Discrimination in metal-ion binding to RNA dinucleotides with a non-bridging oxygen or sulfur in the phosphate diester link

Article abstract of DOI:10.1002/chem.200701491

Replacement of a non-bridging oxygen in the phosphate diester bond by a sulfur has become quite popular in nucleic acid research and is often used as a probe, for example, in ribozymes, where the normally essential Mg²⁺ is partly replaced by a

Full text of DOI:10.1002/chem.200701491

DOI: 10.1002/chem.200701491
Discrimination in Metal-Ion Binding to RNADinucleotides with a Non-
Bridging Oxygen or Sulfur in the Phosphate Diester Link+
Bernd Knobloch,[a] Barbara Nawrot,[b] Andrzej Okruszek,[b, c] and Roland K. O. Sigel*[a]
Abstract: Replacement of a non-bridg-
ing oxygen in the phosphate diester
bond by a sulfur has become quite pop-
ular in nucleic acid research and is
often used as a probe, for example, in
ribozymes, where the normally essen-
tial Mg2+ is partly replaced by a thio-
philic metal ion to reactivate the
system. Despite these widely applied
rescue experiments no detailed studies
exist quantifying the affinity of metal
ions to such terminal sulfur atoms.
Therefore, we performed potentiomet-
ing pUpU3ꢀ complexes. The primary
binding site in both dinucleotides is the
terminal phosphate group. Theoretical-
ly, also the formation of 10-membered
chelates involving the terminal oxygen
or sulfur atoms of the (thio)phosphate
bridge is possible with both ligands.
The results show that Mg2+ and Mn2+
exist as open (op) isomers binding to
both dinucleotides only at the terminal
as the S-coordinated macrochelate,
Cd
(pUp(S)U)ꢀcl=PS Zn2+ forms with
pUp three isomeric species, that
(pUp(S)U)ꢀop, Zn(pUp(S)U)cꢀl=PO
and Zn
(pUp(S)U)ꢀcl=PS, which occur to
about 33, 12 (O-bound), and 55%, re-
spectively. Pb2+ forms the 10-mem-
bered chelate with both nucleotides in-
volving only the terminal oxygen atoms
of the (thio)phosphate bridge, that is,
no indication of S binding was discov-
ered in this case. Hence, Zn2+ and
Cd2+ show pronounced thiophilic prop-
erties, whereas Mg2+, Mn2+, and Pb2+
coordinate to the oxygen, macrochelate
formation being of relevance with Pb2+
only.
A
.
3ꢀ
U
(S)
is, Zn
A
G
,
AHCTREUNG
phosphate group. Whereas Cd
only exists as Cd
(pUpU)ꢀop,
Cd(pUp(S)U)ꢀ is present to about 64%
A
AHCTREUNG
ric pH titrations to determine the bind-
3ꢀ
ing properties of pUp
U
(S)
towards
Keywords:
dinucleotides
·
Mg2+, Mn2+, Zn2+, Cd2+, and Pb2+
,
metal ions · ribozymes · RNA ·
thiophosphates
and compared these data with those
previously obtained for the correspond-
1. Introduction
popular alteration is the substitution of one of the oxygens
by a sulfur atom, such compounds being first synthesized in
1966.[3] Initially these derivatives were employed in studies
regarding the mechanisms of enzymatic reactions[4,5] for
The phosphate group of nucleotides has been altered in
many ways (see citations in refs. [1,2]). Probably the most
[a] Dr. B. Knobloch, Prof. Dr. R. K. O. Sigel##
Institute of Inorganic Chemistry, University of Zürich
Winterthurerstrasse 190, 8057 Zürich (Switzerland)
Fax : (+41)44-635-6802
E-mail: roland.sigel@aci.uzh.ch
[b] Prof. Dr. B. Nawrot,# Prof. Dr. A. Okruszek##
[##] Correspondence regarding the synthesis of pUp(S)U.
[+] Abbreviations and definitions (see also Figures 1 and 2): I, ionic
strength; Ka, general acidity constant; L, general ligand; M2+, general
divalent metal ion; MeOPS2ꢀ, methyl thiophosphate; pUpU3ꢀ and
pUp(S)U
3ꢀ, see Figure 1; RibMP2ꢀ, d-ribose 5-monophosphate; R-
PO32ꢀ, simple phosphate monoester or phosphonate ligand with R
Department of Bioorganic Chemistry
Centre of Molecular and Macromolecular Studies
Polish Academy of Sciences, Sienkiewicza 112
representing a non-interacting residue; UMP, uridine 5’-monophos-
phate; UMPS2ꢀ
, uridine 5’-O-thiomonophosphate. Species written
without a charge either do not carry one or represent the species in
general (i.e., independent of their protonation degree); which of the
two possibilities applies is always clear from the context. A formula
like (pUp(S)UꢀH)4ꢀ means that the dinucleotide has lost a proton
´
90-363 Łódz (Poland)
E-mail: bnawrot@bio.cbmm.lodz.pl
okruszek@bio.cbmm.lodz.pl
[c] Prof. Dr. A. Okruszek##
from one of its (N3)H sites and that it has to be read as “pUp(S)
U
minus H”.
Institute of Technical Biochemistry
Faculty of Biotechnology and Food Sciences
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the authors.
´
Technical University of Łódz, Stefanowskiego 4/10
´
90-924 Łódz (Poland)
[#] Correspondence regarding the equilibrium measurements.
3100
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3100 – 3109

FULL PAPER
which purpose they are still used.[6,7] However, more recent-
ly, with the development of the antisense strategy[8,9] and the
observation that oligonucleotides with a sulfur modification
at the phosphodiester linkage are usually more resistant
toward nuclease degradation than natural oligonucleo-
tides,[10,11] their application has widened including also
drugs.[12] A further important aspect is the increasing focus
on RNA-catalyzed reactions.[13–16] Here, thio analogues are
important tools[17] for studying, for example, the role of
metal ions in ribozyme folding and activity[18–20] as well as in
attempts to understand the mechanisms of ribozyme reac-
tions[21] and to identify those phosphate oxygen atoms which
are important for catalysis.[22,23] Similar experiments have
also been recently conducted with a thioderivative of the
10–23 DNAzyme to identify a metal-ion binding site within
the catalytic core.[24]
Considering the above indicated wide interest in thiophos-
phate derivatives, surprisingly few investigations have been
conducted to determine the basic differences between phos-
phate and thiophosphate groups, such as the effects of hy-
drogen bonding to the sulfur on the ribozyme cleavage reac-
tion[25] or the stabilizing effects on DNA–RNA triplex for-
mation.[26,27] In all instances metal ions need to be present at
the very least for charge neutralization. With regard to ki-
netically labile divalent metal ions (M2+) and thio deriva-
tives of nucleotides only a few reports exist, that is, of ade-
nosine 5’-mono-[1,2], di-, and triphosphate.[28] In addition, a
more recent study dealt with the interaction of M2+ with ur-
idine 5’-O-thiomonophosphate (UMPS2ꢀ) and methyl thio-
phosphate (MeOPS2ꢀ).[29] These experiments proved that
the uracil residue is not involved in metal-ion binding in
placed by a S atom. The main binding site in such a dinucle-
otide is the terminal phosphate group, but chelate formation
with the neighboring thiophosphate diester bridge is at least
theoretically possible and therefore information can be
gained on the metal-ion affinity of this second site.
In order to make the interpretation of the experimental
data unequivocal, we selected a dinucleotide with nucleo-
bases of low metal-ion affinity. As indicated above, the
uracil residue is such a nucleobase and hence we studied
now the metal-ion binding properties of pUp(S)U3ꢀ towards
Mg2+, Mn2+, Zn2+, Cd2+, and Pb2+. Comparison with the
experimental data obtained earlier[36] for pUpU3ꢀ allowed us
to answer questions regarding the selectivity of biologically
relevant metal ions like Mg2+ or Zn2+ for the S atom of the
thiophosphate bridge. Here, we quantify the formation de-
grees of the 10-membered chelates formed by the different
metal ions in their complexes with the mentioned two dinu-
cleotides (Figure 1) and determine the extent of O or S co-
Figure 1. Chemical structures of uridylyl-(5’!3’)-[5’]-uridylate (pUpU3ꢀ
)
M
A
R
and of its thio derivative P-thiouridylyl-(5’!3’)-[5’]-uridylate
(pUp(S)U
3ꢀ). The two uridine units in each structure are shown in their
means that there will be no interference of the uracil moiet-
ies.
dominating anti conformation.[37,38] For the charge distribution in the
thiophosphate diester bridge see also Section 2.1.
