Metal ion-binding properties of 9-(4-phosphonobutyl)adenine (dPMEA), a sister compound of the antiviral nucleotide analogue 9-[2-(phosphonomethoxy)ethyl]adenine (PMEA), and quantification of the equilibria involving four Cu(PMEA) isomers

Article abstract of DOI:10.1039/b001588l

The acidity constants of the threefold protonated acyclic 9-(4-phosphonobutyl)adenine, H₃(dPMEA)⁺, as well as the stability constants of the M(H;dPMEA)⁺ and M(dPMEA) complexes with the metal ions M²⁺ = Mg²⁺, Ca²⁺, Sr²⁺, Ba²⁺, Mn²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺ or Cd²⁺, have been determined by potentiometric pH titrations, in aqueous solution at I= 0.1 M (NaNO₃) and 25 °C. Application of previously determined straight-line plots of log K_M(R-PO3)^M versus PK_H(R-PO3)^H for simple phosph(on)ate ligands, R-PO₃^2-, where R represents a residue without an affinity for metal ions, proves that the primary binding site of dPMEA^2- is the phosphonate group with all the metal ions studied; in fact, in most instances the stability is solely determined by the basicity of the phosphonate residue. Only for the Ni(dPMEA), Cu(dPMEA) and Cd(dPMEA) systems a stability increase due to macrochelate formation with the adenine residue occurs; the formation degrees are 21 ± 15%, 31 ± 14% and 29 ± 18%, respectively. In these three instances the additional interaction of the phosphonate-coordinated M²⁺ occurs most probably with N7; hence, dPMEA^2- is more similar in its metal ion-binding properties to the parent nucleotide adenosine 5′-monophosphate (AMP^2-) than to the antivirally active and structurally more related dianion of 9-[2-(phosphonomethoxy)ethyl]-adenine (PMEA^2-). This result agrees with the observation that replacement of the ether O atom in PMEA by a CH₂ unit leads to a compound, i.e. dPMEA, devoid of any biological activity. In addition, use is made of the stability enhancement obtained for the Cu(dPMEA) system due to macrochelate formation to analyze the equilibria regarding the four isomeric complex species possibly formed in the Cu(PMEA) system. It is shown that a macrochelated isomer involving N7 of the adenine residue occurs with Cu(PMEA) only in trace amounts; the important isomers in this system involve the ether oxygen (formation degree ca. 34%) and also N3 of the adenine moiety (ca. 41%). The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/b001588l

View Article Online / Journal Homepage / Table of Contents for this issue
Metal ion-binding properties of 9-(4-phosphonobutyl)adenine
(dPMEA), a sister compound of the antiviral nucleotide
analogue 9-[2-(phosphonomethoxy)ethyl]adenine (PMEA), and
quantification of the equilibria involving four Cu(PMEA) isomers
Raquel B. Gómez-Coca,a,b Larisa E. Kapinos,a Antonín Holý,*c Rosario A. Vilaplana,b
Francisco González-Vílchezb and Helmut Sigel*a
a Institute of Inorganic Chemistry, University of Basel, Spitalstrasse 51, CH-4056 Basel,
Switzerland. E-mail: Helmut.Sigel@unibas.ch
b Departamento de Química Inorgánica, Facultad de Química, Universidad de Sevilla,
E-4107 Sevilla, Spain
c Institute of Organic Chemistry and Biochemistry, Academy of Sciences, CZ-16610 Prague,
Czech Republic
Received 28th February 2000, Accepted 3rd May 2000
Published on the Web 13th June 2000
The acidity constants of the threefold protonated acyclic 9-(4-phosphonobutyl)adenine, H3(dPMEA)ϩ, as well as
the stability constants of the M(H;dPMEA)ϩ and M(dPMEA) complexes with the metal ions M2ϩ = Mg2ϩ, Ca2ϩ
,
Sr2ϩ, Ba2ϩ, Mn2ϩ, Co2ϩ, Ni2ϩ, Cu2ϩ, Zn2ϩ or Cd2ϩ, have been determined by potentiometric pH titrations, in aqueous
solution at I = 0.1 M (NaNO3) and 25 ЊC. Application of previously determined straight-line plots of log KMM(R-PO )
3
versus pKHH(R-PO ) for simple phosph(on)ate ligands, R-PO23Ϫ, where R represents a residue without an affinity for metal
3
ions, proves that the primary binding site of dPMEA2Ϫ is the phosphonate group with all the metal ions studied; in
fact, in most instances the stability is solely determined by the basicity of the phosphonate residue. Only for the
Ni(dPMEA), Cu(dPMEA) and Cd(dPMEA) systems a stability increase due to macrochelate formation with the
adenine residue occurs; the formation degrees are 21 15%, 31 14% and 29 18%, respectively. In these three
instances the additional interaction of the phosphonate-coordinated M2ϩ occurs most probably with N7; hence,
dPMEA2Ϫ is more similar in its metal ion-binding properties to the parent nucleotide adenosine 5Ј-monophosphate
(AMP2Ϫ) than to the antivirally active and structurally more related dianion of 9-[2-(phosphonomethoxy)ethyl]-
adenine (PMEA2Ϫ). This result agrees with the observation that replacement of the ether O atom in PMEA by a
CH2 unit leads to a compound, i.e. dPMEA, devoid of any biological activity. In addition, use is made of the stability
enhancement obtained for the Cu(dPMEA) system due to macrochelate formation to analyze the equilibria regarding
the four isomeric complex species possibly formed in the Cu(PMEA) system. It is shown that a macrochelated isomer
involving N7 of the adenine residue occurs with Cu(PMEA) only in trace amounts; the important isomers in this
system involve the ether oxygen (formation degree ca. 34%) and also N3 of the adenine moiety (ca. 41%).
CH2 group3 led to compounds devoid of any antiviral activity;
similarly, shortening or lengthening of the chain between the
phosphonyl group and the adenine residue resulted in inactive
products;12 the same is observed with most chain-substituted
derivatives of PMEA.13,14
1. Introduction
Nucleotides like adenosine 5Ј-monophosphate (AMP2Ϫ
)
1
and its diphosphorylated product, adenosine 5Ј-triphosphate
(ATP4Ϫ) as well as their 2Ј-deoxy derivatives (dAMP2Ϫ and
dATP4Ϫ), are at the crossroad of many metabolic processes.2
Therefore, the idea to exploit nucleotide analogues as
therapeutic agents is old. Among the many attempts, the acyclic
nucleoside phosphonates proved promising3,4 and one of
the better known compounds of this class is 9-[2-(phosphono-
methoxy)ethyl]adenine (PMEA), also named Adefovir.5 The
dianion of PMEA can be viewed as an analogue of AMP2Ϫ
(Fig. 1)6–9 or dAMP2Ϫ. The oral prodrug of PMEA, i.e. the
diester bis(pivaloyloxymethyl)–PMEA (also called Adefovir
Dipivoxil), is currently being evaluated in patients infected with
the human immunodeficiency virus (HIV) or the hepatitis B
virus (HBV).4,5 Moreover, very recently it was discovered that
the same prodrug also possesses antiarthritic properties.10
In the course of studies focussed on the antiviral properties
of compounds closely related to PMEA, it soon became
evident3 that the ether oxygen separated from the phosphonyl
group by a methylene unit is compulsory for a biological activ-
ity. For example, replacement of this oxygen by sulfur11 or a
In order to be antivirally active, PMEA must be phosphoryl-
ated in the cell15 to the diphosphate derivative, PMEApp4Ϫ, an
analogue of (d)ATP4Ϫ, it is then recognized by DNA poly-
merases as a substrate and thus incorporated into the growing
nucleic acid chain, which is then terminated due to the lack
of a 3Ј-hydroxy group.15,16 Indeed, PMEApp4Ϫ is initially an
excellent substrate for several polymerases; for example, even
in the presence of a 20-fold excess of dATP, in vitro DNA
synthesis by avian myeloblastosis-virus reverse-transcriptase is
depressed to 50% within 5 minutes.3,13
Knowing that for the activation of a nucleoside 5Ј-triphos-
phate two metal ions have to be coordinated to the triphosphate
chain17 and that for the transfer of a nucleotidyl unit one metal
ion must be bound to the β,γ-phosphate groups and the other
to the α-group, it was recently concluded18,19 that PMEApp4Ϫ is
initially a better substrate than dATP4Ϫ because of the higher
basicity of the phosphonyl group (compared to a phosphoryl
group) and the possibility of a coordinated metal ion (e.g.,
DOI: 10.1039/b001588l
J. Chem. Soc., Dalton Trans., 2000, 2077–2084
This journal is © The Royal Society of Chemistry 2000
2077

View Article Online
The buffer solutions (pH 4.00, 7.00, 9.00 based on the NBS
scale; now NIST) for calibration (Section 2.3) were from
Metrohm AG, Herisau, Switzerland.
