+[c]
60.6[d]
18.6
[b]
4 mm Cu2+[b]
59.1
[c]
20.6
4 mm Mn2+[c]
53.2[d]
20.9
[b]
Figure 4. Sequences of the examined oligonucleotides. L1 =1 and L2 =2.
D=tetrahydrofuran spacer (1’,2’-desoxyribose).
4 mm Cu2+[b]
56.5[e]
21.0
[c]
6 mm Mn2+[c]
57.8[d]
35.4
[b]
salen complexes.[12] The chelate rings exhibit a D configura-
tion that resembles surprisingly closely the propeller twist of
a native Watson–Crick base pair. The dihedral angle, qP, be-
tween the planes defined by the aromatic rings was found to
be +228. This is slightly larger than in natural base pairs
(ca. +108) but of the same sign. The salicylic aldimine moi-
eties are oriented in an anti conformation with respect to
the sugars, and the distance between the C1’ atoms of the
two sugars is 11.47 Š, which is close to the values of 10.44
and 10.72 Š for AT and GC Watson–Crick base pairs, re-
spectively. The angle between the C-glycosidic bond and the
line connecting the C1’ atoms (558) is in excellent agreement
with that of a normal base pair. The 2’-desoxyribose sugar
rings exhibit a C2’-endo (“south”) conformation, which is
common for B-type DNA.[13] The molecule has a C2 axis
that passes through the copper atom and the middle of the
ethylene bridge. The superposition depicted in Figure 3b
shows the high geometrical match of the desoxyribosyl-sub-
stituted Cu–salen complex and a normal AT Watson–Crick
base pair. No circular dichroism was observed for a solution
of 8 in water.
14-L2
100 mm en
100 mm en
52.2
76.5
36.0
[b]
14-L2
4 mm Cu2+[b]
[c]
14-L2
[c]
14-L2
100 mm en
100 mm en
100 mm edh
51.6
14-L2
4 mm Mn2+[c]
70.3[d]
73.4[f]
[b]
9-L2-a/b
[a]For sequences, see Figure 4. All samples contained 3 mm DNA (duplex
or hairpin) and 150 mm NaCl. Melting profiles were measured from 0 to
858C (958C for Cu2+) with a slope of 0.58Cminꢀ1. [b]All experiments
using Cu2+ and corresponding controls were carried out in 10 mm Ches
(N-cyclohexyl-2-aminoethanesulfonic acid) buffer at pH 9.0. [c]All ex-
periments using Mn2+ and corresponding controls were carried out in
10 mm Hepes (N-(2-hydroxyethyl)piperazine-N’-(2-ethanesulfonic acid))
buffer at pH 9.0. [d]Reproducible differences in de- and renaturing pro-
files due to thermal instability of the Mn complex. The given melting
points correspond to the denaturing profiles. [e]Additional transition of
low intensity (entry 12: 23.88C; entry 20: 16.08C). [f]edh =O,O’-ethyl-
enedihydroxyamine. Entries 5–10 are reproduced from reference .[8]
the cross-linking of both strands by ethylenediamine is, how-
ever, surprisingly small. The reason for this is that formation
of the imine linkage in water is highly reversible, which
causes rapid hydrolysis of the cross-link during the melting-
point experiment.[14] The better geometrical fit of the salen
complex based on nucleobase 2 is particularly obvious after
addition of only Cu2+ ions (no ethylenediamine). Only the
perfectly preoriented system 9-L2-a/b accepts the metal,
which results in a strong stabilization. Duplex 9-L1-a/b, in
contrast, shows no stabilizing effect upon addition of Cu2+
ions, thus indicating that in a duplex where two salicylic al-
dehydes 1 face each other as a base pair, metal coordination
between them is impossible (Table 1, entries 3 and 7). How-
ever, addition of ethylenediamine and copper results in dra-
matic melting-point increases for both duplexes (Table 1, en-
tries 4 and 8), thus showing the strong cooperativity of the
complex formation in DNA. The complexed metal prevents
hydrolysis of the imine bonds. The stability of the rigid salen
complex is seemingly so dominating that its formation
Thermal de- and renaturation studies: To determine the
thermal stability of DNA duplexes containing the ligand
precursors 1 and 2, melting-point measurements in the ab-
sence and presence of diamines and metal ions were per-
formed. The results are summarized in Table 1.
From the melting-point data it is evident that all oligonu-
cleotides containing ligand 1 are less stable than those with
nucleoside 2. The base analogue 2 seems to orient the sali-
cylic aldehydes inside the DNA duplex in a way that offers
perfect preorganization for the desired complex formation.
The melting point of duplex 9-L2-a/b (containing ligand 2) is
increased by about 5 K upon addition of only ethylenedi-
ACHTREUNG
Chem. Eur. J. 2006, 12, 8708 – 8718
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8711

T. Carell et al.
occurs even when the preorganization of the salicylic alde-
hyde precursors in DNA is not optimal (as in duplex 9-L1-a/
b). The energetic gain upon forming the complex is so large
that the system can tolerate the distortion of the DNA
double helix. In this situation, complex formation in DNA
occurs first and hence controls the final duplex geometry.
The degree of stabilization is again higher for duplex [9-
L2-a/b+en+Cu], which is another indication that the salen
ligand based on salicylic aldehyde 2 fits better into the
duplex. The discussed melting points are compared in
Figure 5.
Figure 6. Comparison of the thermal stabilities of duplexes 9-L2-a/b, 10-
L2-a/b, and 11-L2-a/b without any additives, with en and Mn2+, and with
en and Cu2+
.
and manganese or copper as the metal. Complex formation
even works when the salicylic aldehydes are shifted as in 11-
L2-a/b but separated by an AT base pair, as in the duplex 9-
L2-a/12-L2-b. The addition of ethylenediamine and Cu2+ to
this duplex leads to a complex melting behavior that differs
significantly from the melting curve of the pure duplex (not
shown). Furthermore, mass spectrometric analysis shows the
quantitative formation of the duplex containing one copper–
salen complex (Table 2, entry 11). Consequently, the forma-
tion of the salen complex in the duplex [9-L2-a/12-L2-
b+en+Cu]must have broken the AT base pair between the
two salicylic aldehydes. The formation of the salen complex
is obviously so strong that it forces the DNA duplexes to
accept unfavorable double helical structures and even one
broken base pair. That the double helix plays a role in com-
plex formation became obvious when we analyzed the
single-strand composition of the duplex as we were unable
to detect any homoduplexes (a/a or b/b) by mass spectro-
Figure 5. Comparison of the thermal stability of duplexes 9-L1-a/b and 9-
L2-a/b upon addition of ethylenediamine and/or Cu2+
.
