Complete oxidation of methanol in biobattery devices using a hydrogel created from three modified dehydrogenases

Article abstract of DOI:10.1002/anie.201207423

Enzyme catalysis: Three dehydrogenases have been engineered to self-assemble into a hydrogel that supports a synthetic metabolic network. The new catalytic biomaterial was used as an anode modification in two enzymatic biobatteries capable of the complete oxidation of methanol to CO₂ (see picture). Copyright

Full text of DOI:10.1002/anie.201207423

Angewandte
Chemie
DOI: 10.1002/anie.201207423
Protein Engineering
Complete Oxidation of Methanol in Biobattery Devices Using
a Hydrogel Created from Three Modified Dehydrogenases**
Yang Hee Kim, Elliot Campbell, Jiang Yu, Shelley D. Minteer, and Scott Banta*
Protein engineering involves the manipulation of amino acids
to improve the properties of proteins. Breakthroughs are still
being reported in the design and improvement of enzymes as
well as efforts to improve structural proteins for biomaterials
applications. Various protein and peptide domains have been
engineered to create new functional materials for a variety of
Here we extend this approach to create an enzymatic
hydrogel that supports a functional synthetic metabolic
pathway. Three NAD(H)-dependent dehydrogenase enzymes
from different sources were modified for self-assembly. The
first enzyme was a tetrameric alcohol dehydrogenase (ADH)
from Bacillus stearothermophilus which oxidizes methanol to
[
1]
[5]
applications. Here we report an advancement of this
approach where we create a new catalytic biomaterial by
engineering of three dehydrogenase enzymes for self-assem-
bly. When combined, the resulting new catalytic biomaterial is
able to fully oxidize methanol to carbon dioxide and we
demonstrate the application of this material as an anode
formaldehyde. The second enzyme was a tetrameric human
aldehyde dehydrogenase (ALDH2) which oxidizes formal-
[
6]
dehyde to formate. The final enzyme, a dimeric formate
dehydrogenase (FDH1) from Saccharomyces cerevisiae, oxi-
[7]
dizes formate to CO2. When combined these enzymes
produce a synthetic metabolic pathway capable of the
[2]
[8]
modification in two types of enzymatic biobattery devices.
complete oxidation of methanol. A schematic diagram of
Hydrogels can be created from proteins and peptides by
outfitting them with cross-linking domains. Pioneering work
by Tirrell and co-workers demonstrated that alpha-helical
leucine zipper domains could be used to create peptides that
self-assemble into hydrogels through coiled-coil interac-
this reaction is as shown in Figure 1a.
An alpha-helical leucine zipper domain (H) and randomly
structured soluble peptide domain (S) were genetically
appended to the N-termini of each of the three dehydrogen-
ase genes. The three new bifunctional enzyme constructs
(HSADH, HSALDH2, and HSFDH1) were overexpressed in
E. coli and purified as described in the Supporting Informa-
tion. HSADH and HSFDH1 were readily expressed and
purified, while the HSALDH2 enzyme required the addition
of the maltose binding protein (MBP) to enable functional
expression. An intein domain was added between the MBP
and HSALDH2 such that it spontaneously cleaved after
expression within the cells and thus the HSALDH2 protein
could be purified as though no fusion protein had been
[
3]
tions, and we have expanded on this line of research by
demonstrating that these domains can be appended to
[
4]
globular proteins. These hydrogel constructs are cross-
linked through both the coiled-coil motifs formed by the
appended leucine zipper domains and through additional
protein/protein interactions because of the quaternary struc-
ture of the proteins. So far, we have described the addition of
[4a]
helical appendages to fluorescent proteins, a thermostable
[
4b]
alcohol dehydrogenase,
an organophosphate hydrolase
[
4c]
[4d]
[9]
enzyme, and a small laccase enzyme.
When the latter
included.
enzyme was combined with osmium-modified peptides, a bio-
electrocatalytic hydrogel was formed that could reduce
oxygen to water and could function as a cathode modification
for a biobattery or enzymatic biofuel cell. In almost every
case, the addition of the helical appendages has had a minimal
impact on the catalytic activity of the enzymes, and robust
hydrogels have been demonstrated.