The exact concentrations of the stock solutions of the
divalent metal ions were determined by potentiometric pH
titrations via their EDTA complexes. The exact concentration
of the ligand solutions was in each experiment newly deter-
mined by the evaluation of the corresponding titration pairs
described below in Section 2.4.
2.2 Synthesis of 9-(4-phosphonobutyl)adenine
The first step in the synthesis of 9-(4-phosphonobutyl)adenine
[= 4-(adenin-9-yl)butylphosphonic acid] was heating of a
mixture of 1,4-dibromobutane (100 g, 0.46 mol) and triethyl
phosphite (90 mL, 1.125 equiv.) for 2 days at 110 ЊC; the result-
ing ethyl bromide was distilled off. The portion volatile at
100 ЊC/2 kPa was removed. The residue was utilized in the next
step without further purification.
The mixture of adenine (20.0 g, 0.148 mol), 1,8-diazabicyclo-
[5.4.0]undec-7-ene (30 mL) and crude diethyl 4-bromobutyl-
phosphonate (65 mL) in dimethylformamide (200 mL) was
heated at 100 ЊC for 5 h and cooled down. The crystalline
material was filtered off, washed successively with dimethyl-
formamide and acetone, and recrystallized from 80% aqueous
ethanol. Yield, 12.4 g (25.6%) diethyl 4-(adenin-9-yl)butyl-
phosphonate. This compound (8.2 g, 25 mmol) in acetonitrile
(100 mL) was stirred with bromotrimethylsilane (15 mL) and
the solution was set aside overnight at ambient temperature.
The volatiles were evaporated in vacuo and the residue co-
distilled with toluene (2 × 25 mL). Water (150 mL) was added
and, after 20 min, the mixture was basified with conc. aqueous
ammonia and evaporated. The residue in minimum water was
applied to a column of Dowex 50 × 8 (Hϩ-form) (250 mL) and
the column was washed with water to the loss of acidity
and UV-absorption of the eluate (monitored at 254 nm). Sub-
sequent eluation of the column with diluted (1:10) aqueous
ammonia gave a UV-absorbing eluate which was evaporated to
dryness in vacuo. This residue in minimum water was basified
with ammonia to pH 10 and applied onto a column of Dowex
1 × 2 (acetate form) prewashed with water (200 mL); the
column was eluted with water (1 L) and 0.25 M acetic acid (500
mL). The resin was then stirred batchwise with 1 M formic acid
(500 mL), filtered off and washed with boiling water (four 500
mL portions). The combined eluates were evaporated in vacuo,
the residue codistilled with water (five 50 mL portions) and
crystallized from water. Yield, 4.3 g (64%), mp 287 ЊC. For
C9H14N5O3PؒH2O (289.2) calculated 37.37% C, 5.58% H,
24.21% N, 10.71% P; found 37.20% C, 5.44% H, 24.00% N,
Fig. 1 Chemical structures of the dianions of 9-[2-(phosphonometh-
oxy)ethyl]adenine (= PMEA2Ϫ = Adefovir) and 9-(4-phosphonobutyl)-
adenine (= dPMEA2Ϫ = 3Ј-deoxa-PMEA2Ϫ
) and of their parent
nucleotide adenosine 5Ј-monophosphate (AMP2Ϫ), which is shown
in its dominating anti conformation.6–8 The orientation of PMEA2Ϫ in
solution9 is similar to the anti conformation of AMP2Ϫ
.
Mg2ϩ) to form a five-membered chelate with the ether oxygen.
These two properties facilitate the M(α)–M(β,γ) coordination
pattern needed for the enzyme-catalyzed incorporation of the
substrate in the growing nucleic acid chain and thus favour
PMEApp4Ϫ over dATP4Ϫ (for details see ref. 19).
This conclusion is in accord with the above mentioned obser-
vation that the ether oxygen of the acyclic chain is compulsory
for obtaining antivirally active nucleotide analogues. Hence,
it was of interest to study the metal ion-binding properties of
the carba analogue of PMEA, i.e. of 9-(4-phosphonobutyl)-
adenine which is also known as 3Ј-deoxa-PMEA (dPMEA) or
4-(adenin-9-yl)butylphosphonic acid (Fig. 1). Indeed, the results
of the present study show that the properties of the metal
ion complexes formed with PMEA2Ϫ and dPMEA2Ϫ differ
considerably.
2. Experimental
2.1 Materials
1
11.00% P. H-NMR (D2O ϩ NaOD, 500 MHz): δ 8.11 s, 1H
Twofold protonated 9-(4-phosphonobutyl)adenine, i.e.
H2(dPMEA) , was synthesized (by modification of an earlier
procedure)20 as described below in Section 2.2 with 1,4-
dibromobutane, triethyl phosphite, 1,8-diazabicyclo[5.4.0]-
undec-7-ene, bromotrimethylsilane and dimethylformamide,
which were purchased from Sigma-Aldrich, Prague, Czech
Republic. The aqueous stock solutions of the ligand were
freshly prepared daily just before the experiments by dissolving
the substance in deionized, ultrapure (MILLI-Q185 PLUS;
from Millipore S.A., 67120 Molsheim, France) CO2-free water
and adding 2 equivalents of NaOH.
The disodium salt of 1,2-diaminoethane-N,N,NЈ,NЈ-tetra-
acetic acid (Na2H2EDTA), potassium hydrogen phthalate,
HNO3, NaOH (Titrisol), and the nitrate salts of Naϩ, Mg2ϩ,
Ca2ϩ, Sr2ϩ, Ba2ϩ, Mn2ϩ, Co2ϩ, Ni2ϩ, Cu2ϩ, Zn2ϩ and Cd2ϩ
(all pro analysi) were from Merck AG, Darmstadt, FRG. All
solutions for the potentiometric pH titrations were prepared
with ultrapure CO2-free water. Dimethylformamide and
acetonitrile, used for the synthesis (Section 2.2), were also
obtained from Merck; they were dried by distillation from
phosphorus pentoxide and stored over molecular sieves.
and 8.10 s, 1H (H-2 and H-8); 4.17 t, 2H, J(1Ј,2Ј) = 7.3 (H-1Ј);
1.88 qnt, 2H, J(2Ј,1Ј) = J(2Ј,3Ј) = 7.3 (H-2Ј); 1.54 m, 2H and
1.44 m, 2H (H-3Ј and H-4Ј).
2.3 Potentiometric pH titrations
The pH titration curves for the determination of the equi-
librium constants in H2O were recorded with a Metrohm E536
potentiograph connected to a Metrohm E535 dosimat and a
Metrohm 6.0222.100 combined macro glass electrode. The pH
calibration of the instrument was done with the buffers
mentioned above. The titer of the NaOH used was determined
with potassium hydrogen phthalate.
The direct pH meter readings were used in the calculations of
the acidity constants; i.e. these constants determined at I = 0.1
M (NaNO3) and 25 ЊC are so-called practical, mixed or Brønsted
constants.21 They may be converted into the corresponding
concentration constants by subtracting 0.02 from the listed pKa
values;21 this conversion term contains both the junction poten-
tial of the glass electrode and the hydrogen ion activity.21,22 It
should be emphasized that the ionic product of water (Kw) and
2078
J. Chem. Soc., Dalton Trans., 2000, 2077–2084

View Article Online
the mentioned conversion term do not enter into our calcula-
tion procedures because we evaluated the differences in NaOH
consumption between a pair of solutions, i.e. with and without
ligand (see below in Section 2.4). The stability constants deter-
mined are, as usual, concentration constants.
