Efficient and selective cleavage of RNA oligonucleotides by calix[4]arene-based synthetic metallonucleases

Article abstract of DOI:10.1021/ja0737366

Di- and trinuclear copper(II) complexes of [12]aneN₃ macrocycles anchored at the upper rim of cone calix[4]arenes in 1,2-, 1,3-, and 1,2,3-positions were investigated as cleaving agents of 6-, 7-, and 17-meric oligoribonucleotides. A kinetic investigation of the cleavage reactions was carried out using gel electrophoresis to separate and analyze reactants and products having a radioactive phosphate label in the terminal opposition. The degree of cooperation was assessed on the basis of a comparison with rates of cleavage by mononuclear controls. A remarkable selectivity of cleavage of the CpA phosphodiester bond was observed for all metal complexes, in sharp contrast with the UpU and UpG selectivity previously observed in the cleavage of diribonucleoside monophosphates by the same metal complexes. The highest rate acceleration, brought about in the cleavage of the 5′-pCpA bond in hexanucleotide 9 by 50 μM trinuclear complex 5-Cu₃ (water solution, pH 7.4, 50°C), amounts to 5 × 10⁵-fold, as based on the estimated background reactivity of the CpA dimer. Selectivity in the cleavage of oligoribonucleotides by copper(II) complexes closely resembles that experienced by ribonuclease A and by a number of metal-independent RNase A mimicks. The possible role of the dianionic phosphate at the 5′-terminal positions as a primary anchoring site for the metal catalyst is discussed.

Full text of DOI:10.1021/ja0737366

Published on Web 09/20/2007
Efficient and Selective Cleavage of RNA Oligonucleotides by
Calix[4]arene-Based Synthetic Metallonucleases
Roberta Cacciapaglia,† Alessandro Casnati,‡ Luigi Mandolini,*,† Alessio Peracchi,*,#
David N. Reinhoudt,*,§ Riccardo Salvio,† Andrea Sartori,§ and Rocco Ungaro*,‡
Contribution from the Dipartimento di Chimica and IMC - CNR Sezione Meccanismi di
Reazione, UniVersita` La Sapienza, Box 34 - Roma 62, 00185 Roma, Italy, Dipartimento di
Chimica Organica e Industriale and Dipartimento di Biochimica e Biologia Molecolare
UniVersita` degli Studi di Parma, V.le G. P. Usberti 17/A, 43100 Parma, Italy, and Laboratory of
Supramolecular Chemistry and Technology, MESA+ - Research Institute, UniVersity of Twente,
P.O. Box 217, 7500 AE Enschede, The Netherlands
Received May 24, 2007; E-mail: luigi.mandolini@uniroma1.it; rocco.ungaro@unipr.it
Abstract: Di- and trinuclear copper(II) complexes of [12]aneN3 macrocycles anchored at the upper rim of
cone calix[4]arenes in 1,2-, 1,3-, and 1,2,3-positions were investigated as cleaving agents of 6-, 7-, and
17-meric oligoribonucleotides. A kinetic investigation of the cleavage reactions was carried out using gel
electrophoresis to separate and analyze reactants and products having a radioactive phosphate label in
the terminal 5′-position. The degree of cooperation was assessed on the basis of a comparison with rates
of cleavage by mononuclear controls. A remarkable selectivity of cleavage of the CpA phosphodiester
bond was observed for all metal complexes, in sharp contrast with the UpU and UpG selectivity previously
observed in the cleavage of diribonucleoside monophosphates by the same metal complexes. The highest
rate acceleration, brought about in the cleavage of the 5′-pCpA bond in hexanucleotide 9 by 50 µM trinuclear
complex 5-Cu3 (water solution, pH 7.4, 50 °C), amounts to 5 × 105-fold, as based on the estimated
background reactivity of the CpA dimer. Selectivity in the cleavage of oligoribonucleotides by copper(II)
complexes closely resembles that experienced by ribonuclease A and by a number of metal-independent
RNase A mimicks. The possible role of the dianionic phosphate at the 5′-terminal positions as a primary
anchoring site for the metal catalyst is discussed.
Introduction
efficiency (cleavage rates exceeding the rate of spontaneous
RNA cleavage by >108-fold) and substrate-selectivity peculiar
RNA transesterification reactions are fundamental and ubiq-
uitous processes in biochemistry. In living organisms, RNA is
cleaved by a diverse array of ribonucleases (RNases) for
purposes that include RNA processing and maturation, regula-
tion of gene expression, host-cell defense, and nucleotide
turnover.1 The majority of natural ribonucleases are proteins,
but an important subset is constituted by RNA enzymes
(ribozymes).2 Prototypical examples of protein RNases are
ribonuclease A and ribonuclease T1, whose mechanisms of
action rely on acid-base catalysis operated by amino acid side
chains.3 However, other protein enzymes and several ribozymes
employ divalent metal ions as cofactors in their catalytic
mechanisms.4 Natural RNases show the exceptional catalytic
to biocatalysts. In order to be selectively cleaved, RNA
molecules must contain specific sequences and/or possess
specific structural features. Ribozymes in general achieve
selectivity by sequence complementarity, i.e., by base-pairing
to specific sequences in their substrates.2 Protein enzymes,
lacking such possibility, achieve selectivity by appropriate
organization of amino acid side chains in their substrate-binding
sites.
The development of artificial ribonucleases capable of
mimicking the action of natural RNases5,6 is a long-standing
goal of bioorganic chemistry, with implications for biotechnol-
ogy and medicine. Small molecules able to cleave RNA
efficiently, in a nondegradative manner, and with high levels
of selectivity will offer many applications for the design of
structural probes and perhaps the development of novel thera-
peutics.
† Universita` di Roma.
‡ Universita` di Parma, DCOI.
# Universita` di Parma, DBBM.
§ University of Twente.
