Characterization of complexes of nucleoside-5′-phosphorothioate analogues with zinc ions

Article abstract of DOI:10.1021/ic400878k

On the basis of the high affinity of Zn²⁺ to sulfur and imidazole, we targeted nucleotides such as GDP-β-S, ADP-β-S, and AP₃(β-S)A, as potential biocompatible Zn²⁺-chelators. The thiophosphate moiety enhanced the stability of the Zn²⁺- nucleotide complex by about 0.7 log units. ATP-α,β-CH ₂-γ-S formed the most stable Zn²⁺-complex studied here, log K 6.50, being ～0.8 and ～1.1 log units more stable than ATP-γ-S-Zn²⁺ and ATP-Zn²⁺ complexes, and was the major species, 84%, under physiological pH. Guanine nucleotides Zn²⁺ complexes were more stable by 0.3-0.4 log units than the corresponding adenine nucleotide complexes. Likewise, AP₃(β-S)A-zinc complex was ～0.5 log units more stable than AP₃A complex. ¹H- and ³¹P NMR monitored Zn²⁺ titration showed that Zn ²⁺ coordinates with the purine nucleotide N7-nitrogen atom, the terminal phosphate, and the adjacent phosphate. In conclusion, replacement of a terminal phosphate by a thiophosphate group resulted in decrease of the acidity of the phosphate moiety by approximately one log unit, and increase of stability of Zn²⁺-complexes of the latter analogues by up to 0.7 log units. A terminal phosphorothioate contributed more to the stability of nucleotide-Zn²⁺ complexes than a bridging phosphorothioate.

Full text of DOI:10.1021/ic400878k

Article
pubs.acs.org/IC
Characterization of Complexes of Nucleoside-5′-Phosphorothioate
Analogues with Zinc Ions
Alon Haim Sayer, Yehudit Itzhakov, Noa Stern, Yael Nadel, and Bilha Fischer*
Department of Chemistry, Bar Ilan University, Ramat-Gan 52900, Israel
S
* Supporting Information
ABSTRACT: On the basis of the high affinity of Zn2+ to
sulfur and imidazole, we targeted nucleotides such as GDP-β-
S, ADP-β-S, and AP3(β-S)A, as potential biocompatible Zn2+-
chelators. The thiophosphate moiety enhanced the stability of
the Zn2+-nucleotide complex by about 0.7 log units. ATP-α,β-
CH2-γ-S formed the most stable Zn2+-complex studied here,
log K 6.50, being ∼0.8 and ∼1.1 log units more stable than
ATP-γ-S-Zn2+ and ATP-Zn2+ complexes, and was the major
species, 84%, under physiological pH. Guanine nucleotides
Zn2+ complexes were more stable by 0.3−0.4 log units than the
corresponding adenine nucleotide complexes. Likewise, AP3(β-
S)A-zinc complex was ∼0.5 log units more stable than AP3A
1
complex. H- and 31P NMR monitored Zn2+ titration showed that Zn2+ coordinates with the purine nucleotide N7-nitrogen
atom, the terminal phosphate, and the adjacent phosphate. In conclusion, replacement of a terminal phosphate by a
thiophosphate group resulted in decrease of the acidity of the phosphate moiety by approximately one log unit, and increase of
stability of Zn2+-complexes of the latter analogues by up to 0.7 log units. A terminal phosphorothioate contributed more to the
stability of nucleotide-Zn2+ complexes than a bridging phosphorothioate.
to ATP-α-S, ATP-β-S, and ADP-α-S is higher than that of Mg2+
1. INTRODUCTION
by ∼0.5−1.4 log units (log K 4.47, 4.04, and 3.66 vs 4.92, 5.44,
Nucleoside 5′-diphosphates (NDPs) and nucleoside 5′-
and 4.95, respectively).14 Later, stability constants of adenosine
triphosphates (NTPs) form stable complexes with Zn2+ with
5′-monothiophosphate (AMP-α-S) with Zn2+ ion and Mg2+ ion
were determined by potentiometric titrations (log K 2.52 vs
1.28, respectively).15
log K values between 4 and 4.51,2 (Zn2+-NDPs complexes), and
5−5.53 (Zn2+-NTPs complexes). Zn2+-nucleotide complexes
are ∼1−1.5 log units more stable than their corresponding
Mg2+-nucleotide and Ca2+-nucleotide complexes, but are ∼1−2
log units less stable than their corresponding Cu2+-nucleotide
complexes.1−4 Crystal structure data on Zn2+-nucleotide
complexes is limited.5 A crystal structure of dimeric [Zn-
(H2ATP)(bpy)]2 complex exhibits distorted octahedral geom-
etry,6 where Zn2+ coordinates with ATP through Pγ, Pβ, and
Pα oxygen atoms. Yet, Zn−O (Pα) bonding is relatively weak
as indicated by the long bond distances.6
The limited data on nucleoside 5′-phosphorothioate
analogues as Zn2+ chelators prompted us to explore the
properties of nucleoside-phosphorothioate-Zn2+ complexes.
Specifically, we report on their acidity constants and stability
constants as determined by potentiometric titrations. These
constants were compared to those of the natural nucleotides,
and species distribution at physiological pH was predicted. In
1
addition, Zn2+ coordination sites were identified by H NMR-
and 31P NMR- monitored Zn2+ titrations.
Zn2+ is classified as “intermediate” Lewis-acid, according to
Hard-Soft Acid−Base theory,7 and coordinates preferentially
atoms such as nitrogen and sulfur.8 Moreover, Zn2+ has affinity
for thiolate second only to Cu2+ throughout biological
systems.9,10 For instance, Zn2+-dithiothreitol complex exhibits
log K of 11.06,11 where a 7-membered ring is formed because
of Zn2+ coordination with two thiol groups. The “thiophilic”
character of Zn2+ is also exemplified in Zn2+-2-thiocytidine
complex with log K of 2.53, being ∼2.3 log units more stable
than the parent complex, Zn2+-cytidine.12
2. RESULTS AND DISCUSSION
We targeted the evaluation of purine-nucleotide analogues 1−6
(Figure 1) as potentially promising biocompatible Zn2+-
chelators. We chose purine nucleotides 1−6, because of their
imidazole moiety, a borderline base, and their “soft” sulfur
atom, both of which preferably coordinate Zn2+.16,17
2.1. NMR Monitored Zn2+ Titrations. To determine Zn2+-
coordination sites in nucleoside 5′-phosphorothioate analogues,
Yet, the Zn2+ vs Mg2+-selectivity of phosphorothioate
analogues was barely reported. Stability constants of several
nucleoside 5′-phosphorthioate analogues with Mg2+ and Cd2+
were determined by 31P NMR,13,14 indicating that Cd2+ affinity
Received: April 9, 2013
Published: September 19, 2013
© 2013 American Chemical Society
10886
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
Figure 1. Nucleoside 5′-phosphorothioate analogues studied here.
Figure 2. Zn2+-titration of 6 mM GDP(β-S) in D2O at pD 7.34 at 300 K. 31P NMR spectrum was measured at 243 MHz.
we performed Zn2+-titration of analogues 1−3 and 6,
tides were identified by the difference, Δδ = δfree ligand −
1
monitored by H and 31P NMR spectroscopy.
δZn‑ligand complex between the chemical shifts of nucleotide Na+
salt and the corresponding Zn2+ complex.
The shift of NMR signals as well as their line-broadening
indicates which atoms in the ligands are involved in Zn2+-
chelation. Solutions of analogues 1−3 and the corresponding
parent nucleotides were titrated by 2 M Zn(NO3)2 solution
(0.1−10 equiv of Zn2+) at pD 7.4 and 300 K, and monitored by
1H and 31P NMR at 500 or 600, and 202.5 or 243 MHz,
respectively. Relatively low nucleotide concentrations (3 or 6
mM) were used to avoid intermolecular base stacking (π-
stacking interactions) in aqueous solutions.18 Chemical shifts
(δH, δp) of these compounds were measured, at different Zn2+
concentrations. Zn2+ coordination sites in nucleoside phos-
phorthioate analogues and the corresponding natural nucleo-
The data are reported in Supporting Information, Table S1
(31P NMR) and Supporting Information, Table S2 (1H NMR).
