Rapid hydrolytic cleavage of the mRNA model compound HPNP by glycine based macrocyclic lanthanide ribonuclease mimics

Article abstract of DOI:10.1039/b205349g

The lanthanide ion based macrocyclic complexes 1·Ln mimic the hydrophobic nature of ribonucleases, where the lanthanide ions induce the formation of a hydrophobic cavity for 1, giving rise to a large order of magnitude enhancement in the hydrolytic cleavage of HPNP.

Full text of DOI:10.1039/b205349g

Rapid hydrolytic cleavage of the mRNA model compound HPNP by
glycine based macrocyclic lanthanide ribonuclease mimics†
a
b
c
b
Thorfinnur Gunnlaugsson,* R. Jeremy H. Davies, Mark Nieuwenhuyzen, Clarke S. Stevenson,
a
a
Romain Viguier and Sinead Mulready
Department of Chemistry, Trinity College Dublin, Dublin 2, Ireland. E-mail: gunnlaut@tcd.ie; Fax: 00
53 1 671 2826; Tel: 00 353 1 608 3459
School of Biology and Biochemistry, Queen’s University of Belfast, Belfast, Northern Ireland, UK BT9 7BL
School of Chemistry, Queen’s University of Belfast, Belfast, Northern Ireland, UK BT9 5AG
a
3
b
c
Received (in Cambridge, UK) 6th June 2002, Accepted 5th August 2002
First published as an Advance Article on the web 22nd August 2002
The lanthanide ion based macrocyclic complexes 1·Ln
mimic the hydrophobic nature of ribonucleases, where the
lanthanide ions induce the formation of a hydrophobic
cavity for 1, giving rise to a large order of magnitude
enhancement in the hydrolytic cleavage of HPNP.
1·Ln, as ribonuclease mimics. We predicted that the tetraamide-
derived cyclen complex, with its concave structure and central
lanthanide ion, would be an ideal synthetic platform for the
integration of amino acid cofactors. The ligand, 1, was
synthesized in a two step synthesis in 68% overall yield (See
1
ESI†). ‡ The H NMR in CDCl
3
indicated C
4
symmetry with
Currently, there is a great interest in the development of robust
catalytic systems that can effectively mimic important enzy-
matic reactions under physiological conditions. The develop-
only five resonances being observed. The corresponding
cationic lanthanide complexes 1·La, 1·Ce, 1·Pr, 1·Nd, 1·Eu,
1·Gd, 1·Tb and 1·Lu were formed by reacting an equimolar
1
ment of such catalysts with the aim of achieving fast and site-
selective hydrolytic cleavage of the phosphate ester bonds of
RNA over DNA is of particular interest.1 This is of prime
importance for the development of novel antisense drugs,3
ribonuclease and ribozyme mimics and gene technology.
Ribonucleases often possess one or more metal ion centers at
amount of 1 with the appropriate Ln(SO
dry CH CN or MeOH. We were able to grow crystals of 1·Eu
(1·EuK
ESI†) from methanol–CHCl
3
CF
3 3
) salt in refluxing
3
,2
+
2
(CF SO ) , the K is bonding to the glycine ester,
3 3 5
3
suitable for crystallographic
investigation (See ESI for structure†).§ 10 This showed the Eu
ion placed in the centre of the cavity, coordinated by the four
nitrogens of cyclen and the oxygens of the carboxyamides
3+
1
their active site, which in combination with basic amino acids
…
…
such as histidine and lysine accelerate their reactions dramat-
ically. This has encouraged the development of synthetic
(average N Eu and O Eu bond lengths were 2.661(4) and
2.369(3) Å, respectively). A single triflate molecule occupies
the ninth coordination site, giving an overall monocapped
4
ribonucleases where transition and lanthanide metal ion
complexes have been employed.1
,2,3,5
Such complexes accel-
square antiprism (CSAP) geometry, with enantiomeric con-
11
erate the rate of hydrolysis via Lewis acid or nucleophilic
activation, often through the synergic action of several metal
centres.2 However, the use of cofactors such as amino acids,
in conjunction with these metal centres has been less explored.6
We are interested in developing physiologically stable ribonu-