ꢀ
dominating phosphate group in nucleic acids and corre-
sponding thio substitutions, that is, -O-P(O)(S)ꢀ-O-, are
widely employed in ribozyme studies.[15,16,19–23,31] Hence, it is
surprising to find that, to the best of our knowledge, not a
single study exists where the intrinsic affinity of a terminal
(non-bridging) sulfur atom in such a bridge towards divalent
metal ions was quantified. Considering further our own in-
terests in metal-ion–ribozyme interactions,[32–35] we made
now an effort to measure such affinities.
ordination, respectively, in those cases where macrochelate
formation occurs. Indeed, the observed selectivity and dis-
crimination is considerable in several respects: For example,
there is no chelate formation in Mg
A
Zn
AHCTREUNG
chelates (S and O bound), a formation degree significantly
larger than the one observed[36] for Zn
amounts to about 26% only.
A
It is very difficult to directly measure the metal-ion affini-
ty of such a -O-P(O)(S)ꢀ-O- site in a polymer or even in an
oligonucleotide, such as UpUpU2ꢀ. The reason being that
competition for binding between a metal ion and a proton
occurs neither within the experimentally accessible nor
within the physiological pH range because the primary pro-
tons of phosphate or thiophosphate residues are released
with pKa values of about one (or even below).[29,36] There-
fore, we measured the metal-ion affinity of such a thiophos-
phate diester unit by using a dinucleotide containing both a
terminal 5’-phosphate group and a 5’!3’-phosphate diester
bridge in which one of the two terminal O atoms was re-
2. Results and Discussion
2.1. Acid–base properties of pUp(S)
U
3ꢀ: This dinucleotide
with a terminal S atom at the diester bridge (Figure 1) can
accept a total of three protons at its (thio)phosphate groups.
Two of these protons are released at very low pH, which fol-
lows from pKa = 1.0Æ0.3 estimated for the release of a
proton from the P(O)(OH)2 group of H2
(pUpU)ꢀ.[36] For
A
H
ACHTREUNG
Chem. Eur. J. 2008, 14, 3100 – 3109
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3101

R. K. O. Sigel et al.
H3
(pUp(S)U) because substitution of an O atom in a phos-
The small increase in acidity by about 0.1 pK unit for the
phate group by an S atom increases the acidity.[1,29,39] The
proton is thereby most likely bound at the terminal oxygen
atom of this bridge.[2,39] After deprotonation the negative
charge of the thiophosphate bridge is mainly located at the
sulfur atom,[39–41] as is depicted in Figure 1.
terminal -P(O)2(OH)ꢀ group in going from H(pUpU)2ꢀ to
N
H
(pUp(S)U)2ꢀ is more difficult to explain and probably a sol-
vation effect: The nearby S atom of the thiophosphate
bridge is expected to be less effective in hydrogen bonding
than it is the case with an O atom.
A
It follows from the above that for the present study three
Interestingly, the difference in acidity between the two
(N3)H sites is small and the same within the error limits for
deprotonation reactions of the H
(pUp(S)U)2ꢀ species are of
A
relevance; that is, the release of the “second” proton from
the terminal phosphate group as well as the two protons
from the uracil (N3)H sites. This leads to the following de-
protonation equilibria:
the two dinucleotides pUpU3ꢀ (DpKa(O)
=
(9.63Æ
3ꢀ
0.08)ꢀ(8.99Æ0.03) = 0.64Æ0.09) and pUp
U
(DpKa(S)
=
(S)
(9.98Æ0.12)ꢀ(9.29Æ0.04)=0.69Æ0.13) (see also Table 1).
These differences are identical within the error limits to the
one expected for symmetrical diprotonic acids (H2L), where
the difference amounts to 0.6 pK units.[36] The observation
2ꢀ
HðpUpðSÞUÞ Ð pUpðSÞU3ꢀ þ Hþ
ð1aÞ
ð1bÞ
ð2aÞ
ð2bÞ
ð3aÞ
that the two uracil residues in
a given dinucleotide
KHHðpUp
¼ ½pUpðSÞ
U
3ꢀꢁ½Hþꢁ=½HðpUpðSÞUÞ2ꢀꢁ
(Figure 1) have practically identical acidic properties means
that these residues do not significantly affect each other.
Another interesting aspect is the observation that the
presence of the S atom in the diester bridge increases the
basicity of all (N3)ꢀ sites as the comparisons of the pKa/2
and pKa/3 values according to Equilibria (2a) and (3a)
shows: DpKa/2 = (9.29Æ0.04)ꢀ(8.99Æ0.03) = 0.30Æ0.05
UÞ
ðSÞ
pUpðSÞU3ꢀ Ð ðpUpðSÞUꢀHÞ þ Hþ
4ꢀ
KHpUp ¼ ½ðpUpðSÞUꢀHÞ4ꢀꢁ½Hþꢁ=½pUpðSÞ
U
3ꢀꢁ
U
ðSÞ
4ꢀ
5ꢀ
ðpUpðSÞUꢀHÞ ÐðpUpðSÞUꢀ2 HÞ þ Hþ
and DpKa/3
=
(9.98Æ0.12)ꢀ(9.63Æ0.08)
=
0.35Æ0.14
KHðpUp
¼ ½ðpUpðSÞUꢀ2 HÞ5ꢀꢁ½Hþꢁ=½ðpUpðSÞUꢀHÞ4ꢀꢁ
(Table 1). This effect is most likely due to the larger hydro-
phobicity of pUp(S)U, compared to that of pUpU, as well as
to the localization of the negative charge of the diester
bridge mainly on the S atom in pUp(S)U whereas in pUpU
this charge is equally distributed between two O atoms
(Figure 1). Hence, the introduction of a bridging thiophos-
phate group into a nucleic acid also affects the acid–base
properties of the neighboring nucleobases to a certain
degree. Interestingly, this effect is only observed with bridg-
ing (thio)phosphate groups, but not with a terminal phos-
phate group (where the charge distribution differs less), as
can be seen by comparison of UMP2ꢀ (pKUHMP = 9.45Æ0.02;
Table 1) and UMPS2ꢀ (pKUHMPS = 9.47Æ0.02).[29] In the
latter case, only the release of the proton from the phos-
phate group is strongly affected, that is, pKHHðUMPÞ = 6.15Æ
0.01 versus pKHHðUMPSÞ = 4.78Æ0.02.[29]
UꢀHÞ
ðSÞ
ð3bÞ
The expression (pUp(S)UꢀH)4ꢀ in Equilibrium (2a) should
be read as “pUp(S)U minus H”, meaning that one of the two
(N3)H sites has lost a proton, without defining which one.
Analogously, in the species (pUp(S)Uꢀ2H)5ꢀ both (N3)H
sites are deprotonated.
The acidity constants for the Equilibria (1a), (2a), and
(3a) were measured by potentiometric pH titrations (see Ex-
perimental Section). The results are listed in Table 1 where-
by the site attributions are evident from the given related
data.[30,36,42,43]
With regard to the release of the final proton from the
phosphate groups, the data show that H
A
basic than H
ACHTREUNG
to 6.15). This increased basicity of the terminal phosphate
group of pUpU3ꢀ compared to that of UMP2ꢀ is clearly a
charge effect and in accord with related observations.[44]
2.2. Stabilities of M
stants of several M
potentiometric pH titrations. All experimental data can be
perfectly explained by taking Equilibrium (1a) as well as the
following complex-forming Equilibrium (4a) into account:
ACHTREUNG
AHCTREUNG
Table 1. Negative logarithms of the acidity constants for the deprotona-
tion of the P(O)2(OH)ꢀ and (N3)H sites in H(pUp(S)U)2ꢀ [Eqs. (1–3)], to-
G
gether with some related data, as determined by potentiometric pH titra-
tions in aqueous solution (258C; I=0.1m, NaNO3).[a,b]
M2þ þ pUpðSÞU3ꢀ Ð MðpUpðSÞUÞꢀ
ð4aÞ
ð4bÞ
Acids
pKa of the sites
Refs.