All equilibrium constants were calculated by curve-fitting
procedures in the way and with the equipment described
recently.23
diazabicyclo[5.4.0]undec-7-ene (DBU) in dimethylformamide24
proceeded smoothly under formation of diethyl 4-(adenin-9-
yl)butylphosphonate which, in turn, gave on transsilylation
with bromotrimethylsilane followed by hydrolysis the target
compound. The purification was performed by desalting on
cation ion-exchanger followed by anion-exchange chrom-
atography and crystallization from water.
Since derivatives of purines are well known to undergo self-
association via π-stacking,25 a concentration26 of 3 × 10Ϫ4
M
2.4 Determination of equilibrium constants
was used in this study for dPMEA in the potentiometric pH
titration experiments to ascertain that only the properties of
the monomeric species are quantified.
The acidity constants KHH (dPMEA) and KHH(dPMEA) of H2(dP-
2
MEA) , where one proton is at the base moiety and the other at
the phosphonate group, were determined by titrating 50 mL of
aqueous 0.9 mM HNO3 (I = 0.1 M, NaNO3; 25 ЊC) in the
presence and absence of 0.3 mM dPMEA2Ϫ under N2 with 1.55
mL of 0.03 M NaOH. The differences in NaOH consumption
between such a pair of titrations were used for the calculations.
The pH range evaluated was determined by the lowest point of
neutralization reached under these experimental conditions
(ca. 30% for the equilibrium H2(dPMEA) /H(dPMEA)Ϫ) and
the neutralization degree of about 98% for the equilibrium
H(dPMEA)Ϫ/dPMEA2Ϫ. The results are the averages of 38
pairs of independent titrations.
3.1 Acidity constants of H3(dPMEA)؉
From the structure of dPMEAϪ (see Fig. 1) it is evident that
this species can accept three protons, two at the phosphonate
group and one at the N1 site of the adenine residue.6 Further
protonations of an adenine residue are possible at N7 and N3,
but these protons are released with pKa < 0,9,27,28 and therefore,
they are not considered in this study. At pH > 0 the strongest
acid that can be derived in aqueous solution from dPMEA2Ϫ
is H3(dPMEA)ϩ. Hence, the following three deprotonation
reactions, in which dPMEA2Ϫ is abbreviated as PA2Ϫ, need to be
considered:
The acidity constant for the release of the first proton from
the phosphonate group in H3(dPMEA)ϩ, KHH (dPMEA), was
3
H3(PA)ϩ
H2(PA) ϩ Hϩ
(1a)
(1b)
(2a)
(2b)
(3a)
(3b)
determined by titrating 20 mL of aqueous 0.03 M HNO3
(I = 0.1 M, NaNO3; 25 ЊC) in the presence and absence of 1.66
mM dPMEA2Ϫ under N2 with 6 mL of 0.1 M NaOH and by
using the differences in NaOH consumption between such a
pair of titrations for the calculations. The highest formation
degree reached for H3(dPMEA)ϩ is close to 50%. The result is
the average of 9 pairs of titrations.
KHH (PA) = [H2(PA) ] [Hϩ]/[H3(PA)ϩ]
3
H2(PA)
H(PA)Ϫ ϩ Hϩ
KHH (PA) = [H(PA)Ϫ] [Hϩ]/[H2(PA) ]
2
The stability constants KMM(H;dPMEA) and KMM(dPMEA) of
M(H;dPMEA)ϩ and M(dPMEA) were determined under
H(PA)Ϫ
PA2Ϫ ϩ Hϩ
the same conditions as the acidity constants KHH (dPMEA) and
2
KHH(PA) = [PA2Ϫ] [Hϩ]/[H(PA)Ϫ]
KHH(dPMEA), but NaNO3 was partly or fully replaced by M(NO3)2
(I = 0.1 M, 25 ЊC). The M2ϩ/dPMEA ratios were 111:1 (Mg2ϩ,
Ca2ϩ, Sr2ϩ, Ba2ϩ, Mn2ϩ), 89:1 (Mg2ϩ), 56:1 (Co2ϩ, Ni2ϩ, Cd2ϩ),
28:1 (Co2ϩ, Ni2ϩ, Zn2ϩ, Cd2ϩ), 14:1 (Cd2ϩ), 11:1 (Cu2ϩ, Zn2ϩ),
5.6:1 (Cu2ϩ) and 5:1 (Zn2ϩ).
Indeed, all the experimental data obtained from the
potentiometric pH titrations in aqueous solution (25 ЊC; I = 0.1
M, NaNO3) could be excellently fitted by taking into account
equilibria (1) to (3). Furthermore, in the pH range above 3.6
only equilibria (2) and (3) need to be considered. The acidity
constants obtained in the present study for H3(dPMEA)ϩ are
given in Table 1, together with some related data.7,9,26,29–32
Comparison of the values listed in Table 1 confirms the above
statement that the first and third proton of H3(dPMEA)ϩ are
from the phosphonate group and the second proton is released
from the N1 site of the adenine residue. However, more com-
parisons are possible; a few are emphasized below:
The stability constants were calculated by considering Hϩ,
H2(dPMEA) , H(dPMEA)Ϫ, dPMEA2Ϫ, M2ϩ, M(H;dPMEA)ϩ
and M(dPMEA) and by collecting the experimental data every
0.1 pH unit from about 2% complex formation of M(H;dP-
MEA)ϩ to a neutralization degree of about 90% with respect to
the species H(dPMEA)Ϫ, or until the beginning of the
hydrolysis of M(aq)2ϩ, which was evident from the titrations
without ligand. The maximal formation degree for the M(dP-
MEA) complexes varies between 30 and 70%, depending on the
metal ion. However, for the Zn(dPMEA) and Cd(dPMEA)
complexes only a formation degree of about 2.2% and 19% was
reached, respectively; hence, the value for Zn(dPMEA) must be
considered as an estimate.
(i) From the comparisons of entry 2 with 6 and of 3, 4 with 5
it follows that the protonated phosphonate groups are less
acidic than a corresponding phosphate group due to the lower
electronegativity of a C atom compared to an O atom.
(ii) The ether oxygen, due to its electron-withdrawing effect,
makes the protonated phosphonate species of PMEA more
acidic than those of dPMEA (entries 3, 4).
The individual results for the stability constants showed no
dependence on pH or on the excess of metal ion concentration
used. The results are in each case the averages of at least 5
independent pairs of titration curves. However, most of the
constants given for the M(H;dPMEA)ϩ complexes must also be
considered as estimates (see Table 2 below), since the formation
degree of these species reached only about 4 to 10% in the
maximum.
(iii) In fact, this acidification is equal for both the H(PA)Ϫ and
H3(PA)ϩ species as the following comparison shows (entries
3,
4):
∆pKa/1 = pKHH(dPMEA) Ϫ pKHH(PMEA) = (7.69 0.01) Ϫ
(6.90 0.01) = 0.79 0.01
and
∆pKa/3 = pKHH (dPMEA) Ϫ
3
pKHH (PMEA) = (1.98 0.13) Ϫ (1.22 0.13) = 0.76 0.18.
3
(iv) This observation contrasts with the acid–base properties of
N1 which remain unaffected by the presence of the oxygen
atom in the acyclic chain (see column 4, entries 3, 4).
3. Results and discussion
The ligand 9-(4-phosphonobutyl)adenine [=4-(adenin-9-
yl)butylphosphonic acid] was synthesized by modification of a
described procedure:20 1,4-dibromobutane gave by reaction
with triethyl phosphite diethyl 4-bromobutylphosphonate
which was used in the following step without purification;
alkylation of adenine with this synthon in the presence of 1,8-
(v) A further remarkable observation is that the following dif-
ferences are all equal within their error limits: pKHH(dPMEA) Ϫ
pKHH (dPMEA) = (7.69 0.01) Ϫ (1.98 0.13) = 5.71 0.13 (entry
3
3), pKHH(PMEA) Ϫ pKHH (PMEA) = (6.90 0.01) Ϫ (1.22 0.13) =
5.68 0.13 (entry 4), 3 and pKHH(AMP) Ϫ pKHH (AMP) = (6.21
3
0.01) Ϫ (0.4 0.2) = 5.8 0.2 (entry 5).
