Enzyme mediated silicon-oxygen bond formation; The use of Rhizopus oryzae lipase, lysozyme and phytase under mild conditions

Article abstract of DOI:10.1039/c0dt00151a

The potential for expanding the variety of enzymic methods for siloxane bond formation is explored. Three enzymes, Rhizopus oryzae lipase (ROL), lysozyme and phytase are reported to catalyse the condensation of the model compound, trimethylsilanol, formed in situ from trimethylethoxysilane, to produce hexamethyldisiloxane in aqueous media at 25 °C and pH 7. Thermal denaturation and reactant inhibition experiments were conducted to better understand the catalytic role of these enzyme candidates. It was found that enzyme activities were significantly reduced following thermal treatment, suggesting a potential key-role of the enzyme active sites in the catalysis. Similarly, residue-specific modification of the key-amino acids believed to participate in the ROL catalysis also had a significant effect on the silicon bio-catalysis, indicating that the catalytic triad of the lipase may be involved during the enzyme-mediated formation of the silicon-oxygen bond. E. coli phytase was found to be particularly effective at catalysing the condensation of trimethylsilanol in a predominantly organic medium consisting of 95% acetonitrile and 5% water. Whereas the use of enzymes in silicon chemistry is still very much a developing and frontier activity, the results presented herein give some grounds for optimism that the variety of enzyme mediated reactions will continue to increase and may one day become a routine element in the portfolio of the synthetic silicon chemist. The Royal Society of Chemistry.

Full text of DOI:10.1039/c0dt00151a

View Article Online / Journal Homepage / Table of Contents for this issue
This article is published as part of the Dalton Transactions themed issue entitled:
New Horizon of Organosilicon Chemistry
Guest Editor: Professor Mitsuo Kira
Tohoku University, Japan
Published in issue 39, 2010 of Dalton Transactions
Image reproduced with the permission of Akira Sekiguchi
Articles in the issue include:
PERSPECTIVES:
Silylium ions in catalysis
Hendrik F. T. Klare and Martin Oestreich
Dalton Trans., 2010, DOI: 10.1039/C003097J
Kinetic studies of reactions of organosilylenes: what have they taught us?
Rosa Becerra and Robin Walsh
Dalton Trans., 2010, DOI: 10.1039/C0DT00198H
π-Conjugated disilenes stabilized by fused-ring bulky “Rind” groups
Tsukasa Matsuo, Megumi Kobayashi and Kohei Tamao
Dalton Trans., 2010, DOI: 10.1039/C0DT00287A
Organosilicon compounds meet subatomic physics: Muon spin resonance
Robert West and Paul W. Percival
Dalton Trans., 2010, DOI: 10.1039/C0DT00188K
HOT ARTICLE:
Novel neutral hexacoordinate silicon(IV) complexes with two bidentate monoanionic benzamidinato
ligands
Konstantin Junold, Christian Burschka, Rüdiger Bertermann and Reinhold Tacke
Dalton Trans., 2010, DOI: 10.1039/C0DT00391C
Visit the Dalton Transactions website for more cutting-edge inorganic and organometallic research
www.rsc.org/dalton

PAPER
www.rsc.org/dalton | Dalton Transactions
Enzyme mediated silicon–oxygen bond formation; the use of Rhizopus oryzae
lipase, lysozyme and phytase under mild conditions†
a
a
b
a
a
Vincenzo Abbate, Alan R. Bassindale,* Kurt F. Brandstadt, Rachel Lawson and Peter G. Taylor*
Received 16th March 2010, Accepted 1st June 2010
DOI: 10.1039/c0dt00151a
The potential for expanding the variety of enzymic methods for siloxane bond formation is explored.
Three enzymes, Rhizopus oryzae lipase (ROL), lysozyme and phytase are reported to catalyse the
condensation of the model compound, trimethylsilanol, formed in situ from trimethylethoxysilane, to
◦
produce hexamethyldisiloxane in aqueous media at 25 C and pH 7. Thermal denaturation and
reactant inhibition experiments were conducted to better understand the catalytic role of these enzyme
candidates. It was found that enzyme activities were significantly reduced following thermal treatment,
suggesting a potential key-role of the enzyme active sites in the catalysis. Similarly, residue-specific
modification of the key-amino acids believed to participate in the ROL catalysis also had a significant
effect on the silicon bio-catalysis, indicating that the catalytic triad of the lipase may be involved during
the enzyme-mediated formation of the silicon–oxygen bond. E. coli phytase was found to be
particularly effective at catalysing the condensation of trimethylsilanol in a predominantly organic
medium consisting of 95% acetonitrile and 5% water. Whereas the use of enzymes in silicon chemistry is
still very much a developing and frontier activity, the results presented herein give some grounds for
optimism that the variety of enzyme mediated reactions will continue to increase and may one day
become a routine element in the portfolio of the synthetic silicon chemist.
Introduction
Given the mild conditions of biological Si–O bond formation it
is of major interest to attempt to transfer biological methodologies
into inorganic and organometallic chemistry to devise new,
greener, more sustainable and more economical ways to synthesise
industrially important organosilicon compounds such as silicones.
Our current interest in this emerging field is the discovery and
analysis of the mode of action of novel enzymes that can facilitate
siloxane bond formation under mild conditions. The eventual aim
is to develop alternative means of silicone and silsesquioxane
production, particularly of ‘designed’ materials—for example
stereoregular silicone polymers—where the usual equilibration
reactions prevent their synthesis. This is an emerging field that is,
to date, not well developed. By contrast there is already significant
interest in organosilicon compounds in biotechnological and
medicinal applications and the subject has also recently been
Silicon oxygen bonds, and particularly disiloxane bonds, Si–O–
Si, are ubiquitous in the biosphere. Silica in organisms such
as diatoms comprises a covalently bonded framework largely
composed of Si–O–Si bonds with some Si–OH bonds. Synthetic
materials such as silicones and related compounds have backbones
1
of long chains, rings or cages of silicon–oxygen bonds.
