Trinuclear copper(II) complex showing high selectivity for the hydrolysis of 2′-5′ over 3′-5′ for UpU and 3′-5′ over 2′-5′ for ApA ribonucleotides

Article abstract of DOI:10.1021/ja020877t

The cooperative action of multiple Cu(II) nuclear centers is shown to be effective and selective in the hydrolysis of 2′-5′ and 3′-5′ ribonucleotides. Reported herein is the specific catalysis by two trinuclear Cu(II) complexes of L3A and L3B. Pseudo firs

Full text of DOI:10.1021/ja020877t

Published on Web 10/29/2002
Trinuclear Copper(II) Complex Showing High Selectivity for
the Hydrolysis of 2′-5′ over 3′-5′ for UpU and
3′-5′ over 2′-5′ for ApA Ribonucleotides
Makoto Komiyama,*,† Shinichiro Kina,† Kazunari Matsumura,† Jun Sumaoka,†
Suzanne Tobey,‡ Vincent M. Lynch,‡ and Eric Anslyn*,‡
Contribution from the Research Center for AdVanced Science and Technology,
The UniVersity of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo, 153-8904 Japan, and
Department of Chemistry and Biochemistry, UniVersity of Texas, Austin, Texas, 78712
Received June 24, 2002
Abstract: The cooperative action of multiple Cu(II) nuclear centers is shown to be effective and selective
in the hydrolysis of 2′-5′ and 3′-5′ ribonucleotides. Reported herein is the specific catalysis by two trinuclear
Cu(II) complexes of L3A and L3B. Pseudo first-order kinetic studies reveal that the L3A trinuclear Cu(II)
complex effects hydrolysis of Up(2′-5′)U with a rate constant of 28 × 10-4 min-1 and Up(3′-5′)U with a
rate constant of 0.5 × 10-4 min-1. The hydrolyses of Ap(3′-5′)A and Ap(2′-5′)A proceed with rate constants
of 24 × 10-4 min-1 and 0.5 × 10-4 min-1 respectively. The L3A trinuclear Cu(II) complex demonstrates
high specificity for Up(2′-5′)U and Ap(3′-5′)A. Similar studies with the more rigid L3B trinuclear Cu(II)
complex shows no selectivity and yields lower rate constants for hydrolysis. The selectivity observed with
the L3A ligand is attributed to the geometry of the ligand-bound diribonucleotide which ultimately dictates
the proximity of the attacking hydroxyl and the phosphoester to a Cu(II) center for activation and subsequent
hydrolysis.
Introduction
complexes showing sufficiently high substrate-specificity in
RNA hydrolysis have been scarce, with only one reported base-
Interest in the hydrolysis of RNA has been rapidly increasing,1
and notable catalysis using lanthanide ions and their complexes
has been documented.2 Furthermore, a number of dinuclear3,4
and trinuclear metal complexes have been reported, thereby
demonstrating that the hydrolysis of RNA can be promoted by
cooperation between multiple metallic centers.6 However, metal
selective cleavage of 3′-5′ diribonucleotides by Hamilton.7 Such
complexes likely require many sites for substrate recognition.
Herein, we report that the trinuclear Cu(II) complex of
N,N,N′,N′,N′′,N′′-hexa[(2-pyridyl)methyl]-1,3,5-tris(aminometh-
yl)benzene (L3A) shows remarkable substrate-specificity in the
hydrolysis of Up(2′-5′)U compared to Up(3′-5′)U and in the
hydrolysis of Ap(3′-5′)A compared to Ap(2′-5′)A.8 The
catalytic activity is not only strongly dependent on the kind of
phosphodiester linkage, but also on the sequence. The catalysis
by trinuclear Cu(II) complex of N,N,N′,N′,N′′,N′′-hexa[(2-
pyridyl)methyl]-1,3,5-tris(aminomethyl)-2,4,6-triethylbenzene
(L3B), which has restricted molecular flexibility compared to
that of the L3A complex, is also reported. The origin of
substrate-specificity is discussed in terms of kinetic and
spectroscopic results.
5
* To whom correspondence should be addressed. E-mail: komiyama@
mkoni.rcast.v-tokyo.ac.jp.
† University of Tokyo.
‡ University of Texas.
(1) (a) Komiyama, M.; Sumaoka, J.; Kuzuya, A.; Yamamoto, Y. Methods
Enzymol. 2001, 341, 455; (b) Komiyama, M.; Sumaoka, J. Curr. Opin.
Chem. Biol. 1998, 2, 751; (c) Oivanen, M.; Kuusela, S.; Lo¨nnberg, H. Chem.
ReV. 1998, 98, 961; (d) Trawick, B. N.; Daniher, A. T.; Bashkin, J. K.
Chem. ReV. 1998, 98, 939; (e) Pratviel, G.; Bernadou, J.; Meunier, B.
AdVances in Inorganic Chemistry; Sykes, A. G., Ed.; Academic Press: San
Diego, 1998, Vol. 45, p 251; (e) Williams, N. H.; Takasaki, B.; Wall, M.;
Chin, J. Acc. Chem. Res. 1999, 32, 485.
