Kinetic and transport analysis of immobilized oxidoreductases that oxidize glycerol and its oxidation products

Article abstract of DOI:10.1016/j.electacta.2009.09.083

Glycerol has drawn increasing attention as a possible fuel, because it has many desirable qualities and is abundant due to the fact that it is a byproduct of biodiesel production. Previous research has shown that non-natural enzyme cascades can be used to create a bioanode that can stepwise oxidize glycerol to carbon dioxide. Two of these enzymes are pyrroloquinoline quinone (PQQ) dependant alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (AldDH) derived from Gluconobacter. The third enzyme, which is responsible for carbon bond cleavage, is oxalate oxidase (OxOx) derived from barley. Previous research has shown that all three enzymes have demonstrated the ability to undergo direct electron transfer to a carbon electrode which allows for a simple and efficient bioanode that completely oxidizes glycerol. In this study, each enzyme was individually immobilized within modified Nafion on a glassy carbon rotating disc electrode (GC-RDE) and voltammetric analysis was performed employing different rotation rates in a solution containing each enzyme's respective substrate. This substrate was glycerol for alcohol dehydrogenase, glyceraldehyde for aldehyde dehydrogenase, and mesoxalic acid for oxalate oxidase. From the voltammograms, Levich plots were produced and the solution diffusion coefficient (D_soln), the membrane diffusion coefficient (D_film), k_CAT, K_M, and V_MAX were determined.

Full text of DOI:10.1016/j.electacta.2009.09.083

Electrochimica Acta 55 (2010) 7679–7682
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Kinetic and transport analysis of immobilized oxidoreductases that oxidize
glycerol and its oxidation products
Robert L. Arechederra, Shelley D. Minteer∗
Department of Chemistry, Saint Louis University, 3501 Laclede Ave., St. Louis, MO 63103, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 8 June 2009
Received in revised form
28 September 2009
Accepted 29 September 2009
Available online 7 October 2009
Glycerol has drawn increasing attention as a possible fuel, because it has many desirable qualities and
is abundant due to the fact that it is a byproduct of biodiesel production. Previous research has shown
that non-natural enzyme cascades can be used to create a bioanode that can stepwise oxidize glycerol
to carbon dioxide. Two of these enzymes are pyrroloquinoline quinone (PQQ) dependant alcohol dehy-
drogenase (ADH) and aldehyde dehydrogenase (AldDH) derived from Gluconobacter. The third enzyme,
which is responsible for carbon bond cleavage, is oxalate oxidase (OxOx) derived from barley. Previous
research has shown that all three enzymes have demonstrated the ability to undergo direct electron
transfer to a carbon electrode which allows for a simple and efficient bioanode that completely oxi-
dizes glycerol. In this study, each enzyme was individually immobilized within modified Nafion® on
a glassy carbon rotating disc electrode (GC-RDE) and voltammetric analysis was performed employ-
ing different rotation rates in a solution containing each enzyme’s respective substrate. This substrate
was glycerol for alcohol dehydrogenase, glyceraldehyde for aldehyde dehydrogenase, and mesoxalic
acid for oxalate oxidase. From the voltammograms, Levich plots were produced and the solution diffu-
sion coefficient (Dsoln), the membrane diffusion coefficient (Dfilm), kCAT, KM, and VMAX were determined.
Keywords:
Biofuel cell
Glycerol
Bioanode
Rotating disc electrochemistry
Direct electron transfer
© 2009 Elsevier Ltd. All rights reserved.
1. Introduction
In previous work, we have developed a three-enzyme cascade
on an electrode that stepwise oxidizes glycerol to carbon dioxide
Enzymatic biofuel cells which use enzymes as catalysts for elec-
trode reactions have shown great potential as an alternative to
traditional precious metal catalyzed fuel cells [1]. Since there are
numerous oxidoreductase enzymes that have the ability to oxi-
dize a wide variety of substrates, the limitations of fuel choices
that are typically imposed by precious metal catalysts are more
easily overcome [2–14]. For example, room temperature precious
metal catalysts and their alloys are typically limited to methanol,
methane, or hydrogen if complete oxidation is to occur [15–18]. Any
fuel molecule that is more complex such as ethanol, propanol, or
glycerol can cause catalyst passivation and is incompletely oxidized
meaning that the oxidation products are typically the aldehyde or
carboxylic acid of the parent molecule [19–21]. The fact that more
complex and more energy dense fuels cannot be efficiently utilized
with precious metals and their alloys limits the usefulness of such
catalysts in applications where low temperature and high energy
density are major issues [22–25].
