Photochemical coenzyme regeneration in an enzymatically active optical material

Article abstract of DOI:10.1021/jp038051g

The photoinduced electron transfer between immobilized thionine and the dinucleotide enzyme cofactors NADH and NADPH in a SiO₂ sol-gel matrix is reported. The electron-transfer quenching of thionine luminescence is used to monitor the rate of NADPH oxidation. Using Stern-Volmer quenching curves, the quenching rates in the silica matrix are 1 to 2 orders of magnitude smaller than those in solution, The rate constants for oxidation of NADPH by thionine were measured to be 9.8(?±2.9) ?? 10^-3 s^-1 in solution and 8.8(?±1.0) ?? 10^-4 s^-1 in the gel. Within the silica matrix, the photoinduced oxidation of NADPH is combined with the enzymatic reaction of isocitrate dehydrogenase, which uses the oxidized cofactor, NADP⁺, as an electron acceptor in the oxidation of isocitrate. The encapsulated isocitrate dehydrogenase is active with a Michaelis-Menten constant, K_M, of 3 ??M and a k_cat of 0.67 ??M/s per mg_enzyme?· Because optical sensors use NADPH fluorescence as an indicator of the presence and relative concentration of enzyme substrate, the successful demonstration of photoinduced regeneration of NADP⁺ makes possible continuous monitoring by the family of dehydrogenase enzymes.

Full text of DOI:10.1021/jp038051g

J. Phys. Chem. B 2004, 108, 9325-9332
9325
Photochemical Coenzyme Regeneration in an Enzymatically Active Optical Material
Jenna L. Rickus,†,§ Pauline L. Chang,‡ Allan J. Tobin,# Jeffrey I. Zink, and Bruce Dunn*,|
Departments of Chemistry and Biochemistry, Materials Science and Engineering, Chemical Engineering,
Neurobiology, Physiological Sciences, Brain Research Institute, Neuroscience IDP Neuroengineering Program,
UniVersity of California, Los Angeles, Los Angeles, California 90095
ReceiVed: December 30, 2003; In Final Form: April 15, 2004
The photoinduced electron transfer between immobilized thionine and the dinucleotide enzyme cofactors
NADH and NADPH in a SiO2 sol-gel matrix is reported. The electron-transfer quenching of thionine
luminescence is used to monitor the rate of NADPH oxidation. Using Stern-Volmer quenching curves, the
quenching rates in the silica matrix are 1 to 2 orders of magnitude smaller than those in solution. The rate
constants for oxidation of NADPH by thionine were measured to be 9.8((2.9) × 10-3 s-1 in solution and
8.8((1.0) × 10-4 s-1 in the gel. Within the silica matrix, the photoinduced oxidation of NADPH is combined
with the enzymatic reaction of isocitrate dehydrogenase, which uses the oxidized cofactor, NADP+, as an
electron acceptor in the oxidation of isocitrate. The encapsulated isocitrate dehydrogenase is active with a
Michaelis-Menten constant, KM, of 3 µM and a kcat of 0.67 µM/s per mgenzyme. Because optical sensors use
NADPH fluorescence as an indicator of the presence and relative concentration of enzyme substrate, the
successful demonstration of photoinduced regeneration of NADP+ makes possible continuous monitoring by
the family of dehydrogenase enzymes.
1. Introduction
other components, i.e., co-immobilized with the enzyme and
the enzyme cofactor. It must be stable in this matrix and retain
The specificity and selectivity of enzymes make them
important components for sensor applications. For optical
sensing, enzymes that use NAD or NADP as enzyme cofactors
are particularly important because the reduced cofactors fluo-
resce. This property makes these enzymatic reactions particularly
useful for solid-state and fiber optic based systems, as the
enzyme-coenzyme is able to provide both molecular recogni-
tion and signal transduction. Optical sensors take advantage of
the NADH or NADPH fluorescence as an optical indicator of
the presence and relative concentration of enzyme substrate.1
The largest class of redox enzymes that utilize NAD or NADP
as the enzyme cofactor are the dehydrogenases that carry out
the general reaction
its excited-state properties. In addition, the photooxidizer must
be compatible with the other components and unreactive.
We have chosen 3,7-diamino-5-phenothiazinium acetate,
thionine or thio+, as the photooxidizing agent because it meets
the above criteria. In solution, excited thionine oxidizes
NADH2,3 and we expect that it will also oxidize NADPH.
Thionine, thio+, absorbs light in the visible range, λmax ) 596
nm (reaction 2). The excited molecule, *thio+, can return to the
ground state via nonradiative processes (reaction 3) or fluores-
cence (reaction 4).4 In the presence of NADH, electron transfer
that oxidizes NADH competes with the fluorescence and
quenches the thionine emission. In this paper we use reaction 5
to regenerate the oxidized form of the enzyme cofactor that was
depleted by the enzyme reaction (reaction 1).
reduced substrate + NAD(P)+ f
oxidized substrate + NAD(P)H + H+ (1)
kd
9
thio+ + hV596
8 *thio+
(2)
This reaction consumes the oxidized form of the coenzyme.
Sensors based on this reaction cannot operate continuously
without an external supply of coenzyme.
The objective of the research in this paper is to determine
whether a photooxidizer can be incorporated in the same solid-
state matrix as the enzyme and cofactor to regenerate, by
oxidation, the reduced cofactor. For the photooxidation process
to be successful, several criteria need to be met. The photooxi-
dizer must be incorporated in the same matrix as that of the
kd
*thio+ 98 thio+
(3)
(4)
kf
*thio+ 98 thio+ + hV620
kq
9
*thio+ + NADH
8 thio + NAD+
(5)
The enzyme isocitrate dehydrogenase (ICDH) was selected
as a representative dehydrogenase. ICDH catalyzes the oxidation
of isocitrate to R-ketoglutarate, using NADP+ as the electron
acceptor (reaction 6).