In order to investigate how the preorganization of the sal-
icylic aldehydes in the duplex effects metal–salen complex
formation in more detail, we systematically varied the posi-
tion of the two salicylic aldehydes in the oligonucleotide se-
quence, as depicted in Figure 6. In these constructs, we
chose
a simple tetrahydrofuran-derived spacer (D in
Figure 4), which does not lead to any unwanted interaction
with the ligands, as counterbase to the aldehydes. A graphi-
cal comparison of the determined thermal stabilities of the
original strand 9-L2-a/b with the sequences 10-L2-a/b and 11-
L2-a/b is displayed in Figure 6
(the sequences are depicted in
Table 2. ESI mass spectrometry experiments with the oligonucleotides 9, 13, and 15.
Figure 4). The melting points
of the bare duplexes decrease
by about 8 K upon loss of the
first and by another 14 K upon
loss of a second AT base pair.
The values for the strands after
assembly of the metal–salen-
base-pair complex follow the
same trend (Table 1, en-
tries 11–14 and 15–18). Howev-
er, duplexes 10-L2-a/b and 11-
L2-a/b, in which the aldehydes
are shifted by one or two posi-
tions, respectively, are still able
to form interstrand salen com-
plexes with ethylenediamine
Entry
Strand(s)[a]
Additive(s)
Species
[Mꢀ4H]4ꢀ
[Mꢀ5H]4ꢀ
[Mꢀ4H]4ꢀ
[Mꢀ5H]4ꢀ
[Mꢀ4H]4ꢀ
[Mꢀ4H]4ꢀ
[Maꢀ4H]4ꢀ
[Mbꢀ4H]4ꢀ
[Mꢀ9H]9ꢀ
[Mꢀ9H]9ꢀ
[Mꢀ9H]9ꢀ
[Mꢀ9H]9ꢀ
Calcd mass
Exptl mass
D [ppm]
1
2
3
4
5
6
7
15-L2
15-L2
15-L2
15-L2
15-L2
15-L2
9-L2-a/b
G
1209.4573
1228.4364
1230.6999
1228.6967
1231.6997
1229.4488
1136.6993
1141.2048
1021.7207
961.3902
1209.4497
1228.4376
1230.6908
1228.6902
1231.6973
1229.4299[b]
1136.6887
1141.1948[c]
1021.7220
961.3792
6.3
1.0
7.4
5.3
1.9
15.4
9.3
8.8
en
en
en
en
en
Mn3+
ACHTREUNG
Cu2+
Fe3+
VO2+
Ni2+
G
U
E
C
AHCTREUNG
T
8
9
10
11
9-L2-a/b
en
Cu2+
U
1.3
13-L2-a/b
15-L2
edh[d]
phen[d]
en
E
11.4
12.7
3.7
Cu2+
Cu2+
U
1242.6998
1020.7297
1242.6840
1020.7259
9-L2-a/12-L2-b
ACHTREUNG
[a]For sequences see Figure 4. All samples contained 30 mm DNA (duplex or hairpin) and 100 mm NH4OAc
(pH 8). DNA strands were first hybridized by slow cooling from 80 to 258C and then incubated for at least
12 h with the diamine and a solution of the metal sulfate at room temperature. [b]Additional peaks for [ 15-
L2+en+2Niꢀ2H2Oꢀ4H+]and [ 15-L2+en+3Niꢀ2H2Oꢀ6H+]were observed. [c]Only single-strand masses
observed. [d]edh =O,O’-ethylenedihydroxyamine, phen=1,2-phenylenediamine.
8712
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8708 – 8718

Metal–Salen-Base-Pair Complexes Inside DNA
FULL PAPER
metric analysis—the metal was always complexed inside the
correct heteroduplexes (a/b). These results indicate that the
two single strands have to form a stable duplex and that this
seems to be a prerequisite for complex formation. In the
duplex, however, complex formation takes place even if the
double helix is distorted afterwards. Formation of [9-L2-a/
12-L2-b+en+Cu]demonstrates that the complex formation
is even able to override sequence information.
The double helix 13-L2-a/b, in which the facing salicylic al-
dehydes are flanked by the spacers D on either side, shows
a
similar behavior in the melting-point experiments
(Table 1, entries 19–22).
Hairpin 14-L2 carries the metal–base-pair complex right
next to a TTTT loop such that the metal complex closes the
hairpin and presents the metal to the core of a chiral cavity.
In this case formation of metal–salen complexes with either
Mn2+ or Cu2+ can also be observed in the thermal de- and
renaturing curves.
Finally, we also varied the diamine bridge. Because imine
formation in the absence of a coordinated metal is reversi-
ble, we used hydroxylamine derivatives as they give much
more stable oximes. These oximes indeed result in cross-
linking, even in the absence of any coordinating metal ions.
For example, addition of O,O’-ethylenedihydroxyamine to
duplex 9-L2-a/b caused bridging of the two strands and an
increase of TM by 33.5 K (Table 1, entry 29). In addition, the
molecular ion of the duplex was observed in the ESI mass
spectrum (see below). 1,2-Phenylenediamine can also be
used as the cross-linking bridge, provided oxygen is exclud-
ed (Table 2, entry 10).