[4,5]. The motivation for developing this three-enzyme system for
a bioanode was because glycerol is an abundant fuel source due
to it being the major byproduct of the large quantities of biodiesel
that are produced, it has a high energy density, and it can be used
at high concentrations, because it does not have the solvent prop-
erties of methanol and ethanol. This three-enzyme cascade is not
the natural metabolic process for glycerol oxidation in living cells,
which involves the glycolytic pathway and the Kreb’s cycle, but
instead is a simpler enzyme cascade that employs the ability of
PQQ-dependent dehydrogenases and oxalate oxidase to oxidize
non-natural substrates.
In order to better understand these enzyme-modified electrodes
and the effect of non-natural substrates on the electrochem-
istry, rotating disk electrochemistry was performed with each
immobilized enzyme in the modified Nafion® membranes that
have been shown to have the highest activity with oxidoreduc-
tase enzymes [26]. These two polymers are tetrabutylammonium
bromide modified Nafion® (TBAB) and triethylhexylammonium
bromide modified Nafion® (TEHA). This allowed for the eval-
uation of mass transport and kinetics in these enzymatic
bioanodes and the comparison to the literature informa-
tion.
∗
Corresponding author. Tel.: +1 314 977 3624; fax: +1 314 977 2521.
E-mail address: minteers@slu.edu (S.D. Minteer).
0013-4686/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2009.09.083

7680
R.L. Arechederra, S.D. Minteer / Electrochimica Acta 55 (2010) 7679–7682
Table 1
Comparison of the moles of loaded enzyme onto the electrode versus the moles of the loaded enzyme that are electrochemical accessible to the carbon electrode.
Enzyme
Polymer
Cast (moles of enzyme)
Accessible (moles of enzyme)
% Accessible enzyme
PQQ-ADH
PQQ-ADH
PQQ-AldDH
PQQ-AldDH
OxOx
TBAB
TEHA
TBAB
TEHA
TBAB
TEHA
5.43 × 10−13
5.43 × 10−13
3.33 × 10−12
3.33 × 10−12
2.35 × 10−10
2.35 × 10−10
2.97 0.38 × 10−13
2.18 0.14 × 10−13
4.11 0.22 × 10−14
3.07 1.01 × 10−14
1.11 0.03 × 10−13
1.28 0.12 × 10−13
54.69
40.14
1.23
0.92
0.05
0.05
OxOx
2. Experimental
the substrate for PQQ-AldDH was glyceraldehyde, and the substrate
for OxOx was mesoxalic acid. The electrodes were allowed to soak
in their respective substrate solutions for at least 1 h prior to being
rotated to ensure that they were equilibrated with the substrate
solution. Each electrode was tested with a platinum mesh counter
and a saturated calomel reference electrode. The rotator was a Pine
Instruments model AFM-SRX. Rotation rates were varied and cyclic
voltammetry was performed at each rotation rate. All experiments
were performed in at least triplicate and the uncertainties reported
correspond to one standard deviation.
2.1. Materials
Glycerol (Sigma), oxalate oxidase (0.71 U/mg from barley,
Sigma), mesoxalic acid (Sigma), sodium phosphate monoba-
sic (Sigma), sodium nitrate (Sigma), sodium hydroxide (Sigma),
tetrabutylammonium bromide (Sigma), triethylhexylammonium
bromide (Sigma), Nafion® 1100EW suspension (Aldrich), ethanol
(Sigma), and glyceraldehyde (Sigma) were used as received. PQQ-
ADH and AldDH were obtained from Gluconobacter sp.33 which
was grown, extracted, and purified according to the protocol pre-
viously established [5,27]. Modified Nafion® was also prepared by
a procedure that has been previously established [28].
3. Results and discussion
Previous research on glycerol bioanodes for biofuel cells
employed multi-enzyme cascades where multiple reactions were
occurring. However, in order to be able to separate the effects of
each enzyme on the system, in this study, we studied the oxida-
tion of one susbstrate at an electrode containing one immobilized
enzyme. The information determined in this study should provide
information about which enzyme/substrate system is most limiting
in the glycerol bioanode.