* Address correspondence to this author. E-mail: bdunn@ucla.edu.
† Neuroscience IDP Neuroengineering Program.
§ Present address: Department of Agricultural and Biological Engineer-
ing, Department of Biomedical Engineering, Purdue University, West
Lafayette, IN 47907.
isocitrate + NAD(P)+ f
‡ Chemical Engineering.
R-ketoglutarate + CO2 + NAD(P)H + H+ (6)
# Neurobiology, Physiological Science, Brain Research Institute.
Chemistry and Biochemistry.
| Materials Science and Engineering.
While most dehydrogenase reactions are reversible, the change
10.1021/jp038051g CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/26/2004

9326 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Rickus et al.
in Gibbs free energy of the ICDH reaction strongly favors the
oxidation of isocitrate, thereby reducing experimental complica-
tions from back reactions.5
2.3. Kinetic Experiments. Time-resolved fluorescence mea-
surements were taken with a Fluoroskan Ascent microplate
fluorometer. Excitation and emission wavelengths for NADH
and NADPH were 355 and 460 nm. Excitation and emission
wavelengths for thionine were 584 and 620 nm. NADH and
NADPH fluorescence was measured with an integration time
of 20 ms while thionine fluorescence was measured with an
integration time of 100 ms.
Steady-state quenching experiments were conducted with
thionine-doped gels and thionine solutions with final concentra-
tions of 8 and 5 µM, respectively. At t ) 0, 20 µL of NADPH
or NADH was injected into the sample well, and thionine
fluorescence was measured over time. Co-encapsulation of the
coenzyme and thionine was not practical because coenzyme
oxidation occurred during the gelation period. For all samples
the quenched intensity ratio, Io/I, was defined as the average
intensity in the absence of quencher divided by the intensity
minimum at steady state after the injection of NADH or
NADPH.
When thionine is co-immobilized with the enzyme, the
NADPH produced by ICDH in the presence of isocitrate should
be converted back to NADP+ when the gel is exposed to visible
light. Excitation of the co-immobilized thionine should allow
for a self-sustaining supply of NADP+ for further isocitrate
oxidation.
The matrix we selected for study is sol-gel derived silica.
The sol-gel process leads to a transparent, porous glass that is
formed at room temperature and is able to physically immobilize
a wide variety of molecules.6-10 Of particular relevance for the
present work is the immobilization of dehydrogenases including
glucose dehydrogenase,10-12 lactate dehydrogenase,13-15 and
alcohol dehydrogenase.16,17 In all of these examples, the enzyme
cofactor (NAD+ or NADP+) used in the transduction was
consumed and not regenerated.
In this paper, we first show that NADPH is oxidized by
photoexcited thionine. The oxidations of the cofactors, NADH
and NADPH, are compared both in a silica sol-gel matrix and
in solution. The rates are characterized by steady-state quenching
experiments and by fluorescence measurements. Second, ICDH
is encapsulated in the sol-gel matrix and the kinetics of the
encapsulated enzyme are characterized. Finally, the thionine-
NADPH reaction is used to regenerate the NADP+ that was
consumed in the enzymatic oxidation of isocitrate by isocitrate
dehydrogenase. All the components were immobilized in the
optically transparent silica gel matrix. The result is a solid-state
system capable of being regenerated by light.
For NADH and NADPH oxidation experiments, gel and
solution samples with final NADPH concentrations of 33, 66,
and 100 µM were produced in triplicate. At t ) 0, 20 µL of 60
µM thionine was injected into each sample. Thionine was
excited over five intervals of 200 s, and the thionine fluorescence
emission was collected during each interval. NADPH fluores-
cence was measured before and after each thionine excitation
interval.
For enzyme characterization experiments, ICDH-doped gels
(240 ng/100 µL) were produced and aged for 24 h at 4 °C.
Kinetic experiments were conducted in the same 96-well plate
in which the gels were prepared. The gels were allowed to
equilibrate in a phosphate-buffered solution of NADP+, so that
the final cofactor concentration of NADP+ was 2.67 mM.
Isocitrate was injected into each well at varying concentrations
at time ) 0. The production of NADPH by the enzyme over
time was determined by measuring NADPH fluorescence over
time. NADPH fluorescence was converted to NADPH concen-
tration with use of a linear calibration curve (data not shown).
The initial reaction velocity was estimated by fitting the linear
portion of the fluorescence increase near time zero with least-
squares methods.
2. Experimental Methods
2.1. Materials Synthesis. NAD+, NADP+, NADH, NADPH,
isocitrate dehydrogenase, and isocitrate were all purchased from
Sigma. Thionine acetate and tetramethoxysilane (TMOS) were
purchased from Fluka. All chemicals were used as purchased.
A 100 µM stock solution of thionine was prepared by dissolving
thionine in buffer (0.02 M sodium phosphate, 0.02 M NaCl,
pH 7) and sonicating it for 10 min.
