Metal ion-promoted cleavage of mRNA 5′-cap models: Hydrolysis of the triphosphate bridge and reactions of the N7-methylguanine base

Article abstract of DOI:10.1039/b108222a

Reactions of mRNA 5′-cap model compounds were studied to evaluate the potential of these reactions in the development of artificial RNases. Diadenosine triphosphate was used as a model for the triphosphate bridge, and its hydrolysis was studied in the pre

Full text of DOI:10.1039/b108222a

View Article Online / Journal Homepage / Table of Contents for this issue
Metal ion-promoted cleavage of mRNA 5Ј-cap models:
hydrolysis of the triphosphate bridge and
7
reactions of the N -methylguanine base
2
Satu Valakoski, Suvi Heiskanen, Sanna Andersson, Mari Lähde and Satu Mikkola*
University of Turku, Department of Chemistry, FIN-20014 Turku, Finland
Received (in Cambridge, UK) 12th September 2001, Accepted 18th December 2001
First published as an Advance Article on the web 25th January 2002
Reactions of mRNA 5Ј-cap model compounds were studied to evaluate the potential of these reactions in the
development of artificial RNases. Diadenosine triphosphate was used as a model for the triphosphate bridge, and its
2ϩ
hydrolysis was studied in the presence of several Cu complexes. The results of the kinetic experiments show that
2ϩ
bifunctional catalysis by phosphate bound Cu complexes is involved. The most efficient catalysis is achieved with
complexes with acidic aqua ligands, and a metal ion-bound hydroxo ligand most probably acts as a nucleophile in the
7
reaction. A detailed mechanism cannot, however, be suggested on the basis of the data. N -methylguanosine and its
Ј-monophosphate and diphosphate were used to study the reactions of the N -methylguanine base of the mRNA
Ј-cap moiety. While Cu complexes efficiently enhance the hydrolysis of the triphosphate bridge, little effect on the
reactions of the N -methylguanine base was observed: neither the cleavage of the imidazole ring or the depurination
of the nucleoside were enhanced to any significant extent.
7
5
5
2ϩ
7
Introduction
a large variety of metal ions promote the hydrolysis of
10–14
nucleotides,
such as ATP, only a few metal ions enhance the
Selective elimination of intracellular mRNA molecules by con-
jugates that recognise their target and actively enhance a chem-
ical modification within it is generally believed to form a novel
method to treat viral or bacterial infections or to prevent the
propagation of tumor cells. The 5Ј-terminal cap-structure of
eukaryotic mRNA molecules, which is essential for stability
and intracellular processing of the molecule, is an attractive
target for chemical modification. Even a modest structural
alteration in the 5Ј-cap moiety has been shown to prevent the
reaction of dialkyl triphosphates. In the case of 5Ј-cap models,
the most efficient catalysis has been achieved with complexes
16,20
2ϩ 15,17–19
of trivalent lanthanide ions
and Cu .
Information
about the catalytic properties of metal ion complexes other
1
2ϩ
than those of lanthanide ions or Cu is scarce. Monometallic
2
2ϩ
Zn species have been mentioned as clearly less active than
3
–7
2ϩ
15
Cu complexes. In all cases constructs that contain more
than one metal ion generally are more efficient catalysts than
8
17,18
mononuclear metal ion complexes.
recognition of the mRNA molecule by the cap-binding
enzymes involved and hence the biosynthesis of the correspond-
ing protein molecule. Specific inhibition of protein biosynthesis
Another potential cleavage site within the 5Ј-cap structure
7
is the N -methylguanine base. The base is electrophilic, and
under alkaline conditions its imidazole ring undergoes a rapid
cleavage as a result of a nucleophilic attack by a hydroxide ion
3ϩ
resulting from the Eu complex promoted cleavage of the
9
5
Ј-cap structure has recently been reported.
21
at C8 (Scheme 2). An attack by a nucleophile other than the
7
The 5Ј-cap structure consists of N -methylguanosine
attached to the terminal nucleotide of the RNA strand via a
5–7
5
Ј,5Ј-triphosphate bridge (Scheme 1). While the hydrolysis
Scheme 2
hydroxide ion has not been reported, even though it could be
presumed that the positively charged form of the base present
under neutral conditions is even more electrophilic than the
zwitterion ion predominating under alkaline conditions
Scheme 1
22
(
pK 7.3 at 25 ЊC).
a
10–14
of nucleoside triphosphates has been extensively studied,
This work studies the hydrolysis of simple mRNA 5Ј-cap
studies on the hydrolysis of dialkyl triphosphates are less
abundant. Results obtained with 5Ј-cap models have shown that
similarly to the hydrolysis of nucleoside triphosphates, the
hydrolysis of dinucleoside triphosphates is promoted by metal
model compounds in the presence of metal ion complexes to
find efficient catalysts for their cleavage, and to evaluate the
potential of the catalysts as constituents of artificial nucleases
targeted towards the 5Ј-cap moiety of mRNA. Hydrolysis of
diadenosine triphosphate (ApppA; 1,) to AMP (2) and ADP (3)
(Scheme 3), and the hydrolysis of diphosphate and tetraphos-
15–20
ions and their complexes.
There appears, however, to be
a difference between these two classes of compounds: while
6
04
J. Chem. Soc., Perkin Trans. 2, 2002, 604–610
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b108222a

View Article Online
material were calculated by using an integrated first-order rate
law. Products were identified by spiking with authentic samples.
In reactions of 1, in the presence of complexes 10–14, the only
products formed were adenosine 5Ј-monophosphate (AMP, 2)
and adenosine 5Ј-diphosphate (ADP, 3). Reactions of the tetra-
phosphate 5 additionally produced adenosine 5Ј-triphosphate
(ATP, 15). AMP was the only product in reactions of 4. For-
mation of adenosine (16) or adenine (17) was not observed.