The amount of electrochemically accessible enzyme was deter-
mined from the cyclic voltammogram for each respective enzyme
immobilized on the glassy carbon electrode in pH 7.15 phosphate
buffer with no added substrate. Both PQQ-dependent enzymes
undergo a 6e-process from the hemes and the PQQ when there
is no substrate present [29]. Oxalate oxidase undergoes a 2e-
process when substrate is not present [30]. The amount of enzyme
cast and the amount of enzyme electrochemically accessible for
each is shown in Table 1. As can be seen in Table 1, PQQ-ADH
has the highest percentage of immobilized protein that is elec-
trochemically accessible and OxOx has the lowest percentage of
immobilized protein that is electrochemical accessible. This will
be important in further optimizations of glycerol bioanodes to
ensure optimal metabolic flux of glycerol oxidation and minimize
2.2. Electrode fabrication
Enzyme-modified rotating disk electrodes were fabricated by
mixing 1.0 mg of one particular enzyme with 100 ␮l of 18 Mꢀ
water in a microcentrifuge tube. In a separate microcentrifuge
tube, 100 ␮l of either TBAB or TEHA modified Nafion® was added,
followed by the addition of 50 ␮l of the enzyme solution. The poly-
mer/enzyme solution was mixed for 30 s with a vortex mixer on the
highest setting. Then, 20 ␮l of the polymer enzyme solution was
cast onto each glassy carbon rotating disk electrode (5 mm diame-
ter, Pine Instruments) and spread evenly across the surface of the
glassy carbon to ensure a complete coating. Once the solution was
cast, the rotating disk electrodes were allowed to dry for 3 h at room
temperature. Then, the electrodes were soaked overnight in pH 7.15
10 mM phosphate buffer with 100 mM sodium nitrate electrolyte
to fully equilibrate the electrodes. The membranes on the rotating
disk electrodes were measured with a profilometer to determine
their thickness.
2.3. Cyclic voltammetry
After equilibration, cyclic voltammetry was performed on each
electrode in the soaking buffer solution with a platinum mesh
counter electrode and a saturated calomel reference electrode to
determine the amount of enzyme that was electrochemically acces-
sible. For PQQ-ADH and PQQ-AldDH modified electrodes, the scan
rate was 0.5 V s−1 and the potential window was −0.8 V to +0.6 V.
For OxOx modified electrodes, the scan rate was 0.5 V s−1 and the
potential window was 0.0 V to +0.8 V. Electrochemical measure-
ments were taken with a CH Instruments model 620A potentiostat
interfaced to a PC. All experiments were performed in triplicate and
the uncertainties reported correspond to one standard deviation.
2.4. Rotating disk electrochemistry
Rotating disc voltammetry was performed on each electrode
in 100 ␮M substrate solution that contained 100 mM NaNO3 elec-
trolyte and pH 7.15 10 mM phosphate buffer. The substrates were
selected based on the oxidation of glycerol and its products that
have been determined previously by carbon-13 NMR intermediate
and product analysis [4]. The substrate for PQQ-ADH was glycerol,
Fig. 1. Representative rotating disk voltammograms for oxalate oxidase enzyme
immobilized in TEHA modified Nafion® at a glassy carbon rotating disk electrode in
0.1 mM mesoxalic acid for a variety of rotation rates.

R.L. Arechederra, S.D. Minteer / Electrochimica Acta 55 (2010) 7679–7682
7681
Table 4
Kinetic values for each enzyme in each polymer determined from RDE experiments.
Enzyme
Polymer
VMAX
(␮M min−1 mg−1
kCAT (s−1
)
KM (mM)
)
PQQ-ADH
PQQ-ADH
PQQ-AldDH
PQQ-AldDH
OxOx
TBAB
TEHA
TBAB
TEHA
TBAB
TEHA
44.8 0.6
30.1 0.4
438 10
558 30
15.8 0.2
14.5 0.5
105
70
1023
1302 68
34
31
1
1
2
0.025 0.002
0.011 0.001
0.068 0.063
0.028 0.001
0.010 0.001
0.003 0.001
1
1
OxOx
plot and the Levich equation was used:
0.62nFAD2/3
ω
C0
1/2
soln
slope = −
ꢁ1/6
where n is the number of electron transfer, F is Faraday’s constant,
A is the area of the electrode, ꢁ is the kinematic viscosity, and C0 is
the concentration of substrate in solution. The diffusion coefficient
in the film (Dfilm) was calculated from the intercept of the Levich
plot with:
Fig. 2. Levich plots of limiting current of substrate at each enzyme immobilized in
each polymer combination.