TMOS was hydrolyzed by sonication in water (silica:water
molar ratio of ∼1:2) and 0.5 mM HCl for 15 min. The
hydrolyzed TMOS sol was combined with buffer (0.02 or 0.1
M sodium phosphate, 0.02 or 0.1 M sodium chloride, pH 7.0)
in a 2:3 volumetric ratio. Emission spectra samples were cast
into 4.5 mL of polystyrene cuvettes, covered with Parafilm and
foil, and aged for 24 h. For kinetic fluorescence measurements,
100 µL of doped sol was cast into the wells of a 96-well Nunc
microplate. Dopants (thionine, NADH, or ICDH) were added
to the buffer prior to the addition of TMOS sol. Gels were
covered with a small volume of buffer to prevent drying during
the aging time. Drying leads to slow kinetics and therefore only
aged wet gels were investigated. Thionine- and NADH-doped
monoliths were aged for 1 h at room temperature. ICDH-doped
monoliths (240 ng/100 µL) were aged for 24 h at 4 °C. Solution
controls were made for comparison by adding buffer in place
of hydrolyzed TMOS.
The reaction of thionine with NADPH was coupled to the
enzyme reaction of isocitrate dehydrogenase (ICDH) by produc-
ing samples codoped with thionine and ICDH (144 ng/100 µL).
Samples were incubated with NADP+ so that the final enzyme
cofactor concentration was 16 mM. At time ) 0, 40 µL of 4
mM isocitrate was injected into the sample well. NADPH
production by ICDH was measured by fluorescence for ∼1600
s. The thionine was then excited at 584 nm for 200 s. Three
NADPH fluorescence measurements were made during the
thionine excitation period. Following the thionine excitation
period, isocitrate was again injected into the sample well. The
NADPH fluorescence again was measured for 1600 s, and
thionine was excited for 200 s. This process was repeated 2
more times without any washing steps.
3. Results and Discussion
2.2. Spectroscopic Measurements. Fluorescence emission
spectra were taken with a SPEX Fluorolog spectrometer with a
right angle configuration. Samples were excited at 590 nm and
emission intensity was measured from 610 to 700 nm. Measure-
ments were taken in the same polystyrene cuvettes in which
the gels were prepared. Only minimal shrinkage of the gels
occurred over the 24-h aging period.
3.1. Enzyme Cofactor Regeneration Scheme. The goal of
this work is to establish and characterize the oxidation of
NADPH with excited thionine within the pores of a silica matrix,
and to then couple this excited-state redox reaction to the
enzymatic reduction of NADPH by ICDH. Figure 1 illustrates
how the coupling of these reactions could be used for the
continuous sensing of isocitrate. As isocitrate is oxidized to

Photochemical Coenzyme Regeneration
J. Phys. Chem. B, Vol. 108, No. 26, 2004 9327
Figure 1. Enzyme cofactor regeneration for continuous isocitrate
oxidation. The photochemical oxidation of NADPH by thionine is
coupled to the enzyme reaction of isocitrate dehydrogenase (ICDH).
ICDH oxidizes isocitrate to R-ketoglutarate, using NADP+ as the
electron acceptor. When thionine is excited, it reacts with NADPH to
reform NADP+. The regenerated NADP+ can participate in another
round of isocitrate oxidation.
R-ketoglutarate by ICDH, NADP+ is reduced to NADPH. If
the resulting NADPH can be converted back to NADP+ by
excited thionine, then the cofactor can again participate in the
enzyme reaction. This system would have a self-sustaining
enzyme cofactor source for ongoing isocitrate detection (in the
presence of oxygen) without requiring additional reagents. Only
light need be supplied to the sensor material.
3.2. Formation of Doped Glasses. All glasses were formed
by using the sol-gel method described previously for encap-
sulating proteins while retaining the biological activity of the
protein.18,19 The precursor tetramethoxysilane, TMOS, is hy-
drolyzed under acidic conditions in a sonication bath. The
addition of phosphate buffer to the resulting hydrolyzed sol
increases the pH to promote condensation of the precursor to
form a three-dimensional porous gel. Because the dopant
molecule is introduced with the buffer, it is protected from the
harsh environment of the acidic sol.
In this work, thionine-doped gels, ICDH-doped gels, and
thionine/ICDH codoped gels are all synthesized. Gelation
occurred within 1 min, producing clear gels that were either
deep purple or colorless depending on the presence or absence
of thionine. Characterization of the reaction between encapsu-
lated thionine and NADH/NADPH in this “protein-friendly”
glass allowed the reaction to then be coupled to the ICDH
reaction in the same material by codoping gels with both
thionine and the enzyme.
3.3. ICDH Enzyme Kinetics in Aged Gels. Many enzymes
follow a steady-state Michaelis-Menten kinetic model in
solution.5 The model is described by the Michaelis-Menten
equation where Vo is the initial reaction velocity, [S] is the initial
substrate concentration, Vmax is the maximum reaction velocity,
and KM is the Michaelis-Menten constant (eq 7). The equation
results in two parameters that characterize the kinetics of the
reaction. The initial velocity of the reaction (the reaction velocity
before any significant product accumulation occurs) increases
with substrate concentration and approaches a maximum veloc-
ity, Vmax, as the enzyme becomes saturated with substrate. The
Michaelis-Menten constant, KM, is the substrate concentration
at which the reaction occurs at half the maximum velocity.