clease mimics based on the use of 1,4,7,10-tetraazacyclodode-
cane (cyclen) derived lanthanide complexes where the cofactors
are integrated covalently into the cyclen framework as pendant
formation for the ring NC-CN as dddd, and the pendant arms
11
arranged in an anticlockwise, L fashion. Significant to our
,4,5
design, the four GlyGly moieties (one of which has its amino
terminus as a part of the macrocyclic ring) form the walls of a
1
cavity. Importantly, H NMR in CD
3
OD indicated that the
CSAP geometry was also the major isomer ( > 95%) in
solution.11
All HPNP cleavage studies were carried out at 37 °C and at
pH 7.4 in 50 mM HEPES buffer (in 96+4 H O–MeOH) with an
2
7
arms. Herein we discuss the synthesis, physical characteristics,
and cleavage evaluation towards the RNA model compound
equimolar amount of catalyst (0.173 mM). The HPNP hydroly-
sis was monitored by the appearance of a new absorption band
at 400 nm, corresponding to the formation of the p-ni-
trophenolate anion (Scheme 1). The cleavage of HPNP by 1·Ln
was in all cases much faster than anticipated, given the fact that
glycine lacks the extra cooperative sites found in basic amino
acids such as lysine and histidine. Moreover, since the glycine
esters extend the size of the cavity, it might have been expected
to inhibit the approach of the complex. For 1·La the rate of
hydrolysis was found to display pseudo first order kinetics with
2-hydroxypropyl p-nitrophenyl phosphate (HPNP) of 1·Ln,
which incorporates rather simple pseudo ‘GlyGly’ cofactors,
and show that they give rise to a remarkable enhancement in the
rate of hydrolysis of HPNP.
Inspired by the work of Morrow and coworkers who used a
2
2
21 8
La(III) complex to hydrolyse HPNP with k = 5.8 3 10
h ,
and from our recent investigations into the use of amide based
cyclen lanthanide complexes as luminescent switches and
sensors, we developed 1 and its lanthanide ion complexes
9
2
1
1
⁄
2
k = 0.41 h and t of 1.7 h, Table 1. This is a remarkable
3
400-fold rate enhancement (kobs) compared to the uncatalyzed
reaction. Moreover, this is seven times faster than the kinetics
observed for Morrow’s La3 complex. Table 1 summarizes
these results for all the 1·Ln complexes. It is particularly
important to note that 1·Eu shows a significant rate enhance-
+
2
1
ment with k of 0.15 h , and kobs = 1250. In fact, all the
complexes showed remarkable cleavage ability and a clear trend
emerges from Table 1: the larger ions are more potent than the
smaller. This can in part be explained by higher coordination
†
Electronic supplementary information (ESI) available: full experimental
details, Fig. S1–S6. See http://www.rsc.org/suppdata/cc/b2/b205349g/
Scheme 1 The mechanism for the hydrolytic cleavage of HPNP.
2
136
CHEM. COMMUN., 2002, 2136–2137
This journal is © The Royal Society of Chemistry 2002

Table 1 Results of the hydrolysis of HPNP using Ln·1
6725-fold enhancement, and almost twice that seen at pH 7.4.
This also implies that the most active form of the catalyst has at
least one hydroxy ion bound to the lanthanide ion centre. It is
also possible that at higher pH (pH > 8.5), this hydroxy group
binds too strongly to the phosphate and this in turn could inhibit
further coordination of HPNP. Preliminary investigation has
also shown that most of the complexes efficiently cleaved a 24
mer-mRNA sequence from the GAG-HIV gene at pH 7.4 and 37
°C after 4 h of incubation (See ESI†). Complexes such as 1·La
and 1·Eu induce cleavage at every base pair, whereas upon
incubation with 1, no cleavage is observed, indicating the vital
role of the Lewis acid centre in the hydrolysis (this cleavage was
not quantified, ESI†). We are currently developing analogue
compounds by incorporating other cofactors into the cyclen
structure and incorporating these complexes into oligonucleo-
tides as potential antisense agents.
