Incorporation of Cobalt-Cyclen Complexes into Templated Nanogels Results in Enhanced Activity

Article abstract of DOI:10.1002/chem.201503946

Recent advances in nanomaterials have identified nanogels as an excellent matrix for novel biomimetic catalysts using the molecular imprinting approach. Polymerisable Co-cyclen complexes with phosphonate and carbonate templates have been prepared, fully characterised and used to obtain nanogels that show high activity and turnover with low catalytic load, compared to the free complex, in the hydrolysis of 4-nitrophenyl phosphate, a nerve agent simulant. This work demonstrates that the chemical structure of the template has an impact on the coordination geometry and oxidation state of the metal centre in the polymerisable complex resulting in very significant changes in the catalytic properties of the polymeric matrix. Both pseudo-octahedral cobalt(III) and trigonal-bipyramidal cobalt(II) structures have been used for the synthesis of imprinted nanogels, and the catalytic data demonstrate that: i) the imprinted nanogels can be used in 15 % load and show turnover; ii) the structural differences in the polymeric matrices resulting from the imprinting approach with different templates are responsible for the molecular recognition capabilities and the catalytic activity. Nanogel P1, imprinted with the carbonate template, shows >50 % higher catalytic activity than P2 imprinted with the phosphonate.

Full text of DOI:10.1002/chem.201503946

DOI: 10.1002/chem.201503946
Full Paper
&
Molecularly Imprinted Nanogels
Incorporation of Cobalt-Cyclen Complexes into Templated
Nanogels Results in Enhanced Activity
Ana Rita Jorge,[a] Mariya Chernobryva,[a] Stephen E. J. Rigby,[b] Michael Watkinson,*[a] and
Marina Resmini*[a]
Abstract: Recent advances in nanomaterials have identified
nanogels as an excellent matrix for novel biomimetic cata-
lysts using the molecular imprinting approach. Polymerisable
Co-cyclen complexes with phosphonate and carbonate tem-
plates have been prepared, fully characterised and used to
obtain nanogels that show high activity and turnover with
low catalytic load, compared to the free complex, in the hy-
drolysis of 4-nitrophenyl phosphate, a nerve agent simulant.
This work demonstrates that the chemical structure of the
template has an impact on the coordination geometry and
oxidation state of the metal centre in the polymerisable
complex resulting in very significant changes in the catalytic
properties of the polymeric matrix. Both pseudo-octahedral
cobalt(III) and trigonal-bipyramidal cobalt(II) structures have
been used for the synthesis of imprinted nanogels, and the
catalytic data demonstrate that: i) the imprinted nanogels
can be used in 15% load and show turnover; ii) the struc-
tural differences in the polymeric matrices resulting from the
imprinting approach with different templates are responsible
for the molecular recognition capabilities and the catalytic
activity. Nanogel P1, imprinted with the carbonate template,
shows >50% higher catalytic activity than P2 imprinted
with the phosphonate.
Introduction
mentally relevant pH and their enhanced reactivity presents
a risk of indiscriminate damage to cellular components in
vivo.[6] In parallel, small molecule complexes of transition
metals continue to be widely studied.[7] Although the rational
design of simple metal complexes is a key element in the opti-
misation of their catalytic activity, it has been demonstrated
that placing these complexes in low polarity environments
that closely mimic the dielectric constant of the active sites of
enzymes is very important. Brown et al. showed that triaza-
macrocyclic bimetallic ZnII complexes can accelerate the cleav-
age of certain phosphate substrates by 12 orders of magnitude
in the presence of methanol,[8] with accelerations approaching
those of natural enzymes.[2,8] More recently the incorporation
of such small molecule catalysts into supramolecular architec-
tures has been shown to significantly enhance their catalytic
properties.[9,10] Of particular interest to us was the report by
Scrimin et al., demonstrating that self-organisation of ZnII com-
plexes of azamacrocyclic ligands on gold nanoparticles result-
ed in remarkable increases in cleavage efficiency when com-
pared to the free complexes in aqueous solution, an effect at-
tributed to the decrease in polarity of the reaction site.[4a]
We have successfully demonstrated that colloidal microgels
and nanogels can be tuned to achieve high catalytic activity
using the molecular imprinting approach, and are able to cata-
lyse chemical reactions, such as carbonate hydrolysis,[11] aldol
condensations[12] and Kemp eliminations[13] with good effi-
ciency and selectivity. Given our additional expertise in the
synthesis of functionalised macrocyclic amine systems[14] and
the use of their metal complexes as biomimetic catalysts[15] as
well as sensors,[16] the incorporation of these systems into
Catalysis of energetically demanding reactions, such as phos-
phate ester hydrolysis, has challenged researchers for many
years.[1,2] Whilst Nature has evolved enzymes that efficiently
promote such reactions, their applications are often limited.[3]
Among the different approaches investigated for the develop-
ment of synthetic alternatives, metal complexes have attracted
considerable interest and have resulted in a number of impres-
sive achievements.[4] For example, lanthanide-based systems
containing CeIV are some of the most active inorganic catalysts
reported,[5] but their complexes are often insoluble at environ-
[a] Dr. A. R. Jorge, M. Chernobryva, Prof. M. Watkinson, Prof. M. Resmini
Department of Chemistry and Biochemistry
School of Biological and Chemical Sciences
Queen Mary University of London, Mile End Road, London, E1 4NS (UK)
E-mail: m.watkinson@qmul.ac.uk
m.resmini@qmul.ac.uk
[b] Dr. S. E. J. Rigby
Faculty of Life Sciences, Manchester Institute of Biotechnology
131 Princess Street, Manchester, M1 7DN (UK)
Supporting information and ORCID(s) from the author(s) for this article are
available on the WWW under http://dx.doi.org/10.1002/chem.201503946.
ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons At-
tribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
Part of a Special Issue “Women in Chemistry” to celebrate International
Women’s Day 2016. To view the complete issue, visit: http://dx.doi.org/
chem.v22.11.
Chem. Eur. J. 2016, 22, 3764 – 3774
3764 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
novel nanomaterials via the imprinting approach appeared to
have considerable potential.
most active mimetic systems for phosphate hydrolysis. Chin
et al.[24] first established 4d as an active synthetic phosphatase
hydrolysis catalyst and reported hydrolysis rate enhancement
factors of up to 107, while supercoiled DNA was hydrolysed in
the presence of a functionalised derivative of 4.[25] However, in
all cases turnover was only achieved in the presence of
a second metal unit, to release the product.[26]
The use of metal-templated sites for molecular recognition
and sensing has been previously investigated.[17–19] An essential
feature for successful imprinting in polymeric systems is the
stability and integrity of the template–monomer complex.[20] In
the case of metal complexes, ligand exchange and fluxionality
of metal centres can lead to dynamic changes to the coordin-
ation environment of the metal. In-depth studies of the coord-
ination geometry of the templated complex prior to polymer-
isation are seldom undertaken,[21] with a few exceptions,[22] and
it is normally assumed that the complex expected to form,
based on the coordination chemistry observed in small mol-
ecule analogues, will also be maintained in the polymer-based
macromolecular structure.[20]
Several features make these systems attractive for the devel-
opment of artificial phosphatases. The size of the metal ion
and the cyclen ligand cavity results in a rigid and non-fluxional
structure that forces only cis-coordination sites to be present
at the cobalt(III) centre.[24] In contrast, analogous trans-com-
plexes of the closely related cyclam ligand are inactive.[27] Fur-
thermore, although such cis-complexes have been reported to
be active in the hydrolysis of phosphate esters, cobalt(III) com-
plexes are generally considered to be substitutionally inert and
should provide a stable interaction with any target template,
as well as a very well-defined primary coordination sphere. We
viewed this aspect as a key feature in our imprinting strategy
to prevent unwanted ligand exchange and fluxionality of the
metal centres which could lead to dynamic changes to the co-
ordination environment of the metal. Additionally, after poly-
merisation, the template can be readily removed from the
cobalt centre under experimental conditions that are well
documented.[23]
Here we report the synthesis of polymerisable Co-cyclen sys-
tems that have been complexed with strategically chosen tem-
plates to obtain imprinted nanogels, where the three-dimen-
sional polymeric matrix around the metal atom mimics the sec-
ondary coordination sphere of natural metalloenzymes. The
catalytic activity of the different polymers in the hydrolysis of
4-nitrophenyl phosphate 1 is evaluated, in comparison with
the free complexes, and analysis of the results used to high-
light the impact of the polymeric matrix on the catalytic activ-
ity.
