Acid-base properties of the nucleic-acid model 2′- deoxyguanylyl(5′→3′)-2′-deoxy-5′-guanylate, d(pGpG)3-, and of related guanine derivatives

Article abstract of DOI:10.1039/b517904a

The dinucleotide d(pGpG) is an often employed DNA model to study various kinds of interactions between DNA and metal ions, but its acid-base properties were not yet described in detail. In this study the six deprotonation reactions of H₄[d(pGpG)]⁺ are quantified. The acidity constants for the release of the first proton from the terminal P(O)(OH)₂ group (pK_a = 0.65) and for one of the (N7)H⁺ sites (pK _a = 2.4) are estimated. The acidity constants of the remaining four deprotonation reactions were measured by potentiometric pH titrations in aqueous solution (25°C; I = 0.1 M, NaNO₃): The pK_a values for the deprotonations of the second (N7)H⁺, the P(O) ₂(OH)^-, and the two (N1)H sites are 2.98, 6.56, 9.54 and 10.11, respectively. Based on these results we show how to estimate acidity constants for related systems that have not been studied, e.g. pGpG, which is involved in the initiation step of a rotavirus RNA polymerase. The relevance of our results for nucleic acids in general is briefly indicated. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b517904a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Acid–base properties of the nucleic-acid model 2ꢀ-deoxyguanylyl(5ꢀ→3ꢀ)-
2ꢀ-deoxy-5ꢀ-guanylate, d(pGpG)3−, and of related guanine derivatives†‡
Bernd Knobloch,a,b Helmut Sigel,b Andrzej Okruszekc,d and Roland K. O. Sigel*a
Received 16th December 2005, Accepted 10th January 2006
First published as an Advance Article on the web 26th January 2006
DOI: 10.1039/b517904a
The dinucleotide d(pGpG) is an often employed DNA model to study various kinds of interactions
between DNA and metal ions, but its acid–base properties were not yet described in detail. In this study
the six deprotonation reactions of H4[d(pGpG)]+ are quantified. The acidity constants for the release of
the first proton from the terminal P(O)(OH)2 group (pKa = 0.65) and for one of the (N7)H+ sites
(pKa = 2.4) are estimated. The acidity constants of the remaining four deprotonation reactions were
measured by potentiometric pH titrations in aqueous solution (25 ◦C; I = 0.1 M, NaNO3): The pKa
values for the deprotonations of the second (N7)H+, the P(O)2(OH)−, and the two (N1)H sites are 2.98,
6.56, 9.54 and 10.11, respectively. Based on these results we show how to estimate acidity constants for
related systems that have not been studied, e.g. pGpG, which is involved in the initiation step of a
rotavirus RNA polymerase. The relevance of our results for nucleic acids in general is briefly indicated.
synthesis leading to the conclusion that this dinucleotide is the
initiator of replication.10
1
Introduction
The anticancer drug Cisplatin, cis-(NH3)2PtCl2, is well known
to interact with DNA.1 Due to the preferential binding of cis-
(NH3)2Pt2+ to the N7 sites of two consecutive guanine residues
in DNA, the dinucleotide d(pGpG) has relatively often been
studied as model compound.1,2 There are also studies with
a novel antitumor-active dirhodium(II,II) complex3 and fur-
ther platinum(II) compounds4 as well as with closely related
oligonucleotides.5 In addition, the interaction of d(pGpG) with
Na+ and K+,6 the antibiotic actinomycin D,7 as well as the human
immunodeficiency virus type 1 integrase8 have received attention.
Whereas the d(pGpG) unit serves as the primary target in DNA
for Cisplatin and related compounds, its ribose relative pGpG does
occur in living cells. The cyclic dimer of GMP, c-di-GMP plays a
critical role in bacterial cell signaling and is thereby hydrolyzed to
pGpG and finally to GMP by a phosphodiesterase.9 Interestingly,
this hydrolysis reaction is strongly Mg2+ dependent. The formation
of pGpG has also been observed during the replication of the
viral RNA by the rotavirus RNA-dependent polymerase:10 During
replication, the dinucleotide pGpG (as well as ppGpG) is formed
and serves subsequently as a specific primer for the (−) strand
In the course of our attempts to reveal the interrelations between
metal ions and nucleic acids,11,12 we started to use dinucleotides
as models to quantify the metal ion-binding properties of single-
stranded RNA and DNA.13 Besides the phosphate diester bridge,13
the N7 site of guanine residues is especially important for metal
ion binding to nucleic acids.14–16 Thus, we selected d(pGpG)3− (see
Fig. 1) as a ligand to be studied because of its wide use as a DNA-
model compound1–3,6–8 as well as in various other investigations;17
furthermore, its relative pGpG occurs in nature as indicated above.