Figure 2. Lineweaver-Burke plot of ICDH after encapsulation in aged
gels (A) and in free solution (B). The inverse of the initial velocity of
the reaction is plotted against the inverse of the initial isocitrate
concentration. A linear fit of the data to the Michaelis-Menten constant
results in a KM of 3 µM and a Vmax of 0.17 µM/s for the encapsulated
enzyme and a KM of 29 µM and a Vmax of 0.05 µM/s for the free
enzyme.
values of these constants include diffusion and geometric
parameters as well as the intrinsic kinetic parameters of the
enzyme.20,21
To evaluate ICDH as a model dehydrogenase for co-
encapsulation with thionine, we produced ICDH-doped gels and
measured their response to varying isocitrate concentrations at
saturating NADP+ levels. Isocitrate was injected onto each
sample, and the ICDH reaction was monitored via the production
of NADPH. For each sample tested, NADPH fluorescence
increased with time after isocitrate injection and then leveled
off to a final steady value as is expected for an enzyme reaction.
Figure 2 shows the initial velocity of each reaction as a function
of the initial isocitrate concentration. The inverse of the initial
velocity correlates linearly with the inverse of the initial
isocitrate concentration as is predicted by the Michaelis-Menten
equation. A linear fit results in a Michaelis-Menten constant,
KM, of 3 µM and a maximum initial velocity, Vmax, of 0.17
µM/s. The strong correlation between reaction velocity and
isocitrate concentration demonstrates both the activity of the
encapsulated ICDH and the material’s ability to distinguish
between different isocitrate concentrations in a predictable
manor.
V
max [S]
Vo )
(7)
KM + [S]
For comparison, ICDH samples in buffer solution were also
evaluated at the same isocitrate concentrations and the same
saturating NADP+ concentration. Again an increase in NADPH
fluorescence was observed for each sample after the addition
Encapsulated enzymes also often follow similar kinetics.
However, if diffusion limitations are present, the characteristic
constants are apparent constants rather than intrinsic ones. The

9328 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Rickus et al.
Figure 3. Fluorescence emission spectra of thionine in buffer solution
and thionine encapsulated in aged silica gels. The samples were excited
at 590 nm.
of isocitrate. A linear fit to the Michaelis-Menten equation
resulted in a KM of 29 µM and a Vmax of 0.05 µM/s. The change
in the KM after encapsulation may be a result of diffusion
limitations, changes in the local pH or ion concentrations, or
changes in binding affinity of the enzyme for the substrate.
Because Vmax is a linear function of enzyme concentration, the
measured values must be adjusted for enzyme concentration to
compare the maximum catalytic rates, kcat, of the encapsulated
and free enzyme. The kcat for the encapsulated enzyme was 0.67
µM/s per mgenzyme while the kcat for the free enzyme was 3.8
µM/s per mgenzyme. The catalytic and binding rates are often
reduced for encapsulated enzymes compared to free enzyme in
solution.22-25 As the kinetic measurements were conducted in
bulk gel materials, it is likely that diffusion is contributing
significantly to the reduction in kcat.
Figure 4. Quenching curves for thionine by NADPH and NADH in
phosphate buffer solution. Steady-state intensity ratios for thionine in
buffer solution are plotted as a function of cofactor concentration. A
linear regression of the thionine intensities as a function of NADH
concentration resulted in a fit function of Io/I ) 1 + (2.1 ×
10-2)[NADH], r2 ) 0.99. A nonlinear regression of the thionine
intensities as a function of NADPH concentration resulted in a fit
function of Io/I ) (6.4 × 10-4)[NADPH]2 + (2.3 × 10-2)[NADPH] +
1, r2 ) 0.99.
thionine in the absence of quencher, and τ is the fluorescence
lifetime in the presence of quencher.
In the case of pure dynamic quenching, the relative reduction
in emission intensity is equal to the relative reduction in the
fluorescence lifetime, and the relationship between relative
intensity and quencher concentration is linear:
3.4. Emission Spectra of Thionine-Doped Gels. Comparison
of thionine emission spectra in solution and in the sol-gel
environment reveals almost no difference between the two
(Figure 3). The fluorescence emission peak for 20 µM thionine
in the gel is at 624 nm and is slightly red shifted to 626 nm for
40 µM thionine gels. A red shift with increasing thionine
concentration is consistent with behavior in aqueous solution
and with previous work.26 While it is known that thionine tends
to dimerize in water (Kdimer ) 4 × 103 M-1),27 there is no
discernible second peak at 665 nm, indicating that thionine
dimers are not a significant factor in the gels.26
3.5. Quenching in Solution and in Aged Gels. NADH has
been shown to quench thionine fluorescence in aqueous solution
in a reaction that results in the net production of NAD+.2,3 To
determine whether NADPH undergoes a reaction similar to that
of NADH with thionine, we examined their relative abilities to
quench thionine fluorescence. Thionine in phosphate buffer was
exposed to varying concentrations of NADPH or NADH. The
thionine fluorescence intensity was measured as a function of
time. After addition of NADPH or NADH, the thionine
fluorescence decreased quickly to a new pseudo-steady-state
value. Given enough time, however, the thionine fluorescence
returned to baseline, indicating thionine was not consumed
during the reaction, but merely quenched.
Io
I
τ
)
o ) 1 + KSV[Q]
(10)
τ
where Io is the fluorescence intensity in the absence of quencher
and I is the fluorescence intensity in the presence of quencher.
However, when static quenching mechanisms also occur
through the complexing of the ground state donor and quencher
molecules, the relative reduction in emission intensity is greater
than the relative reduction in fluorescence lifetime. In this case
a quadratic relationship between relative intensity and quencher
concentration is predicted:
I
o ) 1 + (KSV + Keq)[Q] + KeqKSV[Q]2
(11)
I
where Keq is the equilibrium constant for the binding of the
donor and quencher.
The quenching experiment for NADH and thionine in solution
was repeated three times to provide a better estimate of the
Stern-Volmer and quenching rate constants for the reaction.