Figure 7. a) Electronic absorption bands of duplexes 9-L2-a/b, [9-L2-a/
b+en], and [9-L2-a/b+en+Cu]; b) Titration of [9-L2-a/b+en](30 mm
UV/Vis and CD spectroscopy: Further insight into the for-
mation of the interstrand salen ligand and the complexation
of divalent metal ions was obtained by monitoring of the as-
sembly process by means of UV/Vis spectroscopy. The
duplex 9-L2-a/b has an absorption maximum at l=260 nm,
as expected for a double strand consisting primarily of natu-
ral nucleobases.[13] The salicylic aldehydes give rise to an ad-
ditional absorption at l=330 nm due to the p!p* transi-
tion of the aromatic chromophore.[15] Addition of an excess
DNA, 1 mm en, 100 mm NH4OAc(aq),
pH 8) with Cu2+ in steps of
0.1 equiv; thick line: 1.0 equiv Cu2+; inset: plot of Abs360 against the ratio
[Cu2+]/[9-L2-a/b+en].
points at l=334 and 395 nm. The plot of the absorption at
l=360 nm against the copper concentration shows a linear
increase up to a duplex to Cu2+ ratio of about 1:1. The UV
spectra of 9-L2-a/b, [9-L2-a/b+en], and [9-L2-a/b+en+Cu]
are shown in Figure 7a.
of ethylenediamine resulted in the appearance of a new
A
band at l=410 nm. At the same time, the absorption of the
salicylic aldehyde at l=330 nm decreased over 20 min. The
absorption at l=410 nm matches the reported values for a
deprotonated salen ligand. The existence of isosbestic points
at l=325 and 358 nm indicates an immediate formation of
the salen ligand when an ethylenediamine molecule encoun-
ters the preorganized salicylic aldehydes (not shown). In this
model, the formation of the first imine bond is rate-deter-
mining and the second imine bond formation is accelerated
for entropic reasons.
In contrast to our observation that an aqueous solution of
the homochiral Cu–salen complex 8 shows no CD signal in
the range between l=300 and 700 nm, the duplex [9-L2-a/
b+en+Cu]features a strong CD signal in the range of the
absorption of the p!p* transition (Figure 8). The observed
signal has a positive sign in the high-energy region and a
negative sign for the low-energy part and corresponds, ac-
cording to studies by Downing et al., to a D configuration of
the metal chelate inside the duplex.[16]
The salen complex adopts the same absolute configura-
tion inside the DNA duplex as in the crystal (Figure 3).
Concerning the metal-based d–d transition at around l=
570 nm, only a small CD effect is observed (Figure 8).
Coordination of Cu2+ ions by the preformed salen ligand
in DNA results in a shift of the absorption band to l=
360 nm. In addition, a new band appears at l=570 nm,
which is typical for the N2O2–Cu chromophore.[16] A titra-
tion of Cu2+ ions against the DNA duplex [9-L2-a/b+en]is
shown in Figure 7b. The overlaid curves show isosbestic
HPLC analysis of the copper–salen-containing duplexes:
The unusually high stability of the copper–salen complex in
Chem. Eur. J. 2006, 12, 8708 – 8718
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8713

T. Carell et al.
ESI-ICR mass spectrometry: We used a Thermo Finnigan
LTQ-FT ESI-ICR mass spectrometer for the mass spectro-
metric characterization of the metal-containing duplexes.[17]
The metal–DNA samples were prepared by hybridizing
equimolar amounts of both single strands in ammonium ace-
tate buffer (pH 8, 100 mm) and subsequent incubation with
the diamine and the corresponding metal salt overnight at
room temperature. In all cases, the experimentally found
masses are in excellent agreement with the values calculated
for the hairpins or duplexes containing one molecule of di-
A
lecular weights of the lowest-weight isotopomers along with
the measured values (for m/z with z=ꢀ4 or ꢀ9). In each
case, one molecule of diamine condenses with both salicylic
aldehydes of the DNA strands, with the loss of two water
molecules, to form the cross-linking ligand that binds the
metal ion.
Figure 8. Circular dichroism spectra of 9-L2-a/b, [9-L2-a/b+en], and [9-
L2-a/b+en+Cu](30 mm DNA, 1 mm en, 30 mm CuSO4, 100 mm NH4OAc,
pH 8).
All the molecular weights obtained prove the presence of
only one metal ion in the duplexes or hairpins (Table 2, en-
tries 2–5). Interestingly, only in the case of Ni2+ were molec-
ular weights obtained that indicate the presence of [15-
the DNA duplex has a great influence on the chromato-
graphic behavior of the double strand 9-L2-a/b (Figure 9).
L2+en+2Ni2+ꢀ2H2Oꢀ4H+]and of
[
15-L2+en+3Ni2+
ꢀ2H2Oꢀ6H+]as well as the formation of the expected
mono-Ni2+ adduct [15-L2+en+Ni2+ꢀ2H2Oꢀ2H+]. These
are a sign of further unspecific, but rather tight, binding of
additional Ni2+ to the oligonucleotide once the salen ligand
is saturated with metal.
Addition of Mn2+ and Fe2+ to the ligand-containing du-
plexes and hairpins resulted in oxidation to give Mn3+ and
Fe3+ ions, as clearly proven by the m/z values.[18] The charge
of the coordinated metal can be deduced from the observed
m/z value by comparison with the simulated isotope pattern.
Only the peaks expected for the coordination of one
iron
(III) ion to the assembled salen ligand, along with Na+
G
and K+ adducts, appear in the spectrum of [15-L2+en+Fe3+
ꢀ2H2Oꢀ3H+](Figure 10a).
The mass spectrum of the duplex [9-L2-a/b+en+Cu2+
ꢀ2H2Oꢀ2H+]is shown in Figure 10b as an example. Only
the peaks calculated for the Cu–salen-containing duplex are
Figure 9. Comparison of HPL chromatograms of a) 30 mm 9-L2-a/b in
100 mm NH4OAc (pH 8) and b) the same sample after incubation with
1 mm ethylenediamine and 100 mm Cu2+. Eluent: 2 mm NHEt3OAc in
H2O/CH3CN; gradient: 0–40% CH3CN in 40 min.
+
observed, along with some Na+, K+, and NHEt3 adducts
of it. No uncomplexed single strands are visible, and not
more than one copper atom is complexed to the duplex.
The reaction of oligonucleotide duplex 13-L2-a/b with
O,O’-ethylenedihydroxyamine (edh) in the absence of metal
ions results in quantitative cross-linking to give the bis-
oxime compound [13-L2-a/b+edhꢀ2H2O](Figure 10c).