To our surprise we discovered that the acid–base properties in
aqueous solution of these dinucleotides have never been described
in detail.18
aInstitute of Inorganic Chemistry, University of Zu¨rich, Winterthurerstrasse
190, CH-8057, Zu¨rich, Switzerland. E-mail: roland.sigel@aci.unizh.ch
bDepartment of Chemistry, Inorganic Chemistry, University of Basel,
Spitalstrasse 51, CH-4056, Basel, Switzerland
cDepartment of Bioorganic Chemistry, Center for Molecular & Macromolec-
ular Studies, Polish Academy of Sciences, PL-90-363, Lo´dz, Poland
dInstitute of Technical Biochemistry, Faculty of Biotechnology & Food
Sciences, Technical University of Lo´dz, PL-90-924, Lo´dz, Poland
† Electronic supplementary information (ESI) available: Details of the
procedures how the various estimations were carried out. See DOI:
10.1039/b517904a
‡ This study is dedicated to Professor Bernhard Lippert, University of
Dortmund, Germany, on the occasion of his 60th birthday, with the very
best wishes of the authors for all his future endeavors and with deep
appreciation for his friendship, unselfish help and advice over the years to
H.S. and R.K.O.S.
Fig. 1 Chemical structure of the trianion of 2ꢀ-deoxyguanylyl(5ꢀ→3ꢀ)-
2ꢀ-deoxy-5ꢀ-guanylic acid, i.e., 2ꢀ-deoxyguanylyl(5ꢀ→3ꢀ)-2ꢀ-deoxy-5ꢀ-
guanylate, abbreviated as d(pGpG)3−, and also known6 as 2ꢀ-deoxy[5ꢀ-
phosphate-guanylyl-(3ꢀ-5ꢀ)-guanosine]. The two guanosine units are shown
in their dominating anti conformation.14 Note, species written in the text
without a charge, e.g., d(pGpG), 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.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 1085–1090 | 1085
©

View Article Online
H2[d(pGpG)]− ꢀ H[d(pGpG)]2− + H+
(3a)
(3b)
(4a)
(4b)
(5a)
(5b)
(6a)
(6b)
Here, we report the six equilibrium constants for the stepwise
deprotonation of nearly fully protonated d(pGpG)3−, i.e. of the
H4[d(pGpG)]+ species. Encompassing the pH range of about 10.5
to 1, our study also includes the acidity constants for the release
of the protons from the two (N1)H sites of the guanine residues.
The site attribution for the various deprotonation reactions could
be achieved in an unequivocal way by comparisons with (mostly)
other guanine derivatives. One of the surprising conclusions of
this study is that the two nucleobase residues in d(pGpG) react
rather independently and thus, do not “feel” much of each other.
Regarding nucleic acids this is a remarkable result and different
from the observations made with pUpU,13 where the mutual
nucleobase effects are also minor but where the deprotonation
of the (N3)H sites of the pyrimidine bases is somewhat shifted
towards a lower pH range. Furthermore, the here presented
new data, now allows by sophisticated comparisons to estimate
the corresponding acidity constants for H4(pGpG)+ and other
dinucleotides, which so far have also not been described.
KHH
= [H[d(pGpG)]2−][H+]/[H2[d(pGpG)]−]
H[d(pGpG)]2− ꢀ d(pGpG)3− + H+
[d(pGpG)]
2
KHH[d(pGpG)] = [d(pGpG)3−][H+]/[H[d(pGpG)]2−]
d(pGpG)3− ꢀ d(pGpG–H)4− + H+
KHd(pGpG) = [d(pGpG–H)4−][H+]/[d(pGpG)3−]
d(pGpG–H)4− ꢀ d(pGpG–2H)5− + H+
KHd(pGpG–H) = [d(pGpG–2H)5−][H+]/[d(pGpG–H)4−]
It should be noted that d(pGpG–H)4− in equilibrium (5a) is to be
read as “d(pGpG) minus H”, meaning that one of the two (N1)H
sites has lost a proton, without defining which one. Analogously,
in the species d(pGpG–2H)5− both (N1)H sites are deprotonated.
Acidity constants for equilibria (1a) and (2a) could only be
estimated (see Table 1 and section 2S in the ESI), whereas the
values for the other four equilibria were measured by potentio-
metric pH titrations. The results are listed in Table 1 together with
the acidity constants for several related species.24–30 Comparisons
of the given data allows clear-cut site attributions for the various
deprotonation reactions of H4[d(pGpG)]+ as is for example evident
when comparing the acid–base values for the equally charged
d(pGpG)3− (entry 5) and GDP3− (entry 11) species.
2
Results
The dinucleoside monophosphates GpG and d(GpG), i.e. without
a 5ꢀ-terminal phosphate group, are known for their tendency to
undergo aggregate formation via self-association by nucleobase
stacking and guanine–guanine hydrogen bonding.19,20 Such an
aggregation is much smaller with the here investigated d(pGpG)3−,
as a comparison of the self-association properties of d(GpG)− and
d(pGpG)3− has shown.6 Evidently, the addition of a phosphate
group to the 5ꢀ-OH of the 2ꢀ-deoxyribose residue significantly
inhibits this tendency. This agrees with observations made for
guanosine and GMP2− where the equilibrium constants defined
according to the isodesmic model for an indefinite noncooperative
self-association21 in aqueous solution are K = 8 M−1 (cf.21) and
1.3 M−1 (cf.22), respectively. Calculations for various guanine
derivatives reveal that with the ligand concentrations used in this
study, i.e., 0.15 mM (section 4.3), more than 99% of the species are
present in their monomeric form.23 Hence, the following results
refer in all instances to the monomeric species.
