Screening of inhibitors using enzymes entrapped in sol-gel-derived materials

Article abstract of DOI:10.1021/ac026370i

In recent years, a number of new methods have been reported that make use of immobilized enzymes either on microarrays or in bioaffinity columns for high-throughput screening of compound libraries. A key question that arises in such methods is whether immobilization may alter the intrinsic catalytic and inhibition constants of the enzyme. Herein, we examine how immobilization within sol-gel-derived materials affects the catalytic constant (k_cat), Michaelis constant (K_M), and inhibition constant (K_I) of the clinically relevant enzymes Factor Xa, dihydro-folate reductase, cyclooxygenase-2, and γ-glutamyl transpeptidase. These enzymes were encapsulated into solgel-derived glasses produced from either tetraethyl orthosilicate (TEOS) or the newly developed silica precursor diglyceryl silane (DGS). It was found that the catalytic efficiency and long-term stability of all enzymes were improved upon entrapment into DGS-derived materials relative to entrapment in TEOS-based glasses, likely owing to the liberation of the biocompatible reagent glycerol from DGS. The K_M values of enzymes entrapped in DGS-derived materials were typically higher than those in solution, whereas upon entrapment, k_cat values were generally lowered by a factor of 1.5-7 relative to the value in solution, indicating that substrate turnover was limited by partitioning effects or diffusion through the silica matrix. Nonetheless, the apparent K_I value for the entrapped enzyme was in most cases within error of the value in solution, and even in the worst case, the values differed by no more than a factor of 3. The implications of these findings for high-throughput screening are discussed.

Full text of DOI:10.1021/ac026370i

Anal. Chem. 2003, 75, 2382-2391
Screening of Inhibitors Using Enzymes Entrapped
in Sol-Gel-Derived Materials
Travis R. Besanger,† Yang Chen,† Anil K. Deisingh,† Richard Hodgson,† Wen Jin,† Stanislas Mayer,‡
Michael A. Brook,† and John D. Brennan*,†
Department of Chemistry, McMaster University, Hamilton, Ontario, L8S 4M1, Canada, and Department of Chemistry, York
University, Toronto, Ontario, M3J 1P3, Canada
accessibility,4,8 reaction kinetics,2,9 activity,10-20 and stability,21-28
and their many applications for catalysis, sensing, and affinity
chromatography.29 However, few reports exist describing the use
of sol-gel entrapped enzymes for inhibition studies, or as potential
targets for high-throughput drug-screening (HTS). This is a
particularly important issue, given the emergence of HTS methods
based on immobilized enzymes.30,31
In recent years, a number of new methods have been
reported that make use of immobilized enzymes either on
microarrays or in bioaffinity columns for high-throughput
screening of compound libraries. A key question that
arises in such methods is whether immobilization may
alter the intrinsic catalytic and inhibition constants of the
enzyme. Herein, we examine how immobilization within
sol-gel-derived materials affects the catalytic constant
(kcat), Michaelis constant (KM), and inhibition constant
(KI) of the clinically relevant enzymes Factor Xa, dihydro-
folate reductase, cyclooxygenase-2 , and γ-glutamyl trans-
peptidase. These enzymes were encapsulated into sol-
gel-derived glasses produced from either tetraethyl ortho-
silicate (TEOS) or the newly developed silica precursor
diglyceryl silane (DGS). It was found that the catalytic
efficiency and long-term stability of all enzymes were
improved upon entrapment into DGS-derived materials
relative to entrapment in TEOS-based glasses, likely owing
to the liberation of the biocompatible reagent glycerol from
DGS. The KM values of enzymes entrapped in DGS-
derived materials were typically higher than those in
solution, whereas upon entrapment, kcat values were
generally lowered by a factor of 1 .5 -7 relative to the value
in solution, indicating that substrate turnover was limited
by partitioning effects or diffusion through the silica
matrix. Nonetheless, the apparent KI value for the en-
trapped enzyme was in most cases within error of the
value in solution, and even in the worst case, the values
differed by no more than a factor of 3 . The implications
of these findings for high-throughput screening are dis-
cussed.
To date, most studies of sol-gel-entrapped enzymes have made
use of the silane precursors tetramethyl orthosilicate (TMOS) or
(4) Edmiston, P. L.; Wambolt, C. L.; Smith, M. K.; Saavedra, S. S. J. Colloid
Interface Sci. 1 9 9 4 , 163, 395.
(5) Jordan, J. D.; Dunbar, R. A.; Bright, F. V. Anal. Chem. 1 9 9 5 , 67, 2436.
(6) Gottfried, D. S.; Kagan, A.; Hoffman, B. M.; Friedman, J. M. J. Phys. Chem.
B 1 9 9 9 , 103, 2803.
(7) Doody, M. A.; Baker, G. A.; Pandey, S.; Bright, F. V. Chem. Mater. 2 0 0 0 ,
12, 1142.
(8) Wambolt, C. L.; Saavedra, S. S. J. Sol-Gel Sci. Technol. 1 9 9 6 , 7, 53.
(9) Shen, C.; Kostic, N. M. J. Am. Chem. Soc. 1 9 9 7 , 119, 1304.
(10) Braun, S.; Shtelzer, S.; Rappoport, S.; Avnir, D.; Ottolenghi, M. J. Non-Cryst.
Solids 1 9 9 2 , 147, 739.
(11) Avnir, D.; Braun, S.; Lev, O.; Ottolenghi, M. Chem. Mater. 1 9 9 4 , 6, 1605.
(12) Wang, R.; Narang, U.; Prasad, P. N.; Bright, F. V. Anal. Chem. 1 9 9 3 , 65,
2671.
(13) Ellerby, L. M.; Nishida, C. R.; Nishida, F.; Yamanaka, S. A.; Dunn, B.;
Valentine, J. S.; Zink, J. I. Science, 1 9 9 2 , 225, 1113.
(14) Wu, S.; Ellerby, L. M.; Cohan, J. S.; Dunn, B.; El-Sayed, M. A.; Valentine, J.
S.; Zink, J. I. Chem. Mater. 1 9 9 3 , 5, 115.
(15) Dave, B. C.; Soyez, H.; Miller, J. M.; Dunn, B.; Valentine, J. S.; Zink, J. I.
Chem. Mater. 1 9 9 5 , 7, 1431.
(16) Yamanaka, S. A.; Nishida, F.; Ellerby, L. M.; Nishida, C. R.; Dunn, B.;
Valentine, J. S.; Zink, J. I. Chem. Mater. 1 9 9 2 , 4, 495.
