Photopatterning enzymes on polymer monoliths in microfluidic devices for steady-state kinetic analysis and spatially separated multi-enzyme reactions

Article abstract of DOI:10.1021/ac070705k

A method for photopatterning multiple enzymes on porous polymer monoliths within microfluidic devices has been developed and used to perform spatially separated multienzymatic reactions. To reduce nonspecific adsorption of enzymes on the monolith, its pore surface was modified by grafting poly(ethylene glycol), followed by surface photoactivation and enzyme immobilization in the presence of a nonionic surfactant. Characterization of bound horseradish peroxidase (HRP) was carried out using a reaction in which the steady-state profiles of the fluorescent reaction product could be measured in situ and then analyzed using a plug-flow bioreactor model to determine the observed maximum reaction rate and Michaelis constant. The Michaelis constant of 1.9 μmol/L agrees with previously published values. Mass-transfer limitations were evident at relatively low flow rates but were absent at higher flow rates. Sequential multienzymatic reactions were demonstrated using the patternwise assembly of two- and three-enzyme systems. Glucose oxidase (GOX) and HRP were patterned in separate regions of a single channel, and product formation was analyzed as a function of flow direction. Significant product formation occurred only in the GOX to HRP direction. A three-enzyme sequential reaction was performed using invertase, GOX, and HRP. All possible arrangements of the three enzymes were tested, but significant product formation was only observed when the enzymes were in the correct sequential order. Photopatterning enzymes on polymer monoliths provides a simple technique for preparing spatially localized multiple-enzyme microreactors capable of directional synthesis.

Full text of DOI:10.1021/ac070705k

Anal. Chem. 2007, 79, 6592-6598
Photopatterning Enzymes on Polymer Monoliths in
Microfluidic Devices for Steady-State Kinetic
Analysis and Spatially Separated Multi-Enzyme
Reactions
Timothy C. Logan and Douglas S. Clark*
Department of Chemical Engineering, University of California, Berkeley, California 94720
Timothy B. Stachowiak, Frantisek Svec, and Jean M. J. Fre´ chet*
College of Chemistry, University of California, Berkeley, California 94720-1460, and The Molecular Foundry,
Lawrence Berkeley National Laboratory, Berkeley, California 94720
analysis,4-7 measurement of enzyme kinetics,8-12 biochemical
analyte sensing,13-16 and, more recently, combinatorial biosyn-
thesis.17,18 The use of enzymes in microfluidic devices is motivated
by the significant reduction in the amount of reactant and catalyst
required compared to bench-scale protocols and is often per-
formed with the goal of integration within a larger lab-on-a-chip
device. The utility of enzymatic reactions in microscale devices
has been espoused in several recent reviews.19-23 The use of
immobilized enzymes is of particular interest because immobiliza-
tion facilitates reuse of enzymes and simplifies product separation,
A method for photopatterning multiple enzymes on porous
polymer monoliths within microfluidic devices has been
developed and used to perform spatially separated mul-
tienzymatic reactions. To reduce nonspecific adsorption
of enzymes on the monolith, its pore surface was modified
by grafting poly(ethylene glycol), followed by surface
photoactivation and enzyme immobilization in the pres-
ence of a nonionic surfactant. Characterization of bound
horseradish peroxidase (HRP) was carried out using a
reaction in which the steady-state profiles of the fluores-
cent reaction product could be measured in situ and then
analyzed using a plug-flow bioreactor model to determine
the observed maximum reaction rate and Michaelis
constant. The Michaelis constant of 1.9 µmol/L agrees
with previously published values. Mass-transfer limita-
tions were evident at relatively low flow rates but were
absent at higher flow rates. Sequential multienzymatic
reactions were demonstrated using the patternwise as-
sembly of two- and three-enzyme systems. Glucose oxi-
dase (GOX) and HRP were patterned in separate regions
of a single channel, and product formation was analyzed
as a function of flow direction. Significant product forma-
tion occurred only in the GOX to HRP direction. A three-
enzyme sequential reaction was performed using inver-
tase, GOX, and HRP. All possible arrangements of the
three enzymes were tested, but significant product forma-
tion was only observed when the enzymes were in the
correct sequential order. Photopatterning enzymes on
polymer monoliths provides a simple technique for pre-
paring spatially localized multiple-enzyme microreactors
capable of directional synthesis.
(2) Andersson, H.; van der Wijngaart, W.; Stemme, G. Electrophoresis 2001,
22, 249-57.
(3) Lagally, E. T.; Emrich, C. A.; Mathies, R. A. Lab Chip 2001, 1, 102-7.
(4) Figeys, D.; Pinto, D. Electrophoresis 2001, 22, 208-16.
(5) Bousse, L.; Mouradian, S.; Minalla, A.; Yee, H.; Williams, K.; Dubrow, R.
Anal. Chem. 2001, 73, 1207-12.
(6) Peterson, D. S.; Rohr, T.; Svec, F.; Fre´chet, J. M. J. Anal. Chem. 2002, 74,
4081-8.
(7) Ekstrom, S.; Onnerfjord, P.; Nilsson, J.; Bengtsson, M.; Laurell, T.; Marko-
Varga, G. Anal. Chem. 2000, 72, 286-93.
(8) Hadd, A. G.; Raymond, D. E.; Halliwell, J. W.; Jacobson, S. C.; Ramsey, J.
M. Anal. Chem. 1997, 69, 3407-12.
(9) Gleason, N. J.; Carbeck, J. D. Langmuir 2004, 20, 6374-81.
(10) Seong, G. H.; Heo, J.; Crooks, R. M. Anal. Chem. 2003, 75, 3161-7.
