Microchip Device for Performing Enzyme Assays

Article abstract of DOI:10.1021/ac970192p

An automated enzyme assay was performed within a microfabricated channel network. Precise concentrations of substrate, enzyme, and inhibitor were mixed in nanoliter volumes using electrokinetic flow. Reagent dilution and mixing were controlled by regulating the applied potential at me terminus of each channel, using voltages derived from an equivalent circuit model of the microchip. The enzyme β-galactosidase (β-Gal) was assayed using resorufin β-D-galactopyranoside (RBG), a substrate that is hydrolyzed to resorufin, a fluorescent product Reaction kinetics were obtained by varying the concentration of substrate on-chip and monitoring the production of resorufin using laser-induced fluorescence. Derived Michaelis-Menten constants compared well between an on-chip and a conventional enzyme assay. Bias in the derived K_m and k_cat was primarily due to the limited solubility of RBG and the associated lack of measurements at substrate concentrations exceeding the K_m. A K_i of 8 μM for the inhibitor phenylethyl β-D-thiogalactoside (PETG) was determined from plots of initial rate versus substrate concentration obtained at three concentrations of PETG. The relative inhibition of β-Gal by lactose, p-hydroxymercuribenzoic acid, and PETG was determined by varying the inhibitor concentration with constant enzyme and substrate concentration. An enzyme assay performed on the microchip within a 20-min period required only 120 pg of enzyme and 7.5 ng of substrate, reducing the amount of reagent consumed by 4 orders of magnitude over a conventional assay.

Full text of DOI:10.1021/ac970192p

A C R e s e a r c h
Anal. Chem. 1997, 69, 3407-3412
Accelerated Articles
Microchip Device for Performing Enzyme Assays
Andrew G. Hadd,† Daniel E. Raymond,† John W. Halliwell,‡ Stephen C. Jacobson,† and
J. Michael Ramsey*,†
Chemical and Analytical Sciences Division, and Instrumentation and Controls Division, Oak Ridge National Laboratory,
P.O. Box 2008, Building 4500S, MS 6142, Oak Ridge, Tennessee 37831-6142
network etched into a planar glass substrate using standard
photolithography, wet chemical etching, and bonding techniques.
Microchips represent the ability to miniaturize current “bench-
top” experiments with the advantages of speed, automation, and
volumetric reduction of sample and waste. Analytical applications
using microchips have been primarily demonstrated with electri-
cally driven separation techniques, such as capillary electro-
phoresis,3-8 synchronized cyclic electrophoresis,9 free-flow elec-
trophoresis,10 open-channel electrochromatography,11 and capillary
gel electrophoresis.12-14 Monolithic integration of a chemical
reaction and electrophoretic separation on single devices has been
demonstrated using pre- and postseparation derivatization.15,16
Biochemical applications of integrated microchips have been
demonstrated with a DNA restriction digest combined with
electrophoretic sizing17 in PCR amplification with product analy-
An automated enzyme assay was performed within a
microfabricated channel network. P recise concentrations
of substrate, enzyme, and inhibitor were mixed in nano-
liter volumes using electrokinetic flow. Reagent dilution
and mixing were controlled by regulating the applied
potential at the terminus of each channel, using voltages
derived from an equivalent circuit model of the microchip.
The enzyme â-galactosidase (â-Gal) was assayed using
resorufin â-D-galactopyranoside (RBG), a substrate that
is hydrolyzed to resorufin, a fluorescent product. Reac-
tion kinetics were obtained by varying the concentration
of substrate on-chip and monitoring the production of
resorufin using laser-induced fluorescence. Derived
Michaelis-Menten constants compared well between an
on-chip and a conventional enzyme assay. Bias in the
derived Km and kcat was primarily due to the limited
solubility of RBG and the associated lack of measure-
ments at substrate concentrations exceeding the Km. A
(2) Jacobson, S. C.; Ramsey, J. M. In Handbook of Capillary Electrophoresis;
Landers, J. P., Ed.; CRC Press, Inc.: Boca Raton, FL, 1997.
(3) Harrison, D. J.; Manz, A.; Fan, Z.; Lu¨ di, H.; Widmer, H. M. Anal. Chem.
1 9 9 2 , 64, 1926-1932.