The choice of template in molecular imprinting plays a key
role in influencing the recognition and catalytic properties of
the polymers. Traditionally, transition-state analogues (TSA)
have been shown to be good templates, however this is de-
pendent on the catalytic mechanism being mimicked. In this
work three compounds were chosen as suitable templates.
The first, 4-nitrobenzyl phosphonate 6 is the cognate substrate
analogue to 1 but is not a substrate for the metal complex
and provides almost identical chemical features, particularly
the p-nitro-substituent, previously shown to play a key role in
molecular recognition.[28] The other two structurally distinct
templates were selected to evaluate the impact of the chem-
ical structure on the molecular recognition properties of the
Results and Discussion
We decided to focus our initial investigation on the hydrolysis
of 4-nitrophenyl phosphate 1 (NPP), a model substrate for
nerve agents, shown in Figure 1a, that when hydrolysed re-
leases 2 and the highly chromophoric unit 4-nitrophenol (3),
which can be easily monitored by UV/Vis spectroscopy. The
choice of catalyst was dictated by our interests in complexes
of azamacrocycles,[15,16] in particular cobalt(III) complexes of the
type cis-[Co(cyclen)(X)2]n+ [4,23]
and their ease of synthesis from
,
precursor 5. Such complexes have been extensively used as
ATPase and phosphatase models and remain amongst the
Figure 1. a) Hydrolysis of model substrate 1; b) structures of model complexes 4 and 5; c) templates used for the imprinting approach: phosphonates 6 and
7 and carbonate 8.
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3765 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
polymer matrix. Phenyl phosphonate 7 has a similar structure
to substrate 1 but is lacking the p-nitro-substituent; it is elec-
tronically richer at the oxygen donors and is sterically less de-
manding. The carbonate template is significantly different, al-
though still a good mimic in terms of its coordination geom-
etry, as its coordination mode in cobalt(III) cyclen complexes is
established to be cis-bidentate,[23] resulting in a pseudo-octahe-
dral geometry.
dentate phosphonate 10 (59%, Ka =1.8”10À2 mÀ1) and chelat-
ed phosphonate 11 (39%) can be found in solution, with
a small amount of dimerisation occurring to give 12 (2%). At
708C, the temperature at which polymerisation occurs, this
equilibrium is entirely shifted towards 11 (100%) (see Figure S1
in the Supporting Information). Importantly, addition of HCl re-
sults in complete substitution of 7 in the coordination sphere,
indicating that this template can be readily removed at room
temperature. Further characterisation of 4a, 5, 9 and 10 by
UV/Vis, IR spectroscopy, and high-resolution mass spectrometry
supported the identity of all of the complexes and all data
were consistent with the same pseudo-octahedral geometry
around the cobalt centre as observed in 5. UV/Vis spectra gave
d–d bands with larger than normal extinction coefficients for
octahedral CoIII complexes, as previously reported for com-
plexes of this type.[34]
Non-polymerisable complexes of [Co(cyclen)(L)]n+ were pre-
2À
pared, where L=(Cl)2 (4a), CO3 (5), 4-NPPA (9), and PPA (10)
(Figure 2a). For 4a and 5 the coordination complexes were
readily obtained using reported methods,[23] whilst 9 and 10
were prepared by addition of cobalt(II) chloride and the phos-
phonate to a solution of cyclen under nitrogen, and the result-
ing mixture subsequently allowed to oxidize in air.
To better understand the behaviour and stability of the com-
plexes in solution, template 7 was added to 4a and the
nature, and extent of binding in 10 measured in a range of dif-
ferent solvents, at different temperatures and pH, using
31P NMR spectroscopy. Depending on the conditions used, co-
ordination of one, as in 10, or chelation of both oxygen atoms,
as in 11, to the metal could be favoured; dimerization of co-
balt(III) centres also occurs as in 12. It was found that in water
at pD 9, when the complex is in its dihydroxo form [Co(cyc-
len)(OH)2]+ namely 4c and 7 is a di-anion, a mixture of mono-
For the synthesis of the polymerisable derivative of 4d both
vinyl and styrene groups were evaluated. Aoki and co-workers
previously reported the synthesis of 15b,[35] but instead we
chose to introduce the alkene bonds via alkylation of the tetra-
cyclic bis-aminal intermediate by modification of a reported
procedure,[29,36] (Figure 2b). The first polymerisable complex to
be synthesised and characterised was 16. The coordination
geometry appeared to be consistent with the desired pseudo-
octahedral complexes that had been found for the non-poly-
Figure 2. a) Structure of complexes 9–12; b) synthesis of polymerisable and non-polymerisable ligands 15; i) glyoxal (40% aq.), MeOH, 72%;[29] ii) allyl brom-
ide,[30] 1-(chloromethyl)-4-vinylbenzene or p-ethyl benzyl bromide[29, 31,32] toluene 93% and 94% and acetonitrile 91%; iii) NH2OH (50% aq.), EtOH, 15a 73%,
15b 96%, 15c 18%;[29,33] c) co-polymerisable complexes with carbonate templates 16 and 17 and cognate phosphonate template 18.
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3766 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
merisable analogue with the UV/Vis spectra displaying a single
d–d band with a lmax at 535 nm. This complex led to unsatis-
factory yields in the polymerisation and it was replaced with
the styrene unit, that although bringing additional steric bulk
to the structure, is known to be more active in radical polymer-
isation.[37,38] Complex 17 was prepared and UV/Vis analysis was
also consistent with the geometry observed in 5 (see Fig-
ure S2).
Coordination of 6 and 7 to the cobalt–styrene monomer
was successfully achieved using the same procedure adopted
for the preparation of 9 and 10, although analysis by UV/Vis
and 31P NMR spectroscopies displayed some anomalies. For 7
the UV/Vis spectrum showed a band at significantly higher
wavenumber (580 nm) and a meaningful 31P NMR spectrum
could not be recorded. Nonetheless high-resolution mass spec-
tra and IR spectra appeared to be consistent with the forma-
tion of the expected molecular species [Co(15b)(7)(OH)]+. The
complex with 6 presented the same effect in the NMR studies
but exhibited even more pronounced differences in its UV/Vis
spectrum, with the presence of new d–d bands in the region
600–650 nm (Figure S3). This behaviour appeared to be con-
sistent with the presence of a paramagnetic cobalt(II) species.
Previous work by Sarther and Blinn,[39] reported a tetra-benzyl-
ated derivative of cyclen that was coordinated to cobalt(II) re-
sulted in a complex that was resistant to oxidation by electro-
chemical means, direct aeration or by strong oxidants such as
H2O2.
Figure 3. EPR spectra of frozen DMSO solutions of: a) 18, b) a solution of
nanogel P2 prior to template removal, c) nanogel solution P2 after template
removal at 18 K.
The complex was shown to have the rare five-coordinate
trigonal-bipyramidal geometry and displayed the same charac-
teristic UV/Vis spectra observed for 18 suggesting that this
complex was also five-coordinate cobalt(II) (Figure 2c). To fur-
ther substantiate this hypothesis, magnetic moments were
measured using the Evans’ NMR method.[40] This showed that
both complexes of ligand 15b with templates 6 and 7 con-
tained significant quantities of paramagnetic species in solu-
tion, the latter, 18, apparently being almost exclusively high-
spin cobalt(II) based on its magnetic moment (meff =4.0 mB).
Analysis of both complexes by X-band CW-EPR spectroscopy
clearly showed them to contain high-spin cobalt(II). Spectra re-
corded at 18 K, using 0.5 mW of microwave power and 0.5 mT
field modulation are broad and axial with g values of approxi-
mately 4 and 2 (see Figures 3a–c and Figure S4 in the Support-
ing Information)). This is typical of frozen solution spectra aris-
ing from the S=3/2 d7 spin system of the high-spin cobalt(II)
ion having a zero-field splitting (D) greater than hu at the X
band (u=9.38 GHz).[41]
tion state of the metal centre. Importantly, under the polymer-
isation conditions (DMSO, 708C) the characteristic UV/Vis
bands for the five-coordinate cobalt(II) centre in 18 are un-
altered, indicating that this coordination geometry is main-
tained (see Figure S5). Given that the ligand complex with 7
appeared as a mixture of species, it was not used for the prep-
aration of the nanogels.