Significant Δδ values of terminal thiophosphate in ATP-γ-S,
ADP-β-S and GDP-β-S (Figure 2) were observed only after
addition of 0.1 equiv or 0.2 equiv of Zn2+. Addition of 0.1 equiv
of Zn2+ caused upfield shifts of Pβ of GDP-β-S (9.33 ppm) and
of Pγ of ATP-γ-S (6.95 ppm). In contrast the Δδ of Pβ of ADP-
β-S could not be measured after addition of 0.1 equiv of Zn2+
because of line-broadening of Pβ, resulting from dynamic
equilibrium between the free ADP-β-S and the Zn2+-ADP-β-S
complex. However for GDP-β-S this equilibrium is less
10887
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
dynamic possibly because of outersphere coordination of Zn2+
with O6 of the guanine oxygen.19 After addition of 0.2 equiv of
Zn2+, the signal of Pβ sharpened and shifted upfield to 8.92
ppm. Pα of GDP-β-S and Pβ of ATP-γ-S also showed a large
change upon addition of 0.1 equiv of Zn2+ resulting in
downfield shifts of 1.37 ppm and 2.13 ppm, respectively. This
effect was observed also for Pα signal of ADP-β-S, yet was
smaller (shift of 0.75 ppm), and further shifted by 1.36 ppm
upon addition of 0.2 equiv of Zn2+.
ing that Mg2+ binds only with Pβ of nucleoside triphosphates
and Pα of nucleoside diphosphates at pD ≅ 8. Another report
on NMR monitored Mg2+ titration at pD ≅ 5−7 indicated that
ATP binds Mg2+ through Pβ and Pγ, and ADP through Pα and
Pβ, due to the Δδ of their respective chemical shifts in the
presence and absence of Mg2+.20 In this pH range
deprotonation of the terminal phosphate (Pβ/Pγ) occurs, and
NMR data may result from a mixture of protonated and
nonprotonated species.
Here, the 31P NMR monitored Zn2+ titration of nucleoside
5′-thiophosphate analogues was performed at pD ≅7.4. Hence,
the terminal thiophosphate was not protonated (pKa ≅ 5.5),
especially in presence of Zn2+ ions which further lower the pKa
value of the terminal thiophosphate. Namely, the observed shift
of chemical shift of thiophosphate moiety was due to
coordination with zinc ions and not acid−base equilibrium of
the terminal thiophosphate.
Additional equivalents of Zn2+ (0.5, 1, 2, 10 equiv) did not
shift any further the Pα signal of ADP-β-S and GDP-β-S, and
Pβ signal of ATP-γ-S (Figure 2). Pα of ATP-γ-S almost did not
shift during the titration implying that Zn2+ did not coordinate
with Pα of ATP-γ-S, similar to ATP (Δδ ≅ 0.5 ppm upon
addition of 1 equiv of Zn2+). Addition of Zn2+ to ADP and
GDP resulted in gradual shifts, as the number of equivalents
was increased (Supporting Information, Table S1). Unlike
nucleoside 5′-phosphorthioate, addition of 0.1 equiv of Zn2+ to
ADP, GDP, and ATP affected immediately the Δδ value
(Supporting Information, Table S1).
Previously, ATP was found to coordinate metal ions mainly
through Pβ and Pγ.20−22 This observation supports our finding
that Pα of ATP/ATP-γ-S did not coordinate with Zn2+. In
addition, our findings are consistent with the crystal structure of
[Zn(H2ATP)(bpy)]2 complex showing that Zn−O (Pα)
bonding is relatively weak.6 Whereas Pα and Pβ of ADP
were found as the coordination sites of Mg2+ based on 25 Mg-21
and 17O NMR spectroscopy,22 here, we found that nucleoside
5′-phosphorthioate analogues chelate Zn2+ mainly through the
terminal thiophosphate group (Pγ of ATP-γ-S and Pβ of ADP-
β-S and GDP-β-S) based on their large Δδ values (∼7−9 ppm)
after addition of only 0.1−0.2 equiv of Zn2+.
Signals of terminal phosphate of natural nucleotides (ATP,
ADP, and GDP) were less affected even after addition of 10
equiv of Zn2+ (ΔδPβ (ADP) = +1.22, ΔδPβ (GDP) = +0.48,
ΔδPγ (ATP) = 2.23). Similar data were found for phosphate
groups adjacent to the terminal thiophosphate (Pα of ADP-β-S
and GDP-β-S) and for corresponding phosphate group of the
natural nucleotides after addition of 10 equiv of Zn2+
(ΔδPα (ADP) = +1.23 vs ΔδPα (ADP‑β‑S) = +1.16 and, ΔδPα (GDP)
= +1.48 vs ΔδPα (GDP‑β‑S) = +1.42). While after addition of 1
equiv of Zn2+, Δδ of Pβ of ATP was slightly larger than Δδ of
Pβ ATP-γ-S (+2.89 vs +1.61). This may indicate that the major
chelation of ATP-γ-S with Zn2+ occurs through Pγ, unlike ATP
where Zn2+-chelation occurs through Pγ and Pβ.
Dinucleotides, AP3A and AP3(β-S)A Zn2+-complexes also
exhibited a Δδ difference due to the sulfur modification. 31P
NMR spectra of AP3(β-S)A with and without Zn2+ ions are
shown in Supporting Information, Figure S1. AP3(β-S)A
spectrum showed two Pα signals (PαA and PαB) due to the
sulfur modification in Pβ and the chirality of the ribose,
removing symmetry through S−P−O plane in Pβ.
For both dinucleotides, AP3A and AP3(β-S)A, a gradual shift
of phosphate signals was observed due to Zn2+ addition, yet, a
more significant shift was observed for the modified
dinucleotide, 6. After addition of 10 equiv of Zn2+ to AP3(β-
S)A, its Pβ signal shifted downfield by ∼3 ppm (Supporting
Information, Figure S1), unlike Pβ of AP3A which shifted
downfield by only ∼1 ppm. The enhanced binding of Zn2+ to 6
is further supported by the similar Δδ of Pα in AP3A and
AP3(β-S)A upon addition of Zn2+ (Δδ +0.67 and 0.5 ppm
respectively).