Injection of a hybridized probe containing 9-L2-a/b in
NH4OAc buffer (100 mm) onto a C18-RP column resulted in
complete denaturation of the duplex. Consequently, two
peaks, one for each single strand, were observed. When we
incubated the duplex sample with an excess of ethylenedi-
amine and Cu2+ prior to injection, only one peak was ob-
Conclusion
served. Analysis of this peak by using UV spectroscopy
during the HPLC run revealed a bathochromic shift of the
p!p* band, which is indicative of the presence of the
copper–salen complex. LC–MS analysis of the peak con-
firmed the exclusive presence of the Cu–salen duplex. This
result shows that the Cu–salen-containing DNA duplexes
are so stable that they can be isolated and purified by
HPLC.
We have synthesized DNA duplexes containing a new
metal–base-pair complex derived structurally from the well-
known salen ligand. The metal–salen-base-pair complex is
assembled in oligonucleotide duplexes containing two sali-
cylic aldehyde bases by addition of a diamine such as ethyle-
nediamine and various transition-metal ions such as Cu2+
Mn2+, Ni2+, Fe2+, or VO2+. Assembly of the salen com-
,
8714
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8708 – 8718

Metal–Salen-Base-Pair Complexes Inside DNA
FULL PAPER
Figure 10. Selected ESI mass spectra and comparison of experimental data with calculated molecular weights. a) [15-L2+en+Fe3+ꢀ2H2Oꢀ3H+] ; b) [9-
L2-a/b+en+Cu2+ꢀ2H2Oꢀ2H+] ; c) [13-L2-a/b+edhꢀ2H2O]; Adducts: * =[M+Na+Kꢀ2H], # =[M+NEt3].
Chem. Eur. J. 2006, 12, 8708 – 8718
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8715

T. Carell et al.
2-(5-Bromo-2-hydroxyphenyl)-1,3-dioxane (3a): 5-Bromosalicylic alde-
hyde (3; 6.0 g, 30 mmol) was mixed with triethyl orthoformate (5.20 mL,
plexes resulted in tremendous stabilization of the duplex, as
shown by HPLC and melting-point experiments. Two differ-
ent isomers of salicylic aldehyde nucleobases were examined
and it was found that the copper–salen complex derived
from nucleoside 2 resulted in the highest duplex stabiliza-
tion. A base pair formed by two of these ligands is able to
complex metals even in the absence of the corresponding di-
amine. In accordance with these results, we found that the
X-ray structure of 8 shows a very good geometrical fit with
natural Watson–Crick base pairs. UV measurements confirm
the reaction of two salicylic aldehyde nucleobases that face
each other in the duplex with ethylenediamine to give the
salen ligand, and addition of metal ions allows assembly of
the metal complex inside the duplex. The CD spectra of the
monomeric complex 8 and the interstrand complex show
that the salen complex exists in the DNA as a D chelate.
The sequence context of the oligonucleotide strands was
also changed to test how the preorganization geometries
affect metal complex formation. To our surprise, we ob-
served that the salen complex formation has such a strong
driving force that complex formation occurs even when the
salicylic aldehydes are not directly facing each other in the
duplex. High-resolution ESI-ICR mass spectra confirm the
formation of discrete species consisting of the DNA double
helix, one molecule of diamine, and one metal ion. The ex-
perimentally determined molecular weights match the calcu-
lated values to within a few parts per million. The reported
findings bring us another step closer to being able to exploit
the metal–salen-base-pair complex in bio-inspired nanotech-
nology.
31.2 mmol) and 1,3-propanediol (8.60 mL, 120 mmol).
A catalytic
amount of tetra-n-butylammonium tribromide (1.5 g, 3.1 mmol) was
added and the mixture was stirred for 24 h at room temperature. The re-
action was ended by adding saturated, aqueous NaHCO3 until pH 7 was
reached. The mixture was extracted twice with ethyl acetate (50 mL) and
the combined organic extracts were washed with dilute, aqueous
NaHCO3 and dried with Na2SO4. The solvents were removed in vacuo
and the raw material purified by using column chromatography (silica
gel; hexane/EtOAc=10:1) to give a colorless oil (7.46 g, 28.9 mmol,
96%). 1H NMR (400 MHz, CDCl3): d=1.51 (d, J=13.8, 1H), 2.19–2.31
(m, 1H), 4.00 (ddd, J=12.3, 12.3, 2.6 Hz, 2H), 4.30 (ddd, J=12.0, 5.0,
1.4 Hz, 2H), 5.60 (s, 1H), 6.78 (d, J=9.3 Hz, 1H), 7.28–7.34 (m, 1H),
7.86 ppm (s, 1H); 13C NMR (100 MHz, CDCl3): d=25.6, 67.5, 101.9,
111.2, 119.1, 124.0, 130.4, 133.1, 154.1 ppm; IR (film): n˜ =2972 (m), 2927
(w), 2863 (m), 2727 (w), 1724 (w), 1651 (w), 1618 (w), 1583 (w), 1485 (s),
1428 (m), 1390 (s), 1353 (m), 1277 (m), 1255 (s), 1237 (s), 1175 (m), 1151
(s), 1120 (s), 1095 (s), 1048 (w), 990 cmꢀ1 (s); EI-HRMS: m/z calcd for
C10H11BrO3 [M]+: 257.9892; found: 257.9900.