The dinucleotide d(pGpG)3− (Fig. 1) can accept three protons at
its phosphate groups; however, two of these protons are released at
a very low pH. For the 5ꢀ-terminal P(O)(OH)2 group, a pKa value
can be estimated (see below), though it needs to be emphasized that
the pH range of the deprotonation reaction of this proton certainly
overlaps with that of the proton released from the phosphate
diester bridge, for which in a first approximation a similar pKa
value is expected. Because one proton each can be accepted at
the N7 sites of the two guanine residues, the six deprotonation
reactions considered here begin with H4[d(pGpG)]+ and terminate
with d(pGpG–2H)5−, i.e. the species where the two (N1)H sites
have also lost their proton. This then leads to the following six
deprotonation reactions:
3
Discussion
The here-determined acid–base properties of H4[d(pGpG)]+ do
not stand alone. Together with the additional data on related
compounds, as summarized in Table 1, we can now analyze and
quantify in detail the mutual effects of the distinct subunits on each
other, e.g. of the nucleobase, the 2ꢀ-OH group or the phosphate
residue. Many comparisons and conclusions are possible, some of
which are discussed in the following paragraphs.
At first sight it is somewhat surprising that the 5ꢀ-P(O)2(OH)−
group of H(pUpU)2− (Table 1; entry 6; column 5) is slightly more
acidic than the same group of H[d(pGpG)]2− (entry 5; column 5),
i.e., DpKa = pKHH[d(pGpG)] − pKHH(pUpU) = (6.56
0.03) − (6.44
0.02) = 0.12
0.04. However, from a structural point of view,
two differences between pUpU3− and d(pGpG)3− are immediately
obvious: Not only the nucleobase moieties, but also the sugar
residues differ in the two dinucleotides.
Indeed, it is well known that the replacement of a ribose
residue by a 2ꢀ-deoxyribose unit in a nucleotide leads to a slight
basicity increase.28,31 This increase occurs irrespective of the type
of nucleobase attached to the sugar moiety as the following
comparisons of H(dGMP)− with H(GMP)− and d(CMP)− with
H(CMP)− show: DpKa/deoxy = pKHH(dGMP) − pKHH(GMP) = (6.29
0.01) − (6.25 0.02) = 0.04 0.02 (Table 1; entries 8,9; column
6) and pKHH(dCMP) − pKHH(CMP) = (6.24 0.01) − (6.19 0.02) =
0.05 0.02 (cf. ref. 31).
H4[d(pGpG)]+ ꢀ H3[d(pGpG)] + H+
= [H3[d(pGpG)]][H+]/[H4[d(pGpG)]+]
(1a)
(1b)
(2a)
(2b)
KHH
[d(pGpG)]
4
H3[d(pGpG)] ꢀ H2[d(pGpG)]− + H+
= [H2[d(pGpG)]−][H+]/[H3[d(pGpG)]]
Similarly, from a comparison of the acid–base properties of
H(UMP)− and H(GMP)−, it follows that the guanine residue
causes an increase in basicity at the 5ꢀ-terminal phosphate group:
KHH
[d(pGpG)]
3
1086 | Org. Biomol. Chem., 2006, 4, 1085–1090
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Table 1 Negative logarithms of the acidity constants for the deprotonation of the P(O)(OH)2, (N7)H+, and (N1)H sites in H4[d(pGpG)]+ [entry 5;
eqn (1–6)], together with some related dataa,b
pKa of the sites
Number
Acidsc
P(O)(OH)2
(N7)H+
P(O)2−(OH)
(N1)H
Ref.
1
2
H(Guo)+
2.11 0.04
2.30 0.04
1.49 0.03/2.51 0.03
1.69 0.10/2.71 0.10
2.4 0.2f/2.98 0.13
9.22 0.01
9.24 0.03
9.34 0.07/10.38 0.10
9.37 0.03/10.39 0.07
9.54 0.08/10.11 0.14
8.99 0.03/9.63 0.08g
9.45 0.02g
9.49 0.02
9.56 0.02
9.45 0.02
9.56 0.03
23
24
25,26
25
—
13
27
23
28
H(dGuo)+
3d
4e
5
H2(GpG)+
H2[d(GpG)]+
H4[d(pGpG)]+
H2(pUpU)−
H2(UMP)
0.65 0.3f
1.0 0.3
0.7 0.3
0.3 0.2
0.35 0.2h
6.56 0.03
6.44 0.02
6.15 0.01
6.25 0.02
6.29 0.01
6.14 0.01
6.38 0.01
6.52 0.01
6
7
8
9
H3(GMP)+
H3(dGMP)+
H2(3ꢀdGMP)
H3(GDP)
2.48 0.04
2.69 0.03
2.29 0.04
2.67 0.02
10i
11
12j
—
29
—
0.77 0.20
0.6 0.3
H4(pGpG)+
2.2 0.2/2.80 0.16
9.50 0.09/10.10 0.18
a All constants were determined by potentiometric pH titrations in aqueous solution (25 ◦C; I = 0.1 M, NaNO3) except for a few, which were estimated
(see below). The errors given are three times the standard error of the mean value or the sum of the probable systematic errors, whichever is larger.