(17) Dave, B. C.; Dunn, B.; Valentine, J. S.; Zink, J. I. Anal. Chem. 1 9 9 4 , 66,
1120A.
(18) Blyth, D. J.; Aylott, J. W.; Richardson, D. J.; Russell, D. A. Analyst 1 9 9 5 ,
120, 2725.
(19) Aylott, J. W.; Richardson, D. J.; Russell, D. A. Analyst 1 9 9 7 , 122, 77.
(20) Williams, A. K.; Hupp, J. T. J. Am. Chem. Soc. 1 9 9 8 , 120, 4366.
(21) Braun, S.; Rappoport, S.; Zusman, R.; Avnir, D.; Ottolenghi, M. Mater. Lett.
1 9 9 0 , 10, 1.
(22) Heichal-Segal, O.; Rappoport, S.; Braun, S.; Biotechnology 1 9 9 5 , 13, 798.
(23) Reetz, M. T.; Zonta, A.; Simpelkamp, J. Biotechnol. Bioeng. 1 9 9 6 , 49, 527.
(24) Narang, U.; Prasad, P. N.; Bright, F. V.; Kumar, K.; Kumar, N. D.; Malhotra,
B. D.; Kamalasanan, M. N.; Chandra, S. Chem. Mater. 1 9 9 4 , 6, 1596.
(25) Narang, U.; Prasad, P. N.; Bright, F. V.; Ramanathan, K.; Kumar, N. D.;
Malhotra, B. D.; Kamalasanan, M. N.; Chandra, S. Anal. Chem. 1 9 9 4 , 66,
3139.
An emerging method for immobilization of proteins is their
entrapment into a porous, inorganic silicate matrix that is formed
via a low-temperature sol-gel processing method.1 Numerous
reports have appeared describing both fundamental aspects of
entrapped proteins, such as their conformation,2-4 dynamics,5-7
* To whom correspondence should be addressed. Tel: (905) 525-9140 (ext.
27033). Fax: (905) 522-2509. E-mail: brennanj@mcmaster.ca. Internet: http:/ /
www.chemistry.mcmaster.ca/ faculty/ brennan.
(26) Jordan, J. D.; Dunbar, R. A.; Bright, F. V. Anal. Chim. Acta 1 9 9 6 , 332, 83.
(27) Yamanaka, S. A.; Dunn, B.; Valentine, J. S.; Zink, J. I. J. Am. Chem. Soc.
1 9 9 5 , 117, 9095.
(28) Kauffmann, C.; Mandelbaum, R. T. J. Biotechnol. 1 9 9 8 , 62, 169.
(29) Jin, W.; Brennan, J. D. Anal. Chim. Acta 2 0 0 2 , 461, 1.
(30) Schriemer, D. C.; Bundle, D. R.; Li, L.; Hindsgaul, O. Angew. Chem., Int.
Ed. Engl. 1 9 9 8 , 37, 3383.
† McMaster University.
‡ York University.
(1) Gill, I. Chem. Mater. 2 0 0 1 , 13, 3404.
(2) Zheng, L.; Reid, W. R.; Brennan, J. D. Anal. Chem. 1 9 9 7 , 69, 3940.
(3) Zheng, L.; Brennan, J. D. Analyst 1 9 9 8 , 123, 1735.
(31) Macbeath, G.; Schreiber, S. L. Science 2 0 0 0 , 289, 1760.
2382 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003
10.1021/ac026370i CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/08/2003

tetraethyl orthosilicate (TEOS).29 However, the production of
methanol or ethanol is often detrimental to entrapped proteins
and can lead to significant changes in the properties of the
enzyme, including the Michaelis constant (KM), catalytic constant
(kcat), and inhibition constant (KI).29 The latter is of particular
importance in terms of using entrapped enzymes for drug-
screening purposes. Recently, a number of new biocompatible
silane precursors and processing methods have been reported
that are based on glycerated silanes,1 sodium silicate starting
materials,32 or aqueous processing methods that involve removal
of alcohol byproducts by evaporation before the addition of
proteins.33 Furthermore, our group has reported on the develop-
ment of the new silane precursor diglyceryl silane,34 which is
capable of maintaining entrapped enzymes in an active state for a
significant amount of time. A question that arises is whether
entrapment of enzymes using this precursor may result in their
catalytic and inhibition constants being similar to those in solution.
In this paper, we demonstrate the successful sol-gel encap-
sulation of the clinically relevant enzymes Factor Xa, dihydrofolate
reductase (DHFR), cyclooxygenase-2 (COX-II), and γ-glutamyl-
transpeptidase (γ-GT) into DGS-derived glasses and compare the
catalytic and inhibition properties of the entrapped enzymes to
those obtained in solution. Factor Xa is a serine protease that plays
a central role in the blood coagulation cascade,35,36 and is being
promoted as a potentially efficacious target for antithrombosis
drug development.37 DHFR catalyzes the NADPH-dependent
reduction of dihydrofolate (DHF) to tetrahydrofolate, which is then
used as a cofactor in the biosynthesis of thymidylate, purines, and
several amino acids.38-40 DHFR is an essential enzyme in the cell
and is the target for antifolate drugs, which act by inhibiting the
enzyme in malignant or parasitic cells.41 COX-II binds to a variety
of nonsteroidal antiinflammatory drugs (NSAIDs), which are
commonly used for the treatment of pain and inflammation.42-45
Recently, two specific COX-II inhibitors have been approved,
rofecoxib (Vioxx) and celecoxib (Celebrex), which do not alter
the function of COX-I, reducing gastrointestinal irritation. γ-GT
is a key membrane-associated enzyme involved in glutathione
(GST) homeostatis.46 γ-GT catalyzes the initial step of the transfer
of the γ-glutamyl moiety of γ-glutamyl-derived peptides, such as
GST, to an acceptor molecule,46 initiating a cascade of enzymatic
reactions that break the tripeptide glutathione into its constituent
amino acids so that they can be transported through the cell
membrane. It has been reported that some rapidly growing cancer
cells express high levels of γ-GT so that their intracellular cysteine
requirements can be accommodated.46-,49 Therefore, inhibition of
γ-GT may provide a route to slow the growth of certain metastatic
cells.
Herein, we report on catalytic and inhibition studies of the four
enzymes in solution and when entrapped in sol-gel-derived
materials derived from DGS or TEOS. Given that entrapment can
often alter the kinetic rate constant, Michaelis constant, and
inhibition constants of enzymes, a primary goal of this study was
to ascertain if entrapment into biocompatible silica would result
in the maintenance of the solution behavior of the enzyme. In
particular, it was important to determine if partitioning of inhibitor
between the solution and silica matrix or mass transport limitations
would lead to changes in the apparent KI values of the entrapped
proteins and whether such changes would be problematic in terms
of the use of immobilized proteins for drug screening.