P(O)2(OH)ꢀ
(N3)H
[29,45]
[46]
(RibMP)ꢀ
uridine
6.24Æ0.01
KMMðpUp
¼ ½MðpUpðSÞUÞꢀꢁ=ð½M2þꢁ½pUpðSÞ
U
3ꢀꢁÞ
H
UÞ
ðSÞ
9.18Æ0.02
9.45Æ0.02
[29,45]
[34]
H
H
H
C
6.15Æ0.01
6.44Æ0.02
6.32Æ0.03
8.99Æ0.03/9.63Æ0.08
The data evaluation was restricted to the pH range in which
neither hydroxo complexes are formed nor (N3)H is depro-
tonated, the pH range in which the formation of hydroxo
complexes occurs being evident from the titrations in the
absence of ligand (see Section 4.4).
9.29Æ0.04/9.98Æ0.12
–
[a] The errors given are three times the standard error of the mean value
or the sum of the probable systematic errors, whichever is larger. [b] So-
called practical, mixed, or Brønsted acidity constants[45] are listed (see
also Section 4.3).
3102
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3100 – 3109

Metal-Ion Binding to RNA Dinucleotides
FULL PAPER
The stabilities of the five metal-ion complexes studied are
listed for comparison in column 3 of Table 2, along with the
corresponding values of the M
A
dently, the stabilities of the M
E
Mg2+ and Mn2+ are hardly affected by the presence or ab-
sence of an S atom in the diester bridge, whereas the com-
plexes of Zn2+ or Cd2+ become more stable if S is present
and those with Pb2+ show a reduced stability.
To be able to explain the above observations, a more rig-
orous evaluation procedure is required to elucidate the
structures of the M
N
N
solution. It is well known that straight lines are obtained for
a series of related ligands by plotting logKMMðLÞ versus
[1,47]
pKHHðLÞ
Such correlation lines are available for complexes
formed between several divalent metal ions (M2+) and
simple phosphate monoester or phosphonate ligands
2ꢀ [42,48]
(R-PO3
)
and the parameters according to the straight-
line Equation (5), have been listed.[36,42,46]
logKMMðR-PO Þ ¼ pKHHðR-PO Þ ꢂ m þ b
ð5Þ
3
3
The data pairs of the M
A
,
Zn2+, and Pb2+ systems, on which the parameters for Equa-
tion (5) are based, are shown together with the correspond-
ing data points for the M2+/pUp(S)U3ꢀ and pUpU3ꢀ systems
in Figure 2. For all six complexes an increased stability is ob-
served. However, this stability varies considerably depend-
ing on both the metal ion and the dinucleotide involved in
complex formation: In the case of Zn2+ the thio derivative
forms the more stable complex whereas with Pb2+ the
pUpU3ꢀ complex is more stable.
*
Figure 2. Evidence for an enhanced stability of some M
and
logKMMðR-PO and pKHHðR-PO for M
(pUp(S)U)ꢀ
(
)
M
ACHTREUNG
AHCTREUNG
3Þ
3Þ
2ꢀ
*
phosphate monoester and phosphonate ligands (R-PO3 ) ( ): (from left
to right) 4-nitrophenyl phosphate (NPhP2ꢀ), phenyl phosphate (PhP2ꢀ),
uridine 5’-monophosphate (UMP2ꢀ), d-ribose 5-monophosphate
(RibMP2ꢀ), thymidine [1-(2’-deoxy-b-d-ribofuranosyl)thymine 5’-mono-
phosphate] (dTMP2ꢀ), n-butyl phosphate (BuP2ꢀ), methanephosphonate
(MeP2ꢀ), and ethanephosphonate (EtP2ꢀ). The least-squares lines
[Eq. (5)] are drawn through the corresponding eight data sets (*) taken
from ref. [30] for the phosphate monoesters and from ref. [48] for the
phosphonates. The corresponding straight-line parameters are listed in
refs. [36,46] and [48]. The data points due to the M2+/H+/pUpU3ꢀ sys-
tems (ꢃ) are from ref. [36] and those for the M2+/H+/pUp(S)U3ꢀ systems
*
A more quantitative evaluation is possible by applying
Equation (5) with its parameters[36] together with
pKHHðpUp UÞ = 6.32 (Table 1). The results for these calcula-
ðsÞ
(
) are based on the constants in Tables 1 and 2. The vertical broken
tions are listed in column 4 of Table 2 representing the sta-
lines emphasize the stability differences from the reference lines as de-
fined by Equation (6) (see also Table 2, column 5). All plotted equilibri-
um constants refer to aqueous solutions at 258C and I=0.1m (NaNO3).
bility constants logKMMðR-PO Þ of M
C
3
Table 2. Comparison of the stability constants of the M
(pUp(S)U)ꢀ complexes between the measured stability
A
group that has the basicity of
constants [Eq. (4)] and the calculated stability constants for M(R-PO3) species, based on the basicity of the ter-
E
minal phosphate group of pUp(S)U3ꢀ (pKHHðpUp UÞ =6.32) and the reference-line Equation (5) with its corre-
the terminal phosphate group
in pUp(S)U3ꢀ, that is, no addi-
tional interaction occurs. Com-
parison of these data with the
measured stabilities demon-
strates an enhanced stability for
ðSÞ
sponding parameters,[36, 46] together with the stability differences logDM/pUp
as defined in Equation (6). For
(S)U
comparison the corresponding data for the M2+/pUpU3ꢀ systems are also listed (aqueous solution; 258C; I=
0.1m, NaNO3).[a]
Ligand
M2+
logKMMðpUp
logKMMðR-PO
logDM=pUp
U
ðSÞ
UÞ
3Þ
ðSÞ
Mg2+
Mn2+
Zn2+
Cd2+
Pb2+
1.85Æ0.08
2.42Æ0.07
2.88Æ0.07
3.16Æ0.07
3.49Æ0.11
1.59Æ0.03
2.19Æ0.05
2.16Æ0.06
2.48Æ0.05
2.99Æ0.08
0.26Æ0.08
0.23Æ0.09
0.72Æ0.09
0.68Æ0.09
0.50Æ0.14
all five M
studied.
A
pUp(S)U3ꢀ
2.3. Quantification of the en-
hanced stabilities of the
(pUp(S)U)ꢀ complexes and
Mg2+
Mn2+
Zn2+
Cd2+
Pb2+
1.84Æ0.04
2.49Æ0.05
2.57Æ0.03
2.75Æ0.03
4.45Æ0.20
1.61Æ0.03
2.22Æ0.05
2.20Æ0.06
2.52Æ0.05
3.05Æ0.08
0.23Æ0.05
0.27Æ0.07
0.37Æ0.07
0.23Æ0.05
1.40Æ0.22
M
ACHTREUNG
pUpU3ꢀ
extent of their chelate forma-
tion: The stability differences
between the measured values
[a] For the error limits, see footnote [a] of Table 1. The error limits (3s) of the derived data in column 5 were
calculated according to the error propagation after Gauss.
for the M
(pUp(S)U)ꢀ complexes
A
Chem. Eur. J. 2008, 14, 3100 – 3109
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3103

R. K. O. Sigel et al.
and the calculated values for the M
tained by using Equation (6):
G
tion of the intramolecular Equilibrium (8) of the
M
(pUp(S)U)ꢀ species, between an open (op) and a closed
A
(cl) or chelated isomer, varies depending on the metal ion
involved.
logDM=pUp
¼ logKMMðpUp
ꢀ logKMMðR-PO Þ
ð6Þ
U
UÞ
ðSÞ
ðSÞ
3
MðpUpðSÞUÞꢀop Ð MðpUpðSÞUÞꢀcl
ð8Þ
These differences are listed in column 5 of Table 2 together
with the results obtained previously[36] for the corresponding
The position of Equilibrium (8) is defined by the dimension-
less intramolecular equilibrium constant KI [Eq. (9)]:
M
ACHTREUNG
A
and Cd2+, the stability enhancements are identical within
the error limits (see Table 2, column 5 in the lower part).