Nanogels were synthesised by high-dilution radical polymer-
ization of complexes 17 and 18, following our previously re-
ported protocol,[11] using 80% ethylene bis-acrylamide (EBA) as
the crosslinker, 10% acrylamide and 10% of the polymerisable
catalyst and DMSO as the solvent. Template molecules 6 and 8
were then removed from the matrix by washing with dilute
HCl. Further dialysis in water resulted in the generation of the
catalytically active CoIII aqua-hydroxo species (Figure S6). Two
sets of nanogels were prepared: P1 imprinted with the carbon-
ate template 8 (i.e. complex 17), and P2 imprinted with the
phosphonate template 6 (i.e. complex 18) (Table 1). All nano-
gels were fully characterised and were found to be highly sol-
uble in water, P1: 9 mgmLÀ1 H2O; P2: 7 mgmLÀ1 H2O, as ex-
pected, given their hydrophilic building blocks. Particle sizes
were determined in diluted aqueous solutions of the nanogels
(0.05 mgpolymermLÀ1) by dynamic light scattering (DLS) (Fig-
ure S7). For the carbonate imprinted polymer (P1) and the
phosphonate imprinted nanogel (P2) the particle dimensions
were 12Æ2 nm and 6Æ1 nm, respectively, measured at 258C
and 1 mgmLÀ1. The preparations were also shown to have low
polydispersity. To determine the incorporation of the mono-
mers into the polymeric matrix and estimate an upper limit for
The trigonal-bipyramidal geometry is very rare for cyclen
compounds, and the use of such strong field ligands normally
results in the easy oxidation of cobalt(II) to cobalt(III), as seen
for the non-polymerisable ligands. However, the common octa-
hedral cis-cyclen cobalt(III) geometry already contains strained
metal to donor bonds which are presumably further strained
through the incorporation of the N-benzyl pendant arm and
the phosphonate, resulting in the ligand now behaving as an
unusually weak polyamine ligand.[40] The data suggest that
subtle changes in ligand and phosphonate structure result in
significant changes to the coordination geometry and oxida-
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3767 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
cobalt(III) cyclen complexes of this type resulted in 25% reduc-
tion in activity.[42] Complex 21, a mimic of the active site within
the nanogel, was therefore prepared. The most convenient
route required the preparation of 22, which was purified by re-
crystallization and then converted into 23, which was again re-
crystallized. The dichloride species 23 was then converted in
situ to the catalytically active complex 21 in HEPES buffer with
two equivalents of 0.1m NaOH. As expected the activity of 21
towards the hydrolysis of 1 was lower, with 67% of the sub-
strate being hydrolysed after 48 h compared to 75% for 4d,
Table 1. Chemical composition and characterisation of MIPs P1 and P2.
Polymer Template Acrylamide EBA CM Size Solubility %Co
[%]
[%] [%] [nm] H2O
content
[mgmLÀ1
]
P1
P2
8
6
10
10
80 0.5 12Æ2
80 0.5 6Æ1
9
7
36
20
EBA=ethylene bis-acrylamide; CM =critical monomer concentration.
the number of active sites, the cobalt content of the nanogel when both complexes were used in excess over the substrate
solutions was analysed by flame atomic absorption spectros- that is, in the complex:substrate ratio 3:1 (Figure S8).
copy (FAAS) and the data showed 36% incorporation for P1
However, if the complexes are not used in excess there is no
and 20% incorporation for P2 (Table S1). These data were then evidence of catalytic activity. Figure 5a shows the total product
used to determine the concentration of polymer to be used in 3 formed after nine days in the uncatalysed reaction compared
the kinetic studies.
to the one catalysed by 4d and 21, respectively. Substrate
To establish whether polymerisation had affected the oxida- 1 was kept constant at 442 mm, while both catalysts were used
tion state of the cobalt in the polymers imprinted with 6, at 100 mm, therefore with a 23% catalytic load. There is no sig-
nanogel solutions of P2 were analysed by EPR spectroscopy nificant difference between the catalysed reactions and the
before and after template removal. The spectra were then background, suggesting that product inhibition occurs, a find-
compared to the spectrum of the monomer complex 18 (Fig- ing consistent with previously reported data for these sys-
ure 3a–c). The EPR data clearly show that the crude P2 con- tems.[43]
taining the phosphonate template still displays a signal at g=
The next step focused on the evaluation of the catalytic ac-
4 consistent with the presence of five-coordinate cobalt(II).[41] tivity of the cobalt–cyclen nanogels P1 and P2. As the charac-
Following template removal, P2 became EPR-silent. Complete terisation data clearly indicate (see Table S1 in the Supporting
release of the template was also confirmed by 31P NMR spec- Information), the content of cobalt differs for the two prepar-
troscopy and supported the formation of the active CoIII aqua- ations, therefore in order to ensure accurate data, experiments
hydroxo species.
were carried out using different amounts of polymers to
The first step towards the kinetic studies focused on the ensure equal concentration of active sites. The reactions were
evaluation of the hydrolytic activity of 4d (Figure 4a) com- carried out under identical experimental conditions, with
pared to an analogue of the imprinted complexes, as it was 442 mm and 664 mm of substrate 1 and 100 mm [Co] for both
previously reported that alkylation at a single nitrogen atom of P1 and P2, which represent 23% and 15% catalytic load re-
Figure 4. a) Mechanism of phosphate hydrolysis catalysed by 4d;[24] b) non-polymerisable analogues 21–23.
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3768 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 5. a) Total product 3 (mm) formed after 9 days in the hydrolysis of phosphate 1 (442 mm), without catalyst, catalysed by 21 (100 mm) and by 4d
(100 mm); b) total product 3 (mm) formed after 14 days in the hydrolysis of 1 (442 mm and 664 mm) without catalyst, catalysed by P1 (100 mm [Co]) and by P2
(100 mm [Co]); c) total product 3 (mm) formed after 9, 14 and 19 days in the hydrolysis of 1 (442 mm), without catalyst, catalyzed by P1 (100 mm [Co]) and by
P2 (100 mm [Co]); d) net product 3 (mm) formed by P1 (100 mm [Co]) and P2 (100 mm [Co]) in the hydrolysis of substrate 1 (442 mm) on day 9, 14 and 19; the
data have been corrected for the uncatalysed reaction.
spectively compared to substrate. Figure 5b shows the total
amount of product 3 formed after 14 days with both nanogel
preparations compared to the background reaction. It is very
clear that both polymers lead to higher product formation,
compared to the uncatalysed reaction, at both substrate con-
centrations. The data also suggest that nanogel P1 is more effi-
cient. To establish whether this trend was consistent over time,
a kinetic experiment was carried out by monitoring total prod-
uct formation at different time intervals for the background re-
action and the ones catalysed by P1 and P2. Figure 5c shows
the results obtained at [S]=442 mm, with [Co]=100 mm for
both nanogels at day 9, day 14 and day 19. In all three sets of
data it can be clearly seen that both nanogels lead to higher
product concentration compared to the background reaction,
and P1, the nanogel imprinted with the carbonate template,
consistently performs better than P2, imprinted with the cog-
nate template. If the amount of product formed as a result of
the background reaction is subtracted from the total product
formed, the net product formed as a result of the catalytic
complex can be obtained. The data, shown in Figure 5d, clearly
demonstrate that not only P1 is the most efficient polymer but
that it operates with turnover, with 130 mm and 158 mm prod-
uct being formed at 14 and 19 days.
the different activity of the two polymers, used with identical
cobalt concentration, can only be the result of morphological
changes occurring during the imprinting process as a result of
the structures of the templates.
This work demonstrates that the nature of the polymerisable
unit and the structure of the template used determine the co-
ordination geometry and oxidation state of the complexes;
both pseudo-octahedral cobalt(III) and trigonal-bipyramidal co-
balt(II) structures can be prepared and their coordination
geometry is maintained throughout the polymerisation pro-
cess. Following the removal of the template, both nanogels
contain catalytically active cobalt(III) aqua-hydroxo species in
the cavities. Given the variation in catalytic activity observed,
we can deduce that the morphological differences in the poly-
meric matrices, resulting from the imprinting approach, are
responsible for the different molecular recognition capabilities.