EXPERIMENTAL SECTION
Chemicals. The enzymes Factor Xa (from bovine plasma) and
γ-glutamyl transpeptidase (lyophilized powder from bovine kidney)
were obtained from Sigma (Toronto, ON). Dihydrofolate reductase
(from E. coli) was provided by Dr. Eric Brown (McMaster
University). Cyclooxygenase-II (holoenzyme in a 1:1 complex with
hematin) was donated by Merck Frosst Inc. (Montreal, Canada).
Tetraethyl orthosilicate (TEOS, 99.999%) and benzamidine were
obtained from Aldrich (Oakville, ON). Diglyceryl silane precursors
were prepared from TEOS as described below. The chromogenic
substrate S-2222 (for Factor Xa activity assays) was obtained from
DiaPharma (West Chester, OH). NADPH, -glutamic acid γ-(p-
L
nitroanilide) (GPN), acivicin, glycylglycine, dihydrofolic acid
(DHF), trimethoprim, pyrimethamine, folate, aminopterin, dithio-
threitol, arachidonic acid (sodium salt), N,N,N,N′-tetramethyl-p-
phenylenediamine (TMPD), and Triton X-100 were obtained from
Sigma (Oakville, ON). 1-Palmitoyl-2-oleoyl-sn-glycero-3-[phospho-
rac-(1-glycerol)] (POPG) and 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphocholine (POPC) were purchased from Avanti Polar Lipids
(Alabaster, AL). Analogues of celecoxib (termed V2-01, V2-05,
V2-06, and V2-09) that were used as inhibitors of COX-II were
donated by Dr. Michael Organ (York University, Canada). All
water was distilled and deionized using a Milli-Q synthesis A10
water purification system. All other reagents were of analytical
grade and were used as received.
(32) Bhatia, R. B.; Brinker, C. J.; Gupta, A. K.; Singh, A. K. Chem. Mater. 2 0 0 0 ,
12, 2434.
(33) Ferrer, M. L.; del Monte, F.; Levy, D. Chem. Mater. 2 0 0 2 , 14, 3619.
(34) Brook, M. A.; Chen, Y.; Guo, K.; Zhang, Z.; Brennan, J. D. Chem. Mater.
2 0 0 3 , submitted.
(35) Dave, E. W.; Fujikawa, K.; Kisiel, W. Biochemistry 1 9 9 1 , 30, 10363.
(36) Kaiser, B. Drugs Future 1 9 9 8 , 23, 423.
(37) Maignan, S.; Guilloteau, J.-P.; Pouzieux, S.; Sledeki, Y. M. C.; Becker, M.
R.; Klein, S. I.; Ewing, W. R.; Pauls, H. W.; Spada, A. P.; Mikol, V. J. Med.
Chem. 2 0 0 0 , 43, 3226.
(38) Hitchings, G. H. Angew. Chem., Int. Ed. Engl. 1 9 8 9 , 29, 879.
(39) Stone, S. R.; Morrison, J. F. Biochemistry 1 9 8 8 , 27, 5493.
(40) Benkovic, S. J.; Fierke, C. A.; Naylor, A. M. Science 1 9 8 8 , 239, 1105.
(41) Polshakov, V. I.; Biekofsky, R. R.; Birdsall, B.; Feeney, J. J. Mol. Struct. 2002,
602-603, 257.
P rocedures. Preparation of DGS. TEOS was distilled to
remove any residual water, and a neat mixture of the anhydrous
TEOS (2.08 g, 10.0 mmol) and glycerol (1.84 g, 20.0 mmol) was
heated at 130 °C for 36 h, during which time EtOH was distilled
off. Complete removal of EtOH and unreacted starting materials
at 140 °C in vacuo gave DGS as a solid compound that was not
contaminated with residual ethanol. Structural characterization of
DGS by NMR and the properties of DGS-derived silica are
reported elsewhere.34
(42) Vane, J. R. Nature 1 9 7 1 , 231, 232.
(43) Lipsky, P. E. Am. J. Med. 2 0 0 1 , 110, Supplement 3A, 1S.
(44) Hawkey, C. J.; Jackson, L.; Harper, S. E.; Simon, T. J.; Mortensen, E.; Lines,
C. R. Aliment. Pharmacol. Ther. 2 0 0 1 , 15, 1.
(47) Djavaheri-Mergny, M.; Accaoui, M.; Rouillard, D.; Wietzerbin, J. Mol. Cell.
Biochem. 2 0 0 2 , 232, 103.
(45) Chandrasekharan, N. V.; Dai, H.; Roos, K. L. T.; Evanson, N. K.; Tomsik,
J.; Elton, T. S.; Simmons, D. L. Proc. Natl. Acad. Sci. U.S.A. 2 0 02 , 99, 13926.
(46) Hanigan, M. H. Chem. Biol. Int. 1 9 9 8 , 111, 333.
(48) Bluvshtein, E.; et al. Eur. J. Biochem. 1 9 9 9 , 260, 844.
(49) Antczak, C.; Bauvois, B.; Monneret, C.; Florent, J. C. J. Controlled Release
2 0 0 1, 74, 255.
Analytical Chemistry, Vol. 75, No. 10, May 15, 2003 2383

Scheme 1
Entrapment of Enzymes: Buffered solutions of the enzymes
were prepared as follows: Factor Xa was dissolved at a concentra-
tion of 0.56 µg mL-1 in 50 mM Tris-HCl buffer at pH 8.3
containing 0.5 M NaCl and 5 mM CaCl2; DHFR was dissolved at
a concentration of 20 nM in a buffer composed of 100 mM HEPES,
pH 8.0, containing 1 mM DTT and 0.01% Triton X-100; COX-II
was dissolved at a concentration of 40 µg mL-1 in a buffer
composed of 20 mM MOPS, pH 7.5, containing 1 mM phenol and
0.01% octyl glucoside. γ-Glutamyl transpeptidase was prepared as
a lipid suspension in buffer. Liposomes were first prepared by
dispensing 50 mg mL-1 chloroform stocks of POPG and POPC
in a disposable glass vial and removing the bulk of the organic
solvent under a dry nitrogen stream. The remaining solvent was
then evaporated under vacuum for 1 h. The dried lipid films were
then rehydrated to final lipid concentrations of 2.2 µM with a buffer
composed of 200 mM glycylglycine, 20 mM MgCl2, and 50 mM
Trizma HCl containing 13 µM of γ-GT to create large multilamellar
liposomes with embedded γ-GT. This liposome suspension was
then extruded with an Avanti MiniExtruder through 600-nm pores
to create a suspension of unilamellar liposomes with a mean
diameter of 600 nm.
