Preorganized bis-zinc phosphodiester cleavage catalysts possessing natural ligands: A lesson pertinent to bimetallic artificial enzymes

Article abstract of DOI:10.1002/chem.200390082

Two preorganized bis-zinc receptors (2 and 3) were synthesized wherein the metals were ligated with ligands present in natural phosphodiesterases: imidazoles and carboxylates. The intrametallic distance is near 45 A, that found in natural nucleases and ot

Full text of DOI:10.1002/chem.200390082

FULL PAPER
Preorganized Bis-Zinc Phosphodiester Cleavage Catalysts Possessing Natural
Ligands: A Lesson Pertinent to Bimetallic Artificial Enzymes
Karen Worm, Feiya Chu, Kazunari Matsumoto, Michael D. Best, Vincent Lynch, and
Eric V. Anslyn*[a]
Abstract: Two preorganized bis-zinc re-
ceptors (2 and 3) were synthesized
wherein the metals were ligated with
ligands present in natural phosphodies-
terases: imidazoles and carboxylates.
The intrametallic distance is near
4.5 ä, that found in natural nucleases
and other successful artificial nucleases.
With only two imidazoles (2), the zinc
binding affinities were not high enough
to achieve cooperativity. Yet, with a
third ligand, a carboxylate (3), coopera-
tivity was found in the cleavage of
HPNPP. The preorganization of 3 was
achieved using a ™steric gearing∫ strat-
egy. The enhancement was 80-fold for
cooperation between the two metals
relative to a mono-metallic analogue
(5). However, there was no observable
enhancement in the hydrolysis ofRNA
using 3 relative to 5. Therefore, we
conclude that placing two zinc atoms
that are ligated with natural ligands at
the appropriate distance for catalysis is
not sufficient to enhance the cleavage of
RNA, but is successful for activated
RNA substrate mimics.
Keywords: enzyme models ¥ hydro-
lysis ¥ phosphoester ¥ receptors
and the use ofthese ligands in the development ofa practical
synthetic nuclease is certainly appropriate.
Introduction
A large number ofbimetallic catalysts have been studied for
the enhancement ofphosphate ester cleavage and/or hydrol-
ysis.[1] Many ofthese catalysts possess linkers between the
ligands for the individual metals that have degrees of
conformational freedom.[2] When rigid linkers are used, one
can confidently predict the distances between the metals
during the catalytic event. Correlations between metal
distances and catalytic activity have been made. To date, the
evidence suggests that the best catalysts are those where the
distance between the metal centers is nearly the same as that
found in bimetallic enzymes.[3]
Alternatively, we are developing a series ofcatalysts
modeled after and possessing similar intrametallic distances
as 1, but that use the same ligands and metals as nucleases:
carboxylates/imidazoles and zinc respectively. In this manner,
we seek to test iftwo zinc atoms ligated with natural ligands
and possessing intrametallic distances similar to the enzymes
and 1 would also show good catalysis. In this manner, we are
mimicking a known artificial enzyme, not a particular natural
enzyme. Here, we report that simply placing the ZnII ions
ligated with natural ligands in a geometry mimicking 1 does
not give an efficient catalyst for RNA hydrolysis. Therefore,
Avery good example is from the Chin laboratory (1).[4] This
the appropriate metal metal distance may be necessary, but is
À
À
catalyst has a Cu Cu distance near 4.5 ä. Although 1 is
not sufficient, for good catalysis.
among the best synthetic catalysts in terms ofrate enhance-
ments for the cleavage/transesterification of RNA, the
catalyst does not employ the same ligands and metals as in
natural enzymes. In fact, a very large fraction of all artificial
enzymes for phosphoester hydrolysis contain ligands other
than those used by nature, such as amines, pyridines, and
alkoxides.[1, 2] This is often an important synthetic advantage,
H
H
N
N
+2
+2
N
N
Cu
Cu
H
H
N
N
1
[a] Prof. E. V. Anslyn, Prof. K. Worm, F. Chu,
K. Matsumoto, M. D. Best, V. Lynch
Department ofChemistry and Biochemistry
The University ofTexas at Austin
Austin, TX 78712 (USA)
Results and Discussion
Design criteria: As mimics ofcatalyst 1 but using ZnII and
natural ligands, compounds 2 and 3 were designed. Structure 2
uses fused rings in a tricyclic scaffold to place four imidazole
Fax : (‡1)512471-7791
E-mail: anslyn@ccwf.cc.utexas.edu
Chem. Eur. J. 2003, 9, No. 3
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0903-0741 $ 20.00+.50/0
741

E. V. Anslyn et al.
FULL PAPER
boxylate ligand via the linker, orienting the carboxylate
toward the zinc center.
N
N
O
O
O
O
N
N
N
N
N
N
N
N
N N
ZnII
N N
ZnII
Synthesis: The syntheses of 2 and 3 are shown in Schemes 1
and 2, respectively. The synthesis of 2 commences from
methyl methoxyacetoacetate (6) and formaldehyde. Cycliza-
tion of 7 to 8 occurs upon treatment with ammonia, followed
by aromatization using DDQ to give 9. Addition ofofur
equivalents oflithium methylimidazole ( 10) results in 11. The
methyl groups are removed with BBr3 giving 12, which can be
cyclized in the presence ofphosphoric acid to give apo- 2.
Stirring with ZnCl2 in water yields 2.
N N
N N
O
O
ZnII
ZnII
O
O
3
2
4.5 °A
all natural ligands
steric gearing
N
CH3
CH
3H3C
OH
N
O
N
N
N
N N
ZnII
4
N N
ZnII
Scheme 2 shows the synthetic route used to obtain 3. The
synthesis of 3 starts with the combination of 13[6] with four
equivalents oflithium methylimidazole ( 10) to give 14.