Ki of 8 µM for the inhibitor phenylethyl â-D-thiogalactoside
(P ETG) was determined from plots of initial rate versus
substrate concentration obtained at three concentrations
of P ETG. The relative inhibition of â-Gal by lactose,
p-hydroxymercuribenzoic acid, and P ETG was deter-
mined by varying the inhibitor concentration with constant
enzyme and substrate concentration. An enzyme assay
performed on the microchip within a 2 0 -min period
required only 1 2 0 pg of enzyme and 7 .5 ng of substrate,
reducing the amount of reagent consumed by 4 orders of
magnitude over a conventional assay.
(4) Manz, A.; Harrison, J.; Verpoorte, E. M. J.; Fettinger, J. C.; Paulus, A.; Lu¨ di,
H.; Widmer, H. M. J. Chromatogr. 1 9 9 2 , 593, 253-258.
(5) Seiler, K.; Harrison, D. J.; Manz, A. Anal. Chem. 1 9 9 3 , 65, 1481-1488.
(6) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C. S.; Manz, A.
Science 1 9 9 3 , 261, 895-897.
(7) Jacobson, S. C.; Hergenro¨ der, R.; Koutny, L. B.; Warmack, R. J.; Ramsey,
J. M. Anal. Chem. 1 9 9 4 , 66, 1107-1113.
(8) Jacobson, S. C.; Hergenro¨ der, R.; Koutny, L. B.; Ramsey, J. M. Anal. Chem.
1 9 9 4 , 66, 1114-1118.
(9) Burggraf, N.; Manz, A.; Effenhauser, C. S.; Verpoorte, E.; de Rooij, N. F.;
Widmer, H. M. J. High Resolut. Chromatogr. 1 9 9 3 , 16, 594.
(10) Raymond, D. E.; Manz, A.; Widmer, H. M. Anal. Chem. 1 9 9 4 , 66, 2858-
2865.
(11) Jacobson, S. C.; Hergenro¨ der, R.; Koutny, L. B.; Ramsey, J. M. Anal. Chem.
1 9 9 4 , 66, 2369-2373.
(12) Effenhauser, C. S.; Paulus, A.; Manz, A.; Widmer, H. M. Anal. Chem. 1 9 9 4 ,
Microfabricated devices have been recently developed for
performing a variety of chemical measurements.1,2 A class of these
devices, known as microchips, are designed with a channel
66, 2949-2953.
(13) Woolley, A. T.; Mathies, R. A. Proc. Natl. Acad. Sci. U.S.A. 1 994 , 91, 11348.
(14) Woolley, A. T.; Matthies, R. A. Anal. Chem. 1 9 9 5 , 67, 3676.
(15) Jacobson, S. C.; Hergenro¨ der, R.; Moore, A. W., Jr.; Ramsey, J. M. Anal.
Chem. 1 9 9 4 , 66, 4127-4132.
† Chemical and Analytical Sciences Division.
‡ Instrumentation and Controls Division.
(16) Jacobson, S. C.; Koutny, L. B.; Hergenro¨ der, R.; Moore, A. W., Jr.; Ramsey,
(1) Manz, A.; Harrison, D. J.; Verpoorte, E.; Widmer, H. M. Adv. Chromatogr.
1 9 9 3 , 33, 1-66.
J. M. Anal. Chem. 1 9 9 4 , 66, 3472-3476.
(17) Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1 9 9 6 , 68, 720-723.
S0003-2700(97)00192-3 CCC: $14.00 © 1997 American Chemical Society
Analytical Chemistry, Vol. 69, No. 17, September 1, 1997 3407

sis,18 and multichannel restriction fragment separations.19
In these microchip systems, fluid flow and reagent mixing is
achieved using electrokinetic transport phenomena (electroos-
mosis and electrophoresis). Electrokinetic flow is controlled by
regulating the applied potentials at the terminus of each channel
of the microchip. Within the channel network, cross intersections
and mixing tees are used for valving and dispensing fluids with
high volumetric reproducibility (0.3% RSD). The mixing tee can
also be used to mix proportionately two fluid streams in any ratio
from 0 to 100% from either stream simply by varying the relative
field strengths in the two channels. As an example, dilution of
disodium fluorescein with buffer at a microchannel T-intersection
has been demonstrated.6 This precise fluid control can be applied
to the study of reaction systems in which reagent concentrations
are dynamically changed within a single experiment.