J. Chem. Soc., Dalton Trans., 2000, 2077–2084
2079

View Article Online
Table 1 Negative logarithms of the acidity constants of H3(dPMEA)ϩ [equilibria (1) to (3)] together with the corresponding values of some related
systems in aqueous solution (25 ЊC; I = 0.1 M, NaNO3)a,b
No.
Protonated species
pKHH (PA): P(O)(OH)2
pKHH (PA): (N1)Hϩ
pKHH(PA): P(O)2(OH)Ϫ
Ref.
3
2
1
2
3
4
5
6
H(9MeAde)ϩ
CH3P(O)(OH)2
H3(dPMEA)ϩ
H3(PMEA)ϩ
4.10 0.01
29
30
—
2.10 0.03
1.98 0.13
1.22 0.13c
0.4 0.2c
1.1 0.2
7.51 0.01
7.69 0.01
6.90 0.01
6.21 0.01
6.36 0.01
4.17 0.02
4.16 0.02
3.84 0.02
9,c 31
H3(AMP)ϩ
7,c 26
32
CH3OP(O)(OH)2
a The error limits 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 constants are listed (see also Section 2.3). c Determined by 1H-NMR shift measurements.
Table 2 Logarithms of the stability constants of the M(H;dPMEA)ϩ
Finally, it needs to be pointed out that due to the different
[eqn. (4b)] and M(dPMEA) complexes [eqn. (5b)], together with the
affinities of the phosph(on)ate groups for protons the form-
negative logarithms of the acidity constants of the protonated com-
ation degree of the free ligands, PA2Ϫ, differs considerably in the
physiological pH range around 7.5: Under these conditions
AMP2Ϫ is formed to about 95% whereas PMEA2Ϫ still reaches
about 80%, yet dPMEA2Ϫ occurs only to about 40%.
plexes [eqns. (6b) and (7)] in aqueous solution at 25 ЊC and I = 0.1 M,
(NaNO3)a
M2ϩ
b
log KMM(H;dPMEA)
log KMM(dPMEA)
pKMH(H;dPMEA)
3.2 Stability constants of the M(H;dPMEA)؉ and M(dPMEA)
complexes
Mg2ϩ
Ca2ϩ
Sr2ϩ
0.3 0.4
0.2 0.4
0.1 0.4
0.0 0.4
0.4 0.4
0.7 0.3
1.0 0.3
1.78 0.13
1.4 0.3
1.4 0.3
1.84 0.04
1.57 0.05
1.30 0.04
1.22 0.06
2.47 0.03
2.31 0.06
2.41 0.06
3.72 0.07
2.75 0.2b
3.08 0.10b
6.2 0.4
6.3 0.4
6.5 0.4
6.5 0.4
5.6 0.4
6.1 0.3
6.3 0.3
5.75 0.15
6.3 0.4
6.0 0.3
The experimental data of the potentiometric pH titrations (see
Section 2.3) allow the determination of the stability constants
defined by equilibria (4a) and (5a):
Ba2ϩ
Mn2ϩ
Co2ϩ
Ni2ϩ
Cu2ϩ
Zn2ϩ
Cd2ϩ
M2ϩ ϩ H(PA)Ϫ
M(H;PA)ϩ
(4a)
KMM(H;PA) = [M(H;PA)ϩ]/([M2ϩ] [H(PA)Ϫ]) (4b)
a For the error limits see footnote a of Table 1. The error limits (3σ) of
the derived data, in the present case for column 4, were calculated
according to the error propagation after Gauss. b The constants listed
for the M(H;dPMEA)ϩ complexes are estimates (see last paragraph in
Section 2.4). The experiments with Zn2ϩ were significantly hampered by
precipitation; i.e. the pH range accessible for the evaluation of the con-
stants was severely restricted. The same problem, though less severe,
occurred also with Cd2ϩ (see also Section 2.4).
M2ϩ ϩ PA2Ϫ
M(PA)
(5a)
(5b)
KMM(PA) = [M(PA)]/([M2ϩ] [PA2Ϫ])
Overall, equilibria (2)–(5) are sufficient to obtain excellent
fitting of the titration data, provided the evaluation is not car-
ried into the pH range where formation of hydroxo species
occurs, which was evident from titrations without ligand. Of
course, equilibria (4a) and (5a) are also connected via equi-
librium (6a)
the proton and the metal ion. At first one may ask where the
proton is located because binding of a metal ion to a proton-
ated ligand commonly leads to an acidification of the
ligand-bound proton.41 Indeed, the acidity constants of the
M(H;dPMEA)ϩ complexes given in column 4 of Table 2 are
by about 1.2 to 2 log units smaller than pKHH(dPMEA) (Table 1),
M(H;PA)ϩ
M(PA) ϩ Hϩ
(6a)
(6b)
KHM(H;PA) = [M(PA)][Hϩ]/[M(H;PA)ϩ]
but 1.4 to 2.3 log units larger than pKHH (dPMEA). This com-
2
parison shows that the proton in M(H;dPMEA)ϩ is bound
to the phosphonate group, hence, one may tentatively assume
that the metal ion is bound preferentially to the nucleo-
base, since a monoprotonated phosphonate group is only
a weak binding site. Indeed, this suggestion agrees with
evidence obtained previously for other related M(H;PA)ϩ
species.31,40,42
and the corresponding acidity constant [eqn. (6b)] may be
calculated with eqn. (7):
pKHM(H;PA) = pKHH(PA) ϩ log KMM(H;PA) Ϫ log KMM(PA) (7)
The results are listed in Table 2; the stability constants given
for the M(H;dPMEA)ϩ complexes are only estimates since the
formation degree of these species was low (see Section 2.3). The
stability constants of the M(dPMEA) complexes show the
usual trends. For the alkaline earth ions the stability of the
complexes decreases with increasing ionic radii indicating that
metal ion binding at the phosphonate group is (at least) in part
inner-sphere. For the divalent 3d metal ions the long-standing
experience33,34 is confirmed that the stabilities of phosph(on)ate–
metal ion complexes often do not strictly follow23,26,31–38 the
Irving–Williams sequence,39 an observation in accord with the
fact that in ligands of this kind the phosph(on)ate group is
always the main binding site23,26,36–38,40 in M(PA) complexes (see
Section 3.3).
Furthermore, the stability constants of the M(H;dPMEA)ϩ
complexes are within the error limits identical with the values
determined31 for the corresponding M(H;PMEA)ϩ species.
Considering that the basicity of the N1 site in H(dPMEA)Ϫ and
H(PMEA)Ϫ is also identical (see Table 1; column 4, entries 3,4)
and that evidence has been provided for the M(H;PMEA)ϩ
complexes31,40 that the metal ion is mainly located at the
nucleobase residue, one may not only conclude that in the
M(H;dPMEA)ϩ complexes the proton is at the phosphonate
group, but also that the metal ion is mainly at the adenine
residue. The N1 versus N7 dichotomy for metal ion binding to
the adenine residue is well known43 though there are indications
that binding to N7 dominates.43,44 In any case, the fact that the
stabilities of the M(H;dPMEA)ϩ complexes follow the Irving–
Williams sequence39 (in contrast to phosph(on)ate complexes)
also supports34 the above conclusion that metal ion binding
occurs preferably to a nitrogen atom.
As far as the M(H;dPMEA)ϩ complexes are concerned, it is
evident that the evaluation of potentiometric pH titration data
only allows the determination of their stability constants.
Further information is required to detect the binding sites of
2080
J. Chem. Soc., Dalton Trans., 2000, 2077–2084

View Article Online
Table 3 Stability constant comparisons for the M(dPMEA) com-
plexes between the measured stability constants (exptl; Table 2, column
3) and the calculated stability constants (calcd) based on the basicity of
the phosphonate group in dPMEA2Ϫ (pKHH(dPMEA) = 7.69; Table 1) and
the baseline equations established previously26,31,38 [see eqn. (8) and Fig.
2] together with the stability differences log ∆M/dPMEA, as defined by eqn.