Table 2
ꢀ
ꢁ
Levich slopes and intercepts for each enzyme/polymer combination in 0.1 mM
substrate.
DfilmÄcA0
ϕ
intercept = nFA
Enzyme
Polymer Levich slope (A cm−2 s1/2 Levich intercept (A cm−2
)
PQQ-ADH
PQQ-ADH
PQQ-AldDH TBAB
PQQ-AldDH TEHA
OxOx
OxOx
TBAB
TEHA
2.59 0.16 × 10−7
1.51 0.10 × 10−7
4.89 4.58 × 10−8
3.93 0.20 × 10−7
4.81 0.67 × 10−8
8.21 1.07 × 10−8
5.57 0.34 × 10−6
2.60 0.17 × 10−6
7.52 6.99 × 10−7
6.04 0.31 × 10−6
6.05 0.85 × 10−7
5.071 0.66 × 10−7
where k is the partition coefficient, which is assumed to be 1, and
the film thickness (ϕ) was determined with profilometry to be
10.0 0.2 ␮m. The diffusion coefficients in the film and in solution
are shown in Table 3. Although diffusion coefficients in aqueous
solutions are not known for glyceraldehyde and mesoxalic acid,
they are known for glycerol (1.1 × 10−5 cm2 s−1) [31]. The diffusion
of the substrate molecules in solution determined from the RDE
experiments is in general agreement with the literature values [31]
for similar compounds at infinite dilution which is a good indication
that the experiments were successful and the values determined
for other constants are accurate. The diffusion of the substrates
through the modified Nafion® membranes was approximately two
orders of magnitude slower than in solution. Similar phenomena is
observed with the diffusion of ethanol or water in solution verses
their diffusion through a Nafion® membrane [32].
TBAB
TEHA
the enzymatic bottlenecks. The differences in electrochemically
accessible enzyme for PQQ-ADH, PQQ-AldDH, and OxOx are due
to a number of factors including chemical effects (charge and
hydrophobicity) that result in orientation differences on carbon
electrode surfaces as well as structural differences between the
enzymes which affect their ability to undergo direct electron trans-
fer with the electrode. It is also important to note that although
TBAB modified Nafion® membranes provide higher percentage of
available enzyme that is electrochemical accessible for both of the
dehydrogenase enzymes, TEHA modified Nafion® membranes pro-
vided higher percentage of electrochemical accessible enzyme for
oxalate oxidase.
Limiting currents were determined for each of the rotating
disc experiments at each rotation rate. Note: rotation rates were
kept low to ensure stability of the modification layer, but this
increases the uncertainty in these measurements. Representative
voltammograms are shown in Fig. 1 for several rotation rates. Both
Koutecky–Levich and Levich plots were graphed for each enzyme
and polymer combination. Koutecky–Levich plots were not linear.
However, the Levich plots were linear with non-zero intercepts.
The Levich plot which is the limiting current versus the square
root of the rotation rate (w) is shown in Fig. 2. From this data, the
slope and the intercept for the Levich plot was determined for each
enzyme/polymer combination, which is shown in Table 2. To calcu-
late the solution diffusion coefficient (Dsoln), the slope of the Levich
Kinetic evaluation of the bioelectrodes was also performed. KM
,
kCAT, and VMAX were determined by fitting a line to the plot of
current versus rotation rate and then extrapolating to the value
where it approached its maximum [33,34]. This maximum cur-
rent is defined as the catalytic current (iCAT) which can be directly
converted into kCAT and VMAX, if the amount of electrochemically
accessible enzyme is known and the mass of the enzyme. Also,
KM can be calculated from this plot by determining the current at
½VMAX and then solving the Levich equation for the effective con-
centration at that current, which is KM. The values determined from
the RDE experiments for the enzyme kinetics are shown in Table 4.