The error limits of differences between constants as they appear in the text were calculated according to the error propagation after Gauss. b So-called
practical, mixed or Brønsted acidity constants30 are listed (section 4.2). c Definitions: dGMP2− (= 5ꢀdGMP2−), 2ꢀ-deoxyguanosine-5ꢀ-monophosphate;
d(GpG)−, 2ꢀ-deoxyguanylyl(3ꢀ→5ꢀ)-2ꢀ-deoxyguanosine; 3ꢀdGMP2−, 2ꢀ-deoxyguanosine-3ꢀ-monophosphate; dGuo, 2ꢀ-deoxyguanosine; GDP3−, guanosine-
5ꢀ-diphosphate; GMP2−, guanosine-5ꢀ-monophosphate; GpG−, guanylyl(3ꢀ→5ꢀ)guanosine, Guo, guanosine; pGpG3−, guanylyl(5ꢀ→3ꢀ)-5ꢀ-guanylate;
pUpU3−, uridylyl-(5ꢀ→3ꢀ)-5ꢀ-uridylate; UMP2−, uridine-5ꢀ-monophosphate. d The values for the (N7)H+ sites are from ref. 26. e The values for the
(N7)H+ sites are estimates; see ref. 25. f Estimated values; see text in sections 2 and 3. g These protons are released from the (N3)H sites of the uracil
residues. h This constant was estimated by correcting the value of H3(GMP)+ for the effect of the 2ꢀ-deoxyribose residue which amounts to 0.05 0.02
log units (see text in section 3). i B. Song, J. Zhao, H. Sigel, results to be published. j For details see section 3S in the ESI.
i.e., DpKa/NMP = pKHH(GMP) − pKHH(UMP) = (6.25 0.02) − (6.15
0.01) = 0.10 0.02 (Table 1; entries 7,8; column 5).
DpKa = 0.2 more basic. On the other hand, when considering the
deprotonation of the (N1)H site in the same pairs of compounds
(column 6), the effect of the 2ꢀ-OH group is very minor; on
By using the two increments described above for 2ꢀ-OH and
nucleobase substitution, one can estimate the acidity constant
average the 2ꢀ-deoxy compounds are only by 0.04
0.015 (1r)
for H[d(pGpG)]2−: pKHH[d(pGpG)]estimate = pKHH(pUpU) + DpKa/deoxy
+
log units more basic. Interestingly, the acidifying effect of the
2ꢀ-OH group on the first and second deprotonation step of the
5ꢀ-terminal phosphate group is with DpKa values of about 0.04
also very small but of the same order as observed at the (N1)H
site (Table 1; entries 5,12 and 8,9; columns 3 and 5). Overall, these
results indicate that the electron-withdrawing influence of the 2ꢀ-
OH is strongest experienced at N7 of the nearby imidazole moiety,
which is attached to C1ꢀ, and it has only very little influence on
the more distant (N1)H and 5ꢀ-terminal phosphate positions.
Another interesting observation is that the release of the two
(N1)H protons from d(pGpG)3− (entry 5; column 6) occurs with
DpKa/NMP = (6.44 0.02) + (0.04 0.02) + (0.10 0.02) = 6.58
0.03. This estimated acidity constant is in excellent agreement
with the measured one, pKHH[d(pGpG)] = 6.56 0.03 (Table 1; entry
5; column 5), thereby proving the internal consistency of the
constants listed in Table 1.
Application of comparisons as described above also allows cal-
culation of pKa values for other (2ꢀ-deoxy)dinucleotides based on
the acidity constants available for the corresponding nucleoside 5ꢀ-
monophosphates.18,23,27,28 For example, the differences (see Table 1)
pKHH[d(pGpG)] − pKHH(dGMP) = (6.56 0.03) − (6.29 0.01) = 0.27
0.03 and pKHH(pUpU) − pKHH(UMP) = (6.44 0.02) − (6.15 0.01) =
0.29 0.02, which are on average DpKa/DN,NMP = 0.28 0.03, may
be used to estimate acidity constants for H(pApA)2−, H(pCpC)2−,
a DpKa difference of 0.57
0.16, and this is within the error
limits identical with the statistically expected value of 0.6 for
a symmetrical diprotonic acid.13 Exactly the same observation
is made for pUpU3− where DpKa = 0.64
0.08 (entry 6).