(2) (a) Komiyama, M.; Matsumura, K.; Matsumoto, Y. Chem. Commun. 1992,
640; (b) Morrow, J. R.; Buttrey, L. A.; Shelton, V. M.; Berback, K. A. J.
Am. Chem. Soc. 1992, 114, 1903.
(3) (a) Young, M. J.; Chin, J. J. Am. Chem. Soc. 1995, 117, 10 577; (b) Yashiro,
M.; Ishikubo, A.; Komiyama, M. Chem. Commun. 1995, 1793; (c) Yashiro,
M.; Hattori, M.; Ishikubo, A.; Komiyama, M. Nucleic Acids, Symp. Ser.
1996, 35, 143.
Results and Discussion
Synthesis of the Ligands L3A and L3B. The structures of
the ligands used in this study are presented in Scheme 1.
(4) (a) Koike, T.; Inoue, M.; Kimura, E.; Shiro, M. J. Am. Chem. Soc., 1996,
118, 3091; (b) Ragunathan, K. G.; Schneider, H. J. Angew. Chem., Int.
Ed. Engl. 1996, 35, 1219; (c) Molenveld, P.; Kapsabelis, S.; Engbersen, F.
J.; Reinhoudt, D. N. J. Am. Chem. Soc. 1997, 119, 2948; (d) the papers
cited in ref 1.
(5) (a) Yashiro, M.; Ishikubo, A.; Komiyama, M. Chem. Commun. 1997, 83;
(b) Molenveld, P.; Engbersen, F. J.; Reinhoudt, D. N. Angew. Chem., Int.
Ed. Engl. 1999, 38, 3189; (c) Fritsky, I. O.; Ott, R.; Pritzkow, H.; Kramer.
R. Chem. Eur. J. 2001, 1221.
(7) Liu, S.; Hamilton, A. Chem. Commun. 1999, 587.
(8) Hydrolysis of phosphoesters by Cu(II) ion and its complexes were
reported: (a) Morrow, J. R.; Trogler, W. C. Inorg. Chem. 1988, 27, 3387;
(b) Stern, M. K.; Bashkin, J. K.; Sall, E. D. J. Am. Chem. Soc. 1990, 112,
5357; (c) Modak, A. S.; Gard, J. K.; Merriman, M. C.; Winkeler, K. A.;
Bashkin, J. K.; Stern, M. K, J. Am. Chem. Soc. 1991, 113, 283; (d) Bashkin,
J. K.; Jenkins, L. A. J. Chem. Soc., Dalton Trans. 1993, 3631; (e) Burstyn,
J. N.; Deal, K. A. Inorg. Chem. 1993, 32, 3585; (f) Wall, M.; Hynes, R.
C.; Chin, J. Angew. Chem., Int. Ed. Eng. 1993, 32, 1633; (g) Deal, K. A.;
Hengge, A. C.; Burstyn, J. N.; J. Am. Chem. Soc. 1996, 118, 1713; (h) ref
3a.
(6) Many enzymes involved in the hydrolysis of phosphoesters possess two or
more metal ions in their active sites: Stra¨ter, N.; Lipscomb, W. N.;
Klabunde, T.; Krebs, B. Angew, Chem. Int. Ed. Engl. 1996, 35, 2024.
9
10.1021/ja020877t CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 13731-13736
13731

A R T I C L E S
Komiyama et al.
Scheme 1. Ligands Used in the Present Study
Figure 1. View of L3B showing the atom labeling scheme. Displacement
ellipsoids are scaled to the 30% probability level. Hydrogen atoms have
been removed for clarity.
Table 1. Pseudo-first-order Rate Constants (in 10-4 min-1) for the
Hydrolysis of (2′-5′)- and (3′-5′)-diribonucleotides by the
Trinuclear Cu(II)/L3A Complex at pH 7.0 and 50°Ca,b
linkage
Trinuclear ligand L3A was prepared from 1,3,5-tris(bromo-
methyl)benzene and 2,2′-dipicolylamine. Tris(bromomethyl)-
benzene was prepared by irradiating UV light on a mixture of
1,3,5-trimethylbenzene, N-bromosuccinimide, and benzoyl per-
oxide in carbon tetrachloride.9 Subsequently, tris(bromomethyl)-
benzene was treated with 2,2′-dipicolylamine, and N,N-diiso-
propylamine in acetonitrile. This yielded the desired L3A ligand.
Dinuclear ligand L2 was synthesized in a fashion similar to
L3A, using 1,3-di(bromomethyl)benzene and 2,2′-dipicolyl-
amine. The L3B was prepared by combining 1,3,5-tris(chlo-
romethyl)-2,4,6-triethylbenzene,10 2,2′-dipicolylamine and po-
tassium carbonate in dry acetonitrile.