Considering the different coordinating properties[49–51] of
these three metal ions it is evident that their increased sta-
bility is simply due to the charge effect by going from
KI ¼ ½MðpUpðSÞUÞꢀclꢁ=½MðpUpðSÞUÞoꢀpꢁ
ð9Þ
which is related to the stability enhancement logD* [Eq. (7)]
by Equation (10):[1,47]
M
(R-PO3) to M
(pUpU)ꢀ. In other words, the metal ion co-
N
KI ¼ 10logD ꢀ1
*
ð10Þ
ordinated to the terminal phosphate group in pUpU3ꢀ
“feels” the presence of the negative charge located on the
neighboring phosphate diester bridge. This charge effect is
represented by the average of the logDM/pUpU values, defined
in analogy to Equation (6), for the Mg2+, Mn2+, and Cd2+
systems, logDM/pUpU/charge =0.24Æ0.04. It is reasonable to
assume that exactly the same charge effect is present in the
Knowledge of KI allows calculation of the formation degree
of the closed species in Equilibrium (8) by using Equation
(11):
% MðpUpðSÞUÞꢀcl ¼ 100 ꢂ KI=ð1 þ KIÞ
ð11Þ
The results for KI and % M
(pUp(S)U)ꢀcl are listed in Table 3
in columns 6 and 7, respectively.
G
corresponding M
stability increase in M
must be attributed to an additional interaction of the metal
ACHTREUNG
N
ACHTREUNG
ion already coordinated to the terminal phosphate group of
As already stated, in the case of several M
(pUp(S)U)ꢀ and
G
3ꢀ
pUp
U
(S)
or pUpU3ꢀ. This increase is defined for
M
(pUpU)ꢀ systems, the values of logD* are zero within the
U
M
(pUp(S)U)ꢀ complexes by Equation (7):
error limits. However, within these error limits, traces of
chelated species might form with the corresponding upper
limits as given in parenthesis in Table 3 (column 6). In con-
*
logD ¼ logDM=pUp
ꢀ logDM=pUpU=charge
ꢀ ð0:24 Æ 0:04Þ
ð7aÞ
ð7bÞ
U
U
ðSÞ
trast, chelates for the M
C
*
logD ¼ logDM=pUp
ðSÞ
esting to note that the formation degree for Cd
(close to) zero but reaches about 64% for Cd
ACHTREUNG
As discussed in the Introduction, the only other available
binding site in both pUpU3ꢀ and pUp(S)U3ꢀ is the (thio)-
phosphate diester bridge, which allows the formation of a
10-membered chelate (Figure 1) involving both, the terminal
phosphate and the bridging (thio)phosphate groups.
The values for logD* according to Equation (7) are listed
in column 4 of Table 3 for the
ACHTREUNG
Similarly, Zn
A
about 67% which is much larger than the approximately
26% determined earlier[36] for Zn(pUpU)cꢀl. This proves the
G
known high thiophilicity of these two metal ions.[49–51] The
situation with Pb2+ is more surprising because % Pb-
M
A
M
(pUpU)ꢀ
R
Table 3. Extent of chelate formation in M
A
complexes. As expected, these
values are zero within the error
hancement logD* [Eq. (7)] and quantified by the dimensionless equilibrium constant KI [Eqs. (9, 10)], and the
percentage of the chelated isomer [Eq. (11)]. The corresponding data of the M
ACHTREUNG
limits for the M
C
listed for comparison (aqueous solution; 258C; I=0.1m, NaNO3).[a]
,
Mn2+
, and
[b]
Ligand
M2+
logDM/pUp
logD*
ACHTREUNG
(S)U
Cd2+. This also holds for the
Mg2+ and Mn2+ complexes of
the thio analogue pUp(S)U3ꢀ, il-
lustrating that Mn2+ does not
Mg2+
Mn2+
Zn2+
Cd2+
Pb2+
0.26Æ0.08
0.23Æ0.09
0.72Æ0.09
0.68Æ0.09
0.50Æ0.14
0.02Æ0.09
ꢀ0.01Æ0.10
0.48Æ0.10
0.44Æ0.10
0.26Æ0.15
ꢄ0
ꢄ0 (<22)
ꢄ0 (<19)
67Æ8
ꢄ0
pUp(S)U3ꢀ
2.02Æ0.68
1.75Æ0.62
0.82Æ0.61
64Æ8
45Æ18
remarkably
coordinate
the
sulfur atom. In all other instan-
ces the logD* values are clearly
positive. The different stability
enhancements for the various
Mg2+
Mn2+
Zn2+
Cd2+
Pb2+
0.23Æ0.05
0.27Æ0.07
0.37Æ0.07
0.23Æ0.05
1.40Æ0.26
ꢀ0.01Æ0.06
0.03Æ0.08
0.13Æ0.08
ꢀ0.01Æ0.06
1.16Æ0.26
ꢄ0
ꢄ0
ꢄ0 (<11)
ꢄ0 (<22)
26Æ14
pUpU3ꢀ
0.35Æ0.25
ꢄ0
ꢄ0 (<11)
M
(pUp(S)U)ꢀ and
M
(pUpU)ꢀ
E
13.45Æ8.65
93Æ4
complexes mean that the posi- [a] For the error limits see footnote [a] of Table 2. [b] These values are from column 5 of Table 2.
3104
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3100 – 3109

Metal-Ion Binding to RNA Dinucleotides
FULL PAPER
A
(pUp(S)U)ꢀcl (Table 3; see also discussion
Or, in other words, logD* encompasses the total (tot)
amount of all chelated complexes and the definition given
in Equation (13) holds.
in Section 2.4), that is, macrochelate formation is diminished
upon introduction of the sulfur atom.
2.4. Amore detailed appraisal of the M
(pUp(S)U)ꢀ chelates:
G
½MðpUpðSÞUÞꢀcl=totꢁ ¼ ½MðpUpUðSÞUÞꢀcl=POꢁ þ ½MðpUpðSÞUÞꢀcl=PSꢁ
ð13Þ
Considering that the thiophosphate diester bridge involved
in chelate formation as discussed in Section 2.3 has not only
a terminal S atom but also a terminal O atom (see
Figure 1), it is evident that the situation is ambiguous with
regard to which of the two atoms participates in the chelate.
In this context the following three points need to be noted:
i) The stacking tendency of uracil residues is very small[52]
and therefore, no preferred stacked conformer of pUp(S)U3ꢀ
is expected to occur in aqueous solution. ii) In accord here-
In analogy to Equation (9), one can then define Equa-
tion (14):
½MðpUpðSÞUÞꢀcl=totꢁ
KI=tot
¼
¼
ð14aÞ
½MðpUpðSÞUÞꢀopꢁ
½MðpUpðSÞUÞꢀcl=POꢁ þ ½MðpUpðSÞUꢁcꢀl=PSꢁ
with, the acidity constants of the (N3)H sites indicate that
ð14bÞ
ð14cÞ
½MðpUpðSÞUÞꢀopꢁ
3ꢀ
the two uracil residues in pUp
U
(S)
do not “feel” each
other (see Section 2.1). iii) The thiophosphate diester bridge
¼ 10logD ꢀ1
*
3ꢀ
in pUp
U
(S)
involves five single bonds which means that
the two nucleotide residues may rotate rather freely around
these five bonds. Hence, one has to conclude that both, the
S or the O atom of the thiophosphate bridge can be in-
volved in chelate formation and that the right part of the
following intramolecular Equilibrium (12) needs to be con-
sidered:
As KI/tot equals KI [as defined in Eq. (10)] and Equation (11)
providing % M
(pUp(S)U)ꢀcl is still valid, this means that in
AHCTREUNG
accord with Equilibrium (12) the following definitions can
be written:[53]
KI=PO ¼ ½MðpUpðSÞUÞꢀcl=POꢁ=½MðpUpðSÞUÞoꢀpꢁ
KI=PS ¼ ½MðpUpðSÞUÞꢀcl=PSꢁ=½MðpUpðSÞUÞoꢀpꢁ
KI=tot ¼ KI=PO þ KI=PS
ð15Þ
ð16Þ
ð17Þ
By subtracting % M
tions (14a) and (14b), from 100% the formation degree of
the open species, M
(pUp(S)U)ꢀop, can be calculated. The re-
sults for KI/tot, % M
(pUp(S)U)ꢀcl=tot, and % M(pUp(S)U)oꢀp are
G
In this equilibrium M
(pUp(S)U)ꢀop designates the “open”
N
listed in Table 4 in columns 2, 3, and 4, respectively. In those
cases where one of the closed isomers, for example,
complex [see also Eq. (8)], in which the metal ion is only
bound to the terminal phosphate group. The chelated or
“closed” isomers can involve either the O or S atom of the
thiophosphate diester bridge, giving the species
M
(pUp(S)U)ꢀcl=PO, is not formed, KI/PO [Eq. (15)] becomes
A
zero and Equation (17) reduces to a two-isomer problem as
discussed in Section 2.3.