Scheme 3
phate analogues of 1 (4 and 5, respectively) was studied to
obtain detailed mechanistic information about the metal
ion-promoted hydrolysis of the triphosphate bridge. These
compounds can be regarded as relevant models, since metal
ion complexes most probably interact with these substrates
only through the triphosphate moiety. A bipyridine ligand has
been reported to prevent an interaction of Cu with the base
moiety and a bidentate binding of Cu -terpyridine to two
phosphate groups of adenosine diphosphate has been shown by
2ϩ
23
2ϩ
24
an X-ray structure. The effect of the metal ion complexes
on the ring-opening reaction of the N -methylguanine base
7
(
(
Scheme 2) was studied by using the corresponding nucleoside
7
7
N -mGuo; 6) and its 5Ј-monophosphate (N -mGMP; 7) and
7
diphosphate (N -mGDP; 8) as model compounds.
In HEPES [HEPES = N-(2-hydroxyethyl)piperazine-NЈ-
2ϩ
ethanesulfonic acid] buffers the Cu TerPy promoted hydrolysis
of 1, 5 and 15 was accompanied by a formation of a by-product
that most probably is a HEPES adduct of 5Ј-AMP. Consistent
2ϩ
Results
with the assignment, reactions of Cu TerPy with 1 in CHES
[CHES = 2-(N-cyclohexylamino)ethanesulfonic acid] buffers of
Hydrolysis of the triphosphate bridge
comparable basicity only gave AMP and ADP. In the CHES
buffers these products were formed in an equimolar concen-
tration, while in HEPES buffers the concentration of AMP
observed was only 60% of the concentration of ADP. The
rate constant of disappearance of 1 in the presence of 10 mM
1
3
The hydrolysis of P ,P -diadenosine triphosphate (1) (Scheme
1
2
1
4
3
), P ,P -diadenosine diphosphate (4) and P ,P -diadenosine
tetraphosphate (5) was studied in the presence of different
metal ion complexes (9–14) at 60 ЊC under neutral and slightly
alkaline conditions. Aliquots withdrawn from the reaction
solutions were analysed either by capillary zone electrophoresis
2ϩ
Cu –terpyridine in HEPES buffer is 20 % larger than a rate
constant of the same reaction in a CHES buffer. Results
of HPLC-MS/MS characterisation (M ϩ 1 value of 567), the
UV-spectrum obtained by a diode array detector and the
electrophoretic behaviour of the product are also consistent
with the suggestion.
(
CZE) or reversed-phase HPLC. In both analysis methods, the
starting material and all the products could be separated under
the conditions described in the Experimental section. Pseudo-
first-order rate constants of the disappearance of the starting
J. Chem. Soc., Perkin Trans. 2, 2002, 604–610
605

View Article Online
complex: the maximal rate constants obtained differ by a factor
2ϩ
of 7. The pH–rate profile of the Zn [12]aneN (13) promoted
3
reaction consists only of four points, but it appears that the
2ϩ
maximal rate is obtained at the same pH as with Cu TerPy, but
the levelling off may be slightly less pronounced.
Even though the shape of the pH–rate profiles of the cleavage
of 4 and 5 seem to be similar to those determined for the
cleavage of 1 under the same conditions, the absolute values of
the rate constants are different. The cleavage of diadenosine
2ϩ
tetraphosphate (5) by the Cu –TerPy complex is two to four
2ϩ
times faster than the Cu TerPy-promoted cleavage of 1, which
may well result from the stronger electrostatic attraction by the
more anionic tetraphosphate moiety. The rate constants of
the cleavage of diadenosine diphosphate (4) are approximately
one order of magnitude smaller than those of the cleavage of
the triphosphate 1. The reactivity difference in the presence
2ϩ
of Cu BiPy appears to be of the same magnitude. The rate
2ϩ
constant of the Cu BiPy (2.0 mM) promoted hydrolysis of 4 at
Ϫ5 Ϫ1
pH 7.5 and at 60 ЊC is (1.09 ± 0.02) × 10 s .
2ϩ
The effect of pH on the rate constants of the Cu –
terpyridine (Cu TerPy, 9) promoted hydrolysis of diadenosine
The rate-enhancement by the complexes is rather significant.
The spontaneous hydrolysis of 1 was followed over a pH range
from 6.0 to 10.0 for more than four months. A rate constant
2ϩ
triphosphate (1), diphosphate (4) and tetraphosphate (5),
2
ϩ
2ϩ
2ϩ
and of the Cu –bipyridine (Cu BiPy; 10) and Zn –1,5,9-
could only be obtained at pH 6.0, where a value of 7.5 ×
2ϩ
triazacyclododecane (Zn [12]aneN , 13) promoted reaction of
Ϫ8 Ϫ1
3
1
0
s
was determined. At pH higher than that, less than 5%
1
over a limited pH range from 5.5 to 8.5 is shown in Fig. 1.
of the starting material was cleaved in four months at 60 ЊC,
Ϫ9 Ϫ1
suggesting that the rate constants are of order of 5 × 10
s
or less. At least a 20000-fold rate-enhancement can hence be
2ϩ
obtained with 2 mM Cu BiPy complex under neutral and
2ϩ
slightly alkaline conditions. The bimetallic Cu 2–bisdien com-
plex (14) is a slightly more efficient catalyst, but rate constants
could only be obtained under neutral conditions. The rate con-
stant obtained in the presence of 1 mM complex at pH 5.7 and
Ϫ5 Ϫ1
6
0 ЊC was (7.4 ± 0.2) × 10 s . At a higher pH a precipitate
was formed during the kinetic run. The complex also appears to
be rapidly inactivated under the experimental conditions, which
has previously been attributed to irreversible formation of
25
strong phosphate complexes. Consistent with previously
1
4
2ϩ
reported results, Zn [12]aneN is a rather poor catalyst in
3
2ϩ
comparison to Cu complexes. Over the pH range studied,
2ϩ
Cu BiPy promoted reactions, for example, are approximately a
hundred times more efficient than those promoted by 13. An
efficient reaction of the bisdien ligand with ATP has previously
1
3,26
been reported
but no such activity was observed with
1
: bisdien in 2 mM concentration did not induce any reaction
to any significant extent in three weeks at pH 7.0 and 60 ЊC.