Formation of Trinuclear Cu(II) Complexes of L3A and
L3B. When L3A was added to an aqueous solution of Cu(II)
at pH 7.0, a new absorption band appeared between 550 and
800 nm. The absorption maximum λmax was 662 nm. As the
concentration of L3A was increased, ([Cu(II)]0 was kept
constant at 1.2 mM), the absorbance for this band linearly
increased until [Cu(II)]0/[L3A]0 ) 3, at which point a plateau
was attained. A clear break was observed at the mole ratio 3.
Virtually the same results were obtained for the Cu(II)/L3B
system (λmax ) 674 nm). The trinuclear Cu(II) complexes, of
both the L3A and the L3B ligands, were formed almost
quantitatively in aqueous solutions. According to a potentio-
metric titration, the pKa values of the metal-bound water ligands
in the Cu(II)/L3A complex are 6.9, 7.4, and 7.8, respectively.
These values are close to each other, and comparable with the
corresponding values for analogous Cu(II) complexes.11 Ap-
parently all of the three Cu(II) ions in this complex are in similar
environments, and the metals act independently in the pH
titrations.
ribonucleotide
2′−5′
3′−5′
specificity c
UpU
GpG
ApA
CpC
28 (0.6)
0.3 (0.2)
0.5 (2.0)
0.5 (0.7)
0.3 (0.4)
24 (6.0)
7.0 (2.0)
56
1
0.021
a [Cu(II)]0 ) 6 and [L3A]0 ) 2 mM. bThe values for the dinuclear Cu(II)/
L2 complex are presented in parentheses ([Cu(II)]0 ) 6 and [L2]0 ) 3
mM). cThe ratio of rate constant for the hydrolysis of (2′-5′)-diribonucle-
otide to the corresponding value for the (3′-5′)-isomer.
lography on the trinuclear Cu(II)/L3B complex is presented in
Figure 1. Each of the three 2,2′-dipicolylamine moieties of L3B
binds one Cu(II) ion. The angles for N(pyridine)-Cu-
N(aliphatic) and N(pyridine)-Cu-N(pyridine) are 81.9-83.0°
and 163.5-166.0°, respectively. The Cu-N(aliphatic) distance
is 2.2026-2.034 Å, and the Cu-N(pyridine) distance is 1.967-
1.982 Å. The structure reveals that each of the three 3,3′-
dipicolylamine moieties complexes to one Cu(II) ion and they
reside on one face of the triethylbenzene scaffold. However,
each of the Cu(II) binding moieties points away from the interior
of the cavity. This would indicate that there is little flexibility
available in the receptor for the binding of ribonucleotide
substrates.
Hydrolysis of Diribonucleotides by Trinuclear Cu(II)/L3A
Complex. Table 1 lists the results of diribonucleotide hydrolysis
by the trinuclear Cu(II)/L3A complex at pH 7.0 (50 mM HEPES
buffer) and 50 °C. This complex rapidly hydrolyzes Up(2′-
5′)U to a 1:1 mixture of uridine and uridine monophosphate.
Alternatively, Up(3′-5′)U is hydrolyzed at a significantly slower
rate by this trinuclear Cu(II) complex.13 The pseudo-first-order
rate constant of the hydrolysis of the 2′-5′ isomer is 56-fold
larger than that of the 3′-5′ isomer. It is interesting to note
that the substrate specificity is reversed in the case of the ApA
The X-ray structure of a trinuclear Cu(I) complex of L3B
has previously been reported.12 The result of the X-ray crystal-
(9) Vo¨gtle, F.; Zuber, M.; Lichtenthaler, R.G. Chem. Ber. 1973, 106, 717.
(10) Fuson, R. C.; House, H. O.; Melby, L. R. J. Am. Chem. Soc. 1953, 75,
5954; Kilway, K. V.; Siegel, J. S. Tetrahedron 2001, 57, 3615; Hanes, R.;
Anslyn, E.; Kilway, K.; Siegel, J. Submitted to Organic Synthesis.
(11) The pKa of the water bound to the Cu(II) ion in the bipyridine and
terpyridine complexes are 7.0 and 8.08, respectively (ref 8a,d). The water
bound to free Cu(II) ion has the pKa 5.4-6.8: Burgess, J. Metal Ions in
Solution; Horwood, Chichester, 1978.
(12) Walsdorf, C.; Park, S.; Kim, J.; Heo, J.; Park, K.; Oh, J.; Kim, K. J. Chem.
Soc., Dalton Trans. 1999, 923.
(13) The intrinsic activities of the 2′-5′ diribonucleotides and the 3′-5′ isomers
are comparable with each other. Consistently, Up(2′-5′)U and Up(3′-
5′)U are hydrolyzed at comparable rates under alkaline conditions and also
in the presence of Mg(II), Mn(II), Co(II), Ni(II), and Zn(II): Kuusela, S.;
Lo¨nnberg, H. J. Phys. Org. Chem. 1993, 6, 347.
9
13732 J. AM. CHEM. SOC. VOL. 124, NO. 46, 2002

Trinuclear Copper(II) Complex
A R T I C L E S
hydrolysis. Ap(3′-5′)A is hydrolyzed with a rate constant that
is 48-fold larger than that of the Ap(2′-5′)A substrate.