The KSV was estimated to be 4.5((0.8) × 104 M-1 based on
the slopes of three quenching curves including the one shown
in Figure 4. This estimate is consistent with that previously
reported by Sharma and Arnold, 6.2((0.2) × 104 M-1 2
.
In homogeneous media, quenching can be described with
steady-state Stern-Volmer kinetics:
Figure 4 shows that Io/I for thionine increased linearly with
NADH concentration as is predicted by the Stern-Volmer
relation (eq 10) for steady-state dynamic quenching. However,
the quenching rate constant, kq, that is calculated from the
τ
o ) 1 + KSV[Q]
(8)
(9)
τ
measured Stern-Volmer constant (1.4((0.3) × 1014 M-1 s-1
,
using a thionine florescence lifetime, τo, of 320((20) × 10-12
s) is faster than what is possible for diffusion-controlled
reactions.4 The large magnitude of the estimated kq is consistent
with literature values and indicates that some level of static
quenching must also be occurring.2 Nonlinearities in the
KSV ) kqτo
where KSV is the Stern-Volmer constant (M-1), [Q] is the
concentration of quencher (M), kq is the bimolecular quenching
rate constant (M-1 s-1), τo is the fluorescence lifetime of

Photochemical Coenzyme Regeneration
J. Phys. Chem. B, Vol. 108, No. 26, 2004 9329
quencher curve, however, are not apparent for NADH. Evidence
for the formation of complexes between thionine and other
mono- and polynucleotides also supports a static quenching
process.28,29
The quenching curve for NADPH, on the other hand, is
nonlinear and is better fit by a quadratic function than a linear
one (Figure 4). This difference indicates that static mechanisms
may be playing a larger role in the quenching of thionine by
NADPH than by NADH. It is possible that the addition of the
phosphate group may change the binding constant of the
dinucleotide and thionine.
The quenching experiments were repeated for immobilized
thionine in silica gels. The shape of the quenching curves
changed after immobilization to a nonlinear relationship with a
downward curvature (Figure 5a,b). This is consistent with
observations from other immobilized systems using a hetero-
geneous matrix.30,31 Such quenching curves are typically fit with
a two-site model function (eq 12) or by a power-law function
(eq 13).30 The quencher curves for NADH and NADPH in
Figure 5 were fit by using the power-law function.
Io
I
1
)
(12)
f01
f02
+
1 + KSV1[Q] 1 + KSV2[Q]
Io
) 1 + a[Q]b
(13)
I
The two-site model assumes two discrete subpopulations of
fluorphore where the f0i values are the fraction of emission from
each subpopulation and the KSVi values are the Stern-Volmer
constants for each subpopulation.32 Each subpopulation interacts
Figure 5. Quenching curves for encapsulated thionine by NADH (A)
and NADPH (B) in silica gels. The thionine fluorescence intensity ratio
is plotted as a function of cofactor quencher concentration in aged gels.
Both plots were fit to the power function, Io/I ) 1 + a[Q]b. For NADH,
a ) 7.8 × 10-3 and b ) 0.68 with r2 ) 0.87. For NADPH, a ) 3.1
× 10-2 and b ) 0.42 with r2 ) 0.91.
differently with the matrix and therefore has a different KSV
.
This model can also be expanded to a multisite model or
distribution model that assumes a higher number of discrete
populations or a distribution of dopant populations.31 The power-
law function has no physical correlate; a and b are simply fitting
parameters. This relationship, however, can be useful in the
design and calibration of sensors despite its lack of physical
meaning.
the physical constraint of a subpopulation of molecules in
confining regions that have restricted access to the bulk solution
and therefore not participate in the electron transfer.34 In
addition, a second subpopulation of thionine molecules close
to the silica pore interface could have a reduced quenching rate
due to interactions such as electrostatic adhesion to the wall.7,33
This type of interaction could reduce both dynamic processes
through a reduction in collision rate as well as static processes
through a reduction in nucleotide binding. Although not shown
here, leaching experiments conducted in the same buffer and
ionic strength reveal that there is, in fact, a strong interaction
between the dye and the surface of the glass. Only small
amounts of thionine leach from doped gels even with lengthy
and repeated washes.
Although the quenching rates for the encapsulated system
are 1 to 2 orders of magnitude slower than those for solution,
this reduction would still result in kq values that are faster than
what is expected for diffusion-controlled reactions, indicating
that static quenching mechanisms are still playing a role after
encapsulation.
A comparison of the quenching curves for the cofactors in
solution (Figure 4) and in the gel (Figure 5) reveals that between
1 and 2 orders of magnitude more cofactor is required to achieve
the same level of quenching in the gel compared to solution.
As the dinucleotides were added after the formation of the
gels, one possible explanation for the reduction in the observed
KSV is pore exclusion of the dinucleotides. Charged molecules
can be partially excluded from the pore interior at low ionic
strengths such as those used in these experiments.33 The
quencher plots were constructed assuming that the concentration
of enzyme cofactor in the pores was equal to the concentration
in bulk. It is likely that the negatively charged enzyme cofactor
may not fully penetrate the sol-gel pores resulting in a lower
concentration than assumed. If the concentration of the co-
enzyme is lower in the gel than in solution, then the observed
quenching rates in Figure 5 represent a lower limit. This would
result in a shift in the quenching curves to higher bulk
concentrations.
3.6. NADPH Oxidation Rates. When thionine was exposed
to NADPH, its fluorescence was quenched to a constant value.