(11) Kerby, M. B.; Legge, R. S.; Tripathi, A. Anal. Chem. 2006, 78, 8273-80.
(12) ) Zhou, X. Z.; Medhekar, R.; Toney, M. D. Anal. Chem. 2003, 75, 3681-7.
(13) Murakami, Y.; Takeuchi, T.; Yokoyama, K.; Tamiya, E.; Karube, I.; Suda,
M. Anal. Chem. 1993, 65, 2731-5.
(14) Dodge, A.; Fluri, K.; Verpoorte, E.; de Rooij, N. F. Anal. Chem. 2001, 73,
3400-9.
(15) Banu, S.; Greenway, G. M.; McCreedy, T.; Shaddick, R. Anal. Chim. Acta
2003, 486, 149-57.
(16) Richter, T.; Shultz-Lockyear, L. L.; Oleschuk, R. D.; Bilitewski, U.; Harrison,
D. J. Sens. Actuators, B 2002, 81, 369-76.
(17) Srinivasan, A.; Bach, H.; Sherman, D. H.; Dordick, J. S. Biotechnol. Bioeng.
2004, 88, 528-35.
(18) Ku, B. S.; Cha, J. H.; Srinivasan, A.; Kwon, S. J.; Jeong, J. C.; Sherman, D.
H.; Dordick, J. S. Biotechnol. Prog. 2006, 22, 1102-7.
(19) Urban, P. L.; Goodall, D. M.; Bruce, N. C. Biotechnol. Adv. 2006, 24, 42-
57.
(20) Weigl, B. H.; Bardell, R. L.; Cabrera, C. R. Adv. Drug Delivery Rev. 2003,
55, 349-77.
Enzymes have been used in microfluidic systems to enable a
variety of biochemical applications including DNA1-3 and protein
* Corresponding authors. E-mail: clark@berkeley.edu. Phone: (510) 642-
2408. or E-mail: frechet@berkeley.edu. Phone: (510) 643-3077.
(21) Delamarche, E.; Juncker, D.; Schmid, H. Adv. Mater. 2005, 17, 2911-33.
(22) Wang, J. Electrophoresis 2002, 23, 713-8.
(1) Kopp, M. U.; de Mello, A. J.; Manz, A. Science 1998, 280, 1046-48.
(23) Miyazaki, M.; Maeda, H. Trends Biotechnol. 2006, 24, 463-70.
6592 Analytical Chemistry, Vol. 79, No. 17, September 1, 2007
10.1021/ac070705k CCC: $37.00 © 2007 American Chemical Society
Published on Web 07/21/2007

as well as improves enzymatic reaction rates.24 A final advantage
of immobilization arises from the ability to localize the enzyme in
a microfluidic environment whereby the direction and rate of fluid
flow imparts the ability to control the sequence and residence
time of multienzymatic reactions. Such a configuration also allows
sequential reactions to occur within well-defined spatial reaction
zones. Therefore, to realize all of these benefits, we sought to
develop a method for patterning multiple enzymes within microf-
luidic reactors.
The complexity of confining multiple enzymes to distinct
locations within a microfluidic system has thus far limited the
development of spatially separated multienzymatic reactors. The
first demonstration of a spatially separated enzymatic reaction in
a microfluidic format was performed by Cremer and co-workers,
where two enzymes, glucose oxidase (GOX) and horseradish
peroxidase (HRP), were covalently immobilized to the surface of
two separate PDMS channels. The two channels were connected
externally using plastic tubing in order to carry out a sequential
enzymatic reaction;25 a similar approach was later used by Ku et
al.18 Crooks and co-workers immobilized the same two enzymes
to the surface of microbeads that were retained in separate
sections of a single microfluidic channel by microfabricated
weirs.26 While this method is extremely flexible, it does not scale
well for multiple enzymes because each additional enzyme
requires the fabrication of an auxiliary channel and weir to pack
and retain the microbeads in the channel.
An attractive and readily scalable alternative involves the use
of photochemistry to pattern enzymes in microfluidic channels.
Photochemical approaches allow enzymes to be patterned in fully
fabricated microscale devices via a process similar to photolithog-
raphy. This process has been previously demonstrated for co-
valently immobilizing proteins or enzymes to the surfaces of empty
channels.27-29 Because the entire microfluidic channel is exposed
to a solution of enzyme during this process, nonspecific adsorption
of enzyme outside of the desired immobilization region can
prevent effective patterning of the enzyme. These groups limited
nonspecific adsorption of enzyme through the use of blocking
protein. Another limitation of this technique is that the amount
of immobilized enzyme is limited by the available surface area of
the channel. Several groups have attempted to increase the
amount of immobilized enzyme by entrapping enzyme within
photopatterned hydrogels.30-33 By this method, enzyme is encap-
sulated within a gel rather than attached covalently. The gels must
be dense enough to retain protein, thus rendering the enzyme
inaccessible to convective flow. Consequently, although more
enzyme is immobilized, not all of the immobilized enzyme is
necessarily available for reaction. Furthermore, the use of dense
gels can significantly increase the flow resistance in a microchan-
nel.