Entrapment of enzymes into TEOS-derived materials was
performed according to a previously reported procedure.2 Briefly,
4.5 mL of TEOS, 1.4 mL of water, and 0.l mL of 0.1 N HCl were
mixed and sonicated for 1 h to produce a homogeneous solution.
Equal volumes of the hydrolyzed TEOS solution and buffered
solutions of the appropriate enzyme were mixed in a microtiter
well (total volume, 50 µL) and allowed to gel (typically 5-20 min).
Entrapment of enzymes into DGS-derived materials was ac-
complished by dissolving 0.20 g of solid DGS in 600 µL of water
at pH 7 and sonicating at room temperature (22 ( 2 °C) for 45
min. A 25-µL portion of the DGS solution was then mixed with 25
µL of the appropriate buffered enzyme solution in the well of a
microtiter plate. Gelation typically occurred within 10 min. In all
cases, the microtiter plate was covered with Parafilm, and a hole
was punched through the Parafilm on the top of each well to allow
slow drying of the gel to occur. The plate was then stored at 4 °C
until tested.
varying levels of benzamidine (0-600 µM) for 10 min before the
addition of the substrate solution. An incubation time of 10 min
was found to be sufficient to achieve reproducible inhibition
constants; further incubation did not alter the KI values.
DHFR uses two substrates in its reaction, and thus, two
separate activity assays were performed. To obtain KM values for
NADPH, the enzyme activity was measured using 50 µM DHF
and 0-50 µM NADPH in a total volume of 200 µL with a total of
384 fmol of DHFR/ well in all assays. To obtain the KM values for
DHF, enzyme activity was measured using 50 µM NADPH and
0-50 µM DHF. To obtain the KI values of the different inhibitors,
the enzyme activity was determined in the presence of different
inhibitor concentrations (preincubated with the enzyme for 10
min) using 50 µM NADPH and 0-25 µM DHF. In all cases, the
absorbance change of NADPH at 340 nm was monitored for 10
min for solution assays and 30 min for sol-gel-based assays and
was converted to a rate of substrate turnover using ꢀmax ) 12 000
M-1 cm-1
.
Solution- and sol-gel-based assays of COX-II utilized 1 µg of
protein/ well (for solution, this corresponded to 6.25 µL of buffer
containing 160 µg mL-1 enzyme), 95 µL of arachidonic acid (0-
100 µM), and 95 µL of TMPD (final concentration of 85 µM).
Note: these solutions needed to be prepared fresh daily because
arachidonic acid is air-, light-, and moisture-sensitive, and TMPD
is photosensitive. The absorbance changes of TMPD were
monitored at 590 nm for 5 min for solution assays and 30 min for
sol-gel-based assays and converted to substrate turnover rates
using ꢀmax ) 11 000 M-1 cm-1. For studies of inhibition, 10 µL of
a solution containing 0-50 µM of the appropriate inhibitor was
added to the sample and allowed to equilibrate for 10 min before
introduction of the arachidonic acid (0-95 µM) and TMPD (85
µM).
Solution assays of γ-GT activity were performed by mixing 175
µL of a solution containing varying concentrations of GPN (0.25-
3.0 mM) with 5 µL of the enzyme solution (13 µM) to each well.
For assays of sol-gel-entrapped γ-GT, the activity was monitored
after addition of 150 µL of GPN to the top of a monolith that
contained an amount of enzyme identical to that used in the
solution assays. In both cases, the absorbance change at 410 nm
was then monitored for the next 35 min, and the change in
absorbance with time was used to determine the rate of product
formation using ꢀmax ) 10 200 M-1 cm-1. Inhibition assays were
performed in a similar manner, except that various concentrations
(0-78 µM) of the γ-GT inhibitor acivicin were equilibrated with
the enzyme (0-180 min) before adding the substrate.
Enzyme Assays: All enzyme assays and inhibition studies were
performed in 96-well plates using a TECAN Safire absorbance/
fluorescence plate reader operated in absorbance mode. In all
cases, sol-gel entrapped enzyme samples were washed three
times with buffer solution to remove excess glycerol before the
assay commenced. The enzymatic activity of free and entrapped
Factor Xa was measured by adding 200 µL of a solution containing
varying concentrations of the substrate S-2222. The reaction
catalyzed by Factor Xa is shown in Scheme 1. For solution assays,
2 µL of a 5.6 µg mL-1 solution of Factor Xa in buffer was also
added. The absorbance change at 405 nm was then monitored
for the next 20 min (in solution) or 60 min (for entrapped Factor
Xa) and was converted to a rate of substrate turnover using a
molar extinction coefficient of 10 200 M-1 cm-1. For inhibition
studies, the free and entrapped enzymes were incubated with
For all assays, the initial rate of product formation (V ) was
0
evaluated from at least 10 data points over the first 5 min of
reaction, where the slope of the response curve was linear. The
error margins for the determination of V values were typically
0
(10%, as reflected in the errors in the kcat values. For all enzymes
studied, the KM and kcat values were calculated by generating
double reciprocal (Lineweaver-Burk) plots relating V -1 to
0
[substrate concentration]-1, and fitting these to a linear model.
2384 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

Figure 1. Typical absorbance changes vs time obtained from Factor Xa in solution (panel A) and in DGS-derived glasses (panel B).
Figure 2. Typical kinetic data obtained from Factor Xa, DHFR, COX-II, and γ-GT in solution (b), DGS-derived glasses (O) and TEOS-derived
glasses (1) after 3 days of aging. The activity of Factor Xa in TEOS-derived materials after 7 days of aging is also shown (∇).
Inhibition constants were calculated by assessing the changes in
the KM values of the enzyme in the presence of varying levels of
inhibitor, according to the equation
could be obtained, and it is clear that the slope of the response
curve increases with substrate concentration, as expected. Similar
absorbance-concentration plots were obtained for all other
enzymes tested, both in solution and in DGS-derived materials.
In no case did we observe a lag time before the increase in
absorbance, indicating that the rate of diffusion of substrate into
the silica matrix was relatively fast. Examination of the activity of
the solution surrounding the silica matrix provided no absorbance
changes (within error), ruling out the possibility of the responses
being due to protein that had leached from the matrix.