H[d(pCpC)]2− and H[d(pTpT)]2−. The results are pKHH(pApA)
=
6.49 0.03, pKH
= 6.47 0.04, pKHH[d(pCpC)] = 6.52 0.03,
These results indicate that the two nucleobase residues present
in each of the two nucleotides react rather independently and
do not “feel” much of each other. This is different for the two
dinucleoside monophosphates GpG− and d(GpG)− where the
acidity differences are about 1.0 pK unit (Table 1; entries 3,4). This
is an indication that in the two latter cases some intramolecular
stacking occurs between the guanine residues whereas, once a
5ꢀ-phosphate group is present, this appears no longer to be the
case. It follows that d(pGpG) occurs in dilute solution in an open
form. This conclusion is in perfect agreement with the mentioned
observation (section 2; first paragraph) that d(GpG)3− stacks much
better than d(pGpG)3−.6
and pKH[d(pTpT)] =H6(p.C6p4C) 0.03 (25 ◦C; I = 0.1 M, NaNO3).§
H
As the nature of the nucleobase residue has an effect on the
acidity of the 5ꢀ-terminal phosphate group, one can expect that
changes at the phosphate–sugar moiety will also have an effect
on the acid–base properties of the nucleobase. Comparison of the
acid–base properties of the various guanine derivatives containing
either a ribose or a 2ꢀ-deoxy unit (Table 1; entries 1,2; 8,9; column
4), reveals that the N7 site of the deoxy compounds are by
§ The above values follow from pKHH(pApA) = pKHH(AMP) + DpKa/DN,NMP
=
(6.21 0.01) [cf.23] + (0.28 0.03) = 6.49 0.03, pKHH(pCpC) = pKHH(CMP)
+ DpKa/DN,NMP = (6.19
0.02) [cf.27] + (0.28
0.03) = 6.47
0.04,
With the above considerations in mind, and seeing in Table 1 for
H2(GpG)+ that DpKa for the two (N1)H and the two (N7)H+ sites
is identical, one may also estimate the pKa value for the release
pKHH[d(pCpC)] = pKHH(dCMP) + DpKa/DN,NMP = (6.24
0.01) [cf.31] + (0.28
0.03) = 6.52 0.03, and pKHH[d(pTpT)] = pKHH(dTMP) + DpKa/DN,NMP = (6.36
0.01) [cf.28] + (0.28 0.03) = 6.64 0.03.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 1085–1090 | 1087
©

View Article Online
of the first proton from the two (N7)H+ sites in H3[d(pGpG)] by
deducting the (N1)H difference (0.57 0.16; Table 1; entry 5;
column 6) from the pKa value of the second (N7)H+ site (2.98
0.13) to give pKHH
5 of Table 1 (column 4) (see also section 3S in the ESI). It
should be noted in this context that pKHH[d(pGpG)] = 2.98 0.13 for
the deprotonation of the second (N7)H+ site was experimentally
determined only at pH ≥ 3.6 (see section 4.3) which means that in
this pH range any contribution of the first (N7)H+ site is negligibly
small.
Application of the discussed systematic variations allows an es-
timation of the various acidity constants for the ribose-containing
dinucleotide H4(pGpG)+, which has widely been studied.9,32 Be-
cause to the best of our knowledge no such values exist in the
literature, the estimations are listed in entry 12 of Table 1 (for
details see section 3S in the ESI).
At this point is seems appropriate to ask which of the two
(N7)H+ or (N1)H sites, respectively, is deprotonated first, the 5ꢀ or
the 3ꢀ guanine residue? Interestingly, addition of the 3ꢀ-P(O)2(OH)−
group to dGuo has no effect on the basicity of N7; both pKa
values are within their error limits identical (Table 1; entries 2,10;
column 4), whereas addition of the same group to the 5ꢀ-position
phosphate group, and thus a negative charge, can be determined
by comparing the pKa values of 3ꢀ-dGMP2− (9.45 0.02) and 5ꢀ-
dGMP2− (9.56 0.02) with the first one of the d(pGpG)3− species
(9.54 0.08). As these values are rather similar, this indicates that
the difference in charge between the two mononucleotides and
the dinucleotide as well as the presence of two guanine residues
in the latter, has little influence on the release of the first proton
from the d(pGpG)3− species. This observation differs from the one
made13 with pUpU3− where the two neighboring uracil residues
show an increased acidity relative to that of an isolated uracil
group (see also entries 6 and 7 in Table 1, column 6). Consequently,
predictions about the mutual influence of neighboring nucleobases
in nucleic acids are difficult to achieve at this stage and clearly,
more data for comparisons are needed to reach this goal.
= 2.4 0.2; this value is listed in entry
[d(pGpG)]
3
4
Experimental
4.1 Materials
The synthesis of 2ꢀ-deoxyguanylyl(5ꢀ→3ꢀ)-2ꢀ-deoxy-5ꢀ-guanylate,
i.e., of the trisodium salt of d(pGpG)3−, was achieved by the in-
solution phosphoramidite methodology.34 Thus, 5ꢀ-O-dimethoxy-
trityl-N2-isobutyryl-2ꢀ-deoxyguanosine-3ꢀ-O-(2-cyanoethyl)-N,N-
diisopropylphosphoramidite was reacted with 3ꢀ-O-acetyl-N2-
isobutyryl-2ꢀ-deoxyguanosine in the presence of 1H-tetrazole
in CH3CN solution to give after I2–pyridine–H2O oxidation
of the intermediate phosphite, the fully protected dinucleoside
monophosphate. After selective removal of the dimethoxytrityl
group with 3% dichloroacetic acid in CH2Cl2, the dinucleoside
monophosphate was 5ꢀ-O-phosphorylated with bis-O,O-(2-
cyanoethyl)-N,N-diisopropylphosphoramidite reagent followed
by oxidation of the phosphite intermediate with I2–pyridine–
H2O.35 The obtained fully protected 5ꢀ-O-phosphorylated
dinucleotide was purified by silica gel chromatography and
deprotected by a 20 h treatment with 30% aqueous ammonia
at 55 ◦C. The crude product was purified by ion-exchange
chromatography on DEAE Sephadex A-25 (elution with a linear
gradient of triethylammonium bicarbonate from 0.05 to 0.5 M).