The rate constants of the cleavage of 1 as a function of the
2ϩ
Cu complex concentration are shown in Fig. 2. The concen-
tration of the complexes ranges from 2 to 10 mM. The narrow
concentration range is determined by practical reasons: to
maintain pseudo-first-order conditions, the concentration of
the catalyst had to be at least 10 times higher than that of the
substrate. On the other hand, the complexes tend to precipitate
if the concentration is too high. It can be seen in Fig. 2 that
there is a major difference between the complexes studied:
reactions catalysed by different complexes show a dissimilar
dependence of the rate on the metal ion concentration. The
Fig. 1 The effect of pH on the metal ion promoted hydrolysis of the
phosphate bridge of 5Ј-cap model compounds. Legend: (᭺) ApppA
2ϩ
2ϩ
(
(
(
1)–2mM Cu BiPy; (᭿) AppppA (5)–2 mM Cu TerPy; (᭹) ApppA
2
ϩ
2
ϩ
1)–2 mM Cu TerPy; (᭜) AppA (4)–2 mM Cu TerPy; (ꢀ) ApppA
2ϩ
1)–2 mM Zn [12]aneN . pH adjusted with 0.1 M HEPES buffer,
3
ionic strength 0.1 M adjusted with NaNO3.
Under these conditions the hydrolysis of 1 and 5 exhibited clear
first-order kinetics, the plots ln A vs. t being linear. In reactions
of 4 some curvature was observed on a longer reaction time.
This, however, most probably results from slow dissociation of
the complexes rather than a second-order dependence of the
reaction rate on [4]. As the reactions of 4 are slower than the
hydrolysis of 1 or 5, the inactivation of the catalysts is only
observed in this case.
2ϩ
Cu TerPy promoted reaction shows a higher than first-order
dependence on the complex concentration. The slopes of the
Ϫ1
log (k/s ) vs. log (c/M) plots were 1.5 and 2.0 at pH 6.0 and 8.0,
respectively (Table 1). In contrast, values of 0.9 and 0.8 were
2ϩ
obtained for the Cu BiPy promoted hydrolysis. A similar
2
ϩ
As can be seen in Fig. 1, the hydrolysis of 1, 4 and 5 in
difference was observed between Cu [12]aneN (12) and
3
2ϩ
2ϩ
the presence of Cu TerPy exhibits a first-order dependence
on hydroxide ion concentration at pH below 8, but the
rate-increase levels off to zero-order dependence as the pH is
Cu [9]aneN (11) ([9]aneN = 1,4,7-triazacyclononane). The
slopes of the plots log (k/s ) vs. log (c/M) were 1.9 and 0.8,
3
3
Ϫ1
respectively (Table 1). The slope also seems to be dependent on
2ϩ
2ϩ
further increased. The optimal pH of the Cu BiPy-promoted
the substrate; in Cu [12]aneN promoted reaction of AppA
3
cleavage of 1 is lower than that of the corresponding reaction
(4), the slope is clearly smaller than in reaction of 1 (Table 1).
2ϩ
in the presence of Cu TerPy: The maximal rate is reached
The steep dependence of the rate constants on the concen-
2ϩ
2ϩ
2ϩ
around pH 7. Under these conditions the Cu BiPy -promoted
tration of Cu TerPy and Cu [12]aneN shows that more than
3
2ϩ
reactions are faster than those promoted by the Cu TerPy-
one metal ion probably is involved in the rate-limiting step of
6
06 J. Chem. Soc., Perkin Trans. 2, 2002, 604–610

View Article Online
Table 1 The effect of the copper complex concentration on the hydrolysis of the phosphoanhydride bridge mRNA 5Ј-cap models 1 and 4 at 60 ЊC
a
b
Ϫ1
Substrate
Catalyst
pH
Slope of the plot log (k/s ) vs. log (c/M)
2
ϩ
ApppA (1)
ApppA (1)
ApppA (1)
ApppA (1)
ApppA (1)
ApppA (1)
AppA (4)
Cu TerPy
6.0
8.0
6.0
8.0
7.5
8.0
7.5
1.5 ± 0.1
2.0 ± 0.1
0.9 ± 0.1
0.8 ± 0.1
1.9 ± 0.1
0.8 ± 0.1
1.3 ± 0.1
2
ϩ
Cu TerPy
2
ϩ
Cu BiPy
2
ϩ
Cu BiPy
2
ϩ
Cu [12]aneN3
2
ϩ
Cu [9]aneN3
2
ϩ
Cu [12]aneN3
a
b
pH adjusted with 0.1 M HEPES buffer, I = 0.1 M with NaNO . The pH values reported refer to 20 ЊC. Plots are shown in Fig. 2.
3
2
ϩ
Table 2 The effect of Mg ions (added as the nitrate salt) on the
2
ϩ
a
Cu TerPy (2.0 mM) promoted reaction of 1 at pH 7.5 and 60 ЊC
Ϫ5 Ϫ1
[
Mg(NO ) ]/M
k/10
s
3
2
None
No reaction in ten days
2.6 ± 0.1
2.04 ± 0.04
1.78 ± 0.05
1.28 ± 0.03
0
1
2
5
0
.5
.0
.0
.0
2
0.80 ± 0.02
a
Adjusted with 0.1 M HEPES buffer. pH refers to 20 ЊC. Ionic strength
was adjusted to 0.1 M with NaNO3.
7
Cleavage of the N -methylguanine base
2
ϩ
7
The effect of Cu complexes on the cleavage of the N -guanine
base appears to be very modest (Table 3). Only a 5 to 10-fold
rate-enhancement of the cleavage of the nucleoside 6 was
2ϩ
2ϩ
observed in 10 mM Cu TerPy and Cu BiPy solutions.
Ј-Phosphorylation that might allow complexation with the
Cu catalyst and hence an intramolecular attack of the
hydroxo ligand on C8, does not seem to enforce the catalysis.