Alkylation of 14 with 15 results in 16, followed by hydrolysis
ofthe cyano groups yields apo- 3, which is simply allowed to
stir with ZnCl2 in water to give 3. A very similar route was
used in the synthesis ofthe mono-zinc
O
O
5
ligands in such a manner that the two zinc atoms are located
approximately 4.5 ä apart. The fusing of the rings in the
scaffold leads to a highly preorganized binding cleft, which
positions the two metal centers at the required distances.
Compound 3 also places two zinc atoms at the distances
required for our study, but is more flexible. This compound
has two imidazoles and one carboxylate ligand for each metal
center, and therefore has higher affinity constants for the zinc
metals (as described below). In
O
O
N
receptor 5 (Scheme 3).
O
O
As described below, crystals suitable
for crystallography could not be ob-
tained for 2 and 3, and therefore we
sought another derivative. Specifically,
as described below, compound 20 was
N
N
N
N
N N
N N
2+
2+
Zn
Zn
20
OMe
OO
OMe
OMe
OMe
OMe
order to introduce a third li-
gand to each zinc, as in 3
compared to 2, the scaffold in
2 could not be used. To incor-
porate another factor for pre-
organizing the zinc metals,
thereby obtaining the intrame-
H
N
2
O
CH2O
Et2NH
NH3
MeO
OMe
OMe
MeO
MeO
OMe
O
O
O
O
O
7
8
6
OMe
N
OMe
OMe
N
N
N
4
Li
10
tallic distance required for our
analysis, we turned to a method
we have been exploring for the
creation ofreceptors and cav-
ities, referred to as ™steric gear-
ing∫.[5] Here, alternating steric
bulk is used to bias conforma-
tional preferences to those that
possess the desired geometries.
In the case of 3, the lowest
energy conformation found by
molecular mechanics is with the
methyl groups on the imida-
zoles placed above and below
the methyl groups on the cen-
tral pyridine spacer (see the
analogous methyl groups ex-
plicity drawn in structure 5).
This ™meshing∫ ofthe methyl
groups was predicted to impart
the desired conformation to the
overall structure, leading to the
HO
N
OH
N
15% BBr3
MeO
OMe
DDQ
N
N
O
O
N N
N N
11
9
OH
OH
N
N
O
O
ZnII
3% H3PO4
HO
N
OH
N
N
N
2
N
N
N
N
N N
N N
N N
N N
apo-2
12
Scheme 1. Synthetic route to 2.
N
2
N
Br
N
4
Li
HO
N
OH
N
N
O
O
N
15
N
CN
10
O
O
N N
N N
13
14
N
N
O
O
N
ZnII
O
O
NaOH
iPrOH/H2O
N
N
N
3
N
N
N
N
À
required Zn Zn distance. Fur-
N N
N N
N N
N N
OH
HO
ther, the methyl groups on the
pyridine spacer should impart a
conformational bias to the car-
NC
CN
O
O
apo-3
16
Scheme 2. Synthetic route to 3.
742
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0903-0742 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 3

Bimetallic Artificial Enzymes
741 747
zinc to 4 is 9 Â 104 mÀ1.[6] Due to the electrostatic repulsion
between the two adjacent zinc atoms in 2, and the fact that
only two imidazoles ligate each zinc, the second zinc binding
constant to 2 must be significantly lower. The placement of
only one zinc in apo-2 then likely leads to a dimer in solution,
as found for 20 in the solid state. The dimer possesses
coordinately saturated zinc centers, and would be inactive as a
catalyst for RNA hydrolysis.
As mentioned above, to improve upon the design of 2,
compound 3 was developed. Now three ligands, two imida-
zoles, and one carboxylate, ligate each zinc. This increases the
affinity for zinc, and impedes dimerization because the
coordination number ofeach
N
Br
4
Li
10
OH
N
15
N
O
N
O
17
CN
N N
18
O
O
NaOH
ZnII
N
N
N
5
N
N N
iPrOH/H2O
N N
NC
19
O
H
O
apo-5
Scheme 3. Synthetic route to 5.
metal is greater. An affinity
O
O
1)
N
N
N
constant of7.2 Â 106 mÀ1 for zinc
O
O
4
Li
O
N
O
with 5 was determined using a
O
O
ZnII
N
H3PO4
N
N
10
2) H3PO4
N
20
previously reported competi-
tion assay,[8] and we measured
a Zn-bound water pKa of8.3 for
that complex. Given the near
100-fold enhancement in zinc
N N
N N
O
O
21
apo-20
Scheme 4. Synthetic route to 20.
crystalline. The synthesis ofcompound 20 is shown in
Scheme 4. Addition ofofur equivalents of 10 to 21 gave
binding to apo-5 relative to apo-4, we felt that the study of 3
was warranted.
apo-20 in 30% yield, along with the trans-isomer in 65%
yield. Again, simply stirring with ZnCl2 gives 20.
The catalytic activities of 3 and 5 were examined using
HPNPP and RNA. Compound 5 behaved similar to the
Structural analyses: We could not obtain X-ray quality
crystals for either 2 or 3, and hence we synthesized 20, which
was found to be crystalline. The crystal structure of 20 is
shown in Figure 1A. It confirms the predicted distance
between the zinc atoms in the kinds ofstructures reported
herein, with a 3.98 ä separation, where a fluoride ligand
bridges the two zinc atoms. However, we find that two
equivalents ofapo- 20 dimerize through the bridging oftwo
bound zinc atoms, as shown in Figure 1B . This occurs
although two full equivalents of ZnCl2 per apo-20 were added
to the solution. This dimerization is a likely a structural motif
for 2 also, and could in part explain the low activity of 2 in the
hydrolysis ofphosphoesters (see below).
In the design criteria, the possible steric gearing ofthe
methyl groups in 3 (shown in 5) was discussed as a means to
preorganize the receptor and hold the metals in proximity.
The proposed ™meshing∫ ofthe methyl groups was confirmed
with an NOE study on 3, which showed a 15% enhancement
ofeither set ofmethyl groups when the corresponding
neighboring methyls were irradiated. This is best explained
by the methyl groups being placed in the geometry for 3,
shown above, where the groups are near van der Waals
distances.