1
Data of H NMR monitored Zn2+-titrations are presented in
Supporting Information, Table S2 and in Figure 3 and
While addition of Zn2+ to nucleoside phosphorothioate
analogues resulted in downfield shift of Pα or Pβ adjacent to
thiophosphate, δ of the terminal thiophosphate group shifted
upfield. These observations are supported by previous reports
on 31P NMR Zn2+-monitored titrations of adenosine
thiophosphate analogues at a much lower magnetic field
(24.3 MHz) and basic conditions.23 1:1 Zn2+:ADP-β-S and
Zn2+:ATP-γ-S terminal thiophosphate shifted upfield of the δ of
the terminal thiophosphate by 8.2 and 9 ppm, respectively.23
Titration with Ca2+ and Mg2+ ions shifted downfield the
chemical shift of the terminal thiophosphate of ADP-β-S and
ATP-γ-S by ∼1−2 ppm.23 This phenomenon is probably
because of coordination of Ca2+ and Mg2+ through the oxygen
atom of the terminal thiophosphate of adenosine phosphor-
othioate analogues, whereas Zn2+ is bound through the sulfur
atom.24
Figure 3. Zn2+-titration of 6 mM GDP(β-S) in D2O at pD 7.34 at 300
K. H NMR spectrum was measured at 600 MHz.
1
Supporting Information, Figure S2. Two sets of protons are
1
observed for AP3(β-S)A in H NMR spectrum (Supporting
Information, Figure S2), as for 31P NMR (Supporting
Information, Figure S1), because of asymmetry of the molecule
as mentioned above. The data in Supporting Information,
Table S2 show a similar trend to that in Supporting
Information, Table S1. H8/H2 signals of analogues 1−3
Related NMR-monitored Mg2+-titrations of nucleoside
diphosphates/triphosphates were reported before,25,26 indicat-
10888
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
shifted substantially after addition of only 0.1/0.2 equiv of Zn2+,
unlike for H8/H2 signals of the corresponding parent
nucleotides. For instance, H8 of ATP-γ-S and ATP shifted
downfield by 0.12 ppm and 0.01 ppm, respectively, upon
addition of 0.2 equiv of Zn2+ (Supporting Information, Table
S2). On the other hand, the H2 signals of ADP, ADP-β-S, and
ATP-γ-S moved upfield upon addition of Zn2+, whereas the
chemical shift of H2 of ATP did not change significantly. The
upfield shifts of H2 in the presence of zinc ions possibly result
from stacking interactions27 due to formation of dimeric
species, ([ML]2), as reported for [Cd(ADP)2], [Zn(ADP)2]27
and [M2+(ATP)]2 or [M2+(ATP)]2(OH−)28,29 (M2+ = Zn2+,
Cd2+, Ni2+, and Cu2+). It was also reported that self-stacking
decreases in the series, N > NMP2− > NDP3− > NTP4‑ because
of repulsion of the increasing number of phosphate negative
charges.29 This evidence may explain the higher upfield shifts of
H2/H8 in ADP-β-S (Δδ= −0.14/−0.20 ppm upon addition of
0.2 equiv of Zn2+) as compared to upfield shifts of H2 in ATP-
γ-S (Δδ = −0.03 ppm upon addition of 0.2 equiv of Zn2+).
The chemical shifts of H8 of dinucleotides AP3A and AP3(β-
S)A shifted in a similar way, whereas their H2 chemical shifts
did not change (Supporting Information, Table S2). The
downfield shifts of H8 imply that N7 is a coordination site of
Zn2+.
In addition to shift of signals, line-broadening was observed
as well as during zinc titration. This phenomenon was more
significant for nucleoside phosphorothioate analogues as
compared to the corresponding parent compounds (see
Supporting Information, Table S3). For example, the H8 signal
of GDP-β-S broadened 3-fold as compared to H8 of GDP,
upon addition of 0.1 equiv of Zn2+. Likewise, the H2 signal of
ATP-γ-S broadened 5-fold more than ATP. On the other hand,
signals of adenine base protons of dinucleotides, AP3A and
AP3(β-S)A (Supporting Information, Figure S2), did not
broaden significantly as compared to nucleotides and nucleo-
side phosphorothioate analogues. The major broadening effect
was observed for the terminal phosphate groups. This fact may
indicate that Zn2+ has a stronger affinity to terminal phosphate
or thiophosphate groups.
Purine-nucleotides can form macrochelates with metal ions
like Zn2+ or Cd2+ through the interaction of the phosphate-
coordinated metal ion with N7.30−32 In solution purine-
nucleotide complexes can form two species: an “open” form in
which only the phosphate chain coordinates the metal ion, and
a “closed” form in which the phosphate chain and N7 of the
purine nucleobase coordinate simultaneously the metal ion.3
Furthermore, there are at least two types of macrochelates of
M2+-(purine nucleotide) complexes: one in which the metal ion
binds directly to N7 of the purine residue (innersphere
coordination), and one in which M2+-coordination involves a
water molecule between N7 and M2+, an outersphere
interaction.3
Zn(ADP) exists mostly (80%) as the “open” form. The
“closed” form, (Zn(ADP)cl), is composed of 33% of the
outersphere and 67% of innersphere form.1 Zn(ATP) exists
mostly (70%) as the “open” form. The minor “closed” species
consists of 50:50 outersphere: innersphere coordinated
complexes. Hence, adenine nucleotide-Zn2+ complexes adopt
mainly the “open” form, implying that the major species of
Zn(ADP-β-S) or Zn(ATP-γ-S) complexes probably adopt the
“open” form as well. Since the basicity of N7 site in guanosine
is higher33−35 than that of adenosine, larger amounts of the
macrochelate are expected for guanine over adenine nucleo-
tides. Indeed, M2+(GMP)19,36 species were found to form
substantially larger amounts of macrochelates than
M2+(AMP)32,36 species. Zn(GTP) and Zn(GMP) species
include 68% and 72% “closed” conformation. Macrochelates
of Zn(GTP)cl and Zn(GMP)cl species are mainly inner-
sphere.19,30 Hence, we assume that Zn(GDP-β-S) complex
adopts a “closed” macrochelate form which is mainly
innersphere.
3. CHARACTERIZION OF NUCLEOSIDE
5′-PHOSPHOROTHIOATE-ZN2+ COMPLEXES BY
POTENTIOMETRIC PH-TITRATIONS
To characterize nucleoside 5′-phosphorothioate-Zn2+ com-
plexes we determined the pKa values of nucleotide analogues
1−6, their stability constants with Zn2+, log KZn
and log
ZnL
KZnZnLH, and the acidity constants of their complexes, pKaH
.
ZnLH
In addition, we predicted the species distribution of nucleoside
phosphorothioate complexes under physiological pH. Potentio-
metric pH-titrations were performed for compounds 1, 2, and
4−6 as Na+ salts, and compound 3 Li+ salt. The determination
of the stability constants of Zn2+ complexes were performed at
a 1:1 Zn2+:L ratio.4,37
3.1. Acidity Constants of Nucleoside 5′-Phosphor-
othioate Analogues 1−6. The protonation equilibria of
nucleoside-5′-phosphorothioate (NPnS) and nucleoside-5′-
phosphate (NPn) analogues are expected to occur in the pH
range 2.5−10. 8-SH-ATP, 5, and GDP-β-S, 2, undergo three
protonation equilibria. Thiophosphate analogues, 1, 3, 4, and 6,
and the corresponding nucleotides, undergo only two
protonation equilibria. The following protonation equilibria
apply for NPnS and the corresponding NPn analogues where
the charge of their protonated species is m = n − 1 (n = 2−3):
H3(NP S)m− ⇄ H2(NP S)(m+1)− + H+
(1a)
n
n
[H2(NP S)(m+1)−][H+]
n
KHH(NP S)
=
[H3(NP S)m−
]
3
n
(1b)
(2a)
n
H2(NP S)(m+1)− ⇄ H(NP S)(m+2)− + H+
n
n
[H(NP S)(m+2)−][H+]
n
KHH (NP S)
=
2
n
[H2(NP S)(m+1)−
]
(2b)
(3a)
n
H(NP S)(m+2)− ⇄ (NP S)(m+3)− + H+
n
n
[(NP S)(m+3)−][H+]
n
KHH(NP S)
=
n
[H(NP S)(m+2)−
]
(3b)
n
For guanosine 5′-phosphate or phosphorothioate, the first
deprotonation occurs at the guanine N7−H+ position (eqs 1a,
1b), the second deprotonation occurs at the terminal phosphate
group (eqs 2a, 2b), and the third deprotonation occurs at
guanine N1−H position (eqs 3a, 3b). For adenosine 5′-
phosphate or phosphorothioate, the first deprotonation occurs
at the adenine N1−H+ position (eqs 2a, 2b), and the second
deprotonation occurs at the terminal phosphate group (eqs 3a,
3b). For 8-SH-ATP the first two deprotonation steps occur at
N1−H+ and the terminal phosphate position (eqs 1a, 1b, 2a,
2b). The third deprotonation occurs at 8-SH position (eqs 3a,
3b).
The following protonation equilibria for the dinucleotide,
AP3(β-S)A, are expected to occur in the pH range 3−5:
10889
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
b
Table 1. Acidity Constants of Nucleoside 5′-Phosphate and Phosphorothioate Analogues, 1−6
compound no.
compound
ADP1
pKaH
pKaH
pKaH
LH
LH3
LH2
3.92 0.02(N1)
4.07 0.01(N1)
4.00 0.05(N1)
6.38 0.01(Pβ)
6.41 0.01(Pβ)
5.49 0.02(Pβ)
4.00 0.01(N1)
4.00 0.02(N1)
4.14 0.01(N1)
4.26 0.04(N1)
6.36 0.01(Pγ)
3.41 0.01(N1)
6.40 0.01(Pβ)
6.40 0.01(Pβ)
5.56 0.03(Pβ)
9.56 0.03(N1)
9.47 0.01(N1)
9.53 0.01(N1)
6.47 0.01(Pγ)
6.52 0.01(Pγ)
5.63 0.01(Pγ)
5.86 0.02(Pγ)
7.28 0.01(SH)
4.42 0.01(N1)
1
ADP
ADP-β-S
GDP38
2.67 0.02(N7)
3.23 0.03(N7)
3.14 0.03(N7)
GDP
2
GDP-β-S
ATP4
ATP
3
4
5
6
ATP-γ-S
ATP-α,β-(CH2)-γ-S
8-SH-ATP
AP3(β-S)A
a
3.27 0.02(N1)
a
b
Titrations were performed in triplicates or quadruplicates, except for compound 4, the titration of which was performed in duplicate. The
deprotonation positions appear in brackets.