The kinetics of the purified bifunctionalized enzymes, all
of which follow the ordered bi-bi kinetic mechanism, were
measured in dilute solution to determine the impact of
modifications on the kinetic parameters (Table 1 and Fig-
ure S3 in the Supporting Information). The kinetic parame-
ters of the HSALDH2 enzyme were similar to those reported
in the literature, while unexpectedly the kinetic parameters of
both the HSADH and HSFDH1 enzymes were both found to
be improved by the addition of the helical appendages. Both
modified enzymes showed significant increases in catalytic
efficiency (k /K ) as compared to the published values for
[4d]
[*] Dr. Y. H. Kim, Dr. E. Campbell, Prof. S. Banta
Department of Chemical Engineering, Columbia University
500 West 120th Street, New York, NY 10027 (USA)
cat
m
E-mail: sbanta@columbia.edu
Homepage: http://www.columbia.edu/~sb2373
J. Yu, Prof. S. D. Minteer
Department of Chemistry and Materials Science and
Engineering, University of Utah
the wild-type enzymes. The Michaelis constant (K ) for the
m
substrate of HSADH was three orders of magnitude smaller
than reported for the unmodified enzyme while the kcat value
was found to increase by 120-fold. As a result, the catalytic
efficiency (k /K ) of HSADH was increased six orders of
cat
m
315 S 1400 E, Salt Lake City, UT 84112 (USA)
magnitude. The catalytic efficiency of HSFDH1 was found to
be increased by two orders of magnitude compared to
[
**] This work was funded by the NSF and the AFOSR. We thank Prof.
Prather at MIT for the FDH1 plasmid peAF and Prof. Wood at
Princeton University for the intein gene contained in plasmid
pE1OPD. We also thank Hoang D. Lu for technical assistance.
literature values. The change in K for the substrate was not
m
significantly different, but the kcat value was two orders of
magnitude higher than for the unmodified enzyme. We have
previously observed that the addition of the helical appen-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201207423.
Angew. Chem. Int. Ed. 2013, 52, 1437 –1440
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1437

.
Angewandte
Communications
Figure 1. a) Schematic diagram of the methanol/O enzymatic bio-
2
battery. b) Left: Schematic representation of the physically cross-linked
enzyme hydrogel (red: ADH, purple: ALDH2, and green: FDH1).
Right: Schematic representation of the protein modification where an
a helix and random coil are added to the N-terminus of each
monomer. c) The methanol/Prussian blue biobattery. d) The methanol/
O enzymatic biobattery developed to use the self-assembled protein
hydrogel on the anode.
2
Figure 2. Representative polarization (a) and power curves (b) for the
type 1 biobattery and polarization curve (c) for the type 2 biobattery
when different fuels are added to the anode. Anode and cathode
dages to both ends of a different alcohol dehydrogenase
enzyme resulted in almost no impact on the kinetic param-
2
electrode areas for the type 1 biobattery were 0.785 cm and the scan
ꢀ1
rate was 1 mVs . Anode and cathode areas for the type 2 biobattery
[
4b]
2
ꢀ1
eters. We have also observed that the addition of an N-
terminal helical appendage can dramatically improve the
functional expression of an organophosphate hydrolase
were 0.125 cm and the scan rate was 100 mVs .
shown to be able to form a hydrogel (3.4 mm protein in
100 mm phosphate buffer pH 7.0) as assessed by an inversion
test as well as by the ability of the material to retain its shape
(Figure S5 in the Supporting Information). The hydrogels
formed by the three enzymes as well as a hydrogels formed by
a mixture of the three enzymes were very similar to the
hydrogels formed by all other globular proteins we have
previously investigated. Since all three enzymes are multi-
[
4c]
enzyme.
A similar increase in specific activity could
partially explain the improvements in the kcat values observed,
as this modification could result in an increase in the amount
of active enzymes that are purified. However, further experi-
ments will be needed to prove this hypothesis.
Each of the newly bifunctionalized enzymes was purified,
concentrated, (Figure S5 in the Supporting Information) and
1
438
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1437 –1440

Angewandte
Chemie
Table 1: Kinetic parameters of hydrogel-forming enzymes for the oxidation of methanol (HSADH),
formaldehyde (HSALDH), and formate (HSFDH).