Complexa
k/h21bcg
t
1
/h
k
obs
d
⁄
2
1
1
1
1
1
1
1
1
·Lae
·Ce
·Pr
0.41
0.37
0.30
0.20
0.15
0.12
0.0935
0.072
1.69
1.87
2.31
3.46
4.62
5.77
7.41
9.63
3417
3083
2500
1667
1250
1000
779
·Nd
f
·Eu
·Ga
f
·Tb
·Lu
600
a
Measured using an Agilent 8453 spectrophotometer fitted to circulating
temperature controlled water bath, and water driven mechanical stirring, in
b
5
0 mM HEPES buffer, at pH 7.4 and at 37 °C. Average over three
c
measurement and three half-lifetimes. k values were determined by fitting
the data to first order rate kinetics using Biochemical Analysis Software for
d
Agilent ChemStation. Errors are within ±10%. These enhancement factors
In conclusion, the various lanthanide complexes 1·Ln show
significant rate enhancement in the cleavage of HPNP. To the
best of our knowledge, 1·La displays one of the largest rate
accelerations observed for HPNP by trivalent, non-redoxactive
lanthanide complexes. We thank Enterprise Ireland, Dublin
Corporation (postgraduate scolarship to S. M.), and TCD for
financial support, Dr Hazel M. Moncrieff for helpful discussion
and Dr John E. O’Brien for assisting with NMR.
are obtained from the ratio of the catalyst vs. the uncatalysed reaction using
ucat = 0.00012 h2 (R. Breslow and D-L Huang, Proc. Natl. Acad. Sci.
1
k
e
USA, 1991, 88, 4080). The 1·La complex did not cleave the DNA model
compound bis(p-nitrophenyl) phosphate (BNPP) over 24 h. We were
unable to determine the hydration number q accurately. 20% EDTA
f
g
affected the rate of hydrolysis slightly but we believe that this is due to
binding of one or more of the EDTA carboxylic acids to the metal ion
complex rather than metal extraction. 1·Eu and 1·Tb were stable to
competitive Cu(II) (sulfate) exchange in water over a week at pH 6.5
(measured by UV/VIS). We are currently investigating the stability of the
other complexes.
Notes and references
+
‡
1: Found for C28
H
49
N
8
O
12 (MH ): 689.3470. Calc.: 689.3469. Found for
KC28
14.9.
49 8
H N O12: C 46.21; H 6.65; N 15.40; Calc.: C 46.64, H 6.88; N
number requirements for the larger ions that are fulfilled by the
additional water molecules,7 which are important for both
inner and outer sphere catalytic activation modes. In contrast,
a Cu(II) complex of 1 was found to be inactive, i.e. no
measurable cleavage of HPNP was observed over 24 h.
Potentiometric pH titration for 1·La revealed that ca. 1.5–2 base
equivalents were needed to deprotonate the water molecules of
,11
§
Data were collected on a Bruker SMART diffractometer with graphite
1,2
monochromated Mo-Ka radiation using omega/phi scans. The structure
was solved using direct methods and refined with the SHELXTL program
package. The two triflate anions are disordered; firstly the anion associated
with the K centres has been modelled over two sites with the major
component being 76(1)% occupancy. The second anion is bound to the Eu
centre and is disordered about a four-fold symmetry element and has been
modelled as 25% occupancy for each position. Crystal data for
1
·La. We estimated these pK
respectively, but we were unable to determine them accurately
see ESI†) even when the titrations were repeated at different
concentrations. In contrast, a single pK of 7.38 was measured
a
s to be ca. 8.2 and 8.5,
33 48 8 15 5 2
C H N F S K Eu: M = 1664.25, tetragonal, space group P4/ncc, a = b
(
3
=
17.2721(8), c = 21.0475(14) Å, U = 6279.0(6) Å , Z = 4, m = 1.425
a
2
1
23
c
mm , F(000) = 3336, D = 1.761 g cm , Rint = 0.0408, transmission
very accurately for 1·Eu. For 1·La, we predict that one of the
two water molecules can be rapidly displaced upon binding to
the phosphate ester, revealing extra binding sites over that of
range (max., min.) = 1.000, 0.804, crystal size = 0.33 3 0.28 3 0.21 mm.
A total of 51944 reflections were measured for the angle range 3 < 2q < 58
and 3844 independent reflections were used in the refinement. The final
parameters were wR2 = 0.1366 and R1 = 0.0479 [I > 2sI].CCDC 177867.
See http://www.rsc.org/suppdata/cc/b2/b205349g/ for crystallographic data
in CIF or other electronic format.