In processes collectively called biosilicification, structural silica
is produced in diatoms and other similar organisms from the
low concentration of silicic acid in the aqueous environment. The
mechanisms of these Si–O–Si bond forming reactions are not yet
completely understood but polypeptides and proteins have been
2,3
shown to participate actively in such biosilicification processes ,
In silica biosynthesis, proteins believed to be responsible for the
4
deposition of silica in two different families, diatoms and marine
6
reviewed. Since the 1970’s Tacke has led the field in modifying
3,4
sponges have been isolated and partially characterized. In par-
the action of selected drugs by the strategic replacement of carbon
3
ticular, Morse and coworkers have isolated proteins (silicateins)
from the T. Aurantia marine sponge that catalyze the in vitro
polycondensation of tetraethoxysilane (TEOS). Based on site-
directed mutagenesis the enzymatic active site of these silicatein
proteins was found to catalyze the hydrolysis and/or condensation
of alkoxysilanes during biosilicification reactions.
7,8
by silicon, for which he has coined the term ‘silicon switch’. Thus,
there is a rapidly growing interest in silicon compounds in a wide
variety of biological, medicinal and biotechnological contexts.
Several examples of enzyme-mediated reactions involving
organosilicon compounds have been reported; however, most of
these transformations do not directly involve the silicon atom. For
example lipases have been used to study the selective synthesis of
5
9
a
organosilicon esters under mild reaction conditions. In another
Department of Chemistry and Analytical Sciences, The Open University,
example, optically active silylmethanol derivatives were prepared
Walton Hall, Milton Keynes, UK MK7 6AA. E-mail: a.bassindale@
open.ac.uk; Tel: +44 1908 653363
10
by esterification with hydrolases and papain was found to be
particularly enantioselective. Hydrolases have also been used
successfully to catalyze the esterification of trimethylsilylpropanol
b
Dow Corning Corporation, 2200 W. Salzburg Rd., Midland, MI 48686,
USA
†
Electronic supplementary information (ESI) available: Sequence align-
11
isomers with 5-phenylpentanoic acid under mild conditions.
ment for phytases Aspergillus ficuum, Aspergillus niger and E. coli phytases.
See DOI: 10.1039/c0dt00151a
Since b-hydroxyalkylsilanes are unstable in the presence of acid
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9361–9368 | 9361

or base, the mild conditions of the enzymic reaction allowed the
ester to be formed without decomposition. Acrylate modification
of organosilicon compounds was achieved using hydrolases such
and different analytical methods might possibly explain such
apparently different biocatalytic activities.
We have recently carried out a more extensive trawl of the
available enzymes to screen and identify novel biocatalysts for the
12
as esterases, lipases and proteases. For example, NovozymeꢀR
23
4
35 (N435, a commercial source of Candida antarctica lipase B
cleavage of Si–OR bonds and the formation of Si–O–Si bonds.
immobilized on acrylic resin beads) was used to catalyze the ester-
ification of a silicone polyether. Advantages of the methodology
included elimination of unwanted side reactions and improved
safety of the work-up procedure as opposed to conventional
routes. Enzymes were used for the synthesis of carbohydrate-
This led to the discovery of at least three novel potential enzymatic
candidates for the biocatalysis of Si–O–Si bond formation. The
enzymes that were found to be particularly good at catalyzing the
reaction of interest are; a lipase from the fungus Rhizopus oryzae
(ROL); lysozyme and several phytases from different sources.
The aim of the work presented here was to further investigate
the catalytic activity of the selected enzymes in forming Si–O–Si
bonds by condensation of two Si–OH units, and to elucidate the
possible mechanism underlying this catalysis.
13
functional amphiphilic siloxanes. These “sweet silicones” possess
exceptional properties, depending on the attached carbohydrate.
For example, they can be used as surfactants and emulsifiers,
adhesion promoters, or chiral templates. Potato phosphorylase
◦
and a-D-glucose-1-phosphate were used in a citrate buffer at 37 C
to extend a synthetic maltoheptose-functionalized PDMS. More
recently, Gross and co-workers reported the synthesis of a
PDMS-ethyl glucoside copolymer using an immobilized lipase
Results and discussion
14
Reactions in aqueous solutions
(
N435). Hydrolase enzymes were used to enable the synthe-
15
In this study of the effects of enzymes on silicon–oxygen bond
formation, monofunctional silanes were chosen as the reactants
to limit complexity and enable the study of the formation of
molecules with only a single siloxane bond. Monoalkoxysilanes
are neither obviously suitable starting materials for the preparation
of silicone or silsesquioxane compounds, nor do they resemble
the natural silicaceous precursor used by microorganisms (i.e.
free silicic acid or silicates). However, this system was chosen,
as previously reported, as a simplified model to measure quan-
titatively the extent of siloxane bond formation under mild and
controlled conditions. As a starting point, trimethylethoxysilane
(TMES) is an ideal model compound as it has a single Si–O
bond, so there are no competing reactions and the system can
be followed without ambiguity. The possible products of the
reaction of monoalkoxysilanes with enzymes are trimethylsilanol
sis of organosilicon amides under mild reaction conditions.
Lipoprotein lipase A and cholesterol esterase type A from
Pseudomonas sp. were determined to have the highest activity for
catalyzing the formation of an amide bond between octanoic acid
and aminomethyltrimethylsilane in wet 2,2,4-trimethylpentane at
3
◦
0 C after two days of reaction.
There are cases where the silicon atom is directly implicated
in an enzymic reaction. The first reported case of direct siloxane
18
16
biocatalysis concerned the hydrolysis of silatranes. An esterase
obtained from the yeast Rhodotorula mucilaginosa was found
to catalyze the complete hydrolysis of the silatranes in Tris–
HCl buffer (pH 6.8) over 4 h at 30 C. Deactivated protein,
after thermal denaturation or treatment with EDTA, SDS and
◦
HgCl , resulted in loss of activity, suggesting a catalytic role
2
17
for the enzyme. In another example, Nishino reported an
enzymatic silicone oligomerization catalyzed by a lipid-coated
lipase. Rhizopus delemar lipase was found to be able to catalyze,
with proposed active site participation, the polycondensation of
(
(
TMSiOH) obtained by hydrolysis and the hexamethyldisiloxane
HMDS) obtained by condensation as shown in Scheme 1.
diethoxydimethylsilane in isooctane containing 2 wt% water at
◦
4
0 C over 20 h.
18
Our group has recently reported the first detailed, unam-
biguous example of biocatalysis at silicon, showing how the
active site of trypsin, a serine protease enzyme, was able to
catalyze siloxane bond formation under mild conditions during
the in vitro hydrolysis and condensation of model, monofunc-
tional alkoxysilanes. The reaction and conditions were carefully
chosen to allow differentiation between chemical and enzymatic
catalysis. The data suggested that homologous lipase and protease
enzymes also promoted the formation of siloxane bonds under the
same conditions. Similarly, Zelisko and co-workers reported that
certain proteases could mediate the sol–gel processing of TEOS,
Scheme 1 Enzyme-catalyzed siloxane bond formation.