The purified dinucleotide was then transformed into its trisodium
salt by passing through Dowex 50 W × 8 (Na+ form) and
lyophylized to give a white solid in 34% yield (based on the
protected nucleoside). The compound gave with analytical
reversed phase HPLC a single peak. The structure of d(pGpG)
was confirmed by using spectroscopic methods: proton-decoupled
31P NMR (Bruker Avance, 200 MHz; D2O) showed two singlets
at d = −0.40 ppm (internucleotide phosphorus) and 4.35 ppm
(terminal phosphate); FAB MS (Finnigan MAT 95; negative
ions) gave m/z 675.0 (calculated MW 676.42 for free acid). The
observed 31P NMR chemical shifts are fully consistent with the
data reported for d(pGpG) salts;36 other synthetic routes37 have
also been applied.3
enhances the basicity of N7 remarkably, i.e., DpKa = pKHH
−
ꢀ
(5 dGMP)
2
pKHH(dGuo) = (2.69 0.03) − (2.30 0.04) = 0.39 0.05. Presumably
there are two reasons for this behavior: (i) the negative charge of
the P(O)2(OH)− group is closer to (N7)H+ in H2(5ꢀ-dGMP) than
in H2(3ꢀ-dGMP) . (ii) The proton at N7 is sterically in a position to
form a hydrogen bond (possibly also involving a water molecule)
with the 5ꢀ-P(O)2(OH)− group. Both effects will inhibit the release
of the proton from the (N7)H+ site in H2(5ꢀGMP) . It may be
added that the formation of similar hydrogen bonds between NH
sites and phosph(on)ate groups is known.33 Clearly, formation of
the indicated hydrogen bond is not possible in H2(3ꢀ-GMP) for
steric reasons.
However, when pKHH
= 2.69 is compared with the average
ꢀ
(5 dGMP)
2
pKa value [= 2.69 = (1/2)(2.4 + 2.98)] of the two (N7)H+ sites in
H3[d(pGpG)], one notes with surprise that the values are identical.
However, one should point out that the charge effect and hydrogen
bond formation can also operate in this case as described above.
Consequently, these comparisons indicate that the two N7 sites
in H[d(pGpG)]2− have the same basicity and this conclusion is
in accordance with the mentioned statistical difference of 0.6
pK units between the two pKa values of the (N7)H+ sites (Table 1;
entry 5; column 4). Furthermore, because the same difference of
0.6 pK units also applies to the pKa values for the two (N1)H sites
of d(pGpG)3− (column 6), one has to conclude that the acid–base
properties of the two guanine residues in d(pGpG)3− are identical
because they behave as is expected for symmetrical diprotonic
acids. This is an important conclusion regarding nucleic acids.
Finally, we want to discuss the effect of a phosphate group
on the deprotonation reaction of a (N1)H unit. Addition of a
PO3 group to the 3ꢀ or 5ꢀ site of dGuo to give 3ꢀ-dGMP2− or
All the other materials and reagents were the same as used
previously and the NaOH stock solutions were also prepared as
described.13 The aqueous stock solutions of d(pGpG) 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 con-
sumption between solutions with and without ligand (see below).
2−
5ꢀ-dGMP2− enhances the basicity of the (N1)− sites by DpKa =
0.21 0.04 or 0.32 0.04, respectively.¶ The effect of a further
¶ The above differences result from pKH3ꢀdGMP − pKdHGuo = (9.45 0.02) −
(9.24
0.03) = 0.21
0.04 (see Table 1; entries 2,10; column 6) or
pKH5ꢀdGMP − pKHdGuo = (9.56
0.02) − (9.24
0.03) = 0.32
0.04 (see
entries 2,9).
1088 | Org. Biomol. Chem., 2006, 4, 1085–1090
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
4.2 Potentiometric pH titrations
Inagaki, Y. Kidani and J. Reedijk, Inorg. Chem., 1992, 31, 4656–4661;
(e) M. J. Bloemink, J. M. J. Pe´rez, R. J. Heetebrij and J. Reedijk, J. Biol.
Inorg. Chem., 1999, 4, 554–567.
5 (a) N. Farrell, Met. Ions Biol. Syst., 1996, 32, 603–639; (b) M. J.
Bloemink and J. Reedijk, Met. Ions Biol. Syst., 1996, 32, 641–685;
(c) G. Natile and M. Coluccia, Met. Ions Biol. Syst., 2004, 42, 209–250;
(d) N. Farrell, Met. Ions Biol. Syst., 2004, 42, 251–296.
6 M. M. Kawasaki and J. A. Walmsley, J. Biomol. Struct. Dyn., 1999, 17,
561–566.
7 (a) T. R. Krugh, E. S. Mooberry and Y.-C. C. Chiao, Biochemistry,
1977, 16, 740–747; (b) Y.-C. C. Chiao and T. R. Krugh, Biochemistry,
1977, 16, 747–755.
The pH titrations were carried out with a Metrohm E536
potentiograph connected to a Metrohm E665 dosimat and a
Metrohm 6.0253.100 Aquatrode-plus combined double-junction
macro glass electrode. This equipment was calibrated with
buffer solutions (pH 4, 7, 9; all based on the NBS scale,
now U.S. National Institute of Science and Technology, NIST)
obtained from Metrohm AG, Herisau, Switzerland. The acidity
constants determined at I = 0.1 M (NaNO3) and 25 ◦C are
so-called practical, mixed or Brønsted constants13,30 which may
be converted into the corresponding concentration constants by
subtracting 0.02 from the measured and listed pKa values.30 The
ionic product of water (KW) and the conversion term mentioned
do not enter into the calculations because the differences in NaOH
consumption between solutions with and without ligand (see
below) are evaluated.