2-(5-Bromo-2-(triisopropylsilyloxy)phenyl)-1,3-dioxane (4): Compound
3a (6.7 g, 26 mmol) was dissolved in dry dichloromethane (50 mL) and
NEt(iPr)2 (11.2 mL, 65.0 mmol) was added. Triisopropylsilyl triflate
G
(8.4 g, 22 mmol) was then added dropwise at 08C. After stirring for 12 h
at room temperature, water was added and the mixture was extracted
twice with dichloromethane (50 mL). The combined organic extracts
were washed with water and saturated, aqueous NaCl and dried with
Na2SO4 . After removal of the solvents in vacuo, the resulting oil was
subjected to column chromatography (first pure hexane to elute excess
silyl reagent, then hexane/EtOAc (9:1)). The resulting colorless oil
(10.3 g, 24.8 mmol, 95%) was intensively dried under high vacuum
1
before it was used for the next step. H NMR (400 MHz, CDCl3): d=1.11
(d, J=7.3 Hz, 18H), 1.30 (sept, J=7.3 Hz, 3H), 1.42 (d, J=13.5 Hz, 1H),
2.16–2.28 (m, 1H), 3.94 (ddd, J=12.3, 12.3, 2.5 Hz, 2H), 4.23 (dd, J=
10.7, 5.0 Hz, 2H), 5.81 (s, 1H), 6.66 (d, J=8.7 Hz, 1H), 7.26 (dd, J=8.7,
2.7 Hz, 1H), 7.71 ppm (d, J=2.7 Hz, 1H); 13C NMR (100 MHz, CDCl3):
d=13.0, 18.0, 25.8, 67.5, 96.7, 113.2, 119.9, 130.5, 130.9, 132.3, 152.3 ppm;
IR (KBr): n˜ =2946 (s), 2892 (m), 2867 (s), 2724 (w), 1730 (w), 1597 (w),
1578 (w), 1486 (s), 1468 (s), 1428 (w) 1409 (m), 1391 (s), 1277 (s), 1238
(m), 1181 (m), 1151 (m), 1127 (m), 1102 (s), 1006 (s), 957 (w), 917 (s),
881 cmꢀ1 (s); ESI-HRMS: m/z calcd. for C19H32BrO3Si [M+H]+:
417.1304; found: 417.1279.
Experimental Section
Protected nucleoside 6: A solution of compound 4 (5.23 g, 12.6 mmol) in
freshly distilled diethyl ether (20 mL) was cooled to ꢀ788C and tBuLi
(16.9 mL, 26.5 mmol) in pentane (1.7m) was added dropwise over one
hour. The reaction was kept at ꢀ788C, with stirring, for 3 h and subse-
quently transferred with a cannula to a precooled (ꢀ788C) suspension of
copper(I) bromide/dimethyl sulfide complex (1.3 g, 6.3 mmol) in diethyl
ether (10 mL). The reaction mixture was carefully warmed to ꢀ308C for
20 min, whereupon the solids dissolved to give a yellow solution, which
was immediately cooled to ꢀ788C and transferred with a cannula to a
precooled solution of a-3’,5’-bistoluyl-1’-ribosyl chloride (5)[19] (1.64 g,
4.20 mmol) in dry dichloromethane (20 mL). The reaction mixture was
allowed to warm up to room temperature overnight, then saturated
NH4Cl solution (20 mL), 2m ammonia (1 mL), and diethyl ether
(100 mL) were added and the organic phases were separated. The aque-
ous phases were extracted twice with diethyl ether (100 mL) and the or-
ganic phases combined. After washing twice with water, once with satu-
rated NaCl solution, and drying over Na2SO4, the solvents were removed
in vacuo and the resulting oil was purified by using flash column chroma-
tography (silica gel; hexane/EtOAc=10:1) to give the desired b-anomer
(303 mg, 0.44 mmol, 10%), which eluted shortly before the a-anomer
(yield not determined). 1H NMR (400 MHz, CDCl3): d=1.11 (d, J=
7.3 Hz, 18H), 1.29 (sept, J=7.3 Hz, 3H), 1.37 (d, J=13.3 Hz, 1H), 2.10–
2.25 (m, 2H), 2.39 (s, 3H), 2.42 (s, 3H), 2.49 (dd, J=13.8, 5.0 Hz, 1H),
3.89–3.96 (m, 2H), 4.15–4.20 (m, 2H), 4.47–4.50 (m, 1H), 4.62 (m, 2H),
5.22 (dd, J=11.0, 4.9 Hz, 1H), 5.57 (d, J=5.6 Hz, 1H), 5.84 (s, 1H), 6.73
(d, J=8.4 Hz, 1H), 7.20–7.29 (m, 5H), 7.58 (d, J=2.3 Hz, 1H), 7.94–
7.97 ppm (m, 4H); 13C NMR (100 MHz, CDCl3): d=13.0, 18.0, 21.7, 25.9,
General: Chemicals were purchased from Sigma–Aldrich, ACROS, or
Lancaster and were used without further purification. The solvents used
were of reagent grade and were purified by using standard methods. The
reactions were monitored on Merck Silica 60 F254 TLC plates. Detection
was done by irradiation with UV light (254 nm) and staining with an
acidic 2,4-dinitrophenylhydrazine solution in ethanol. Flash chromatogra-
phy was performed on Silica 60 (Merck, 230–400 mesh). NMR spectra
were recorded on Varian Oxford 200, Bruker AC 300, Varian XL 400,
and Bruker AMX 600 spectrometers. Mass spectra were recorded on Fin-
nigan MAT 95 (EI), Bruker Autoflex II (MALDI-TOF), and Thermo
Finnigan LTQ-FT (ESI-ICR) spectrometers. IR spectra were measured
on a Nicolet 510 FTIR spectrometer in a KBr matrix or with a diamond-
ATR (Attenuated Total Reflection) setup. DNA synthesis was performed
on a PerSeptive Biosystems Expedite 8900 Synthesizer and an ¾kta Oli-
gopilot 10 (Amersham Biosciences). Analytics and purification of the oli-
gonucleotides were performed on Merck LaChrome HPLC systems with
UV and diode-array detectors using 5-m Silica-C18 RP columns and 0.1m
NHEt3OAc in H2O/CH3CN as eluent. UV spectra and melting profiles
were measured on a Cary 100 UV/Vis spectrometer and CD spectra were
measured on a JASCO J 810 CD-spectropolarimeter according to previ-
ously reported protocols.[8] The ESI spectra of DNA strands were meas-
ured in flow injection analysis mode or coupled to chromatographic sepa-
ration (eluent: 2 mm NHEt3OAc in H2O/CH3CN). In flow injection
mode, a 2-mL sample (30 mm DNA, 100 mm NH4OAc) was injected in a
steady flow of H2O/CH3CN (8:2; 200 mLminꢀ1). The capillary tempera-
ture was 3008C, with a spray voltage of 4–5 kV (negative mode).