Figure 4. Keto−enol tautomerism of 8-SH-ATP.
H2(AP (β − S)A)− ⇄ H(AP (β − S)A)2− + H+
lower than that of ATP (5.86 vs 6.52), but was more basic as
compared to ATP-γ-S (by 0.23 log units). The replacement of
the bridging oxygen atom between Pα and Pβ with an electron
donating methylene group, reduces the acidity of the Pγ-
thiophosphate. The pKa value of Pγ of 8-SH-ATP is similar to
that of Pγ of ATP (Table 1).
(4a)
3
3
[H(AP (β − S)A)2−][H+]
3
KHH (AP (β−S)A)
=
−
2
3
[H2(AP (β − S)A )]
(4b)
(5a)
3
H(AP (β − S)A)2− ⇄ (AP (β − S)A)3− + H+
3
3
pKa 7.28 is assigned to SH moiety of 8-SH-ATP (Table 1),
involved in keto−enol tautomerism (Figure 4). This assign-
ment is supported by 13C NMR experiments of Mercapto-
benzimidazole (MBI) and its sodium salt (MBI−Na+),42
suggesting that MBI and its Na+ salt exist as enol and enolate
forms respectively, and not as the keto forms. 8-SH substitution
affects also the pKa value of N1−H+, which is more acidic by
about 0.7 log units than that of N1−H+ in ATP. A similar trend
was reported for the pKa value of N1−H+ position in 8-OH-
adenosine (2.9).43
[(AP (β − S)A)3−][H+]
3
KHH(AP (β−S)A)
=
[H(AP (β − S)A)2−]
3
(5b)
3
For dinucleotide, AP3(β-S)A, there are two deprotonation
steps: the first deprotonation occurs at one of the N1−H+
positions in the adenine bases (eqs 4a, 4b), and the second
deprotonation occurs at the N1−H+ position of the second
nucleobase (eqs 5a, 5b).
Acidity constants of nucleoside 5′-phosphorothioate ana-
logues, determined as described above, were compared to those
of the corresponding nucleotide analogues. The pKa values of
ADP, GDP, and ATP, which we determined, were consistent
with literature (Table 1).1,4,38 pKa values of the protons in the
guanine and adenine nucleobases at N7−H+ and N1−H+
protons, respectively were not affected by the sulfur
modification at the phosphate chain (Table 1). Likewise, pKa
values of AP3(β-S)A were highly similar to the parent
dinucleotide, AP3A.39 However, pKa values of the thiophos-
phate group in ADP-β-S, GDP-β-S, and ATP-γ-S were more
acidic by about 1 log unit as compared to those of their parent
nucleotides (Table 1), because of the higher polarizibility of the
sulfur atom as compared to oxygen, inducing greater stability of
the thiophosphate anion. These results are consistent with the
literature.13,40 Increased acidity of at least 1 log unit was
reported for AMP-α-S and UMP-α-S vs A(U)MP.41 The acidity
constant of the thiophosphate group in ATP-α,β-CH2-γ-S was
3.2. Stability Constants of Zn2+-Nucleoside 5′-Phos-
phorothioate Complexes, 1−6. The equations describing
the formation of 1:1 Zn2+-complexes with NPnS (n = 2−3, m =
n − 1) are as follows.
H(NP S)(m+1)− + Zn2+ ⇄ Zn(H·NP S)(m−1)−
(6a)
n
n
[Zn(H·NP S)(m−1)−
]
Zn
n
K
=
Zn(H·NP S)
n
[Zn2+][(H·NP S)(m+1)−
]
(6b)
(7a)
n
(NP S)(m+2)− + Zn2+ ⇄ Zn(NP S)m−
n
n
[Zn(NP S)m−
]
Zn
n
K
=
Zn(NP S)
n
[Zn2+][(NP S)(m+2)−
]
(7b)
n
The deprotonation of the complex is represented by: eq 8a
Zn(H·NP S)m−1 ⇄ Zn(NP S)m− + H+
(8a)
n
n
10890
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
GDP-β-S, and ATP-γ-S complexes with Zn2+ were more stable
than the parent nucleotide complexes at least by 0.4−0.7 log
units (Table 2). Likewise, the stability constant of AP3(β-S)A-
Zn2+ complex was approximately 0.5 log units higher than that
of AP3A (3.8 vs 3.32). Furthermore, the stability constant of the
complex AP3(β-S)A with zinc was less stable than of the Zn2+-
complexes of nucleoside 5′-phosphorthioate analogues (3.8 for
AP3(β-S)A vs 5.42 and 5.74 for ADP-β-S and ATP-γ-S),
indicating the role of a terminal thiophosphate group in
stabilizing the complex.
[Zn(NP S)m−][H+]
n
KZHn(H·NP S)
=
Zn(H·NP S)m−1
]
n
(8b)
n
The acidity constant of the complex is calculated by the
following equation:4
pKHZn(H·NP S) = pKHH(NP S) + log KZn
Zn(H NPnS
·
)
n
n
− log KZn
Zn(NP S)
(9)
n
The equations describing the formation of 1:1 Zn2+-
complexes with dinucleotide analogues, AP3(β-S)A and AP3A
are the following.
Furthermore, Zn2+ forms a more stable complex with ATP-γ-
S than ADP-β-S (5.74 vs 5.42), as observed for ATP vs ADP
(5.37 vs 4.72), implying that a longer phosphate chain increases
the stability constant of Zn2+ complexes. These results are
supported by literature.1
H(AP (β − S)A)2− + Zn2+ ⇄ Zn(H·AP (β − S)A)
3
3
(10a)
GDP-β-S complex with Zn2+ is more stable than the
corresponding ADP-β-S complex (5.75 vs 5.42), as observed
also for GDP-Zn2+ complex vs ADP-Zn2+ complex (5.10 vs
4.72). This may be a result of outersphere interaction of Zn2+
with the guanine O6 via a hydrogen bond.19
[Zn(H·AP (β − S)A)]
3
Zn
K
=
Zn(H·AP (β−S)A)
[Zn2+][(H·AP (β − S)A)2−
]
3
(10b)
(11a)
3
(AP (β − S)A)3− + Zn2+ ⇄ Zn(AP (β − S)A)−
3
3
ATP-α,β-(CH2)-γ-S was found to form the most stable Zn2+-
complex studied here (log K 6.50) being approximately 0.8 and
1.1 log units more stable than ATP-γ-S-Zn2+ and ATP-Zn2+-
complexes, respectively. The high stability of this complex is
due to both its terminal thiophosphate moiety and the electron
donating methylene group. The latter makes the phosphate
chain more electron rich, and hence, a better chelator.
C8-SH substitution of ATP did not increase significantly the
stability constant of the complex with Zn2+ (log K 5.54). To
explain this phenomenon, the conformations of 8-SH-ATP and
its Zn2+ complex were determined by NOESY 2D-experiment
(Figure 5). syn Conformation of the free nucleotide was
indicated by cross peaks between H2 and H2′, and H2 and H3′
(Figure 5A). Addition of Zn2+ shifted the purine ring vs the
ribose ring in such a way, that H2 is closer to H1′ as indicated
by cross peaks between H2 and H1′ (Figure 5B). syn
Conformation implies that the major contribution to Zn2+
chelation occurs through the phosphate chain. Moreover, the
signal of H2 of 8-SH-ATP shifted downfield by 0.02 ppm upon
addition of 1 equiv of Zn2+ which may imply that N1 is another
coordination site of Zn2+ in 8-SH-ATP.