This paper describes the use of electrokinetic flow to control
the dilution and mixing of the reagents used in an enzyme assay.
Precise concentrations of a fluorogenic substrate, resorufin â-D-
galactopyranoside (RBG), are mixed with â-galactosidase (â-Gal),
and the kinetics of the reaction are obtained by monitoring the
fluorescence of the hydrolysis product, resorufin, using laser-
induced fluorescence. Michaelis-Menten constants are derived
for the hydrolysis reaction in the presence and absence of a
Figure 1. Schematic of the enzyme analysis chip. The channels
terminate at reservoirs containing the indicated solutions.
competitive inhibitor, phenylethyl â-D-thiogalactoside (PETG).
and 0.10 mM MgCl2 at pH 7.8 using filtered Barnstead Nanopure
water.
These results are compared to values obtained using conventional
measurements. An experiment comparing the relative inhibition
of three compounds is also described. These results are applicable
to the development of microfabricated devices for use in drug
discovery, medical diagnostics, and biochemical detection meth-
ods.
Apparatus. Platinum electrodes in each reservoir provided
electrical contact between each buffer solution and an individual
Ultravolt 10A12-P4 high-voltage power supply (Ultravolt, Ronkonko-
ma, NY). An electrode attached to ground was submersed in the
buffer solution of the waste reservoir. The resistance in each
segment of the channel network was obtained by applying a high
voltage between each of the 10 reservoir pairs and measuring the
current using a Keithley 617 programmable electrometer (Keithley
Instruments Inc., Cleveland, OH). Analysis of the equivalent
circuit was performed using the software P-SPICE (Microsim
Corp., Irvine, CA).
Imaging of fluid flow within the channels was obtained using
laser-induced fluorescence of a 100 µM disodium fluorescein
solution. An argon ion laser (Omnichrome, 488 nm, 100 mW)
was expanded to a 10 mm spot at the chip surface using a lens.
Laser-induced fluorescence was measured using a charge-coupled
device (CCD) mounted on an optical microscope. The fluores-
cence signal was spectrally filtered using a 540DF30 band-pass
filter (Omega Optical, Brattleboro, VT). On-chip reaction kinetics
were monitored using single-point laser-induced fluorescence
detection as described previously.17 A helium-neon laser beam
(Melles-Griot, 543.5 nm, 1.5 mW) was focused on the channel 20
mm downstream from the four-way reaction cross using a
planoconvex fused-silica lens and a mirror to impinge the beam
onto the chip 50° from normal. The fluorescence signal was
collected using a 20× objective lens (NA ) 0.42), followed by
spatial filtering (0.8-mm-diameter pinhole) and spectral filtering
with a 570-nm cutoff filter (Oriel 51310, Stratford, CT). Collected
fluorescence emission was measured using a photomultiplier tube
(PMT, Oriel 77340). A Labview (National Instruments Corp.,
Austin, TX) program written in-house was used to control the high-
voltage power supplies and data acquisition.
EXPERIMENTAL SECTION
Chip Design. The microchip design used for this study is
shown in Figure 1. The fabrication of the microchips using
standard photolithographic, wet chemical etching, and cover plate
bonding techniques has been described.1,2,20 Cylindrical reservoirs
cut from glass tubing were bonded to the edge of the cover plate
at the end of each channel using Epo-tek 353ND epoxy (Epoxy
Technologies, Inc., Billerica, MA). The channels had a depth of
9 µm and a width at half-height of 35 µm, measured using a Tencor
P-10 surface profiler (Tencor, Mountain View, CA). The microchip
was prepared daily by washing the channels with 0.1 N sodium
hydroxide followed by analysis buffer using a vacuum line applied
to one of the channel reservoirs.