[I]
KI )
(1)
(KM′/ KM) - 1
where KM is the Michaelis constant in the absence of inhibitor,
KM′ is the Michaelis constant in the presence of inhibitor, and [I]
is the concentration of inhibitor.
Figure 2 shows typical kinetic data (V0 vs [S]) obtained from
Factor Xa, DHFR, COX-II, and γ-GT in solution, DGS-derived
glasses, and TEOS-derived glasses after 3 days of aging (note:
COX-II and DHFR showed no activity in TEOS-based glasses).
In all cases, the enzymatic reactions followed Michaelis-Menten
kinetics, allowing for the extraction of kinetic data. Tables 1 and
2 provide the kcat and KM values and the overall catalytic efficiency
RESULTS AND DISCUSSION
Catalytic Efficiency of Free and Entrapped Enzymes.
Figure 1 shows plots of absorbance vs time for Factor Xa in
solution (panel A) and when entrapped in DGS-derived silica
(panel B) in the presence of varying amounts of substrate. In both
cases, there is a broad linear range from which initial rate data
Analytical Chemistry, Vol. 75, No. 10, May 15, 2003 2385

values. Since KM is defined as (k2 - k-1)/ k1 for the reaction
Table 1. Kinetic Parameters of Free and Entrapped
Enzymes
k2
E + S yk1z ES
98 E + P, restrictions in mass transport would not
k-1
affect off rates (k-1 and k2) but would be expected to lower the
value of k1 (slower on-rate), causing the value of KM to increase
relative to the solution value, as was observed in upon entrapment
of most of the enzymes.53
KM (µM)
kcat (s-1
)
kcat/ KM (M-1 s-1
)
Factor Xa in solution 360 ( 40
37 ( 3
21 ( 4
27 ( 3
9 ( 2
105
Factor Xa in TEOS
Factor Xa in DGS
COX-II in solution
COX-II in TEOS
COX-II in DGS
γ-GT in solution
γ-GT in TEOS
400 ( 40
500 ( 50
5.0 ( 2.0
no activity no activity
6.0 ( 2.0
1950
3.5 × 104
4.5 × 104
1.8 × 106
no activity
9.5 × 105
2.0 × 107
3.6 × 106
3.0 × 106
Mass transport limitations may also explain why KM values
appeared to be more severely affected upon entrapment into DGS
relative to TEOS. This effect is likely due to residual glycerol in
the DGS samples that lead to decreases in transport of the
substrate to the enzyme. Indeed, examination of enzyme kinetics
in solutions containing 25% glycerol (v/ v) (to increase viscosity
and slow diffusion) confirmed that kcat values were lowered and
KM values increased in the presence of glycerol. For example,
γ-GT in 25% glycerol showed a kcat value of 2.86 × 104 s-1 (70% of
the solution value) and a KM value of 8.2 mM (∼4-fold higher
than in solution), leading to a decrease of 6-fold in catalytic
efficiency. Similar effects were observed for each of the other
enzymes (Factor Xa: KM ) 700 µM, kcat ) 24 s-1 corresponding
to a 3-fold decrease in efficiency; COX-II: KM ∼ 2.3 µM, kcat ) 7
4 ( 2
3.9 × 104
5.2 × 103
1.7 × 104
1420
5690
γ-GT in DGS
Table 2. Kinetic Parameters of Free and Entrapped
DHFR
NADPH
DHF
KM
(µM)
kcat
kcat/ KM
KM
(µM)
kcat
kcat/ KM
(s-1
)
(M-1 s-1
)
(s-1
)
(M -1 s-1
)
solution 3.0 ( 0.5
6.8
2 × 106 2 ( 1
11.8
1.6
5 × 106
8 × 105
TEOS
DGS
no activity
20 ( 2
no activity
1.5
5 × 105 5 ( 3
s-1, ∼2-fold decrease in efficiency; DHFR: KM ) 29 µM, kcat
)
1.6 s-1 for NADPH, 9-fold decrease in efficiency). These results
confirm that the presence of residual glycerol within DGS-derived
materials, coupled with the inherently slow diffusion of materials
though such matrixes,2,9 likely contributed to the decrease in
catalytic efficiency of the enzymes upon entrapment as a result
of mass transport limitations, but at the same time glycerol
improved enzyme stability, thereby resulting in an overall
improvement in enzyme performance relative to TEOS-based
glasses.
(kcat/ KM) for each of the enzymes in solution and in DGS- and
TEOS-derived materials. In all cases, entrapment into a sol-gel-
derived glass reduced the kcat value of the enzymes, which can
be related to a reduced rate of delivery of substrate to the enzyme.
This is expected on the basis of the tortuous path that must be
taken to allow diffusion of small molecules through the porous
network of the silica,9 along with possible substrate-silica interac-
tions, which would also slow the movement of molecules through
the matrix.2 Previous studies have noted similar effects for a series
of enzymes entrapped in TEOS- or TMOS-derived glasses, where
kcat values for the entrapped enzymes, such as alcohol dehydro-
genase,50 bovine carbonic anhydrase II,51 glucose-6-phosphate
dehydrogenase,27 and oxalate oxidase,52 were in the range of
3-4000-fold lower than the corresponding values in solution. An
important point to note from the kinetic data is that the kcat values
of enzymes entrapped in DGS-derived materials were in all cases
within a factor of 7 of the value in solution, and in most cases,
were within a factor of 2 of the solution value (Factor Xa, COX-II,
γ-GT). The only enzyme that appears to be significantly affected
immediately after entrapment into DGS is DHFR, and even in this
case, the catalytic constants are within an order of magnitude of
those in solution. On the other hand, entrapment into TEOS-
derived glasses either led to complete protein denaturation (COX-
II, DHFR) or produced decreases in kcat values beyond those
obtained in DGS-derived materials, providing evidence for im-
proved enzyme performance in DGS-derived materials.
Another potential effect that could alter the observed KM values
is partitioning of the substrate between the solution and the silica
matrix. Silica is polar and anionic; thus, partitioning would be
expected to occur for charged analytes (increased partitioning of
cations, partial exclusion of anions). Partitioning will also depend
on ionic strength. In this work, we used >0.1 M ionic strength in
all assays to minimize partitioning effects (ionic strength was 0.55
(Factor Xa), 0.1 (DHFR), 0.12 (COX II), or ∼ 0.3 M (γ-GT)).