M
A
E
As a consequence the charge-corrected stability enhance-
ments logD* [Eq. (7)] (see also Section 2.3 and column 4 of
With those metal ions, where all three isomers are formed
according to Equilibrium (12), KI/tot and the concentration
Table 3) reflect all possible additional interactions that a
fractions of M
G
N
metal ion coordinated to the terminal phosphate group in
determined [Eqs. (14) and (11)] as shown above. KI/PS values
can be calculated by assuming that the extent of chelate for-
3ꢀ
pUp
U
(S)
may experience according to Equilibrium (12).
Table 4. Formation degrees of the isomeric species M
C
G
G
M2+
KI/tot
[Eq. (14)[b]
% M
G
M
G
KI/PO
[Eq. (15)][e]
KI/PS
[Eqs.
% M
A
% M
(pUp(S)U)ꢀcl=PS
E
A
A
E
G
(16, 17)]
A
ACHTREUNG
Mg2+
Mn2+
Zn2+
Cd2+
Pb2+
ꢄ0
ꢄ0
ꢄ0
ꢄ0
100
100
33Æ8
36Æ8
55Æ18
ꢄ0
ꢄ0
ꢄ0
ꢄ0
ꢄ0
ꢄ0
ꢄ0
ꢄ0
2.02Æ0.68
1.75Æ0.62
0.82Æ0.61
67Æ8
64Æ8
45Æ18
0.35Æ0.25
ꢄ0
1.67Æ0.72
1.75Æ0.62
ꢄ0
12
55
64
0
ꢄ0
45[g
(13.45Æ8.65)[f]
[a] See footnote [a] of Table 2. [b] Values from column 5 of Table 3 (upper part); see text. [c] From column 6 in Table 3 (upper part). [d] These values
follow from 100ꢀ% M
(pUp(S)U)ꢀcl=tot. [e] These values are from column 5 of Table 3 (lower part); see text. [f] This value is larger than KI/tot =0.86; hence,
no meaningful calculation is possible. [g] See discussion in Section 2.4.
ACHTREUNG
Chem. Eur. J. 2008, 14, 3100 – 3109
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3105

R. K. O. Sigel et al.
mation involving the oxygen of the bridging (thio)phosphate
group is the same in M
(pUpU)ꢀcl and M(pUp(S)U)cꢀl=PO spe-
cies. Values for KI/PS (Table 4) are then obtained from Equa-
tion (17) by using KI/PO values of the M
(pUpU)ꢀcl species
sulfur. Mn2+ is not suitable because it does not own a pro-
nounced thiophilicity.[1,15,16] In the few cases, where Mn2+
has an effect in such experiments, this is most likely due to
its larger global stability constants observed for phosphate
complexes compared to those of the corresponding Mg2+
complexes (see also Table 2). These conclusions are con-
firmed by the metal-ion-promoted hydrolytic cleavage reac-
tion of the Sp and Rp diastereomers of the phosphoromono-
G
ACHTREUNG
AHCTREUNG
(see lower part in Table 3).
Our experimental data described in Section 2.3 shows that
Mg2+ and Mn2+ form only M(pUp(S)U)oꢀp isomers. The same
T
is true for Cd
covered.[36] Hence, the complete stability enhancement ob-
served for Cd
(pUp(S)U)ꢀ is to be attributed to the formation
of the Cd
(pUp(S)U)ꢀcl=PS isomer, or in other words, in this
case only the lower pathway of Equilibrium (12) operates.
For Zn
(pUp(S)U)ꢀ clearly both pathways of Equilibrium
(12) are in action and the three isomers Zn
(pUp(S)U)ꢀop,
Zn
(pUp(S)U)ꢀcl=PO, and Zn(pUp(S)U)cꢀl=PS occur with formation
A
thioate analogues of uridylyl
A
U
cleavage (to 2’,3’-cUMPS) is significantly accelerated by
Zn2+ and Cd2+, the rate enhancements being almost equal
with the Sp and Rp diastereomers. The effect of Mn2+ and
Mg2+ on the cleavage rate is, in turn, very modest.[54]
ACHTREUNG
ACHTREUNG
E
The high affinity of Pb2+ towards
O sites is well
A
N
known[49–51,55] and, for example, exhibited in the large stabili-
ty of the G-quadruplex formed by guanine residues into
which a Pb2+ ion is inserted.[56] In the resulting (G)8–Pb2+
coordination pattern Pb2+ sits between two G-quartets
being coordinated to eight (C6)O carbonyls.[56] As a conse-
quence, Pb2+ also shows a pronounced ability to interact
strongly with two neighboring phosphate sites and it is not
surprising to find that leadzymes could be isolated[57,58] and
degrees of about 33, 12, and 55%, respectively (Table 4),
confirming the pronounced preference for S over O, that is,
the thiophilicity of Zn2+
.
The situation with Pb2+ is considerably more complicated
because from the entries in Table 3 it follows that the stabili-
ty enhancement logD* is larger for Pb
A
that for Pb
ACHTREUNG
ingful calculation is possible because KI/PO > KI/tot. From
this observation it follows that in these complexes the affini-
ty of Pb2+ for O is higher than for S. Considering that in Pb-
that Pb2+-dependent DNAzymes[59] exist. Because Pb2+ is a
[49,50,60]
well known mimic of Ca2+
,
one may add that for this
alkaline earth metal ion it is not expected that it has any af-
finity towards sulfur, but due to its size it could be that, in
contrast to Mg2+, it may bind favorably to two neighboring
phosphate groups in a nucleic acid, just like Pb2+ does.
Finally, Pb2+ exhibits one more fascinating characteristic
that is revealed by its coordination chemistry to thiophos-
phate derivatives. Pb2+ appears to be a chameleon-like
metal ion because its binding properties seem to depend on
the first strongly coordinating site, that is, a directing effect
(pUp(S)U)ꢀ only one terminal O in the thiophosphate bridge
is available for chelate formation compared to two in
Pb
(pUpU)ꢀ, the stability enhancement of the latter, logD*
= 1.16Æ0.26 (Table 3, column 4) may be reduced by the
ACHTREUNG
*
statistical factor of one half, that is, the expected logDexp for
Pb
(pUp(S)U)ꢀ amounts then to 0.86Æ0.26. However, this
value is still much larger than the determined value, logD*=
0.26Æ0.15 (Table 3), for Pb
(pUp(S)U)ꢀ. The reason for this
ACHTREUNG
AHCTREUNG
discrepancy can be found in the fact that the negative
charge of the thiophosphate bridge is mainly located at the
terminal S atom and not on the terminal O atom (see also
Section 2.1). Consequently, the affinity of this O atom to-
wards Pb2+ is lower than that of the two non-bridging
oxygen atoms in pUpU3ꢀ which each carry a charge of ꢀ0.5.
Hence, in all likelihood Pb2+ forms with pUp(S)U3ꢀ a 10-
membered chelate involving only oxygen atoms.
of the first ligand is observed: In such cases, where only one
2þ
binding site is present, Pb shows a preference for sulfur, as
aq
evident from the comparison of, for example, UMPS2ꢀ with
UMP2ꢀ.[2,29] In contrast, when two binding sites are present
in a ligand enabling macrochelate formation, as is the case
with pUpU3ꢀ or pUp(S)U3ꢀ, no such thiophilic behavior is
observed anymore: If coordination occurs first to a phos-
phate group then oxygen binding is further favored despite
the presence of the neighboring sulfur in the bridging thio-
phosphate group of pUp(S)U3ꢀ.