(1) (a) Carmell, M. A.; Hannon, G. J. Nat. Struct. Mol. Biol. 2004, 11, 214-
218. (b) Yoshida, H. Ribonucleases 2001, 341, 28-41. (c) Nicholson, A.
W. FEMS Microbiol. ReV. 1999, 23, 371-390. (d) Beintema, J. J. Cell
Mol. Life Sci. 1998, 54, 763-765. (e) Beintema, J. J.; Kleineidam, R. G.
Cell Mol. Life Sci. 1998, 54, 825-832.
(2) Cech, T. R. Biochem. Soc. Trans. 2002, 30, 1162-1166.
(3) (a) Steyaert, J. Eur. J. Biochem. 1997, 247, 1-11. (b) Nogues, M. V.;
Vilanova, M.; Cuchillo, C. M. Biochim. Biophys. Acta 1995, 1253, 16-
24.
(4) (a) Weston, J. Chem. ReV. 2005, 105, 2151-2174. (b) Lilley, D. M. J.
Trends Biochem. Sci. 2003, 28, 495-501. (c) Pyle, A. M. J. Biol. Inorg.
Chem. 2002, 7, 679-690. (d) Jedrzejas, M. J.; Setlow, P. Chem. ReV. 2001,
101, 607-618. (e) Cowan, J. A. Chem. ReV. 1998, 98, 1067-1087. (f)
Brautigam, C. A.; Steitz, T. A. J. Mol. Biol. 1998, 277, 363-377. (g)
Wilcox, D. E. Chem. ReV. 1996, 96, 2435-2458. (h) Strater, N.; Lipscomb,
W. N.; Klabunde, T.; Krebs, B. Angew. Chem., Int. Ed. Engl. 1996, 35,
2024-2055.
9
12512
J. AM. CHEM. SOC. 2007, 129, 12512-12520
10.1021/ja0737366 CCC: $37.00 © 2007 American Chemical Society

Synthetic Cu(II)−Nucleases for Oligo-RNA Cleavage
A R T I C L E S
Chart 1
Among the chemical models of natural ribonucleases a special
place is occupied by artificial metallonucleases6 where one or
more metal ions (usually lanthanides or transition metals)
complexed to suitable synthetic ligands perform their catalytic
function in the cleavage of phosphodiester bonds. With few
exceptions6a,b,7-9 synthetic metallonucleases as such are char-
acterized by low base or sequence selectivity, the latter being
in general achieved upon conjugation of the active catalysts to
antisense oligonucleotides, peptides, or small molecules which
act as RNA recognition agents.5a,b,d,f,o,10-12
diribonucleoside monophosphate (NpN′) cleavage.6a In particu-
lar, the bimetallic 3-Cu2 and the trimetallic 5-Cu3 catalysts
showed a remarkable base selectivity for the cleavage of
substrates having a uracil base at the 5′-hydroxyl terminus,
namely for UpU and UpG. These results prompted us to test
our copper(II) complexes toward longer oligoribonucleotides,
which are of higher biological interest but have seldom been
used as substrates of synthetic metallonucleases.13
Here we report on the catalytic activity of the bimetallic
complexes 3-Cu2 and 4-Cu2 and of the trimetallic complex
5-Cu3 in the cleavage of ribonucleotides 6-1214 in water, pH
7.4, at 50 °C. The activity of the monometallic complexes 1-Cu
and 2-Cu was also investigated for comparison purposes.
We have recently reported on the synthesis of a series of
water-soluble copper(II) complexes of calix[4]arene-based
nitrogen ligands (Chart 1) and on their catalytic activity of
(5) For review articles, see: (a) Niittymaki, T.; Lo¨nnberg, H. Org. Biomol.
Chem. 2006, 4, 15-25. (b) Lo¨nnberg, T.; Lo¨nnberg, H. Curr. Opin. Chem.
Biol. 2005, 9, 665-673. (c) Mancin, F.; Scrimin, P.; Tecilla, P.; Tonellato,
U. Chem. Commun. 2005, 2540-2548. (d) Morrow, J. R.; Iranzo, O. Curr.
Opin. Chem. Biol. 2004, 8, 192-200. (e) Kuznetsova, I. L.; Sil’nikov, V.
N. In Artificial Nucleases; Zenkova, M. A., Ed.; Springer-Verlag: Heidel-
berg, 2004; pp 111-127. (f) Franklin, S. J. Curr. Opin. Chem. Biol. 2001,
5, 201-208. (g) Cowan, J. A. Curr. Opin. Chem. Biol. 2001, 5, 634-642.
(h) Komiyama, M.; Sumaoka, J.; Kuzuya, A.; Yamamoto, Y. In Ribonu-
cleases: Methods in Enzymology; Nicholson, A. W., Ed.; 2001; Vol. 341,
pp 455-468. (i) Zenkova, M.; Beloglazova, N.; Sil’nikov, V.; Vlassov,
V.; Giege, R. In Ribonucleases: Methods in Enzymology; Nicholson, A.
W., Ed.; 2001; Vol. 341, pp 468-490. (j) Molenveld, P.; Engbersen, J. F.
J.; Reinhoudt, D. N. Chem. Soc. ReV. 2000, 29, 75-86. (k) Kimura, E.
Curr. Opin. Chem. Biol. 2000, 4, 207-213. (l) Williams, N. H.; Takasaki,
B.; Wall, M.; Chin, J. Acc. Chem. Res. 1999, 32, 485-493. (m) Bashkin,
J. K. Curr. Opin. Chem. Biol. 1999, 3, 752-758. (n) Hegg, E. L.; Burstyn,
J. N. Coord. Chem. ReV. 1998, 173, 133-165. (o) Komiyama, M.;
Sumaoka, J. Curr. Opin. Chem. Biol. 1998, 2, 751-757.