The present study develops a method of photopatterning
multiple enzymes in microfluidic devices at high densities in order
to enable multistep enzymatic reactions whose reaction sequence
is determined by the direction of fluid flow. The immobilization
method employed in this work was the covalent attachment of
enzyme to the surface of a porous polymer monolith. Porous
polymer monoliths have been developed over the past decade as
support materials for a variety of analytical applications.34 Polymer
monoliths increase the available surface area in a microchannel
and are thus particularly useful for high-density enzyme immobil-
ization.6,35-39 In addition, porous polymers are easy to fabricate
in-device via a UV-initiated polymerization process whereby liquid
monomers polymerize in the presence of porogenic solvents to
form a porous solid support that can be patterned using a
photomask.40,41 Furthermore, because they are covalently an-
chored to the microchannel walls, polymer monoliths do not
require any retaining structures such as the microfabricated weirs
that are needed to retain microbeads. Recently, we introduced a
photografting method for the in situ modification of these porous
polymers.42,43 This surface modification process is also photoini-
tiated, allowing the polymer surface to be photopatterned with
specific chemical functionalities such as azlactone for the covalent
attachment of a variety of proteins.38,42 In particular, we have used
this technique to fabricate a simple monolithic device with dual
functionality that combined solid-phase extraction and tryptic
digestion.38 However, this early approach enabled only limited
control over the nonspecific adsorption of enzyme throughout the
device. Recently, we extended our photografting technique to poly-
(ethylene glycol) methacrylates in order to obtain protein adsorp-
tion-resistant coatings.44
In this work, we demonstrate the preparation of monolithic
porous polymer supports that suppress the nonspecific adsorption
of enzymes and enable the photopatterning of spatially separated
multienzymatic microreactors. Kinetic parameters for an im-
mobilized enzyme reaction are determined by measuring product
formation on device and analyzing the results using a plug-flow
reactor model. Finally, we demonstrate model sequential reactions
involving two and three enzymes, and we present an example of
directional biosynthesis.
EXPERIMENTAL SECTION
Materials. Butyl methacrylate (99%, BuMA), ethylene dimeth-
acrylate (98%, EDMA), 1-dodecanol (98%), cyclohexanol (99%),
(34) Svec, F., Tennikova, T. B., Deyl, Z., Eds. Monolithic materials: preparation,
properties, and applications; Elsevier: Amsterdam, 2003.
(35) Petro, M.; Svec, F.; Fre´chet, J. M. J. Biotechnol. Bioeng. 1996, 49, 355-63.
(36) Xie, S. F.; Svec, F.; Fre´chet, J. M. J. Biotechnol. Bioeng. 1999, 62, 30-5.
(37) Peterson, D. S.; Rohr, T.; Svec, F.; Fre´chet, J. M. J. J. Proteome Res. 2002,
1, 563-8.
(24) Blanch, H. W.; Clark, D. S. Biochemical Engineering; Marcel Dekker: New
York, 1996.
(25) Mao, H. B.; Yang, T. L.; Cremer, P. S. Anal. Chem. 2002, 74, 379-85.
(26) Seong, G. H.; Crooks, R. M. J. Am. Chem. Soc. 2002, 124, 13360-1.
(27) Balakirev, M. Y.; Porte, S.; Vernaz-Gris, M.; Berger, M.; Arie, J. P.; Fouque,
B.; Chatelain, F. Anal. Chem. 2005, 77, 5474-9.
(38) Peterson, D. S.; Rohr, T.; Svec, F.; Fre´chet, J. M. J. Anal. Chem. 2003, 75,
5328-35.
(39) Svec, F. Electrophoresis 2006, 27, 947-61.
(40) Rohr, T.; Yu, C.; Davey, M. H.; Svec, F.; Fre´chet, J. M. J. Electrophoresis
(28) Holden, M. A.; Jung, S. Y.; Cremer, P. S. Anal. Chem. 2004, 76, 1838-43.
(29) Nakajima, H.; Ishino, S.; Masuda, H.; Nakagama, T.; Shimosaka, T.;
Uchiyama, K. Anal. Chim. Acta 2006, 562, 103-9.
2001, 22, 3959-67.
(41) Throckmorton, D. J.; Shepodd, T. J.; Singh, A. K. Anal. Chem. 2002, 74,
784-9.
(30) Zhan, W.; Seong, G. H.; Crooks, R. M. Anal. Chem. 2002, 74, 4647-52.
(31) Heo, J.; Crooks, R. M. Anal. Chem. 2005, 77, 6843-51.
(32) Koh, W. G.; Pishko, M. Sens. Actuators, B 2005, 106, 335-42.
(33) Zimmermann, S.; Fienbork, D.; Flounders, A. W.; Liepmann, D. Sens.
Actuators, B 2004, 99, 163-73.
(42) Rohr, T.; Hilder, E. F.; Donovan, J. J.; Svec, F.; Fre´chet, J. M. J.
Macromolecules 2003, 36, 1677-84.
(43) Hilder, E. F.; Svec, F.; Fre´chet, J. M. J. Anal. Chem. 2004, 76, 3887-92.
(44) Stachowiak, T. B.; Svec, F.; Fre´chet, J. M. J. Chem. Mater. 2006, 18, 5950-
7.
Analytical Chemistry, Vol. 79, No. 17, September 1, 2007 6593

benzophenone (99%, BP), 3-(trimethoxysilyl)propyl methacrylate,
ethanolamine (98%), and 2,2-dimethoxy-2-phenylacetophenone
(99%) were purchased from Sigma-Aldrich Co. (St. Louis, MO).
BuMA and EDMA were purified by distillation under reduced
pressure to remove inhibitors. Ethoxylated hydroxyethyl meth-
acrylate (PEGMA, average Mn ) 570 g/mol, containing 10-11
ethylene glycol units) was purchased from Sartomer (Exton, PA).
PEGMA was purified by passing it through a column containing
inhibitor remover beads of basic alumina (Aldrich). 2-Vinyl-4,4-
dimethylazlactone (vinyl azlactone) was a gift from the 3M Co.
(St. Paul, MN). UV-transparent fused-silica capillary (100-µm i.d.)
was purchased from Polymicro Technologies (Phoenix, AZ).
Fused-silica capillaries served as model microfluidic channels in
this study.
Glucose oxidase (EC 1.1.3.4, ∼200 units/mg from Aspergillus
niger), horseradish peroxidase (EC 1.11.1.7, 250-330 units/mg
from horseradish), and invertase (INV, EC 3.2.1.26, >300 units/
mg from Saccharomyces cerevisiae) were purchased from Sigma-
Aldrich Co. (St. Louis, MO). Recombinant green fluorescent
protein (GFP) was produced by the method described by White-
head et al.45 Sodium phosphate monobasic, sodium chloride, and
sodium hydroxide were purchased from Fisher Scientific (Hamp-
ton, NH). Resorufin dye, 3 wt % hydrogen peroxide, and dimethyl
sulfoxide (DMSO) were purchased from Sigma-Aldrich. The
concentration of hydrogen peroxide was regularly assessed with
a hydrogen peroxide kit from Palintest (Erlanger, KY). Amplex
Red (10-acetyl-3,7-dihydroxyphenoxazine) was purchased from
Invitrogen Corp. (Carlsbad, CA). A stock solution of 10 mmol/L
Amplex red was prepared by dissolving 5 mg of Amplex Red in
anhydrous DMSO; the resultant solution was stored protected
from light at -20 °C.
Instrumentation. An OAI model 30 deep UV collimated light
source (San Jose, CA) fitted with a 500-W HgXe lamp was used
for UV exposures. The irradiation power was calibrated to 15.0
mW/cm2 using an OAI model 306 UV power meter with a 260-
nm probe head. Microscopic evaluations were performed using a
Nikon TE200 inverted fluorescence microscope (Scientific Instru-
ment Co., Sunnyvale, CA). Images were acquired with a Micro-
publisher 5.0 RTV CCD camera from QImaging (Burnaby, BC,
Canada). Green fluorescence was observed with Nikon filter B-1A;
red fluorescence was observed with Nikon filter TRITC HYQ.
Image analysis and quantification was performed with image
analysis program ImageJ, provided by the NIH. Parameter fitting
was performed by linear multiple regression performed using
scientific graphing and data analysis software, Origin (OriginLab
Corp., Northampton, MA). Off-chip fluorescent assays were
analyzed with a SpectraMax M2 fluorescent microplate reader
(Molecular Devices Corp., Sunnyvale CA).
acid for 30 min at a flow rate of 0.5 µL/min, rinsed with water,
and finally flushed with ethanol. A 20 wt % solution of 3-(tri-
methoxysilyl)propyl methacrylate in 95% ethanol with pH adjusted
to 5 using acetic acid was pumped through the capillary at a flow
rate of 0.5 µL/min for 60 min. The capillary were then washed
with acetone, dried in a stream of nitrogen, and left at room
temperature for 24 h.
Monolith Formation. The polymerization mixture contained
the monomers BuMA and EDMA and porogens 1-decanol and
cyclohexanol. The polymerization mixture consisted of 24 wt %
BuMA, 16 wt % EDMA, 50 wt % 1-decanol, 10 wt % cyclohexanol,
and 1 wt % DMPA (with respect to monomers) and was purged
with nitrogen for 10 min prior to use. A 10-cm-long capillary that
had been previously activated was filled with this mixture, placed
under the light source, and irradiated with UV light for 10 min at
a distance of 30 cm from the UV source. After photopolymeriza-
tion, the porogenic solvents were flushed from the polymer
monolith by pumping methanol through the column at a flow rate
of 0.5 µL/min for 8 h. The average pore size of the monolith was
∼1.6 µm, as measured by mercury intrusion porosimetry.42
Photografting of PEG. Poly(ethylene glycol) (PEG) was
photografted onto the surface of the methacrylate polymer
monolith to render it hydrophilic and prevent the nonspecific
adsorption of protein to the monolith surface.44 Briefly, BP was
initially grafted to the surface of the monolith by rinsing the
monolith with a 5 wt % BP solution in methanol at a flow rate of
0.5 µL/min for 30 min and then exposing the monolith to UV
irradiation for 2 min. After photografting BP, the monolith was
rinsed with methanol at a flow rate of 0.5 µL/min for 20 min to
remove unbound initiator. Next, a 0.10 mol/L solution of PEGMA
monomer in water was pumped at a flow rate of 0.5 µL/min
through the monolith for 30 min. The monolith was then exposed
to UV irradiation for 2 min. The monolith was then rinsed with
water at a flow rate of 0.5 µL/min for 7 h to remove any unreacted
hydrophilic monomers.
Photografting of Vinyl Azlactone. To enable the covalent
immobilization of proteins, the surface of the polymer monolith
was first activated by photografting with the reactive monomer
2-vinyl-4,4-dimethylazlactone, using a procedure modified from
previous reports.36,44 The photografting solution consisted of 15
wt % vinyl azlactone and 0.22 wt % BP dissolved in a solution of
75:25 wt % tert-butyl alcohol/water. Vinyl azlactone was photo-
grafted onto the surface of the monoliths using a 5-min exposure
through a photomask; the mask defined the area where protein
was to be immobilized. After photografting vinyl azlactone, the
monolith was rinsed with acetone for 1 h at a flow rate of 0.5 µL/
min to remove excess reagents.
Protein Immobilization. Protein was covalently immobilized
onto the surface of a monolith already photografted with vinyl
azlactone by pumping a protein solution through the monolith at
a flow rate of 0.5 µL/min for 1-1.5 h. The azlactone functionality
reacts with the amine groups of proteins, primarily the ꢀ-amino
group of lysine, forming a covalent bond between the protein and
the support surface. The monolith was then rinsed with 1.0 mol/L
ethanolamine at 0.5 µL/min for 1 h to quench unreacted azlactone
functionalities. Finally, the capillary was rinsed with 50 mmol/L
phosphate buffer, pH 7.50, at 1.0 µL/min for 1 h to remove excess
reagents.
Capillary Surface Activation. The inner surface of fused-silica
capillaries was first activated with 3-(trimethoxysilyl)propyl meth-
acrylate; this process enables the polymer monolith to covalently
attach to the capillary surface through the resulting pendent vinyl
groups. The inner capillary surface was first quickly rinsed with
acetone and water, then flushed with 0.2 mol/L sodium hydroxide
for 30 min at a flow rate of 0.5 µL/min, and quickly rinsed with
water. Next, the capillary was flushed with 0.2 mol/L hydrochloric
(45) Whitehead, T. A.; Boonyaratanakornkit, B. B.; Hollrigl, V.; Clark, D. S. Protein
Sci. In press.
6594 Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

Two-Enzyme Assay. HRP was initially patterned on 1 cm of
a 9-cm-long monolith via the previously described vinyl azlactone
photografting and enzyme immobilization process. GOX was then
patterned on the other 8 cm of the monolith, by performing a
second round of patterned vinyl azlactone photografting, followed
by GOX immobilization. The GOX immobilization solution con-
sisted of 10 mg/mL GOX and 0.01% (v/v) Tween-20 in 50 mmol/L
phosphate buffer, pH 7.50. After each enzyme was immobilized,
the monolith was rinsed with 1.0 mol/L ethanolamine at 0.5 µL/
min for 1 h to quench any unreacted azlactone functionalities.
Finally, the monolith was rinsed with phosphate buffer at 1.0 µL/
min for 1 h. A 1.0 mg/mL dextrose solution was created by
dissolving crystalline dextrose in phosphate buffer. This solution
was allowed a minimum of 12 h to establish equilibrium between
the isomers of glucose. Immediately prior to use, the dextrose
solution was saturated with pure oxygen for a minimum of 15 min.
To 1.00 mL of this dextrose solution, 10 µL of the 10 mmol/L
stock Amplex Red solution was added such that the final
concentration was 100 µmol/L Amplex Red and 1.0% (v/v) DMSO.
A 50-µL sample was collected and the fluorescence was analyzed
as previously described. Between trials the monolith was rinsed
with 50 µL of phosphate buffer at 1.0 µL/min to remove excess
reagents.
Three-Enzyme Assay. Similar to the two-enzyme procedure,
a 9-cm-long monolith was initially patterned with 1 cm of HRP.
GOX was then patterned on 4 cm, and finally INV was patterned
to the remaining 4 cm of the column, using a third round of
photopatterned vinyl azlactone grafting, followed by INV im-
mobilization. The INV immobilization solution consisted of 10 mg/
mL INV and 0.01% (v/v) Tween-20 in 50 mmol/L phosphate buffer,
pH 7.50. After each enzyme was immobilized, the monolith was
rinsed with 1.0 mol/L ethanolamine at 0.5 µL/min for 1 h to
quench unreacted azlactone functionalities. Finally, the monolith
was rinsed with phosphate buffer at 1.0 µL/min for 1 h. A 10.0
mg/mL sucrose solution was prepared by dissolving crystalline
sucrose in phosphate buffer. Immediately prior to use, the sucrose
solution was saturated with pure oxygen for a minimum of 15 min.
To 1.00 mL of this sucrose solution, 10 µL of the 10 mmol/L stock
Amplex Red solution was added such that the final concentration
was 100 µmol/L Amplex Red and 1.0% (v/v) DMSO. A 50-µL
sample was collected at a flow rate of 0.10 µL/min, and the
fluorescence was analyzed as previously described. Between trials,
the monolith was rinsed with 50 µL of phosphate buffer at 1.0
µL/min to remove excess reagents.
Figure 1. Schematic diagram of the photopatterning process. (A)
Protein is immobilized to the surface of a polymer monolith in
patterned regions within a microfluidic channel. (B) PEG is grafted to
the surface of the polymer monolith to prevent nonspecific protein
adsorption. Vinyl azlactone is photopatterned onto the PEG surface
and activates the surface for protein immobilization. (C) Azlactone
functionality reacts with amines of proteins to form a covalent amide
bond between the protein and the polymer monolith surface.
a 1:1 stoichiometry.46 Monitoring any fluorescence produced
within the microchannels provides a sensitive means of detecting
the presence of active enzyme.
We employed azlactone attachment chemistry for covalent
immobilization of enzyme to the polymer monolith supports,
whereby reactive azlactone groups are introduced onto the
monolith surface by photografting the surface with vinyl azlac-
tone.38,42 The resulting azlactone functionalities react with amines
on proteins, as illustrated in Figure 1; the coupling reaction is
relatively rapid and not overly sensitive to hydrolysis.48 No leaving
groups are produced during the immobilization reaction since
azlactone reacts via a ring-opening, nucleophilic addition. In
addition, the azlactone attachment chemistry does not require any
modifications of the enzyme prior to immobilization such as those
required with the frequently used streptavidin-biotin bioconju-
gation techniques.25,28 An important benefit of photografting is that
it is possible to pattern the location of grafted azlactone and,
subsequently, the location of immobilized protein (Figure 1).
Furthermore, multiple enzymes can be immobilized in different
locations within a single device by repeating the azlactone
photografting and subsequent enzyme immobilization process for
each enzyme.
RESULTS AND DISCUSSION
Minimizing Nonspecific Protein Adsorption. Successful
patterning of immobilized enzymes requires the suppression of
nonspecific protein adsorption, which can result in the immobiliza-
tion of active enzyme outside of a patterned region. Minimizing
nonspecific protein adsorption is particularly important with high-
surface-area support materials such as porous polymer monoliths.
To this end, horseradish peroxidase was employed to investigate
the nonspecific adsorption of protein onto the surface of the
polymer monoliths and to evaluate the effectiveness of attempts
to pattern enzyme immobilization. The activity of HRP was
detected using the fluorogenic substrate, Amplex Red. HRP
catalyzes the oxidation of nonfluorescent Amplex Red to fluores-
cent resorufin, employing hydrogen peroxide as a cosubstrate in
In order to determine appropriate immobilization conditions,
pairs of identical polymer monolith columns were prepared in
(46) Zhou, M. J.; Diwu, Z. J.; PanchukVoloshina, N.; Haugland, R. P. Anal.
Biochem. 1997, 253, 162-8.
(47) Towne, V.; Will, M.; Oswald, B.; Zhao, Q. J. Anal. Biochem. 2004, 334,
290-6.
(48) Heilmann, S. M.; Drtina, G. J.; Haddad, L. C.; Rasmussen, J. K.; Gaddam,
B. N.; Liu, J. J.; Fitzsimons, R. T.; Fansler, D. D.; Vyvyan, J. R.; Yang, Y. N.;
Beauchamp, T. J. J. Mol. Catal. B: Enzym. 2004, 30, 33-42.
Analytical Chemistry, Vol. 79, No. 17, September 1, 2007 6595

Table 1. Effect of Surface Treatment on the Concentration of Resorufin Produced from Immobilized HRP on
Patterned and Nonpatterned Monolithic Columnsa
patterned monolith
nonpatterned monolith
ratio of [resorufin]
monolith type
[resorufin] (µmol/L)
[resorufin] (µmol/L)
patterned/nonpatterned
initial monolith
PEG-grafted monolith
7.8 ( 1.3
6.4 ( 1.1
7.6 ( 1.5
0.88 ( 0.02
1.0
7.3
a The reported error is the standard deviation of three separate trials. Conditions: patterned and nonpatterned monolith columns, 4.5 cm in
length; HRP was immobilized to 1.0 cm of the patterned monolith; enzyme immobilization solution, 1.0 mg/mL HRP in 50 mmol/L phosphate
buffer, pH 7.50; substrate solution, 8.80 µmol/L hydrogen peroxide, 100 µmol/L Amplex Red, and 0.10% (v/v) DMSO in 50 mmol/L phosphate
buffer, pH 7.50; substrate flow rate, 1.0 µL/min; 50 µL of sample was collected and analyzed.
Table 2. Effect of HRP Concentration in Immobilization
which one column contained a photopatterned region of grafted
vinyl azlactone for enzyme immobilization, as in Figure 1A and
Solutiona
patterned
monolith
nonpatterned
monolith
ratio of
B, while the other column was not patterned with azlactone. Then,
both columns were treated with the enzyme immobilization
solution. Thus, enzyme was covalently immobilized to the pho-
topatterned region of one column, whereas the only enzyme
present in the other column was the result of nonspecific
adsorption. With unmodified polymer monolith supports, both the
patterned and nonpatterned columns exhibited a high level of
enzyme activity (Table 1), indicating that nonspecific adsorption
onto the polymer monolith was significant and that much of the
adsorbed enzyme remained active.
Several attempts were made to disrupt the surface interactions
that resulted in the high degree of nonspecific HRP adsorption.
Following enzyme immobilization, the monoliths were rinsed with
a high ionic strength solution (1.0 mol/L NaCl) to disrupt ionic
interactions possibly responsible for nonspecifically adsorbed
protein and rinsed with solutions of pH 6.0 and pH 9.0 to vary the
net charge of the HRP. Neither of these steps significantly
decreased the observed activity of the nonpatterned monoliths,
indicating that nonspecific adsorption of HRP was not due
primarily to ionic interactions.
[resorufin]
patterned/
nonpatterned
[HRP]
[resorufin]
(µmol/L)
[resorufin]
(µmol/L)
(mg/mL)
1.0
0.10
0.010
6.4 ( 1.1
3.6 ( 0.62
0.54 ( 0.04
0.88 ( 0.02
0.32 ( 0.06
0.28 ( 0.06
7.3
11
1.9
a For conditions, see Table 1.
Table 3. Effect of Surfactant Concentration in
Immobilization Solutiona
patterned
monolith
nonpatterned
monolith
ratio of
[resorufin]
patterned/
nonpatterned
[Tween-20]
(%) (v/v)
[resorufin]
(µmol/L)
[resorufin]
(µmol/L)
0.10
0.010
0.62 ( 0.02
5.6 ( 0.12
0.50 ( 0.02
0.32 ( 0.04
1.2
18
a The concentration of HRP in the immobilization solution was 0.10
mg/mL. For other conditions, see Table 1.
Next, to examine the impact of nonionic interactions, specifi-
cally hydrophobic interactions, PEG was grafted to the surface of
the monolith in the first step followed by photografting vinyl
azlactone on the top of the PEG layer. We have demonstrated
the utility of this two-layer approach previously.43 PEG grafting
imparts a hydrophilic character to the surface of the monolith and
has previously been shown to reduce the nonspecific adsorption
of protein on polymer monoliths.44 The PEG surface treatment
reduced the amount of resorufin produced from both columns,
but a much larger reduction was observed for the nonpatterned
column. The benefit of PEG grafting on reducing protein adsorp-
tion is clearly reflected in the ratios of the amount of resorufin
produced from the patterned columns over that produced by the
nonpatterned columns. The value for unmodified monoliths was
1.0, while the PEG-grafted columns exhibited a ratio of 7.3.
Next, we investigated the effect of enzyme concentration,
which influences both the immobilization of enzyme and the extent
of nonspecific adsorption. To evaluate this effect, patterned and
nonpatterned PEG-grafted monolithic columns were prepared
using immobilization solutions with different concentrations of
HRP. The greater the concentration of HRP in the immobilization
solution, the larger the amount of resorufin produced for both
the patterned and nonpatterned monolith columns (Table 2).
Maximizing the activity of the patterned column through the use
of a higher HRP concentration must be balanced against minimiz-
ing nonspecific protein adsorption, which is facilitated by employ-
ing a lower HRP concentration in the immobilization solution. For
the conditions investigated, the HRP concentration of 0.10 mg/
mL best balanced these divergent trends and maximized the ratio
of patterned/nonpatterned product formation.
Finally, Tween-20, a common surfactant, was included in the
immobilization solution to further disrupt hydrophobic interactions
and reduce the nonspecific adsorption of HRP on the PEG-grafted
polymer monoliths. The activity of the patterned column remained
high when the Tween-20 concentration was 0.010% (v/v). How-
ever, at 0.10% (v/v), the activity of the patterned column was
significantly reduced (Table 3). The use of a high concentration
of Tween-20 likely had an adverse effect on the activity of the
immobilized HRP.49 However, PEG grafting in combination with
the appropriate HRP and Tween-20 concentrations (0.10 mg/mL
and 0.010% (v/v), respectively) resulted in a patterned/nonpat-
terned column ratio of 18. Thus, these techniques for reducing
nonspecific protein adsorption enabled the effective patterning of
protein to spatially defined regions within a microfluidic channel.
In Situ Measurement of Enzyme Kinetics Using Continu-
ous Flow. In order to measure in situ steady-state fluorescent
product profiles, HRP and GFP were coimmobilized to a small,
patterned region on the surface of a PEG-grafted monolith. GFP
(49) Kenny, G. E.; Dunsmoor, C. L. Isr. J. Med. Sci. 1987, 23, 732-4.
6596 Analytical Chemistry, Vol. 79, No. 17, September 1, 2007

Figure 3. Consecutive patterning of multiple GFP patches on a
polymer monolith in a 100-µm capillary. The top image shows the
location of GFP simultaneously immobilized in multiple patches; each
GFP patch is ∼750 µm in length. The lower image shows the
fluorescence from GFP patches after a second round of GFP
immobilization on the same monolith. The plot shows profiles of the
GFP fluorescence along the length of the capillary for the above
images; the fluorescence intensity at each point in the column was
determined by averaging the fluorescence intensity across the cross
section of the monolith. In the images above, the diameter of the
monolith has been increased over the actual aspect ratio to more
clearly show the patch positions.
resorufin fluorescence prior to the microreactor increases with
decreasing flow rates. This phenomenon can be attributed to the
activity of remaining nonspecifically adsorbed HRP upstream of
the microreactor. The steady-state product profiles were used to
determine the relationship between reaction conversion and
residence time for the immobilized HRP. Detailed description of
the measurements and their results are presented in Supporting
Information.
Directional Synthesis in Spatially Separated Multienzy-
matic Reactions. Photopatterning allows multiple proteins to be
independently positioned within a single microfluidic channel.
Such a configuration enables multistep sequential enzymatic
reactions to occur within well-defined spatial reaction zones with
a reaction sequence that is determined by the direction of fluid
flow. We demonstrate the ability to pattern protein by consecu-
tively photopatterning multiple patches of GFP, with one set of
patches slightly offset from the other, as shown in Figure 3. The
fluorescent intensity of the patterned GFP revealed that the second
round of photopatterning did not denature the initial GFP patches,
nor did it result in additional GFP binding to the original patches.
Patterning enzymes in microfluidic systems also enables
control of the sequence of multienzymatic reactions. We demon-
strate directional synthesis using a model two-enzyme reaction
involving GOX and HRP. In the first step of the reaction, GOX
Figure 2. (A) Steady-state formation of fluorescent resorufin from
Amplex Red catalyzed by photopatterned HRP immobilized on
polymer monolith in a 100-µm capillary (substrate solution flow is from
left to right). The top fluorescent image shows the green fluorescence
of GFP, which coincides with the location of the coimmobilized
patterned HRP. The other fluorescent micrographs show the red
fluorescence of resorufin produced in the microreactor. Conditions:
substrate solution, 8.80 µmol/L hydrogen peroxide, 100 µmol/L
Amplex Red, and 1.0% (v/v) DMSO in 50 mmol/L phosphate buffer,
pH 7.50. (B) Steady-state resorufin fluorescence intensity profiles
taken from the images in (A). The fluorescence intensity at each point
in the microreactor column was determined by averaging the
fluorescence intensity across the cross section of the monolith. The
fluorescence profile of GFP indicates the location of the coimmobilized
HRP.