Therefore, the catalysis is selective with respect to both the type
of phosphodiester linkage and the base sequence. Further, Gp-
(3′-5′)G and Gp(2′-5′)G are not hydrolyzed to a significant
extent, and the 2′-5′/3′-5′ specificity is negligible, yet Cp-
(3′-5′)C is hydrolyzed at a reasonable rate. Each of the reactions
described proceeds via an intramolecular attack by the 3′-OH
(or 2′-OH) of the ribose. This is further confirmed by the fact
that the 2′-deoxyribonucleotides are not hydrolyzed to a
measurable extent by this complex. However, the intermediate
2′,3′-cyclic monophosphates do not accumulate, but are rapidly
hydrolyzed to the final product ribonucleoside monophosphate.
For the purposes of comparison, the results with the dinuclear
Cu(II) complex of N,N,N′,N′-tetra[(2-pyridyl)methyl]-1,3-
bisaminoethylbenzene (L2) are also reported in Table 1. In the
hydrolysis of Up(2′-5′)U, the trinuclear L3A complex is 47-
fold more active than the dinuclear L2 complex. For Ap(3′-
5′)A hydrolysis, the L3A complex is 4-fold more active than
that of the L2 complex. Both the mononuclear Cu(II) complex
of N,N-bis[(2-pyridyl)methyl]-(aminoethyl)benzene (L1) and
Cu(II) ion are still less active than the dinuclear L2 complex.14,15
In the case of the other diribonucleotides, the differences in the
activity between the trinuclear complex and the dinuclear
complex are smaller. This indicates that the presence of the third
Cu(II) ion in the trinuclear Cu(II)/L3A complex significantly
accelerates the hydrolysis of Up(2′-5′)U and Ap(3′-5′)A.
Catalysis by Trinuclear Cu(II)/L3B Complex. In contrast
to the efficient and substrate-specific catalyses by trinuclear Cu-
(II)/L3A complex, the trinuclear Cu(II)/L3B complex demon-
strates lower catalytic activity in the hydrolysis of the diribo-
nucleotide substrates investigated. The pseudo-first-order rate
constants are 0.1 × 10-4 min-1 for Up(2′-5′)U, 0.1 × 10-4
min-1 for Up(3′-5′)U, 0.02 × 10-4 min-1 for Ap(2′-5′)A, 0.1
× 10-4 min-1 for Ap(3′-5′)A, and 0.1 × 10-4 min-1 for Cp-
(3′-5′)C. As observed in the catalysis by the L3A complex,
Ap(3′-5′)A is more susceptible to hydrolysis than is Ap(2′-
5′)A. However, no specificity was observed between Up(2′-
5′)U and Up(3′-5′)U. The ligand L3B is similar in structure to
L3A, the only difference being the restricted motions of the
dipicolylamine moieties that result from the presence of the
alternating ethyl groups on the benzene scaffold. Apparently,
the molecular flexibility inherent in the L3A ligand is essential
for the trinuclear Cu(II)/L3A complex to effect efficient and
substrate-specific catalysis.
Figure 2. Plots of the hydrolysis rates of Up(2′-5′)U (open circles) and
Ap(3′-5′)A (closed circles) vs the [Cu(II)]0/[L3A]0 ratio at 50°C and pH
7.0: [Cu(II)]0 is kept constant at 0.6 mM for Up(2′-5′)U and 6 mM for
Ap(3′-5′)A.
Figure 3. Dependencies of the hydrolysis rates of Up(2′-5′)U (open
circles) and Ap(3′-5′)A (closed circles) on the concentration of the
trinuclear Cu(II)/L3A complex ([Cu(II)]0: [L3A]0 ) 3:1) at 50°C and pH
7.0: The solid and the dotted lines are theoretical ones calculated by using
the parameters in Table 2.
Table 2. Kinetic parameters for the trinuclear Cu(II)/L3A
complex-induced hydrolysis of Up(2′-5′)U and Ap(3′-5′)A at pH
7.0 and 50°Ca
substrate
kcat (10-4 min-1
)
Km (mM)
Up(2′-5′)U
Ap(3′-5′)A
27
210
0.21
17
a Experimental error is estimated to be around 20%.
Kinetic Studies on the Catalysis by Trinuclear Cu(II)/L3A
Complex. When the [Cu(II)] is kept constant and the [Cu(II)]0/
[L3A]0 ratio is varied, the rate of Up(2′-5′)U hydrolysis results
in a maximum value at a ratio of 3 (open circles in Figure 2).
A similar result is obtained for Ap(3′-5′)A hydrolysis (closed
circles). At [Cu(II)]0/[L3A]0 ratios greater than three, the
trinuclear Cu(II) complex is quantitatively formed in proportion
to the added ligand L3A, and the hydrolysis rate monotonically
increases. On further addition of L3A, the fractions of mono-
nuclear and dinuclear Cu(II) complexes gradually increase,
diminishing the catalytic activity. These results confirm that the
catalytic activity is primarily due to the Cu(II)/L3A complex.