The hydrolysis of alkoxy- or aryloxysilanes is readily accom-
plished in aqueous solution at pH 7 in the presence of a wide variety
of peptides and proteins and is relatively rapid and non-selective.
1
8
Surprisingly the condensation of silanols at pH 7 is not facile. To
date only one or two enzymes have been shown to facilitate the
condensation of silanols. The involvement of the enzyme active
19
20
phenyltrimethoxysilane and other organosilicon substrates.
We have also demonstrated that certain enzymes were able
to accelerate silica deposition from silicic acid with resultant
formation of silica-enzyme composites while retaining enzymatic
6,18,20
site is strongly indicated in the studies so far reported.
Silanol
condensation taking place in the presence of enzymes at accessible
rates at neutral pH, together with the even slower reverse reaction,
hydrolysis of disiloxanes, is the key reaction in making silicones
under non-equilibration conditions.
21
activity. By contrast, Ansorge, Schumacher and co-workers have
22
reported that the hydrolysis of a model monoalkoxysilane was
not accelerated by the presence of a variety of protease and lipase
enzymes, including trypsin; different reaction conditions from
those reported by us, such as monomer and enzyme concentration
In this study, a variety of proteins were reacted in a buffered,
◦
pH 7, aqueous solution at 25 C saturated with TMES, and with
9
362 | Dalton Trans., 2010, 39, 9361–9368
This journal is © The Royal Society of Chemistry 2010

Table 1 The effect of various proteins on the hydrolysis and condensation
◦
of trimethylethoxysilane in buffered water (pH 7.0) after 24 h at 25 C
a
Normalised yield (%)
Entry
TMES
TMSiOH
HMDS
negative control
bovine serum albumin
10.9
3.2
0.0
3.7
0.0
0.0
0.0
0.0
0.0
0.0
88.1
95.8
98.8
93.5
54.6
68.2
74.9
94.4
38.3
85.0
1.0
1.1
1.2
2.8
45.4
31.8
25.1
5.6
61.7
15.0
poly (L-lysine) hydrochloride
poly (allylamine) hydrochloride
Aspergillus ficuum phytase
Aspergillus niger phytase
chicken egg white lysozyme
E. coli phytase
Fig. 1 ROL- and lysozyme-mediated formation of HMDS after 24 h in
◦
a neutral media at 25 C. Substantial hydrolysis was observed with the
Rhizopus oryzae lipase—ROL
Trichoderma reesei phytase
negative control and non-specific protein and peptide over 24 h, therefore
only the extent of condensation is shown.
a
Actual measured total yields of TMES, TMSiOH aand HMDS varied
between 94–102%.
of lysozyme where the condensation yield was reduced from 24%
to around 5%. These observations strongly suggest that the tertiary
structures of the two enzymes play a significant role in the catalysis
of siloxane bond formation, and therefore, that the observed
reactions may be active-site mediated.
the excess TMES forming a separate layer; after 24 h the amounts
of TMES, TMSiOH and HMDS were measured quantitatively by
gas chromatography.
Although neutral pH may not be the optimum condition for the
use of certain proteins, it is essential for our model studies so that
acid- and base-catalyzed alkoxysilane hydrolysis and condensation
is minimized and enzyme impact can be differentiated from the
chemical effects of acidity or basicity.
Lysozyme is a small (14.3 kDa) but robust protein consisting
24
of a single 129 amino acid polypeptide chain, and containing a
region of extended b-sheet with a non-polar interior. Lysozyme
catalyzes the hydrolysis of the b-1,4 glycosidic bonds of the
polysaccharides that comprise the peptidoglycan layer of bacterial
cell walls. Hydrolysis of the glycosidic bond occurs with general
acid catalysis involving the carboxylate of glutamate 35, with
The results are given in Table 1. Following non-specific hy-
drolysis of TMES, as is normally observed in the presence of a
20
25
wide variety of proteins, ROL, lysozyme and several phytases
promoted the formation of HMDS. These proteins (enzymes)
were therefore able to catalyze the condensation of the resulting
TMSiOH into HMDS to a greater extent than the negative
control and non-specific protein (i.e. BSA) or peptide (e.g. poly-
L-lysine) reactions. Substantial hydrolysis was observed with
the negative control and non-specific protein and peptide over
aspartate 52 stabilizing the transition state.
Rhizopus oryzae lipase (ROL) is a small globular protein of
approximately 32 kDa that exists as a monomer in solution. ROL
catalyzes the hydrolysis of triglycerides to yield the free fatty acid.
It favours triglycerides with alkyl chains in the region of 8 to 18
carbons long, whereas poor activity has been seen with short-
26
chain fatty esters. Synthesized as a pre-proenzyme, it contains
an N-terminal signal sequence plus a prosequence that keeps it
2
4 h. We were not able to differentiate between enzymatic versus
27
chemically catalyzed hydrolysis over this time period. Notably,
no condensation product was observed in the absence of the
previously noted enzymes.
in an inactive state until it is exported from the cell. The crystal
28
˚
structure of ROL has been solved to 2.6 A, and the protein is
found be a typical a/b type protein with a central 8-stranded mixed
b-pleated sheet, and with three disulfide bonds as predicted from
its high homology (50% amino acid identity) with Rhizomucor
The catalytic activity of ROL and lysozyme was further inves-
tigated by conducting thermal denaturation experiments. Specifi-
cally, the enzymes were boiled for twenty minutes in Tris-buffered
water (following which treatment they were inactive towards
the control substrates) before adding the reactant to the cooled
solution in order to determine whether the catalysis was affected
by the above treatment. The results are summarized in Fig. 1.
The treatment of ROL by boiling the enzyme for twenty minutes
29
miehei lipase. The active site of lipases is similar to that in serine
proteases; a triad consisting of histidine, serine and aspartic acid
residues.