Catalytic activity: We first analyzed compound 2 for catalysis
ofthe cleavage of2-hydroxypropyl- p-nitrophenylphosphate
(HPNPP). Relative to control structure 4, whose catalytic
activity we have previously reported,[7] there was no rate
enhancement in the cleavage ofHPNPP or RNA, but instead
even lower activity. Titrations ofZnCl 2 into solutions ofapo- 2,
followed by UV/Vis spectroscopy, indicated that a second zinc
is not bound in 2 to any significant extent at low millimolar
concentrations ofreceptor and zinc. The binding constant ofa
Figure 1. View ofthe Zn complex 20 showing a partial atom labeling
scheme. Displacement ellipsoids are scaled to 30% probability level.
A) Ortep diagram showing one unit of 20. B) Ortep diagram showing the
dimer found in the crystal structure of 20.
Chem. Eur. J. 2003, 9, No. 3
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0903-0743 $ 20.00+.50/0
743

E. V. Anslyn et al.
FULL PAPER
previously reported activity of 4 in the catalysis ofHPNPP
cleavage. We found a pH dependence on the rate constant for
cleavage ofHPNPP indicating general base catalysis (Fig-
ure 2A). A plateau was found above pH 8.7, close to the pKa
ofthe zinc bound water (see above).
N
O
O
N
N
O
N
N
N N
N N
O
O
ZnII
ZnII
OH
O
O
O
P
H
O
O
O
O2N
Figure 3. Hypothetical mechanism for the transesterification of HPNPP.
further confirms the preorganization and promixity of the zinc
ions in 3.
However, we found there is essentially no difference in the
rate ofcleavage ofRNA when 3, 4, or 5 are used. At pH 8.4
(10 mm HEPES), the cleavage ofUpU to give Up and U
shows essentially identical time course plots in initial rate
kinetic measurements. The lack ofactivity of 3 must be due to
an improper orientation in the binding ofRNA to the catalyst
to promote cleavage/transesterification. Therefore, what ever
geometry is achieved upon binding is insufficient to activate
the phosphate ofRNA for subsequent nucleophilic attack.
Conclusion
In summary, although compound 3 possesses similar intra-
metallic distances as bimetallic enzymes and 1, a statement we
can make based upon its rigid and preorganized structure, it is
a significantly worse catalyst than 1 for RNA cleavage. This is
despite the fact that we found it does act as a good bimetallic
general acid/general base catalyst with an activated substrate
(HPNPP). Therefore, other features of enzymes besides just
the metal distance, such as proper orientation ofthe bound
substrate and other activating groups, must also play major
roles in the catalysis. Further, as noted by Chin,[4] compound 1
is likely a good synthetic catalyst not only due to the
Figure 2. Kinetic runs (378C) showing kobs versus pH profiles for 3 and 5 in
methanol/water 1:2 solutions (MeOH added for solubility purposes, 50 mm
tris buffer). All runs are with 0.4 mm HPNPP. A) Compound 3 at 3.8 Â
10À4 m. B) Compound 5 at 2.6 Â 10À3 m.
intrametallic distance, but also due to the higher electro-
II
philicity ofCu
compared with ZnII, and the unnatural
synthetic ligands which enhance metal binding relative to
natural ligands. Our studies support his previous conclusions.
To create rates ofcleavage ofHPNPP with 3 and 5 that are
similar, a more dilute solution of 3 was required because it is
significantly more active. With a dilute solution of 3 (see
caption to Figure 2B), we found similar rates constants, but
now a bell-shaped pH versus rate constant profile was found.
Such bell-shaped curves are common in the cleavage of
phosphoesters, reflecting a combined general-acid/general-
base catalyzed mechanism.[9] The peak ofthe bell-shaped
curve is once again similar to the pKa value ofwater bound to
the zinc atom in 5. Our proposed catalytic step is shown in
Figure 3. The phosphate is activated by either one or two
metals, and a zinc-bound hydroxide acts to deliver the
intramolecular alcohol nucleophile.
Experimental Section
General considerations: 1H and 13C NMR spectra were obtained in CDCl3
or CD3OD used as purchased. NMR spectra were recorded on a Bruker
AC250 and a Varian Unity Inova (500 MHz, NOESY) spectrometer. Low
resolution and high resolution mass spectra were measured with Finnigan
TSQ70 and VG Analytical ZAB2-E instruments. Preparative flash
chromatography was performed on Scientific Absorbents Incorporated
Silica Gel 40 mm. Reversed-phase liquid chromatography was performed
on RP18 modified silica gel 55-105 mm using a Pharmacia LKB-FRAC-100
LC system. Solvents and reagents were purchased from Aldrich, Sigma and
Mallinckrodt and used without purification. Tetrahydrofuran (THF) was
distilled from sodium/benzophenone.
Using the relative concentrations ofcompounds employed
in the studies illustrated in Figure 2, and a factor of 2 due to
the presence oftwo zinc atoms in 3, we find an 80-fold
advantage imparted to the catalysis ofcleavage ofHPNPP
using 3 relative to 5. This enhancement is significant, and
pKa Determination: Potentiometric titrations ofthe ligands apo- 3 and apo-
5 were performed in the presence and absence of ZnII. The chloride salts
(obtained by lyophilizing HCl acidic solutions ofthe ligands) were
dissolved in volumetric standard sodium chloride solution (0.0973m,
744
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0903-0744 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 3

Bimetallic Artificial Enzymes
741 747
Aldrich) and 30% methanol. The base (0.0975m NaOH standard from
Aldrich) was delivered using a programmable Harvard Apparatus syringe
infusion pump 22. The pH readings were taken on an Orion Model 720A
pH meter with an Orion Ross model 8103 combination pH electrode. The
volume oftitrant added, as well as the pH, was recorded during the
experiment with the Graphit program on a PC.
chromatography (100% EtOAc ! EtOAc:MeOH(sat.NH3) 80:20), then
on an ion-exchange column (50 mm NH4Oac, pH 6.8), yielding the title
1
coompound (396 mg, 37%). H NMR (300 MHz, CDCl3): d ˆ 3.24 (s, 6H,
CH2OCH3), 3.33 (s, 12H, NCH3), 4.64 (s, 4H, CH2OCH3), 5.69 (s, 2H, OH),
6.75 (s, 4H, CH CH), 6.82 (s, 4H, CH CH), 7.24 (s, 1H, Ar); 13C{1H} NMR
(75 MHz, CDCl3): d ˆ 34.4, 58.4, 74.9, 75.9, 123.4, 126.1, 135.6, 135.9, 146.8,
154.8; HRMS-CI: m/z: calcd for C27H34N9O4: 548.2734; found: 548.2731
[M‡H] .