Increasing metal ion concentration only slightly enhances
5
2ϩ
7
the reaction of N -methylguanosine 5Ј-monophosphate (7)
Fig. 2 The effect of catalyst concentration of the metal ion promoted
hydrolysis of the triphosphate bridge of 5Ј-cap model compounds.
(
Table 3). Rate-enhancement of the same magnitude was
7
2
ϩ
2ϩ
observed also with N -methylguanosine 5Ј-diphosphate (8).
The maximal rate enhancement by 10 mM metal ion complexes
is approximately 10-fold.
Legend: (ᮀ) ApppA (1)–Cu BiPy, pH 8.0; (᭿) ApppA (1)–Cu BiPy,
2
ϩ
pH 6.0; (᭺) ApppA (1)–Cu TerPy, pH 8.0; (᭜) ApppA (1)–
2
ϩ
2ϩ
Cu [9]aneN , pH 8.0; (ꢀ) ApppA (1)–Cu [12]aneN , pH 7.5; (᭢)
3
3
2ϩ
2ϩ
AppA (4)–Cu [12]aneN , pH 7.5; (᭹) ApppA (1)–Cu TerPy, pH 6.0.
pH adjusted with HEPES buffer. The values reported refer to 20 ЊC.
3
The products of these reactions were identified by comparing
7
them with products formed in well-known reactions of N -
methylguanine derivatives. Under alkaline conditions a hydrox-
ide ion attacks the base at C8, which results in a rapid cleavage
of the imidazole ring (Scheme 2). Mixing aliquots from
alkaline reaction solution and a neutral solution containing
Cu complex, and analysing the mixture by RP-HPLC shows
that the same products are formed in both reactions. In the
I adjusted to 0.1 M with NaNO . Slopes of the plots are collected
3
in Table 1.
21
the reaction. To study the role of individual metal ions in the
catalysis, experiments were carried out where a replacement of
a metal ion by a co-catalyst was attempted. When ethylamine
2ϩ
2ϩ
2ϩ
was added as a potential external nucleophile in Cu TerPy
promoted reaction at pH 8.0, it was observed that an increasing
concentration of ethylamine slightly decreased the rate of the
hydrolysis of 1. The rate constants obtained were (3.0 ± 0.1) ×
presence of Cu complexes, the formation of the initial ring-
opening product (I in Scheme 2) is apparently slow and a
significant amount of subsequent products is formed during
the reaction. In contrast to this, in the alkaline cleavage I is
the predominant product and its subsequent reactions are
slow in comparison to its formation. Under the experimental
Ϫ5
Ϫ5
Ϫ5 Ϫ1
1
0 , (2.4 ± 0.1) × 10 and (2.1 ± 0.1) × 10 s , at an ethyl-
amine concentration of 5, 25 and 50 mM, respectively. A rate
Ϫ5 Ϫ1
7
constant of (3.8 ± 0.1) × 10
s
was obtained in the absence of
conditions, N -methylguanosine is also depurinated (Scheme 4),
ethylamine.
which is shown by a corresponding comparison analysis using
acid catalysed depurination as a reference reaction. Metal ion
2ϩ
28
Table 2 shows the results of experiments where Mg ions
were added into reaction solutions to study whether the cata-
lytic efficiency of Cu complexes could also be enhanced by
complexes do not however seem to enhance the depurination,
but the rate appears to be comparable to that of the spon-
taneous reaction under neutral conditions. In contrast, a
2ϩ
2ϩ
another metal ion. The pK of the Mg -bound aqua ligand
a
27
2ϩ
2ϩ
7
is 12.8 (30 ЊC and I = 0.1 M) and consequently Mg cannot
provide hydroxide ion catalysts in the reaction, but it can
enhance the reaction by coordinating to the phosphate and
slight enhancement by Cu aqua ions of depurination of N -
methylguanosine was previously observed under slightly acidic
19
conditions.
2ϩ
hence making it more electrophilic. Mg was added at a
concentration equal to or less than that of Cu TerPy, and in
ten-fold excess. The results in Table 2 show that under these
2ϩ
Discussion
2ϩ
2ϩ
conditions Mg itself is inactive as a catalyst and the addition
Cu complexes efficiently promote the hydrolysis of the tri-
2ϩ
2ϩ
of Mg ion slightly retards the Cu TerPy promoted hydrolysis
rather than enhances it.
phosphate bridge in dialkyl oligophosphates. More than 20000-
fold rate-enhancement by these complexes was observed under
J. Chem. Soc., Perkin Trans. 2, 2002, 604–610
607

View Article Online
2
ϩ
2ϩ
7
Table 3 The effect of Cu TerPy and Cu BiPy on the cleavage of N -methylguanine base at 60 ЊC
a
Ϫ6 Ϫ1
Substrate
Catalyst
Conditions
k/10
s
7
N -mGuo (6)
None
2 mM Cu TerPy
10 mM Cu TerPy
pH 7.9, I = 0.15 M
pH 7.9, I = 0.15 M
pH 7.9, I = 0.15 M
5.8 ± 0.8
14 ± 2
32 ± 2
7
2ϩ
N -mGuo (6)
7
2ϩ
N -mGuo (6)
7
N -mGMP (7)
None
2 mM Cu TerPy
5 mM Cu TerPy
10 mM Cu TerPy
pH 7.0, I = 0.10 M
pH 7.0, I = 0.15 M
pH 7.0, I = 0.15 M
pH 7.0, I = 0.10 M
pH 7.0, I = 0.15 M
2.2 ± 0.1
6.3 ± 0.3
9.0 ± 0.4
10.1 ± 0.3
8.5 ± 0.7
7
2ϩ
N -mGMP (7)
7
2ϩ
N -mGMP (7)
7
2ϩ
N -mGMP (7)
7
2ϩ
N -mGMP (7)
10mM Cu BiPy
a
pH adjusted with 0.1 M HEPES buffer. The values reported refer to 20 ЊC. Ionic strength adjusted with NaNO3.