Reactions with ROL were further investigated, specifically to
elucidate the catalytic mechanism for the lipase-catalyzed HMDS
formation. First, reactant inhibition was studied by measuring
the activity of ROL with the pseudo-natural substrate para
nitrophenyllaurate (pNPL) shown in Scheme 2. In a spectropho-
tometric assay pNPL was reacted first with ROL in the absence
(
denaturation) drastically reduced its activity from about 62%
condensation of TMSiOH to less than 5% condensation. A similar,
but less dramatic, result was observed after thermal denaturation
Scheme 2 ROL-catalyzed hydrolysis of pNPL.
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9361–9368 | 9363

(
“negative control”) of TMES, and subsequently in the presence
aqueous ROL-catalyzed hydrolysis and condensation of TMES
was followed for 24 h at 25 C (Fig. 3). The concentrations of
reactant, intermediate and product were all measured over 24 h
study period.
◦
of the alkoxysilane (TMES), with which it had been incubated
for different time periods before addition of the pNPL. As the
TMES is known to be hydrolysed rapidly, any effect on the
rate of reaction can be assumed to be a result of TMSiOH
interaction with the ROL active site. A lower hydrolysis rate of
the pNPL in the presence of the silicon monomer would suggest
a competitive binding near or in the lipase active site. Analyzing
1
2
the half life (t ) of these reactions (based on the change in UV
absorbance at 405 nm, (Fig. 2), the enzyme activity towards
p-NPL decreased after various periods of incubation with the
alkoxysilane. Therefore, the silicon compound seemed to interact
within or near the active site region, thus causing a reduced
hydrolytic activity of the lipase towards the pNPL substrate.
Fig. 3 Time study of the ROL-catalyzed hydrolysis and condensation of
ethoxytrimethylsilane.
The hydrolysis of ethoxytrimethylsilane appeared to be com-
plete after 4–5 h, after which hexamethyldisiloxane began to form
and the yield increased smoothly. Since condensation did not occur
to any great extent until the completion of hydrolysis, this may
reflect the need for the silanol to achieve a sufficient concentration
to act as a nucleophile on the enzyme–substrate complex. This
hypothesis would be consistent with the formation of a silylated
serine intermediate (enzyme–OSiMe ) in or close to the lipase
3
active site as postulated as one explanation of the results of the
inhibition study (Fig. 2). The addition of TMES to the ROL
showed a strong inhibition of the enzyme activity towards pNPL
after 2–5 h of incubation with the alkoxysilane, corresponding to
the presence of fully hydrolyzed ethoxysilane (see Fig. 3).
Fig. 2 Half lives of ROL-catalyzed pNPL hydrolysis after various
incubation times with ethoxytrimethylsilane.
A reasonable inference is that the same active site is implicated
in the silanol condensation and the pNPL hydrolysis. Recently
Further insights into the lipase action in catalysing siloxane
bond formation may be obtained by selective chemical modifi-
cation of certain residues in the ROL. Both serine residues and
histidines may be modified chemically. In the case of ROL the
serine residue in the active site acts as the nucleophilic residue,
attacking the carbonyl group of its natural substrate to form
a tetrahedral intermediate. The nucleophilicity of this serine is
enhanced by an interaction with a histidine residue that holds it
in a constrained conformation. This histidine is in turn hydrogen
bonded to an aspartate residue, and it is this chain of interactions
30
Reetz has discovered that there can be different active sites in
the same enzyme with each site acting independently in different
catalytic reactions. This so-called alternate-site promiscuity does
not appear to be operating in this case as the TMSiOH has an
effect on the ROL active site for hydrolysis. However, given the
slow reduction in the rate of hydrolysis of pNPL, this may not be
competitive binding either, as that would be expected to operate
immediately. There are several other possible explanations for this
slow inhibition. It could be slow, irreversible covalent inhibition
through the formation of TMSiO-enzyme bonds at or very close
to the active site. There are some potential problems with this
interpretation as the condensation of TMSiOH, formed from
TMES in the presence of ROL, proceeds smoothly and there is no
evidence of “suicide inhibition” by TMSiOH. Further possibilities
33
(i.e. catalytic triad ) that confers on the enzyme its catalytic
properties. Covalent modification of histidines was accomplished
using the reagent diethylpyrocarbonate (DEPC), and the serines
were modified by treatment with diisopropylfluorophosphate
34
(DFP). The amount of reagent was calculated from the primary
sequence information for each protein. Pre- and pro-protein
sequences were ignored, based on the assumption that this protein
is produced from its natural host organism and therefore should
be correctly processed. As a result, ROL is expected to contain
seven histidines and thirty serines. A large excess of modifiers
(excess 10 and 100 fold molar for DFP and DEPC reactions,
respectively) was added to ensure all accessible relevant residues
were modified. In the modification of histidine, the nucleophilic
histidine attacks the carbonate group of the DEPC to give an
ethyl carboxylate modified histidine residue with the release of
carbon dioxide. In the neutral Tris buffer system employed, this
31
are a reversible allosteric effect where the conformation of
the enzyme is slowly altered by the TMSiOH binding which
would have to affect the pNPL hydrolysis but not the TMSiOH
condensation. It is also possible that other slow-binding inhibi-
32
tion mechanisms are operating. We carried out some further
investigation aimed at elucidating this mechanism, however it is
reasonable to hypothesize that the ROL active site has a potential
strong involvement in the biocatalyzed Si–O reaction.
In part, to confirm that slow enzyme inhibition is not a feature of
the Si–O–Si bond forming reactions in the presence of ROL, the
9
364 | Dalton Trans., 2010, 39, 9361–9368
This journal is © The Royal Society of Chemistry 2010

Table 2 Effect of DEPC and DFP modification of ROL on pNPL hydrolysis and HMDS formation
Reaction
pNPL hydrolysis
Result
Average units
Protein
% activity relative to control
ROL control for DEPC reaction
ROL DEPC treated
ROL control for DFP reaction
ROL DFP treated
645
26.0
1636
918
4
56
TMES hydrolysis
and condensation
% condensation relative to
ROL control
% yield TMSiOH
% yield HMDS
ROL control
38
64
96
88
99
62
36.