Comparing the electrochemical KM, kCAT, and VMAX to the lit-
erature values for free enzyme in solution is more difficult, since
the literature values are for the natural substrate and, in this sys-
tem, we were studying non-natural substrates for each of the
enzymes. PQQ-dependent enzymes are known for their broad sub-
strate lack of specificity, so it is not surprising that the VMAX values
Table 3
Diffusion coefficients determined for solution (Dsoln) and membrane (Dfilm) for each substrate/polymer/enzyme combination.
Enzyme/substrate
Polymer
Dfilm (cm2 s−1
)
Dsoln (cm2 s−1
)
PQQ-ADH/glycerol
PQQ-ADH/glycerol
PQQ-AldDH/glyceraldehyde
PQQ-AldDH/glyceraldehyde
OxOx/mesoxalic acid
OxOx/mesoxalic acid
TBAB
TEHA
TBAB
TEHA
TBAB
TEHA
2.9 0.2 × 10−7
1.4 0.1 × 10−7
3.9 3.6 × 10−8
3.1 0.2 × 10−7
3.1 0.4 × 10−8
2.6 0.3 × 10−8
2.3 0.2 × 10−5
1.0 0.1 × 10−5
1.9 2.6 × 10−6
4.2 0.3 × 10−5
1.8 0.4 × 10−6
4.0 0.8 × 10−6

7682
R.L. Arechederra, S.D. Minteer / Electrochimica Acta 55 (2010) 7679–7682
determined for glycerol and glyceraldehyde are consistent with
the literature values for ethanol at PQQ-ADH (52 ␮mole/min/mg)
[35] and acetaldehyde at PQQ-AldDH (430 ␮mole/min/mg) [36].
Oxalate oxidase shows a VMAX for oxalate oxidase in TBAB and
TEHA modified Nafion® membranes of 15.8 ␮mole/min/mg and
14.2 ␮mole/min/mg, respectively, which is lower than the VMAX
for oxalate at oxalate oxidase in solution (34 ␮mole/min/mg) [30].
This shows that the kinetics of oxalate oxidase are different for
non-natural substrates. KM values were determined for each of the
enzymes in both polymers. The literature value for KM for the natu-
ral substrate (ethanol) and PQQ-ADH enzyme is 34 mM [35]. The KM
value for acetaldehyde substrate and PQQ-AldDH enzyme has been
shown to be 3.3 mM in the literature [36]. The literature value for
KM for the natural substrate oxalic acid and oxalate oxidase enzyme
is 0.78 mM [30]. The KM values determined for all three immobi-
lized enzyme with their non-natural substrates were lower than
the values reported in the literature for free enzyme in solution
indicating that the KM for the non-natural substrate is different for
these enzymes immobilized on the electrodes. This is expected due
to the fact that the substrate binding for a non-natural substrate is
expected to be different than the natural substrate.
References
[1] G.T.R. Palmore, Trend. Biotechnol. 22 (2004) 99.
[2] C.M. Halliwell, E. Simon, C. Toh, A.E.G. Cass, P.N. Bartlett, Bioelectrochemistry
55 (2002) 21.
[3] R.L. Arechederra, S.D. Minteer, Electrochim. Acta 53 (2008) 6698.
[4] R.L. Arechederra, S.D. Minteer, Fuel Cells 9 (2009) 63.
[5] R.L. Arechederra, B.L. Treu, S.D. Minteer, J. Power Sources 173 (2007)
156.
[6] M. Germain, R.L. Arechederra, S.D. Minteer, J. Am. Chem. Soc. 130 (2008)
15272.
[7] G.T.R. Palmore, H. Bertschy, S.H. Bergens, G.M. Whitesides, J. Electroanal. Chem.
443 (1998) 155.
[8] S.C. Barton, H. Kim, G. Binyamin, Y. Zhang, A. Heller, J. Phys. Chem. B 105 (2001)
11917.
[9] A. Heller, Curr. Opin. Chem. Biol. 10 (2006) 664.
[10] N. Mano, F. Mao, A. Heller, J. Am. Chem. Soc. 125 (2003) 6588.
[11] V. Soukharev, N. Mano, A. Heller, J. Am. Chem. Soc. 126 (2004) 8368.
[12] I. Willner, V. Heleg-Shabtai, R. Blonder, E. Katz, G. Tao, A.F. Buckmann, A. Heller,
J. Am. Chem. Soc. 118 (1996) 10321.
[13] Y. Kamitaka, S. Tsujimura, N. Setoyama, T. Kajino, K. Kano, Phys. Chem. Chem.