The contributions to the catalysis by the dinuclear Cu(II)
complex, the mononuclear Cu(II) complex, and the Cu(II) ion
are marginal.
As shown in Figure 3, the rate of hydrolysis increases with
increasing concentration of the Cu(II)/L3A complex with
eventual saturation.16 The values of kcat and Km, determined from
Lineweaver-Burk plots, are presented in Table 2. Significantly,
the Km value for Up(2′-5′)U hydrolysis is much smaller than
the corresponding value for Ap(3′-5′)A hydrolysis, but the kcat
is significantly smaller. Thus, the efficient hydrolysis of Up-
(2′-5′)U by the trinuclear Cu(II)/L3A is primarily attributed
(14) The rate constants for the hydrolysis of Up(3′-5′)A and Ap(3′-5′)U by
the Cu(II)/L1 complex are 3.0 × 10-4 and 1.0 × 10-4 min-1, respectively.
(15) Hydrolysis of bis(2,4-dinitrophenyl)phosphate by the Cu(II) complex of
bis(2-pyridyl)methylamine was reported: Young, M. J.; Wahnon, D.;
Hynes, R. C.; Chin, J. J. Am. Chem. Soc. 1995, 117, 9441.
(16) Here, the [L3A]0/[Cu(II)]0 ratio is kept constant at 1/3. Note that the complex
formation is almost quantitative, as described in the text.
9
J. AM. CHEM. SOC. VOL. 124, NO. 46, 2002 13733

A R T I C L E S
Komiyama et al.
bonds, these structures are rather rigid because of their ring-
forms. The L3A ligand is flexible enough to allow the formation
of this substrate-catalyst adduct and the conformational change
required for the catalysis.21 By contrast, the rigidity of the L3B
ligand resulting from the alternating ethyl groups likely gives a
less tightly held substrate-catalyst adduct and reduces the
efficiency of hydrolysis for both Up(2′-5′)U and Up(3′-5′)U.
The L3B complex is a poor catalyst for promoting the hydrolysis
of all other diribonucleotides.
The dinuclear L2 complex is less active in the hydrolysis of
Up(2′-5′)U because only one Cu(II) ion is available for
complex formation with the uracil moieties. For efficient
transformation, at least one Cu(II) ion must bind the phosphodi-
ester linkage to activate it. These arguments are supported based
on the following experiments using Zn(II) ions which are known
to strongly bind uracil.22 In the hydrolysis of Up(2′-5′)U, the
trinuclear Zn(II) complex of L3A is 14 times more active than
that of the dinuclear Zn(II) complex of L2. For Ap(3′-5′)A
hydrolysis, however, the trinuclear Zn(II) complex is almost as
active as the dinuclear Zn(II) complex.
Figure 4. Schematic drawings of the adducts between the trinuclear Cu-
(II)/L3A complex and various dinucleotides: (A) Up(2′-5′)U and (B) Up-
(3′-5′)U.
In the case of the hydrolysis of Ap(3′-5′)A by the trinuclear
Cu(II)/L3A complex, the adenine moieties may hydrophobically
stack with the aromatic groups in the L3A ligand, because direct
coordination of adenine to Cu(II) ion is inefficient.20 To
maximize the overlap between these apolar moieties and
stabilize the substrate-catalyst adduct, the trinuclear Cu(II)
complex is pulled toward the phosphodiester linkage. This
enables coordination of the third Cu(II) ion to the phosphodi-
ester, activating it for hydrolysis. The hydrolysis of the (2′-5′)
isomer is less efficient due to unfavorable steric factors.23 The
proposed importance of steric factors is supported by the fact
that the trinuclear Cu(II)/L3A complex is less active than the
dinuclear Cu(II)/L2 complex only in this substrate (see Table
to the strong binding of this substrate to this complex, whereas
the large kcat value is responsible for the rapid hydrolysis of
Ap(3′-5′)A.
Possible Mechanisms for Substrate-Specific Hydrolysis.
The hydrolysis of Up(2′-5′)U by the trinuclear Cu(II)/L3A
complex is characterized by a very small Km value (0.21 mM:
Table 2). This value is far smaller than the value (47 mM) for
the coordination of 4-nitrophenyl ethyl phosphate to the 2,2′-
bipyridine/Cu(II) complex.8a,17 Furthermore, it is even smaller
than the value (0.63 mM)18 for the coordination of HPO4 to
-2
1). Consistently, both Gp(3′-5′)G and Gp(2′-5′)G are not
20
Cu(II) ion, although this dianionic ligand should be more
favorable for the coordination than is UpU, from the viewpoint
of electrostatic interactions. For the hydrolysis of Ap(3′-5′)A,
however, substrate-binding is not promoted (Km ) 17 mM).