The nanogel imprinted with the phosphonate in which the
metal centre is five-coordinate increases the activity of the
catalyst significantly compared to the free catalyst. It seems
plausible that this is the result of the template site disfavour-
ing formation of the inactive post-hydrolysis complex 20, Fig-
ure 4a, which is known to inhibit the catalytic cycle. The nano-
gel imprinted with the carbonate P1, shows even higher cata-
lytic efficiency and turnover, despite the carbonate being
considerably smaller than the cognate template and contain-
ing a pseudo-octahedral active site. Initially this result appears
to be rather counter intuitive and surprising, however, analysis
of the crystal structure of an analogue of [Co(15b)(CO3)]Cl re-
This result is very significant because it demonstrates that
the same catalyst, which is required in large excess to acceler-
ate the chemical reaction, when incorporated into a polymeric
matrix via the molecular imprinting approach, can be success-
fully used in catalytic quantities giving turnover. Interestingly,
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3769 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
veals the O-Co-O angle to be rather strained (O(1)-Co(1)-O(2)=
68.54(9)8),[44] which is very similar to that seen in other co-
balt(III) complexes containing bidentate carbonate ligands in
the Cambridge Crystallographic Database.[45] In contrast, the
equivalent angle in cobalt(III) phosphate chelates is consider-
ably larger at 768.[46]
This suggests that imprinting with carbonate produces an
active site in the polymer, which further enhances catalytic ac-
tivity. A combination of factors is likely to contribute to these
results. The first is manifested in both nanogels and is due to
the lower polarity within the polymeric matrices, which leads
to higher product formation compared to 21. The second
effect results from the significantly more strained coordination
geometry for 17, which results from the tight bite angle of the
carbonate ligand; this may cause a difference in stability of the
hydrolysis product 2–polymer complexes, which would then
lead to increased activity.
solvent purification system. To remove traces of water from
EtOH, the solvent was dried over 3 Š molecular sieves under
N2. Distilled water was obtained from an Elga Purelab Option
system. Reagents were purchased commercially form Sigma–
Aldrich and CheMatech. Nitrogen used for inert atmosphere
was oxygen-free grade. All glassware, glass syringes and metal
needles were oven-dried and cooled prior to experimental use.
Infrared spectra were recorded in the range 4000–600 cmÀ1
,
obtained directly from the compound as a solid on a Bruker
Tensor 37 FTIR spectrometer. 1H NMR, 13C NMR, and 31P NMR
spectra were recorded using three different NMR instruments:
a Jeol JNM-EX spectrometer (1H NMR : 270 MHz; 13C NMR:
67.5 MHz; 31P NMR: 162 MHz), a Bruker AV400 or a Bruker
AMX400 (1H NMR : 400 MHz; 13C NMR: 100 MHz; 31P NMR:
162 MHz). For 31P NMR spectra, an external reference spectrum
was acquired before each experiment using H3PO4 (85.0%) as
standard (0.00 ppm). Deuterated solvents and equipment used
to record the spectra are stated before each set of data. Chem-
ical shifts are reported in ppm and referenced to residual
protonated solvent. Multiplicity is given as follows: s=singlet,
d=doublet, t=triplet, q=quartet, dd=doublet of doublets,
m=multiplet, bs=broad signal, and coupling constants meas-
ured in Hertz (Hz) and reported to 1 d.p.
Conclusion
This work represents the first example of a complex of [Co(cyc-
len)(OH)(OH2)]2+, incorporated into a polymeric nanogel with
the molecular imprinting approach, that catalyses the hydroly-
sis of a nerve agent mimic with good turnover, using a low
catalytic load (15%). The increased activity of the polymerically
embedded complexes when compared to the free catalyst ap-
pears to result from a combination of factors. The local coord-
ination environment at the cobalt centre clearly plays a key
role, as evidenced by the significant differences in activity be-
tween P1 and P2. However, the fact that both imprinted poly-
mers display catalytic turnover is also likely to be a result of
the different local environment within the polymeric nanogel
which has a lower polarity compared to bulk aqueous solution;
an effect that has been previously observed as an enhancing
factor in related systems,[4a,8] Furthermore this study clearly
demonstrates the significant variation that can occur in the co-
ordination chemistry and geometry of the active metal ion as
result of interactions with structurally similar analogues. These
ultimately determine the molecular recognition characteristics
and properties of the nanoparticles. In previous reports, in-
depth studies of the coordination geometry of the complex
prior to polymerisation have seldom been undertaken,[21] so
that factors that we have shown to occur such as ligand ex-
change, altered coordination geometry and even oxidation
state have not been robustly investigated. These results pro-
vide evidence that polymerised biomimetic metal complexes,
with well-defined coordination chemistry, have the potential to
deliver viable alternatives when natural proteins are not avail-
able.
UV/Vis spectra were obtained on a HP 8453 spectrophotom-
eter, absorption maxima (lmax) are expressed in nm, the molar
extinction coefficients (e) are expressed in mÀ1 cmÀ1. Electro-
spray ionisation mass spectrometry was carried out by the
EPSRC National Mass Spectrometry Service, University of Wales,
Swansea on a Thermofisher LTQ Orbitrap XL. Melting points
were measured on a Stuart SMP3 melting point apparatus and
are uncorrected. Continuous wave (CW) X-band EPR spectra
were obtained at 18Æ0.2 K using a Bruker ELEXSYS E500 EPR
spectrometer equipped with an Oxford Instruments ESR900
helium flow cryostat. Additional experimental parameters
were: 0.5 mW microwave power; 100 KHz field modulation;
0.5 mT field modulation. Kinetic studies were carried out by
UV/Vis spectroscopy using a HP 8453 spectrophotometer. p-
Ethyl benzyl alcohol and p-ethyl benzyl bromide were pre-
pared according to a modified procedure.[31,32]
Synthesis
Perhydro-2a,4a,6a,8a-tetraazacyclopenta acenaphthylene
(bis-aminal): Cyclen 13 was protected as its bis-aminal by
a slight modification of the reported procedure.[29] A solution
of glyoxal (as a 40% aqueous solution, 6.00 mL, 52.0 mmol) in
MeOH (15.0 mL) was added dropwise at 08C to a solution of
cyclen (8.00 g, 46.4 mmol) in MeOH (80.0 mL). The mixture was
stirred at this temperature for 1 h, heated at 558C for a further
2 h, and finally stirred at room temperature for 16 h. The solu-
tion was concentrated in vacuo to afford a dark orange oil,
which was then triturated with Et2O (5”50.0 mL). The organic
fractions were combined and the solvent was removed in
vacuo to give the bis-aminal as a white powder (7.43 g, 83%).
All data were consistent with those previously reported.
1H NMR (400 MHz, CDCl3): d=2.50–2.54 (m, 4H, CH2), 2.65 (bs,
Experimental Section
Materials
All commercially available reagents and solvents were used
without further purification, unless otherwise stated. Anhyd-
rous solvents were obtained using an MBraun MB SPS-800
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3770 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
4H, CH2), 2.90–2.99 (m, 8H, CH2), 3.10 ppm (bs, 2H, CHaminal);
1H, CHaminal), 4.02 (d, J=4.0 Hz, 1H, CHaminal), 4.18–4.23 (m, 1H,
CH2), 4.68–4.86 (m, 2H, residual HOD solvent peak overlaps
signal, N-CH2), 7.44 (d, J=8.0 Hz, 2H, Ar-H), 7.51 ppm (d, J=
8.0 Hz, 2H, Ar-H); 13C NMR (100 MHz, D2O): d=14.7, 28.0, 43.8,
47.5, 48.3, 51.3, 57.2, 61.2, 61.4, 62.5, 71.7, 82.3, 90.5, 124.0,
129.0, 132.5, 148.1 ppm; IR: nmax =2961, 2883, 2848, 1612,
1514, 1433, 1266, 1219, 1184, 1134, 1053, 1027, 982, 929,
846 cmÀ1; HRMS (EI) calcd for C19H29N4 [M+H]+ 313.2387,
found: 313.2383.
13C NMR (DEPT 135, 100 MHz, CDCl3): d=49.6, 50.4, 76.8 ppm.
General procedure for mono-alkyation of the bis-aminal as
demonstrated by the synthesis of 14a (Precursor to 15a)
(1RS,13SR,14RS)-1-allyl-4,7,10-triaza-1-azoniatetracyclo tetra-
decane bromide: An excess of allyl bromide (R-X halide)
(1.50 mL, 18.0 mmol) was added under an inert atmosphere to
a solution of the bis-aminal (1.75 g, 9.00 mmol) dissolved in
the minimum volume of dry toluene (10.0 mL). The mixture
was maintained at room temperature for 2 h until a precipitate
formed. The resulting white solid was collected by filtration
and was washed using freshly distilled toluene (2”5 mL). The
collected solids were dried in vacuo to afford the desired prod-
uct as a white amorphous solid (hygroscopic) (2.90 g, 94%).