Chemicals. Sodium resorufin (7-hydroxy-3H-phenoxazin-3-
one) and resorufin â-
D
-galactopyranoside (ꢀ469 ) 1.9 × 104 L mol-1
cm-1) were purchased from Molecular Probes (Eugene, OR). Tris-
(hydroxymethyl)aminomethane (Tris), o-nitrophenyl â-
pyranoside (ONPG), phenylethyl â- -thiogalactoside (PETG),
p-hydroxymercuribenzoic acid (PHMB), -lactose, â-galactosidase
D-galacto-
D
D
(â-Gal, Escherichia coli, 780 units/ mg, 540 kDa), rhodamine B
(90%), disodium fluorescein, and mercaptoethanol were obtained
from Sigma (St. Louis, MO). All other chemicals were of analytical
grade. Stock solutions of â-Gal, 1.0 mg/ mL, and RBG, 340 µM,
were prepared in an analysis buffer of 100 mM Tris, 2.0 mM KCl,
(18) Woolley, A. T.; Hadley, D.; Landre, P.; deMello, A. J.; Mathies, R. A.;
Northrup, M. A. Anal. Chem. 1 9 9 6 , 68, 4081-4086.
(19) Woolley, A. T.; Sensabaugh, G. F.; Mathies, R. A. Anal. Chem. 1 9 9 7 , 69,
2181-2186.
On-Chip Enzyme Assay. A continuous-flow assay, monitored
with the single-point setup, was used to observe the product of
the on-chip reaction of â-Gal and RBG. The general reaction is
(20) Moore, A. W., Jr.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1 9 9 5 , 67,
4184-4189.
3408 Analytical Chemistry, Vol. 69, No. 17, September 1, 1997

shown in eq 1 for the â-Gal catalyzed hydrolysis of the nonfluo-
through that channel, microfluidic control is realized by adjusting
the applied potentials at each channel reservoir. With the
channels of the microchip filled with the Tris analysis buffer, the
resistances (in MΩ) for each channel section were R1 ) 60.3, R2
) 75.0, Rmix ) 5.5, R3 ) 23.5, R4 ) 20.0, and Rrxn ) 83.0, with an
average error of 0.1 MΩ. The subscripts 1-4 designate the
channel section connected to the reservoirs, and “mix” and “rxn”
designate the mixing and reaction channels as indicated in Figure
1. The placement of the cover plate affected the actual channel
length, contributing to the slight differences between R1 and R2
and R3 and R4.
(1)
In order to accurately partition fluid flow in the channels, the
junction potentials at the mixing tee, Vmix, and reaction cross, V
,
rxn
were maintained at constant values. By setting V to 625 V, for
rxn
a field strength of 220 V/ cm, the current in the reaction channel,
I
rxn, was defined, and the currents in the other channels were
calculated as a percent of Irxn. For the enzyme assay, the currents
were partitioned such that Imix ) 0.40Irxn and I3 ) I4 ) 0.30Irxn
rescent RBG (I) to resorufin (II, λem ) 585 nm) and D-galactose
(III). Reagents were mixed using electrokinetic pumping. In a
typical experiment, a stock solution of 340 µM RBG added to
reservoir 1 (Figure 1) was mixed in the mixing channel with
various amounts of buffer. The flow of diluted substrate was
continuously mixed at the reaction cross with a flow of â-Gal
(0.625-2.50 µg/ mL in reservoir 3) and either inhibitor or buffer
from reservoir 4. Fluid flow from the mixing channel and the
two side channels (channels 3 and 4) was calculated using current
measurements as described in the results section. A mixing ratio
of 40% flow from the mixing channel and 30% flow from each of
the side channels (3 and 4) was used for all experiments.
Detection of resorufin was obtained 20 mm downstream from the
reaction cross, which corresponded to a reaction channel volume
of 6.5 nL. The reagent flow rate and mixing time were controlled
by modifying the electric field strength in the reaction channel,
which was 220 V/ cm for most experiments. Initial reaction rates
were calculated by subtracting the fluorescence signal from the
background signal (the signal of RBG without enzyme present)
and dividing by the transit time. The fluorescence signal was
calibrated using a standard solution of resorufin. Enzyme assays
were performed at an ambient temperature of 21 °C.
.
The voltages applied to each reservoir according to those current
settings were calculated using Ohm’s law.22 Incremental changes
in V and V were used to change the concentration of substrate
1
2
in the mixing channel. The values for V and V were constrained
1
2
by keeping Vmix and Imix constant and choosing current values for
I1 and I2 from -3 to +103% Imix. The 3% “over-shoot” in the current
values compensated for diffusional and hydrodynamic bleeding
of solution from either channel at 0 and 100% flow.