(6) For recent contributions see, for example, the following papers and
references therein: (a) Cacciapaglia, R.; Casnati, A.; Mandolini, L.;
Reinhoudt, D. N.; Salvio, R.; Sartori, A.; Ungaro, R. J. Am. Chem. Soc.
2006, 128, 12322-12330. (b) Wang, Q.; Lo¨nnberg, H. J. Am. Chem. Soc.
2006, 128, 10716-10728. (c) O’Donoghue, A.; Pyun, S. Y.; Yang, M. Y.;
Morrow, J. R.; Richard, J. P. J. Am. Chem. Soc. 2006, 128, 1615-1621.
(d) Neverov, A. A.; Lu, Z.-L.; Maxwell, C. I.; Mohamed, M. F.; White, C.
J.; Tsang, J. S. W.; Brown, R. S. J. Am. Chem. Soc. 2006, 128, 16398-
16405. (e) Boseggia, E.; Gatos, M.; Lucatello, L.; Mancin, F.; Moro, S.;
Palumbo, M.; Sissi, C.; Tecilla, P.; Tonellato, U.; Zagotto, G. J. Am. Chem.
Soc. 2004, 126, 4543-4549. (f) Livieri, M.; Mancin, F.; Tonellato, U.;
Chin, J. Chem. Commun. 2004, 2862-2863.
Results
The activity of copper(II) complexes based on [12]aneN3
(Chart 1) in the cleavage of 0.05-0.5 nM oligoribonucleotides
6-12 was investigated in water (15 mM 1,4-piperazinediethane
sulfonic acid (PIPES) buffer, pH 7.4, I ) 0.07, 50 µM sodium
dodecyl sulfate (SDS), 50 °C). To avoid contamination by
natural nucleases, all of the laboratory equipment which came
into contact with solutions of oligoribonucleotides, calixarenes,
and buffer were sterilized and used only once. Experiments were
(7) Molenveld, P.; Engbersen, J. F. J.; Reinhoudt, D. N. Angew. Chem., Int.
Ed. 1999, 38, 3189-3192.
(8) Liu, S. H.; Hamilton, A. D. Chem. Commun. 1999, 587-588.
(9) Yashiro, M.; Ishikubo, A.; Komiyama, M. Chem. Commun. 1997, 83-84.
(10) Niittymaki, T.; Lo¨nnberg, H. Bioconjugate Chem. 2004, 15, 1275-1280.
(11) Sakamoto, S.; Tamura, T.; Furukawa, T.; Komatsu, Y.; Ohtsuka, E.;
Kitamura, M.; Inoue, H. Nucleic Acids Res. 2003, 31, 1416-1425.
(12) Kuzuya, A.; Mizoguchi, R.; Morisawa, F.; Machida, K.; Komiyama, M. J.
Am. Chem. Soc. 2002, 124, 6887-6894.
(13) Jancso´, A.; Mikkola, S.; Lo¨nnberg, H.; Hegetschweiler, K.; Gajda, T. Chem.
Eur. J. 2003, 9, 5404-5415.
(14) It is highly improbable that these oligomers adopt any regular secondary
structure at 50 °C, in view of their reduced size. Algorithms commonly
used to compute the stability of RNA structures, such as RNAfold and
sFold, predict compounds 6-12 to be devoid of secondary structures.
9
J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007 12513

A R T I C L E S
Cacciapaglia et al.
Figure 1. Transesterification of a 5′-32P-labeled ribonucleotide by intramolecular attack of the 2′-OH. (B ) nucleobase).
carried out in order to ascertain the possible presence of
contaminating nucleases. The most reactive substrates (incubated
at 50 °C in 15 mM PIPES, 50 µM SDS, in the absence and
presence of 150 µM CuCl2) did not show any cleavage after 20
h. Furthermore, experiments at 25 °C in the presence of 5-Cu3
resulted in rates of cleavage lower than those determined at
50 °C, thereby arguing against the presence of contaminating
nucleases which should be at least partially denaturated at the
latter temperature.
The oligoribonucleotide cleavage was followed by gel elec-
trophoresis. To allow detection of the substrate and of the
cleavage products, a radioactive phosphate (32P) was introduced
into the 5′-terminal position by reaction with *γ-ATP and T4
polynucleotide kinase (T4 PNK). Product formation was fol-
lowed by analyzing on the gel small aliquots of the reaction
mixture quenched at proper time intervals. Detectable fragments
are only those containing the 5′-32P-end-label, whereas other
fragments not containing any radioactive labeling are lost (Figure
1). Identification of the cleavage products was carried out by
comparison with the products obtained in the aspecific hydroly-
sis of the oligonucleotides in 0.1 M NaOH. The basic hydrolysis
gives two labeled fragments for each phosphate cleavage, one
with a terminal 2′,3′-cyclic phosphate coming from transesteri-
fication and the other with a terminal acyclic 2′- or 3′-phosphate
produced by hydrolysis of the cyclic phosphate. The two
isomeric acyclic phosphates move on the gel slightly more
rapidly than the cyclic phosphate.15 In the experiments carried
out in the presence of calix[4]arene-copper(II) complexes, only
the fragments corresponding to the 2′,3′-cyclic phosphate were
observed in all cases. This indicates that formation of the cyclic
phosphate via intramolecular attack of 2′-OH (Figure 1) is not
followed by hydrolysis under the conditions of the cleavage
experiments. The bands on the gel corresponding to reactants
and products were quantitatively analyzed to evaluate the mole
fractions at different time intervals over a total reaction time of
4-6 h. Typical gel electrophoretic autoradiograms are shown
in Figures 2, 4, and 6. Time-concentration data of the fastest
reactions were subjected to time-course kinetic analysis by
means of appropriate integrated rate equations to obtain catalytic
pseudo-first-order rate constant kobs (s-1). An initial rate method
was applied to the slowest reactions, for which the reaction
percent in the 4-6 h time window was low. For the sake of
convenience, we shall report first on oligoribonucleotides 6-8,
then on oligoribonucleotides 9-11, and finally on the hepta-
decamer oligoribonucleotide 12. The metal complex concentra-
tion in the cleavage of 6-11 was 50 µM, and in the cleavage
of 12 it was 10 µM.16
Figure 2. Gel electrophoretic autoradiograms corresponding to the cleavage
of oligoribonucleotides 6-8 promoted by 50 µM 5-Cu3. Lanes 1, 8, and
14 show the aspecific hydrolysis in 0.1 M NaOH.