[Zn(AP (β − S)A)−]
3
Zn
K
=
Zn(AP (β−S)A)
[Zn2+][(AP (β − S)A)3−]
3
(11b)
3
The deprotonation of the complex is represented by eq 12a.
Zn(H·AP (β − S)A) ⇄ Zn(AP (β − S)A)− + H+
(12a)
3
3
[Zn(AP (β − S)A)−][H+]
3
KZHn(AP (β−S)A)
=
3
[Zn(H·AP (β − S)A)]
(12b)
3
The acidity constant of the complex is calculated by the
following equation.
pKH
Zn(H·AP (β−S)A)
3
= pKHH(AP (β−S)A) + log KZn
Zn(H AP (β−S)A)
·
3
3
− log KZn
Zn(AP (β−S)A)
(13)
3
All titrations were performed on 1:1 ligand:Zn2+ mixtures,
and the resulting stability constants of nucleoside 5′-
phosphorothioate-Zn2+ complexes were compared to those of
natural nucleotides (Table 2).1,4,44 We found that ADP-β-S,
ATP coordinates Zn2+ ions mainly by the phosphate chain
and in addition by N7-nitrogen atom. However, 8-SH-ATP
coordinates Zn2+ by the phosphate chain and possibly N1-
nitrogen atom. Yet, the contribution of N7−Zn2+ binding in
ATP is relatively small36 since the adenosine N1 site is more
basic than N7 by 5 orders of magnitude. This may explain the
minor increase of log K of 8-SH-ATP-Zn2+ as compared to
ATP-Zn2+ complex (5.54 vs 5.37). The shift of the purine ring
vs the ribose ring of 8-SH-ATP-Zn2+ complex (Figure 5B),
possibly reduced slightly the pKa of the complex as compared
to that of the free ligand (6.80 vs 7.28, C-8-SH).
Table 2. Stability Constants of Zn2+-NPn, Zn2+-NPnS, and
Zn2+-NPnN Complexes
compound
ADP1
log K
log K
pKH
Zn
ZnL
Zn
ZnLH
ZnLH
4.28 0.05
4.72 0.03
5.42 0.02
4.52 0.03
5.10 0.03
5.75 0.02
5.16 0.06
5.37 0.03
5.54 0.09
5.74 0.03
6.50 0.05
3.80 0.02
3.32 0.02
2.31 0.20
3.19 0.02
3.49 0.08
4.43 0.21
4.87 0.02
3.62 0.06
ADP
ADP-β-S
GDP44
GDP
3.45 0.01
3.54 0.04
2.86 0.11
3.14 0.04
5.06 0.07
4.05 0.01
4.31 0.05
3.03 0.03
4.76 0.03
3.28 0.02
4.17 0.13
4.29 0.01
6.80 0.02
3.94 0.01
3.67 0.01
3.66 0.02
The log KZn
values of the protonated zinc complexes
ZnLH
GDP-β-S
ATP4
were lower than log KZn
by 1.7−2.3 log units because of
ZnL
protonation on the phosphate chain, whereas log KZnZnLH of 8-
SH-ATP (5.06) and AP3(β-S)A (3.03) were less stable by only
∼0.5−0.8 log units because of protonation on the nucleobase.
This finding also supports the notion that the major
contribution to coordination is the phosphate chain. As
expected, the acidity constants of the Zn2+-complexes were
more acidic by at least 1.5 log units as compared to the free
ligand. However, the acidity constants of 8-SH-ATP and AP3(β-
ATP
8-SH-ATP
ATP-γ-S
ATP-α,β-(CH2)-γ-S
AP3(β-S)A
AP3A
a
n.o.
a
n.o. = not observed.
10891
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
1
1
Figure 5. Determination of the conformation of 5. (A) H,1H-NOESY spectrum of compound 5 Na+ salt. (B) H,1H-NOESY spectrum of
compound 5 in presence of 1 equiv of Zn2+.
Figure 6. Simulation of pH titration of 1:1 GDP and GDP-β-S complexes with Zn2+. (A) GDP species distribution with Zn2+. (B) GDP-β-S species
distribution with Zn2+.
S)A complexes with zinc were more acidic only by 0.48 and
0.76 log units, respectively, as compared to the free ligands.
To analyze the composition of the nucleoside 5′-phosphor-
othioate complexes with Zn2+ in titration solutions, as
compared to their parent nucleotide complexes, we performed
titration simulations using the Hyss program45 (Figures 6 and
7).
Under physiological pH (7.4) the major species is GDPZn−
(∼48%, Figure 6A) and [GDP-β-SZn(OH)]2− (∼76%, Figure
6B), whereas both GDP3− (Figure 6A) and GDP-β-S3− (Figure
6B) are approximately 9% and 3% of all species, respectively.
Likewise, the most stable complex studied here, ATP-α,β-
(CH2)-γ-SZn2−, is the dominant species under physiological
pH (∼84%), and only ∼2% of ATP-α,β-(CH2)-γ-S4− exist as a
free ligand (Figure 7C).
10892
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
Figure 7. Simulation of pH titration of 1:1 ATP, ATP-γ-S, and ATP-α,β-(CH2)-γ-S Zn2+complexes species distribution of (A) ATP, (B) ATP-γ-S,
and (C) ATP-α,β-CH2-γ-S.
The major species of ATP-γ-S and ATP in the titration
solution of pH 7.4 are also [ATP-γ-SZn]2− (∼49%) and
[ATPZn]2− (∼62%), yet they are less dominant than ATP-α,β-
(CH2)-γ-SZn2− (Figure 7). Nucleoside 5′-phosphorothioate-
Zn2+ complexes form hydroxo species much more than their
parent nucleotide. For instance, [GDP-β-S-Zn-(OH)]2− is
∼76% of all species under physiological pH, whereas GDP-Zn-
(OH)2− is ∼42%. Similarly, [ATP-γ-S-Zn-(OH)]3− vs [ATP-
Zn-(OH)]3− make ∼47% and ∼31%, respectively of all species.
The same trend was reported before for [AMP-α-S-Zn]
complex.46
resulting from an outersphere interaction of Zn2+ with the
guanine O6 via hydrogen bond. We concluded that a terminal
thiophosphate group is required to achieve a highly stable Zn2+-
complex. Indeed, Zn2+-AP3(β-S)A is a less stable complex than
Zn2+-complexes of nucleoside phosphorothioate analogues.
Furthermore, replacement of an α,β-bridging oxygen in
nucleoside-5′-phosphorothioate by a CH2 group further
increase affinity to Zn2+ ions. The pKa values of the
thiophosphate group in ATP-γ-S, ADP-β-S, and GDP-β-S
decreased by ∼1 log unit as compared to natural nucleotides.
Furthermore, pKa value of the thiophosphate group in ATP-
α,β-CH2-γ-S was more acidic by ∼0.7 log units as compared to
ATP, but was more basic as compared to ATP-γ-S because of
the positive inductive effect of the methylene group. C8-SH
substitution of ATP did not increase significantly the stability
constant of the nucleotide-Zn2+complex, because of syn
conformation of this nucleotide. Shift of the purine ring vs
the ribose ring of 8-SH-ATP in its Zn2+ complex enabled Zn2+
coordination by the N1 site, unlike ATP, where the Zn2+
coordination site in adenine ring is the N7 nitrogen atom. Zn2+
ion was found to coordinate with two phosphates, Pα and Pβ,
or Pβ and Pγ, in nucleoside-5′-di/tri phosphate/thiophosphate,
respectively. In AP3(β-S)A, Pβ is the major coordination site of
Zn2+, in addition to, Pα and N7. In summary, sulfur-modified
nucleotides are promising Zn2+-chelators. Specifically, we
propose ATP-α,β-CH2-γ-S as a promising zinc-ion-chelator
4. CONCLUSIONS
Replacement of a nonbridging oxygen atom in the phosphate
chain of purine nucleotides by sulfur increased the affinity to
Zn2+ as we expected according to HSAB theory. Thus, ADP-β-S
and GDP-β-S complexes with Zn2+ were ∼0.7 log unit more
stable than the parent nucleotide complexes. Likewise, the
stability constant of the AP3-(β-S)A-Zn2+ complex was higher
by ∼0.5 log units than that of AP3A. In addition to sulfur
substitution; an electron donating methylene group replacing a
bridging oxygen atom rendered ATP-α,β-CH2-γ-S the most
stable Zn2+complex studied here (log K = 6.50). Additionally,
Zn2+-GDP-β-S and Zn2+-GDP were found to be more stable
complexes than ADP-β-S and ADP complexes possibly
10893
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
forming a highly stable Zn2+-complex, log K 6.50, being ∼0.8
and ∼1.1 log units more stable than ATP-γ-S-Zn2+ and ATP-
Zn2+ complexes. Furthermore, [ATP-α,β-CH2-γ-S-Zn]2− is the
major species, 84%, under physiological pH.