Given enough time, however, the thionine fluorescence levels
returned to initial levels suggesting that thionine was not
consumed during the quenching. The return of the thionine levels
to baseline signifies the disappearance of the quenching reaction
due to a depletion of NADPH concentration. To validate and
In addition, the matrix structure could be decreasing the
effective quenching rate by interacting with thionine in such a
way as to decrease the population that is able to interact with
cofactor. This is consistent with the multisite model, which is
based on an interpretation involving subpopulations of thionine
with reduced quenching rates. The reduction could be due to

9330 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Rickus et al.
observed reaction rate. From reaction rates alone, it is impossible
to determine the individual contributions of diffusion barriers,
pore exclusion, and multiple dopant populations.
3.7. Reaction Mechanism. While the presented oxidation
experiments address the consumption of NADPH, they do not
directly confirm the production of NADP+. However, data
reported elsewhere by us and by Sharma have verified that the
reaction product of excited thionine and NADH and NADPH
is NAD+ and NADP+, respectively, by its ability to act as an
electron acceptor in NAD+ or NADP+ specific enzyme
reactions.35-37 These results indicate that the quenching of
thionine is at least partially due to an excited-state redox process
that results in a net electron transfer from NADH or NADPH
to thionine.
Although the exact mechanism for thionine quenching by
NAD(P)H is not known, it has been suggested that short-lived
semithionine, thio(H)+, and leukothionine, thio(H2)+, intermedi-
ates are likely (reactions 16 and 17).3
Figure 6. Disappearance of NADPH during exposure to excited
thionine in buffer solution and in aged gels. NADPH levels, measured
via NADPH fluorescence intensity, are plotted as a function of exposure
time. The NADPH concentration is expressed as a percentage of the
initial NADPH concentration at time ) 0. Open symbols show the
decrease of NADPH in aged gels. Closed symbols show the decrease
of NADPH in solution. Three starting NADPH concentrations were
used for each sample type: O, aged gel, 120 µM NADPH; 0, aged
gel, 80 µM NADPH; 4, aged gel, 40 µM; b, buffer solution, 120 µM
NADPH; 9, buffer solution, 80 µM NADPH; 2, buffer solution, 40
µM NADPH. The data for the six conditions were fit individually to a
single-exponential decay, y ) 100e-bt. The decay parameter, b, for the
aged gels and the solution samples was 8.8((1.0) × 10-4 and 9.8-
((2.9) × 10-3 s-1 respectively. The r2 for each fit ranged between
0.86 and 0.99 with a mean of 0.97.
*thio+ + NADPH + H+ f thio(H2)+ + NADP+ (16)
2*thio+ + NADPH + H+ f 2thio(H)+ + NADP+ (17)
2thio(H2)+ + O2 f 2H2O + 2thio+
4thio(H)+ + O2 f 2H2O + 4thio+
(18)
(19)
Two electrons must be removed from NADPH to form NADP+.
Reaction 16 would result if both electrons were accepted by
one thionine molecule, whereas reaction 17 would result if the
two electrons were accepted by two different thionine molecules.
The samples in this study were not deoxygenated and the
suggested nonfluorescent intermediates are likely to be quickly
converted back to ground-state thionine through reaction with
oxygen in the solution (reactions 18 and 19).
characterize the consumption of NADPH, we monitored the
NADPH fluorescence levels during the quenching of thionine.
When NADPH-doped gels and NADPH solutions were exposed
to excited thionine, the NADPH fluorescence decreased with
time, but unlike thionine, the NADPH fluorescence levels did
not return to baseline. These results validate that NADPH is
consumed in the quenching reaction with thionine, as is the case
with NADH in solution.2,3
In addition, because the quenching rate of thionine is much
faster that the rate of disappearance of NADPH or NADH, we
suggest that the electron transfer between thionine and enzyme
cofactor is reversible. Only a fraction of the electron-transfer
events result in a net production of oxidized cofactor, but
thionine fluorescence is quenched each time regardless of the
presence of the back reaction.
Figure 6 shows the NADPH concentration, expressed as a
percentage of the initial NADPH concentration, over time in
the presence of excited thionine. NADPH fluorescence is linearly
proportional to concentration within the range used. The thionine
fluorescence quickly dropped to a reduced level and remained
relatively constant for the duration of the reaction. For a constant
*thio+ level, the NADPH consumption would be expected to
follow a single-exponential decay (eqs 14 and 15) with a pseudo-
first-order decay constant, koxidize, equal to the product of the
quenching constant, kq, and [*thio+]. The data were fit to an
exponential decay function and the koxidize was estimated for
each sample.