ˆ
ˆ
Kinetic measurements on HPNPP: The studies were conducted in 50 mm
Tris buffer, 30% methanol at 378C in a Beckman DU640 spectropho-
tometer fitted with a temperature controller. Readings were taken at
401 nm every 30 min for 72 h. The buffer was prepared with distilled/
deionized water and the pH adjusted with HCl before adding the methanol.
The samples were prepared in 10 mm path length cuvettes which could be
sealed with teflon stoppers. A typical sample would involve 900 mL buffer,
100 mL ligand ‡ ZnII and 500 mL methanol, equilibrated for 30 min at
37 8C followed by addition of 10 mL substrate stock solutions.
‡
2,6-Hydroxymethyl-3,5-bis[bis(1-methyl-2-imidazolyl)hydroxymethyl]pyr-
idine (12): A round bottom flask was charged with 11 (86 mg, 0.16 mmol) in
dry CH2Cl2 (5 mL) The reaction mixture was cooled to À788C and BBr3
(1m in CH2Cl2; 0.47 mL, 0.47 mmol) was then added dropwise. After the
mixture was stirred for an hour, additional BBr3 (0.47 mL) was added. The
mixture was allowed to warm up to room temperature overnight. The
mixture was then quenched with MeOH and then pH was adjusted to 7
using Na2CO3 (aq.). The mixture was concentrated and the residue was
purified by silica gel chromatography (EtOAc:MeOH 80:20) to give 12
(12.9 mg, 15.8%). 1H NMR (300 MHz, CD3OD): d ˆ 3.38 (s, 12H, NCH3),
pH Studies: The pH studies were run at 2.6 Â 10À3 m catalyst for 5 and 3.8 Â
10À4 m for 3 in the range ofpH 7.6 9.2 and 7.8 9.2, respectively.
Synthesis
ˆ
4.56 (s, 4H, C-CH2-O), 6.32 (s, 1H, Ar-H), 6.84 (s, 4H, CH CH), 7.07 (s,
Dimethyl 1,4-dihydro-2,6-dimethoxymethyl-3,5-pyridinedicarboxylate (8):
Methyl methoxyacetoacetate (6; 56.45 g, 0.39 mol) was added to a 40%
aqueous formaldehyde solution (15.25 g). The reaction mixture was cooled
to 08C, then diethylamine (3 mL) was added dropwise. The reaction
mixture was kept at 08C for 4 h, and then allowed to slowly warm to room
temperature, followed by stirring for another 2 d. During this time, two
layers formed, which were separated. The aqueous layer was extracted with
three 100 mL portions ofdiethyl ether. The ether layers were combined
with the previous organic layer, dried over sodium sulfate, and the solvents
removed by rotary evaporation. The residue (7) was dissolved in absolute
ethanol (100 mL), and cooled on an ice bath. The solution was saturated
with ammonia gas, and the mixture was allowed to stir overnight. The
ethanol was evaporated and the residue purified by silica gel chromatog-
raphy (EtOAc:CH2Cl2 15:85), and then by recrystallization from ethanol
(47% yield). 1H NMR (300 MHz, CDCl3): d ˆ 3.22 (s, 2H, C-CH2-C), 3.32
(s, 6H, CH2OCH3), 3.59 (s, 6H, CO2CH3), 4.47 (s, 4H, CH2OMe), 7.79 (s,
1H, NH); 13C{1H} NMR (75 MHz, CDCl3): d ˆ 24.2, 50.8, 58.8, 69.4, 96.8,
146.2, 167.5; HRMS-CI: m/z: calcd for C13H19NO6: 284.1134; found:
4H, CH CH); 13C{1H} NMR (75 MHz, CD3OD): d ˆ 35.2, 64.5, 76.5, 125.6,
ˆ
126.7, 135.47, 136.1, 148.0, 158.2; HRMS-CI: m/z: calcd for C25H29N9O4:
‡
520.2421; found: 520.2401 [M‡H] .
Bis[2,2-bis(1-methyl-2-imidazolyl) furol][3.4-b:3',4'-e] pyridine (apo-2): A
flask containing 85% H3PO4 (3 mL) was heated at 1508C for 30 min. To
this flask was added 12 (35 mg, 0.067 mmol). The mixture was heated an
additional 1.5 h, resulting in a greenish brown solution. After cooling the
mixture to room temperature, water (3 mL) were added and the reaction
mixture was basified to pH 8 by adding Na2CO3. The aqueous solution was
extracted with five 10 mL portions of CH2Cl2. The organic layer was dried
over Na2SO4 and the solvent was removed under reduced pressure. The
residue was purified by two silica gel chromatography (first
CH2Cl2:MeOH(sat.NH3) 95:5; then EtOAc:MeOH 85:15) to give the title
compound (1.12 mg, 34%). 1H NMR (300 MHz, CDCl3): d ˆ 3.53 (s, 12H,
ˆ
ˆ
NCH3), 5.10 (s, 4H, CCH2O), 6.86 (s, 4H, CH CH), 6.86 (s, 4H, CH CH),
7.57 (s, 1H, Ar-H); 13C{1H} NMR (75 MHz, CDCl3): d ˆ 34.4, 71.0, 85.2,
123.7, 127.1, 130.6, 133.5, 145.0, 161.0; HRMS-CI: m/z: calcd for
‡
‡
C25H25N9O2: 484.2187 [M‡H] ; found: 484.2199.