Cleavage of 6-8. We first investigated the cleavage of
oligoribonucleotides 6-8 in which all of the 16 possible
diribonucleotide combinations are present, thereby enabling a
preliminary screening of the efficiency and selectivity of the
cleaving agents.
The gel electrophoretic autoradiograms related to the reactions
of the three oligoribonucleotides in the presence of 50 µM
trimetallic complex 5-Cu3 (Figure 2)17 show that cleavage of
6 is much faster than cleavage of 7 and 8. In the cleavage of 6,
which takes place in the CpA and CpC bonds (Figure 2, left-
hand panel), the disappearance of the band corresponding to
the reactant is accompanied by the appearance of the bands
(15) Kierzek, R. Nucleic Acids Res. 1992, 20, 5079-5084.
9
12514 J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007

Synthetic Cu(II)−Nucleases for Oligo-RNA Cleavage
A R T I C L E S
Table 1. Cleavage of Oligoribonucleotides 6-8 Promoted by
Copper(II)-Calixarene Complexes; Pseudo First-Order Rate
Constants (106 k, s-1
)
in Water, pH 7.4, 50 °Cb,c
a
Figure 3. Reaction of 5′-p*ACCAUC (6) promoted by 50 µM 5-Cu3.
Mole fractions of reactant (1), and of fragments 5′-p*ACC (O; CpA
cleavage), and 5′-p*AC (b; CpC cleavage) vs time.
Scheme 1. Parallel First-Order Reactions
corresponding to the ACC and AC fragments. Plots of the mole
fractions against time are reported in Figure 3. Data points show
a reasonable adherence to exponential curves based on the same
first-order rate constant k ) 1.2 × 10-4 s-1, which is consistent
with a reaction scheme in which a reactant R undergoes two
parallel first-order reactions leading to products P1 and P2
(Scheme 1 and eqs 1 and 2,
a
Rate constants for the cleavage of 6 by 3-Cu2 and 5-Cu3 from time-
course kinetics, error (10%. All other k values from initial rates, error
b
(15%. Runs carried out on <0.2 nM substrate solutions, in the presence
c
of 50 µM catalyst and 50 µM SDS. The width of the arrows pointing to
k1
d
the scissile bonds gives an indication of the relative reactivity. In the
[P1]/[R]0 ) (1 - e-kt
)
)
(1)
(2)
presence of 50 µM monometallic catalyst 1-Cu, kCA ) 4.6 × 10-7 s-1
.
k
e
No cleavage of the CpC bond was observed in 5 h. No reaction in 5 h.
k2
[P2]/[R]0 ) (1 - e-kt
not act as independent units but show variable degrees of
cooperation among metal ions. Interestingly, the 1,3-distal
regioisomer 4-Cu2 is more effective than its 1,2-vicinal
regioisomer 3-Cu2 in the cleavage of 7, but the reverse holds
in the cleavage of 6 and 8. The N-benzylated mononuclear
complex 1-Cu (see footnote d to Table 1) cleaves the CpA
bond of 6 one order of magnitude more slowly than the
mononuclear calixarene complex 2-Cu and is inactive toward
the less reactive CpC bond. This finding suggests that the
calixarene moiety plays an active role in the catalysis. A possible
reason for this is the hydrophobicity of the calixarene moiety
which can enhance the binding to the aromatic nucleobases.
Cleavage of 9-11. Since the reactivity of a given phosphate
bond depends on the position in the oligomer, as well as on the
base sequence of the oligomer itself, it was felt that investigation
of substrates with different structural features was appropriate.
Accordingly, the catalytic cleavage of oligoribonucleotides 9-11
was investigated under the same conditions used for the cleavage
of 6-8. The CpA phosphodiester bond, which is the most
k
with k ) k1 + k2 and [P1]/[P2] ) k1/k2). The central and right-
hand panels in Figure 2 show that, whereas cleavage of 7 occurs
in the CpG, CpU, and UpG bonds, 8 is cleaved selectively in
the UpA bond. Rate constants for such reactions, as for other
slow reactions, were calculated from initial rates.
The rate constant k for the cleavage of all scissile bonds under
the reaction conditions are collected in Table 1. We first note
that the copper(II)-calixarene complexes behave as effective
and selective artificial ribonucleases. The CpA sequence turns
out to be the most reactive, independent of catalyst identity.
More in general, the cleavage patterns appear to be independent
of catalyst identity, and so does the reactivity order 6 > 7 ≈ 8,
with the notable exception of 3-Cu2 which is totally inactive
toward 7.
There is an undeniable tendency for the catalytic efficiency
to increase with the number of metal units in a way suggesting
that the subunits in the trimetallic and bimetallic complexes do
(16) From the log K value of 12.63 for binding of copper(II) to [12]aneN3 (see:
Zompa, L. J. Inorg. Chem. 1978, 17, 2531-2536) and the value of 12.6
for the pKa of the ligand (see: Kimura, E.; Shiota, T.; Koike, T.; Shiro,
M.; Kodama, M. J. Am. Chem. Soc. 1990, 112, 5805-5811) we calculate
that at pH 7.4 and 50 µM ligand and copper(II) ion, the fraction of ligated
metal ion is 0.97 and is 0.94 at a concentration of 10 µM.
reactive bond in the first set of oligonucleotides, is in the 5′-
terminal position in 9, and in the 3′-terminal position in 10.