S-2222 and DPN are both zwitterionic and should show no
partitioning; thus, only mass transport should affect the turnover
of substrates by Factor Xa and γ-GT. Given that the kcat values
for these enzymes are not tremendously altered (less that a factor
of 2 in DGS vs solution), mass transport limitations should be
minimal, and the KM value should be relatively similar to solution,
as is observed for both enzymes.
Cox II used arachadonic acid (anionic), which should be
excluded from the matrix, leading to a somewhat higher apparent
KM value, since more substrate is needed in solution to provide
an adequate level of entrapped substrate for the reaction to
proceed (kcat and Vmax are not affected by partitioning). However,
partial charge screening likely plays a role in offsetting this effect;
thus, the KM value is not much higher than that obtained in
solution. The most dramatic changes in KM values were for DHFR,
which used NADPH (dianionic owing to the phosphate groups)
and dihydrofolic acid (anionic) as substrates. Exclusion of anions
The strength of substrate binding, as reflected by the KM value,
was also altered significantly upon entrapment. In general, the
KM values of entrapped enzymes increased relative to solution,
indicative of weaker binding of substrates to the enzyme. This is
expected, owing to the reduction in the rate of transport of the
substrate to the enzyme, and is consistent with the reduced kcat
(50) Williams, A. K.; Hupp, J. T. J. Am. Chem. Soc. 1 9 9 8 , 120, 4366.
(51) Badjic, J. D.; Kostic, N. M. Chem. Mater. 1 9 9 9 , 11, 3671.
(52) Yamanaka, S. A.; Nguyen, N. P.; Dunn, B.; Valentine, J. S.; Zink, J. I. J.
Sol-Gel Sci. Technol. 1 9 9 6 , 7, 117.
(53) Trevan, M. D. Immobilized Enzymes: An Introduction and Applications in
Biotechnology; John Wiley and Sons: New York, 1980.
2386 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

Figure 3. Changes in activity of the enzymes with storage time in DGS- (b) and TEOS-derived (9) materials for Factor Xa, DHFR, COX-II,
and γ-GT.
from the matrix leads to a lower than expected level of substrate
in the vicinity of the enzyme and an increase in KM, as was
observed for both DHF and NADPH. Thus, partitioning effects
likely contribute significantly to changes in KM values, suggesting
that higher ionic strength should be employed when possible to
offset this effect.
The activity of enzymes entrapped into DGS-derived materials
was in most cases far superior to the activity obtained in TEOS-
based materials. For example, both COX-II and DHFR, which were
inactive in TEOS-based materials, showed at least some activity
in DGS-derived glasses. COX-II was by far the least stable of the
enzymes tested, showing complete loss of activity in as little as 5
days, even in DGS-derived materials. However, this enzyme was
found to be inherently unstable, even in solution, showing a
complete loss of activity in as little as 1 day when stored at 4 °C
in solution; thus, the activity in DGS-derived materials is actually
Long-Term Stability of Entrapped Enzymes. The data
presented above provide insight into how enzyme performance
is affected at early times after entrapment. Another issue in the
use of entrapped enzymes is how their performance is altered
over time after entrapment. Figure 3 shows changes in the relative
activity of the enzymes with storage time in DGS- and TEOS-
derived materials. In both cases, the samples were stored in buffer
solution at 4 °C until tested. The activity was measured as the
initial rate of production of product, and was normalized to the
value obtained during the initial test of enzyme activity 3 days
after entrapment. No detectable leaching of any enzyme from
either of the sol-gel matrixes was observed during the time
course of the aging study; thus, all changes in activity are due to
alterations in the catalytic efficiency of the entrapped enzyme.
In general, all enzymes showed increased activity when
entrapped in DGS-derived glasses when compared to their activity
in TEOS-derived materials. Indeed, neither COX-II nor DHFR
showed any activity in TEOS-derived materials, whereas Factor
Xa lost all activity after only 1 week of aging. The only enzyme
that demonstrated significant activity in TEOS-derived materials
was γ-GT, and even then, the activity was in most cases a factor
of 2 lower than in DGS-derived materials. This may reflect the
fact that γ-GT is inherently membrane-associated and, thus, may
be less susceptible to denaturation by hydrophobic species. Hence,
the protein is more likely to maintain its active conformation, even
in the presence of ethanol.
significantly improved relative to solution. COX-II contains Fe3+
-
protoporphyrin IX as a cofactor, which can dissociate from the
protein during storage, leading to a mixture of apo- and holoen-
zymes.54 To obtain maximal activity, hematin is added to the
mixture to minimize the loss of the protoporphyrin group. It is
likely that the presence of ethanol (in TEOS-derived materials)
either promoted dissociation of the protoporphyrin, or deactivated
the hematin, leading to loss of function of the COX-II enzyme.
In the case of DHFR, the enzyme retained up to 60% of its
solution activity in DGS-derived materials for a period of 3 weeks,
at which point it underwent a catastrophic loss in activity. It is
not clear why the activity was lost so suddenly, but it is important
to note that the ability to maintain significant activity over a period
of 3 weeks is still likely to be sufficient for many applications,
such as the development of sol-gel-based microarrays. Finally,
in the case of entrapped Factor Xa, the enzyme remained stable
for a period of at least two months when entrapped in DGS-derived
materials, showing that such materials have the potential to serve
(54) Kurumbail, R. G.; Stevens, A. M.; Gierse, J. K.; McDonald, J. J.; Stegeman,
R. A.; Pak, J. Y.; Gildehaus, D.; Miyashiro, J. M.; Penning, T. D.; Siebert,
K.; Isakson, P. C.; Stallings, W. C. Nature 1 9 9 6 , 384, 644.
Analytical Chemistry, Vol. 75, No. 10, May 15, 2003 2387

Figure 4. Typical inhibition plots for Factor Xa (benzamidine inhibition), DHFR (trimethoprim inhibition), COX-II (V2-01 inhibition), and γ-GT
(acivicin inhibition) in solution DGS-derived glasses after 3 days of aging. The concentration of inhibitor used is shown on the plots.
as useful platforms for the development of immobilized enzyme
columns or microarrays.
of the protein.58,59 Glycerol may also act as a humectant, providing
greater retention of entrapped water, and it may also coat the walls
of the pores within the silica material, reducing direct contact of
the protein with the sol-gel matrix. It may, therefore, be
concluded that DGS is a more suitable precursor for sol-gel
matrixes than TEOS for the encapsulation of the enzymes.