Based on this finding in the Pb
ACHTREUNG
may conclude that also in the above discussed Zn
ACHTREUNG
species the affinity of Zn2+ towards the neutral non-bridging
oxygen of the thiophosphate is reduced (compared to
Zn
for Zn
ered an upper limit and consequently the value for
Zn
(pUp(S)U)ꢀcl=PS a lower limit.
(pUpU)ꢀ). Hence, the formation degree of 12% given
Experimental Section
A
4.1. Synthesis of P-thiouridylyl-(5’!3’)-[5’]-uridylate, pUp(S)U3ꢀ (1): This
dinucleotide was synthesized in three different ways with all three com-
pounds showing the same metal-ion binding properties. Dinucleotide 1
was obtained as a mixture of two diastereomers in ca. 2:1 ratio (synthetic
route I) and 1:1 ratio (synthetic routes II and III), having opposite con-
figuration at the phosphorus atom.
ACHTREUNG
3. Conclusion
4.1.1. Synthetic route I: The trisodium salt of pUp(S)U3ꢀ was prepared by
a multistep synthesis (see Scheme S1, Supporting Information) using the
phosphoramidite methodology, with the 1-(2-fluorophenyl)-4-methoxypi-
peridin-4-yl (Fpmp) group for protection of the 2’-hydroxy functionali-
ty.[61,62] Thus, 5’-O-dimethoxytrityl-2’-O-Fpmp uridine-3’-O-(2-cyanoethyl-
From this study it follows that in so-called thio rescue ex-
periments,[15,16] being often conducted in ribozyme chemis-
try,[18–23] only Zn2+ and Cd2+ can effectively be applied as
rescuing agents due to their pronounced affinity towards
3106
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3100 – 3109

Metal-Ion Binding to RNA Dinucleotides
FULL PAPER
N,N-diisopropylphosphoramidite) (2) was prepared as previously de-
scribed[62] and characterized by 31P NMR (Bruker Avance spectrometer,
200 MHz) showing two singlets at d=150.24 and 152.13 ppm (in CD3CN)
(ratio of diastereomers of 3:2). Compound 2 was reacted with 2’,3’-di-O-
acetyluridine in the presence of 1H-tetrazole in CH2Cl2 solution to yield
the fully protected dinucleoside phosphorothionate triester (3) upon ad-
dition of elemental sulfur to the intermediate phosphite (31P NMR of 3
(in CD3CN); d=68.40 and 68.77 ppm (ratio 2:3 of phosphorothionate
diastereomers)). After selective removal of the dimethoxytrityl (DMT)
group with 2% dichloroacetic acid in CH2Cl2, the resulting 5’-hydroxyl
compound was 5’-O-phosphorylated with bis-O,O-(2-cyanoethyl)-N,N-di-
isopropylphosphoramidite reagent and the phosphite intermediate subse-
quently oxidized with I2/pyridine/H2O.[63] The resulting fully protected 5’-
O-phosphorylated dinucleoside-phosphorothionate triester (4) was puri-
MALDI-TOF signal at m/z 812 corresponding to incompletely deprotect-
ed 9. The latter product was again treated with aqueous ammonia (28%,
2 mL) for 48 h at +48C and purified as before to give also 10. Both frac-
tions of 10 were combined and, after concentration to dryness, treated
with (C2H5)3N·3HF (0.2 mL, 1.2 mmol; Aldrich, Steinheim, Germany)
for 24 h at room temperature. The reaction mixture was dissolved in
water and HPLC purified on a C18 reverse phase column (PRP-1, Ham-
ilton). Pure thio-dinucleotide 1 was isolated by elution with buffer A (1m
TEAB buffer, pH 7.4) with a gradient of buffer B (40% acetonitrile in
1m TEAB buffer, pH 7.4), 0–100% in 30 min. The obtained yield was
5.5 mg (25%). MALDI-TOF MS gave a single signal of m/z 645.3 in neg-
ative ions (MW 646.41).
4.1.3. Synthetic route III: Compound 9 was alternatively synthesized by
routine automated solid phase synthesis[66] using standard 2’-TBDMS
phosphoramidite CE monomers and a chemical phosphorylating reagent
(Glen Research, Sterling VA, USA). Deprotection of 9 to give pure 1
was then done according to the procedure described for Synthetic route
II.
fied by silica gel chromatography (31P NMR of
4 (CD3CN); d=
ꢀ1.32 ppm (phosphotriester group) 67.85 and 68.22 ppm (ratio 2:3, phos-
phorothionate diastereomers)). Compound 4 was then subjected to step-
wise deprotection, including 16 h incubation in 30% aqueous ammonia at
558C (removal of 2-cyanoethyl and acetyl groups) followed by 12 h treat-
ment with 0.01m HCl (pH 2.0) at room temperature (removal of the
Fpmp group).
4.2. Other materials: Nitric acid (HNO3), the nitrate salts of Na+, Mg2+
,
Mn2+, Zn2+, Cd2+, and Pb2+, disodium ethylenediamine-N,N,N’,N’-tet-
raacetate dihydrate (Na2H2EDTA·2H2O), potassium hydrogen phthalate
(all pro analysi), and sodium hydroxide (NaOH) solution (Titrisol) were
purchased from Merck, Darmstadt, Germany. The buffer solutions used
(pH 4.00, 7.00, 9.00) were traceable to standard reference materials
(SRM) of the US National Institute of Science and Technology (NIST)
and purchased from Metrohm, Herisau, Switzerland. All solutions were
prepared using deionised, ultra pure (Milli-Q185 Plus; from Millipore,
Molsheim, France) CO2-free water.
The crude dinucleotide 1 was purified by ion-exchange chromatography
on DEAE Sephadex A-25 (elution with a linear gradient of triethylam-
monium bicarbonate from 0.1 to 0.6m). Purified 1 was then transformed
into its trisodium salt by passing through Dowex 50Wx8 (Na+ form) and
lyophylized to give a white solid in 17% overall yield (based on 2’,3’-di-
O-acetyluridine). The structure of 1 was confirmed by spectroscopic
methods: 31P NMR (D2O), d=0.74 ppm (phosphate group), 56.45, and
56.69 ppm (ratio ca. 2:1, phosphorothioate diastereomers); FAB MS (Fin-
nigan MAT 95): m/z: 645.2 (negative ions), calculated MW 646.41 for the
free acid. Analytical RP HPLC of the product 1 showed two peaks with
retention times of 13.32 and 14.05 min in a ca. 2:1 ratio.
The concentrations of the NaOH solutions were determined with potassi-
um hydrogen phthalate, those of the stock solutions of divalent metal
ions by potentiometric pH titrations via their EDTA complexes. The
stock solutions of (pUp(S)U)3ꢀ were freshly prepared daily and the pH of
the solutions was adjusted close to 8.0 with sodium hydroxide. The exact
concentration of the ligand solutions was determined in each experiment
by evaluation of the corresponding titration pair, that is, the differences
in NaOH consumption between solutions with and without ligand (see
below).
4.1.2. Synthetic route II (see also Scheme S2, Supporting Information):
Anhydrous 2’,3’-di-O-acetyluridine (6) (93 mg, 0.28 mmol) was dissolved
in anhydrous acetonitrile (1.5 mL) and mixed with ethylthio-1H-tetrazole
(50 mg, 0.38 mmol), to which a solution of 2-cyanoethyl-N,N-diisopropyl-
phosphoramidite of 2’-O-tert-butyldimethylsilyl-5’-O-dimethoxytrityluri-
dine (5) (270 mg, 0.32 mmol; Glen Research, Sterling VA, USA) in anhy-
drous acetonitrile (1.5 mL) was added dropwise. After stirring the reac-
tion mixture for 2 h anhydrous sulfur (13.5 mg, 0.053 mmol S8) was
added. The reaction mixture was left overnight, the solvent evaporated
and the residue subsequently chromatographed on a silica gel 60H short
column with chloroform elution. 270 mg (95%) of pure 7 were obtained.
This compound (200 mg, 0.18 mmol) was treated with 50% acetic acid
(15 mL) for 30 min and after concentration purified by means of silica
gel column chromatography. Dimer 8 was obtained in 58% yield (80 mg;
spectral data: FAB MS: m/z: 763.3 [MꢀH]+; MW 764).