(10). The previously determined stability enhancements, log ∆M/PMEA
,
for M(PMEA) complexes are given for comparison31 (aqueous solu-
tion; 25 ЊC; I = 0.1 M, NaNO3)a
log KMM(dPMEA)
M2ϩ
exptl
calcd
log ∆M/dPMEA
log ∆M/PMEA
Mg2ϩ
Ca2ϩ
Sr2ϩ
1.84 0.04
1.57 0.05
1.30 0.04
1.22 0.06
2.47 0.03
2.31 0.06
2.41 0.06
3.72 0.07
2.75 0.2c
1.87 0.03
1.64 0.05
1.36 0.04
1.29 0.04
2.51 0.05
2.27 0.06
2.31 0.05
3.56 0.06
2.64 0.06
Ϫ0.03 0.05
Ϫ0.07 0.07
Ϫ0.06 0.06
Ϫ0.07 0.07
Ϫ0.04 0.06
0.04 0.08
0.10 0.08
0.16 0.09
0.1 0.2
0.16 0.05
0.11 0.07
0.07 0.05
0.08 0.06
0.21 0.08
0.28 0.07b
0.30 0.07b
0.77 0.07b
0.30 0.10d
0.33 0.06
Ba2ϩ
Mn2ϩ
Co2ϩ
Ni2ϩ
Cu2ϩ
Zn2ϩ
Cd2ϩ
3.08 0.10c 2.93 0.05
0.15 0.11
a Regarding the error limits (3σ) see footnote a of Tables 1 and 2.
b These values contain also a contribution40 of an N3 interaction;19,31
for Co(PMEA) see ref. 19. c See footnote b of Table 2. d Estimated value.
Ni2ϩ and Cu2ϩ, as examples, with the data points (empty circles)
of the eight simple ligand systems used31 for the determination
of the straight baselines. The four solid points refer to the
corresponding M(dPMEA) complexes; those for the Ni2ϩ and
Cu2ϩ species are somewhat above their reference lines, thus
proving an increased stability for these two complexes, whereas
the data points for the Mg2ϩ and Ba2ϩ complexes fit on the lines
(or are even slightly below). In contrast, all of the data points31
of the M(PMEA) complexes (crossed circles) are above their
reference lines indicating that all these species are more stable
than is expected on the basis of the basicity of the phosphonate
group of PMEA2Ϫ.
This increased stability of the M(PMEA) complexes has been
proven19,31 to be largely due to the involvement of the ether
oxygen atom in metal ion binding (see Fig. 1) which gives rise to
the formation of 5-membered chelates and thus, to the intra-
molecular equilibrium (9):19,31,36
Fig. 2 Evidence for an enhanced stability of the Ni(dPMEA) and
Cu(dPMEA) complexes, and for the lack of such an enhanced stability
of the Ba(dPMEA) and Mg(dPMEA) species (᭹), together with the
corresponding metal ion complexes of PMEA2Ϫ (ꢀ), which all show an
enhanced stability, based on the relationship between log KMM(R-PO ) and
3
pKHH(R-PO ) for M(R-PO3) complexes of some simple phosphate
3
monoester and phosphonate ligands (R-PO23Ϫ) (᭺): 4-nitrophenyl phos-
phate (NPhP2Ϫ), phenyl phosphate (PhP2Ϫ), uridine 5Ј-monophosphate
(UMP2Ϫ), D-ribose 5-monophosphate (RibMP2Ϫ), thymidine [=1-
(2-deoxy-β-D-ribofuranosyl)thymine] 5Ј-monophosphate (TMP2Ϫ),
n-butyl phosphate (BuP2Ϫ), methanephosphonate (MeP2Ϫ), and ethane-
phosphonate (EtP2Ϫ) (from left to right). The least-squares lines
[eqn. (8)] are drawn through the corresponding 8 data sets (᭺) taken
from ref. 46 for the phosphate monoesters and from ref. 31 for the
phosphonates. The points due to the equilibrium constants for the M2ϩ
/
dPMEA systems (᭹) are based on the values listed in Tables 1 and 2;
those for the M2ϩ/PMEA systems (ꢀ) are from ref. 31. The vertical
broken lines emphasize the stability differences from the reference lines;
they equal log ∆M/dPMEA as defined in eqn. (10) for the M(dPMEA)
complexes. All the plotted equilibrium constants refer to aqueous
solutions at 25 ЊC and I = 0.1 M (NaNO3).
(9)
3.3 Evaluation of the stabilities of the M(dPMEA) complexes
In certain systems, like Ni(PMEA) and Cu(PMEA), a third
isomer exists which involves in addition N3 of the adenine
residue19,31,36,40 (see also Section 3.5).
The question that arises for the M(dPMEA) complexes is: does
the adenine residue also participate in metal ion binding next to
the phosphonate group? Should such an additional interaction
with the adenine residue occur then it has to be reflected in an
increased complex stability.45 Hence, it is necessary to define the
stability of a pure –PO32Ϫ/M2ϩ interaction. This can be done by
applying the previously defined31,38 straight-line correlations
Naturally, in complexes of dPMEA2Ϫ, whose side-chain is
devoid of the ether oxygen atom (Fig. 1), equilibrium (9)
cannot exist and therefore, any possibly observed stability
increase has to be attributed to an interaction with the adenine
residue. Stability enhancements like those seen in Fig. 2 can be
quantified by the differences between the experimentally (exptl)
measured stability constants and those calculated (calcd)
according to eqn. (8); this difference is defined in eqn. (10)
which are based on log KMM(R-PO ) versus pKHH(R-PO ) plots for
3
simple phosphate monoesters46 and phosphonat3es;31 these
ligands are abbreviated as R-PO32Ϫ, where R represents a non-
coordinating residue. The parameters for the corresponding
straight-line equations, which are defined by eqn. (8), have been
log∆M/PA = log KMM(PA) Ϫ log KMM(PA)
(10a)
(10b)
exptl
calcd
= log KMM(PA) Ϫ log KMM(PA) (= log ∆)
log KMM(R-PO ) = m ؒ pKHH(R-PO ) ϩ b
(8)
op
3
3
tabulated,26,31,36a,38 i.e. the slopes m and the intercepts b with the
y-axis. Hence, with a known pKa value for the deprotonation of
a –P(O)2(OH)Ϫ group an expected stability constant can be
calculated for any phosph(on)ate–metal ion complex.
where the expressions log KMM(PA) and log KMM(Pa) are syn-
calcd
op
onymous because the calculated value equals the stability
constant of the ‘open’ isomer, M(PA)op [see e.g. equilibrium
(9)], in which only a –PO32Ϫ/M2ϩ interaction occurs. In columns
2–4 of Table 3 the values for the three terms of eqn. (10)
are listed. It is evident that only the log ∆M/dPMEA values for
The plots of log KMM(R-PO ) versus pKHH(R-PO ) according to eqn.
3
3
(8) are shown in Fig. 2 for the 1:1 complexes of Ba2ϩ, Mg2ϩ,
J. Chem. Soc., Dalton Trans., 2000, 2077–2084
2081

View Article Online
Ni(dPMEA), Cu(dPMEA) and Cd(dPMEA) are positive. The
observed stability increases are small but certainly real (note
that three times the standard error is given); hence, in these
instances equilibrium (11) has to be considered:
Cu(dPMEA) (Table 3).48 The answer to the question raised is
clearly “no” as shown below; such a species forms only in trace
amounts in the Cu(PMEA) system.
It has previously been proven40a that three isomers are
important for the Cu(PMEA) system.19 (i) An ‘open’ isomer,
Cu(PMEA)op, in which the metal ion is solely coordinated to
the phosphonate group; (ii) an isomer which involves the ether
oxygen (see Fig. 1) as shown in equilibrium (9), designated
as Cu(PMEA)cl/O; and (iii) an isomer in which not only a 5-
membered chelate but in addition a 7-membered one involving
(11)
3.4 Extent of macrochelate formation in the M(dPMEA)
systems of Ni2؉, Cu2؉ and Cd2؉
N3 exists, i.e. Cu(PMEA)cl/O/N3
.