Microfluidic control was verified using fluorophore-doped
buffer. A qualitative view of the flow of solution at the two channel
junctions is shown in Figure 2. White-light photomicrographs of
the mixing channel T-intersection and the reaction cross are
shown in Figure 2a and c. The voltages applied at each channel
reservoir are indicated. Dilution of a disodium fluorescein doped
buffer is shown in Figure 2b, with 50% flow from channels 1 and
2 into the mixing channel. Figure 2d shows the flow of disodium
fluorescein-doped buffer from channels 3 and 4 into the reaction
channel. The channels in Figure 2b and d are shown labeled with
the reagents used for the enzyme assays.
Mixing of the reagents and buffer streams at channel intersec-
tions was diffusion controlled due to the very low Reynolds
numbers (<10-2) associated with electrokinetically induced flow.
The 34-µm channel width allowed for rapid diffusion of reagents
downstream from an intersection. In the case of fluorescein, with
Standard Enzyme Assay. In order to compare results of the
on-chip assay with a conventional method, hydrolysis kinetics of
â-Gal and RBG were obtained by monitoring the absorbance of
resorufin at 571 nm (ꢀ571 ) 4.7 × 104 L mol-1 cm-1) with a Cary
1E UV-visible spectrophotometer. The reaction was initiated by
adding 50 µL of 10.4 µg/ mL â-Gal to 950 µL of a prepared RBG
solution for a final enzyme concentration of 520 µg/ L. The linear
portion of the absorbance versus time plot was used to determine
the initial reaction rate. The activity of â-Gal was calibrated by
an estimated diffusion coefficient of 3 × 10-6 cm2 s-1 23
, diffusion
across 17 µm of the mixing channel occurred in ∼0.5 s. In the
reaction channel of the microchip, diffusion of â-Gal (D ≈ 2.7 ×
24
10-7 cm2 s-1
)
across 70% of the channel width required ∼11 s.
The slower diffusion of â-Gal relative to the substrate had a
minimal effect on the observed reaction kinetics.
monitoring the hydrolysis of o-nitrophenyl â-D-galactopyranoside
at 410 nm (3500 L mol-1 cm-1) in 80 mM phosphate, 20 mM
Controlled fluid flow within the microchip channels was
evaluated by diluting a stock rhodamine B solution on-chip and
monitoring the fluorescence signal 20 mm downstream from the
reaction cross. The fluorescence signal of a rhodamine B solution
stepped in equal increments from 0.32 to 1.68 µM is shown in
Figure 3 for three runs. Each plateau in the signal versus time
plot represents the increased amount of flow from channel 1
containing the fluorophore-doped buffer. The average signal of
mercaptoethanol, pH 7.3, buffer at 37 °C.
RESULTS AND DISCUSSION
Microfluidic Control. A circuit model of the microchip was
used to derive the voltages applied at each channel reservoir
necessary for controlling fluid flow. The channels in the microchip
function as electrical resistors, and when measured, the potential
at any point within a channel can be estimated by application of
Kirchoff’s rules and Ohm’s law.21 Since the current through a
channel approximates the electrokinetic transport of material
(22) Seiler, K.; Fan, Z. H.; Fluri, K.; Harrison, D. J. Anal. Chem. 1994, 66, 3485-
3491.
(23) Magde, D.; Elson, E. L.; Webb, W. W. Biopolymers 1 9 7 4 , 13, 29-61.
(24) Wallenfels, K.; Weil, R. In The Enzymes; Boyer, P. D., Ed.; Academic
Press: New York, 1972; Vol. 7, pp 617-663.
(21) Giancoli, D. G. In General Physics; Prentice Hall, Inc.: Englewood Cliffs,
NJ, 1984; pp 515-531.
Analytical Chemistry, Vol. 69, No. 17, September 1, 1997 3409

Figure 4. Fluorescence signal of resorufin obtained by mixing
stepped-increments of RBG with (b) 190 , (2) 370, and (9) 740 µg/L
â-Gal. Each plateau represents a step increase of RBG from 14 to
122 µM. Every fiftieth data point is indicated with a symbol. Condi-
tions: 100 mM Tris, 2.0 mM KCl, and 100 µM MgCl2 pH 7.8 buffer at
21 °C.