The reverse holds for CpC. Hexanucleotide 11, which differs
from 6 for the presence of UpA in place of CpA was chosen to
compare the reactivity of different pairs in otherwise identical
oligoribonucleotide structures.
(17) The trimetallic complex 5-Cu3 is still catalytically active at 5 µM
concentration. The rate of cleavage of the CpA bond in 6 at 5 µM is 1
order of magnitude lower than that at 50 µM, which provides an indication
that the conditions of the catalytic experiments are well below saturation.
This implies that the affinity of the catalyst for the oligoribonucleotide
does not exceed 104 M-1
.
9
J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007 12515

A R T I C L E S
Cacciapaglia et al.
Scheme 2. Cleavage of 10 Promoted by 3-Cu2
Table 2. Cleavage of Oligoribonucleotides 9-11 Promoted by
Copper(II)-Calixarene Complexes; Pseudo-First-Order Rate
a
Constants (106 k, s-1
)
in Water, pH 7.4, 50 °Cb,c
Figure 4. Gel electrophoretic autoradiograms corresponding to the cleavage
of oligoribonucleotides 9-11 promoted by 50 µM 4-Cu2. Lanes 1, 7, and
14 correspond to the aspecific hydrolysis in 0.1 M NaOH.
Figure 5. Reaction of 5′-p*CCGGCA (10) promoted by 50 µM 1,2-vicinal
bimetallic complex 3-Cu2. Mole fractions of reactant (1), and of fragments
5′-p*C (3; CpC cleavage), 5′-p*CC (b; CpG cleavage), and 5′-p*CCGGC
(0; CpA cleavage) vs time. The points are experimental, and the curves
are calculated from integrated rate equations derived from Scheme 2 and
best fit values of rate constants reported in Table 2.
a Rate constants for the cleavage of 9 by 3-Cu2, 4-Cu2, and 5-Cu3, of
10 by 3-Cu2, and of 11 by 4-Cu2 and 5-Cu3 from time-course kinetics,
error (10%. All other k values from initial rates, error (15%. b Runs carried
out on <0.2 nM substrate solutions, in the presence of 50 µM catalyst and
The autoradiograms related to the cleavage of 9-11 in the
presence of the 1,3-distal bimetallic complex 4-Cu2 are shown
in Figure 4. The CpA bond in 9 is cleaved very effectively and
with high selectivity. The cleavage of the same bond in 10 is
much slower, and is accompanied by rupture of the CpC and
CpG bonds. In the case of 11, the only scissile bond is UpA,
which is cleaved with good efficiency.
c
50 µM SDS. The width of the arrows pointing to the scissile bonds gives
d
an indication of the relative reactivity. No reaction in 5 h.
rate constants: k1 ) 7.6 × 10-5 s-1; k2 ) 1.7 × 10-5 s-1; k3 )
4.7 × 10-5 s-1; k4 ) 5.4 × 10-5 s-1. The curves calculated
with the optimized values of the parameters show a close
adherence to data points (Figure 5).
The 1,2-vicinal bimetallic complex 3-Cu2 is more effective
than its 1,3-distal regioisomer 4-Cu2 by 5-fold in the cleavage
of all the scissile bonds in 10. Time-concentration data for the
reaction of 10 in the presence of 3-Cu2 are plotted in Figure
5. It is apparent that data points deviate from the behavior
expected on the basis of simple parallel first-order reactions.
The concentration of the fragment CCGGC increases in the early
stages of the reaction, then reaches a maximum, and eventually
decreases, thus indicating that CCGGC undergoes further
cleavage. Considering that cleavage of the CpC bond in 10 is
the fastest, we assume that CCGGC undergoes further cleavage
at the CpC site, as shown in Scheme 2. The set of differential
equations derived from Scheme 2 are amenable to analytical
integration (see Supporting Information). Best fit of data points
to the integrated rate equations gave the following values of
Rate constants for the cleavage of all the scissile bonds in
9-11 are collected in Table 2. Many of the reactivity and
selectivity features emerged from the reactions of oligoribo-
nucleotides 6-8 are also observed in the reactions of 9-11.
The much lower efficiency of the monometallic catalyst 2-Cu
compared with 3-Cu2 and 4-Cu2 clearly indicates that the two
metal ions in the bimetallic complex act in a synergic fashion.
There are cases in which the efficiency of bimetallic complexes
is either superior (cleavage of 10 by 3-Cu2) or virtually identical
(cleavage of 11 by 4-Cu2 and cleavage of the CpU bond in 7
by 4-Cu2) to that of the trimetallic complex 5-Cu3. There are
also cases in which 5-Cu3 is the best cleaving agent in the lot,
but the advantage due to the third metal ion hardly exceeds a
factor of 2, showing that a convincing evidence of the operation
of trimetallic catalysis is lacking.
9
12516 J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007

Synthetic Cu(II)−Nucleases for Oligo-RNA Cleavage
A R T I C L E S
Figure 7. Reaction of 5′-p*GCAAGCACAGACAUCAG (12) promoted
by 10 µM 1,2-vicinal bimetallic complex 3-Cu2. Mole fractions of reactant
(1) and fragments 5′-p*GC (O; R-CpA cleavage), 5′-p*GCAAGC (3;
â-CpA cleavage), and 5′-p*GCAAGCAC (0; γ-CpA cleavage) vs time.
The points are experimental, and the curves are calculated from integrated
rate equations derived from Scheme 3 and best fit values of rate constants
reported in Table 3.
Figure 6. Gel electrophoretic autoradiograms corresponding to the cleavage
of 12 in the presence of 10 µM 3-Cu2 (lanes 5-8) or 10 µM 4-Cu2 (lanes
9-12). Quenching times were 110, 195, 270, and 375 min in both
experiments. Lane 1 corresponds to the aspecific hydrolysis in 0.1 M NaOH.