In the present context it is important to emphasize that for
steric reasons no macrochelate involving only N3 can be formed
with PMEA2Ϫ and Cu2ϩ.36a If one tries to form such a species
with molecular models, one automatically forces the ether
oxygen into the coordination sphere of Cu2ϩ giving rise to the
already mentioned Cu(PMEA)cl/O/N3 isomer.36a Hence, we may
postulate for Cu(PMEA) that the whole stability increase
observed for Cu(dPMEA), if relevant for Cu(PMEA), has to
be attributed to macrochelate formation [eqn. (11)] of the
phosphonate-bound Cu2ϩ with N7 of the adenine residue
leading to the postulation of a fourth isomer designated as
Cu(PMEA)cl/N7. With these reasonings in mind the simple
equilibrium (5a) must be replaced for the Cu(PMEA) system by
the rather complicated equilibrium scheme (13):
The position of the concentration-independent equilibrium
(11) between the simple phosphonate-bound species, which we
designate as the ‘open’ isomer, M(dPMEA)op, and the macro-
chelated isomer involving the adenine residue, designated as the
‘closed’ species, M(dPMEA)cl, is defined by the intramolecular
and hence, dimensionless equilibrium constant KI [eqn. (12)]:
KI = [M(dPMEA)cl]/[M(dPMEA)op]
(12)
Application of log ∆M/dPMEA (Table 3, column 4) allows one
to define the position of equilibrium (11) as shown previ-
ously.26,31,38,45 For the Ni(dPMEA) system one calculates
∆
0.10
for KI = 10log
Ϫ 1 = 10
Cu(dPMEA) system KI = 10log
0.45 0.30 and for the Cd(dPMEA) system KI =
0.08 Ϫ 1 = 0.26 0.23, for the
∆ 0.16
Cu/dPMEA
Ni/dPMEA
Ϫ 1 = 10
0.09 Ϫ 1 =
(13)
10log
Ϫ 1 = 10
–1 = 0.41 0.36; hence, the form-
∆
0.15 0.11
Cd/dPMEA
ation degree of the macrochelated or closed species amounts
to 21 15%, 31 14% and 29 18% for Ni(dPMEA)cl,
Cu(dPMEA)cl and Cd(dPMEA)cl, respectively. For all the
other M(dPMEA) systems the formation degree of the
closed species in equilibrium (11) is zero within the error limits
(see the log ∆M/dPMEA values in Table 3). In contrast, in the
M(PMEA) complexes the 5-membered chelates seen in equi-
librium (9) are important for all metal ions as is evident from
the log ∆M/PMEA values given in column 5 of Table 3 [for details
and the formation degrees of M(PMEA)cl see refs. 31 and 36;
slightly revised values are given in ref. 19] (see also Section 3.5).
From the increased complex stabilities observed for the Ni2ϩ,
Cu2ϩ and Cd2ϩ complexes of dPMEA2Ϫ, it follows that at least
for these three metal ions, which have the most pronounced
affinity for N sites among the ten metal ions studied,34
equilibrium (11) operates and that the adenine residue is
involved in metal ion binding. From the three nitrogen atoms
available in the purine ring, i.e. N1, N3 and N7 (see Fig. 1),
N1 cannot be reached by a metal ion already coordinated to
the phosphonate group;36 hence, N3 and N7 remain for macro-
chelate formation. At this stage one cannot distinguish with
certainty between these two possibilities, but we strongly favor
an interaction of the phosphonate-coordinated metal ion with
N7 because N7 is known to be more basic than N3.9,27,28
Furthermore, macrochelate formation in the M(AMP) com-
plexes occurs with certainty via N7;26,47 hence, it appears that
M(dPMEA) resembles in this respect the parent complexes,
i.e. M(AMP), more than M(PMEA), though the extent of
macrochelate formation is more pronounced in the M(AMP)
systems26,38 than in the M(dPMEA) species.
The four equilibrium constants of scheme (13) are defined by
eqns. (14)–(17):
KMM(PA) = [M(PA)op]/([M2ϩ][PA2Ϫ])
(14)
(15)
(16)
(17)
op
KI/N7 = [M(PA)cl/N7]/[M(PA)op]
KI/O = [M(PA)cl/O]/[M(PA)op]
KI/O/N3 = [M(PA)cl/O/N3]/[M(PA)cl/O
]
With these definitions the measured overall stability constant
[eqn. (5b)] can be redefined as in eqns. (18a)–(18d):
[M(PA)]
[M2ϩ][PA2Ϫ
KMM(PA)
=
(18a)
]
(18b)
[M(PA)op] ϩ [M(PA)cl/N7] ϩ [M(PA)cl/O] ϩ [M(PA)cl/O/N3]
=
[M2ϩ][PA2Ϫ
= KMM(PA) ϩ KI/N7ؒKMM(PA) ϩ KI/OؒKMM(PA)
KI/O/N3ؒKI/OؒK
]
ϩ
op
op
op
M
(18c)
(18d)
M(PA)op
= KMM(PA) (1 ϩ KI/N7 ϩ KI/O ϩ KI/OؒKI/O/N3
)
op
The connection between the overall intramolecular equi-
librium constant19,31,36 KI/tot and the accessible stability
enhancement [eqn. (10)] is given by eqns. (19a)–(19e):
3.5 Lesson learnt from the properties of Cu(dPMEA) regarding
the formation degree of the isomeric complexes formed in the
Cu(PMEA) system. A four-isomer problem
KMM(PA)
(19a)
KI/tot
=
=
=
Ϫ 1 = 10log ∆ Ϫ 1
M
K
M(PA)op
Since it is highly likely that the enhanced complex stability
observed for Cu(dPMEA) has to be attributed to macrochelate
formation [eqn. (11)] of the phosphonate-bound Cu2ϩ with
N7 of the adenine residue (Section 3.4), one may ask: is such
a macrochelate formation involving N7 also important in
the Cu(PMEA) complex system? We are concentrating the
attention here on the Cu2ϩ systems since the most pronounced
stability enhancement (log ∆ = 0.16 0.09) is observed for
[M(PA)cl/tot
[M(PA)op]
]
(19b)
(19c)
[M(PA)cl/N7] ϩ [M(PA)cl/O] ϩ [M(PA)cl/O/N3
[M(PA)op]
]
= KI/N7 ϩ KI/O ϩ KI/O/N3ؒKI/O
= KI/N7 ϩ KI/O (1 ϩ KI/O/N3
(19d)
(19e)
)
2082
J. Chem. Soc., Dalton Trans., 2000, 2077–2084

View Article Online
Table 4 Intramolecular equilibrium constants (KI) for the formation of the various Cu(PMEA) isomers as defined in the equilibrium scheme (13),
together with the percentages in which the isomers occur in aqueous solution at 25 ЊC and I = 0.1 M (NaNO3)a
No.
eqns. used
log ∆
KI
Formation degrees in percentages
1a
1b
(10)
(19a), (19b)
log ∆Cu/PMEA
KI/tot
% Cu(PMEA)cl/tot
83 3c
% Cu(PMEA)op
17 3d
0.77 0.07b
4.89 0.95
KI/O
2.02 0.49
KI/N7
0.45 0.30
2a
2b
(10), (16)
[analogous to (19a)] (16)
log ∆Cu/PME-R
% Cu(PMEA)cl/O
34 10f
0.48 0.07e
3a
3b
(10), (15)
see Section 3.4
log ∆Cu/dPMEA
% Cu(PMEA)cl/N7
7.7 5.3h
0.16 0.09g
4a
4b
(17)
(19e)
KI/O/N3
% Cu(PMEA)cl/O/N3
41 12j
1.20 0.73i
a Regarding the error limits (3σ) see footnote a of Tables 1 and 2. Note, in this table always entries (a) and (b) go together. b From ref. 31 [see also refs.
19, 36 and 40(a)]. c Calculated with the equation: % Cu(PMEA)cl/tot = 100 × KI/tot/(1 ϩ KI/tot). d This percentage follows from 100 Ϫ % Cu(PMEA)cl/tot
.
e From ref. 49 (see also ref. 19). f Calculated with KI/O and % Cu(PMEA)op by application of eqn. (16). g From Table 3. h Calculated with KI/N7 and %
Cu(PMEA)op by application of eqn. (15). i This value follows from eqn. (19e) since all the other intramolecular equilibrium constants are now known.
j This value follows from the difference % Cu(PMEA)cl/tot Ϫ % Cu(PMEA)cl/O Ϫ % Cu(PMEA)cl/N7; it could also be calculated with KI/O/N3 and %
Cu(PMEA)cl/O by application of eqn. (17).