NADH and NADPH are dinucleotide cofactors with one
nicotinamide base and one adenine base. Adenine nucleotides
have been shown to quench thionine fluorescence while the site
of NADH and NADPH oxidization is on the nicotinamide
ring.5,28 It is therefore possible that the dinucleotide has two
sites of interaction with thionine and that the adenine nucleotide
can quench thionine independent from the cofactor oxidation
reaction. To investigate this possibility we have measured the
quenching rate of thionine by adenosine 5′-monphosphate
(AMP) using the same steady-state methods presented here and
have found the kq to be 3 orders of magnitude smaller than for
NADH or NADPH. This result is consistent with literature
values and suggests that the quenching by the adenine nucleotide
is secondary to the interaction at the nicotinamide site.28
3.8 Reaction Coupling. Figure 1 illustrated how the photo-
chemical oxidation of NADPH can be coupled to the enzymatic
oxidation of isocitrate. ICDH was chosen as the model dehy-
drogenase because the change in Gibbs free energy of the
reaction greatly favors the forward oxidation reaction.38 The
lack of a significant reverse reaction allows for the ICDH
reaction to go close to completion before starting the regenera-
d[NADPH]
) koxidize[NADPH(t)]
(14)
(15)
dt
[NADPH(t)] ) [NADPH(t)0)]e-k
oxidize
The rate of disappearance of NADPH is 1 order of magnitude
slower in the sol-gel, koxidize ) 8.8((1.0) × 10-4 s-1, compared
with solution, koxidize ) 9.8((2.9) × 10-3 s-1. This rate reduction
is consistent with the rate reduction seen in the steady-state
quenching experiments. In this case, however, the cofactor was
immobilized and the thionine was diffused into the gel. Thionine
is positively charged and is likely to have strong interactions
with the negatively charged silica matrix, which makes pore
exclusion a less likely contributor to reduced rates. It is possible
that diffusion restrictions are playing a role in the decrease in

Photochemical Coenzyme Regeneration
J. Phys. Chem. B, Vol. 108, No. 26, 2004 9331
Figure 7. Repeated ICDH reactions in a sol-gel monolith. ICDH, thionine, and NADP+ are encapsulated in a silica gel. Isocitrate is injected onto
the gel (at times indicated by arrows) and the reaction is monitored via NADPH fluorescence. After each reaction thionine is excited (at times
indicated by solid bars) to induce the oxidation of NADPH back to NADP+. After the NADPH levels have returned close to the baseline, isocitrate
is injected for another ICDH reaction.
tion. This enables easy cycling between the ICDH reaction and
the regeneration reaction in a closed system.
ICDH and thionine were co-encapsulated in a silica gel, which
was aged and then incubated with NADP+. The thionine
concentration was increased from previous experiments to
reduce the time required to oxidize the NADPH. Isocitrate was
injected into the sample at four different times. As seen in Figure
7, at each injection time the ICDH reaction is observed by the
production of NADPH fluorescence. The NADPH fluorescence
increases with time and levels off, at which point the ICDH
reaction is assumed to be at equilibrium. These curves are
consistent with those observed for gels doped only with ICDH.
Between each enzymatic reaction, the encapsulated thionine was
excited until the NADPH fluorescence levels returned to close
to baseline. This regeneration required about 200 s of excitation
each time.
Figure 8. Adjusted NADPH concentration as a function of cycle
number. The NADPH concentration was estimated based on the
NADPH fluorescence measurements. The concentration was adjusted
by background subtraction for each cycle. The injected isocitrate
concentration decreased with cycle number resulting in a smaller
NADPH production.
The raw data shown in Figure 7 were used to estimate the
NADPH produced after each isocitrate injection. The fluores-
cence measurements were converted to NADPH concentration
by using a linear calibration curve and the background was
subtracted from each enzyme reaction. Figure 8 shows the
minimum and maximum adjusted NADPH concentrations for
each enzyme reaction over the four cycles. Figures 7 and 8
demonstrate the ability to cycle between the enzymatic reduction
of NADP+ to NADPH by ICDH and the photochemical
oxidation of NADPH to NADP+ by thionine.
The four isocitrate injections were of decreasing isocitrate
concentration. Figure 9 shows the relationship between the
maximum NADPH produced as a function of the isocitrate
concentration. A linear relationship between NADPH production
and isocitrate concentration was observed for the investigated
range. This experimental relationship is useful for sensor
calibration. The physical meaning of this relationship is more
difficult to understand, however, because it is not clear how
well separated the two coupled reactions are in time. Overlap
of the reactions and therefore simultaneous reduction of NADP+
by ICDH and oxidation of NADPH by thionine are likely to
complicate the net NADPH profile. Further exploration of the
overlap will be required for the design of optimized continuous
sensors.
The data presented in Figures 7, 8, and 9 support the potential
use of this coupled photochemical and enzymatic system for
repeated measurements with an enzyme-doped gel without any
additional washing steps or added reagents. The coupling of
the two reactions was not explored beyond four cycles.
Optimized sensor design may require further work to understand
any potential effects of the presence of the excited thionine on
the enzyme and the enzyme kinetics.
4. Conclusions
We have demonstrated that excited thionine reacts with
NADPH resulting in thionine quenching and oxidation of the
NADPH to NADP+. In addition we have demonstrated that the
reactions can be conducted in the sol-gel environment allowing
for photochemical coenzyme regeneration in the sol-gel matrix.
The reaction rates have been characterized both in solution and
in the sol-gel monoliths. The rates are 1 to 2 orders of
magnitude slower in the bulk gel. The reasons for this reduction
are unknown, but it is possible that a transition to thin films
may improve the reaction velocities.

9332 J. Phys. Chem. B, Vol. 108, No. 26, 2004
Rickus et al.
S., Rowe-Taitt, C. A., Eds.; Elsevier: Amsterdam, The Netherlands, 2002;
p 427.
(9) Avnir, D.; Braun, S.; Lev, O.; Ottolenghi, M. Chem. Mater. 1994,
6, 1605.
(10) Gill, A.; Ballesteros, A. J. Am. Chem. Soc. 1998, 120, 8587.
(11) O’Neill, H.; Angley, C. V.; Hemery, I.; Evans, B. R.; Dai, S.;
Woodward, J. Biotechnol. Lett. 2002, 24, 783.
(12) Yamanaka, S. A.; Dunn, B.; Valentine, J. S.; Zink, J. I. J. Am.
Chem. Soc 1995, 117, 9095.