284.1134 [M‡H] .
2,6-Dimethyl-3,5-bis-[di(1-methyl-2-imidazolyl)hydroxymethyl]pyridine (14):
A solution of1-methylimidazole (5.6 mL, 70.25 m m) in dry THF (70 mL)
was cooled to À788C. A solution of nBuLi (42 mL, 1.6m in hexanes,
67.2 mm) was added dropwise under argon. The reaction mixture was
stirred for 1 h at À788C and a solution of3,5-ethylcarboxy-2,6-dimethyl-
pyridine (13; 1.2 g, 4.78 mm) was then added. The reaction mixture was
allowed to warm up to room temperature over night and was cooled again
to À788C followed by very slow addition of MeOH (60 mL). The solvent
was evaporated and the yellow residue was purified by silica gel
chromatography (Rf ˆ 0.35; CH2Cl2:MeOH (sat. NH3) 95:5). Yield 70%.
1H NMR (250 MHZ, MeOD): d ˆ 2.27 (s, 6H, CH3), 3.33 (s, 12H, Im-CH3),
Dimethyl 2,6-dimethoxymethyl-3,5-pyridinedicarboxylate (9):
Method A: A mixture ofH 2O (21.6 g), concentrated nitric acid (sp. gr. 1.42,
5.8 g) and concentrated sulfuric acid (6.25 g) were added to 8 (18.05 g,
0.063 mol). The mixture was heated until all solids had dissolved, then
heating was continued an additional 3 min; the reaction mixture turned
yellowish orange. The mixture was then cooled on ice, while ice (10 g) was
added; then the mixture was made alkaline (pH 9) by adding ammonium
hydroxide. It was then extracted with CH2Cl2 and the combined extracts
were dried over sodium sulfate. The residue after concentration was
purified by silica gel chromatography (CH2Cl2: EtOAc 7:3) to yield the
product (9.99 g, 56%).
6.57 (s, 1H, Ar-H), 6.83/6.84 and 7.06/7.07 (each d, 4H Im-CH ); 13C NMR
ˆ
Method B: A round bottom flask was charged with 8 (420 mg, 1.47 mmol)
and 2,3-dichloro-5,6-dicyanobenzoquinone (DDQ; 401 mg, 1.77 mmol) in
THF (20 mL). The mixture was stirred under nitrogen for 5 h, at which time
TLC showed no starting material. The reaction mixture was then passed
through the neutral alumina to remove excess DDQ and hyroquinone, then
the alumina was washed with a large quantity ofmethylene chloride. The
filtrate was concentrated and the residue purified by silica gel chromatog-
raphy (EtOAc:CH2Cl2 3:7), yielding the product (304 mg, 73%). M.p. 45
478C; 1H NMR (300 MHz, CDCl3): d ˆ 3.35 (s, 3H, CH2OCH3), 3.36 (s,
3H, CH2OCH3), 3.84 (s, 3H, COCH3), 3.85 (s, 3H, COCH3), 4.83 (s, 2H,
CH2OMe), 4.84 (s, 2H, CH2OMe), 8.50 (s, 1H, Ar-H); 13C{1H} NMR
(75 MHz, CDCl3): d ˆ 52.1, 58.6, 73.6, 124.2, 139.9, 159.9, 165.3; HRMS-CI:
(62.9 MHz, MeOD): d ˆ 23.31 (CH3), 35.09 (Im-CH3), 76.14 (C-OH),
125.28, 126.56 (Im-CH), 148.31 (Im-C), 134.64, 135.46, 158.38 (Ar-C); MS
‡
‡
(CI): calcd for C25H29N9O2: 487; found: 487 [M] , 489 [M‡2] .
2,6-Dimethyl-3,5-bis-[di(1-methyl-2-imidazolyl)(3-cyanobenzyloxy)meth-
yl] pyridine (16): The reaction conditions were similar to 18, but the residue
after lyophilization was purified by silica gel chromatography (first wash
with 100% CH2Cl2, then eluted with CH2Cl2:MeOH (sat. NH3) 90:10, Rf ˆ
0.52). Yield 65%. 1H NMR (250 MHz, CDCl3): d ˆ 2.52 (s, 6H, CH3Ar),
3.08 (s, 12H, CH3-Im), 4.67 (s, 4H, Ar-CH2), 6.74 and 6.91 (each s, 4H, Im-
ˆ
CH ), 7.36, 7.52, 7.76 (4H, AR-CH); MS (CI): calcd for C41H39N11O2:
‡
‡
717.833); found: 717, 719 [M‡2] , 587 [M À OBnzCN] ; HR-MS: calcd for:
‡
‡
718.3366; found: 718.3367 [M‡H] .
m/z: calcd for C13H17NO6: 284.1134; found: 284.1134 [M‡H] .
2,6-Dimethoxymethyl-3,5-bis[di(1-methyl-2-imidazolyl)hydroxymethyl] py-
ridine (11): A solution of nBuLi (15.2 mL, 24.2 mmol ofa 1.6 m solution in
hexane) was added dropwise to a solution of1-methylimidazole (1.92 g,
22 mmol) in THF (30 mL) at À788C for an hour, and then 9 (0.632 g,
2.21 mmol) in THF (200 mL) was added, which resulted in a dark red
solution. The reaction mixture was allowed to warm to room temperature
overnight. The reaction mixture was cooled to À788C and methanol
(100 mL) was then added slowly. The solvent was removed in vacuo,
Di(1-methyl-2-imidazolyl)(3-cyanobenzyloxy)methyl-2-methylbenzene (19):
Alcohol 18 (220 mg) and KOH (360 mg) were ground together and dry
DMSO (3 mL) was added. The light yellow mixture was stirred for 10 min
at room temperature. Then, the supernatant was taken and a-bromotolu-
nitrile (15; 211 mg) was added. The now dark yellow solution was stirred
for 5 h at RT. H2O (100 mL) and 1n HCl (2 mL) were added and the
mixture was lyophilized three times, adding water (50 mL) each time. The
residue was purified by reversed phase liquid chromatography using a
gradient mixer, column 1.0 Â 35 cm, flow rate 3 mLminÀ1). Yield 85%.