8716
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8708 – 8718

Metal–Salen-Base-Pair Complexes Inside DNA
FULL PAPER
41.7, 65.0, 67.5, 77.4, 80.7, 82.8, 97.4, 118.4, 125.2, 127.0, 128.6, 129.1,
129.2, 129.7, 129.8, 132.8, 143.6, 144.0, 152.8, 166.1, 166.5 ppm; IR (KBr):
n˜ =2947 (m), 2867 (m), 1720 (s), 1613 (m), 1500 (m), 1466 (w), 1377 (w),
1275 (s), 1178 (m), 1150 (w), 1098 (s), 1001 (m), 906 (m), 884 cmꢀ1 (w);
ESI-HRMS: m/z calcd for C40H53O8Si [M+H]+: 689.3510; found:
689.3498.
added, and the mixture stirred for 3 h at room temperature. Then, con-
centrated HCl (200 mL) and one drop of water were added and the mix-
ture stirred for another 2 h. A further 10 mL of water was added and the
mixture was extracted three times with Et2O (20 mL). The combined or-
ganic extracts were dried with Na2SO4, the solvents removed in vacuo,
and the raw product purified by using flash column chromatography
(silica gel; CHCl3/CH3OH=9:1). The resulting brown solid was purified
by recrystallization from EtOAc to yield colorless needles (15 mg,
0.06 mmol, 32%). 1H NMR (400 MHz, CD3OD): d=1.90 (ddd, J=13.1,
10.4, 5.9 Hz, 1H), 2.26 (ddd, J=13.1, 10.4, 5.9 Hz, 1H), 3.68 (pseudo-
sept, J=5.1, 11.6 Hz, 2H), 3.98 (dt, J=5.1, 2.4 Hz, 1H), 4.32 (dt, J=5.9,
1.9 Hz, 1H), 5.13 (dd, J=10.4, 5.6 Hz, 1H), 7.03 (s, 1H), 7.06 (dd, J=8.0,
1.4 Hz, 1H), 7.66 (d, J=8.0 Hz, 1H), 9.98 ppm (s, 1H); 13C NMR
(100 MHz, CD3OD): d=44.81, 64.01, 74.28, 80.85, 89.50, 115.13, 118.50,
121.87, 134.11, 153.82, 162.79, 196.86 ppm; IR (diamond-ATR): n˜ =3262
(m), 2897 (m), 1650 (s), 1628 (s), 1434 (m), 1348 (m), 1309 (s), 1177 (m),
1153 (s), 1087 (s), 1051 (s), 988 (s), 956 (m), 874 (m), 810 (s), 680 cmꢀ1
(m); ESI-HRMS: m/z calcd for C12H13O5 [MꢀH]ꢀ: 237.0757; found:
237.0771.
Deprotected nucleoside 6a: The b-anomer of 6 (303 mg, 0.44 mmol) was
dissolved in dry methanol (7 mL) and K2CO3 (134 mg, 0.97 mmol) was
added. The suspension was stirred for 2 h at room temperature until all
the solids had dissolved. The yellow solution was diluted with chloroform
(50 mL) and water (50 mL). The aqueous phase was separated and ex-
tracted three times with chloroform (50 mL). The combined organic ex-
tracts were washed with saturated, aqueous NaCl and dried with Na2SO4.
After removal of the solvents in vacuo, the raw material was purified by
using flash column chromatography (silica gel; CHCl3/CH3OH=10:1) to
yield
a
colorless oil (86 mg, 0.19 mmol, 43%). 1H NMR (400 MHz,
CDCl3): d=1.11 (d, J=7.3 Hz, 18H), 1.29 (sept, J=7.3 Hz, 3H), 1.41 (d,
J=13.5 Hz, 1H), 1.95–2.04 (m, 1H), 2.13 (dd, J=10.2, 5.6 Hz, 1H), 2.17–
2.28 (m, 1H), 2.80 (s, 2 OH), 3.63–3.73 (m, 2H), 3.88–3.97 (m, 3H), 4.20–
4.24 (m, 2H), 4.27–4.30 (m, 1H), 5.07 (dd, J=10.2, 5.6 Hz, 1H), 5.86 (s,
1H), 6.75 (d, J=8.4 Hz, 1H), 7.16 (dd, J=8.4, 2.4 Hz, 1H), 7.54 ppm (d,
J=2.4 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=13.0, 18.0, 25.8, 43.6,
63.3, 67.6, 73.6, 79.9, 87.20, 97.4, 118.2, 125.2, 127.7, 128.5, 133.4,
152.8 ppm; IR (KBr): n˜ =2946 (m), 2868 (m), 1654 (m), 1618 (m), 1500
(m), 1466 (w), 1389 (w), 1279 (m), 1150 (w), 1127 (w), 1095 (m), 1051
(w), 1000 cmꢀ1 (w); ESI-HRMS: m/z calcd for C24H40ClO6Si [M+Cl]ꢀ:
487.2283; found: 487.2257.
Salen ligand 2b: The fully deprotected ligand 2 (45 mg, 0.19 mmol) was
dissolved in dry methanol (10 mL) and ethylenediamine (0.5 equiv,
6.32 mL, 0.095 mmol) was added. The color of the solution changed to
yellow and a microcrystalline yellow material precipitated over several
days. The reaction was also carried out in CD3OD in an NMR tube and
quantitative conversion was observed by NMR spectroscopy. 1H NMR
(400 MHz, CD3OD): d=1.90 (ddd, J=13.2, 10.4, 5.9 Hz, 1H), 2.21 (dtd,
J=12.5, 5.2, 1.7 Hz, 1H), 3.53–3.76 (m, 3H), 3.95 (s, 2H), 4.30 (m, 1H),
5.07 (dd, J=10.4, 5.6 Hz, 1H), 6.83–6.93 (m, 2H), 7.30 (dd, J=19.2,
8.0 Hz, 1H), 8.43 ppm (s, 1H); 13C NMR (100 MHz, CD3OD): d=44.85,
59.59, 63.83, 73.95, 80.63, 88.71, 114.38, 115.96, 117.88, 131.75, 147.55,
162.14, 166.12 ppm; IR (diamond-ATR): n˜ =3253 (w), 2890 (w), 2853
(w), 2428 (s), 1984 (w), 1627 (s), 1429 (m), 1372 (m), 1265 (m), 1186 (w),
1138 (m), 1089 (m), 1055 (s), 1021 (s), 976 (s), 938 (m), 898 (m), 866 (m),
816 (s), 808 (s), 755 cmꢀ1 (m); ESI-HRMS: m/z calcd for C26H33O8N2
[M+H]+: 501.2231; found: 501.2229.