These results indicate that the two uridine residues of UpU are
coordinated to the Cu(II) ions in the L3A complex.19 In fact,
literature precedent substantiates coordination of uracil to Cu-
(II) ion at its N3 atom.20
The proposed structures of the adducts between UpU and
the Cu(II)/L3A complex are schematically depicted in Figure
4. The third Cu(II) ion of this trinuclear complex is the catalytic
center. In the adduct of Up(2′-5′)U (A), the scissile phosphodi-
ester linkage is located nearer to this Cu(II) ion than in the
adduct of Up(3′-5′)U (B), simply because the uracil is bound
to the 1′-carbon atom of the ribose. Despite many rotatable
hydrolyzed by this complex despite the fact
that guanine
coordinates Cu(II) ions at N7. The coordination of the guanine
moieties to two of the Cu(II) ions in the L3A complex orient
the phosphodiester linkage away from the third Cu(II) ion,
thereby impeding catalytic hydrolysis. Similar recognition of
nucleic bases by dinuclear Cu(II) complexes was previously
proposed for the hydrolysis of 2′,3′-cyclic monophosphates of
ribonucleosides.24 The base-selectivities were ascribed to the
difference in the strength of hydrogen-bonding and/or stacking
interaction between the complex and the substrate.
(21) This argument does not imply that the conformation in which all the three
dipicolylamine-Cu moieties exist in the same plane with respect to the
central benzene is the most stable one. Rather, the flexible nature of this
complex allows it to take this conformation for the cooperative catalysis.
According to the molecular modeling, the distance between the phosphorus
atom and the third Cu(II) ion in 2′-5′ UpU is 8.2 Å, which is by 0.6 Å
smaller than that in its 3′-5′ counterpart. For the chemical transformation
in the catalysis, the system should show conformational change to place
the third Cu(II) ion closer to the scissile phosphodiester linkage. In this
calculation, the structure of the adduct between the trinuclear Cu complex
and UpU was first optimized under the assumptions that (1) both the two
uracil residues and the phosphate are coordinated to the three Cu(II) ions,
and (2) the coordination modes of these Cu(II) complexes are square-planar.
Then, the phosphate was removed from the corresponding Cu(II), to which
a water molecule was bound instead. Finally, the whole system was
optimized.
(17) The Km for the coordination of bis(2,4-nitrophenyl)phosphate to the Cu-
(II) complex of a modified bipyridine-ligand is 13 mM in 19/1 ethanol/
water mixture; Ko¨va´ri, E.; Kra¨mer, R. J. Am. Chem. Soc. 1996, 118, 12 704.
(18) Sigel, H.; Decker, K.; McCormick, D. B., Biochim. Biophys. Acta, 1967,
89, 1987.
(19) With regards to these results, the possibility that the third picolylamine-
Cu moiety in the Cu(II)/L3A complex simply forces the other two moieties
to the same side of the central benzene ring is unlikely. If this were the
case, and only one of the three picolylamine-Cu moieties were binding
UpU (another moiety must function as the catalyst), then the UpU binding
by this complex could never be so unusually efficient.
(22) The binding constant of uridine(at the N3 atom) to Zn(II) is 103.57 (data
from ref 20).
(20) The binding constants for the coordination of uridine (N3) and guanosine
(N7) are 105.9 and 105.58, respectively: Biocoordination Chemistry: Co-
ordination Equilibria in Biologically ActiVe Systems, Ka´lma´n Burger, 1990,
Ellis Harwood: New York. The value for adenine is 102.25 (the coordination
site is unclear).
(23) According to the molecular modeling study, the distances between the P
atom and the third Cu(II) ion are 4.6 Å for Ap(3′-5′)A and 5.8 Å for
Ap(2′-5′)A, respectively.
(24) Liu, S.; Luo, Z.; Hamilton, A. D. Angew. Chem., Int. Ed. Engl. 1997, 36,
2678.
9
13734 J. AM. CHEM. SOC. VOL. 124, NO. 46, 2002

Trinuclear Copper(II) Complex
A R T I C L E S
filtrated, and recrystallized from chloroform/petroleum ether to needle-
shaped crystals (2.5 g).
Conclusion
Trinuclear Cu(II) complex of L3A hydrolyzes 2′-5′ and 3′-
5′ ribonucleotides with remarkable substrate specificity. Up-
(2′-5′)U is hydrolyzed more efficiently than is Up(3′-5′)U,
but Ap(3′-5′)A is overwhelmingly preferred to Ap(2′-5′)A.
Neither Gp(2′-5′)G nor Gp(3′-5′)G is hydrolyzed to any
extent. This complex is, to the best of our knowledge, the first
multinuclear metal complex that clearly distinguishes between
2′-5′ phosphodiester linkages and 3′-5′ phosphodiester link-
ages. The corresponding dinuclear Cu(II) complex of L2 shows
little substrate-specificity, indicating that three Cu(II) ions are
necessary. The data reported herein demonstrates that the
cooperation of several metal ions in multinuclear metal com-
plexes can give rise to notable substrate-specificity in addition
to large accelerations of hydrolysis. A design criteria for
catalysis of RNA dinucleotide hydrolysis in a sequence de-
pendent manner has emerged.
To a stirred solution of 1,3,5-tris(bromomethyl)benzene (100 mg,
0.28 mol) in acetonitrile (18 mL), was added 2,2′-dipicolylamine (167
mg, 0.84 mmol) and N,N-diisopropylethylamine (116 mg, 0.90 mmol).