was included to enable direct visualization of the immobilized
protein. Control experiments determined that including 0.10 mg/
mL GFP did not measurably affect the activity of the immobilized
HRP. Fluorescent images of the patterned GFP and the steady-
state production of resorufin at flow rates ranging from 5.0 to 60
µL/h are shown in Figure 2. A significant increase in resorufin
fluorescence coincided with the location of the patterned protein,
as evidenced by the GFP fluorescence, confirming that active HRP
was photopatterned with little nonspecific adsorption. Two ad-
ditional observations were that (i) for a given flow rate the
resorufin fluorescence increased along the length of the microre-
actor and (ii) lower flow rates produced a higher level of resorufin
fluorescence. Both of these observations can be explained by the
effect of the residence time of the substrate, Amplex Red, in the
reactor. For a given flow rate, the residence time increases with
increasing distance along the length of the microreactor; while
for a given position in the microreactor, residence time increases
as the flow rate decreases. One final observation is that the
employs dissolved oxygen to catalyze the oxidation of â-
to δ- -glucono-1,5-lactone and hydrogen peroxide. The hydrogen
D-glucose
D
peroxide is then utilized by HRP in the second step of the reaction
to oxidize Amplex Red to resorufin. The enzymes were patterned
by first immobilizing HRP to one end of a monolith, followed by
immobilization of GOX to the other end. The patterned GOX
section was 8 times longer than the patterned HRP section to
achieve a longer residence time with GOX and compensate for
the relatively slow GOX reaction rate observed in separate
comparisons of the two enzymes (data not shown). A substrate
Analytical Chemistry, Vol. 79, No. 17, September 1, 2007 6597

gen peroxide, which is then used by HRP to oxidize Amplex Red
in the third step. Monolithic microreactors were created such that
the INV and GOX sections were each 4 times longer than the
HRP section. Several microreactors were prepared with all six
possible arrangements of the three enzymes. A substrate solution
consisting of sucrose and Amplex Red was then flushed through
the monoliths and the samples were analyzed; the results are
presented in Figure 5. The correct sequential order of catalyst
(INV- GOX-HRP) produced the largest resorufin fluorescence
by more than 3-fold. Although these studies were carried out
across several weeks, no significant change in the enzymatic
activity was observed.
In this case, a third enzyme was easily added to the same
microchannel. In the same way, additional enzymes could be
photopatterned within a single microchannel to create a multi-
enzyme system. By comparison, including an additional enzyme
in a spatially separated system based on microbeads would require
redesign of the entire microfluidic device.
Figure 4. Directional synthesis with a two-enzyme system: GOX
and HRP. Conditions: monolithic support, 9 cm long; GOX and HRP
patterned to a length of 8 and 1 cm, respectively; product fluorescence
measured in both the forward (GOX f HRP) and reverse (HRP f
GOX) directions; substrate solution, 1.0 mg/mL dextrose, 100 µmol/L
Amplex Red, and 1.0% (v/v) DMSO in 50 mmol/L phosphate buffer,
pH 7.50; pure oxygen was bubbled through this solution for 15 min
prior to use.
SUMMARY AND CONCLUSIONS
We have developed a simple strategy to immobilize several
enzymes in different and well-defined regions of a microfluidic
device by a photopatterning method involving a high-surface-area
polymer monolith support. Photoinitiated grafting of selected
monomers enabled activation of the support for immobilization
and reduction of nonspecific protein adsorption. Use of the in situ
prepared monolithic supports provides the increased surface area
without the need for separate retaining structures required when
using microbeads. The steady-state product profiles of a patterned
enzyme were measured and utilized to determine enzyme kinetics
in situ. It is anticipated that similar measurements could be
performed in a high-throughput format to determine the kinetic
properties for large enzyme libraries. The same patterning
technique was also employed to demonstrate directional biosyn-
thesis with model two- and three-enzyme reactions. Potential
applications of photopatterned enzyme sequences include minia-
turized analytical and diagnostic assays utilizing a specific order
of reactions, high-throughput combinatorial biocatalysis in a
microfluidic format, and the creation and investigation of artificial
metabolic pathways ex situ.
Figure 5. Directional synthesis with a three-enzyme system: INV,
GOX, and HRP. Conditions: monolithic support, 9 cm long; INV,
GOX, and HRP patterned to a length of 4, 4, and 1 cm, respectively;
Product fluorescence was measured from columns prepared with
each possible arrangement of the three enzymes, as indicated in the
figure legend; The substrate solution consisted of 10 mg/mL sucrose,
100 µmol/L Amplex Red, and 1.0% (v/v) DMSO in 50 mmol/L
phosphate buffer, pH 7.50; pure oxygen was bubbled through this
solution for 15 min prior to use. The flow rate was 0.10 µL/min.
ACKNOWLEDGMENT
solution consisting of glucose and Amplex Red was flushed
through the capillary in forward and reverse directions, and the
results of these experiments are plotted in Figure 4. As previously
observed, the concentration of fluorescent product increased with
residence time. More important, the direction of fluid flow
substantially affected the fluorescence. When the substrate solu-
tion passed through the column in the forward direction, i.e., from
GOX to HRP, the fluorescence was much higher than when then
same solution flowed in the reverse direction. The minimal
fluorescence observed in the reverse flow direction was likely due
to the attachment of either GOX or HRP outside of their defined
patterned regions. This attachment may be attributed to effects
such as nonspecific adsorption to the monolith, noncovalent
interactions with previously immobilized enzyme, and covalent
attachment to residual azlactone groups.
Financial support of this work by the National Institute of
Biomedical Imaging and Bioengineering, National Institutes of
Health (EB-006133), is acknowledged with thanks. T.C.L. and
D.S.C. recognize partial support of this work by the UC Discovery
Grant, the IndustrysUniversity Cooperative Research Program,
and Genencor International. Work at the Molecular Foundry was
supported by the Director, Office of Science, Office of Basic
Energy Sciences, Materials Sciences and Engineering Division,
of the U.S. Department of Energy under Contract DE-AC02-
05CH11231. T.C.L. and T.B.S. contributed equally to this work.
SUPPORTING INFORMATION AVAILABLE
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
Received for review April 10, 2007. Accepted June 20,
2007.
A sequential three-enzyme reaction was also performed. In the
first step, INV hydrolyzes sucrose to glucose and fructose; in the
second step, GOX oxidizes glucose to gluconolactone and hydro-
AC070705K
6598 Analytical Chemistry, Vol. 79, No. 17, September 1, 2007