Figure 2. CCD images of (a) the mixing tee and (c) the four-way
reaction cross under white-light illumination. The voltages applied to
the reservoirs of each channel section are indicated; the two junction
potentials are Vmax ) 645 V and Vrxn ) 625 V. (b) The fluorescence
signal of the electrokinetic flow and dilution of disodium fluorescein
(shaded areas) into the mixing channel. (d) The mixing of disodium
fluorescein solution from channels 3 and 4 with buffer from the mixing
channel into the reaction cross. The arrows indicate the direction of
electroosmotic flow, and the channels are labeled with the reagents
used in the enzyme assay.
ability to control the flow of fluid within a channel network by
changing the voltage applied at each reservoir.
Enzyme Assays. Enzyme kinetics were evaluated using the
Michaelis-Menten equation:
Vmax[S]
νo )
(2)
Km + [S]
which relates the initial rate of the enzyme reaction, νï, to the
concentration of substrate, [S], a maximum rate, Vmax, and an
equilibrium constant, Km. The equation describes a rectangular
hyperbola in which the initial rate is first-order in substrate
concentration when [S] , Km and zeroth order in substrate
concentration when [S] . Km. The Km of the reaction is defined
as the amount of substrate required for half-maximal velocity. The
kcat of an enzyme, a first-order rate constant, is related to the total
enzyme concentration, [E]t, by
kcat ) Vmax/ [E]t
(3)
A Lineweaver-Burk plot of 1/ ν versus 1/ [S], a linear form of eq
2, was used to calculate kcat and Km from initial rate versus
substrate concentration data.
The enzyme kinetics were obtained by measuring the fluores-
cence signal of resorufin after a fixed mixing time determined by
the electrokinetic flow rate in the reaction channel. In a typical
assay (Figure 4), the final concentration of RBG was increased in
equal increments from 14 to 122 µM RBG and mixed with enzyme
for a final â-Gal concentration of 190, 370, or 740 µg/ L. The
enzyme concentration was changed by manually refilling reservoir
3 with a new enzyme concentration. . An increase in substrate
concentration resulted in the gradual increase of product reaching
a new plateau, the height of which was proportional to the rate of
enzyme turnover for that substrate concentration. Each increment
of RBG was monitored for 150 s, and the assay was completed in
20 min. At a flow rate of 14 nL/ min, the assay consumed an
Figure 3. Fluorescence intensity versus time for three runs of a
step-gradient increase from 0.32 to 1.68 µM rhodamine B. The left
axis shows the fluorescence signal collected 20 mm downstream from
the mixing cross, and the right axis shows the concentration of
rhodamine B corresponding to each signal plateau.
each step in Figure 3 was compared with the fluorescence signals
obtained by a manual calibration method. For the manual cali-
bration curve, solutions of rhodamine B prepared off-chip were
added to reservoirs 1 and 2 and the fluorescence signal was mea-
sured as in the on-chip dilution method. The two calibration
curves agreed within an average of 2% RSD and confirmed the
3410 Analytical Chemistry, Vol. 69, No. 17, September 1, 1997

Figure 5. Lineweaver-Burk plot of the initial rate versus substrate
concentration for (b) 190 , (2) 370, and (9) 740 µg/L â-Gal.
Figure 6. Fluorescence signal of resorufin produced in the PETG-
inhibited hydrolysis of RBG and â-Gal. Each plateau represents
increments of 14 µM RBG from 28 to 115 µM mixed with (b) buffer,
(9) 6 µM PETG, ([) 12 µM PETG, and (2) 30 µM PETG. Symbols
are shown for 0.5% of the data points. Conditions: 570 µg/L â-Gal
for each reaction in 100 mM Tris, 2.0 mM KCl, and 100 µM MgCl2
pH 7.8 buffer at 21 °C.