DMT-protected nucleoside 6b: Compound 6a (86 mg, 0.19 mmol) was
coevaporated twice with 2 mL of dry pyridine. It was then dissolved in
1 mL of pyridine and stirred over 4 Š molecular sieves for 2 h. 4,4’-Di-
A
tion was stirred for 2 h at room temperature. Subsequently, dry methanol
(2 mL) was added, the mixture stirred for 1 h, filtered, and the solvents
removed in vacuo. Flash chromatography (silica gel; hexane/EtOAc
(9:1)+0.1% pyridine) yielded a colorless oil (79 mg, 0.11 mmol, 55%).
1H NMR (400 MHz, CDCl3): d=1.11 (d, J=7.6 Hz, 18H), 1.29–1.33 (m,
4H), 2.02–2.11 (m, 2H), 2.17 (dd, J=13.1, 5.6 Hz, 1H), 3.25–3.34 (m,
2H), 3.78 (s, 6H), 3.87–3.93 (m, 2H), 4.01 (m, 1H), 4.09–4.18 (m, 2H),
4.38 (m, 1H), 5.10 (dd, J=10.1, 5.6 Hz, 1H), 5.85 (s, 1H), 6.73 (d, J=
8.4 Hz, 1H), 6.82 (d, J=9.0 Hz, 4H), 7.17–7.30 (m, 4H), 7.36 (d, J=
9.0 Hz, 4H), 7.48 (d, J=8.3 Hz, 2H), 7.60 ppm (d, J=2.1 Hz, 1H);
13C NMR (150 MHz, CDCl3): d=13.0, 18.0, 25.8, 43.9, 55.2, 64.6, 67.4,
74.8, 79.8, 86.1, 97.4, 113.1, 118.1, 125.3, 126.7, 127.3, 127.8, 128.3, 128.6,
130.1, 134.0, 136.2, 144.9, 152.5, 158.4 ppm; IR (KBr): n˜ =1719 (m), 1654
(m), 1618 (m), 1560 (w), 1542 (w), 1508 (m), 1458 (w), 1272 (m),
1097 cmꢀ1 (m); ESI-HRMS: m/z calcd for C45H59O8Si [M+H]+: 755.3979;
found: 755.3961.
Cu–salen complex 8: A solution of ligand 2b (50 mg, 0.10 mmol) in dry
methanol (5 mL) was combined with a methanolic solution of [Cu(acac)2]
G
(26 mg, 0.10 mmol) and heated under reflux for 10 min. The color
changed from yellow to green to purple. Slow cooling of a saturated
methanolic solution yielded small, dichroic green-purple crystals which
were used for crystallographic examination. IR (diamond-ATR): n˜ =3305
(m), 2919 (w), 2888 (w), 1634 (s), 1614 (s), 1526 (s), 1482 (m), 1427 (s),
1387 (m), 1322 (s), 1302 (m), 1312 (m), 1187 (m), 1064 (s), 1038 (s), 998
(s), 966 (s), 959 (s), 873 (s), 795 cmꢀ1 (s); ESI-HRMS: m/z calcd for
C26H31CuN2O8 [M+H]+: 562.1371; found: 562.1369. For crystallographic
data, see ref. [11].
Phosphoramidite 7: Compound 6b (66 mg, 87 mmol) was coevaporated
twice with 2 mL of dry THF and finally dissolved in 2 mL of degassed
DNA synthesis, cleavage, and purification: DNA synthesis was per-
formed using Ultramild Bases and reagents (Glen Research) and follow-
ing standard phosphoramidite protocols. The coupling times and amounts
of ligands could be reduced to match the parameters for coupling the
natural bases. The trityl values showed good incorporation of the modi-
fied nucleoside. After additional treatment with 2% dichloroacetic acid
plus 1% H2O in dichloromethane to remove the acetal protecting
groups, the controlled-pore-size glass (CPG) solid support was treated
with concentrated, aqueous NH3/EtOH (3:1) for 12 h at room tempera-
ture to cleave the strands. The solvents were removed in a SpeedVac con-
centrator and the pellet redissolved in doubly distilled water. Analysis
and purification was performed on Merck LaChrome HPLC systems
using 5m Silica-C18 RP columns and 0.1m NHEt3OAc in H2O/CH3CN
(typically 0–40% CH3CN in 40 min) as eluent. Prior to HPLC purifica-
tion, 20% HOAc was added and the mixture incubated for 20 min. The
purified fractions were concentrated, desalted on Waters Sepac-C18 car-
tridges, and concentrated again. The concentration was estimated by UV
spectroscopy following standard procedures, taking into account the
molar extinction coefficient for the ligand 2 (e=10290 Lmolꢀ1 cmꢀ1).
THF. NEt(iPr)2 (31 mL, 170 mmol) and (iPr2N)(NCCH2CH2O)PCl (28 mL,
A
130 mmol) were then added and the reaction mixture was stirred for 2 h.
The solvents were removed in vacuo and the residue was dissolved in de-
gassed EtOAc (1 mL) and purified by using column chromatography
under a protecting gas atmosphere (deactivated silica gel; hexane/EtOAc
(2:1)+0.1% pyridine; all solvents degassed). The solvent was removed
under high vacuum to yield a mixture of diastereomers as a colorless oil
(76 mg, 80 mmol, 92%), which was immediately used in DNA synthesis.
1H NMR (200 MHz, CDCl3): d=1.05–1.20 (m, 30H), 1.25–1.33 (m, 4H),
1.96–2.12 (m, 2H), 2.23–2.29 (m, 1H), 2.44 (t, J=6.5 Hz, 2H), 2.61 (t, J=
6.5 Hz, 2H), 3.21–3.32 (m, 2H), 3.46–3.74 (m, 2H), 3.77 (s, 3H), 3.78 (s,
3H), 3.90–3.97 (m, 2H), 4.11–4.24 (m, 3H), 4.43–4.45 (m, 1H), 5.10 (dd,
J=10.6, 5.0 Hz, 1H), 5.86 (s, 1H), 6.73–6.84 (m, 5H), 7.23–7.31 (m, 4H),
7.35–7.41 (m, 4H), 7.48–7.52 (m, 2H), 7.64 ppm (d, J=2.3 Hz, 1H);
31P NMR (80 MHz, CDCl3): d=149.0, 148.6 ppm; ESI-HRMS: m/z calcd
for C54H76N2O9PSi [M+H]+: 955.5058; found: 955.5083.