2ϩ
2ϩ
concentration of Cu TerPy and Cu [12]aneN complexes
3
ranged from 1.5 to 2. The mixed reaction order most probably
suggests that different metal ion–substrate complexes are
formed, and the most efficient catalysis is achieved with com-
plexes where more than one metal ion is involved in the rate-
limiting step of the reaction. The results in Table 1 also show
Ϫ1
that the slope of the plot log (k/s ) vs. log (c/M) for reaction
2ϩ
of 4 in the presence of Cu [12]aneN is clearly less than in
3
reaction of 1 under the same conditions suggesting that in the
former case the two-metal pathway is of less importance. This
may be due to steric factors: binding of two metal ion catalysts
to a diphosphate function may be less favourable than tri-
nuclear complexing of 1.
The dependence of the rate constants of the cleavage of 1
2ϩ
2ϩ
Scheme 4
promoted by Cu BiPy and Cu [9]aneN on the catalyst con-
3
Ϫ1
centration is different: The slopes of the plots log (k/s ) vs. log
2
ϩ
neutral conditions. Zn [12]aneN is clearly less active a catalyst
(c/M) suggest that only one metal ion complex participates
in the rate-limiting step of the reaction. The difference most
3
2ϩ
and Mg does not enhance the reaction to any significant
extent. The catalytic activity of bimetallic Cu bisdien (14),
2
ϩ
2
2ϩ
37
probably results from dimerisation: Both Cu BiPy and
2ϩ 38
2ϩ
29
with a hydroxo bridge between the two Cu ions, is slightly
higher than that of monometallic complexes. Further studies
on this complex were, however, prevented by the instability of
the complex. The bisdien ligand, which has previously been
shown to react efficiently with ATP through a nucleophilic
attack of an amino function to a phosphate group, did not
enhance the reaction of 1.
Cu [9]aneN3 are known to dimerise under the experimental
2ϩ
conditions, whereas no report on such behaviour for Cu TerPy
2ϩ
or Cu [12]aneN exists. The effect of dimerisation has clearly
3
20
been seen in kinetic studies on RNA hydrolysis. At a low con-
2
ϩ
2ϩ
centration Cu TerPy and Cu BiPy are approximately equally
36
good catalysts in hydrolysis of RNA model compounds.2 At a
ϩ
higher catalyst concentration, the catalytic activity of Cu BiPy
2ϩ
2ϩ
2ϩ
The rate-enhancement by Cu TerPy and Cu BiPy in this
work appears to be larger than in the cleavage of RNA model
compounds. However, the difference strongly depends on the
source of the data and the choice of the model compound.
A meaningful comparison is also made difficult by the fact
that the reaction conditions have different effects on different
reactions. The hydrolysis of RNA model compounds typically
shows a first-order dependence on the concentration of mono-
dramatically drops in comparison to that of Cu TerPy: 10 mM
Cu TerPy at pH 7.0 and 20 ЊC is a hundred times as good a
catalyst as Cu BiPy. The dimer of Cu [9]aneN , which at
2ϩ
2ϩ
30
2ϩ
3
pH 7.2 and 50 ЊC has been reported to be the predominant
species, is also inactive as a catalyst in phosphodiester
38
hydrolysis. In this case a half-order dependence of the rate
of the phosphodiester cleavage on the metal ion complex con-
centration has been observed.
2ϩ
30,31
meric Cu complexes
whereas a second-order dependence
Whether or not the dimer is catalytically active in the
hydrolysis of a triphosphate bridge cannot be concluded on
the basis of the present data, since the stability constants of
the dimers, and therefore the proportions of monomeric and
dimeric species under the experimental conditions, are not
accurately known. This being the case, the slopes of around
0.8 and 0.9 can be explained in three alternative ways. Firstly, if
the dimerisation is complete under the experimental conditions,
a first-order dependence of rate on the complex concentration
shows that the dimer is the catalytically active species. Secondly,
a first-order dependence could also be observed if the dimerisa-
tion is not quite complete under the experimental conditions
and if the dimer is not catalytically active, but two separate
complexes are required for the catalysis. In this case a second-
order dependence on the complex concentration may be com-
pensated by the effect of dimerisation, and over the narrow
concentration range utilised, an approximately first-order
dependence is observed. Alternatively, it can also be suggested
32
is seen in the present work. Results of Stern et al. show that
2ϩ
2ϩ
at pH 7.1, 0.16 mM Cu TerPy and Cu BiPy promote the
cleavage of oligo-A by factors of 40 and 15, respectively. Rate-
enhancement of the cleavage of a chimeric oligonucleotide by
mM Cu TerPy at pH 8.1 appears to be even more modest.
The cleavage of 2Ј,3Ј-cAMP, a simpler model compound, is
promoted to a more significant extent: using the rate con-
stants of the spontaneous cleavage reported by Ora et al.
and the temperature dependence of the rate constants reported
by Eftinkt and Biltonen, an approximately 1000-fold rate-
enhancement by 2 mM Cu TerPy and Cu BiPy at pH 7.5 can
be estimated from the rate constants reported by Liu et al.
Rate-enhancement of the cleavage of dinucleoside mono-
phosphate 3Ј,5Ј-ApA by 10 mM Cu TerPy at pH 8.2 appears
to be of the same order, whereas catalysis by Cu BiPy under
these conditions is prevented by dimerisation.
The results of the present work suggest that metal ion cata-
lysts play more than one role in the hydrolysis of a triphosphate
bridge, and more than one metal ion may be required for an
efficient hydrolysis. The reaction order of the cleavage 1 on the
2ϩ
33
1
34
35
2ϩ
2ϩ
36
2ϩ
2ϩ
30
2ϩ
2ϩ
that the structure of Cu BiPy and Cu [9]aneN allows a
3
single complex to provide the bifunctional catalysis and an
assistance by another complex is not required, as in the case of
6
08
J. Chem. Soc., Perkin Trans. 2, 2002, 604–610

View Article Online
2
ϩ
2ϩ
Cu TerPy and Cu [12]aneN . If this were the case, the pro-
tonating the phosphate group. While this kind of catalysis by
metal ions has actually never been observed in RNA cleavage,
3
portion of the catalytically inactive dimer would have to be low
under the experimental conditions, since only a slight deviation
if any) from a first-order dependence is observed.