A value for KI/tot can be calculated [see eqns. (10b) and (19a)]
4. Conclusions
and its relation to the other three intramolecular equilibrium
The analysis of Section 3.5 indicates that in the Cu(PMEA)
constants follows from eqns. (19b) and (19c). Based on the
system the Cu(PMEA)cl/N7 isomer [eqn. (11)] is likely to be
reasonable assumption19 that the stability of the Cu(PMEA)cl/O
formed in trace amounts. The corresponding calculations are
isomer is well represented by that of the Cu(PME-R)cl/O species,
the first example, to the best of our knowledge, of a detailed
evaluation of the equilibria involving four isomeric complexes.
Of course, the fact that the Cu(PMEA)cl/N7 isomer occurs only
where R in PME-R2Ϫ is a nucleobase residue without the ability
to interact with metal ions,49 a value for KI/O, which defines
the position of equilibrium (9), is also known.19,49 Since the
in low concentration distinguishes this system from the
stability of the Cu(PMEA)cl/N7 isomer is assumed to be repre-
situation in Cu(dPMEA) (Section 3.4) and especially in
sented (but see also below in the terminating paragraph of
Cu(AMP).26,38,47 In accord herewith it was concluded in Section
3.4 that the M(dPMEA) complexes resemble more the parent
M(AMP) complexes than do the M(PMEA) species.
this section) by that of the Cu(dPMEA)cl/N7 species, as dis-
cussed and postulated above, the corresponding KI/N7 value is
also known (see Section 3.4). Hence, a value for KI/O/N3 [eqn.
(17)] can be obtained by applying eqn. (19e), and therefore
Indeed, the properties of the binary M(dPMEA) and
M(PMEA) complexes differ significantly. With PMEA2Ϫ the
the formation degree of all four isomers appearing in scheme
ether oxygen affects the complex structures in solution [equi-
librium (9)]; there is no equivalent in the M(dPMEA) systems.
(13) can be calculated.19,36 The corresponding results are
summarized in Table 4.
If it is assumed that dPMEA is transported to the cell and also
From Table 4 it is evident that the Cu(PMEA)cl/O/N3 species
diphosphorylated, like PMEA,15,16 then it would become under-
with the 5- and 7-membered chelate rings dominates with a
formation degree of 41 ( 12)%; it is revealing to see that this
result is within the error limits identical with the previously
standable why dPMEA lacks3 antiviral activity. A facilitated
M(α) binding (via the ether oxygen) is not possible18 with
dPMEApp4Ϫ and thus the formation of the M(α)–M(β,γ)-
obtained 49 10% where the formation of the fourth isomer,
binding mode, which is crucial for the transfer of a nucleo-
Cu(PMEA)cl/N7, had not been taken into account.19,31 This
tidyl unit in the polymerase reaction,19 would also not be
shows immediately that the Cu(PMEA)cl/N7 isomer must have a
low formation degree; indeed, the present calculations provide
facilitated.
As far as the correct location of dPMEA and PMEA in the
7.7 5.3% (Table 4, entry 3b) proving that it is a minority
active sites of enzymes is concerned, one would not expect a
species.50
significant difference since the hydrogen-bonding properties of
the adenine residue should be identical in both ligands. This
It needs to be emphasized that this value (7.7 5.3%) must
be considered as an upper limit for two reasons. The nucleotide
also seems to be true for the stacking interactions with other
analogue dPMEA2Ϫ contains a pure aliphatic chain, whereas
aromatic-ring systems; the extent of stack formation in the
the chain of PMEA2Ϫ has an ether oxygen incorporated (see
ternary Cu(Arm)(PMEA) and Cu(Arm)(dPMEA) complexes,
Fig. 1), and this difference may have consequences: (i) The
where Arm = 2,2Ј-bipyridine or 1,10-phenanthroline, is quite
alike51 indicating that this property is not significantly
affected by the presence of the ether oxygen atom in the acyclic
chain between the phosphonate group and the adenine residue.
purely aliphatic chain of dPMEA2Ϫ may undergo a weak
hydrophobic interaction with the aromatic imidazole-ring of
the adenine residue upon “folding” of the ligand as needed for
the formation of the Cu(dPMEA)cl/N7 isomer; such an inter-
action would facilitate the formation of the latter species. (ii)
The ether oxygen in the –CH2CH2–O–CH2–PO32Ϫ residue of
PMEA2Ϫ will be solvated by H2O and it will then somewhat
inhibit the “folding” process needed for the formation of
Acknowledgements
The competent technical assistance of Mrs Bela Nováková and
Mrs Rita Baumbusch in the synthetic work and the prepar-
ation of this manuscript, respectively, is gratefully acknow-
ledged. This study was supported by the Swiss National Science
Foundation (H. S.), the Grant Agency of the Czech Republic
(203/96/K001; A. H.) and the General Health Insurance
Agency of the Czech Republic (A. H.) as well as within
the COST D8 programme by the Swiss Federal Office for
Education and Science (H. S.) and the Ministry of Education
of the Czech Republic (A. H.).
Cu(PMEA)cl/N7. Both points mean that the value log
∆
Cu/dPMEA = 0.16 0.09 is somewhat too large, if applied to the
Cu(PMEA) system, and therefore the formation degree of
about 8% calculated for the Cu(PMEA)cl/N7 isomer may well be
too large but certainly it is not too low. At the same time one
may conclude that the present calculations indicate that the
Cu(PMEA)cl/N7 isomer is likely to be formed in trace amounts;
e.g., a calculation with the reduced value, log ∆Cu/N7 = 0.1, gives
a formation degree of about 4% for the Cu(PMEA)cl/N7 isomer.
J. Chem. Soc., Dalton Trans., 2000, 2077–2084
2083

View Article Online
22 H. M. Irving, M. G. Miles and L. D. Pettit, Anal. Chim. Acta, 1967,
Notes and references
38, 475.
1 Abbreviations and definitions: see Fig. 1; ATP4Ϫ, adenosine 5Ј-
23 C. A. Blindauer, T. I. Sja˚stad, A. Holý, E. Sletten and H. Sigel,
J. Chem. Soc., Dalton Trans., 1999, 3661.
triphosphate; dAMP2Ϫ
,
2Ј-deoxyadenosine 5Ј-monophosphate;
dATP4Ϫ, 2Ј-deoxyadenosine 5Ј-triphosphate; I, ionic strength; M2ϩ
,
24 A. Holý, J. Günter, H. Dvoráková, M. Masojídková, G. Andrei,
ˇ
divalent metal ion; 9MeAde, 9-methyladenine; PA2Ϫ, dPMEA2Ϫ or
any other twofold negatively charged nucleotide or nucleotide
analogue; PMEApp4Ϫ, diphosphorylated PMEA; R-PO23Ϫ, simple
phosphate monoester or phosphonate ligand with R representing a
non-coordinating residue (see also legend of Fig. 2). 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. In formulas
like M(H;dPMEA)ϩ, Hϩ and dPMEA2Ϫ are separated by a semi-
colon to facilitate reading, yet they appear within the same paren-
thesis to indicate that the proton is at the ligand without defining its
location (see Section 3.2).
2 (a) Interactions of Metal Ions with Nucleotides, Nucleic Acids and
Their Constituents, vol. 32 of Metal Ions in Biological Systems, eds.
A. Sigel and H. Sigel, Dekker, New York, 1996, pp. 1–814; (b)
Interrelations among Metal Ions, Enzymes and Gene Expression, vol.
25 of Metal Ions in Biological Systems, eds. H. Sigel and A. Sigel,
Dekker, New York, 1989, pp. 1–557.
R. Snoeck, J. Balzarini and E. De Clercq, J. Med. Chem., 1999, 42,
2064.
25 O. Yamauchi, A. Odani, H. Masuda and H. Sigel, Met. Ions Biol.
Syst., 1996, 32, 207 [see ref. 2(a)].
26 H. Sigel, S. S. Massoud and N. A. Corfù, J. Am. Chem. Soc., 1994,
116, 2958.