Inhibition of Free and Entrapped Enzymes. Inhibition
studies were initiated for all enzymes in solution and DGS-derived
glasses. It was not possible to generate useful inhibition data for
either COX-II or DHFR within TEOS-derived glasses, owing to
the poor activity of these enzymes within such materials. However,
inhibition constants were obtained for Factor Xa and γ-GT in
TEOS-derived glasses. In the case of Factor Xa, the KI value for
inhibition by benzamidine was 2.1 mM, a factor of 10 higher than
in solution. In the case of γ-GT entrapped in TEOS-derived
materials, the KI value for inhibition of the enzyme by acivicin
was 450 µM, which is also >10-fold higher than the value obtained
in solution. These results indicate that entrapment of enzymes in
TEOS-derived materials is likely to be inadequate for the purposes
of drug screening. In both cases, it is likely that both changes to
the active site of the protein and interactions of the inhibitors with
the silica matrix may have led to the need for much higher levels
of inhibitor to alter the protein function.
An unexpected finding was that, with the exception of COX-
II, the activity of the entrapped enzyme increased during the first
1-2 weeks of storage, followed by a decrease of the enzyme
activity over a period of several days to weeks. The origin of the
increased activity at early storage times is not clear. All assays
were performed after washing to remove glycerol; thus, the level
of glycerol present during the assay is not likely to be a major
factor in determining the observed activity. The reasons for the
observed increases in activity are likely to be more complex and
may include alterations in surface polarity and charge, pore size,
protein accessibility, molecular crowding, and entrapped water
structure (i.e., protein hydration) as the sample aged.55,56 We are
currently investigating this phenomenon in more detail and will
report on our findings in a future manuscript.
Overall, the long-term stability studies showed that all enzymes
had improved activity in DGS-derived materials relative to TEOS-
derived materials. This is not surprising, since TEOS-derived
materials may contain up to 35% ethanol immediately after
gelation,57 which is likely to promote denaturation of entrapped
enzymes. On the other hand, DGS-derived materials evolve
glycerol, a known stabilizer of proteins, which may cause the
protein to adopt a more compact and rigid conformation as a result
of preferential hydration and, thus, decrease the specific volume
Figure 4 shows representative inhibition plots for each of the
four enzymes when entrapped in DGS-derived materials. Table 3
shows the structures of the inhibitors and the inhibition constants
generated for each enzyme in solution and in DGS-derived
(55) Eggers, D. K.; Valentine, J. S. Protein Sci. 2 0 0 1 , 10, 250. (b) Eggers, D.
K.; Valentine, J. S. J. Mol. Biol. 2 0 0 1 , 314, 911.
(56) Flora, K. K.; Brennan, J. D. Chem. Mater. 2 0 0 1 , 13, 4170.
(57) Flora, K. K.; Dabrowski, M. A.; Musson, S. P.; Brennan, J. D. Can. J. Chem.
1 9 9 9 , 77, 1617.
(58) Priev, A.; Almagor, A.; Yedgar, S.; Gavish, B. Biochemistry 1 9 9 6 , 25, 2061.
(59) Sousa, R. Acta Crystallogr. 1 9 9 5 , D51, 271.
2388 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

Table 3. KI Values for Inhibition Studies in Solution and in DGS-Derived Glasses
materials. Although many of the enzymes, such as Factor Xa and
DHFR, have proteinaceous inhibitors, all inhibitors in this study
were chosen to be small molecules, because these are representa-
tive of the types of molecules obtained from combinatorial libraries
and also because larger protein-based inhibitors would not be able
to enter into the silica matrix and interact with the entrapped
protein. As shown in Table 3, the majority of the inhibition
constants obtained in DGS-based glasses were within error of
those obtained in solution. The only exceptions were the COX-II
inhibitors V2-05 and V2-09, and even in these cases, the largest
difference in inhibition constants was a factor of 3. Overall, the
data show that reliable inhibition constants could be obtained over
a KI range spanning low nanomolar to high micromolar concentra-
tions, suggesting that entrapped enzymes should be useful in high-
throughput drug screening studies.
to a similar degree by mass transport; thus, the ratio should
remain constant regardless of mass transport effects. Thus, any
variations in KI values upon entrapment of an enzyme are likely
to be due to partitioning effects, which can change the equilibrium
concentration of the inhibitor in the vicinity of the enzyme. Such
effects have previously been demonstrated to occur for charged
analytes within sol-gel-derived silica,9 wherein cations were
preconcentrated in the matrix while anions were partially ex-
cluded. Since inhibitors compete with substrate molecules for
binding to the active site, partitioning of both species must be
considered to evaluate the overall affect on inhibition constants.
If the inhibitor and substrate are of the same charge or uncharged,
no differential partitioning occurs, and thus, no changes in
inhibition will be observed relative to solution. If the matrix is
anionic (as for the silica matrix), then inhibition will be greatest
when the inhibitor is cationic and the substrate is anionic;
inhibition will be least effective when using anionic inhibitors and
cationic substrates.
Previous studies of enzymes immobilized on charged surfaces
have shown that mass transport limitations tend not to alter KI
values, provided that sufficient time is allowed for the inhibitor
to reach equilibrium with the enzyme.53 Stated another way, KI is
related to the ratio KM′/ KM. Both KM′ and KM should be altered
However, substrate charge must also be considered. For Factor
Xa, the inhibitor benzamidine is cationic, and the substrate is
Analytical Chemistry, Vol. 75, No. 10, May 15, 2003 2389

zwitterionic. Thus, a partitioning of benzamidine into the matrix
and a concomitant lowering of the inhibition constant would be
expected, as is observed. For DHFR, the inhibitors pyrimethamine
(pKa ) 7.34) and trimethoprim (pKa ) 7.13) would be cationic at
pH 8.3 (the pH at which assays were performed), while the
substrates are anionic; thus, one might expect preferential
partitioning of the inhibitors into the silica, leading to lower
inhibition constants relative to solution. However, at the ionic
strength employed for these studies, the partitioning effect would
be minimized, leading to the reasonable agreement between the
inhibition constants in solution and silica. Folic acid (pKa 4.7)
would be anionic and, thus, would undergo a similar degree of
partitioning relative to dihydrofolate. Thus, the KI values for folic
acid should not be significantly affected by partitioning, as
observed. In the case of γ-GT, both the substrate (DPN) and the
inhibitor (acivicin60) are zwitterionic; thus, no partitioning would
be expected, and indeed, no changes are observed in the inhibition
constants upon entrapment in DGS-derived silica. A further
point to note in the case of γ-GT is that the enzyme is present on
the surface of a phospholipid liposome, and it is possible that the
lipid may act to increase the porosity of the matrix through a
templating effect and may also aid in passivating the silica matrix
through lipid/ silica interactions. Together, these effects would
likely lead to a more “solutionlike” environment for the enzyme
and produce a more reliable inhibition constant.