1-Allyl-1,4,7,10-tetraazacyclododecane 15a:[33] Compound
14a (0.385 g, 1.22 mmol) was dissolved in dry EtOH (6 mL). Hy-
droxylamine degassed in a 50% aqueous solution (0.360 mL,
12.2 mmol) was added dropwise to this solution and the re-
sulting mixture was stirred for 20 min at room temperature
and then for 120 min at about 508C, making the solution turn
from cloudy white to clear pale yellow. After cooling, the reac-
tion mixture was stirred for a further hour at room tempera-
ture before an equal volume of 10% w/w KOH was added. The
mixture was extracted with DCM (5”10.0 mL) and the organic
fractions were combined, dried over MgSO4, and removed in
vacuo to afford the product as a pale yellow oil (0.190 g, 73%).
1H NMR (270 MHz, CDCl3): d=2.50–2.61 (m, 12H, CH2), 2.71–
2.75 (m, 4H, CH2), 3.06 (d, J=6.5 Hz, 2H, NCH2-CH), 5.06–5.13
(m, 2H, CH=CHHcis, CH=CHHtrans), 5.79 ppm (m, 1H, CH=CH2);
13C NMR (67.5 MHz, D2O): d=44.8, 45.2, 45.8, 46.8, 50.8, 51.3,
57.4, 117.4, 135.3 ppm; IR (CHCl2, NaCl): nmax =3658, 3250,
3053, 2961, 2852, 2305, 1451, 1421, 1265, 734 cmÀ1.
1
The salt exhibits extremely complex H NMR spectra, as already
reported.[30] 1H NMR (400 MHz, D2O): d=2.47–2.57 (m, 2H,
CH2), 2.77–2.97 (m, 5H, CH2), 3.18–3.34 (m, 4H, CH2), 3.51–3.53
(m, 1H, CH2), 3.61 (bs, 1H, CHaminal), 3.65–3.81 (m, 3H, CH2),
3.96 (bs, 1H, CHaminal), 4.02–4.03 (m, 1H, CH2), 4.12 (dd, J=13.1,
5.9 Hz, 1H, NCH2-CH), 4.33 (dd, J=13.2, 7.9 Hz, 1H, NCH2-CH),
5.72–5.78 (m, 2H, CH=CHHcis, CH=CHHtrans), 6.03–6.13 ppm (m,
1H, CH=CH2); 13C NMR (100 MHz, D2O): d=43.6, 47.6, 47.8,
48.1, 48.3, 51.2, 57.5, 60.7, 61.6, 71.7, 82.7, 124.2, 129.1 ppm;
ESIMS: m/z (%): 235 (100) [M+]; HRMS (EI) calcd for C13H23N4
[M+H]+ 235.1917, found 235.1918.
(1RS,13SR,14RS)-1-(4-Vinylbenzyl)-4,7,10-triaza-1-azonia-
tetracyclo tetradecane bromide 14b (Precursor to 15b): Fol-
lowing the general procedure, bis-aminal (2.06 g, 10.6 mmol)
and 1-(chloromethyl)-4-vinylbenzene (3.00 mL, 21.2 mmol) with
a 16 h reaction time at room temperature yielded a white solid
1-(4-Vinylbenzyl)-1,4,7,10-tetraazacyclododecane
15b:[33]
Compound 14b (0.500 g, 1.44 mmol) was dissolved in dry
EtOH (6 mL). Hydroxylamine in a 50% aqueous solution
(0.420 mL, 14.4 mmol) was added dropwise to this solution
and the resulting mixture was stirred for 20 min at room tem-
perature and then for 16 h at about 608C, making the solution
turn from cloudy white to clear pale yellow. After cooling, the
reaction mixture was stirred for a further hour at room
temperature before an equal volume of 10% w/w KOH was
added. The mixture was extracted with CHCl3 (5”10 mL) and
the organic fractions were combined, dried over MgSO4, and
removed in vacuo to afford the crude product as a yellow oil
(0.400 g, 96%). This crude product contains some minor impur-
ities and required further purification. The crude oil was sus-
pended in EtOH (1 mL), and HCl (35m, 2 mL) was added drop-
wise to obtain a hydrochloride salt of 15b as a white precipi-
tate. This precipitate was filtered, dissolved in a solution of
KOH (40% aqueous solution, w/v, 10 mL), and extracted with
CHCl3 (5”30 mL). The organic layers were combined, dried
over MgSO4, and removed in vacuo to afford 15b as a yellow
1
(hygroscopic) (3.12 g, 95%). The salt exhibits complex H NMR
spectra, as already reported.[29] 1H NMR (400 MHz, D2O): d=
2.49–2.56 (m, 2H, CH2), 2.76–2.88 (m, 3H, CH2), 2.92–2.99 (m,
1H, CH2), 3.07–3.20 (m, 2H, CH2), 3.27–3.37 (m, 3H, CH2), 3.47–
3.66 (m, 4H, CH2), 3.75 (d, J=4.0 Hz, 1H, CHaminal), 4.02 (d, J=
4.0 Hz, 1H, CHaminal), 4.19–4.23 (m, 1H, CH2), 4.68–5.87 (m, 2H,
residual HOD solvent peak overlaps signal, N-CH2), 5.44 (d, J=
11.0 Hz, 1H, CH=CHHcis), 5.96 (d, J=17.6 Hz, 1H, CH=CHHtrans),
6.85 (dd, J=11.0, 17.6 Hz, 1H, CH=CH2), 7.57 (d, J=8.0 Hz, 2H,
Ar-H), 7.67 ppm (d, J=8.0 Hz, 2H, Ar-H); 13C NMR (100 MHz,
D2O): d=43.9, 47.5, 47.6, 48.2, 48.3, 51.2, 57.2, 61.3, 71.7, 82.5,
90.4, 116.5, 126.5, 127.5, 133.2, 135.9, 140.4 ppm; IR: nmax
=
2945, 2750, 1620, 1135, 1055 cmÀ1. ESIMS: m/z (%): 311 (55)
[M+], 193 (100), 117 (80); HRMS (EI) calcd for C19H27N4 [M+H]+
311.2230, found: 311.2233.
(4-Ethylbenzyl)-4,7,10-triaza-1-azoniatetracyclo tetradecane
bromide 14c (Precursor to 15c): Following general procedure,
bis-aminal (2.00 g, 10.3 mmol) and 4-ethyl benzyl bromide
(2.05 g, 10.3 mmol) with a 20 h reaction time at room tempera-
ture yielded a white solid (3.70 g, 91%). m.p. 165–1688C.
1H NMR (400 MHz, D2O): d=1.23 (t, J=8.0 Hz, 3H, CH3), 2.48–
2.56 (m, 2H, CH2), 2.72 (q, J=8.0 Hz, 2H, CH2), 2.77–2.88 (m,
3H, CH2), 2.92–3.00 (m, 1H, CH2), 3.08–3.19 (m, 2H, CH2), 3.25–
3.37 (m, 3H, CH2), 3.45–3.77 (m, 4H, CH2), 3.77 (d, J=4.0 Hz,
1
oil (60.0 mg,). H NMR (400 MHz, CDCl3): d=2.58–2.83 (m, 16H,
CH2), 3.61 (s, 2H, N-CH2), 5.18 (d, J=10.9 Hz, 1H, CH=CHHcis),
5.69 (d, J=17.7 Hz, 1H, CH=CHHtrans), 6.67 (dd, J=10.9, 17.7 Hz,
1H, CH=CH2), 7.25 (d, residual solvent peak overlaps signal, J=
8.0 Hz, 2H, Ar-H), 7.34 ppm (d, J=8.0 Hz, 2H, Ar-H); 13C NMR
(100 MHz, CDCl3): d=45.2, 46.1, 47.2, 51.3, 59.2, 113.5, 126.2,
129.3, 136.9, 138.3, 146.3 ppm; IR: nmax =2807, 2363, 1628,
1567, 1450, 1348, 1263 cmÀ1.
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3771 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1-(4-Ethylbenzyl)-1,4,7,10-tetraazacyclododecane 15c: Com-
pound 14c (0.800 g, 2.03 mmol) was dissolved in dry EtOH
(10 mL). Hydroxylamine in a 50% aqueous solution (0.550 mL,
20.3 mmol) was added dropwise to this solution and the re-
sulting mixture was stirred for 20 min at room temperature
and then for 16 h at about 608C, making the solution turn
from cloudy white to clear pale yellow. After cooling, the reac-
tion mixture was stirred for a further hour at room tempera-
ture. The reaction mixture was then reduced in volume to
afford a yellow oil, which was dissolved in HCl (5.00m, 2 mL).