The type 2 biobattery was con-
structed to take advantage of the
hydrogel-forming ability of the
engineered proteins (Table 3). To
maintain the hydrogel state of the
proteins (25 wt% protein), only
a small amount of liquid fuel was
added to the hydrogel, and a fast
ꢀ1
+
+
kcat [s
]
Km,S [mm]
Km,NAD [mm]
k ,
i NAD
[mm]
[
a]
ꢀ3
ꢀ3
ꢀ3
HSADH
ADH
7.2ꢁ0.5
(12ꢁ10)ꢀ10
(7.6ꢁ2.6)ꢀ10
(69.5ꢁ72.8)ꢀ10
[
b]
0.06
20
–
–
[
a]
HSALDH2
58.3ꢁ19.8
67.5ꢁ8.3
21.8
21.7ꢁ12.1
0.32ꢁ0.08
0.42ꢁ0.08
0.14ꢁ0.18
0.21ꢁ0.14
[
c]
ALDH2
–
–
–
–
[
d]
ALDH2
ꢀ1
scan rate (100 mVs ) was required
as the high loading of the enzymes
in the anode caused the fuel to be
consumed rapidly. As was seen in
the type 1 biobattery design, the
polarization curves (Figure 2c)
[
a]
HSFDH1
FDH1
1670ꢁ80
6.5ꢁ0.4
7.2ꢁ1.5
5.5ꢁ0.3
0.8ꢁ0.1
1.7ꢁ0.6
[
e]
0.036ꢁ0.005
–
[
5]
[6a]
[
a] T=378C, pH 8.5. [b] T=378C, pH 7.0. [c] T=258C, pH 9.5. [d] T=ambient temperature,
[6b] [7]
pH 9.5. [e] T=308C, pH 7.0.
mers, the hydrogels are cross-linked through protein/protein
interactions in addition to the coiled-coil interactions added
by the helical appendages (Figure 1b).
Table 2: Average values for the maximum power density, current density,
and open circuit voltage measured with three fuels in biobattery type 1
ꢀ1
(
2.5 wt% protein, Prussian Blue cathode, 1 mVs scan rate).
To create the metabolic pathway, a mixture of the three
enzymes was made with equal activity of each enzyme (1 U
Maximum power Maximum current Open circuit
density
[mWcm
density
[mAcm ]
potential
[V]
ꢀ
2
ꢀ2
ꢀ
1
]
each, where 1 U = 1 mmolmin ) and this mixture was used to
create anodes for use in two different enzymatic biobattery
configurations. The first type of biobattery (type 1) is based
methanol
32.7ꢁ0.2
577ꢁ26
379ꢁ16
317ꢁ24
0.195ꢁ0.015
0.146ꢁ0.002
0.075ꢁ0.007
formaldehyde 15.9ꢁ0.1
formate
6.7ꢁ0.1
[10]
on a previously published design where the enzyme mixture
in a dilute, non-hydrogel-forming concentration (2.5 wt%) is
+
mixed with methylene blue, NAD , Na SO , carbon black,
2
4
Table 3: Average values for the maximum current density measured with
and carbon fiber and applied to a stainless-steel mesh, and this
anode is combined with a Prussian blue cathode in a battery
test cell with a Nafion proton exchange membrane (PEM;
Figure 1c). The type 2 device was newly developed to take
advantage of the hydrogel-forming ability of the enzymes. A
small hydrogel (25 wt%) was formed by dissolving the
three fuels in biobattery type 2 (25 wt% hydrogel, air breathing platinum
ꢀ1
cathode, 100 mVs scan rate).
Maximum current density
ꢀ
2
[
mAcm ]
methanol
formaldehyde
formate
26.4ꢁ1.8
16.5ꢁ3.8
4.82ꢁ1.44
+
lyophilized enzymes along with NAD , methylene blue, and
Na SO . This material was spread on a stainless-steel mesh
2
4
anode. A commercially prepared air-breathing cathode was
assembled with a Nafion PEM and pressed to the anode
(
Figure 1d). Methylene blue is used to lower the overpoten-
increased with the different fuels, and the maximum current
ꢀ
2
tial required for the electrochemical oxidation of NADH at
the electrode (Figure 1a).
density (26.4 ꢁ 1.8 mAcm ) was obtained with methanol
fuel. There is a shape change between the polarization curves
of the two biobatteries, as well as a difference in the open
circuit potentials. The open circuit potential of the type 2
biobattery is considerably higher, because the oxygen cathode
has a higher standard reduction potential than the Prussian
blue cathode. The differences in shapes between the polar-
ization curves are interesting. There is a kinetic polarization
in the type 2 biobattery associated with the poor kinetics of
oxygen reduction on platinum, but this is not observed in the
Prussian blue cathode. On the other hand, the type 1
biobattery has higher resistance resulting in lower current
densities.