2ϩ
41
Zn aqua ions have been shown to act as general acid-
catalysts that protonate the poor leaving group in the cleavage
of RNA model compounds. It is also possible that the other
apparent role of the metal ion complexes is not catalytic, but
(
The results in Fig. 1 show that the catalytic activity of the
complexes increases as the pH increases. The increase also
seems to level off at a higher pH. This suggests that a metal ion
co-ordinated hydroxide ion plays a role in the catalysis. Also the
2ϩ
structural. This is the case in Zn promoted cleavage of
short 3Ј-phosphorylated oligonucleotides that are folded in
a structure that is stabilised by a co-ordination of another
2ϩ
fact that Mg aqua ions that are not markedly deprotonated
27
2ϩ
42,43
under the experimental conditions, and are therefore inactive,
supports a mechanism that involves a metal ion bound hydroxo
ligand. No clear correlation can, however, be seen between
the catalytic activity (or the pH optimal for the catalysis) of a
Zn ion.
Similarly to the results of the present work, higher
than first-order dependence of the rate constants on the
43
2ϩ
catalyst concentration was observed, and addition of Mg
ion resulted in a slower cleavage.
given metal ion complex and the pK of the metal ion bound
The formation of a HEPES–AMP adduct, observed in the
presence of Cu TerPy complexes, is not inconsistent with the
a
2ϩ
2ϩ
water ligand. The pK of a Zn [12]aneN bound aqua ligand
a
3
3
9
(
7.3 reported by Zompa ) is clearly smaller in than those in
Cu BiPy and Cu TerPy (pK values of 7.8 and 8.2, respec-
tively, have been reported ), yet the complex is approximately
discussion above, even though the formation of such a product
implies that nucleophilic attack of a HEPES ion on the phos-
phate group may take place during the reaction. The adduct
2
ϩ
2ϩ
a
3
0
2
ϩ
2ϩ
1
00-fold poorer a catalyst than Cu BiPy, for example. Among
formation does not seem to compete with the Cu TerPy pro-
2ϩ
2ϩ
Cu complexes, Cu [9]aneN (pK value of 7.3 reported by
moted hydrolysis, but the proportion of the adduct remains
3
a
38
40
2ϩ
Deal and Burstyn and a value of 7.7 by Itoh et al. ) is more
the same when the Cu TerPy concentration is changed. It can
2ϩ
2ϩ
acidic than Cu TerPy or Cu BiPy. The lack of correlation has
tentatively be suggested that the by-product is formed through
18
2ϩ
been reported before and it has been suggested that it results
from the effect of the highly charged triphosphate moiety on
the deprotonation of an aqua ligand, even though no experi-
mental data on this has been reported.
an additional reaction pathway of 1 that binds two Cu TerPy
2ϩ
complexes. The adduct formation only takes place with Cu -
TerPy in HEPES buffers, which suggests that there may be a
2ϩ
specific interaction between the buffer constituent and Cu -
TerPy complex, and an entropically favoured intracomplex
nucleophilic attack by the HEPES ion on the phosphate group
takes place.
In the absence of quantitative data on the strength of the
metal ion–phosphate complexes, and on the acidity of the metal
bound water ligands in these complexes, no firm mechanistic
conclusions can be drawn on the basis of comparisons between
different metal ion complexes. The results of the present work
are, however, consistent with previous suggestions that metal
ions assist the hydrolysis as bifunctional catalysts and a nucleo-
philic attack by a metal ion-bound hydroxo ligand is assisted
by co-ordination of another metal ion. This mechanism is
Conclusions
2ϩ
Cu complexes significantly enhance the hydrolysis of the tri-
phosphate bridge in an mRNA 5Ј-cap model. The catalysis may
involve a single catalyst species, two metal ion complexes as
a dimer or two separate metal ion complexes as in the case of
2ϩ
very similar to that one often suggested for the Cu complex-
promoted cleavage of RNA model compounds, where a single
metal ion may provide both the electrophilic assistance and the
attacking hydroxide nucleophile. It may, however, be suggested
that in the case of the cap-model compounds studied in the
present work, the catalysis is more complex than that. The cata-
2ϩ
2ϩ
Cu TerPy or Cu [12]aneN . Independent of the composition
3
of the catalyst, bifunctional catalysis by phosphate-bound
metal ion complexes is involved. Active catalysts contain acidic
aqua ligands, and when deprotonated, a hydroxo ligand most
probably acts as a nucleophile that attacks the phosphate.
An acidic aqua ligand may play another role in the catalysis,
possibly as a general acid catalyst. Alternatively, the other role
of a metal ion catalyst may be structural. In any case more data
is still required to draw any final conclusions on the details of
2ϩ
lysis by Cu complexes appears to be a very specific process,
2ϩ
and external nucleophiles or electrophiles cannot replace Cu
complexes as catalysts. Addition of an amine as external nucleo-
2ϩ
phile did not enhance the catalytic activity of Cu TerPy, even
though a nucleophilic attack by an amino group on a phosphate
2ϩ
the mechanism. In contrast to the triphosphate hydrolysis, Cu
complexes do not enhance the cleavage of N -methylguanine.
14,26
7
function has previously been reported.