yielding
a
dark red residue which was first purified by silica gel
Chem. Eur. J. 2003, 9, No. 3
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0903-0745 $ 20.00+.50/0
745

E. V. Anslyn et al.
FULL PAPER
1H NMR (250 MHz, CDCl3): d ˆ 2.14 (s, 3H, CH3Ar), 3.27 (s, 6H, CH3-
smoothed curve ofthe intensities ofthese check relfections was used to
scale the data. The scaling factor ranged from 0.940 to 1.00. The data were
corrected for Lp effects but not absorption. Data reduction and decay
correction were performed using the SHELXTL-Plus software package.[10]
The structure was solved by direct methods[11] and refined on F2 by full-
ˆ
Im), 4.70 (s, 2H, Ar-CH2), 6.88 and 7.03 (each br, 2H, Im-CH ), 7.11, 7.23,
7.37, 7.52, 7.71, (m, 8H, Ar-CH); 13C NMR (62.9 MHz, CDCl3): d ˆ 21.03,
34.57, 67.14, 82.12, 112.11,118.84, 123.29, 125.03, 126.23, 127.88, 128.82,
128.90, 130.84, 131.13, 132.06, 132.95, 135.91, 139.54, 140.18, 145.52;
‡
matrix least-squares[12] with anisotropic thermal parameters for the non-H
MS(CI): calcd for C24H23N50: 397; found: 397, 398 [M‡H] , 265 [M À
‡
À
Obnz À CN] .
atoms ofthe Zn complex. Because only three BF ions were observed in
4
the asymmetric unit and the ligands were neutral, the bridging atom
between the Zn2‡ ions had to have a À1 charge. Models were refined with
the bridging ion as O, F and Cl. The best refinement as judged by the
reasonableness ofthe isotropic temperature parameter was with F as
bridge. The Uiso for the Cl refinement was 0.175, for O 0.002 and for F it was
0.033. The Uiso for F was close to the values observed for the other atoms of
the complex. The atoms ofthe Zn dimer were well resolved and were
refined anisotropically. Of the three BF4À ions, one was well behaved, one
was disordered by rotation about an oxygen-boron bond and the third was
CN-Hydrolysis–General Procedure
Di(1-methyl-2-imidazolyl)(3-carboxybenzyloxy)methyl]-2-methyl benzene
and 2,6-dimethyl-3,5-bis-[di(1-methyl-2-imidazolyl)(3-carboxylbenzyloxy)-
methyl] pyridine) (apo-3 and apo-5): The cyano compound (16 or 19),
300 mol% ofNaOH and isopropanol/H 2O (5:1) were stirred in a sealed
tube at 1308C for 12 h. The mixture was neutralized with HCl in an ice bath.
The organic phase was evaporated and the residue was purified via
reversed phase liquid chromatography using a gradient of90% CH 3CN
(100 mL each in
a
gradient mixer, column 1.0 Â 35 cm, flow rate
À
badly disordered. This third BF4 ion and several solvate molecules could
3 mLminÀ1 Yields: 93 and 91%, respectively. Apo-5: 1H NMR
.
not be refined. The contribution to the observed structure factor
amplitudes due to the disordered entities was removed by using the utility,
(250 MHz, CDCl3/MeOD, 1:1): d ˆ 2.03 (s, 3H, CH3Ar), 3.37 (s, 6H,
ˆ
CH3-Im), 7.94 and 7.98 (each s, 2H, Im-CH ), 7.14 7.45, 7.85 7.91 (m, 8H,
2
2
SQUEEZE, in PLATON98.[12] The function, Sw(jFo j Àj Fc j )2, was
‡
Ar-CH); MS(CI): calcd for C24H24NO3: 416; found: 416, 417 [M‡H] , 418
2
2
minimized, where w ˆ 1/[(s(Fo))2‡(0.02P)2] and P ˆ (jFo j ‡2 j Fc j )/3.
Rw(F2) refined to 0.173, with R(F) equal to 0.0980 and a goodness ofift,
S ˆ 1.04. Definitions used for calculating R(F), Rw(F2) and the goodness of
fit, S, are given below.[13] The data were checked for secondary extinction
effects but no correction was necessary. Neutral atom scattering factors and
values used to calculate the linear absorption coefficient are from the
International Tables for X-ray Crystallography.[14] All figures were
generated using SHELXTL/PC.[11] Tables ofpositional and thermal
parameters, bond lengths and angles, torsion angles, figures and lists of
observed and calculated structure factors are available from the CCDC.[15]
‡
‡
[M‡2] ; HR-MS: calcd for: 417.192666; found: 417.191467 [M‡H] .
Apo-3: 1H NMR (250 MHz, CDCl3/MeOD): d ˆ 2.35 (s, 6H, CH3Ar), 3.30
(s, 12H, CH3-Im), 4.50 (s, 4H, Ar-CH2), 6.91 and 7.00 (each s, 4H, Im-
ˆ
CH ), 7.27(br, 4H), 7.48 (s, 1H), 7.85 (br, 4H, all Ar-CH); MS(CI): calcd for
‡
C41H41N9O6: 755; found: 755, 756 [M‡H] ; HR MS: calcd for: 756.3258,
‡
found: 756.3268 [M‡H] .