4
1
1
—
58
6
ROL DEPC treated
ROL DFP treated
Negative control
negative control + DEPC
3
5
covalent adduct is stable over many hours. The modification
of the histidine means that it is unable to act as a base or as a
nucleophile, and its participation in hydrogen bonding networks
should be reduced, thus weakening the charge-relay system of
the catalytic triad. Diisopropylfluorophosphate (DFP) is a very
potent inhibitor of serine esterases. It covalently inactivates all the
serine residues, thereby blocking the catalytic nucleophilicity of
their hydroxyl functionalities. The results of the modifications of
ROL on its activity towards pNPL and TMSiOH are summarised
in Table 2
Table 3 Enzyme-catalyzed
hydrolysis
and
condensation
of
trimethylethoxysilane in 5% (v/v) buffered water in tert-butanol
◦
after 24 h at 25 C
a
Normalized yields (%)
Entry
TMES
TMSiOH
HMDS
negative control
34.7
68.3
77.9
10.1
9.7
10.1
3.8
64.4
30.9
21.4
78.1
83.4
78.1
71.0
1.0
0.8
0.7
11.8
6.9
11.8
25.2
bovine serum albumin
poly (L-lysine) hydrochloride
Chicken egg white lysozyme
Aspergillus ficuum phytase
chicken egg white lysozyme
E. coli phytase
Although formation of HMDS was observed with the DEPC-
treated ROL, the enzyme activity was halved following the treat-
ment, suggesting at least a partial involvement of the histidines
in the catalysis. DFP-treatment almost completely abolished the
catalytic activity of the lipase. This result strongly highlights the
essential role of serine residues in the reaction mechanism for
disiloxane formation.
In summary, in aqueous solutions ROL was found to catalyze
in vitro siloxane bond formation under mild conditions of pH and
temperature. While there remains some degree of uncertainty the
experiments reported here suggest that the active site of ROL, the
catalytic triad, is intimately involved in the catalysis of siloxane
bond formation. It also seems to be more than coincidence that
the only proteins that we and others have found that catalyse
silanol condensation are enzymes with some kind of hydrolase
a
Actual measured total yields of TMES, TMSiOH aand HMDS varied
between 94–102%.
deactivated under these conditions and only a few retain their
activity. Nevertheless, many examples of biotransformations in
organic media have been reported.
37–40
Enzymes were chosen for this study that are known to be
active in more hydrophobic environments and these are shown
in Table 3. Following the familiar non-specific hydrolysis of the
model alkoxysilane, TMES, certain enzymes (Table 3) catalyzed
the condensation of the resulting TMSiOH to give HMDS to a
greater extent than the negative control and non-specific proteins
such as BSA or peptides. Substantial hydrolysis was observed with
the negative control and a non-specific protein and peptide over
6,18
function. However, more work is in progress to gain further
insights into these examples of enzyme promiscuity involving
silicon.
24 h. Again, we are not able to differentiate between enzymatic
and chemically catalyzed hydrolysis over this time period.
Notably, no significant condensation product was observed in
the absence of enzymes. As illustrated (Tables 2 and 3), several
phytases promoted HMDS formation. In this case, different
activities were observed, depending on the enzyme source and
the solvent used.
Specifically, A. niger, and A. ficuum phytases preferentially
catalyzed the reaction in water, whereas E. coli phytase gave a
high yield of the HMDS product only in wet organic media and
particularly so in wet acetonitrile. The results are summarized in
Fig. 4.
Reactions in organic media
In industrial contexts water is a desirable solvent but it is not
always suitable. The use of organic solvents particularly in silicone
fabrication is very convenient in process terms. Accordingly, the
hydrolysis and condensation study of TMES was also carried out
in organic media, particularly as many enzymes naturally function
in hydrophobic environments, such as those in the presence of, or
36
bound to, a membrane. The TMES studies were carried out by
37
using monophasic aqueous-organic solutions, consisting of 5%
v/v Tris-buffered water (pH = 7) in water-miscible organic co-
solvent such as tert-butanol or acetonitrile. Many proteins are
The phytases are examples of histidine acid phosphatases. These
show broad substrate specificity and hydrolyze phytate under
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9361–9368 | 9365

where available, for each phytase (see ESI† for data). Data for T.
reesei was not available.
The data for Aspergillus ficuum, Aspergillus nigera and E.coli
phytases is given in the ESI. There is 99% homology between
Aspergillus ficuum, and Aspergillus nigera which is consistent with
their similar activity in aqueous solution. Surprisingly there is
only a 4% homology between Aspergillus ficuum, Aspergillus nigera
phytases on the one hand and E.coli phytase on the other. This
again is fully consistent with the different reactivity patterns of
E.coli phytase relative to the Aspergillus phytases. In future work
to disclose the nature of the observed biocatalysis at silicon centre
by phytases we aim to carry out further analyses of the probable
active sites in the different phytases, possibly followed by genetic
modification of the proteins.
In summary, phytases, a class of acid phosphatases, were
also observed to catalyze HMDS formation under mild reaction
conditions. Depending on the phytase source and solvent system,
different activities were found. Notably, E. coli phytase gave a high
yield of disiloxane product in wet organic media. This reaction
may be of interest in silicone production, where the syntheses of
silicon-based materials are accomplished in water-poor systems.
Furthermore, silicon monomers are often insoluble in water.
Therefore, the discoveries of novel biocatalysts that catalyze the
siloxane bond formation in organic solvents present advantages.
◦
Fig. 4 Phytase-catalyzed HMDS formation after 24 h at 25 C.
acidic pH conditions to produce myo-inositol monophosphate as
41
the final product.
The phytase from E. coli is a 44.6 kDa protein that consists
of a highly conserved a/b-domain and a more varied a-domain
with the active site located between the two domains. Mutagenesis
studies alongside crystallographic analyses have been used to
identify key residues within the active site that are involved in
Conclusion
42
both the catalytic process and the tight binding of the substrate.
ROL, lysozyme and four different phytase enzymes were observed
to promote in vitro siloxane bond formation under mild reaction
conditions. Thermal denaturation of ROL and lysozyme strongly
suggested that the enzymatic tertiary structure may play a key-role
in the catalysis. Moreover, reactant inhibition studies coupled with
selective amino acid residue modification in ROL demonstrated
the potential role of the active site at catalyzing the HMDS
formation. The four phytases showed different activity depending
on enzyme source and solvent probably due to different molecular
structures and this is the subject of further research. Theses
enzymes may be very useful when designing novel organosilicon
biotransformations.
Whereas the use of enzymes in silicon chemistry is still very much
a developing and frontier activity, the results presented herein give
some grounds for optimism that the variety of enzyme mediated
reactions will continue to increase and may one day become a
routine element in the portfolio of the synthetic silicon chemist.
The active site histidine is known to be the residue responsible for
catalysis through its nucleophilic attack on the phosphorus of the
natural substrate. Morse has proposed a mechanism for catalysis
of siloxanes bond formation by silicatein in diatoms in which
nucleophilic attack on silicon by a histidine imidazole nitrogen
3–5
is a key step.
The difference in behaviour of the various phytase enzymes
initially proved difficult to understand. For example in aqueous
solution the E. coli phytase was significantly less reactive than the
other phytases (Table 1). Phytase enzymes have optimum activity
at acidic pH and this may account for the poor activity in neutral
water in the E. coli phytase reactions. although that does not
easily explain the difference between the enzymes. A. niger and T.
reesei phytases were supplied as liquid samples, and consequently,
for miscibility reasons, were only tested in aqueous solutions
without investigating their activities in organic solvents. Their
matrix solutions were found to be slightly acidic (pH = ~5). The
acidity of the medium might have influenced the catalysis relative
to the neutral buffered media (pH = 7.0) used in the reaction using
E. coli phytase (supplied as a solid sample) if there was also residual
acidity in that case. However, control reactions at the same pH of
the liquid phytase matrices showed no particular condensation
over 24 h (results not shown). A. ficum phytase (supplied as a solid
sample) showed better activity in water as opposed to organic
media. The differences between the aqueous activities of E. coli
and A. ficum phytases are not easily explained even though not all
of the phosphatase enzymes would be expected to show identical
activity profiles.
Experimental section
Materials
Aspergillus niger phytase was purchased from Interspex inc.
(San Mateo, CA). Trichoderma reesei phytase (Finase L) was
purchased from AB Enzymes (Darmstadt, Germany). E. coli
phytase was supplied by Genencor International, Inc. (Palo
Alto, CA). Acetonitrile, Aspergillus ficum phytase, bovine serum
albumin (BSA), chicken egg white lysozyme, dodecane, hexam-
ethyldisiloxane (HMDS), poly-(allylamine hydrochloride), poly-
L-lysine, Rhizopus oryzae lipase, sodium chloride, tert-butanol,
trimethylethoxysilane (TMES) and trizma pre-set crystals (pH 7.0)
were purchased from Sigma Aldrich (Poole, UK). Trimethylsilanol
We have therefore tried to understand the reasons for the
different activity towards the model alkoxysilane reaction in the
phytases by examining the homology of the amino acid sequences,
9
366 | Dalton Trans., 2010, 39, 9361–9368
This journal is © The Royal Society of Chemistry 2010

(
(
TMSiOH) was obtained from the Dow Corning Corporation
Midland, MI, USA). All the materials were used as received
DEPC was added (100-fold excess) and the suspension mixed by
rotation (30 min, 20 C, 120 rpm). Excess DEPC was removed
◦
without further modification. Ultra high purity water (UHP) was
obtained from a Milli-Q system at The Open University (Milton
Keynes, UK).
by centrifugation (13,000 rpm, 2 min) and the top aqueous layer
was filtered through a gel-filtration column (NAP-10, sephadex G-
50) to remove any remaining unreacted DEPC. The concentration
of protein in the samples was measured by UV absorbance. The
activity of the protein towards pNPL hydrolysis and TMSiOH
condensation was assessed via the relevant enzymatic activity assay
and the standard HMDS formation reaction, respectively.
Biocatalysis study
The reactions in the presence of enzymes were formulated with
a 5 : 1 alkoxysilane (100 mg) to enzyme (20 mg) weight ratio in
0
.5 g of the appropriate solvent system. The monophasic aqueous-
organic solution consisted of 5% (v/v) Tris buffered water (pH
7.0) in tert-butanol. Milli-Q water (‘water’) was buffered with
00 mM Tris–HCl buffer, pH 7.0 (‘buffered pH 7’). Based on
the estimated solubility of trimethylsilanol in water (42.56 mg
Diisopropylfluorophosphate (DFP) modification of serines
=
Protein was added to a solution of Tris buffer (980 mL, 50 mM,
-
1
1
pH 7.0) to give a solution of 10 mg mL . DFP was added (10-fold
◦
molar excess) and the suspension mixed by rotation (30 min, 20 C,
-
1
25
mL ), the concentration of trimethylethoxysilane (~200 mg
mL ) saturated the aqueous media and created a two-phase
reaction mixture. The time study of the Rhizopus oryzae lipase-
catalyzed hydrolysis and condensation of trimethylethoxysilane
was conducted at 25 C for defined periods of time over 24 h.
The reactions were conducted in silylated glass vials, which
provided an inert surface. Trimethylsilylated vials have been shown
120 rpm). Excess DFP and any precipitated protein were removed
by centrifugation (13,000 rpm, 2 min) and the top aqueous layer
was passed through a gel-filtration column (NAP-10, sephadex G-
50) to remove any remaining unreacted DFP. The concentration
of protein in the samples was measured by UV absorbance. The
activity of the protein was assessed via the activity assay and
HMDS formation.
-
1
◦
18a
43
not to contaminate the reactions with TMSiOH and HMDS.
Prior to analysis, the aqueous reactions were extracted with an
organic solvent, while the organic reactions were directly analyzed.
THF was selected as the extracting solvent, based on a previous
Gas chromatography-flame ionization detection
The gas chromatography (GC) analysis was performed with
an Agilent 6890 Series injector on an Agilent 6890 plus gas
chromatograph (GC) with a flame-ionization detector (FID). The
system was configured as follows: carrier gas, 99.9995% ultra high
43
extraction efficiency study on silicon compounds, as well as on its
documented use during the extraction a variety of silicones from
44
biological matrices. Sodium chloride was used to change the ionic
strength of the aqueous-phase in order to separate the miscible
THF-phase. The gas chromatography analysis was performed as
described below. The samples were prepared at ~1% (w/w) product
in a THF solution containing 1% (w/w) dodecane. The reactions
◦
purity (UHP) helium; GC inlet, split, 250 C, split ratio = 100 : 1,
-
1
constant flow (rate = 1.0 mL min .), Flame ionization detector at
◦
-1
-1
2
75 C, H
2
= 40 mL min , Make up N
2
= 45 mL min .; Column,
HP-5MS crosslinked 5% phenylmethylsiloxane film (30 m ¥
0.25 mm, 0.25 mm film); GC temperature program, 50(2) → 250
◦
were conducted in a neutral aqueous medium (pH 7.0) at 25 C
◦
-1
(
8) @ 10 C min , 30 min total run time; internal standard,
for 24 h.
~
1%(w/w) dodecane in THF; Data system, Agilent Technologies
The closed (screw-capped) two-phase reactions were conducted
◦
ChemStation. Dodecane was used as an internal standard to
quantitate gravimetrically the chromatographic analyses. The
samples were prepared at ~1% (w/w) product in a THF solution
containing 1% (w/w) dodecane as internal standard (IS). Based
in silanizeds glass vials at 25 C with magnetic stirring for 24 h.
The reaction products were isolated and quantitatively analyzed by
GC-FID. Prior to analysis, the aqueous reactions were extracted
with 1 ml of THF in the presence of NaCl and filtered through a
Whatman AutovialꢀR 5 or 12 0.45-mm TeflonꢀR filter, while the
organic reactions were directly filtered and analyzed. Control
reactions are defined as non-enzymatic reactions. Specifically,
experiments conducted in the absence of a protein are defined as
negative control reactions. Proteinaceous and peptide molecules
such as bovine serum albumin and poly-L-lysine, respectively, were
used to study non-specific protein/peptide catalysis.
45
on triplicate measurements, the response factors for the analytes
were calculated (eqn (1)), and found to be linear as a function of
concentration over four orders of magnitude (i.e. 0.01–10% w/w).
RFanalyte = ([analyte]/Areaanalyte) ¥ (AreaIS/[IS]) ¥ RFIS
(1)
where, RFanalyte = response factor for the analyte, [analyte] =
concentration of the analyte, Areaanalyte = peak area of the
analyte, AreaIS = peak area of the internal standard, [IS] =
concentration of the internal standard, RFIS = response factor
for the internal standard = 1. Subsequently, eqn (1) was used
to calculate quantitatively the concentration of an analyte in the
presence of an internal standard.
Thermal denaturation
In the thermal denaturation experiments, the enzymes (10 mg
-
1
mL ) were boiled for 20 min in 100 mM Tris–HCl buffered Milli-
◦
Q water (pH 7.0) and cooled to 25 C before the addition of the
alkoxysilane. Thereafter, the reaction mixtures were treated and
analyzed in an identical manner as described above.
Spectrophotometric analysis: enzyme activity assay
4
-Nitrophenyl laurate (pNPL) was used to study the li-
pase/esterase activity of Rhizopus oryzae lipase. As detailed in
Table 4, the assays were formulated by transferring the reagents,
with the exception of the pNPL solution (in order to avoid
non-enzymatic hydrolysis), into a glass cuvette. After mixing the
Diethylpyrocarbonate (DEPC) modification of histidines
Protein was added to a solution of Tris buffer (960 mL, 50 mM,
pH 7.0) plus ethanol (20 mL) to give a solution of 10 mg ml .
-
1
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9361–9368 | 9367

Table 4 Rhizopus oryzae lipase spectrophotometric enzymatic activity
assay
17 H. Nishino, T. Mori and Y. Okahata, Chem. Commun., 2002, 2684–
2685.
1
8 (a) A. R. Bassindale, K. F. Brandstadt, T. H. Lane and P. G. Taylor,
J. Inorg. Biochem., 2003, 96, 401–405; (b) A. R. Bassindale, K. F.
Brandstadt, T. H. Lane and P. G. Taylor, Polymer Preprints, 2003,
Reagent
A
Volume
4
4, 570; (c) A. R. Bassindale, K. F. Brandstadt, T. H. Lane and P. G.
0.4 : 10 (v/v) isopropanol : 50 mM
2.955 mL
Taylor, ACS Symp. Ser., 2005, 900, 164–181.
Tris–HCl buffer (pH 8.0), 0.5M CaCl
2
-1
19 M. Frampton, A. Vawda, J. Fletcher and P. M. Zelisko, Chem.
Commun., 2008, 5544–5546.
B
C
ROL solution (0.2 mg mL ) in reagent A
10 mM pNPL in acetonitrile in reagent A
15 mL
30 mL
2
2
0 M. Frampton, R. Simionescu and P. M. Zelisko, Silicon, 2009, 1, 47–56.
1 A. Bassindale, P. G. Taylor, V. Abbate and K. F. Brandstadt, J. Mater.
Chem., 2009, 19, 7606–7609.
2 A. Maraite, M. B. Ansorge, Schumacher, B. Ganchegui, W. Leitner and
G. Grogan, J. Mol. Catal. B: Enzym., 2009, 56, 24–28.
3 V. Abbate, A. R. Bassindale, K. F. Brandstadt and P. G. Taylor,
submitted to J. Inorg. Biochem, Jan 2010.
contents, background spectra were acquired at 405 nm on an
Uvikon L ultraviolet–visible spectrophotometer. Subsequently, the
pNPL solutions were added to the cuvettes, mixed, and the release
of para-nitrophenol followed at 405 nm over 10 min, at 25 C,
detecting 60 points/min.
2
2
◦
24 R. E. Canfield, J. Biol.Chem., 1963, 238, 2698.
2
5 (a) I. Kumagai, K. Maenaka, F. Sunada and S. Takeda, Eur. J. Biochem.,
1
993, 212, 151–156; (b) K. Maenaka, G. Kawai, K. Watanabe and F.
Acknowledgements
Sunada, J. Biol. Chem.y, 1994, 269, 7070–7075.
2
2
6 A. Hiol, M. D. Jonzo, N. Rugani, D. Druet, L. Sarda and L. C. Comeau,
Enzyme Microb. Technol., 2000, 26, 421–430.
We gratefully acknowledge generous financial and technical
support from Dow Corning Corporation, Midland (MI), USA.
7 (a) M. J. Haas, J. Allen and T. R. Berka, Gene, 1991, 109, 107–113;
(
b) H. D. Beer, J. E. McCarthy, U. T. Bornscheuer and R. D. Schmid,
Biochim. Biophys. Acta, Gene Struct. Expression, 1998, 1399, 173–180.
8 U. Derewenda, L. Swenson, Y. Wei, R. Green, P. H. Kobos, R. J o¨ rger,
M. J. Haas and Z. S. Derewenda, J. Lipid Res., 1994, 35, 524–534.
29 L. Brady, A. M. Brzozowski, Z. S. Derewenda, E. Dodson, G. Dodson,
S. Tolley, J. P. Turkenburg, L. Christiansen, B. Huge-Jensen and L.
Norskov, Nature, 1990, 343, 767–770.
30 A. Taglieber, H. H o¨ benreich, J. D. Carballeira, R. G. J. G. Mondiere
and M. T. Reetz, Angew. Chem., Int. Ed., 2007, 46, 8597–8600.
31 D. E. Koshland, G. N e´ methy and D. Filmer, Biochemistry, 1966, 5,
365–368.
32 (a) M. J. Sculley, J. F. Morrison and W. W. Cleland, Biochim. Biophys.
Acta, Protein Struct. Mol. Enzymol., 1996, 1298, 78–86; (b) S. E.
Szedlacsek and R. G. Duggleby, Methods Enzymol., 1995, 249, 144–
180.
2
References
1
2
3
M. A. Brook, Silicon in Organic, Organometallic and Polymer Chem-
istry, John Wiley and Sons Inc, New York, 2000.
N. Kroger, R. Deutzmann and M. Sumper, Science, 1999, 286, 1129–
1
132.
J. N. Cha, K. Shimizu, Y. Zhou, S. C. Christiansen, B. F. Chmelka,
G. D. Stucky and D. E. Morse, Proc. Natl. Acad. Sci. U. S. A., 1999,
9
6, 361–365.
4
(a) N. Kroger, R. Deutzmann, C. Bergsdorf and M. Sumper, Proc.
Natl. Acad. Sci. U. S. A., 2000, 97, 14133–14138; (b) N. Kroger, R.
Deutzmann and M. Sumper, J. Biol. Chem., 2001, 276, 26066–26070;
(
c) K. Shimizu, J. Cha, G. D. Stucky and D. E. Morse, Proc. Natl. Acad.
Sci. U. S. A., 1998, 95, 6234–6238.
33 (a) P. Carter and J. A. Wells, Nature, 1988, 332, 564–568; (b) G. Dodson
and A. Wlodawer, Trends Biochem. Sci., 1998, 23, 347–352.
34 A. Fersht, Enzyme Structure and Mechanism, 1985, 2nd edition,
W.H. Freeman, & Company, New York.
5
Y. Zhou, K. Shimizu, J. N. Cha, G. D. Stucky and D. E. Morse, Angew.
Chem., Int. Ed., 1999, 38, 779–782.
6
7
M. B. Frampton and P. M. Zelisko, Silicon, 2009, 1, 147–163.
(a) W. Bains and R. Tacke, Curr. Op. Drug Disc. Dev., 2003, 6, 526–
35 J. L. Dage, H. Sun and H. B. Halsall, Anal. Biochem., 1998, 257, 176–
5
43; (b) R. Tacke, S. A. Wagner, in The chemistry of organic silicon
185.
compounds, Vol. 2 ed.: Z. Rappoport, and Y. Apeloig, John Wiley &
Sons, West Sussex, England, 1998.
36 B. Borgstr o¨ m and H. Brockman, Lipases, 1984, Elsevier, Amsterdam.
37 K. Faber, Biotransformation in Organic Chemistry, 2000, Springer-
Verlag, New York.
8
9
R. Tacke, V. M u¨ ller, M. W. B u¨ ttner, W. P. Lippert, R. Bertermann,
J. O. Daiss, H. Khanwalkar, A. Furst, C. Gaudon and H. Gronemeyer,
ChemMedChem, 2009, 4, 1797–1802.
38 J. R. Hanson, An introduction to biotransformation in organic chemistry,
1998, W. H. Freeman, Oxford.
T. Kawamoto, K. Sonomoto and A. Tanaka, J. Biotechnol., 1991, 18,
39 Y. Yesiloglu and A. Sagiroglu, Turk. J. Chem., 2002, 26, 529–5534.
40 (a) H. N. Cheng and R. A. Gross, ACS Symp. Ser., 2005, 900, 1–12;
(b) Y. Kimura, M. Shima, T. Notsu and S. Adachi, in Developments in
Food Engineering ed.: T. Yano, R. Matsuno and K. Nakamura, Blackie
Academic & Professional, Chiba Japan, 1993.
8
5–92.
1
1
0 T. Fukui, T. Kawamoto and A. Tanaka, Tetrahedron: Asymmetry, 1994,
5
, 73–82.
1 A. Uejima, T. Fukui, E. Fukusaki, T. Omata, T. Kawamoto, K.
Sonomoto and A. Tanaka, Appl. Microbiol. Biotechnol., 1993, 482–
41 S. Golovan, G. Wang, J. Zhang and C. W. Forsberg, Can. J. Microbiol.,
4
86.
2000, 46, 59–71.
1
1
1
1
1
2 B. Gruning, G. Hills, W. Josten, D. Schaefer, S. Silber, C. Weitemeyer,
and Th. Goldschmidt, 2001, U.S. Patent, 6288129, United States.
3 V. V. B r a u n m u¨ hl, G. Jonas and R. Stadler, Macromolecules, 1995, 28,
42 (a) Z. Jia, S. Golovan, Q. Ye and C. W. Forsberg, Acta Crystallogr.,
Sect. D: Biol. Crystallogr., 1998, 54, 647–649; (b) D. Kostrewa, F.
Grueninger-Leitch, A. D’Arcy, C. Broger, D. Mitchell and A. P. G. M.
Van, Loon, Nat. Struct. Biol., 1997, 4, 185.
43 K. F. Brandstadt, PhD Thesis, The Open University (Milton Keynes),
2003.
1
7–24.
4 B. Sahoo, K. F. Brandstadt, T. H. Lane and R. A. Gross, Org. Lett.,
005, 7, 3857–3860.
5 T. Kawamoto, R. S. So, Y. Masuda and A. Tanaka, J. Biosci. Bioeng.,
999, 87, 607–610.
6 A. N. Fattakhova, E. P. Chirko and E. N. Ofitserov, Biol. Nauki, 1992,
, 100–105.
2
44 S. Varaprath, K. L. Salyers, K. P. Plotzke and S. Nanavati, Anal.
1
Biochem., 1998, 256, 14–22.
45 The Analytical Chemistry of Silicones, A. L. Smith, (Ed.), 1991, Wiley,
New York.
4
9
368 | Dalton Trans., 2010, 39, 9361–9368
This journal is © The Royal Society of Chemistry 2010