In this work we have used a protein engineering approach
to make a new catalytic biomaterial that supports a synthetic
metabolic pathway and can be used as an anode modification
for an enzymatic biobattery. The protein engineering strategy
enabled the enzymes to self-assemble while either retaining
or enhancing the catalytic performance of the enzymes. The
biobatteries resulted in power and current densities compa-
rable to those that have been previously reported in the
[
11]
The three enzyme metabolic pathway was designed to
fully oxidize methanol to CO through formaldehyde and
formate intermediates with a concomitant generation of
6
complete biobatteries with the two intermediates as fuels, the
functionality of all three enzymes in the pathway can be
investigated.
2
electrons per molecule of methanol. By operating the
Biobattery type 1 used the enzymes in a dilute solution
2.5 wt%), which reduces transport issues, and allows for
(
larger volumes of fuels to be used. The polarization curves for
the type 1 biobattery show that the current increases as the
reduction state of the fuel is increased from formate up to
methanol, indicating that all three enzymes are functioning in
the anodic pathway (Figure 2ab and Table 2). The maximum
power density (shown in Figure 2b) also increases as the
number of enzymes involved in the reaction increases. As
ꢀ
2
expected, the highest power density (32.7 ꢁ 0.2 mWcm ) and
ꢀ2
current density (577 ꢁ 26 mAcm ) were observed when
methanol was used in the biobattery.
Angew. Chem. Int. Ed. 2013, 52, 1437 –1440
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1439

.
Angewandte
Communications
literature for similar systems. Palmore et al. reported a meth-
measurements were carried out using an Autolab PGSTAT 302
potentio/galvanostat. Polarization curves were obtained by using
linear sweep voltammetry scanned (100 mVs ) from open circuit
voltage (OCV) to 0 V.
anol/O enzymatic biofuel cell using with the same cascade of
2
ꢀ1
three enzyme with
a
maximum power density of
ꢀ
2 [8a]
0
.68 mWcm .
In later work, Akers et al. constructed
a methanol/oxygen enzymatic biofuel cell and Addo et al.
made methanol/oxygen (air) enzymatic biofuel cells with
Received: September 14, 2012
Published online: December 13, 2012
ꢀ2[8b]
maximum power densities of 1.55 mWcm
and (261 ꢁ
ꢀ
2[8c]
7
.6) mWcm
, respectively. Higher power can certainly be
Keywords: biofuel cells · electrochemistry · enzyme catalysis ·
.
obtained by further optimizing this system.
protein engineering · self-assembly
On the large scale, enzymes can be produced inexpen-
sively, and these hydrogels may be a competitive alternative
to other chemical catalysts for methanol oxidation. This
general approach can also be applied to other enzymes so that
new biocatalytic biomaterials could be used for additional
bio-electrocatalysis applications as well as other technologies
involving heterogeneous biocatalysis.
[1] a) S. Banta, I. R. Wheeldon, M. A. Blenner, Annu. Rev. Biomed.
Eng. 2010, 12, 167 – 186; b) S. Banta, Z. Megeed, M. Casali, K.
Rege, M. L. Yarmush, J. Nanosci. Nanotechnol. 2007, 7, 387 –
401; c) D. N. Woolfson, M. G. Ryadnov, Curr. Opin. Chem. Biol.
2006, 10, 559 – 567; d) W. F. Daamen, J. H. Veerkamp, J. C.
van Hest, T. H. van Kuppevelt, Biomaterials 2007, 28, 4378 –
4398; e) S. A. Maskarinec, D. A. Tirrell, Curr. Opin. Biotechnol.
2005, 16, 422 – 426; f) F. G. Omenetto, D. L. Kaplan, Science
Experimental Section
2010, 329, 528 – 531.
The cloning, expression, and purification of the three modified
dehydrogenases are described in the Supporting Information. The
enzymatic activity of the hydrogel-forming proteins in dilute solution
[2] a) S. C. Barton, J. Gallaway, P. Atanassov, Chem. Rev. 2004, 104,
4867 – 4886; b) S. D. Minteer, B. Y. Liaw, M. J. Cooney, Curr.