27 R. L. Benoit and M. Fréchette, Can. J. Chem., 1984, 62, 995.
28 C. Meiser, B. Song, E. Freisinger, M. Peilert, H. Sigel and
B. Lippert, Chem. Eur. J., 1997, 3, 388.
29 L. E. Kapinos, G. Kampf, R. Griesser, B. Lippert and H. Sigel,
Chimia, 1999, 53, 348.
30 H. Sigel, C. P. Da Costa, B. Song, P. Carloni and F. Gregánˇ, J. Am.
Chem. Soc., 1999, 121, 6248.
31 H. Sigel, D. Chen, N. A. Corfù, F. Gregánˇ, A. Holý and M. Strasák,
ˇ
Helv. Chim. Acta, 1992, 75, 2634.
32 A. Saha, N. Saha, L.-n. Ji, J. Zhao, F. Gregánˇ, S. A. A. Sajadi,
B. Song and H. Sigel, J. Biol. Inorg. Chem., 1996, 1, 231.
33 H. Sigel, K. Becker and D. B. McCormick, Biochim. Biophys. Acta,
1967, 148, 655.
34 H. Sigel and D. B. McCormick, Acc. Chem. Res., 1970, 3, 201.
35 S. A. A. Sajadi, B. Song, F. Gregánˇ and H. Sigel, Inorg. Chem., 1999,
38, 439.
36 H. Sigel, (a) Coord. Chem. Rev., 1995, 144, 287; (b) J. Indian Chem.
Soc., 1997, 74, 261 (P. Ray Award Lecture).
37 H. Sigel, Chem. Soc. Rev., 1993, 22, 255.
3 A. Holý, E. De Clercq and I. Votruba, ACS Symp. Ser., 1989, 401,
51.
4 E. De Clercq, Collect. Czech. Chem. Commun., 1998, 63, 480.
5 (a) L. Naesens, R. Snoeck, G. Andrei, J. Balzarini, J. Neyts and
E. De Clercq, Antiviral Chem. Chemother., 1997, 8, 1; (b)
D. Colledge, G. Civitico, S. Locarnini and T. Shaw, Antimicrob.
Agents Chemother., 2000, 44, 551.
6 R. B. Martin and Y. H. Mariam, Met. Ions Biol. Syst., 1979, 8,
57.
38 H. Sigel and B. Song, Met. Ions Biol. Syst., 1996, 32, 135 [see ref.
2(a)].
39 H. Irving and R. J. P. Williams, Nature (London), 1948, 162, 746;
7 R. Tribolet and H. Sigel, Eur. J. Biochem., 1987, 163, 353.
8 K. Aoki, Met. Ions Biol. Syst., 1996, 32, 91 [see ref. 2(a)].
J. Chem. Soc., 1953, 3192.
40 (a) C. A. Blindauer, A. H. Emwas, A. Holý, H. Dvoráková,
9 C. A. Blindauer, A. Holý, H. Dvoráková and H. Sigel, J. Chem.
ˇ
ˇ
Soc., Perkin Trans. 2, 1997, 2353.
E. Sletten and H. Sigel, Chem. Eur. J., 1997, 3, 1526; (b) C. A.
10 (a) D. Franková, Z. Zídek, E. Buchar, Z. Jiøièka and A. Holý,
Nucleosides Nucleotides, 1999, 18, 955; (b) Z. Zídek, D. Franková
and A. Holý, Eur. J. Pharmacol., 1999, 376, 91.
11 D. Villemin and F. Thibault-Starzyk, Synth. Commun., 1993, 23,
1053.
Blindauer, A. Holý, H. Dvoráková and H. Sigel, J. Biol. Inorg.
ˇ
Chem., 1998, 3, 423.
41 (a) H. Sigel and B. Lippert, Pure Appl. Chem., 1998, 70, 845; (b)
B. Song, J. Zhao, R. Griesser, C. Meiser, H. Sigel and B. Lippert,
Chem. Eur. J., 1999, 5, 2374.
42 M. S. Lüth, L. E. Kapinos, B. Song, B. Lippert and H. Sigel,
J. Chem. Soc., Dalton Trans., 1999, 357.
43 H. Sigel, N. A. Corfù, L.-n. Ji and R. B. Martin, Comments Inorg.
Chem., 1992, 13, 35.
44 R. B. Martin, Met. Ions Biol. Syst., 1996, 32, 61 [see ref. 2(a)].
45 R. B. Martin and H. Sigel, Comments Inorg. Chem., 1988, 6,
285.
12 E. De Clercq, T. Sakuma, M. Baba, R. Pauwels, J. Balzarini,
I. Rosenberg and A. Holý, Antiviral Res., 1987, 8, 261.
ˇ
13 A. Holý, I. Votruba, A. Merta, J. Cerný, J. Veselý, J. Vlach,
ˇ
K. Sedivá, I. Rosenberg, M. Otmar, H. Hrebabecký, M. Trávnícˇek,
ˇ
V. Vonka, R. Snoeck and E. De Clercq, Antiviral Res., 1990, 13, 295.
14 H. Dvorákova, A. Holý and I. Rosenberg, Collect. Czech. Chem.
ˇ
Commun., 1994, 59, 2069.
15 (a) J. Balzarini, Z. Hao, P. Herdewijn, D. G. Johns and E. De Clercq,
Proc. Natl. Acad. Sci. USA, 1991, 88, 1499; (b) B. L. Robbins,
J. Greenhaw, M. C. Connelly and A. Fridland, Antimicrob. Agents
Chemother., 1995, 39, 2304.
46 S. S. Massoud and H. Sigel, Inorg. Chem. 1988, 27, 1447.
47 H. Sigel, S. S. Massoud and R. Tribolet, J. Am. Chem. Soc., 1988,
110, 6857.
48 It may be added that for Co(dPMEA) no enhanced complex
stability is observed (Table 3); hence, the formation of macro-
chelated Co[(d)PMEA] species is not of relevance (see also ref. 19).
Similarly, the enhanced stability of the Cd(PMEA) system19,31,36 is
solely explained by equilibrium (9); i.e., there is no evidence for a
nucleobase–Cd2ϩ interaction. For the Ni[(d)PMEA] systems a
calculation analogous to the one carried out here for the Cu2ϩ
systems could be done, but because of the smaller stability
enhancements (compared to those observed with the Cu2ϩ systems)
and the relatively large error limits no conclusive results are
expected.
16 (a) A. Merta, I. Votruba, I. Rosenberg, M. Otmar, H. Hrebabecký,
ˇ
R. Bernaerts and A. Holý, Antiviral Res., 1990, 13, 209; (b) S. A.
Foster, J. Cerny and Y.-c. Cheng, J. Biol. Chem., 1991, 266, 238;
(c) P. Kramata, I. Votruba, B. Otová and A. Holý, Molecular
Pharmacol., 1996, 49, 1005; (d) G. Birkus, I. Votruba, A. Holý and
B. Otová, Biochem. Pharmacol., 1999, 58, 487.
17 (a) H. Sigel, Coord. Chem. Rev., 1990, 100, 453; (b) H. Sigel, Pure
Appl. Chem., 1998, 70, 969.
18 H. Sigel, B. Song, C. A. Blindauer, L. E. Kapinos, F. Gregánˇ and
N. Prónayová, Chem. Commun., 1999, 743.
19 H. Sigel, Pure Appl. Chem. 1999, 71, 1727.
49 C. A. Blindauer, A. Holý and H. Sigel, Collect. Czech. Chem.
Commun., 1999, 64, 613.
50 Of course, the sum of the now calculated 41% plus the 7.7% give the
20 (a) I. Rosenberg, A. Holý and M. Masojídková, Collect. Czech.
Chem. Commun., 1988, 53, 2753; (b) A. Holý, in Advances in
Antiviral Drug Design, ed. E. De Clercq, JAI Press, Greenwich, CT,
1993, vol. 1, pp. 179–231.
49% obtained previously for Cu(PMEA)cl/O/N3
.
51 R. B. Gómez-Coca, L. E. Kapinos, A. Holý, R. A. Vilaplana,
F. González-Vílchez and H. Sigel, to be submitted for
publication.
21 H. Sigel, A. D. Zuberbühler and O. Yamauchi, Anal. Chim. Acta,
1991, 255, 63.
2084
J. Chem. Soc., Dalton Trans., 2000, 2077–2084