Table 1. Derived Michaelis-Menten Constants
Measured On-Chip with Three Concentrations of â-Gal
Compared with a Cuvette-Based Measurement
concentration of â-Gal, µg/ L
190
370
740
absorbancea
Km (µM)
360 ( 100
54 ( 20
320 ( 80
54 ( 20
660 ( 200
80 ( 20
550 ( 200
70 ( 30
the maximum rate, could not be observed, because the maximum
concentration of RBG was at least 3 times lower than the derived
Km. Therefore, the range of RBG concentration used in this study
limited the run-to-run precision in the derived Michaelis-Menten
constants.
k
cat (s-1
)
a
Initial rates obtained by conventional absorbance assay of resorufin
at 571 nm using 520 µg/ L â-Gal.
Protein absorption to the channel walls contributed to an
increase in the background signal of RBG. At low substrate
concentrations, 28-42 µM RBG, the background signal was 50%
of the signal produced in the enzyme-catalyzed reaction. The
background signal increased from a starting value by 10-20%
during the course of three runs, depending on the concentration
of the enzyme and length of analysis time. An increase in the
background signal during the course of an assay may have
contributed to the observed deviation of the results for the lower
substrate concentrations in the Lineweaver-Burk plot shown in
Figure 5.
average of 120 pg of â-Gal and 7.5 ng of RBG. The dwell time at
each substrate concentration and the corresponding reagent
consumption can be further reduced with additional refinement.
A Lineweaver-Burk plot using the initial rates calculated from
the data in Figure 4 is shown in Figure 5. For a Lineweaver-
Burk plot, the slope of the line is Km/ Vmax, the y-intercept is 1/ Vmax
and the extrapolated 1/ [S] intercept is -1/ Km. Values for Km
and kcat, derived from a weighted least-squares fit to the points
shown in Figure 5, are summarized in Table 1 and compared with
constants obtained using a conventional method. The average
on-chip Km value was 450 ( 200 µM, which compared well with
the 550 ( 200 µM obtained using cuvettes. The turnover rate
for the enzyme, kcat, also agreed between the microchip and
cuvette measured reaction. The values in Table 1 compare
reasonably well with a literature value Km of 380 µM but dif-
Inhibition of â-Galactosidase. Enzyme inhibition was evalu-
ated using phenylethyl â-
D-thiogalactoside, a competitive inhibitor
of â-Gal. Three concentrations of PETG were mixed in the
reaction cross with â-Gal, and the resulting decrease in resorufin
production was monitored over a range of RBG concentrations.
As the concentration of PETG increased, the initial reaction rate
at each RBG concentration decreased, as shown in Figure 6. The
PETG concentration was changed by manually refilling reservoir
4 with a new solution of PETG. Average values for the rate of
resorufin produced for each concentration of RBG and PETG are
shown in Figure 7 as a set of Lineweaver-Burk plots. The plots
shown in Figure 7 are diagnostic for competitive inhibition and a
Ki of 8 µM was derived by plotting the slope of each line versus
the inhibitor concentration.26 The Ki derived using RBG and the
microchip assay is larger than the reported literature values, which
range from 0.98 to 2.5 µM PETG.27-29 A limited substrate
fer by a factor of 10 from a literature value kcat of 700 s-1 25
.
For â-Gal, different rate constants using similar substrates have
been obtained, due to differences in the pH, ionic strength, and
the presence of activators,24 which may account for the dif-
ference between the literature kcat value and those listed in
Table 1.
Reproducibility in the microchip enzyme assay was mainly
affected by the solubility of the substrate, with minor contributions
from protein absorption to the channel walls. In an ideal enzyme
assay, a substrate concentration range is chosen to encompass
values above and below the approximate Km. In this experiment,
saturation kinetics, in which the initial reaction rate approaches
(26) Engel, P. C.; Dickinson, F. M.; Cornish-Bowden, A. In Enzymology; Engel,
P. C., Ed.; Academic Press: San Diego, 1996; pp 77-113.
(25) Hofmann, J.; Sernetz, M. Anal. Chim. Acta 1 9 8 4 , 163, 67-72.
(27) de Bruyne, C. K. D.; Yde, M. Carbohydr. Res. 1 9 7 7 , 56, 153-164.
Analytical Chemistry, Vol. 69, No. 17, September 1, 1997 3411

Figure 8. Effect of (b) 0-40 µM PETG, (9) 0-4.0 mM lactose,
and (2) 0-2 mM PHMB on the observed relative inhibition of the
â-Gal-catalyzed hydrolysis of RBG. Each data point represents a 20%
increase in the concentration of inhibitor ranging from 0 to 100%.