The resulting solution was stirred at room temperature overnight, and
the solvent was removed to yield a brown product which was purified
by silica gel chromatography (acetonitrile/triethylamine ) 9/1 (v/v)).
1H NMR (500 MHz, CDCl3 / TMS) δ 8.50 (d, J ) 5.0 Hz, 6H, Py-
H6), 7.54-7.59 (m, 12H, Py-H3, H4), 7.36 (3H, Bz-H), 7.11 (m, 6H,
Py-H5), 3.81 (12H, NCH2Py), 3.68 (6H, NCH2Bz); 13C NMR (125.6
MHz, CDCl3/TMS) δ 159.9, 149.0, 139.2, 136.3, 128.0, 122.7, 121.9,
60.2, 58.7.
(2) N,N,N′,N′,N′′,N′′-Hexa[(2-pyridyl)methyl]-1,3,5-tris(amino-
methyl)-2,4,6-triethylbenzene (L3B). 1,3,5-Tris(chloromethyl)-2,4,6-
10
triethylbenzene (640 mg, 2.1 mmol) was dissolved in acetonitrile
(25 mL), and to this solution 3,3′-dipicolylamine (1.65 g, 8.3 mmol)
and then potassium carbonate (862 mg, 6.3 mmol) were added. The
reaction mixture was refluxed for 72 h and filtered through a bed of
diatomaceous earth using dichloromethane. By removing the solvent,
a brown solid was obtained. The crude mixture was triturated with
cold ethyl acetate and decanted. The light brown solid was recrystallized
using ethyl acetate/hexane to yield white crystals (148 mg, 60%
yield): 1H NMR (300 MHz, CD3OD/TMS) δ 8.33 (d, J ) 4.8, 6H),
7.51 (dt, J ) 1.5, J ) 7.8, 6H), 7.22 (d, J ) 8.1, 6H), 7.11 (m, 6H),
3.65 (s, 12H), 3.63 (s, 6H). 2.92 (q, J ) 7.5, 6H) 0.65 (t, J ) 7.2, 9H);
13C NMR (CD3OD, 300 MHz) δ 160.5, 149.1, 146.6, 138.4, 125.1,
123.6, 60.5, 51.9, 23.2, 16.0; HRMS (CI, m/z) calcd for C51H57N9
796.48, found 796.48.
Experimental Section
Materials. Diribonucleotides were purchased from Seikagaku Ko-
gyo. Highly purified water (the specific resistance > 18.3 ohm‚cm)
was sterilized at 120 °C immediately before use. Other chemicals were
commercially obtained. Throughout the present study, great care was
taken to avoid contamination by natural enzymes and other metal ions.
The dinuclear ligand N,N,N′,N′,-tetra[(2-pyridyl)methyl]-1,3-bisami-
noethylbenzene (L2) was synthesized and purified according to the
literature method.24
Preparation of the Ligands L3A, L3B, and L2. (1) N,N,N′,N′,-
N′′,N′′-Hexa[(2-pyridyl)methyl]-1,3,5-tris(aminomethyl)benzene (L3A).
The starting material 1,3,5-tris(bromomethyl)benzene was synthesized
as follows.9 Mesitylene (35 mL, 0.25 mol) and N-bromosuccinimide
(1.42 g, 0.80 mol) were dissolved in 600 mL of carbon tetrachloride,
and benzoyl peroxide (10 mg) was added with a bamboo spoon. After
the solution was refluxed for 96 h under irradiation from a UV lamp,
the precipitates were removed by suction filtration and the filtrate was
evaporated. The solid material was dissolved in 200 mL of chloroform
and washed with sodium bicarbonate solution (150 mL) and then with
distilled water (450 mL). The organic layer was dried with sodium
sulfate and evaporated. The product was dissolved in petroleum ether,
Hydrolysis of Diribonucleotides. The reaction mixtures were
prepared by adding an acetonitrile solution of the ligand to 50 mM
HEPES buffer containing Cu(ClO4)2. The pH was adjusted to 7.0 using
aliquots of NaOH. A stock solution of the substrate in water (10 mM)
was added, and the reaction was carried out at 50 °C. The initial
concentration of the substrate was 0.1 mM. At appropriate intervals, a
small portion of the reaction mixture was taken and combined with an
aqueous EDTA (100 mM) solution. The specimen was treated with a
pretreatment filter (Tosoh, W-3-2) and analyzed by reverse-phase
HPLC (Merck LiChrospher RP-18(e) ODS column; water/acetonitrile
) 92/8 (v/v)). The HPLC peaks were assigned by co-injection with
Table 3. Crystal Data and Structure Refinement for L3B
empirical formula
formula weight
temperature
C113 H158 Cl12 Cu6 N18 O11
2751.22
153(2) K
wavelength
crystal system
space group
0.71073 Å
monoclinic
P21
unit cell dimensions
a ) 12.1429(3) Å
b ) 33.2454(8) Å
c ) 15.7008(4) Å
6318.9(3) Å3
2
R) 90°.