2-Methylphenyl bis(1-methyl-2-imidazolyl) methanol (18): A solution of
1-methylimidazole (5.6 mL, 70.25 mm) in dry THF (70 mL) was cooled to
À788C. A solution of nBuLi (42 mL, 1.6m in hexanes, 67.2 mm) was added
dropwise under argon. The reaction mixture was stirred for 1 h at À788C,
at which time methyl-2-methyl benzoate (17; 3.2 mL, 20 mm) was added
very slowly. The reaction mixture was allowed to warm to room temper-
ature over night. It was then cooled again to À788C and MeOH (60 mL)
were added slowly. The solvent was evaporated and the yellow residue was
purified by silica gel chromatography (Rf ˆ 0.43; CH2Cl2:MeOH (sat. NH3)
90:10). Yield 67 72%. 1H NMR (250 MHz, CDCl3): d ˆ 2.06 (s, 3H, CH3),
3.38 (s, 6H, Im-CH3), 6.09 (s, 1H, Ar-H), 6.41 (d, 1H, Ar-H), 7.05 7.20 (m,
Acknowledgement
We gratefully acknowledge support for this work from the National Science
Foundation.
2H, Ar-H), 7.87 and 7.97 (d, 2H each, Im-HC ); 13C NMR (62.9 MHz,
ˆ
Table 1. Crystal data and structure refinement for 20.
CDCl3): d ˆ 19.99 (CH3), 34.79 (Im-CH3), 76.12 (q-C-OH), 125.43, 127.70,
empirical formula
C55H60B3F13N16O13Zn2
ˆ
128.67, 132.81 (Ar-CH), 138.86, 148.66 (Ar-C), 123.15, 125.93 (Im-HC );
MS (CI): calcd for C16Hl8N4O: 282; found: 282, 283 [M‡H] .
‡
Fw
1563.36
T [K]
l [ä]
173(2)
0.71073
1,2,3,5,6,7-Hexahydro-3,3,5,5-tetra(1-methyl-2-imidazolyl)-2,6-dioxa-1,7-s-
indacenedione (apo-20): A solution of nBuLi (3.4 mL, 5.1 mmol ofa 1.5 m
solution in hexane) was added dropwise to a solution of1-methylimidazole
(0.38 g, 4.6 mmol) in THF (20 mL) at À788C for an hour then pyromelli-
tide anhydride (20; 0.2 g, 0.92 mmol) in THF (50 mL) was added dropwise
to the reaction mixture. The mixture was allowed to warm to room
temperature overnight. The mixture was cooled to À788C and methanol
(50 mL) was added slowly. The solvent was removed in vacuo, yielding a
dark red residue which was dissolved in 85% H3PO4 (50 mL) and heated at
1208C for an hour. After cooling the mixture to room temperature, water
(30 mL) was added and the mixture was basified to pH 8 by adding
concentrated NH4OH. The aqueous solution was extracted with three
50 mL portions ofCH 2Cl2. The organic layer was dried over Na2SO4 and
the solvent was removed under reduced pressure. The residue was purified
by silica gel chromatography (CH2Cl2:MeOH(sat.NH3) 97:3) to give 20
(197 mg, 42%). 1H NMR (300 MHz, CDCl3): d ˆ 3.45 (s, 12H, NCH3), 6.85
crystal system
space group
a [ä]
monoclinic
P21/n
18.367(3)
b [ä]
c [ä]
19.038(4)
19.453(4)
a [8]
90
b [8]
99.10(1)
g [8]
90
6717(2)
4
1.546
0.823
3192
V [ä3]
Z
1calcd [MgmÀ3
]
m [mmÀ1
F(000)
]
crystal size [mm3]
0.57 Â 0.23 Â 0.23
2.25 to 22.51
0 ꢀ h ꢀ 19
t range for data collection [8]
index ranges
ˆ
ˆ
(s, 4H, NCH C), 6.89 (s, 4H, NCH C), 8.00 (s, 1H, Ar-H), 8.43 (s, 1H, Ar-
H); 13C{1H} NMR (75 MHz, CD3OD): d ˆ 34.4, 83.7, 123.3, 124.6, 125.7,
127.4, 127.6, 141.4, 153.4, 167.0; HRMS-CI: m/z: calcd for C26H22N8O4:
0 ꢀ k ꢀ 20
À 20 ꢀ l ꢀ 20
‡
511.1842; found: 511.1833 [M‡H] .
reflections collected
8663
X-ray Experimental for [(C26H22N8O4)2Zn2F] Â 3BF4À Â 3CH3OH Â 2H2O
(20): Crystals grew as colorless prisms from CH3OH. The crystals
decomposed rapidly when removed from the mother liquor and had to
be stabilized by immersion in mineral oil. The data crystal had approximate
dimensions; 0.23 Â 0.23 Â 0.57 mm3. The data were collected at 173 K on a
Nicolet P3 diffractometer, equipped with a Nicolet LT-2 low-temperature
device and using a graphite monochromator with MoKa radiation (l ˆ
0.71073 ä). Details ofcrystal data, data collection and structure refinement
are listed in Table 1. Three reflections (3,0, À 5;5,0, À 3;3,2,3) were remeas-
ured every 97 reflections to monitor instrument and crystal stability. A
independent reflections
completeness to t ˆ 22.518
absorption correction
refinement method
data/restraints/parameters
GOF on F2
8449 [R(int) ˆ 0.1003]
96.0 %
none
full-matrix least-squares on F2
8449/673/829
1.050
final R indices [I > 2s(I)]
R indices (all data)
largest diff. peak and hole [eäÀ3
R1 ˆ 0.0980, wR2 ˆ 0.1511
R1 ˆ 0.2292, wR2 ˆ 0.1726
0.813 and À0.532
]
746
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0903-0746 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 3

Bimetallic Artificial Enzymes
741 747
[1] For just a few representative examples see: P. Hurst, B. K. Takasaki, J.
Chin, J. Am. Chem. Soc. 1996, 118, 9982; J. K. Bashkin, E. I. Frolova,
U. Sampath, J. Am. Chem. Soc. 1994, 116, 5981; D. H. Vance, A. W.
Czarnik, J. Am. Chem. Soc. 1993, 115, 12165; N. H. Williams, J. Chin,
Chem. Commun. 1996, 131; M. K. Stern, J. K. Bashkin, E. D. Sall, J.
Am. Chem. Soc. 1990, 112, 5357; B. Linkletter, J. Chin, Angew. Chem.
1995, 107, 529; Angew. Chem. Int. Ed. Engl. 1995, 34, 472; S. Liu, Z.