4.3. Potentiometric pH titrations: The pH titrations were performed with
a E536 potentiograph connected to a E665 dosimat and a 6.0253.100
Aquatrode-plus combined macro glass electrode (all from Metrohm, Her-
isau, Switzerland). The instruments were calibrated using the buffer solu-
tions mentioned above. The acidity constants determined at I=0.1m
(NaNO3) and 258C are so-called practical, mixed or Brønsted con-
stants,[45] which may be converted into the corresponding concentration
constants by subtracting 0.02 from the measured pKa values.[45] The ionic
product of water (Kw) is not included in our calculations because the dif-
ferences in NaOH consumption between solutions with and without
Anhydrous
8 (0.05 mmol, 40 mg) and ethylthio-1H-tetrazole (9 mg,
ligand are evaluated.[45,67] The stability constants of the M
complexes are, as usual, concentration constants.
A
0.07 mmol) were dissolved in anhydrous acetonitrile (1.0 mL). Then a so-
lution of the chemical phosphorylation reagent 2-[2-4,4’-dimethoxytrityl-
oxy)ethylsulfonyl]ethyl-(2-cyanoethyl)-(N,N-diisopropyl)-phosphorami-
dite[64] (35 mg, 0.053 mmol; Glen Research, Sterling VA, USA) in anhy-
drous acetonitrile (0.25 mL) was added. After stirring the reaction mix-
ture for 30 min, a 5m solution of tert-butylperoxide in decane (80 mL,
0.4 mmol) was added. The reaction was chromatographed on a silica gel
60H short column with chloroform/methanol (0!3%) elution yielding
45 mg (64%) of pure 9. The structure of 9 was confirmed by MALDI-
TOF MS: m/z: 1334.2 (MW 1334) and 31P NMR: d=ꢀ2.21, ꢀ2.26, 67.75,
67.29 ppm.
4.4. Determination of the equilibrium constants: The acidity constants
KHHðpUp UÞ, KHpUp U, and KHðpUp
of H
(pUp(S)U)2ꢀ [Eqs. (1–3)] were de-
C
UꢀHÞ
ðSÞ
ðSÞ
The experimental data were evaluated with a curve-fitting procedure
using a Newton–Gauss non-linear least-squares program by employing
every 0.1 pH unit the difference in NaOH consumption between the
mentioned pair of titrations, that is, with and without ligand. The acidity
For deprotection, compound 9 (45 mg, 0.03 mmol) was treated with a
mixture of 20% ammonia in ethanol (1 mL) and mercaptoethanol
(50 mL, 0.72 mmol).[65] After 1 h, more ethanolic ammonia (1 mL) and
aqueous ammonia (28%, 3 mL) were added to the reaction mixture and
left overnight at +48C. A precipitate was filtered off and the remaining
solution concentrated, the residue dissolved in water and purified over a
DEAE Sephadex A-25 column. Two fractions were isolated: The slower
fraction was the main product 10 giving the expected MALDI-TOF MS
constants of H
10.4, corresponding to about 17% neutralization (initial) for the equilib-
rium (pUp(S)U)
2ꢀ (pUp(S)U)3ꢀ and about 72% (final) for
(pUp(S)UꢀH) [Eq. (1)],
4ꢀ (pUp(S)Uꢀ2H)5ꢀ. The final result for KHHðpUP
is the average of eight independent pairs of titrations; those for KpHUp
A
H
A
/
ACHTREUNG
/
ACHTREUNG
UÞ
ðSÞ
U
ðSÞ
m/z signal at 759.2 whereas the faster eluting fraction exhibited
a
[Eq. (2)], and KðHpUP
[Eq. (3)] are the averages of four independent
UꢀHÞ
ðSÞ
Chem. Eur. J. 2008, 14, 3100 – 3109
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3107

R. K. O. Sigel et al.
pairs of titrations only because the low concentrations (0.07 mm) were ti-
trated only up to pH 7.
[10] M. Gilar, A. Belenky, Y. Budman, D. L. Smisek, A. S. Cohen, Anti-
sense Nucleic Acid Drug Dev. 1998, 8, 35–42.
[11] F. Eckstein, Antisense Nucleic Acid Drug Dev. 2000, 10, 117–121.
[12] K. L. Warfield, R. G. Panchal, M. J. Aman, S. Bavari, Curr. Opin.
Mol. Ther. 2006, 8, 93–103.
[13] R. F. Gesteland, T. R. Cech, J. F. Atkins, The RNA World, Cold
Spring Harbor Press, 2006, pp. 1–768.
[14] R. K. O. Sigel, A. M. Pyle, Met. Ions Biol. Syst. 2003, 40, 477–512.
[15] R. K. O. Sigel, Eur. J. Inorg. Chem. 2005, 2281–2292.
[16] R. K. O. Sigel, A. M. Pyle, Chem. Rev. 2007, 107, 97–113.
[17] R. K. O. Sigel, H. Sigel, Met. Ions Life Sci. 2007, 2, 109–180.
[18] P. M. Gordon, J. A. Piccirilli, Nat. Struct. Biol. 2001, 8, 893–898.
[19] M. Boudvillain, A. de Lencastre, A. M. Pyle, Nature 2000, 406, 315–
318.
At the end of each titration a small volume (about 0.8 mL or less) of
0.1m HNO3 was added to the solutions to restore the initial pH of about
3.3. A further comparatively small volume of a M =
(NO3)2 solution (M2+
G
Mg2+, Mn2+, Zn2+, Cd2+, Pb2+) was subsequently added and the titra-
tions were repeated. The total volume of these solutions was approxi-
mately 35 mL with an ionic strength I varying between 0.095 and 0.13m.
This small variation in I has no effect on complex stability[36] as is also
evident from the following titrations. This means that the stability con-
stants of the Mg2+, Mn2+, Zn2+, and Pb2+ systems were additionally de-
termined under the same conditions used for the acidity constants, but
NaNO3 was partly (Mn2+, Zn2+, Pb2+) or fully (Mg2+) replaced by
M(NO3)2 (258C; I=0.1m).
T
[20] M. Boudvillain, A. M. Pyle, EMBO J. 1998, 17, 7091–7104.
[21] T. E. Horton, M. Maderia, V. J. DeRose, Biochemistry 2000, 39,
8201–8207.
[22] S.-O. Shan, A. V. Kravchuk, J. A. Piccirilli, D. Herschlag, Biochem-
istry 2001, 40, 5161–5171.
[23] S.-O. Shan, A. Yoshida, S. Sun, J. A. Piccirilli, D. Herschlag, Proc.
Natl. Acad. Sci. USA 1999, 96, 12299–12304.
[24] B. Nawrot, K. Widera, M. Wojcik, B. Rebowska, G. Nowak, W. J.
Stec, FEBS J. 2007, 274, 1062–1072.
[25] T. Lçnnberg, M. Ora, S. Virtanen, H. Lçnnberg, Chem. Eur. J. 2007,
13, 4614–4627.
[26] P. Guga, M. Boczkowska, M. Janicka, A. Maciaszek, B. Nawrot, S.
Antoszczyk, W. J. Stec, Pure Appl. Chem. 2006, 78, 993–1002.
[27] P. Guga, M. Boczkowska, M. Janicka, A. Maciaszek, S. Kuberski,
W. J. Stec, Biophys. J. 2007, 92, 2507–2515.
[28] V. L. Pecoraro, J. D. Hermes, W. W. Cleland, Biochemistry 1984, 23,
5262–5271.
[29] C. P. Da Costa, A. Okruszek, H. Sigel, ChemBioChem 2003, 4, 593–
602.
The metal-to-ligand ratios in the various titrations, with [pUp(S)U] being
about 0.07 or 0.20 mm, were 233:1 and 170:1 for Mg2+; 285:1 and 113:1
for Mn2+; 116:1, 36:1, and 25:1 for Zn2+; 285:1, 27:1, and 24:1 for Cd2+
,
and 58:1, 22:1, and 15:1 for Pb2+. Even though the metal-to-ligand ratios
vary widely, the calculated stability constants for the M2+ complexes
showed no dependence on the excess of M2+ used.