Conditions: 100 µM RBG; 370 µg/L â-Gal.
Figure 7. Lineweaver-Burk plot of the PETG-inhibited hydrolysis
of RBG showing initial rate and substrate concentration for (b) no,
(9) 6, ([) 12, and (2) 30 µM PETG.
concentration range, with the corresponding uncertainty in the
derived Vmax, may account for the discrepancy between this
experiment and the reported values.
allowed the various reagents to be combined in varying propor-
tions while the extent of the reaction was monitored under
continuous-flow conditions. The â-galactosidase-catalyzed hy-
The relative effect of PETG, lactose, and PHMB on the
inhibition of â-Gal was evaluated by varying the inhibitor concen-
tration with constant substrate and enzyme concentration. The
inhibitor was added to reservoir 1, diluted in the mixing channel,
and mixed with a constant concentration of enzyme and substrate
in the reaction cross. Signal versus time profiles similar to those
in Figure 4 were obtained using 20% increments from 0 to 100%
of the inhibitor concentration contained in reservoir 1. Percent
inhibition, % I, was defined as
drolysis of resorufin â-D-galactopyranoside was used as a model
system for enzyme kinetic and inhibition determinations. This
approach to enzyme assays allows automated studies to be
completed with significant time savings and reduction of reagent
consumption by more than 4 orders of magnitude while delivering
results consistent with conventional approaches. Further de-
creases in measurement time and reagent consumption are
possible. Highly parallel assays are also imaginable due to the
small footprint of the microchip device. The channel network of
the microchip also allowed for a rapid assessment of relative
inhibition over a range of inhibitor concentrations. This technique
represents the ability to rapidly study enzyme inhibition with small
amounts of potentially expensive enzymes, substrates, or inhibi-
tors.
So - Si
So
% I )
× 100
(4)
where So was the signal produced in the absence of an inhibitor,
and Si, the signal at the various inhibitor concentrations. Plots
of % I in the range 0-4 mM lactose, 0-2 mM PHMB, and 0-40
µM PETG are shown in Figure 8 as a function of inhibitor
concentration. By fitting the lactose and PETG inhibition curves
to the Michaelis-Menten equation (eq 2), values for i50, the
inhibitor concentration required to give 50% inhibition, were
determined. With 100 µM RBG, the i50 values were 7.2 µM for
PETG and 4.2 mM for lactose. Since PHMB deactivates the
enzyme by reacting with SH groups, only slight inhibition over
the 0-2 mM range was observed due to the slower rate of
inhibition. Previous inhibition studies of â-Gal have observed
similar effects and concentration ranges for these compounds.24
ACKNOWLEDGMENT
Research sponsored by Central Measurement and Signature
Intelligence Office (MASINT), U.S. Department of Defense. Oak
Ridge National Laboratory is managed by Lockheed Martin
Energy Research Corp. for the U.S. Department of Energy, under
Contract DE-ACO5-96OR22464. This research was supported in
part by an appointment of A.G.H. and D.E.R. to the ORNL
Postdoctoral Research Associates Program administered jointly
by the Oak Ridge Institute for Science and Education and ORNL.
We thank Donald J. Rose for initial discussions, Julius C. Fister
III, and K. Bruce Jacobson for helpful comments, and Justin E.
Daler for fabrication of the microchips.
CONCLUSIONS
An integrated microfabricated device that performs automated
enzymatic assays was evaluated. Active and precise microfluidic
control of reagent transport throughout the interconnected chan-
nel network was achieved using electrokinetic-induced motion.
Computer control of the applied potential at each channel terminus
Received for review February 18, 1997. Accepted June
30, 1997.X
AC970192P
(28) Huang, Z. Biochemistry 1 9 9 1 , 30, 8530-8534.
(29) Russo-Marie, F.; Roederer, M.; Sager, B.; Herzenberg, L. A.; Kaiser, D.
Proc. Natl. Acad. Sci. U.S.A. 1 9 9 3 , 90, 8194-8198.
X Abstract published in Advance ACS Abstracts, August 1, 1997.
3412 Analytical Chemistry, Vol. 69, No. 17, September 1, 1997