The aqueous layer was washed with EtOAc (20 mL ” 2) and
Et2O (20 mL ” 2). The resulting aqueous layer was then adjust-
ed to pH 14.0 with KOH (40% aqueous solution, w/v, 2 mL)
and extracted with CHCl3 (5”30 mL). Organic fractions were
then combined, dried over MgSO4, and removed in vacuo to
afford the crude product as a yellow oil. After suspending the
oil in EtOH (1 mL), HCl (35m, 2 mL) was added dropwise to
obtain a hydrochloride salt of 15c as a white precipitate. This
precipitate was collected by filtration, dissolved in a solution of
KOH (40% aqueous solution, w/v, 10 mL), and extracted with
CHCl3 (5”30 mL). The organic layers were combined, dried
over MgSO4, and concentrated in vacuo to afford 15c as
HRMS (ESI): calcd for C12H24O3N4Co [M+H+] 331.1175, found:
331.1175.
cis-[Co(15b)CO3]HCO3 17: Following the general procedure,
cyclen
derivative
15b
(0.216 g,
0.750 mmol)
and
Na3[Co(CO3)3]·3H2O (0.271 g, 0.750 mmol) with a reaction time
of 16 h yielded complex 17 as a burgundy red powder
1
(0.237 g, 67%). H NMR (400 MHz, D2O): d=2.60–2.76 (m, 3H,
CH2), 2.85-3.23, (m, 12H, CH2), 3.40-3.45 (m, 2H, CH2), 3.83 (d,
JAB =12.0 Hz, 1H, N-CH2), 4.00 (d, JAB =12.0 Hz, 1H, N-CH2), 5.39
(d, J=11.0 Hz, 1H, CH=CHHcis), 5.92 (d, J=17.7 Hz, 1H, CH=
CHHtrans), 6.84 (dd, J=11.0, 17.7 Hz, 1H, CH=CH2) 7.44 (d, J=
8.0 Hz, 2H, Ar-H), 7.58 ppm (d, J=8.0 Hz, 2H, Ar-H). IR: nmax
3089, 2890, 1613, 1449, 1344, 1291, 828, 750 cmÀ1; UV/Vis
(DMSO): lmax (e)=539 nm (213 dm3 molÀ1 cmÀ1); ESIMS m/z (%):
407.1 (43) [M+], 351.2 (55), 701.4 (100); HRMS (EI) calcd for
C18H28O3N4CoÀHCO3 [M+H]+ 407.1488, found 407.1487.
cis-[Co(15c)CO3]HCO3 22: Following the general procedure,
cyclen
derivative
15c
(95.0 mg,
0.330 mmol)
and
Na3[Co(CO3)3]·3H2O (0.120 g, 0.330 mmol) with a reaction time
of 16 h yielded complex 22 as a dark pink powder (0.125 g,
1
1
a yellow oil (0.106 g, 18%). H NMR (400 MHz, CDCl3): d=1.21
81%). m.p. 184–1868C. H NMR (400 MHz, D2O): d=1.23 (t, J=
(t, J=8.0 Hz, 3H, CH3), 2.58–2.69 (m, 14H, CH2), 2.81–2.83 (m,
4H, CH2), 3.59 (s, 2H, N-CH2), 7.13 (d, J=8.0 Hz, 2H, Ar-H),
7.20 ppm (d, J=8.0 Hz, 2H, Ar-H); 13C NMR (100 MHz, CDCl3):
d=15.6, 28.6, 45.5, 46.4, 47.6, 51.3, 59.2, 127.9, 129.1, 136.2,
143.1 ppm; IR: nmax =3280, 2928, 1671, 1511, 1455, 1349, 1263,
1113, 1041, 931, 818 cmÀ1; HRMS (EI) calcd for C17H31N4 [M+
H]+ 291.2543, found 291.2542.
8.0 Hz, 3H, CH3), 2.58–2.77 (m, 6H, CH2), 2.86–3.23 (m, 10H,
CH2), 3.44–3.46 (m, 2H, CH2), 3.84 (d, JAB =16.0 Hz, 1H, N-CH2),
3.98 (d, JAB =12.0 Hz, 1H, N-CH2), 7.39 ppm (m, 4H, Ar-H);
13C NMR (100 MHz, D2O): d=14.9, 27.9, 47.1, 47.5, 48.4, 49.4,
53.1, 55.6, 55.8, 56.9, 63.1, 126.9, 128.2, 132.6, 146.3,
167.0 ppm; IR: nmax =3379, 3112, 2882, 1957, 1450, 1337, 1264,
1058, 1016, 976, 828, 750 cmÀ1; UV/Vis (HPLC MeOH): lmax (e)=
366 (472), 540 nm (542 dm3 molÀ1 cmÀ1); HRMS (EI) calcd for
C18H30CoN4O3 [M+H]+ 409.1644, found 409.1636.
General procedure for the synthesis of CoIIICO3 complexes,
as demonstrated by the synthesis of cis-[Co(cyclen)CO3]HCO3
5:[47] Cyclen 13 (0.193 g, 0.550 mmol) was dissolved in
a MeOH:H2O mixture (1:1) (6 mL) and an equimolar amount of
Na3[Co(CO3)3]·3H2O (0.200 g, 0.550 mmol) was added. The dark
green solution gradually turned burgundy-red and was left to
react for 16 h at 658C. The solution was filtered whilst hot
under suction to separate the liquid from a black solid. The fil-
trate was dried in vacuo, redissolved in MeOH (6 mL), and the
resulting solution was filtered to remove white precipitate. The
resulting filtrate was reduced in volume in vacuo and excess
Et2O (30 mL) was added. The resulting solids were collected by
filtration and washed with Et2O (10 mL) to afford 5 as a pink
General procedure for the synthesis of CoIIICl2 complexes as
demonstrated by the synthesis of cis-[Co(cyclen)Cl2]Cl 4a:[34]
Complex 5 (0.205 g, 0.580 mmol) was dissolved in MeOH
(8 mL) and HCl (35m, 2 mL) was added dropwise to this solu-
tion. The reaction mixture was reduced to dryness and the resi-
due was again dissolved in MeOH (8 mL) and treated with HCl
(35m, 2 mL). This procedure was repeated five times and
a gradual colour change from dark pink to dark violet was ob-
served. The final solution was concentrated in vacuo and the
resulting violet solid was washed with Et2O (20 mL) to afford
4a as
a
violet crystalline solid (0.220 g, 75%). 1H NMR
1
powder (0.170 g, 87%). H NMR (400 MHz, D2O): d=2.47–3.20
(400 MHz, [D6]DMSO): d=2.25–3.45 (m, 16H, CH2), 7.48 (bs, 1H,
NH), 7.59 (bs, 1H, NH), 7.95 ppm (bs, 1H, NH); 13C NMR
(m, 14H, CH2), 3.55–3.62 ppm (m, 2H, CH2); IR: nmax =3157,
3080, 2959, 2896, 1619, 1440, 1405, 1346, 1242, 834 cmÀ1; UV/
Vis (H2O): lmax (e)=525 (271), 368 nm (200 dm3 molÀ1 cmÀ1).