(0.6 g, 4.4 mmol, 8 equiv) were added to form a colorless solution.
After 3 h, the reaction was quenched by addition of EDTA solution
(1.73 g, 4.4 mmol in 66 mL of deionized water). The neutral solution
was then freeze-dried. The resulting white solid was separated on a
Sephadex DEAE-A25 column applying a buffer gradient of water
(1000 mL) to 0.4 M NH4HCO3 (1000 mL). The relevant fraction was
freeze-dried at least 4 times to yield a yellowish solid. The LC
separation was repeated applying a buffer gradient of water (600 mL)
to 0.3 M NH4HCO3 (600 mL). The solution was freeze-dried at least
4 times to yield a yellowish solid. Product 2 was obtained in 55% yield
(152.7 mg) after LC separation as a yellowish solid. The final
purification was achieved by HPLC applying an isocratic elution of 0.1
M TEAA (pH 7)/CH3CN, 96:04, in 10 min (flow rate 5 mL/min).
Retention time of 2 was 6.9 min. The spectral data are consistent with
literature.48
5. EXPERIMENTAL SECTION
General Information. All reactions were carried out in flame-
dried, argon-flushed, two necked flasks sealed with rubber septa, and
the reagents were introduced with a syringe. The progress of reactions
was monitored by thin-layer chromatography (TLC) on precoated
Merck silica gel plates (60F-254) applying elution of iPrOH/H2O/
NH4OH (55%:35%:10%). Nucleotide starting materials were dried
overnight in a vacuum oven with P2O5. Mass spectra were measured
on a Q-TOF Micro Mass Spectrometer instrument (Micromass-
Waters, United Kingdom). Purification of the nucleotides was
achieved on a liquid chromatography (LC) (Isco UA-6) system with
a Sephadex DEAE-A25 column, which was swelled in 1 M NaHCO3 in
the cold for 1 day. Conditions for LC separation are described below.
Final purification of the nucleotides was achieved on an HPLC
(Hitachi Elite LaChrome) system, using a semipreparative reverse-
phase column (Gemini 5u C-18 110A, 250 × 10.00 mm, 5 μm,
Phenomenex, Torrance, U.S.A.). The purity of the nucleotides was
evaluated with an analytical reverse-phase column system (Gemini 5u
C-18 110A, 150 mm × 4.60 mm; 5 μm; Phenomenex, Torrance, CA)
using the following solvent systems: solvent system I, (A) 100 mM
triethylammonium acetate (TEAA), pH 7:(B) CH3CN; solvent system
II, 10 mM PBS buffer, pH 7.4:(B) CH3CN. The details of the solvent
system conditions used for the separation of each product are given
below. To obtain the nucleotides as sodium salts, they were passed,
following final purification, through a column of Sephadex-CM C25
(Na+-form). The purity of the nucleotides was generally ≥95%. ATP-
γ-S, 3, tetralitium salt (≥90% pure) was obtained from Enzo Life
Sciences (Lausen, Switzerland). ADP, GDP, ATP, and AP3A sodium
salts were purchased from Sigma (Steinheim, Germany). Nucleotide 5
was synthesized according to literature.47 AMP·H2O (free acid), GDP
(sodium salt), sodium thiophosphate tribasic hydrate were purchased
from Sigma-Aldrich Co. GMP (disodium salt) was purchased from
Acros Organics.
Adenosine 5′-[Beta-thio]diphosphate, 1. CDI (0.83 g, 5 mmol,
5 equiv) was added to AMP (Bu3NH+, Octyl3NH+) salt (1 mmol, 1
equiv) and anhydrous DMF (10 mL) were stirred in a two-necked
flask to form a colorless solution. The reaction was stirred at room
temperature overnight. Next, dry MeOH (0.22 mL, 5 mmol, 5 equiv)
was added to the reaction flask, and the solution was stirred for 8 min.
Subsequently, [PSO3H]2−(Bu3NH+)2 salt (2.91 g, 6 mmol, 6 equiv)
dissolved in anhydrous DMF (7 mL) and anhydrous ZnCl2 (1.05 g,
7.7 mmol, 7.7 equiv) were added to form a colorless solution. After 3
h, the reaction was quenched by addition of EDTA solution (3.16 g,
8.5 mmol in 127.5 mL deionized water). The neutral solution was then
freeze-dried. The resulting white solid was separated on a Sephadex
DEAE-A25 column applying a buffer gradient of water (1000 mL) to
0.4 M NH4HCO3 (1000 mL). The relevant fraction was freeze-dried at
least 4 times to yield a yellowish solid. The LC separation was repeated
applying a buffer gradient of water (800 mL) to 0.3 M NH4HCO3
(800 mL). The solution was freeze-dried at least 4 times to yield a
yellowish solid. Product 1 was obtained in 34% yield (169.3 mg) after
LC separation as a yellowish solid. The final purification was achieved
by HPLC applying an isocratic elution of 0.1 M TEAA (pH 7)/
CH3CN, 96:04, in 10 min (flow rate 5 mL/min). Retention time of 1
was 7.6 min. The spectral data are consistent with literature.48 ESI-MS
(negative) m/z: calcd for C10H14N5O9P2S−: 442.26 found: 442(M−).
Guanosine 5′-[Beta-thio]diphosphate, 2. CDI (0.461 g, 2.75
mmol, 5 equiv) was added to GMP (Bu3NH+, Octyl3NH+) salt (0.55
mmol, 1 equiv) and anhydrous DMF (5 mL) were stirred in a two-
necked flask to form a colorless solution. The reaction was stirred at
room temperature overnight. Next, dry MeOH (0.12 mL, 2.75 mmol,
5 equiv) was added to the reaction flask, and the solution was stirred
for 8 min. Subsequently, [PSO3H]2−(Bu3NH+)2 salt (3.3 mmol,
6equiv) dissolved in anhydrous DMF (6.5 mL) and anhydrous ZnCl2
P2-Thio P1, P3-Di(adenosine 5′-)triphosphate, 6. A mixture of
PSCl3 (26 μL, 0.24 mmol, 1.2 equiv) and pyridine (0.57 mL) was
stirred in a two-necked flask for 30 min at 0 °C, while AMP (Bu3NH+,
Octyl3NH+) salt (0.2 mmol, 1 equiv) was dissolved in dry pyridine
(0.68 mL) for 25 min at 0 °C. Subsequently, the pyridine solution of
AMP (Bu3NH+, Octyl3NH+) salt was added to the reaction flask and a
turbid solution was obtained. The AMP salt was washed with another
portion of dry pyridine (0.5 mL), and the solution was added to the
reaction flask. Stirring continued for 5 min, then 1 M TEAB (4 mL,
pH = 7.5) was added resulting in a clear solution. After 1 h the
reaction pH was 8−9. The resulting solution was freeze-dried, to give a
white solid residue which was separated on a Sephadex DEAE-A25
column applying a buffer gradient of water (200 mL) to 0.2 M
NH4HCO3 (200 mL) and then from 0.2 M NH4HCO3 (200 mL) to
0.4 M NH4HCO3 (200 mL). The solution was freeze-dried for at least
4 freeze-drying cycles to yield a white solid. Product 6 was obtained in
1
24% yield (19.6 mg) after LC separation as a white solid. H NMR
(700 MHz, D2O) δ: 8.33/8.32 (2s, H8A/H8B), 8.11/8.10 (2s, H2A/
H2B), 6.01/6.01 (2d, J = 4.9 Hz, H1′A/H1′B), 4.65/4.62 (2t, J = 4.9
Hz, H2′A/H2′B), 4.56/4.50 (2t, J = 4.6 Hz, H3′A/H3′B), 4.39/4.32
(2m, H4′A/H4′B), 4.35/4.27 (4m, H5″A, H5″B, H5′A, H5′B) ppm. 31
P
NMR (243 MHz, D2O) δ: 31.04 (t, J = 26.2 Hz, Pβ), −11.55/−11.61
(2d, J = 26.2 Hz PαA/PαB), ppm. 13C NMR (176 MHz, D2O) δ:
155.09 (C6A/C6B), 152.73 (C2A/C2B), 148.42/148.38 (C4A/C4B),
139.32/139.28 (C8A/C8B), 118.02/118.00 (C5A/C5B), 87.29/87.19
(C1′A/C1′B), 83.37/83.26 (d, JCP = 9.2 Hz, C4′A/C4′B), 74.99/74.94
(C2′A/C2′B), 69.97/69.85 (C3′A/C3′B), 64.87/64.72 (d, JCP = 5.1 Hz,
C5′A /C5′B ) ppm. HRMS (MALDI) m/z calcd for
−
C20H26N10O15P3S1 : 771.051, found: 771.051. Purity data obtained
on an analytical column-retention time: 5.83 min (97% purity) using
solvent system I (isocratic elution of 94:6 A:B over 12 min at a flow
rate of 1 mL/min). Retention time: 6.59 min (96% purity) using
solvent system II (isocratic elution of 98:2 A:B over 12 min at a flow
rate of 1 mL/min).
Typical Procedure for the Preparation of Nucleoside 5′-
Thiophosphate Derivatives. AMP (Bu3NH+, Octyl3NH+) salt was
prepared from the corresponding free acid, Bu3N (1 equiv) and
Octyl3N (1 equiv) in EtOH. GMP (Bu3NH+, Octyl3NH+) salt was
prepared from its corresponding sodium salt. The nucleotide sodium
salt was passed through a column of activated Dowex 50WX-8 200
mesh, H+ form. The column eluate was collected in an ice-cooled flask
containing tributylamine (1 equiv), trioctylamine (1 equiv), and
EtOH. The resulting solution was freeze-dried to yield the salt as white
solid. GMP (Bu3NH+, Octyl3NH+) salt was dried by repeated
coevaporation (three times) with EtOH, followed by coevaporation
with anhydrous DMF (three times).
P r e p a r a t i o n o f [ P S O 3 H ] 2 − ( B u 3 N H +
)
S a l t .
2
[PSO3H]2−(Bu3NH+)2 salt was prepared from the corresponding
sodium salt (tribasic hydrate). The thiophosphate sodium salt was
passed through a column of activated Dowex 50WX-8 200 mesh, H+
form. The column eluate was collected in an ice-cooled flask
containing tributylamine (1 equiv) and EtOH. The resulting solution
was freeze-dried to yield the salt as a colorless oil. The oil was dried by
repeated coevaporation with absolute EtOH (three times), followed by
10894
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
coevaporation with anhydrous DMF (three times) to form a colorless
oil. [PSO3H]2−(Bu3NH+)2 salt was stored in a desiccator at −20 °C.
NMR Experiments. Compounds were characterized by nuclear
magnetic resonance (1H NMR and 31P NMR) on Bruker AC-200 (200
Speciation and titration simulations were achieved with the HYSS2009
software.45
ASSOCIATED CONTENT
* Supporting Information
■
1
and 81 MHz for H and 31P, respectively), Bruker DPX-300 (300
S
1
1
MHz for H), Avance III-500 (500 and 202.5 MHz for H and 31P,
Further details are given in Tables S1−S3 and Figures S1−S2.
This material is available free of charge via the Internet at
http://pubs.acs.org.
respectively), DMX-600 (600 and 243 MHz for 1H and 31P,
respectively), or Avance III-700 (700 and 176 MHz for H, and 13C,
1
1
respectively) spectrometers. H NMR spectra were measured in D2O,
and the chemical shifts are reported in parts per million (ppm) relative
to DSS and D2O (4.8 ppm) as an internal standard. Nucleotides were
characterized also by 31P NMR in D2O, with 85% H3PO4 as an
external reference. The pD values of the NMR samples were adjusted
with diluted NaOD/DCl D2O solutions. Measurements of pD were
performed with a Knick (Berlin, Germany) pH meter equipped with a
Hamilton (Bonaduz, Switzerland) biotrode 238140 or tiptrode 238080
electrode.
AUTHOR INFORMATION
Corresponding Author
*E-mail: bilha.fischer@biu.ac.il. Fax: 972-3-6354907. Phone:
972-3-5318303.
■
Notes
The authors declare no competing financial interest.
Preparation of NMR Samples. The samples were prepared by
dissolving the nucleotide sodium salts in 99.9% D2O, and the pD was
adjusted to a physiological pH (pD = 7.4 0.2, pD = pH + 0.4). NMR
spectra were measured for 6 mM nucleotide or 3 mM dinucleotide
samples at 600 or 500 MHz (1H NMR), and 243 or 202.5 MHz (31P
NMR) at 300 0.5 K using DSS as an internal reference. The sample
concentration was less than 20 mM but more than 2 mM to avoid
intermolecular base stacking, but to obtain spectra with sufficient
intensity.4,31,49
ABBREVIATIONS:
■
A - adenine; G - guanine; N - Nucleoside; NTP - nucleoside 5′-
triphosphate; NDP - nucleoside 5′-diphosphate; M - metal ion;
L - ligand; H - proton; ML - complex of metal and ligand; AMP
- adenosine 5′-monophosphate
REFERENCES
■
NMR-Monitored Zn2+ Titrations of Nucleotides. Nucleoside
5′-phosphorothioate analogues 1−3, 6 and their corresponding parent
nucleotides (ADP, GDP, ATP, and AP3A) in D2O were titrated with
0.1, 0.2, 0.5, 1, 2, and 10 equiv of 2 M ZnNO3 D2O solution and
(1) Bianchi, E. M.; Sajadi, S. A. A.; Song, B.; Sigel, H. Chem.Eur. J.
2003, 9, 881−892.
(2) Sajadi, S. A. A.; Song, B.; Gregan, F.; Sigel, H. Inorg. Chem. 1999,
38, 439−448.
1
measured by 31P NMR spectroscopy at 202.5 or 243 MHz, and H
(3) Sigel, H.; Griesser, R. Chem. Soc. Rev. 2005, 34, 875−900.
(4) Sigel, H.; Tribolet, R.; Malinibalakrishnan, R.; Martin, R. B. Inorg.
Chem. 1987, 26, 2149−2157.
NMR spectroscopy at 500 or 600 MHz. The pD of the samples was
adjusted to approximately 7.4 with NaOD and DCl D2O solutions.
Conformational analysis (including NOESY studies) of analogue 5,
Na+ salt, and its Zn2+ complex was performed at 700 MHz in D2O (pD
= 7.4), using a 10 mM solution of analogue 5, with or without
Zn(NO3)2.
Potentiometric Titrations. Materials. The titrated nucleotides
were mostly ≥95% pure. ATP-γ-S, 3, tetralitium salt, was ≥90% pure.
Five M NaOH, 2 M HNO3, and potassium biphthalate were purchased
from Merck (Darmstadt, Germany). NaNO3 salt (background
electrolyte) and Zn(NO3)2 standard solution were purchased from
Sigma (Steinheim, Germany). All solutions for the titrations were
prepared with deionized water. The concentration of the titer of
NaOH was determined with potassium biphthalate, which was kept in
an oven at 130 °C and cooled to room temperature under a vacuum
before use for NaOH calibration. The concentration of the nucleotide
stock solutions was determined by titration with NaOH.
(5) Aoki, K.; Murayama, K. Met. Ions Life Sci. 2012, 10, 43−102.
(6) Orioli, P.; Cini, R.; Donati, D.; Mangani, S. J. Am. Chem. Soc.
1981, 103, 4446−4452.