Lanes 2-4 are blank experiments carried out in the absence of metal
complex.
Scheme 3. Cleavage of 12 Promoted by 3-Cu2 or 4-Cu2
Cleavage of 12. Substrate 12 is a heptadecamer featuring
six pyrimidine-pN sequences. Five of them are CpA, the most
reactive unit among the previously examined oligonucleotides
(Tables 1 and 2), and one is UpC, which in 6 was not cleaved
by any metal complex. Catalytic experiments on 12 could not
be carried out under the conditions used for oligonucleotides
6-11, due to a poor and irreproducible recovery of the radio-
activity on the electrophoresis gels. Good results were obtained
by increasing the SDS concentration to 150 µM and decreasing
the metal complex concentration to 10 µM. The autoradiograms
corresponding to the 3-Cu2 and 4-Cu2 promoted cleavage of
12 under the given conditions are shown in Figure 6. All the
CpA bonds are cleaved whereas other bonds do not undergo
appreciable scission in more than 6 h. CpA bonds labeled as R,
â, and γ are cleaved much more rapidly than the bonds labeled
as δ and ꢀ, which would suggest that far-away CpA bonds are
cleaved more slowly than bonds closer to the 5′-terminal
position.
Discussion
In the preceding section we have shown that di- and trinuclear
metal complexes 3-Cu2, 4-Cu2, and 5-Cu3 exhibit pronounced
ribonuclease activity - at concentrations as low as 10-50 µM
and under conditions close to physiological - toward oligori-
bonucleotides composed of different monomeric units. When-
ever multiple cleavages were observed, time-concentration data
showed a good adherence to kinetic schemes involving either
competitive or competitive-consecutive first-order reactions.
The results listed in Tables 1-3 offer a varied phenomeno-
logical picture. A prominent feature of the data is the remarkable
selectivity of the cleavage reactions. The most reactive phos-
phodiester bond is that in CpA, followed by UpA and CpC.
Other scissile bonds, namely, CpG, CpU, and UpG are cleaved
less effectively. The reactive bonds always have a pyrimidine
base in the 5′-position, with a marked preference for cytosine
over uracyl. However, not every NpN′ bond in which N is a
pyrimidine base is reactive enough to be cleaved under the
conditions of the catalytic runs, as no reaction of UpU in 8 and
UpC in 6 was observed. The base in 3′-position can be either a
pyrimidine or a purine base, but the decided superiority of
adenine is clearly indicated by the high reactivity of CpA and
UpA phosphodiester bonds.
The time-concentration profiles in Figure 7 are related to
the cleavage of the three most reactive CpA bonds in 12 pro-
moted by 3-Cu2. The shapes of the profiles suggest that frag-
ments resulting from â and γ cleavage undergo further cleavage
of the CpA bond closest to the 5′-terminus, as shown in Scheme
3. With the reasonable assumption that k4)k5)k1, the set of
differential equations consistent with Scheme 3 is amenable to
analytical integration (see Supporting Information). Numerical
values of the rate constants (Table 3), obtained by best fit of
data points to the rate equations, were used to calculate the
curves in Figure 7. A similar treatment was applied to the
cleavage of 12 in the presence of 4-Cu2. As shown in Table 3,
the two isomeric bimetallic complexes exhibit quite similar
catalytic activities, with a modest superiority of 3-Cu2 over
4-Cu2.
The catalytic rate enhancement (k/kbg) shown by the trinuclear
complex 5-Cu3 in the cleavage of CpA bond in 9 with respect
to the estimated value of the background reaction of CpA
9
J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007 12517

A R T I C L E S
Cacciapaglia et al.
Table 3. Cleavage of 12 Promoted by 3-Cu2 and 4-Cu2;
The paucity of quantitative data for the cleavage of single-
stranded oligoribonucleotides by di- and trimetallic complexes13
prevents a comparison of the results from the present work with
strictly analogous data. Still, we cannot help noting striking
similarities with literature reports on nonmetallic cleaving agents.
In a thorough investigation of the nonenzymatic cleavage of
short, single-stranded oligoribonucleotides by polyamines such
as spermidine, putresceine, and spermine, Kierzek19 reported
that the most reactive phosphodiester bonds are NpA and NpC,
where N is a pyrimidine base. Furthermore, ribonuclease A
mimicks composed of a bisquaternary DABCO moiety as RNA
binding domain and a histamine or hystidine residue as cleaving
unit, were reported to cleave oligoribonucleotides primarily in
the CpA and UpA sites, with a moderate preference for the
former bond.20,21 However, it is difficult to decide whether such
similarities are fortuitous, or may be taken as an indication of
similar mechanisms of action, in spite of the widely differing
nature of the cleaving agents.
a
Pseudo-First-Order Rate Constants (106 k, s-1
)
in Water, pH 7.4,
50 °Cb,c
Many artificial ribonucleases are composed of a catalytic
group, such as a ligated metal ion, and a highly structured probe
for substrate binding and sequence recognition.5a In simple
synthetic metalloribonucleases, such as those described in this
work, the tasks of substrate binding (recognition) and activation
of the bound substrate are assigned to ligated metal ions. Most
likely, the primary binding site for our metallocatalysts is the
doubly charged, labeled phosphate group present in the 5′-
terminal position of all investigated oligoribonucleotides.
b
a From time-course kinetics, error (10%. Runs carried out on 0.1 nM
substrate solutions, in the presence of 10 µM catalyst and 150 µM SDS.c
The width of the arrows pointing to the scissile bonds gives an indication
of the relative reactivity.
dinucleoside monophosphate (kbg ) 6.1 × 10-10 s-1, pH 7, T
) 50 °C)6a amounts to ca. 5 × 105-fold. This is one of the
highest values recorded for a synthetic metallonuclease in the
cleavage of a phosphodiester bond in ribonucleotide dimers or
higher oligomers.5a,j,l The corresponding figures for the 3-Cu2
catalyzed cleavage of CpC and CpG in 10 are 2 × 105 and 5 ×
104, respectively.