(13) Ramanathan, K.; Kamalasanan, M. N.; Malhotra, B. D.; Pradhan,
D. R.; Chandra, S. J. Sol-Gel Sci. Technol. 1997, 10, 309.
(14) Chaubey, A.; Singh, G. M.; Malhotra, V. S. Appl. Biochem.
Biotechnol. 2001, 96, 293.
(15) Li, C.-I.; Lin, Y.-H.; C.-L.; S.; Tsaur, J.-P.; Chau, L.-K. Biosens.
Bioelectron. 2002, 17, 323.
(16) Williams, A. K.; Hupp, J. T. J. Am. Chem. Soc. 1998, 20, 4366.
(17) Ramesh, P.; Sivakumar, P.; Sampath, S. J. Electroanal. Chem. 2002,
528, 82.
(18) Ellerby, L. M.; Nishida, C. R.; Nishida, F.; Yamanaka, S. A.; Dunn,
B.; Valentine, J. S.; Zink, J. I. Science 1992, 255, 1113.
(19) Miller, J. M.; Dunn, B.; Valentine, J. S.; Zink, J. I. J. Non-Cryst.
Solids 1996, 202, 279.
Figure 9. Calibration curve generated from repeated measurements,
using the same ICDH-doped gel. The maximum NADPH concentration
produced during each cycle is plotted as a function of isocitrate
concentration. The line is the linear regression of the data points.
(20) Lee, J. Immobilized Enzyme. In Biochemical Engineering; Prentice
Hall: Upper Saddle River, NJ, 1992; p 54.
(21) Shuler, M. L.; Kargi, F. Bioprocess Engineering; PTR Prentice
Hall: Englewood Cliffs, NJ, 1992.
(22) Braun, S.; Shtelzer, S.; Rappoport, S.; Avnir, D.; Ottolenghi, M. J.
Non-Cryst. Solids 1992, 147&148, 739.
(23) Rickus, J. L.; Lan, E.; Tobin, A. J.; Zink, J. I.; Dunn, B. Sol-gel
Optical Sensors for Glutamate; Materials Research Society Fall Meeting,
2000, Boston, MA.
The extension of the thionine NADH reaction to NADPH
opens the possibility of using thionine as a general coenzyme
regeneration strategy for dehydrogenases that use either or both
forms of the enzyme cofactor. The level of excited thionine
can control the reaction rate. Therefore, the pattern and intensity
of thionine excitation provides a design parameter that can be
controlled externally in a functioning sensor.
(24) Husing, N.; Reisler, E.; Zink, J. I. J. Sol-Gel Sci. Technol. 1999,
15, 57.
Acknowledgment. The authors greatly appreciate the sup-
port of their research through the following: the NIH Morris
K. Udall Center of Excellence for Parkinson’s Disease Research
(P50NS38369) and the NSF (DMR 0103952). The authors thank
Dr. Esther Lan for her help in preparing the manuscript.
(25) Wang, R.; Narang, U.; Prasad, P. N.; Bright, F. V. Anal. Chem.
1993, 65, 2671.
(26) Krihak, M.; Shahriari, M. R. Opt. Mater. 1996, 5, 301.
(27) Lai, W. C.; Dixit, N. S.; Mackay, R. A. J. Phys. Chem. 1984, 88,
5364.
(28) Tuite, E.; Kelly, J. M.; Beddard, G. S.; Reid, G. S. Chem. Phys
Lett. 1994, 226, 517.
References and Notes
(29) Tuite, E.; Kelly, J. M. Biopolymers 1995, 35, 419.
(30) Carraway, E. R.; Demas, J. N.; DeGraff, B. A.; Bacon, J. R. Anal.
Chem. 1991, 63, 337.
(31) Draxler, S.; Lippitsch, M. E.; Klimant, I.; Kraus, H.; Wolfbeis, O.
S. J. Phys. Chem 1995, 99, 3162.
(32) Demas, J. N.; DeGraff, B. A.; Xu, W. Anal. Chem 1995, 67, 1377.
(33) Shen, C.; Kostic, N. M. J. Am. Chem. Soc. 1997, 119, 1304.
(34) Castellano, F. N.; Meyer, G. J. J. Phys. Chem 1995, 99, 14742.
(35) Sharma, A.; Quantrill, N. S. M. Spectrochim. Acta 1994, 50A, 1179.
(36) Sharma, A.; Quantrill, N. S. M. Spectrochim. Acta 1994, 50A, 1161.
(37) Sharma, A.; Quantrill, N. S. M. Biotechnol. Prog. 1996, 12, 413.
(38) Stryer, L. Biochemistry, 3rd ed.; W. H. Freeman and Company:
New York, 1988.
(1) Arnold, M. Anal. Chem. 1985, 57, 565.
(2) Sharma, A. Spectrochim. Acta 1992, 48, 647.
(3) Sharma, A. Spectrochim. Acta 1992, 48A, 893.
(4) Archer, M. D.; Ferreira, M. I. C.; Porter, G.; Tredwell, C. J. NouV.
J. Chim. 1977, 1, 9.
(5) Lehninger, A. L.; Nelson, D. L.; Cox, M. M. Principles of
Biochemistry, 2nd ed.; Worth Publishers: New York, 1993.
(6) Dave, B. C.; Dunn, B.; Valentine, J. S.; Zink, J. I. Anal. Chem.
1994, 66, 1120A.
(7) Dunn, B.; Zink, J. I. Chem. Mater. 1997, 9, 2280.
(8) Rickus, J. L.; Dunn, B.; Zink, J. I. Optically Based Sol-Gel
Biosensor Materials. In Optical Biosensors: Present and Future; Ligler, F.