(100 MHz, [D6]DMSO): d=46.3, 49.7, 54.3, 57.5 ppm; IR: nmax
=
3210, 3172, 3050, 2870, 1645, 1477, 1444, 1356, 1112,
1057 cmÀ1; UV/Vis (DMSO): lmax (e)=561 (134), 390 nm (53
dm3 molÀ1 cmÀ1); ESIMS: m/z (%): 301.0 (100) [M+], 358.9 (15),
545.4 (25); HRMS (EI) calcd for C8H20N4CoCl2 [M+H]+ 301.0392,
found 301.0392.
cis-[Co(15a)CO3]HCO3 16: Following the general procedure,
cyclen
derivative
15a
(0.161 g,
0.757 mmol)
and
Na3[Co(CO3)3]·3H2O (0.274 g, 0.757 mmol) with a reaction time
of 16 h yielded complex 16 as a dark red powder (89.0 mg,
1
30%). H NMR (400 MHz, D2O): d=2.61–3.49 (m, 18H, CH2, N-
cis-[Co(15c)Cl2]Cl 21b: Following the general procedure, car-
bonate complex 22 (0.120 g, 0.250 mmol) yielded complex 23
as a dark violet solid (0.129 g, 86%). m.p. 195–1988C; H NMR
CH2), 5.48 (m, 2H, CH=CH2), 6.09 ppm (dd, J=9.8, 17.7 Hz, 1H,
CH-CH2); UV/Vis (H2O): lmax (e)=536 (215), 350 nm (213
dm3 molÀ1 cmÀ1); ESI m/z (%): 331.2 (100) [M+], 269.2 (30);
1
(400 MHz, MeOD): d=1.23 (t, J=8.0 Hz, 3H, CH3), 2.52–2.73 (m,
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3772 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
9H, CH2), 3.32–3.79 (m, 16H, CH2, residual MeOD proton peak
overlaps signal), 4.52 (dd, JAB =4.0 Hz, 16.0 Hz, 2H, N-CH2), 7.27
(d, J=4.0 Hz, 2H, Ar-H), 7.32 ppm (d, J=8.0 Hz, 2H, Ar-H);
13C NMR (100 MHz, D2O): d=16.0, 29.5, 51.0, 51.2, 50.2, 58.3,
59.4, 64.0, 65.1, 129.4, 130.3, 133.5, 146.6 ppm; IR: nmax =3056,
2962, 2870, 1613, 1447, 1098, 1063, 1008, 970, 828 cmÀ1; UV/
Vis (MeOH): lmax (e)=570 (224), 395 nm (264 dm3 molÀ1 cmÀ1);
HRMS (EI) calcd for C17H28CoN4 [M+H]+ 347.1640, found
347.1639.
1437, 1134, 1057, 1005, 920, 752, 697, 616, 810, 712 cmÀ1; UV/
Vis (H2O): lmax (e)=372 (2126), 532 nm (307 dm3 molÀ1 cmÀ1);
UV/Vis (DMSO): lmax (e)=529 nm (155 dm3 molÀ1 cmÀ1); ESI:
m/z (%): 387.1 (100) [M+]; HRMS (EI) calcd for C14H25N4O3CoP
[M+H]+ 387.1007, found 387.0989.
cis-[Co(15b)(6)]Cl 18: Following the general procedure, cyclen
derivative 15b (0.484 g, 1.68 mmol) with phosphonate 6
(0.363 g, 1.68 mmol) and CoCl2·6H2O (0.398 g, 1.68 mmol)
yielded 18 as a blue-green powder (0.869 g, 91%). This proced-
ure results in the reproducible formation of the same mixture
of compounds and the data presented should be regarded as
being diagnostic, rather than corresponding to a single species
with complete bulk purity. m.p. 120–1258C. IR: nmax =1601,
1513, 1344, 1247, 1153, 1109, 1035, 919, 858, 772, 729,
696 cmÀ1; UV/Vis (DMSO): lmax (e)=556 (78), 602 (91), 615 (90),
655 nm (84 dm3 molÀ1 cmÀ1); ESIMS: m/z (%): 737.4 (40) [M+],
562.2 (50), 551.2 (100); HRMS (EI) cacld. for C24H34O5N5CoP [M+
H]+ 562.1624, found 562.1569.
4-Nitrophenyl phosphonate (6):[48] An excess of bromotri-
methylsilane (4.50 mL, 34.1 mmol) was added dropwise at 08C
to a solution of diethyl 4-nitrobenzylphosphonate (3.30 mL,
15.0 mmol) in acetonitrile (20 mL). The ice bath was removed
and the mixture was stirred for 24 h at room temperature. The
resulting dark yellow solution was concentrated in vacuo and
the residue was dissolved in MeOH (20 mL). Stirring was main-
tained for a further 24 h at room temperature, after which
time the solvent was removed in vacuo, leaving a beige solid.
The solid was repeatedly washed with DCM (30 mL) and then
dried in vacuo to give 6 as an off-white solid (3.00 g, 92%).
1
m.p. 224–2268C; H NMR (400 MHz, [D6]DMSO): d=3.16 (d, J=
EPR spectroscopy
22.0 Hz, 2H, CH2), 7.51 (dd, J=8.8, 2.3 Hz, 2H, Ar-H), 8.20 (d,
J=8.3 Hz, 2H, Ar-H), 9.75 ppm (bs, 2H, OH); 13C NMR
(100 MHz, D2O): d=35.1, 123.7, 130.4, 142.1, 146.3 ppm;
31P NMR (161 MHz, [D6]DMSO): d=19.37 ppm; 31P NMR
(161 MHz, CD3OD): d=22.07 ppm; 31P NMR (161 MHz, D2O):
d=22.38 ppm; IR: nmax =2613, 1599, 1520, 1344, 1268, 944,
EPR spectra were recorded at X-band (approximately 9.4 GHz
on the spectrometer employed) using a Bruker ELEXSYS E500/
E580 spectrometer. Temperature control was provided by an
Oxford Instruments ESR900 liquid helium cryostat and an
ITC503 temperature controller. The spectra were recorded at
the Manchester Interdisciplinary Biocentre, University of Man-
chester (UK).
771, 695 cmÀ1
.
General procedure for the synthesis of phosphonate com-
plexes as demonstrated by the synthesis of cis-[Co(cyc-
len)NPPA]Cl 9: Cyclen 13 (48.2 mg, 0.280 mmol) was dissolved
in degassed MeOH (2 mL), and an equimolar solution of
CoCl2·6H2O (66.4 mg, 0.280 mmol) in MeOH (5 mL) was added
under N2, causing the solution to turn brown and shortly after
pale violet. After stirring the mixture for 5 min, an equimolar
solution of the phosphonate 6 (60.5 mg, 0.280 mmol) in MeOH
(5 mL) was added under an inert atmosphere, and the solution
turned to dark purple. Stirring under N2 was maintained for 5 h
at room temperature, after which the solution was bubbled
with compressed air for an equal amount of time. The solvent
was removed in vacuo and the resultant solids were collected
by filtration and washed with Et2O (10 mL), yielding 9 as
a purple solid (108 mg, 84%). m.p. 2408C (decomposes);
Cobalt content characterisation by atomic absorption (FAAS)
Solutions of nanogels (0.30 mgmLÀ1) in at least 3.00 mL of
water were prepared and tested for Co absorbance against
a set of five cobalt standard calibration solutions (1.00–
5.00 mgmLÀ1), showing absorbance values in the 2.00–
4.00 mgmLÀ1 range. The amount of cobalt present in solution
was taken from the absorbance values, which were converted
into concentration using the Beer–Lambert Law. The yield of
metallomonomer incorporated was calculated by dividing the
amount of cobalt found by atomic absorption by the theoret-
ical amount initially introduced into the polymer mixture.
31P NMR (161 MHz, D2O): d=32.1 ppm. IR: nmax =2855, 1597, Acknowledgements
1511, 1452, 1343, 1237, 1058, 1007, 918, 858, 813, 768, 727,
695 cmÀ1
;
UV/Vis
(DMSO):
lmax
(e)=548 nm
(160
The authors wish to acknowledge the financial support of the
European Commission (MCRTN-2006-033873 and MCIAPP-
2009-251307), the EPSRC for the provision of a studentship
(M.C.) and for the National Mass Spectrometry Service, Swan-
sea University. We thank M. Motevalli for his support with X-
ray crystallography data.
dm3 molÀ1 cmÀ1); ESIMS: m/z (%): 446.1 (100) [M+], 663.1 (25);
HRMS (EI) calcd for C15H26N5O5PCo [M+H]+ 446.1015, found
446.098.
cis-[Co(cyclen)(7)]Cl 10: Following the general procedure,
cyclen 13 (29.2 mg, 0.170 mmol) with phosphonate 7 (26.9 mg,
0.170 mmol) and CoCl2·6H2O (40.3 mg, 0.170 mmol) yielded 10
as a purple solid (50.0 mg, 68%). m.p. 224–2258C; 31P NMR
(161 MHz, D2O): d=32.1 ppm. IR: nmax =2858, 1638, 1479,
Keywords: cyclen · enzyme mimics · molecular imprinting ·
nanogels · phosphatase
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3773 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[1] F. Mancin, P. Scrimin, P. Tecilla, U. Tonellato, Chem. Commun. 2005,
2540–2548.
[22] A. N. Cammidge, N. J. Baines, R. K. Bellingham, Chem. Commun. 2001,
2588–2589.