(7) Pearson, R. G. Chemical hardness: Applications from Molecules to
Solids, 1st ed.; John Wiley & Sons: Weinheim, Germany, 1997.
(8) Harding, M. M. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2006,
62, 678−682.
(9) Berthon, G., Ed.; Handbook of Metal-Ligand Interactions in
Biological Fluids: Bioinorganic Medicine; Marcel Dekker: New York,
1995; Vol 1.
(10) Liebeskind, L. S.; Srogl, J.; Savarin, C.; Polanco, C. Pure Appl.
Chem. 2002, 74, 115−122.
(11) Krezel, A.; Lesniak, W.; Jezowska-Bojczuk, M.; Mlynarz, P.;
Brasun, J.; Kozlowski, H.; Bal, W. J. Inorg. Biochem. 2001, 84, 77−88.
(12) Brasun, J.; Matera, A.; Sochacka, E.; Swiatek-Kozlowska, J.;
Kozlowski, H.; Operschall, B. P.; Sigel, H. J. Biol. Inorg. Chem. 2008,
13, 663−674.
Potentiometric pH Titrations. The pH titrations were performed
with a Metrohm 794 basic Titrino potentiometer and a Metrohm
Viscotrode glass electrode. The buffers (pH 4.00 and 7.00, and 9.00)2
used for calibration were purchased from Metrohm (Herisau,
Switzerland). All titrations were performed at 24 °C under argon
atmosphere.
(13) Jaffe, E. K.; Cohn, M. Biochemistry 1978, 17, 652−657.
(14) Pecoraro, V. L.; Hermes, J. D.; Cleland, W. W. Biochemistry
1984, 23, 5262−5271.
(15) Sigel, R. K. O.; Song, B.; Sigel, H. J. Am. Chem. Soc. 1997, 119,
744−755.
Determination of the acidity constants, KHHL, KHLH2, and KHLH3 (L
-ligand), was made by the titration of 4.5 mL of aqueous 4 mL of
HNO3 in NaNO3 (from a 3 mM HNO3/NaNO3 stock solution, I =
0.1 M) and 0.5 mL of ligand (from a 4.5 mM nucleotide stock
solution, I = 0.1 M). Acidity constants were calculated with the
HYPERQUAD software.50 Determination of the stability constants,
KZnZnL, KZnZnH, and pKaHZnLH, was achieved by the titration of 3−3.5
mL of aqueous HNO3 in NaNO3 (from a 3 mM HNO3 stock solution,
I = 0.1 M), 0−0.5 mL NaNO3 (I = 0.1 M), 0.5 mL of ligand (from a
4.5 mM nucleotide stock solution, I = 0.1 M), and 0.5 mL of
Zn(NO3)2 (from a 4.5 mM stock solution, I = 0.1 M). 1:1 Zn2+:ligand
ratio was used. The pH range of all titrations was about 2.6−10.3. The
end points of nucleotide-complex titrations were obtained by the
second derivative method.51 Each titration was repeated up to 4 times.
Stability constants were calculated with the HYPERQUAD software.
(16) Andersson, M.; Hedin, J.; Johansson, P.; Nordstrom, J.; Nyden,
M. J. Phys. Chem. A 2010, 114, 13146−13153.
(17) Sundberg, R. J.; Martin, R. B. Chem. Rev. 1974, 74, 471−517.
(18) Stern, N.; Major, D. T.; Gottlieb, H. E.; Weizman, D.; Fischer,
B. Org. Biomol. Chem. 2010, 8, 4637−4652.
(19) Sigel, H.; Massoud, S. S.; Corfu, N. A. J. Am. Chem. Soc. 1994,
116, 2958−2971.
(20) Cohn, M.; Hughes, T. R., Jr. J. Biol. Chem. 1962, 237, 176−181.
(21) Cowan, J. A. Inorg. Chem. 1991, 30, 2740−2747.
(22) Huang, S. L.; Tsai, M. D. Biochemistry 1982, 21, 951−959.
(23) Jaffe, E. K.; Cohn, M. J. Biol. Chem. 1979, 254, 10839−10845.
(24) Jaffe, E. K.; Cohn, M. J. Biol. Chem. 1978, 253, 4823−4825.
(25) Son, T. D.; Roux, M.; Ellenberger, M. Nucleic Acids Res. 1975, 2,
1101−1110.
10895
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896

Inorganic Chemistry
Article
(26) Tran-Dinh, S.; Neumann, J. M. Nucleic Acids Res. 1977, 4, 397−
403.
(27) Bock, J. L. J. Inorg. Biochem. 1980, 12, 119−130.
(28) Sigel, H. Inorg. Chim. Acta 1992, 198, 1−11.
(29) Sigel, H. Pure Appl. Chem. 1998, 70, 969−976.
(30) Scheller, K. H.; Hofstetter, F.; Mitchell, P. R.; Prijs, B.; Sigel, H.
J. Am. Chem. Soc. 1981, 103, 247−260.
(31) Scheller, K. H.; Sigel, H. J. Am. Chem. Soc. 1983, 105, 5891−
5900.
(32) Sigel, H.; Massoud, S. S.; Tribolet, R. J. Am. Chem. Soc. 1988,
110, 6857−6865.
(33) Martin, R. B. Metal Ions Biol. Syst. 1996, 32, 61−89.
(34) Sigel, H.; Corf, N. A.; Ji, L.-N.; Martin, R. B. Comments Inorg.
Chem. 1992, 13, 35−59.
(35) Kapinos, L. E.; Holy, A.; Gunter, J.; Sigel, H. Inorg. Chem. 2001,
40, 2500−2508.
(36) Sigel, H.; Song, B. Metal Ions Biol. Syst. 1996, 32, 135−205.
(37) Sigel, H.; Bianchi, E. M.; Corfu, N. A.; Kinjo, Y.; Tribolet, R.;
Martin, R. B. Chem.Eur. J. 2001, 7, 3729−3737.
(38) Bianchi, E. M.; Griesser, R.; Sigel, H. Helv. Chim. Acta 2005, 88,
406−425.
(39) Stern, N.; Major, D. T.; Gottlieb, H. E.; Weizman, D.; Sayer, A.
H.; Blum, E.; Fischer, B. J. Biol. Inorg. Chem. 2012, 17, 861−879.
(40) Rosch, P.; Kalbitzer, H. R.; Goody, R. S. FEBS Lett. 1980, 121,
211−214.
(41) Da Costa, C. P.; Krajewska, D.; Okruszek, A.; Stec, W. J.; Sigel,
H. J. Biol. Inorg. Chem. 2002, 7, 405−415.
(42) Sutter, E. m. m.; Ammeloot, F.; Pouet, M. J.; Fiaud, C.;
Couffignal, R. Corros. Sci. 1999, 41, 105−115.
(43) Cho, B. P.; Evans, F. E. Nucleic Acids Res. 1991, 19, 1041−1047.
(44) Sigel, R. K. O.; Sigel, H. Met. Ions Life Sci. 2007, 2, 109−180.
(45) Alderighi, L.; Gans, P.; Ienco, A.; Peters, D.; Sabatini, A.; Vacca,
A. Coord. Chem. Rev. 1999, 184, 311−318.
(46) Song, B.; Sigel, R. K. O.; Sigel, H. Chem.Eur. J. 1997, 3, 29−
33.
(47) Maurizi, M. R.; Ginsburg, A. Biochemistry 1986, 25, 131−140.
(48) Kowalska, J.; Lewdorowicz, M.; Darzynkiewicz, E.; Jemielity, J.
Tetrahedron Lett. 2007, 48, 5475−5479.
(49) Mayo, K. H.; Mvele, O. M.; Puri, R. N. FEBS Lett. 1990, 265,
97−100.
(50) Gans, P.; Sabatini, A.; Vacca, A. Talanta 1996, 43, 1739−1753.
(51) Vogel, A. I.; Jeffery, G. H. Vogel’s textbook of quantitative chemical
analysis, 5th ed.; Longman Scientific & Technical: Essex, U.K., 1989.
10896
dx.doi.org/10.1021/ic400878k | Inorg. Chem. 2013, 52, 10886−10896