It should be borne
in mind, though, that the slight effect observed may also be a
sum of two opposite effects that compensate each other, and
that while the amine is actually a nucleophile, it also prevents
the co-ordination of the substrate by strongly binding the metal
Experimental
Materials
2ϩ
ion catalysts. The fact that Mg ion could not assist the
2ϩ
Diadenosine 5Ј,5Ј-triphosphate (1), diadenosine 5Ј,5Ј-
diphosphate (4), diadenosine 5Ј,5Ј-tetraphosphate (5), adeno-
sine 5Ј-diphosphate (3), adenosine 5Ј-monophosphate (2) and
adenosine 5Ј-triphosphate (15) were products of Sigma and
they were used as received. N -methylguanosine (6) was pre-
pared from guanosine by methylating with methyl iodide as
Cu TerPy promoted reaction by binding to the phosphate
lends, however, additional support to the suggestion that the
2ϩ
catalysis specifically requires two Cu TerPy complexes, which
cannot be replaced.
7
It would seem, on the basis of the discussion above, that the
different roles of a metal ion-catalyst cannot be separated, but
the reaction rather requires two phosphate-bound metal ions
that have acidic water ligands. While a metal ion bound hydroxo
ligand acts as a nucleophile, it could be suggested that another
acidic aqua ligand also has a role in the catalysis. Consistent
44
described before. Dimethyl sulfate was used as methylating
7
agent in the synthesis of N -methylguanosine 5Ј-mono-
45
7
phosphate (7) and N -methylguanosine 5Ј-monophosphate
46
1
31
(
8). The compounds were characterised by H and P NMR
2ϩ
2ϩ
spectroscopy (7 and 8) and EI-MS. Buffer constituents, metal
salts and bipyridine, terpyridine, 1,4,7-triazacyclononane and
1,5,9-triazacyclododecane ligands were of reagent grade. The
with this suggestion, it has been shown that Zn , but not Mg ,
enhances the lanthanum complex promoted hydrolysis of a
20
cap-model compound. An unexpectedly high catalytic activity
25
2ϩ
synthesis of the bisdien ligand has been reported elsewhere.
of a bimetallic Zn complex with no hydroxo ligands under
20
experimental conditions, has also been reported. The most
logical role of an acidic water ligand is that of a general
acid catalyst, and it could be possible that an appropriately
positioned metal ion enhances the nucleophilic attack by pro-
Preparation of reaction solutions
Metal ion complexes were prepared by mixing the metal ion
(as nitrate) and 1.1 equivalents of the ligand. The pH of the
J. Chem. Soc., Perkin Trans. 2, 2002, 604–610
609

View Article Online
solution was adjusted with HEPES buffer [N-(2-hydroxyethyl)-
8 E. Darzynkiewicz, J. Antosiewicz, I. Ekiel, M. A. Morgan,
S. M. Tahara and A. J. Shatkin, J. Mol. Biol., 1981, 153, 451–458.
piperazine-NЈ-ethanesulfonic acid; pK 7.56], with MES buffer
a
9
B. F. Baker, S. S. Lot, J. Kringel, J. S. Cheng-Flournoy, P. Villiet,
H. M. Sasmor, A. M. Siwkowski, L. L. Chappell and J. R. Morrow,
Nucleic Acids Res., 1999, 27, 1547–1551.
(
2-morpholinoethanesulfonic acid; pK 6.15) or with CHES
a
buffer [CHES = 2-(N-cyclohexylamino)ethanesulfonic acid;
pK 9.3]. The pH of the reaction solutions was checked with
a
10 R. M. Milburn, M. Gautam-Basak, R. Tribolet and H. Sigel, J. Am.
Chem. Soc., 1985, 107, 3315–3321.
11 F. Tafesse, S. S. Massoud and R. M. Milburn, Inorg. Chem., 1985,
a pH-meter either at 25 ЊC or at 60 ЊC. Ionic strength was
adjusted with NaNO . Sterilised water was used to prepare the
3
2
4, 2591–2593.
2 S. H. McClauherty and C. M. Grisham, Inorg. Chem., 1982, 21,
133–4138.
3 R. J. Geue, A. M. Sargeson and R. Wijesekara, Aust. J. Chem., 1993,
6, 1021–1040.
4 D. A. Nation, Q. Lu and A. E. Martell, Inorg. Chim. Acta, 1997, 263,
09–217.
5 B. F. Baker, J. Am. Chem. Soc., 1993, 115, 3378–3379.
reaction solutions, and they were handled with sterilised
equipment.
1
1
1
1
4
Kinetic measurements
4
Reactions were carried out in Eppendorf tubes immersed in a
water bath, the temperature of which was thermostated at
2
6
0 ЊC. The reactions were initiated by adding substrate stock
16 B. F. Baker, H. Khalili, N. Wei and J. R. Morrow, J. Am. Chem. Soc.,
1997, 119, 8749–875.
solution (a few microlitres) to give a concentration of 0.1 mM.
The total reaction volume was 1.5 ml. Aliquots (12) of 100 µl
were withdrawn at appropriate intervals to cover approximately
two half-lives of the reaction. The aliquots were immediately
cooled down on an ice-bath, and stored in the freezer until
analysed.
1
7 K. P. McCue, D. A. Voss, Jr., C. Marks and J. R. Morrow, J. Chem.
Soc., Dalton Trans., 1998, 2961–2963.
1
1
8 K. P. McCue and J. R. Morrow, Inorg, Chem., 1999, 38, 6136–6142.
9 Z. Wieczorek, E. Darzynkiewicz, S. Kuusela and H. Lönnberg,
Nucleosides Nucleotides, 1999, 18, 11–21.
20 D. M. Epstein, L. L. Chappel, H. Khalili, R. M. Supkowski,
W. D. Horrocks and J. R. Morrow, Inorg. Chem., 2000, 39,
2
130–2134.
Analysis of aliquots
2
1 E. Darzynkiewicz, J. Stepinski, S. M. Tahara, R. Stolarski, I. Ekiel,
D. Haber, K. Neuvonen, P. Lehikoinen, I. Labadi and H. Lönnberg,
Nucleosides Nucleotides, 1990, 9, 599–6118.
2 E. Darzynkiewicz, L. Labadi, D. Haber, K. Burger and
H. Lönnberg, Acta Chem. Scand., Ser. B, 1988, 86–92.
3 H. Sigel and P. E. Amsler, J. Am. Chem. Soc., 1976, 98, 7390–7400.
4 R. Cini and C. Pifferi, J. Chem. Soc., Dalton Trans., 1999,
699–710.
Aliquots from reactions of 1, 4 and 5 were analysed either
by capillary zone electrophoresis (CZE) or HPLC. Capillary
electrophoretic analysis was carried out in a fused silica
capillary (50 µm id, 77 cm) with 0.1 M boric acid, pH 8.0 as the
run buffer. A voltage of 30 kV was applied. The compounds
were detected at 260 nm (diode array detector). Alternatively,
an HPLC analysis was utilised, particularly in case of 4. The
column used was Thermo Quest, Hypersil HS RP-18 (250 ×
2
2
2
25 S. Mikkola, Q. Wang, Z. Jori, H. Helkearo and H. Lönnberg, Acta
Chem. Scand., 1999, 53, 453–456.
2
6 M. W. Hosseini, J.-M. Lehn and M. P. Mertes, Helv. Chim. Acta,
983, 66, 2454–2466.
4
.6 mm, 5 µm particle size), and the eluent a 0.1 M formic acid
1
buffer, pH 3.0, containing 0.1 M tetrabutylammonium bromide.
UV-analysis at 260 nm was utilised. Aliquots from reactions of
2
7 Ionisation Constants of Inorganic Acids and Bases in Aqueous
Solution, Second Edition, D. D. Perrin (Ed.), Pergamon Press,
Oxford, England, 1982, p. 67.
6
and 7 were analysed by RP-HPLC using the Thermo Quest,
Hypersil HS column mentioned above. The eluent was a
mixture of 0.1 M triethyl amine–phosphoric acid buffer and
methanol. In case of 6, the pH of the buffer was 6.8 and the
methanol content was 0.5 %. With 7, the pH was 3.0 and
methanol content 1.0 %. Compounds were detected at wave-
length of 274 nm (6) or 256 nm (7). Aliquots from reactions of
28 J. A. Zoltewicz, D. F. Clarc, T. W. Sharpless and G. Grahe, J. Am.
Chem. Soc., 1973, 92, 1741–1750.
2
9 R. J. Motekaitis, A. E. Martell, J.-P. Lecomte and J.-M. Lehn, Inorg.
Chem., 1983, 22, 609–614.
3
0 B. Linkletter and J. Chin, Angew. Chem., Int. Ed. Engl., 1995, 34,
4
72–474.
31 J. K. Bashkin and L. A. Jenkins, J. Chem. Soc., Dalton Trans., 1993,
3631–3632.
8
were analysed by CZE. The run buffer was 25 mM boric acid,
3
2 M. K. Stern, J. K. Bashkin and E. D. Sall, J. Am. Chem. Soc., 1990,
12, 5357–5359.
3 L. A. Jenkins, L. A. Bashkin and M. E. Autry, J. Am. Chem. Soc.,
996, 118, 6822–6825.
34 M. Ora, M. Oivanen and H. Lönnberg, J. Org. Chem., 1996, 61,
951–3955.
5 M. R. Eftinkt and R. L. Biltonen, Biochemistry, 1983, 22, 5134–
140.
pH 9.0, and a voltage of 30 kV was applied. The capillary
dimensions were as described above.
1
3
1
Calculation of rate constants
3
The rate constants of the disappearance of the starting material
3
(
either decrease of mole fraction or decrease of the signal area)
5
were calculated by using the integrated first-order rate-law. In
CZE analysis the signal areas were divided by the respective
migration times to normalise the areas.
36 S. Liu, Z. Luo and A. D. Hamilton, Angew. Chem., Int. Ed., 1997,
36, 2678–2680.
3
7 J. R. Morrow and W. C. Trogler, Inorg. Chem., 1988, 27,
387–3394.
3
3
3
4
8 K. A. Deal and J. N. Burstyn, Inorg. Chem., 1996, 35, 2792–2798.
9 L. J. Zompa, Inorg. Chem., 1978, 17, 2531–2536.
0 T. Itoh, H. Hisada, Y. Usui and Y. Fujii, Inorg. Chim. Acta, 1998,
283, 51–60.
References
1
2
3
For a recent review, see B. N. Trawick, A. Daniher and J. K. Bashkin,
Chem. Rev., 1998, 98, 939–960.
41 S. Mikkola, E. Stenman, K. Nurmi, E. Yousefi-Salakdeh,
R. Strömberg and H. Lönnberg, J. Chem. Soc., Perkin Trans. 2,
1999, 1619–1625.
K. G. K. Murthy, P. Park and J. L. Manley, Nucleic Acids Res., 1991,
1
9, 2685–2692.
M. M. Konarska, R. A. Padgett and P. A. Sharp, Cell, 1984, 38,
31–736.
42 S. Kuusela and H. Lönnberg, Nucleosides Nucleotides, 1998, 17,
7
2417–2427.
4
5
6
J. Hamm and I. W. Mattaj, Cell, 1990, 63, 109–118.
43 S. Kuusela, A. Guzaev and H. Lönnberg, J. Chem. Soc., Perkin
Trans. 2, 1996, 1895–1899.
44 J. W. Jones and R. K. Robins, J. Am. Chem. Soc., 1963, 95, 193–201.
45 K. Kamiichi, M. Doi, M. Nabae, T. Ishida and M. Inoue, J. Chem.
Soc., Perkin Trans. 2, 1987, 1739–1745.
A. J. Shatkin, Cell, 1985, 40, 223–224.
R. E. Rhoads, in Progress in Molecular and Subcellular Biology Vol.
9
, H. F. Hahn, D. J. Kopecho and W. E. Muller (Eds.), Springer,
Berlin, 1985, pp. 104–155.
7
N. Sonenberg, Prog. Nucleic Acids Res. Mol. Biol., 1988, 35,
46 S. Hendler, E. Fürer and P. R. Srinivasan, Biochemistry, 1970, 9,
1
73–207.
4141–4153.
6
10
J. Chem. Soc., Perkin Trans. 2, 2002, 604–610