[2] F. Mancin, P. Scrimin, P. Tecilla, Chem. Commun. 2012, 48, 5545–5559.
[3] H. J. Schneider, A. K. Yatsimirsky, Met. Ions Biol. Syst. 2003, 369–462.
[4] a) M. Diez-Castellnou, F. Mancin, P. Scrimin, J. Am. Chem. Soc. 2014, 136,
1158–1161; b) O. Iranzo, A. Y. Kovalevsky, J. R. Morrow, J. P. Richard, J.
Am. Chem. Soc. 2003, 125, 1988–1993; c) A. Tamilselvi, G. Mugesh,
Chem. Eur. J. 2010, 16, 8878–8890.
[5] S. J. Franklin, Curr. Opin. Chem. Biol. 2001, 5, 201–208.
[6] D. R. Edwards, C. T. Liu, G. E. Garrett, A. A. Neverov, R. S. Brown, J. Am.
Chem. Soc. 2009, 131, 13738–13748.
[23] M. Shibata, Modern Synthesis of CoIII Complexes, Springer, New York,
1993.
[24] J. Chin, M. Banaszczyk, V. Jubian, X. Zou, J. Am. Chem. Soc. 1989, 111,
186–190.
[25] R. Hettich, H.-J. Schneider, J. Am. Chem. Soc. 1997, 119, 5638–5647.
[26] C.-S. Jeung, J. B. Song, Y.-H. Kim, J. Suh, Bioorg. Med. Chem. Lett. 2001,
11, 3061–3064.
[27] J. Chin, Acc. Chem. Res. 1991, 24, 145–152.
[28] M. Resmini, R. Vigna, C. Simms, N. J. Barber, E. P. Hagi-Pavli, A. B. Watts,
C. Verma, G. Gallacher, K. Brocklehurst, Biochem. J. 1997, 326, 279–287.
[7] E. GourØ, M. Carboni, A. Troussier, C. Lebrun, J. PØcaut, J.-F. Jacquot, P.
Dubourdeaux, M. ClØmancey, G. Blondin, J.-M. Latour, Chem. Eur. J.
2015, 21, 8064–8068.
ˇ
´
ˇ
[29] J. Rohovec, R. Gyepes, I. Císarovµ, J. Rudovsky, I. Lukes, Tetrahedron Lett.
2000, 41, 1249–1253.
[8] A. A. Neverov, Z.-L. Lu, C. I. Maxwell, M. F. Mohamed, C. J. White, J. S. W.
Tsang, R. S. Brown, J. Am. Chem. Soc. 2006, 128, 16398–16405.
[9] D. Ajami, J. Rebek, Org. Biomol. Chem. 2013, 11, 3936–3942.
[10] M. Zhao, S.-S. Xue, X.-Q. Jiang, L. Zheng, L.-N. Ji, Z.-W. Mao, J. Mol. Catal.
A 2015, 396, 346–352.
[11] a) S. C. Maddock, P. Pasetto, M. Resmini, Chem. Commun. 2004, 536–
537; b) P. Pasetto, S. C. Maddock, M. Resmini, Anal. Chim. Acta 2005,
542, 66–75.
[30] a) H. Morales-Rojas, R. A. Moss, Chem. Rev. 2002, 102, 2497–2522;
b) D. G. Oakenfull, D. I. Richardson, D. A. Usher, J. Am. Chem. Soc. 1967,
89, 5491–5492.
[31] D. E. Ward, C. K. Rhee, Can. J. Chem. 1989, 67, 1206–1211.
[32] C. A. Martin, P. M. McCrann, M. D. Ward, G. H. Angelos, D. A. Jaeger, J.
Org. Chem. 1984, 49, 4392–4396.
[33] J. Rawlings, A. C. Hengge, W. W. Cleland, J. Am. Chem. Soc. 1997, 119,
542–549.
[12] a) D. Carboni, K. Flavin, A. Servant, V. Gouverneur, M. Resmini, Chem.
Eur. J. 2008, 14, 7059–7065; b) M. Resmini, Anal. Bioanal. Chem. 2012,
402, 3021–3026.
[34] J. P. Collman, P. W. Schneider, Inorg. Chem. 1966, 5, 1380–1384.
[35] S. Aoki, A. Jikiba, K. Takeda, E. Kimura, J. Phys. Org. Chem. 2004, 17,
489–497.
[13] a) A. Servant, K. Haupt, M. Resmini, Chem. Eur. J. 2011, 17, 11052–
11059; b) A. Servant, S. Rogers, A. Zarbakhsh, M. Resmini, New J. Chem.
2013, 37, 4103–4109; c) P. Bonomi, A. Servant, M. Resmini, J. Mol. Rec-
ognit. 2012, 25, 352–360.
[36] W. C. Baker, M. J. Choi, D. C. Hill, J. L. Thompson, P. A. Petillo, J. Org.
Chem. 1999, 64, 2683–2689.
[37] T. Alfrey, J. G. Harrison, J. Am. Chem. Soc. 1946, 68, 299–301.
[38] R. C. Laible, Chem. Rev. 1958, 58, 807–843.
[14] J. E. W. Scheuermann, K. F. Sibbons, D. M. Benoit, M. Motevalli, M. Wat-
kinson, Org. Biomol. Chem. 2004, 2, 2664–2670.
[39] C. M. Sarther, E. L. Blinn, Inorg. Chem. 1976, 15, 3083–3086.
[40] D. F. Evans, J. Chem. Soc. 1959, 2003–2005.
[15] a) G. Ilyashenko, G. De Faveri, S. Masoudi, R. Al-Safadi, M. Watkinson,
Org. Biomol. Chem. 2013, 11, 1942–1951; b) K. Shastri, E. W. C. Cheng,
M. Motevalli, J. Schofield, J. S. Wilkinson, M. Watkinson, Green Chem.
2007, 9, 996–1007.
[16] a) K. Jobe, C. H. Brennan, M. Motevalli, S. M. Goldup, M. Watkinson,
Chem. Commun. 2011, 47, 6036–6038; b) E. Tamanini, A. Katewa, L. M.
Sedger, M. H. Todd, M. Watkinson, Inorg. Chem. 2009, 48, 319–324; c) E.
Tamanini, S. E. J. Rigby, M. Motevalli, M. H. Todd, M. Watkinson, Chem.
Eur. J. 2009, 15, 3720–3728; d) J. Pancholi, D. J. Hodson, K. Jobe, G. A.
Rutter, S. M. Goldup, M. Watkinson, Chem. Sci. 2014, 5, 3528–3535.
[17] A. G. Mayes, M. J. Whitcombe, Adv. Drug Delivery Rev. 2005, 57, 1742–
1778.
[41] F. E. Mabbs, D. Collison, Electron Paramagnetic Resonance of d Transition
Metal Compounds, Elsevier, Amsterdam, 1992, pp. 594–595, 710–711.
[42] D. A. Knight, J. B. Delehanty, E. R. Goldman, J. Bongard, F. Streich, L. W.
Edwards, E. L. Chang, Dalton Trans. 2004, 2006–2011.
[43] J. Chin, X. Zou, J. Am. Chem. Soc. 1988, 110, 223–225.
[44] A. Jorge, M. Chernobryva, M. Motevalli, M. Resmini, M. Watkinson, un-
published results.
[45] L. Broge, I. Sotofte, K. Jensen, N. Jensen, U. Pretzmann, J. Springborg,
Dalton Trans. 2007, 3826–3839.
[46] A. S. Kumbhar, M. S. Deshpande, R. J. Butcher, CrystEngComm 2008, 10,
1520–1523.
[47] J. H. Loehlin, E. B. Fleischer, Acta Crystallogr. Sect. A 1976, 32, 3063–
3066.
[48] M. Resmini, PhD Thesis, Universita degli Studi di Milano 1993.
[18] Y. Fujii, K. Matsutani, K. Kikuchi, J. Chem. Soc. Chem. Commun. 1985,
415–417.
[19] H. Kubo, T. N. Player, S. Shinoda, H. Tsukube, H. Nariai, T. Takeuchi, Anal.
Chim. Acta 2004, 504, 137–140.
Received: October 1, 2015
Published online on December 10, 2015
[20] K. Severin, Curr. Opin. Chem. Biol. 2000, 4, 710–714.
[21] N. M. Brunkan, M. R. GagnØ, J. Am. Chem. Soc. 2000, 122, 6217–6225.
Chem. Eur. J. 2016, 22, 3764 – 3774
www.chemeurj.org
3774 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim