The enzymatic cleavage of Si-O bonds: A kinetic analysis of the biocatalyzed hydrolysis of phenyltrimethoxysilane

Article abstract of DOI:10.1016/j.molcatb.2010.04.002

Previously we reported the ability of trypsin to mediate the cross-linking of alkoxysily-functionalized silicone polymers. Although enzymes and silicon-containing compounds are not necessarily incompatible species, the exact mechanism of how enzymes process silicon substrates is not fully understood. The focus of this current work was to examine the reaction kinetics associated with the processing of an alkoxysilane substrate by enzymes using ²⁹Si NMR so as to gain a greater insight into the actual reaction mechanism, especially those involving more complex silicone systems. A series of time course ²⁹Si NMR experiments using D₂O revealed that the trypsin-mediated hydrolysis of a single alkoxy moiety in water is a pseudo-first order reaction. The relative effect of the enzyme was determined to be β = 3.549 while the relative effect of water was γ = 3.325. Prolonged contact with phenyltrimethoxysilane was not sufficiently deleterious to the enzyme and did not induce the complete and irreversible denaturation of trypsin. Computational evidence suggests that while in the active site of the enzyme, serine addition to silicon to forms a pentacoordinate species and is favoured over histidine addition.

Full text of DOI:10.1016/j.molcatb.2010.04.002

Journal of Molecular Catalysis B: Enzymatic 66 (2010) 105–112
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
The enzymatic cleavage of Si–O bonds: A kinetic analysis of the biocatalyzed
hydrolysis of phenyltrimethoxysilane
∗
Mark B. Frampton, Razvan Simionescu, Travis Dudding, Paul M. Zelisko
Department of Chemistry and Centre for Biotechnology, Brock University, 500 Glenridge Avenue, St. Catharines, Ontario, Canada L2S 3A1
a r t i c l e i n f o
a b s t r a c t
Article history:
Previously we reported the ability of trypsin to mediate the cross-linking of alkoxysily-functionalized
silicone polymers. Although enzymes and silicon-containing compounds are not necessarily incompatible
species, the exact mechanism of how enzymes process silicon substrates is not fully understood. The
focus of this current work was to examine the reaction kinetics associated with the processing of an
Received 11 January 2010
Received in revised form 26 March 2010
Accepted 1 April 2010
Available online 9 April 2010
2
9
alkoxysilane substrate by enzymes using Si NMR so as to gain a greater insight into the actual reaction
mechanism, especially those involving more complex silicone systems. A series of time course 29Si NMR
experiments using D2O revealed that the trypsin-mediated hydrolysis of a single alkoxy moiety in water
is a pseudo-first order reaction. The relative effect of the enzyme was determined to be ˇ = 3.549 while the
relative effect of water was ꢀ = 3.325. Prolonged contact with phenyltrimethoxysilane was not sufficiently
deleterious to the enzyme and did not induce the complete and irreversible denaturation of trypsin.
Computational evidence suggests that while in the active site of the enzyme, serine addition to silicon to
forms a pentacoordinate species and is favoured over histidine addition.
Keywords:
Silicon
Biotechnology
Enzyme mediated
Reaction kinetics
Alkoxysilane
©
2010 Elsevier B.V. All rights reserved.
1
. Introduction
vitro condensation of a single siloxane bond [8]. Hydrolysis was
thought to proceed mainly through non-specific surface interac-
tions between the enzyme and the alkoxysilane, and the active site
was a requirement for performing siloxane condensation [8]. This
report appeared to be challenged by Maraite et al. in a 2009 publi-
cation claiming that trypsin did not increase the rate of hydrolysis
of trimethylethoxysilane under similar reaction conditions [12].
Subsequent to the Maraite report it has been demonstrated that
several proteases are capable of mediating the sol–gel processing
of tetra- and tri-alkoxysilanes under solvent-free conditions to pro-
duce hydro-gels, as well as monolithic silica and organosilica [2,3].
Inhibition of trypsin by a Kunitz-type inhibitor resulted in a com-
plete lack of hydrolysis of phenyltrimethoxysilane suggesting that
the enzyme required a catalytically competent active site to carry
out both the hydrolysis and condensation of phenyltrimethoxysi-
lane.
Sol–gel processing of organically modified alkoxysilanes is a
relatively mild method for preparing hybrid organic-inorganic
materials [1]. Several methods have been investigated for this pro-
cess that allow for control over the characteristics (e.g., pore size,
total surface area) of the resulting silsesquioxanes. More recently
biomimetic and enzymatic routes to sol–gel silica and silsesquiox-
anes have been pursued [2–11].
The isolation of silicatein, a silicon-processing protein, from
a variety of marine sponges served as the impetus for sev-
eral research programs aimed at exploring the biotechnological
potential of enzymatic catalysis of silicon frameworks. Morse and
coworkers demonstrated that silicatein was capable of perform-
ing the in vitro hydrolysis and condensation of tetraethoxysilane
and phenyltriethoxysilane [4]. Structural analyses demonstrated
that silicatein has an active site that is structurally similar to
the serine proteases containing a serine–histidine–aspartic acid
Accurate descriptions of the sol–gel processing of alkoxysilanes
have been attempted, unfortunately due to the complicated nature
of the process these models have needed to employ simplifications
to describe the systems under study [13,14]. As an example, it has
been postulated that the hydrolysis and condensation of tetram-
ethoxysilane requires at least 160 rate equations to adequately
(
Ser–His–Asp) catalytic triad. The similarity between these two
classes of proteins prompted not only Morse and coworkers to
examine the possibility of trypsin as a catalyst for in vitro silica
formation [4], but Bassindale et al. also sought to explore the in
29
describe the entire reaction process [14]. Si Nuclear magnetic
resonance (29Si NMR) spectroscopy has proven to be a valuable
technique in describing the reaction kinetics of sol–gel process-
ing and has been used extensively in the last decade [15,16]. 29Si
NMR is an ideal method not only for its quantitative nature, but
also because it allows for the identification of each unique silicon
∗ Corresponding author. Tel.: +1 905 688 5550x4389; fax: +1 905 682 9020.
E-mail addresses: mark.frampton@brocku.ca (M.B. Frampton),
rsimionescu@brocku.ca (R. Simionescu), tdudding@brocku.ca (T. Dudding),
pzelisko@brocku.ca (P.M. Zelisko).
1
381-1177/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2010.04.002

1
06
M.B. Frampton et al. / Journal of Molecular Catalysis B: Enzymatic 66 (2010) 105–112
2.2. Methods
2.2.1. General procedure for 29Si NMR experiments
Scheme 1. The generalized reaction scheme for a single hydrolysis event of
phenyltrimethoxysilane. The hydrolysis rate constant is given by kh.
Trypsin or pepsin was dissolved in 500 ␮L of a 3:1 mix-
ture of D O:H O and subsequently diluted to 7.5–15 mg/mL.
2
2
The enzyme preparations were subsequently combined with
phenyltrimethoxysilane (300–500 ␮L) in fresh Eppendorf tubes
and mixed using a vortex. A 550 ␮L aliquot of the reaction mix-
ture was withdrawn and transferred into a 5 mm glass NMR tube
and spectra were acquired. Enzyme-free control experiments were
carried out at each water concentration as described (vida infra).
During spectral acquisition the tube was spinning at 20 Hz.
resonance [13]. Typically, as alkoxy groups are replaced by silanol
2
9
groups the Si resonances shift down field (becoming more posi-
tive) due to the increasing positive charge that develops on silicon.
This trend appears to be the case for tetra- and tri-alkoxysilanes, but
only under certain circumstances for di-substituted alkoxysilanes.
The hydrolysis of diethoxydimethylsilane under basic conditions
mimics this trend, but under acidic conditions the observed change
in chemical shift is in the opposite direction [17]. Disiloxane bond
formation results in an up-field shift to more negative ı values. This
2
.2.2. Nuclear magnetic resonance (NMR)
29
Si NMR spectra were acquired using a Bruker Avance AV-300
trend is reversed when small cyclic oligomers form, which due to
(59.6 MHz for 29
Si) and AV-600 (119.2 MHz for Si) spectrome-
29
the increasedringstrain, see the 29Siresonances beingshifted down
ters. Spectra were recorded over 10–60 h at 1 h intervals; each
spectrum was acquired with 60 averages by employing a proton
field [18]. The exact amount of change in the chemical shift of the
relative 29Si nucleus is dependent on the degree of substitution at
◦
inverse-gated decoupled sequence with a 30 flip angle and a 60 s
the silicon atom.
relaxation delay to ensure complete relaxation of all 29Si nuclei. Si
NMR spectra were analyzed using the Bruker Topspin v2.0 software
platform. Integration of 29Si resonances was performed over fixed
limits for all experiments.
29
The hydrolysis and condensation of alkoxysilanes generally pro-
ceeds as outlined in Scheme 1. The hydrolysis of an alkoxysilane
can proceed through either nucleophilic or electrophilic catalysis
depending on the catalytic conditions (i.e., acidic or basic) [19]. The
reverse of this reaction, namely silyl ether formation, is more dif-
ficult to quantify but can be suppressed if an excess of water is
included in the reaction; Le Châtelier’s principle will favour the
hydrolysis products under these conditions. While the analysis of a
single hydrolysis event is a simplification of the complete hydrol-
ysis of alkoxysilanes, it is sufficient for the discussion contained
herein.
2.2.3. Determination of trypsin activity
Phenyltrimethoxysilane monoliths were prepared in a similar
manner as the 29Si NMR experiments. After 24 h of incubation at
room temperature the reaction mixture contained a white, gel-
like material. Monoliths were subsequently formed by ageing the
◦
mixtures at 50 C for an additional 24 h. Unfortunately, the long
Using our previously described 29Si NMR-based methodology,
the present work describes the hydrolysis of phenyltrimethoxysi-
lane by two dissimilar enzymes; trypsin (a serine protease) and
pepsin (a carboxypeptidase). These enzymes possess different
modes of catalysis, natural substrate preference (positively charged
basic amino acid side chains for trypsin versus aromatic amino
acid side chains for pepsin) as well as optimal functioning envi-
ronments (pH 8 for trypsin compared to pH 2 for pepsin). In
this series of experiments we examined the fate of trypsin at the
end of the sol–gel process using FTIR spectroscopy and colouri-
metric assays. While the entrapped enzyme appeared to remain
structurally intact, the catalytic competency of the enzyme cannot
be confirmed. Attempts at leaching the enzyme from within the
organosilica monolith were unsuccessful suggesting that the pore
size was restrictive to the mobility of either trypsin or the substrate
reagent.
time scale required to form gels, sometimes greater than 24 h for
lower enzyme concentrations, did not permit an accurate deter-
mination of the time required to achieve each organosilica gel.
Monoliths containing trypsin were ground to a fine powder and
incubated in a 0.2 M phosphate buffer at pH 7 for 24 h. Trypsin was
assayed using the BAEE reagent as described in the literature [20].
A fresh enzyme preparation was prepared to 0.5 mg/mL to pro-
vide a baseline for comparison with the entrapped enzyme. The
enzymatic assay for residual trypsin activity consisted of 2.9 mL
of 0.263 mM BAEE in 0.0067 M phosphate buffer and 0.1 mL of
0.5 mg/mL trypsin prepared in 0.02 M phosphate buffer at pH 7.
The quartz cuvette containing the BAEE reagent was equilibrated
in the photospectrometer for 5 min prior to the addition of the
trypsin preparation. The change in the absorbance at 253 nm was
monitored for 10 min. Colourimetric enzyme kinetics assays were
performed on an Ultrospec pro 2100 photospectrometer equipped
with the Swift II software interface.
2
. Experimental
2.2.4. Fourier transform infrared spectroscopy (FTIR)
FTIR spectroscopy was performed using a Mattson Research
2
.1. Materials
Series infrared spectrometer in transmittance mode. Spectra were
TM
analyzed using the Winfirst
software platform. Solid samples
Trypsin from bovine pancreas (90–100% purity, EC.3.4.21.4),
of silsesquioxane monoliths were ground into fine powders and
pepsin 1:10,000 from porcine stomach mucosa (EC.3.4.23.1),
and benzoyl-l-arginine ethyl ester (BAEE) were acquired from
Sigma–Aldrich (Oakville, Ontario, Canada). Chloroform-d (CDCl3,
◦
aged in a 60 C oven for 24 h, then stored at room temperature in a
TM
dessicator containing Drierite prior to the acquisition of spectra.
Samples were prepared as KBr pellets containing approximately 2%
9
9.8% deuterated) and deuterium oxide (D O, 99.9% deuter-
2
(
w/w) of each enzyme containing monolith. FTIR spectra were the
ated) were acquired from Cambridge Isotope Laboratories, Inc.
Landover, Maryland, USA). Phenyltrimethoxysilane (PTMS, >95%
−
1
average of 64 scans acquired at 2 cm resolution.
(
purity) was obtained from Gelest (Morristown, Pennsylvania, USA).
Sodium phosphate dibasic (minimum 99% purity) was obtained
from Caledon Laboratories Ltd. (Georgetown, Ontario, Canada).
Sodium phosphate monobasic (>98% purity) was obtained from
BDH (Poole, England). Reagents were used as received without fur-
2.2.5. Computational modeling
Computational studies using density functional theory (DFT)
employing the gradient-corrected (B3LYP) hybrid functional of
Becke–Lee–Yang and Parr with a double zeta-potential 6-31G(d)
basis set as implemented by Gaussian 03 have been utilized to
examine the chemistry within the enzymes’ active site.
ther purification or modification. Distilled (dH O) water was used
2
for the preparation of the BAEE reagent and phosphate buffers.

M.B. Frampton et al. / Journal of Molecular Catalysis B: Enzymatic 66 (2010) 105–112
107
Fig. 1. Representative 29Si NMR spectra of the trypsin (left) and pepsin (right) mediated hydrolysis and condensation of PTMS. The spectra shown represent 2 h intervals
beginning at t = 1 h (bottom) and proceed to t = 24 h (top).
1
2
3
. Results and discussion
represent the single and double hydrolysis products (T and T ) of
0 0
the PTMS. Those 29Si resonances for the initial hydrolysis reactions
were not detectable using the acquisition parameters that were
employed. As well the paramagnetic relaxation agent chromium
acetylacetonate could not be used due to its inhibitory effect on
trypsin [3]. Studies employing traditional acid/base catalysis and
paramagnetic relaxation agents have been able to observe the
first and second hydrolysis products, T 01 and T 02 for PTMS. These
have been shown to occur near −54 and −52 ppm, respectively
Biocatalytic- and biomimetic-based sol–gel methods are receiv-
ing an increasing amount of attention in the literature. While only
a single family of enzymes has been shown to naturally process
silica and organosilica precursors [4,5,11], several more enzymes
[
2,3,5,7,8], and other biologically relevant molecules [9,10] have
been studied for their silicon-processing abilities.
29
Si Nuclear magnetic resonance spectroscopy is a common
method for analyzing the polymerization of alkoxysilanes. This is
due in part to the quantitative nature of the technique as well as the
volume of information that can be acquired (e.g., effect of catalyst,
solvent, temperature, and co-monomer ratios) [15]. Several sys-
tems of nomenclature have been employed when describing the
intermediates that are observed during the hydrolysis and con-
densation of alkoxysilanes. The nomenclature from Brinker and
[
21].
After 13 h of spectral acquisition a third resonance began to
2
1
2
1
appear at −61.1 ppm (T T ) which indicated the beginning of dis-
iloxane formation. When the concentration of PTMS was increased,
a fourth resonance appeared after ∼15 h located at −70.5 ppm
2
1
2
(
T T T ) that represented the beginning of small oligomer for-
1 1
2
mation. There are several possibilities as to the identity of this
n
Scherer (T ) will be employed to identify the relevant peaks in the
2 1 2
m
resonance, however, we have assigned it as the T T T oligomer
1 2 1
2
9
Si NMR spectra [1]. Specifically, to facilitate the discussion con-
(
1,3,5-triphenyl-1,1,3,5,5-pentahydroxytrisiloxane) because there
was only spectral evidence for the existence of 1,3-diphenyl-
,1,3,3-tetrahydroxydisiloxane and it is not likely to represent any
tained herein the following will be used: T is used to represent
a silicon atom that is covalently bonded to three oxygen atoms
and one organic moiety; n will be used to represent the num-
ber of silanols attached to a particular silicon resonance; and m
will represent the number of siloxane linkages to the same silicon
environment.
1
partially methoxylated trisiloxanes. However, with the observ-
able line broadening these other candidate siloxane species
may be obscured. The presence of the 1,3,5-triphenyl-1,1,3,5,5-
pentahydroxytrisiloxane oligomer was further confirmed by the
growth of a smaller peak at −60.9 ppm representing the end groups
2
1
1
2
2
1
within the T T T oligomer (Table 1).
3
.1. Trypsin-mediated processing of phenyltrimethoxysilane
The serine protease trypsin facilitated the sol–gel polymer-
ization of PTMS, methyltrimethoxysilane, ethyltrimethoxysilane,
and allyltrimethoxysilane while the carboxypeptidase pepsin pos-
sessed a similar ability to process PTMS under near neutral
conditions [3]. Tetraethoxysilane hydrolysis and condensation was
also attainable by enzymatic means, but through different pro-
cesses [2]. The hydrolysis of phenyltrimethoxysilane under tryptic
and peptic modes of catalysis was typically initiated soon after the
enzyme was charged with the organosilicon substrate. To minimize
experimental error induced by the time lag between reagent mix-
ing and spectral acquisition, less than 5 min were allowed to elapse
before spectral acquisition commenced. After the first hour of spec-
Table 1
Assignment of 29Si resonances for phenyltrimethoxysilane from experimentally
generated spectra.
Peak assignment
ı (ppm)
Structure
3
T
−51
RSi(OH)3
RSi(OMe)3
R(OH)2Si–O–Si(OH)2R
R(OH)2Si–O–SiR(OH)–O–Si(OH)2R
R(OH)2Si–O–SiR(OH)–O–Si(OH)2R
0
0
T
−55
0
2
2
T
T
T
T
T
T
−61.1
−60.9
−70.5
1
1
2
1
2
T
T
1
2
2
1
1
2
1
2
1
29
Table 2
tra acquisition, under trypsin catalysis, Si resonances were clearly
The pseudo-first order hydrolysis rate constants for PTMS. Hydrolysis rate constants
include the standard deviation determined from line fitting.
0
3
visible at −55.3 ppm (T , PTMS) and −51.5 ppm (T , phenylsilan-
0
0
etriol) (Fig. 1). The identification of these resonances was based
on those previously determined for phenyltriethoxysilane [21] and
methyltriethoxysilane [22,23]. A complete list of all observed 29Si
resonances for these experiments can be found in Table 1. It was
Trypsin (h−1)
−1
Pepsin (h )
[
PTMS] mmol
1.468
0.087 ± 0.000667
0.12 ± 0.0324
0.22 ± 0.276
0.424 ± 0.140
0.332 ± 0.154
0.57 ± 0.149
1.180
expected that two 2 Si resonances should be visible which would
9
0.89

1
08
M.B. Frampton et al. / Journal of Molecular Catalysis B: Enzymatic 66 (2010) 105–112
Fig. 2. Representative time course profiles for the hydrolysis and condensation of phenyltrimethoxysilane by trypsin (left) and pepsin (right). For each panel the evolution
and/or disappearance of PTMS (ꢀ), phenylsilanetriol (♦) and disiloxanes (ꢁ) are shown.
3
.2. Pepsin-mediated processing of phenyltrimethoxysilane
ideal for the hydrolysis of PTMS by pepsin and is a potential expla-
nation for the increased rate of hydrolysis with this enzyme.
Following the initial hydrolysis by pepsin there was an accumu-
lation of phenylsilanetriol and disiloxanes. At no point throughout
the NMR experiment did all of the phenylsilanetriol become
consumed. The curves representing phenylsilanetriol and the
disiloxanes in Fig. 2 remained nearly constant suggesting the attain-
ment of an equilibrium condition.
When PTMS was processed employing pepsin catalysis, the
reaction appeared to proceed in a similar manner to those reactions
mediated by trypsin. However, after approximately 3 h of incu-
bation there was no longer any evidence for the existence of the
PTMS (Fig. 2). Furthermore, the condensation of siloxane linkages
was more expedient, occurring after only 3–5 h. The observed 29Si
resonances were similar to those observed under tryptic catalysis.
A representative time course reaction profile for the enzymatic
hydrolysis of PTMS is shown in Fig. 2. The observed trends for
the hydrolysis and condensation reaction agree well with those
previously published for tetraethoxysilane, methyltriethoxysilane,
and vinyltriethoxysilane [24]. The early stages of the reaction are
dominated by hydrolysis events under both catalytic schemes. A
comparison of the lines representing the hydrolysis of PTMS sug-
gests that pepsin has a greater proficiency for processing PTMS.
Pepsin, unlike trypsin, preferentially hydrolyzes amino acids con-
taining aromatic side chains. Within the active site of pepsin there
are two binding pockets that accommodate aromatic side chains
3.3. Comparative enzyme kinetics
Of the two enzymes, namely trypsin and pepsin, pepsin was
the more proficient enzyme at the hydrolysis of PTMS at all of
the concentrations of PTMS that were assayed (Table 2, Fig. 3).
A comparison of the pseudo-first order rate constants and the
concentration of PTMS resulted in a negatively sloped line (data
not shown). It should be noted that in neither the pepsin nor
the trypsin cases was either enzyme following the traditional
Michaelis-Menton kinetics. This is perhaps not surprising given
that neither enzyme has evolved to process silicon-based com-
pounds as their natural substrates.
[
25]. These binding pockets accommodate the amino acid residues
that are proximal and distal to the bond to be cleaved. This arrange-
ment places the scissile peptide bond in good proximity to the
catalytic carboxylate residues [26]. This structural arrangement is
3.4. Hydrolysis kinetics
A description of the kinetics associated with the hydrolysis of
phenyltrimethoxysilane is a complicated process. We have focused
here on describing the first hydrolysis of phenyltrimethoxysilane
under trypsin-catalyzed conditions. Throughout the course of our
studies the solubility of pepsin in our system became an issue
at higher enzyme concentrations. A description of the pepsin-
mediated reaction kinetics will be the subject of a future report.
The relative proportion of the 29Si resonance for PTMS was used
to compare the hydrolysis kinetics for these biocatalytic processes.
In the first spectrum, the integral representing the resonance for
PTMS was set to a value of 1.0. This resonance was used as a cali-
bration point for all subsequent spectra. Despite the complexity of
the hydrolysis and condensation of alkoxysilanes, the hydrolysis of
a single methoxyl group can be described using a set of differential
equations. The rate of hydrolysis should be a third order reaction
that is dependent upon the relative amounts of PTMS, enzyme cat-
alyst, and water. This condition can be represented by Eq. (1).
−
d[PTMS]
dt
˛
ˇ
ꢀ
rate =
= k [PTMS] [Enzyme] [D O]
(1)
h
2
The bracketed quantities represent the molar concentrations of
the reagents and the exponents (␣, ␤, and ␥) are the order of the
reaction with respect to PTMS [17]. As we have suggested previ-
Fig. 3. A comparison of the trypsin- and pepsin-mediated hydrolysis of PTMS. Pepsin
(
open bars) was more active than trypsin (shaded bars) at all PTMS concentrations.
Error bars represent the standard deviation.

M.B. Frampton et al. / Journal of Molecular Catalysis B: Enzymatic 66 (2010) 105–112
109
the hydrolysis of PTMS can be denoted by Eq. (5).
−
d[PTMS]
dt
ˇ
ꢀ
rate =
= k [PTMS][Enzyme] [D O]
(5)
h
2
In examining the effect of the enzyme on the hydrolysis of PTMS
several 29Si NMR experiments were conducted where the amount
of trypsin was varied (86, 115, 230, and 345 nmol) while the other
contributing factors were held constant. The enzyme effect for
trypsin was determined to be ˇ = 3.549 (Fig. 5). In a similar man-
ner, the effect of changing the amount of water was determined
to be ꢀ = 3.325 (Fig. 5). In the absence of any enzyme hydrolysis of
PTMS was not detectable on the timescale of the NMR experiments,
regardless of the amount of water that was incorporated into the
reaction mixture. Therefore the final rate equation describing the
trypsin-mediated hydrolysis of PTMS is represented by (6):
−
d[PTMS]
dt
1
h
3.5486
3.325
rate =
= k [PTMS][Trypsin]
[D O]
2
(6)
Fig. 4. A representative plot of the pseudo-first order rate constant for the trypsin-
mediated hydrolysis of 1.468 mmol of PTMS in D2O:H2O (3:1). Each set of points
corresponds to an individual trial.
3
.5. Residual activity of entrapped enzymes
ously, the amount of the enzyme that is used remains constant,
and the amount of water that is consumed is less than 10%, and
therefore can also be assumed to be a constant. By defining a new
The entrapment of enzymes within silica matrices is a common
method for the preparation of immobilized catalysts. However, in
many of these cases, the enzyme is added after pre-hydrolyzing
the alkoxysilane precursor and the enzyme is subsequently added
as a buffered solution [27,28]. In most cases there was some loss in
the residual activity of the enzyme. Enzymes have also been suc-
cessfully entrapped in silicone elastomers showing good retention
of enzymatic activity [29–31]. The use of colourimetric assays to
determine residual enzymatic activity is common [31,32]. Trypsin
was assayed for residual activity using the BAEE assay as described
previously [20]. This assay monitors the hydrolysis of the ethyl ester
by trypsin using the change in absorbance at 253 nm. After incuba-
tion of the trypsin containing monoliths in phosphate buffer for
between 24 and 72 h, the leachate did not show any residual enzy-
matic activity. The slope of the line representing the hydrolysis of
the BAEE reagent was nearly zero and subsequently understood to
mean that no hydrolysis of the substrate had occurred. Initially we
were unsure if this was indicative of complete protein denaturation,
the inability of the enzyme to be leached from the silsesquioxane
matrix due to restrictive pore sizes, or some other factor.
1
ˇ
ꢀ
rate constant k = k [Enzyme] [D O] , we can simplify the rate
h
2
h
equation to a pseudo-first order reaction (2).
−
d[PTMS]
dt
1
˛
rate =
= k [PTMS]
(2)
h
The hydrolysis of PTMS is thought to be first order in PTMS and
we can therefore set ˛ = 1. Integration of (2) gives (3) followed by
(
4).
ꢀ
ꢀ
t
t
d[PTMS]
1
−
= kh
dt
(3)
(4)
[PTMS]
0
0
1
ln[PTMS] = ln[PTMS] − k t
0
h
If the hydrolysis of PTMS is first order in PTMS then a plot
of ln[PTMS] versus time (t) should yield a straight line, which
was supported by the experimental data (Fig. 4). The experimen-
tal data shows only a minor deviation from the fitted line. The
slope of this line represents the pseudo-first order rate constant
FTIR was used to examine the change in the intensities of
1
−1
k
while the intercept represents ln[PTMS] . From these equations
the vibrational frequencies of the N–H (∼1550 cm ) and N–D
0
h
−
1
we can derive the complete rate equations for the trypsin-mediated
(∼1450 cm ) bending modes. A monolith was prepared from
hydrolysis of PTMS. The value of the exponent ˇ is determined from
PTMS, D O, and trypsin. The FTIR spectrum indicated that the
ꢁ
ꢂ
2
1
a plot of In kh versus ln[Enzyme] [17]. This is accomplished by
changing the enzyme concentration while keeping the other vari-
ables constant; the exponent ꢀ is determined in a similar manner
enzyme was entrapped within the silsesquioxane matrix as evi-
denced by the amide carbonyl stretching modes observable at
−
1
−1
1650 cm
(Fig. 6). Additionally, two bands at 1537 cm
(N–H
ꢁ
ꢂ
1
−1
by plotting In k versus ln[D O] [17]. The final rate equations for
bending) and 1437 cm (N–D bending) indicated that while there
2
h
Fig. 5. Representative graphs depicting the relative enzyme effect, ˇ, (left panel) and D2O, ꢀ, effect (right). The error bars represent the standard deviation.

1
10
M.B. Frampton et al. / Journal of Molecular Catalysis B: Enzymatic 66 (2010) 105–112
to release the cleaved peptide [35]. To date there has been little
experimental or computational evidence presented to support a
mechanistic understanding of how an enzyme processes a silane
substrate. Although computational modeling has been done in
regards to the protein-mediated biosilification of the diatom cell
wall [36,37], the interaction between serine–silicate complexes has
also been modeled [38]. Morse and coworkers proposed a catalytic
scheme for the in vitro hydrolysis of tetraethoxysilane by silicatein
in which the nucleophilic serine residue attacked the silicon cen-
tre with concomitant release of ethanol [4]. The mode of action
of trypsin is likely similar, though not necessarily identical to that
which was previously proposed.
Targeting a long-term program grounded in the development
of enzyme-mediated biosynthetic methodologies for the produc-
tion of designer polymeric silicon materials, we were compelled to
undertake a preliminary investigating into the reaction mechanism
of the trypsin-mediated hydrolysis of alkoxysilanes using accepted
computational methods.
Fig. 6. A representative FTIR spectrum of PTMS-entrapped trypsin. The PTMS mono-
lith was produced under trypsin catalysis, aged in a 60 C oven for 24 h, and then
stored in a dessicator prior to the spectrum acquisition.
◦
The trypsin catalytic triad is structurally similar to that of sili-
catein, but contains an aspartic acid residue in lieu of asparagine.
It became apparent that two possible mechanisms could be used
to account for the hydrolysis of PTMS by trypsin. These are the (1)
formation of a serine–silicon bond, an analogue of traditional ser-
ine protease chemistry, or the (2) formation of a histidine-silicon
bond which would be consistent with solution-phase organosilicon
chemistry [39].
As a first principles approach toward addressing these fun-
damental mechanistic questions, our initial studies used Houk’s
concept of theozymes as truncated active site constructs for
modeling enzyme catalysis [40,41]. To facilitate this work elec-
tronic structure calculations using density functional theory
employing the gradient-corrected (B3LYP) hybrid functional of
Becke–Lee–Yang and Parr with a double zeta-potential 6-31G(d)
basis set as implemented in Gaussian 03 have been utilized. For the
was some degree of protein denaturation, it was not complete. In
−1
fact the band at 1537 cm
was more intense that the band at
−1
1
437 cm suggesting that H–D exchange was not complete and
that a greater proportion of the protein remained in a native con-
formation.
The relevant bands for the silsesquioxane monolith were located
−
1
2
−1
at 3049 and 3045 cm (sp C–H str), 2941, 2917, and 2843 cm
3
−1
(
C
symmetric and asymmetric sp C–H str), 1593 and 1429 cm (sp2
C str). The peak present at 2404 cm was due to the O–D stretch-
ing mode. The broad band centred at 3286 cm was due to the
hydroxyl groups of trypsin that were sequestered on the inside of
the protein and not available for H–D exchange.
−1
−1
The FTIR data provides some support for the hypothesis that
mobility restrictions were inhibiting the successful hydrolysis of
the BAEE reagent by trypsin, and not the irreversible denaturation
of the enzyme itself. During the sol–gel process enzymes can act
as templates around which the growing sol–gel forms [33,34]. This
could result in cavities within the silica matrix with very defini-
tive geometries making it difficult for substrate molecules to reach
the enzyme. Trypsin contains a large number of hydrophilic side
chains that are solvent accessible that may form favourable inter-
actions with the hydrolyzed silicon species. These interactions may
also serve to hinder conformational flexibility and subsequently
catalysis by the entrapped enzyme as was demonstrated with the
attempted BAEE assays.
purposes of these studies silicic acid (Si(OH) ) was used as a model
4
compound.
The transition state for histidine addition (Fig. 8) and serine
addition (Fig. 9) to the silicon atom have both been modeled and
preliminary computational results suggest that serine and histi-
dine addition are mechanistically rate determining. Analysis of the
transition state energies for both possible mechanisms within a
trypsin theozyme revealed that the addition of the serine residue
is 3.72 kcal/mol (15.56 kJ/mol) more stable than the addition of the
histidine residue, suggesting that serine addition to silicon is the
preferred reaction mode for this condensation.
3.6. Mechanistic studies
The latter observation is consistent with known Si–O
(536 kJ/mol, 128 kcal/mol) versus Si–N (401 kJ/mol, 96 kcal/mol)
During peptide cleavage trypsin, and other serine proteases,
bond strengths [39]. Taken together the highlighted bond strengths
and relative transition state energies suggest that the serine addi-
tion pathway is the operative mechanistic mode of catalysis.
Perhaps the most interesting feature of the serine addition model
forms an acyl-enzyme intermediate initiated by a nucleophilic
attack of a catalytic serine residue at a carbonyl centre to form
a tetrahedral acyl-enzyme intermediate followed by hydrolysis
Fig. 7. Proposed SN2-type reaction mechanism for the initial step of the trypsin-mediated hydrolysis of PTMS illustrating the interaction between all three residues of the
catalytic triad and PTMS.

M.B. Frampton et al. / Journal of Molecular Catalysis B: Enzymatic 66 (2010) 105–112
111
a concerted SN2 reaction in which the TS is neither early nor late,
−
and the OH is the leaving group. Conversely, the transition state
for histidine addition appears to be early, and much more SN1-like,
with silicic acid maintaining a tetrahedral arrangement about the
silicon atom (Fig. 9).
4. Conclusions
The use of biological molecules to produce inorganic–organic
hybrid materials based on silicon is well established. Despite this
a description of the reaction kinetics related to these systems has
not been made readily available. While a description of the reac-
tion kinetics of such a system would be complex, we have shown
that the trypsin-mediated hydrolysis of a single alkoxy moiety is
a pseudo-first order reaction. The relative effect of the enzyme
was determined to be ˇ = 3.549 while the relative effect of water
was ꢀ = 3.325. Prolonged contact with phenyltrimethoxysilane was
not sufficiently deleterious to the enzyme and did not induce the
complete and irreversible denaturation of trypsin.
Fig. 8. A computational model of serine addition to silicic acid in a trypsin theozyme.
A preliminary investigation with computational methods indi-
cated that trypsin catalysis proceeds through serine addition
in a manner analogous to traditional peptide chemistry. Using
a theozyme construct of the active site of trypsin and a sin-
gle molecule of silicic acid, serine addition was 3.72 kcal/mol
(15.56 kJ/mol) more favourable than histidine addition. From this,
a reaction mechanism was proposed for the trypsin-mediated
hydrolysis of PTMS. Under trypsin catalysis serine addition to a
tetrahedral silicon centre leads the silicon to acquire a trigonal
bipyramidal geometry suggesting that the silicon atom is becoming
pentacoordinate. We are currently expanding our effort to model
several mechanistic possibilities for these two enzymes so as to fur-
ther understand the enzyme mediated processing of alkoxysilanes.
Acknowledgements
Fig. 9. A computational model of histidine addition to silicic acid in a trypsin
theozyme.
Funding for this work was provided by the Ontario Partner-
ship for Innovation and Commercialization (OPIC) and the Ontario
Centre for Excellence (OCE). MF was supported by scholarships
from OGS (2007–2008) and OGSST (2008–2009, 2009–2010). The
authors wish to thank Sharcnet for their contribution to the com-
putational studies.
is the Brønsted activation of the approaching silicic acid unit
toward nucleophilic hydroxyl addition as a result of favourable
(
N–H(O) hydrogen bonding. Also, the well-defined position of the
histidine residue proximal to the serine hydroxyl group allows for
rapid proton transfer from serine to histidine after nucleophilic
addition. As a result it is postulated that the trypsin-mediated
hydrolysis of PTMS may proceed through a stable siloxy-enzyme
intermediate in a similar manner to the processing of natural
peptide substrates by trypsin. Silicon is known to expand its
valence shell to accommodate extra ligands [39].
Based on the available modeling data, a proposed reaction
mechanism for the nucleophilic addition of serine to PTMS by
trypsin is presented in (Fig. 7). The co-operation of Asp102 and His57
serve to polarize the O–H bond of the nucleophilic Ser195 to pro-
mote serine addition to the silicon centre. At this point a leaving
hydroxyl group may directly abstract the serine hydroxyl proton,
or alternatively, silicon may expand its coordination number to
attain pentacoordinate geometry. However, taking into considera-
tion the polarizing effect of the His57 and Asp102 residues, we feel
that this particular process proceeds in a manner analogous to pep-
tide chemistry with silicon adopting trigonal bipyramidal geometry
References
[
[
[
1] C.J. Brinker, G.W. Scherer, Sol–GelScience, ThePhysicsandChemistry ofSol–Gel
Processing, Academic Press, New York, NY, 1990.
2] M. Frampton, A. Vawda, J. Fletcher, P.M. Zelisko, Chem. Commun. (2008)
5544–5545.
3] M. Frampton, R. Simionescu, P.M. Zelisko, Silicon 1 (2009) 47–56.
[4] Y. Zhou, K. Shimizu, J.N. Cha, G.E. Stucky, D.E. Morse, Angew. Chem. Int. Ed. 38
1999) 779–782.
(
[
5] J.N. Cha, K. Shimizu, Y. Zhou, S.C. Christensen, B.F. Chmelka, G.D. Stucky, D.E.
Morse, Proc. Natl. Acad. Sci. U.S.A. 96 (1999) 361–365.
[6] P. Buisson, E. El Rassy, S. Maury, A.C. Pierre, J. Sol–Gel Sci. Technol. 27 (2003)
73–379.
3
[
[
7] T. Coradin, A. Coupe, J. Livage, Colloids Surf. B 29 (2003) 189–196.
8] A.R. Bassindale, K.F. Brandstadt, T.H. Lane, P.G. Taylor, J. Inorg. Biochem. 96
(2005) 401–406.
[9] C.C. Perry, T. Keeling-Tucker, Chem. Commun. (1998) 2587–2588.
10] N. Kröger, R. Deutzmann, M. Sumper, J. Biol. Chem. 276 (2001) 26066–26070.
11] W.E.G. Müller, U. Schlo␤macher, X. Wang, A. Boreiko, D. Brandt, S.E. Wolf, W.
Tremel, H.C. Schröder, FEBS J. 275 (2008) 362–370.
[
[
(
Fig. 8).
The geometry of silicic acid in the energetically favoured serine-
[12] A. Maraite, M.B. Ansorge-Schumacher, B. Ganchegui, W. Leitner, G. Grogan, J.
Mol. Catal. B: Enzym. 56 (2009) 24–28.
[
[
13] R.J. Hook, J. Non-Cryst. Solids 195 (1996) 1–15.
14] R.A. Assink, B.D. Kay, Ann. Rev. Mater. Sci. 21 (1991) 491–513.
addition TS is highly indicative of a pentacoordinate trigonal
bipyramidal structure: Three hydroxyl groups equatorial with the
silicon atom form angles of 114 , 119 , and 120 , and the attack-
ing hydroxyl of serine and its axial hydroxyl counterpart on the
[15] I. Artaki, M. Bradley, T.W. Zerda, J. Jonas, J. Phys. Chem. 89 (1985) 4399–4404.
[16] C.T.G. Knight, R.J. Balec, S.D. Kinrade, Angew. Chem. Int. Ed. 46 (2007)
◦
◦
◦
8148–8152.
[
17] R. Liu, Y. Xu, D. Wu, Y. Sun, H. Gao, H. Yuan, F. Deng, J. Non-Cryst. Solids 343
(2004) 61–70.
◦
silicon form an angle of 178 . Thus, the geometry is indicative of

1
12
M.B. Frampton et al. / Journal of Molecular Catalysis B: Enzymatic 66 (2010) 105–112
[
18] H. Dong, M. Lee, R.D. Thomas, Z. Zhang, R.F. Reidy, D.W. Meuller, J. Sol–Gel Sci.
[29] A.M. Ragheb, M.A. Brook, M. Hrynyk, Biomaterials 26 (2005) 1653–1664.
[30] A.M. Ragheb, O.E. Hileman, M.A. Brook, Biomaterials 26 (2005) 6973–
6983.
[31] Y. Poojari, A.S. Palsule, S.J. Clarson, R.A. Gross, Silicon 1 (2009) 37–45.
[32] P.M. Zelisko, M.A. Brook, Langmuir 18 (2002) 8982–8987.
[33] I. Gill, A. Ballesteros, Trends Biotechnol. 18 (2000) 282–296.
[34] C.D. Tran, D. Ilieva, S. Challa, J. Sol–Gel Sci. Technol. 32 (2004) 207–217.
[35] H.R. Horton, L.A. Moran, K.G. Scrimgeour, M.D. Perry, J.D. Rawn, Principles
of Biochemistry, fourth ed., Pearson Prentice Hall, Upper Saddle River, NJ,
1993.
[36] K.D. Lobel, J.K. West, L.L. Hench, Mar. Biol. 126 (1996) 353–360.
[37] K.D. Lobel, J.K. West, L.L. Hench, J. Mater. Sci. Lett. 15 (1996) 648–650.
[38] N. Sahai, J.A. Tossell, Geochim. Cosmochim. Acta 65 (2001) 2043–2053.
[39] M.A. Brook, Silicon in Organic, Organometallic, and Polymer Chemistry, John
Wiley & Sons, New York, NY, 2000.
[40] D.J. Tantillo, J. Chen, K.N. Houk, Curr. Opin. Chem. Biol. 2 (1998) 743–750.
[41] G. Ujaque, D.J. Tantillo, Y. Hu, K.N. Houk, K. Hotta, D. Hilvert, J. Comput. Chem.
24 (2003) 98–110.
Technol. 28 (2003) 5–14.
[
[
19] K.D. Keefer, Mater. Res. Soc. Symp. Proc. 32 (1984) 15–24.
20] V. Worthington (Ed.), Worthington Enzyme Manual: Enzymes and Related Bio-
chemicals, Worthington Biochemical Corporation, Lakewood, NJ, 1993.
21] M. Kuniyoshi, M. Takahashi, Y. Tokuda, T. Yoko, J. Sol–Gel Sci. Technol. 39 (2004)
[
[
1
75–183.
22] Y. Sugahara, S. Okada, S. Sato, K. Kuroda, C. Kato, J. Non-Cryst. Solids 167 (1994)
1–28.
2
[
[
23] Y. Sugahara, T. Inoue, K. Kuroda, J. Mater. Chem. 7 (1997) 53–59.
24] A. Vainrub, F. Devreux, J.P. Boiloit, F. Chaput, M. Sarkar, Mater. Sci. Eng. B37
(
1996) 197–200.
[
[
25] M.W. Hunkapiller, J.H. Richards, Biochemistry 11 (1972) 2829–2839.
26] M.G. Williams, J. Wilsher, P. Nugent, A. Mills, V. Dhanaraj, M. Fabri, J. Sedlacek,
J.M. Uusitalo, M.E. Pentilla, J.M. Pitts, T.L. Blundell, Prot. Eng. 10 (1997) 991–997.
27] C. Lin, C. Shih, L. Chau, Anal. Chem. 79 (2007) 3757–3763.
[
[
28] T.R. Besanger, B. Easwaramoorthy, J.D. Brennan, Anal. Chem. 76 (2004)
6
470–6475.