Luo, A. D. Hamilton, Angew. Chem. 1997, 109, 2794; Angew. Chem.
Int. Ed. Engl. 1997, 36, 2678; M. Komiyama, N. Takeda, H. Shigekawa,
Chem. Commun. 1999, 1443; P. Molenved, J. F. J. Engbersen, D. N.
Reinhoudt, Angew. Chem. 1999, 111, 3387; Angew. Chem. Int. Ed.
1999, 38, 3189; K. A. Deal, A. C. Hengge, J. N. Burstyn, J. Am. Chem.
Soc. 1996, 118, 1713; M. J. Young, J. Chin J. Am. Chem. Soc. 1995, 117,
10577; M. Komiyama, K. Yoshinari, J. Org. Chem. 1997, 62, 2155; R.
Breslow, D. Berger, D.-L. Huang, J. Am. Chem. Soc. 1990, 112, 3686;
J. R. Morrow, D. Epstein J. Chem. Soc. Chem. Commun. 1995, 2431; E.
Kimura, Y. Kodama, T. Koike, M. Shiro, J. Am. Chem. Soc.1995, 117,
8304; P. Molenveld, J. F. J. Engbersen, D. N. Reinhoudt, J. Org. Chem.
1999, 64, 6227; Y. Baran, T. W. Hambley, G. A. Lawrance, E. N.
[3] N. Strater, W. N. Lipscomb, T. Klabunde, B. Krebs, Angew. Chem.
1996, 108, 2158; Angew. Chem. Int. Ed. 1996, 35, 2024; D. E. Wilcox,
Chem. Rev. 1996, 96, 2435.
[4] M. J. Young, J. Chin, J. Am. Chem. Soc. 1995, 117, 10577.
[5] G. Hennrich, E. V. Anslyn, Chem. Eur. J. 2002, 8 2218.
[6] R. P. Thummel, D. K. Kohli, Tetrahedron Lett. 1979, 2, 143.
[7] F. Chu, J. Smith, V. M. Lynch, E. V. Anslyn, Inorg. Chem. 1995, 34,
5689.
[8] E. Ohyoshi, Anal. Chem. 1985, 57, 446; E. Ohyoshi, S. Kohata, J. Inorg.
Biochem. 1991, 43, 45.
[9] F. M. Richards, H. W. Wyckoff, The Enzymes, Vol. III (Ed.: P. D.
Boyer), 3rd ed., Academic Press, New York, 1971.
[10] G. M. Sheldrick, SHELXTL-PLUS (Version 4.1), Siemens Analytical
X-ray Instruments, Inc., Madison, Wisconsin (USA), 1990.
[11] G. M. Sheldrick, SHELXTL/PC (Version 5.03)mSiemens Analytical
X-ray Instruments, Inc., Madison, Wisconsin (USA), 1994.
[12] A. L. Spek, PLATON,
A Mul tipurpose Crystallographic Tool,
Utrecht University (The Netherlands), 1998.
2
2
[13] Rw(F2) ˆ {Sw(jFo j Àj Fc j )2/Sw(jFo j )4}1/2
¬
Wilkes, Aust. J. Chem. 1997, 883; E. Kˆvari, R. Kr‰mer J. Am. Chem.
where w is the weight given each reflection. R(F) ˆ S(jFo jÀj Fc j )/S j
2
2
Soc. 1996, 118, 12704; M. Wall, B. Linkletter, D. Williams, A.-M.
Lebuis, R. C. Hynes, J. Chin, J. Am. Chem. Soc. 1999, 121, 4710. For
reviews ofthe area see: J. R. Morrow, Metal Ions in Biological Systems
(Eds.: H. Sigel, A. Sigel), Dekker, NY, pp. 32, Chapter 7, 1996; N. H.
Williams, B. Takasaki, M. Wall, J. Chin, Acc. Chem. Res. 1999, 32, 485;
P. Molenveld, J. F. J. Engbersen, D. N. Reinhoudt, Chem. Soc. Rev.
2000, 29, 75.
Fo j } for reflections with Fo > 4(s(Fo)). S ˆ [Sw(jFo j Àj Fc j )2
(n À p)]1/2, where n is the number ofreflections and p is the number of
refined parameters.
[14] International Tables for X-ray Crystallography, Vol. C, Tables 4.2.6.8
and 6.1.1.4, (Ed.: A. J. C. Wilson), Boston, Kluwer Academic Press,
1992.
[15] Crystallographic data (excluding structure factors) for the structures
reported in this paper have been deposited with the Cambridge
Crystallographic Data Centre: CCDC-197492 contains the supple-
mentary crystallographic data for this paper. These data can be
obtained free of charge via www.ccdc.cam.ac.uk/conts/retrieving.html
(or from the Cambridge Crystallographic Data Centre, 12 Union
Road, Cambridge CB2 1EZ, UK; fax: (‡44)1223-336033; or depos-
it@ccdc.cam.uk).
[2] S. Liu, A. D. Hamilton. Bioorg. Med. Chem. Lett. 1997, 7, 1779; K. P.
McCue, D. A. Voss, Jr., C. Marks, J. Morrow, J. J. Chem. Soc. Dalton
Trans. 1998, 2961; P. Moleveld, J. F. J. Engbersen, H. Kooijman, A. L.
Spek, D. N. Reinhoudt, J. Am. Chem. Soc. 1998, 120, 6726; S. Matsuda,
A. Ishikubo, A. Kuzuya, M. Yashiro, M. Komiyama, Angew. Chem.
1998, 110, 3477; Angew. Chem. Int. Ed. 1998, 37, 3284; M. Yashira, A.
Ishikubo, M. Komiyama, Chem. Commun. 1997, 83; W. H. Chapman,
Jr. R. Breslow,J. Am. Chem. Soc. 1995, 117, 5462.
Received: August 5, 2002 [F4317]
Chem. Eur. J. 2003, 9, No. 3
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0903-0747 $ 20.00+.50/0
747