Phys. 9 (2007) 1793.
[14] B.L. Treu, R.L. Arechederra, S.D. Minteer, J. Nanosci. Nanotechnol. 9 (2009)
2374.
[15] A.B. Anderson, E. Grantscharova, S. Seong, J. Electrochem. Soc. 143 (1996)
2075.
[16] D.R. Rolison, P.L. Hagans, K.E. Swider, J.W. Long, Langmuir 15 (1999) 774.
[17] C. Roth, N. Benker, S. Zils, R. Chenitz, A. Issanin, H. Fuess, Oldenbourg 221 (2007)
1549.
4. Conclusions
[18] H. Igarashi, T. Fujino, Y. Zhu, H. Uchida, M. Watanabe, Phys. Chem. Chem. Phys.
3 (2001) 306.
[19] G. Yildiz, F. Kadirgan, J. Electrochem. Soc. 141 (1994) 723.
[20] M.L. Avramov-Ivic, J.M. Leger, C. Lamy, V.D. Jovic, S.D. Petrovic, J. Electroanal.
Chem. 308 (1991) 309.
[21] A. Kahyaoglu, B. Beden, C. Lamy, Electrochim. Acta 29 (1984) 1489.
[22] S.D. Minteer, in: S.D. Minteer (Ed.), Alcoholic Fuels, CRC Press, New York, 2006,
p. 191.
[23] F. Vigier, C. Coutanceau, A. Perrard, E.M. Belgsir, C. Lamy, J. Appl. Chem. 34
(2004) 439.
[24] W. Zhou, Catal. B: Environ. 46 (2003) 273.
[25] S. Song, L. Chen, J. Lie, Z. Wei, Q. Xin, Dianhuaxue 8 (2002) 105.
[26] S. Topcagic, B.L. Treu, S.D. Minteer, Proc.-Electrochem. Soc. 2004–2018 (2005)
230.
[27] B.L. Treu, D.S. Minteer, Bioelectrochemistry 74 (2008) 73.
[28] T.L. Klotzbach, M.M. Watt, Y. Ansari, S.D. Minteer, J. Membr. Sci. 282 (2006)
276.
[29] M. Torimura, K. Kano, T. Ikeda, T. Ueda, Chem. Lett. 6 (1997) 525.
[30] M.M. Whittaker, H. Pan, E.T. Yukl, J.W. Whittaker, J. Biol. Chem. 282 (2007)
7011.
[31] CRC Handbook of Chemistry and Physics, CRC Press, Inc., Boca Raton, Florida,
2006.
[32] P.T. Inglefield, X. Gong, A. Bandis, A. Tao, G. Meresi, Y. Wang, A.A. Jones, W. Wen,
Polymer 42 (2001) 6485.
Overall, the data determined from these electrochemical experi-
ments will be useful for further development of the glycerol biofuel
cell anodes and yield insight into other immobilized biocatalytic
systems. The PQQ-dependant enzymes when in direct electrical
communication with the carbon electrode exhibit activities that
are similar to enzyme in solution with the natural substrate indi-
cating the enzyme kinetics are mostly unchanged. However, the
enzyme kinetics of oxalate oxidase are different when in direct
electrical communication with a carbon electrode and employing
a non-natural substrate (mesoxalic acid) than for free enzyme in
solution with the natural substrate oxalic acid. The main limiting
factor for this system of immobilization is transport through the
modified Nafion® membrane which is two orders of magnitude
lower than that in solution. In order to enhance current density,
future work will focus on improving mass transport within the
immobilization layer while also engineering the enzymes to handle
non-natural substrates more efficiently.
[33] V.K. Pardo-Yissar, E.I. Willner, A.B. Kotlyar, C. Sandersc, H. Lillc, Faraday Disc
116 (2000) 119.
Acknowledgements
[34] N. Mano, H.-H. Kim, Y. Zhang, A. Heller, J. Am. Chem. Soc. 124 (2002)
6480.
[35] M. Ameyama, E. Shinagawa, K. Matsushita, O. Adachi, Agric. Biol. Chem. 49
(1985) 699.
[36] M. Ameyama, O. Adachi, Methods Enzymol. 89 (1982) 491.
The authors would like to thank the United Soybean Board and
the Air Force Office of Scientific Research for generous funding of
this work.