8 A. Mazumder, H. Uchida, N. Neamati, S. Sunder, M. Jaworska-
Maslanka, E. Wickstrom, F. Zeng, R. A. Jones, R. F. Mandes, H. K.
Chenault and Y. Pommier, Mol. Pharmacol., 1997, 51, 567–575.
9 M. Christen, B. Christen, M. Folcher, A. Schauerte and U. Jenal, J. Biol.
Chem., 2005, 280, 30829–30837.
10 D. Chen and J. T. Patton, RNA, 2000, 6, 1455–1467.
11 R. K. O. Sigel and A. M. Pyle, Met. Ions Biol. Syst., 2003, 40, 477–512.
12 R. K. O. Sigel, Eur. J. Inorg. Chem., 2005, 2281–2292.
13 B. Knobloch, D. Suliga, A. Okruszek and R. K. O. Sigel, Chem. Eur. J.,
2005, 11, 4163–4170.
14 K. Aoki, Met. Ions Biol. Syst., 1996, 32, 91–134.
4.3 Determination of the acidity constants
15 (a) D. J. Klein, P. B. Moore and T. A. Steitz, RNA, 2004, 10, 1366–1379;
(b) A. M. Pyle, J. Biol. Inorg. Chem., 2002, 7, 679–690; (c) P. Auffinger
and E. Westhof, J. Mol. Biol., 2000, 300, 1113–1131; (d) H. Robinson,
Y.-G. Gao, R. Sanishvili, A. Joachimiak and A. H.-J. Wang, Nucleic
Acids Res., 2000, 28, 1760–1766; (e) M. Soler-Lopez, L. Malinina and
J. A. Subirana, J. Biol. Chem., 2000, 275, 23034–23044; (f) J. B. Murray,
D. P. Terwey, L. Maloney, A. Karpeisky, N. Usman, L. Beigelman and
W. G. Scott, Cell, 1998, 92, 665–673; (g) C. C. Correll, B. Freeborn,
P. B. Moore and T. A. Steitz, Cell, 1997, 91, 705–712; (h) J. H. Cate,
R. L. Hanna and J. A. Doudna, Nat. Struct. Biol., 1997, 4, 553–558;
(i) J. H. Cate and J. A. Doudna, Structure, 1996, 4, 1221–1229; (j) H. W.
Pley, K. M. Flaherty and D. B. McKay, Nature, 1994, 372, 68–74.
16 (a) R. K. O. Sigel and B. Lippert, Chem. Commun., 1999, 2167–2168;
(b) R. K. O. Sigel, E. Freisinger and B. Lippert, J. Biol. Inorg. Chem.,
2000, 5, 287–299; (c) H. Sigel and R. Griesser, Chem. Soc. Rev., 2005,
34, 875–900.
The acidity constants KHH [d(pGpG)] , KHH[d(pGpG)], KHd(pGpG), and KdH(pGpG
–H)
2
of H2[d(pGpG)]− [eqn (3–6)] were determined by titrating aqueous
solutions (30 mL) of HNO3 (0.5 mM) (25 ◦C; I = 0.1 M, NaNO3)
under N2 with NaOH (up to 3.5 mL, 0.02 M) in the presence and
absence of d(pGpG)3− (0.15 mM). 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 two mentioned
titrations, i.e., with and without ligand. The acidity constants
of H2[d(pGpG)]− were calculated in the pH range 3.6 to 10.2,
corresponding to 81% neutralization (initial) for the equilibrium
H2[d(pGpG)]−/H[d(pGpG)]2− and about 50% (final) for d(pGpG–
H)4−/d(pGpG–2H)5−. These neutralization degrees explain why
the error limits of the first and the last acidity constant are
17 (a) A. Kanavarioti, J. Org. Chem., 1998, 63, 6830–6838; (b) D. V.
Bugreev, E. L. Vasyutina, V. N. Buneva, Y. Yasui, M. Nishizawa, T.
Andoh and G. A. Nevinsky, FEBS Lett., 1997, 407, 18–20; (c) T. Ito
and M. Saito, Radiat. Phys. Chem., 1991, 37, 681–690.
relatively large. The final results for KHH [d(pGpG)], KHH[d(pGpG)], KHd(pGpG)
18 (a) IUPAC Stability Constants Database, Release 5, Version 5.16,
(compiled by L. D. Pettit and H. K. J. Powell), Academic Software,
Timble, Otley, West Yorkshire, UK, 2001; (b) NIST Critically Selected
Stability Constants of Metal Complexes, Reference Database 46, Version
6.0, (data collected and selected by R. M. Smith and A. E. Martell),
U. S. Department of Commerce, National Institute of Standards and
Technology, Gaithersburg, MD, 2001; (c) Joint Expert Speciation Sys-
tem (JESS), Version 6.4, (joint venture by K. Murray and P. M. May),
Division of Water Technology, CSIR, Pretoria, South Africa, and
School of Mathematical and Physical Sciences, Murdoch University,
Murdoch, Western Australia, 2001.
19 (a) R. Savoie, H. Klump and W. L. Peticolas, Biopolymers, 1978, 17,
1335–1345; (b) R. Ghana, C. Walss and J. A. Walmsley, J. Biomol.
Struct. Dyn., 1996, 14, 101–110.
20 J. A. Walmsley, M. L. Schneider, P. J. Farmer, J. R. Cave, C. R. Toth
and R. M. Wilson, J. Biomol. Struct. Dyn., 1992, 10, 619–638.
21 (a) K. H. Scheller, F. Hofstetter, P. R. Mitchell, B. Prijs and H. Sigel,
J. Am. Chem. Soc., 1981, 103, 247–260; (b) N. A. Corfu` and H. Sigel,
Eur. J. Biochem., 1991, 199, 659–669.
2
and KHd(pGpG H) are the averages of the values from four independent
–
pairs of titrations.
Acknowledgements
Financial support from the Swiss National Science Foundation
(SNF-Fo¨rderungsprofessur to R.K.O.S., PP002-68733/1), the
Universities of Zu¨rich (R.K.O.S.) and Basel (H.S.), and within
the COST D20 programme from the Swiss State Secretariat for
Education and Research (H.S.) is gratefully acknowledged, as
is the competent technical assistance of Mrs Astrid Sigel in the
preparation of the manuscript.
References
22 K. J. Neurohr and H. H. Mantsch, Can. J. Chem., 1979, 57, 1986–1994.
23 H. Sigel, S. S. Massoud and N. A. Corfu`, J. Am. Chem. Soc., 1994, 116,
2958–2971.
1 Cisplatin: Chemistry and Biochemistry of a Leading Anticancer Drug,
ed. B. Lippert, VHCA, Zu¨rich, 1999, Wiley-VCH, Weinheim, 1999, pp.
1–563.
2 J. P. Whitehead and S. J. Lippard, Met. Ions Biol. Syst., 1996, 32, 687–
726.
3 H. T. Chifotides, K. M. Koshlap, L. M. Perez and K. R. Dunbar, J. Am.
Chem. Soc., 2003, 125, 10714–10724.
4 (a) S. J. Berners-Price, J. D. Ranford and P. J. Sadler, Inorg. Chem., 1994,
33, 5842–5846; (b) E. L. M. Lempers, M. J. Bloemink, J. Brouwer, Y.
Kidani and J. Reedijk, J. Inorg. Biochem., 1990, 40, 23–35; (c) C. J.
van Garderen, M. J. Bloemink, E. Richardson and J. Reedijk, J. Inorg.
Biochem., 1991, 42, 199–205; (d) M. J. Bloemink, R. J. Heetebrij, K.
24 H. Sigel, B. Song, G. Oswald and B. Lippert, Chem. Eur. J., 1998, 4,
1053–1060.
25 C. P. Da Costa and H. Sigel, Inorg. Chem., 2003, 42, 3475–3482.
26 N. Ogasawara and Y. Inoue, J. Am. Chem. Soc., 1976, 98, 7048–7053.
27 S. S. Massoud and H. Sigel, Inorg. Chem., 1988, 27, 1447–1453.
28 B. Song and H. Sigel, Inorg. Chem., 1998, 37, 2066–2069.
29 E. M. Bianchi, Comparison of the Stabilities and Solution Structures
of Metal Ion Complexes Formed with 5ꢀ-Di- and 5ꢀ-Triphosphates of
Purine Nucleosides, PhD Thesis, University of Basel, Logos Verlag,
Berlin, 2003, pp. 1–216.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 1085–1090 | 1089
©

View Article Online
30 H. Sigel, A. D. Zuberbu¨hler and O. Yamauchi, Anal. Chim. Acta, 1991,
255, 63–72.
31 B. Song, G. Feldmann, M. Bastian, B. Lippert and H. Sigel, Inorg.
Chim. Acta, 1995, 235, 99–109.
32 (a) A. J. Schmidt, D. A. Ryjenkov and M. Gomelsky, J. Bacteriol.,
2005, 187, 4774–4781; (b) A. Kanavarioti, E. E. Baird, T. B. Hurley,
J. A. Carruthers and S. Gangopadhyay, J. Org. Chem., 1999, 64, 8323–
8333; (c) J. P. Ferris and G. Ertem, Origin Life Evol. Biosphere, 1992,
22, 369–381.
33 (a) M. J. Sa´nchez-Moreno, R. B. Go´mez-Coca, A. Ferna´ndez-Botello,
J. Ochocki, A. Kotynski, R. Griesser and H. Sigel, Org. Biomol.
Chem., 2003, 1, 1819–1826; (b) C. F. Moreno-Luque, E. Freisinger,
B. Costisella, R. Griesser, J. Ochocki, B. Lippert and H. Sigel, J. Chem.
Soc., Perkin Trans. 2, 2001, 2005–2011.
34 (a) M. H. Caruthers, Science, 1985, 230, 281–285; (b) S. L. Beaucage
and R. P. Iyer, Tetrahedron, 1992, 48, 2223–2311.
35 E. Uhlmann and J. Engels, Tetrahedron Lett., 1986, 27, 1023–1026.
36 C. G. Reinhardt and T. R. Krugh, Biochemistry, 1977, 16, 2890–2895.
37 (a) G. A. van der Marel, C. A. A. van Boeckel, G. Wille and J. H.
van Boom, Nucleic Acids Res., 1982, 10, 2337–2351; (b) G. A. van der
Marel, C. A. A. van Boeckel, G. Wille and J. H. van Boom, Tetrahedron
Lett., 1981, 22, 3887–3890.
1090 | Org. Biomol. Chem., 2006, 4, 1085–1090
This journal is
The Royal Society of Chemistry 2006
©