The titration data were evaluated with a curve-fitting procedure using a
Newton–Gauss non-linear least-squares program for each titration pair
(i.e., with and without ligand) by calculating an apparent acidity constant
Ka’. Depending on the metal ion, the evaluation commenced at a forma-
tion degree of the M
(pUp(S)U)ꢀ species of about 4 to 12%, while the
R
upper limit was given either by the hydrolysis of M(aq)2+ or by the for-
mation of complexes, where the (N3)H sites of the uracil residues lost a
proton, which was evident from the titrations without ligand or by the de-
viation of the experimental data from the calculated curves. Representa-
tive examples for the pH ranges employed are in the case of the
M
ACHTREUNG
AHCTREUNG
[30] S. S. Massoud, H. Sigel, Inorg. Chem. 1988, 27, 1447–1453.
[31] N. J. Reiter, L. J. Nikstad, A. M. Allmann, R. J. Johnson, S. E.
Butcher, RNA 2003, 9, 533–542.
A
G
ACHTREUNG
ACHTREUNG
[32] M. C. Erat, O. Zerbe, T. Fox, R. K. O. Sigel, ChemBioChem 2007, 8,
306–314.
[33] R. K. O. Sigel, D. G. Sashital, D. L. Abramovitz, A. G. Palmer III,
S. E. Butcher, A. M. Pyle, Nat. Struct. Mol. Biol. 2004, 11, 187–192.
[34] R. K. O. Sigel, A. Vaidya, A. M. Pyle, Nat. Struct. Biol. 2000, 7,
1111–1116.
A
are the averages of two independent titrations in the case of the Mg2+
and Mn2+ systems, whereas for the Zn2+, Cd2+, and Pb2+ systems three
independent pairs of titrations were performed.
[35] E. Freisinger, R. K. O. Sigel, Coord. Chem. Rev. 2007, 251, 1834–
1851.
Acknowledgements
[36] B. Knobloch, D. Suliga, A. Okruszek, R. K. O. Sigel, Chem. Eur. J.
2005, 11, 4163–4170.
[37] K. Aoki, Met. Ions Biol. Syst. 1996, 32, 91–134.
[38] D. B. Davies, P. Rajani, H. Sadikot, J. Chem. Soc. Perkin Trans. 2
1985, 279–285.
[39] B. Song, R. K. O. Sigel, H. Sigel, Chem. Eur. J. 1997, 3, 29–33.
[40] P. A. Frey, R. D. Sammons, Science 1985, 228, 541–545.
[41] P. A. Frey, Adv. Enzymol. Relat. Areas Mol. Biol. 1989, 62, 119–201.
[42] C. P. Da Costa, H. Sigel, J. Biol. Inorg. Chem. 1999, 4, 508–514.
[43] B. Knobloch, W. Linert, H. Sigel, Proc. Natl. Acad. Sci. USA 2005,
102, 7459–7464.
We thank Professor Helmut Sigel, University of Basel, for helpful sugges-
tions regarding this manuscript and Genowefa Nowak for the excellent
technical assistance during chemical synthesis of the dinucleotides. Finan-
cial support from the Swiss National Science Foundation (SNF-Fçrder-
ungsprofessur to R.K.O.S.) within the COST D39 program from the
Swiss State Secretariat for Education and Research, as well as from the
State Committee for Scientific Research in Poland are gratefully ac-
knowledged.
[44] B. Knobloch, H. Sigel, A. Okruszek, R. K. O. Sigel, Org. Biomol.
Chem. 2006, 4, 1085–1090.
[45] H. Sigel, A. D. Zuberbühler, O. Yamauchi, Anal. Chim. Acta 1991,
255, 63–72.
[46] H. Sigel, L. E. Kapinos, Coord. Chem. Rev. 2000, 200–202, 563–594.
[47] R. B. Martin, H. Sigel, Comments Inorg. Chem. 1988, 6, 285–314.
[1] R. K. O. Sigel, B. Song, H. Sigel, J. Am. Chem. Soc. 1997, 119, 744–
755.
[2] C. P. Da Costa, D. Krajewska, A. Okruszek, W. J. Stec, H. Sigel, J.
Biol. Inorg. Chem. 2002, 7, 405–415.
[3] F. Eckstein, J. Am. Chem. Soc. 1966, 88, 4292–4294.
[4] J. R. Knowles, Annu. Rev. Biochem. 1980, 49, 877–919.
[5] F. Eckstein, Annu. Rev. Biochem. 1985, 54, 367–402.
[6] J. Liu, M. D. Tsai, Biochemistry 2001, 40, 9014–9022.
[7] M. Koziolkiewicz, A. Owczarek, M. Wojcik, K. Domanski, P. Guga,
W. J. Stec, J. Am. Chem. Soc. 2002, 124, 4623–4627.
[8] W. S. Marshall, M. H. Caruthers, Science 1993, 259, 1564–1569.
[9] A. DeMesmaeker, R. Haner, P. Martin, H. E. Moser, Acc. Chem.
Res. 1995, 28, 366–374.
´
ˇ
[48] H. Sigel, D. Chen, N. A. Corfœ, F. Gregan, A. Holy, M. Straak, Helv.
Chim. Acta 1992, 75, 2634–2656.
[49] R. B. Martin in Encyclopedia of Inorganic Chemistry, Vol. 4 (Ed.:
R. B. King), Wiley, Chichester, 1994, pp. 2185–2196.
[50] R. B. Martin in Encyclopedia of Molecular Cell Biology and Molecu-
lar Medicine, Vol. 1 (Ed.: R. A. Meyers), VCH, Weinheim, 1996,
pp. 125–134.
[51] R. B. Martin, Met. Ions Biol. Syst. 1986, 20, 21–65.
3108
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3100 – 3109

Metal-Ion Binding to RNA Dinucleotides
FULL PAPER
[52] H. Sigel, R. Griesser, Chem. Soc. Rev. 2005, 34, 875–900.
[53] B. Knobloch, H. Sigel, A. Okruszek, R. K. O. Sigel, Chem. Eur. J.
2007, 13, 1804–1814.
[54] M. Ora, M. Peltomaki, M. Oivanen, H. Lçnnberg, J. Org. Chem.
1998, 63, 2939–2947.
[55] H. Sigel, C. P. Da Costa, R. B. Martin, Coord. Chem. Rev. 2001,
219–221, 435–461.
[56] F. W. Kotch, J. C. Fettinger, J. T. Davis, Org. Lett. 2000, 2, 3277–
3280.
[61] M. V. Rao, C. B. Reese, V. Schehlmann, P. S. Yu, J. Chem. Soc.
Perkin Trans. 1 1993, 43–55.
[62] B. Beijer, I. Sulston, B. S. Sproat, P. Rider, A. I. Lamond, P. Neuner,
Nucleic Acids Res. 1990, 18, 5143–5151.
[63] E. Uhlmann, J. Engels, Tetrahedron Lett. 1986, 27, 1023–1026.
[64] M. S. Urdea, T. Horn, Tetrahedron Lett. 1986, 27, 2933–2936.
[65] L. Zhang, L. Sun, Z. Cui, R. L. Gottlieb, B. Zhang, Bioconjugate
Chem. 2001, 12, 939–948.
[66] M. H. Caruthers, Science 1985, 230, 281–285.
[57] T. Pan, O. C. Uhlenbeck, Biochemistry 1992, 31, 3887–3895.
[58] T. Pan, O. C. Uhlenbeck, Nature 1992, 358, 560–563.
[59] A. K. Brown, J. Li, C. M. B. Pavot, Y. Lu, Biochemistry 2003, 42,
7152–7161.
[67] M. Bastian, H. Sigel, J. Coord. Chem. 1991, 23, 137–154.
´
[68] C. A. Blindauer, T. I. Sjꢁstad, A. Holy, E. Sletten, H. Sigel, J. Chem.
Soc. Dalton Trans. 1999, 3661–3671.
[60] R. B. Martin in Handbook on Toxicity of Inorganic Compounds
(Eds.: H. G. Seiler, H. Sigel, A. Sigel), Marcel Dekker, New York,
1988, pp. 9–25.
Received: September 19, 2007
Published online: February 12, 2008
Chem. Eur. J. 2008, 14, 3100 – 3109
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3109