â) 94.494(1)°.
γ ) 90°.
volume
Z
density (calculated)
absorption coefficient
F(000)
1.446 Mg/m3
1.308 mm-1
2856
crystal size
theta range for data collection
index ranges
0.22 × 0.10 × 0.04 mm
2.97 to 26.55°.
-15<)h<)14, -40<)k<)33, -18<)l<)18
18056
reflections collected
independent reflections
refinement method
data/restraints/parameters
goodness-of-fit on F2
final R indices [I>2sigma(I)]
R indices (all data)
absolute structure parameter
largest diff. peak and hole
18056
Full-matrix-block least-squares on F2
18056/967/1455
1.867
R1 ) 0.116, wR2 ) 0.164
R1 ) 0.195, wR2 ) 0.180
0.11(2)
1.024 and -0.933 e.Å-3
9
J. AM. CHEM. SOC. VOL. 124, NO. 46, 2002 13735

A R T I C L E S
Komiyama et al.
authentic samples. All reactions satisfactorily showed pseudo-first-order
kinetics, and the pH change during the reactions was less than 0.1 pH
unit.
UV-Visible Absorption Spectra and Potentiometric Titration.
The spectra were measured on a JASCO U-520 spectrometer equipped
with a temperature-controller. The cell path length was 1.0 cm. All of
the potentiometric titration for the Cu(II) complexes was carried out at
25 °C.
Ueq of the attached atom (1.5 × Ueq for methyl hydrogen atoms).
Hydrogen atom positions in the vicinity of the oxygen atoms of the
methanol molecules were not located in a ∆F map. As a result, these
hydrogen atoms were not included in the refinement model. The
absolute structure was determined by the method of Flack.29 The Flack
2
2
x parameter refined to 0.11(2). The function, Σw(|Fo| - |Fc| )2, was
minimized, where w ) 1/[(σ(Fo))2 + (0.02*P)2] and P ) (|Fo|
+
2
2
2|Fc| )/3. Rw(F2) refined to 0.180, with R(F) equal to 0.116 and a
goodness of fit, S, ) 1.87. Definitions used for calculating R(F),-
Rw(F2) and the goodness of fit, S, are given below.30
X-ray Crystallography. The ligand L3B and 3 equiv of CuCl2 were
dissolved in methanol and allowed to evaporate slowly at 25 °C to
yield a blue crystal.
X-ray Experimental for [C51H57N9Cu3Cl3]2-2Cl-- 11CH3OH: Crys-
tals grew as pale blue needles by slow evaporation from methanol.
Data were collected at 153 K on a Nonius Kappa CCD diffractometer
using a graphite monochromator with MoKR radiation (λ ) 0.71073
Å). A total of 594 frames of data were collected using ω-scans with a
scan range of 0.5° and a counting time of 98 s per frame. Details of
crystal data and structure refinement are listed in Table 3. Data reduction
were performed using DENZO-SMN.26 The structure was solved by
direct methods using SIR9227 and refined by full-matrix least-squares
on F2 with anisotropic displacement parameters for the non-H atoms
using SHELXL-97.28 The hydrogen atoms on carbon were calculated
in ideal positions with isotropic displacement parameters set to 1.2 ×
Acknowledgment. The authors thank Professors Masanobu
Hidai and Yoichi Ishii for their kind advice in X-ray crystal-
lography. This work was partially supported by a Grant-in-Aid
for Scientific Research from the Ministry of Education, Science,
and Culture, Japan. We also gratefully acknowledge support
for this work from the National Science Foundation.
Supporting Information Available: X-ray crystallographic
data (CIF). This material is available free of charge via the
Internet at http://pubs.acs.org.
JA020877T
(25) Gulneh, Y.; Ahvazi, B.; Khan, A. R.; Butcher, R. J.; Tuchagues J. P., Inorg.
Chem. 1995, 34, 3633.
(28) Sheldrick, G. M. SHELXL97. Program for the Refinement of Crystal
Structures; University of Gottingen: Germany, 1994.
(26) DENZO-SMN. Otwinowski, Z.; Minor, W. Methods in Enzymology,
276: Macromolecular Crystallography, part A; Carter, C. W., Jr., Sweets,
R. M., Eds.; Academic Press: New York, 1997; 307-326.
(27) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A. SIR92. A
Program for Crystal Structure Solution. J. Appl. Crystallogr. 1993, 26, 343-
350.
(29) Flack, H. D. Acta Crystallogr. 1983, A39, 876-881.
(30) Rw(F2) ) {Σw(|Fo| - |Fc| )2/Σw(|Fo|)4}1/2 where w is the weight given
2
2
each reflection. R(F) ) Σ(|Fo| - |Fc|)/Σ|Fo|} for reflections with Fo
>
2
2
4(σ(Fo)). S ) [Σw(|Fo| - |Fc| )2/(n - p)]1/2, where n is the number of
reflections and p is the number of refined parameters.
9
13736 J. AM. CHEM. SOC. VOL. 124, NO. 46, 2002