The COX-II enzyme provided the most challenging system for
inhibition studies (owing to the poor stability of this enzyme) and
also generated the largest differences in inhibition constants upon
entrapment of the enzyme. For Cox II, the V2 compounds are
cationic, and the substrate is anionic; thus, we may expect
increased inhibition, as is, indeed, observed for V2-05 and V2-09.
However, at the ionic strength used for these assays (0.12 M),
partial charge screening would be expected to occur, decreasing
the partitioning effect. This may explain why V2-01 and V2-06 show
inhibition constants for entrapped enzymes that are within error
of the values obtained in solution. Overall, the inhibition constants
for entrapped COX-II were all within a factor of 3 of the values in
solution (as was the case for all other enzymes tested); thus, it
appears that the use of enzymes entrapped in DGS-derived silica
may be amenable to high-throughput screening studies.
many cases, the inhibition constants for free and entrapped
proteins were within error of each other. It must be noted,
however, that such good agreement in KM and KI values occurred
only in DGS-derived materials; significant deviations in kcat, KM,
and KI values were obtained when TEOS-derived materials were
used to entrap the enzyme, and in many cases, the entrapped
protein did not remain active in these materials. This finding
shows the advantages of using biocompatible silane precursors
for entrapment of proteins.
Overall, the inhibition studies suggest that the encapsulation
of enzymes within DGS-derived materials may be useful for the
development of a HTS platform that is based on immobilized
enzymes. Of the four enzymes tested, all provided useful catalytic
and inhibition data, although the COX-II enzyme proved to be a
problematic system that was not particularly amenable to entrap-
ment, even in DGS-derived materials. However, this enzyme is
inherently unstable, even in solution and, thus, when it is used
for any type of HTS studies, it may require special precautions
be taken, such as short assay times and minimal exposure to room
temperature before use. The small differences in the inhibition
constants of the entrapped enzymes are not likely to be particularly
significant in high-throughput drug screening trials, where typi-
cally, changes of a factor of 10 or more are needed to separate a
“hit” from a nonlead. Indeed, all inhibitors that would be
considered candidates for further study (i.e., those with submi-
cromolar inhibition constants, such as trimethoprim and py-
rimethamine) were correctly identified, indicating that entrapment
did not have a deleterious effect in HTS studies. Given that both
mass transport and partitioning effects can alter KM and KI values,
it is clear that materials with greater porosity should be used to
maximize diffusion rates and that high ionic strength should be
employed to offset partitioning effects. Second generation materi-
als that have surfaces modified to be either neutral or zwitterionic
may also prove advantageous for HTS applications, since these
materials would eliminate electrostatic partitioning effects.
Other requirements for a robust HTS method include good
reproducibility, good stability of the enzymes, ability to distinguish
changes in KM or kcat upon inhibition, and ideally, the potential
for multiple use. As shown herein, the majority of enzymes tested
(except COX-II) are sufficiently stable to be used over a period of
weeks for HTS studies. Furthermore, the kinetic and inhibition
data obtained from the entrapped enzymes was highly reproduc-
ible, providing values that had precision similar to those obtained
in solution. At this stage, all inhibitors tested were competitive
inhibitors; thus, we have examined only changes in KM values
upon inhibition. Further studies on the effects of entrapment on
noncompetitive, allosteric and mixed inhibitors are needed to
better assess the capabilities of the entrapped enzymes for HTS
applications. Finally, multiple use capability, while not examined
herein, has been demonstrated for many sol-gel- entrapped
proteins,12 and thus, we expect that it should be possible to run
multiple assays on the same entrapped enzyme system with good
reproducibility.
Application of Entrapped Enzymes to HTS. An important
finding from this study was that a wide variety of different
substrates and inhibitors of varying size and charge were all able
to enter the DGS-derived glasses and interact with the active site
of the entrapped enzyme. This finding suggests that the enzymes
are essentially fully accessible to externally added reagents and
that the active sites are not significantly altered so as to greatly
affect the binding constants of either substrates or inhibitors.
Rather, it appears that variations in either KM or KI values between
solution and entrapped enzymes are likely to be based on mass
transport limitations (KM) or partitioning of substrates/ inhibitors
between the solution and glass (affecting both KM and KI). Specific
interactions of substrates or inhibitors with the silica matrix may
also play a role, although at the ionic strength employed, such
interactions should be minimized.9
CONCLUSIONS
The clinically relevant enzymes Factor Xa, DHFR, COX-II, and
In general, the inhibition constants obtained for entrapped
proteins were within a factor of 3 of the solution value, and in
γ-glutamyl transpeptidase were successfully entrapped in DGS-
derived materials. In general, the kcat values for all enzymes
decreased by a factor of 2-7, while the KM values of the enzymes
(60) Allen, L. M.; Maria, V. C.; Meinking, T. Chem-Biol. Interact. 1 9 8 1 , 33, 361.
2390 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

increased upon entrapment into DGS-derived materials, leading
to an overall lowering of the catalytic efficiency of the entrapped
enzyme by 2-10-fold. These changes are predominantly due to
slower diffusion and partitioning of the substrate within the
microporous material, suggesting that increases in porosity may
be one route to increase the catalytic efficiency of entrapped
enzymes. Inhibition studies indicated that the interaction of the
inhibitor with the entrapped enzyme was remarkably similar to
that in solution, with all but one KI value being within a factor of
2 relative to the value in solution. The minor changes in inhibition
constants reflect interactions of the inhibitors with the silica
surface that likely alter the concentration of free inhibitor in the
sol-gel matrix. However, in no case was the inhibition constant
for the entrapped enzyme more than a factor of 3 different from
the value obtained in solution, indicating that immobilized
enzymes are likely to be adequate for high-throughput drug
screening. Efforts to both understand and control the partitioning
of small molecules between solution and the matrix are ongoing,
and the results of these studies will be presented in due course.
ACKNOWLEDGMENT
The authors thank the Natural Sciences and Engineering
Research Council of Canada, the Ontario Research and Develop-
ment Challenge Fund, and MDS-Sciex for financial support of this
work. The authors also thank Dr. Eric Brown for providing DHFR,
Merck-Frosst Canada for providing COX-II, and Dr. Michael
Organ for providing the COX-II inhibitors. J.D.B. holds the Canada
Research Chair in Bioanalytical Chemistry.
Received for review November 29, 2002. Accepted March
6, 2003.
AC026370I
Analytical Chemistry, Vol. 75, No. 10, May 15, 2003 2391