We have already touched upon the apparently important role
of the distance of a scissile bond from the 5′-terminus. Reactivity
ratios of 27:10:1 are calculated from the data in Tables 1 and
2 for the 5-Cu3 catalyzed cleavage of the CpA bond in 9, 6,
and 10, respectively. Similarly the CpC bond in 10 is cleaved
by 3-Cu2 16 times more rapidly than in 6, whereas its rate of
cleavage in 9 is not fast enough to compete with the cleavage
of the CpA bond. Although reactivity ratios vary from catalyst
to catalyst, the reactivity orders 9 > 6 g 10 for the CpA bond
and 10 > 6 > 9 for the CpC bond are independent of catalyst
identity. However, this is not always the case. The UpA bond
in 11 is cleaved more rapidly by 5-Cu3 and 4-Cu2 than in 8,
but the rates of cleavage in the presence of 3-Cu2 are almost
the same. Still odder is the cleavage of the CpG bond, whose
reactivity order is 7 > 10 in the presence of 5-Cu3, 7 ≈ 10 in
the presence of 4-Cu2, and 7 , 10 in the presence of 3-Cu2.
Furthermore, the three most reactive bonds in 12slabeled as
R, â, and γsare cleaved by the dinuclear catalysts at very
similar rates (Table 3) without displaying any reactivity decrease
on increasing the distance from the 5′-terminus. A reasonable
conclusion is that the distance of a scissile bond from the 5′-
terminus is indeed an important factor, but other poorly
understood factors are also important.
In our previous work on the cleavage of diribonucleoside
monophosphates CpA, GpA, CpG, CpC, ApG, GpG, GpU,
UpG, and UpU we found that rate enhancements relative to
background brought about by 3-Cu2 and 5-Cu3 (1 mM
catalyst, water solution, pH 7.0, 50 °C) were on the order of
104-fold, largely independent of the base identity, with the
exception of UpU and UpG, for which rate enhancements were
1 order of magnitude higher. Selectivity in the reactions of
dinucleoside monophosphates containing uracyl as the 5′-
nucleobase was explained in terms of a deprotonated uracyl
moiety acting as an anchor site for the metal catalyst.18
Comparison with the data from the present work reveals
completely different selectivity patterns. The highly reactive
CpA and CpC pairs in the oligoribonucleotides display no
enhanced reactivity as dinucleoside monophosphates. Moreover,
the UpG bond in 7 is cleaved quite slowly, whereas the UpU
bond in 8 is not cleaved at all. These observations clearly
indicate that the modes of interaction of the copper(II) com-
plexes with a 5′-labeled oligoribonucleotide and with a diribo-
nucleoside monophosphate are markedly different, which is not
surprising, inter alia, in view of the presence of a doubly charged
binding site in the former.
On the basis of the simple view that binding of a metal cation
to a doubly charged phosphate is expected to be stronger than
to a singly charged phosphate, we speculate that the primary
binding interaction between catalyst and oligoribonucleotide
involves one of the metal centers and the 32P-radiolabeled 5′-
(19) Kierzek, R. Methods Enzymol. 2001, 341, 657-675.
(20) Beloglazova, N. G.; Mironova, N. L.; Konevets, D. A.; Petyuk, V. A.;
Sil’nikov, V. N.; Vlassov, V. V.; Zenkova, M. A. Mol. Biol. 2002, 36,
1068-1073.
(21) Koval’ov, N.; Kuznetsova, I.; Burakova, E.; Sil’nikov, V.; Zenkova, M.;
Vlassov, V. Nucleosides, Nucleotides Nucleic Acids 2004, 23, 977-981.
(18) Enhanced cleavage rates of uracyl-containing dinucleoside monophopshates
by dinuclear zinc(II) complexes of [12]aneN3-based ditopic ligands were
explained along the same lines. See ref 6b.
9
12518 J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007

Synthetic Cu(II)−Nucleases for Oligo-RNA Cleavage
A R T I C L E S
Scheme 4. Catalysis by a Dinuclear Copper(II)-Calixarene
Complex in the Cleavage of a Hexaribonucleotide Labeled at the
5′-Terminus with a 32P Phosphate Dianion
such to unambiguously support the operation of trimetallic
catalysis.
Interestingly, the observed CpA selectivity, which resembles
that shown by natural RNases A, was obtained without the use
of antisense oligonucleotides, generally employed to develop
site-directed RNA cleaving agents. Surprisingly, this selectivity
is completely different from that previously observed in the
cleavage of diribonucleoside monophosphates by the same
complexes, where a high preference for the cleavage of UpU
and UpG sequences was observed.
The reported results may also have a general significance
for the entire class of synthetic metallonucleases because the
base selectivity shown by some of them in the cleavage of
dinucleoside monophosphates, usually ascribed to a preferred
binding of the metal ion to a specific nucleobase, may change
in favor of the intrinsically most reactive NpN′ bonds when using
oligoribonucleotide or RNA substrates having a terminal doubly
charged phosphate group. In these cases, the catalyst-phosphate
interactions and local structural changes may become the
dominant factors.
terminal phosphate dianion, as schematically shown in I
(Scheme 4). One of the possible productive complexes (e.g.,
II) is depicted as a dinuclear metallomacrocycle arising from
intramolecular (intracomplex) binding of one of the singly
charged phosphates of the nucleotide backbone with the other
metal center. A similar mechanism was proposed by Kuusela
et al.22 for the zinc(II)-promoted hydrolysis of oligonucleotides
bearing a phosphate dianion at the 3′-terminus, where the
divalent metal ion was suggested to bridge two phosphate groups
that are several nucleotidic units apart.
Experimental Section
Materials and Methods. Azacrown ligands 1-5 were available from
previous investigations.6a Oligoribonucleotides 6-12 were purchased
from Dharmacon Research (Lafayette, CO) and 32P-5′-end-labeled with
T4 polynucleotide kinase according to standard procedures.25 Concen-
trations of radioactive oligonucleotides were determined from specific
activities, while concentrations of solutions of nonradioactive samples
were determined from absorbance readings at 260 nm and extinction
coefficients were estimated by the nearest-neighbor method.26 Cleavage
experiments were carried out in water (50 µM SDS and 15 mM PIPES
( NaOH, adjusted to pH 7.4, 50 °C), unless otherwise stated. The ionic
strength was adjusted to 0.07 M with NaCl. Solutions of CuCl2 were
used immediately after preparation to minimize formation of insoluble
hydroxides. The aqueous solutions were prepared using deionized
(Millipore) and sterilized distilled water.
Since our reactions have been carried out under conditions
approaching subsaturation,17 the rate of cleavage of a given bond
is a function of both the stability of the productive complex (K′
× K′′) and of its reactivity (kcat). Other things being equal, the
quantity K′′ is expected to decrease on increasing the ring size23
of the metallomacrocycle intermediate II,24 which is well in
keeping with the finding that scissile bonds close to the 5′-
terminal are (usually) cleaved more easily.
Kinetic Measurements. Cleavage experiments were carried out in
sterilized polypropylene micro test tubes stored inside metal boxes that
were immersed in a thermostated water bath. Copper(II) complexes of
ligands 1-5 were formed in situ by addition to the buffer (15 mM
PIPES in water, pH 7.4) of a concentrated ethanol solution of the ligand
and of the calculated stoichiometric amount of the CuCl2 aqueous
solution. The resulting solution was left at room temperature overnight
to allow quantitative formation of the metal complex.27 Reactions were
started by addition of the aqueous solution of the substrate, previously
heated at 90 °C for 2 min (to disrupt potential aggregates) and
equilibrated for 10 min at the reaction temperature. The total volume
of the reaction mixture was 200 µL. At proper time intervals, aliquots
(25 µL) of the reaction mixtures were withdrawn and added to 30 µL
of CH2Cl2. After the addition of 2.5 µL of a 0.5 M EDTA solution
(final concentration 50 mM), the solutions were briefly shaken with a
Vortex mixer and centrifuged. The aqueous phase was separated and
quenched by addition of formamide and excess EDTA. Radiolabeled
substrates and products were separated on 7 M urea/20% polyacryla-
mide gels. The dried gels were exposed to a Kodak phosphor screen
that was analyzed by a Personal Molecular Imager. Quantitative analysis
was carried out using the software Quantity-One (Bio-Rad). Nonlinear
To summarize, di- and trinuclear azacrown-copper(II) com-
plexes implanted on the upper rim of the calix[4]arene scaffold
cleave single-stranded RNA oligonucleotides with high ef-
ficiency and selectivity under pH and temperature conditions
close to physiological. An extensive kinetic investigation of
the cleavage reactions afforded pseudo-first-order rate con-
stants for the cleavage of the most labile phosphodiester bonds.
With few exceptions, substantial levels of cooperation between
metal centers were observed in the reactions promoted by
dinuclear complexes, irrespective of whether the substitution
pattern was 1,2-vicinal or 1,3-distal. However, in no case was
the contribution from a third ligated metal ion at the upper rim
(22) (a) Kuusela, S.; Azhayev, A.; Guzaev, A.; Lo¨nnberg, H. J. Chem. Soc.,
Perkin Trans. 2 1995, 1197-1202. (b) Kuusela, S.; Guzaev, A.; Lo¨nnberg,
H. J. Chem. Soc., Perkin Trans. 2 1996, 1895-1899. (c) Kuusela, S.;
Lo¨nnberg, H. Nucleosides, Nucleotides Nucleic Acids 1998, 17, 2417-
2427.
(23) (a) Mandolini, L. AdV. Phys. Org. Chem. 1986, 22, 1-111. (b) Galli, C.;
Mandolini, L. Eur. J. Org. Chem. 2000, 3117-3125.
(24) Such a decrease is predicted to be substantial, but not dramatic. The number
of rotatable single bonds, including copper(II)-phosphate coordinative
bonds, which are involved in the closure of the smallest dinuclear
metallomacrocycle is 11, and each additional nucleotide unit contributes 5
additional rotatable bonds. On a purely entropic basis, the EM for closure
of a strainless macrocycle decreases from the value of 0.37 for the 11-
rotor case to one of 0.037 M for the 31-rotor case. A further drop to 0.021
M is calculated for the 46-rotor case, which corresponds to the cleavage
of the γ-CpA bond in the 17-mer 12. See ref 23.
(25) Maniatis, T.; Fritsch, E. F.; Sambrook, J. Molecular Cloning: A Laboratory
Manual; Cold Spring Harbor: New York, 1982.
(26) Cantor, C. R.; Warshaw, M. M.; Shapiro, H. Biopolymers 1970, 9, 1059-
1077.
(27) Riedo, T. J.; Kaden, T. A. HelV. Chim. Acta 1979, 62, 1089-1096.
9
J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007 12519

A R T I C L E S
Cacciapaglia et al.
least-squares calculation of kinetic data was carried out using the
program SigmaPlot 2002 for Windows, version 8.0 (SPSS Inc.).
Supporting Information Available: Kinetic treatment of rate
data according to integrated rate equations derived from
Schemes 2 and 3. This material is available free of charge via
the Internet at http://pubs.ac.org.
Acknowledgment. Financial contribution from MUR PRIN
2006 is acknowledged for the work carried out in Parma and
Roma.
JA0737366
9
12520 J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007