1
·Eu. The second water molecule (or bound hydroxide) is then
able to carry out a nucleophilic reaction on the phosphodie-
ster.2 Preliminary P NMR binding studies in buffered H
using 1·Eu and 1·La, and diethylphosphate
(CH O) PO ] (DEP), which lacks the 2-hydroxy group,
,4
31
O
2
2
2
1 A. J. Kirby, Angew. Chem., Int. Ed. Engl., 1996, 35, 707; D. E. Wilcox,
Chem. Rev., 1996, 96, 2435; J. Chin, Acc. Chem. Res., 1991, 24, 145.
[
3
CH
2
2
showed that 1·Eu binds more strongly to DEP than 1·La which
needed 30 equivalents of DEP vs. ca. one for 1·Eu to obtain
saturation for the 31P signal. We predict that similar binding
preferences can be expected for HPNP.
2
M. Komiyama and J. Sumaoka, Curr. Opin. Chem. Biol., 1998, 2, 751;
D. M. Perreault and E. V. Anslyn, Angew. Chem., Int. Ed. Engl., 1997,
3
6, 432.
B. N. Trawick, A. T. Daniher and J. K. Bashkin, Chem. Rev., 1998, 98,
39.
3
It is remarkable that 1·Ln shows such a high rate enhance-
ment despite the fact that the Lewis acid centre is more shielded
from the solvent environment due to the steric effect enforced
by the glycine esters. It is our prediction that these rate
enhancements are due to hydrophobic effects caused by the
arrangement of the amino esters around the metal ion centre,
giving rise to the formation of a hydrophobic pocket around the
ion. This possibly favours the formation of stronger interactions
between the ion and the phosphodiester. Similar observations
have previously been seen for example in Collman’s ‘picket
fence’ porphyrins,12 and in mimicking the active site of
9
4 P. Molenveld, J. F. J. Engbersen and D. M. Reinhoudt, Chem. Soc. Rev.,
2000, 29, 75; P. Gómez-Tagle and A. K. Yatsimirsky, Inorg. Chem.,
2001, 40, 3786.
5
N. H. Williams, B. Takasaki, M. Wall and J. Chin, Acc. Chem. Res.,
999, 32, 485.
1
6
7
A. Roigh and H.-J. Schneider, Eur. J. Org. Chem., 2001, 205.
S. Aime, A. Barge, M. Botta, J. A. K. Howard, R. Kataky, M. P. Lowe,
J. M. Moloney, D. Parker and A. S. de Sousa, Chem. Commun., 1999,
1
047.
8
S. Amin, J. R. Morrow, C. H. Lake and M. R. Churchill, Angew. Chem.,
Int. Ed. Engl., 1994, 33, 773.
13
carbonic anhydrase by Boxwell and Walton. We are currently
investigating these features in more detail. Secondly, our
assumption that HPNP binds more weakly to 1·La than 1·Eu
will contribute to the release of the product from the cavity of
the lanthanide complexes.
9 T. Gunnlaugsson, D. A. Mac Dónaill and D. Parker, J. Am. Chem. Soc.,
2001, 123, 12866; T. Gunnlaugsson, Tetrahedron Lett., 2001, 42, 8901;
T. Gunnlaugsson, D. A. Mac Dónaill and D. Parker, Chem. Commun.,
2
000, 93.
1
1
1
0 SAINT-NT, Brüker AXS Madison, Wisconsin, 1998; G. M. Sheldrick,
University of Göttingen, Göttingen, Germany, 1998.
1 P. Cravan, J. J. Ellison, T. J. McMurry and R. B. Lauffer, Chem. Rev.,
Although the complexes are highly potent at pH 7.4, the pH-
rate profile for 1·La (Fig. S2 in ESI†) shows that the activity (at
1
999, 99, 2293.
3
7 °C) is strongly influenced by pH, with an optimum rate at ca.
2 J. P. Collman, R. R. Gagne, C. A. Reed, T. R. Halbert, G. Lang and W.
2
1
pH 8.5 of k = 0.807 h . This correlates well with the pH
titration of 1·La discussed earlier. This is a remarkable
T. Robinson, J. Am. Chem. Soc., 1975, 97, 1427–1439.
13 C. J. Boxwell and P. H. Walton, Chem. Commun., 1999, 1647–1648.
CHEM. COMMUN., 2002, 2136–2137
2137