Deprotected ligand nucleoside 2: b-1’-(4-[1,3]Dioxan-2-yl-3-(triisopropyl-
silyloxy)phenyl)-2’-desoxyribose (85 mg, 0.17 mmol) was dissolved in dry
THF (2 mL), tetrabutylammonium fluoride (1.7 equiv, 1.1m in THF) was
Chem. Eur. J. 2006, 12, 8708 – 8718
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8717

T. Carell et al.
Feringa, Angew. Chem. 2005, 117, 3294; Angew. Chem. Int. Ed.
2005, 44, 3230.
Acknowledgments
[7]K. Tanaka, A. Tengeiji, T. Kato, N. Toyama, M. Shionoya, Science
2003, 299, 1212.
[8]G. H. Clever, K. Polborn, T. Carell, Angew. Chem. 2005, 117, 7370;
Angew. Chem. Int. Ed. 2005, 44, 7204.
We thank the Volkswagen Foundation (Priority program: Complex Mate-
rials), the DFG (SFB 486), and the Fonds der chemischen Industrie for a
KekulØ Fellowship to G.H.C.
[9]R. X.-F. Ren, N. C. Chaudhuri, P. L. Paris, S. Rumney IV, E. T. Kool,
J. Am. Chem. Soc. 1996, 118, 7671.
[1]J. D. Watson, F. H. C. Crick, Nature 1953, 171, 737.
[10]G. ZemplØn, E. Pascu, Ber. Dtsch. Chem. Ges. 1929, 62, 1613.
[11]CCDC-602317 contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
[2]a) L. Orgel, Science 2000, 290, 1306; b) A. Eschenmoser, R. Krishna-
murthy, S. Guntha, P. Scholz, K.-U. Schçning, Science 2000, 290,
1347.
[3]R. Micura, R. Kudick, S. Pitsch, A. Eschenmoser, Angew. Chem.
1999, 111, 715; Angew. Chem. Int. Ed. 1999, 38, 680.
[12]a) D. Hall, T. N. Waters, J. Chem. Soc. 1960, 2644; b) L. C. Nathan,
J. E. Koehne, J. M. Gilmore, K. A. Hannibal, W. E. Dewhirst, T. D.
Mai, Polyhedron 2003, 22, 887; c) L. Dyers Jr., S. Y. Que, D. Van-
Derveer, X. R. Bu, Inorg. Chim. Acta 2006, 359, 197.
[13]W. Saenger, Principles of Nucleic Acid Structures, Springer, New
York, 1984, pp. 9–27.
[14]S. Akine, T. Taniguchi, T. Nabeshima, Chem. Lett. 2001, 7, 682.
[15]a) S. Zolezzi, A. Decinti, E. Spodine, Polyhedron 1999, 18, 897;
b) R. Klement, F. Stock, H. Elias, H. Paulus, P. Pelikµn, M. Valko,
M. Mazfflr, Polyhedron 1999, 18, 3617.
[16]R. S. Downing, F. L. Urbach, J. Am. Chem. Soc. 1969, 91, 5977.
[17]a) J. H. Banoub, R. P. Newton, E. Esmans, D. F. Ewing, G. Macken-
zie, Chem. Rev. 2005, 105, 1869; b) S. A. Hofstadler, R. H. Griffey,
Chem. Rev. 2001, 101, 377; c) J. L. Beck, M. L. Colgrave, S. F. Ralph,
M. M. Sheil, Mass Spectrom. Rev. 2001, 20, 61.
[4]a) N. J. Leonard, Acc. Chem. Res. 1982, 15, 128; b) H. Liu, J. Gao,
S. R. Lynch, Y. D. Saito, L. Maynard, E. T. Kool, Science 2003, 302,
868.
[5]a) C. Brotschi, C. J. Leumann, Nucleosides, Nucleotides Nucleic
Acids 2003, 22, 1195; b) L. Zhang, E. Meggers, J. Am. Chem. Soc.
2005, 127, 74; c) N. Zimmerman, E. Meggers, P. G. Schultz, J. Am.
Chem. Soc. 2002, 124, 13684; d) E. Meggers, P. L. Holland, W. B.
Tolman, F. E. Romesberg, P. G. Schultz, J. Am. Chem. Soc. 2000,
122, 10714; e) T. Tanaka, A. Tengeiji, T. Kato, N. Toyama, M. Shiro,
M. Shionoya, J. Am. Chem. Soc. 2002, 124, 12494; f) K. Tanaka, Y.
Yamada, M. Shionoya, J. Am. Chem. Soc. 2002, 124, 8802; g) C.
Switzer, S. Sinha, P. H. Kim, B. D. Heuberger, Angew. Chem. 2005,
117, 1553; Angew. Chem. Int. Ed. 2005, 44, 1529; h) C. Switzer, D.
Shin, Chem. Commun. 2005, 1342; i) H. Weizman, Y. Tor, J. Am.
Chem. Soc. 2001, 123, 3375; j) D. Popescu, T. Parolin, C. Achim, J.
Am. Chem. Soc. 2003, 125, 6354; k) Y. Miyake, H. Togashi, M. Ta-
shiro, H. Yamaguchi, S. Oda, M. Kudo, Y. Tanaka, Y. Kondo, R.
Sawa, T. Fujimoto, T. Machinami, A. Ono, J. Am. Chem. Soc. 2006,
128, 2172.
[18]J. F. Larrow, E. N. Jacobsen, J. Org. Chem. 1994, 59, 1939.
[19]M. Hoffer, Chem. Ber. 1960, 93, 2777.
[6]a) S. Klug, M. Famulok, Mol. Biol. Rep. 1994, 20, 97; b) D. Zha, A.
Eipper, M. T. Reetz, ChemBioChem. 2003, 4, 34; c) G. Roelfes, B. L.
Received: April 20, 2006
Published online: September 19, 2006
8718
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8708 – 8718