Opin. Biotechnol. 2007, 18, 228 – 234; c) P. Atanassov, C.
Apblett, S. Banta, S. Brozik, S. C. Barton, M. Cooney, B. Y.
Liaw, S. Mukerjee, S. D. Minteer, Interface 2007, 16, 28 – 31.
[3] a) W. A. Petka, J. L. Harden, K. P. McGrath, D. Wirtz, D. A.
Tirrell, Science 1998, 281, 389 – 392; b) W. Shen, R. G. H.
Lammertink, J. K. Sakata, J. A. Kornfield, D. A. Tirrell, Macro-
molecules 2005, 38, 3909 – 3916; c) W. Shen, K. Zhang, J. A.
Kornfield, D. A. Tirrell, Nat. Mater. 2006, 5, 153 – 158; d) H. R.
Marsden, A. Kros, Angew. Chem. 2010, 122, 3050 – 3068; Angew.
Chem. Int. Ed. 2010, 49, 2988 – 3005.
[4] a) I. R. Wheeldon, S. C. Barton, S. Banta, Biomacromolecules
2007, 8, 2990 – 2994; b) I. R. Wheeldon, E. Campbell, S. Banta, J.
Mol. Biol. 2009, 392, 129 – 142; c) H. D. Lu, I. R. Wheeldon, S.
Banta, Protein Eng. Des. Sel. 2010, 23, 559 – 566; d) I. R.
Wheeldon, J. W. Gallaway, S. C. Barton, S. Banta, Proc. Natl.
Acad. Sci. USA 2008, 105, 15275 – 15280.
+
were measured with varying NAD concentrations and varying the
three different substrate concentrations (methanol, formaldehyde,
and formate). The reaction rates of all three dehydrogenases were
determined by following the change of absorbance from the produced
ꢀ1
ꢀ1
NADH at 340 nm (e = 6.22 mm cm ) using SpectraMax M2 plate
reader (Molecular Devices, Sunnyvale, CA). All assays were per-
formed in 96-well UV-transparent microplate at 378C. The kinetic
parameters were determined by fitting the reaction rates obtained
with varied substrate and cofactor concentration combinations to the
ordered bi-bi kinetic rate equation [Eq. (1)] using the Enzyme
Kinetics Module of SigmaPlot (Systat Software Inc., San Jose, CA).
All measurements were repeated at least three times.
total
þ
½
Eꢂ
k
½NAD ꢂ½Sꢂ
cat
þ
rate ¼ ki;NAD
ð1Þ
þ
þ
KM;S þ K ½NAD ꢂ þ K
M;S
M;NAD
þ
½Sꢂ þ ½NAD ꢂ½Sꢂ
[
5] M. C. Sheehan, C. J. Bailey, B. C. Dowds, D. J. McConnell,
Mixed hydrogels were prepared by combining equal units of each
enzyme and the mixture was lyophilized. The enzyme unit was
determined by following the change of absorbance at 340 nm using
Biochem. J. 1988, 252, 661 – 666.
[6] a) A. A. Klyosov, Biochemistry 1996, 35, 4457 – 4467; b) H.
Glatt, K. Rost, H. Frank, A. Seidel, R. Kollock, Arch. Biochem.
Biophys. 2008, 477, 196 – 205.
[7] A. E. Serov, A. S. Popova, V. V. Fedorchuk, V. I. Tishkov,
Biochem. J. 2002, 367, 841 – 847.
+
a UV/Vis spectrophotometer when 1 mm NAD and 10 mm substrate
ꢀ1
were used, and the unit was defined as 1 mmolmin of activity.
In the type 1 biobattery, a 2.5 wt% protein solution was made
+
with 1 mm methylene blue, 6 mm NAD , 200 mm Trizma (pH 8.5), 2m
[8] a) G. T. R. Palmore, H. Bertschy, S. H. Bergens, G. M. White-
sides, J. Electroanal. Chem. 1998, 443, 155 – 161; b) N. L. Akers,
C. M. Moore, S. D. Minteer, Electrochim. Acta 2005, 50, 2521 –
2525; c) P. K. Addo, R. L. Arechederra, S. D. Minteer, Electro-
analysis 2010, 22, 807 – 812.
[9] a) K. V. Mills, H. Paulus, Homing Endonucleases and Inteins,
Vol. 16 (Eds.: M. Belfort, D. W. Wood, B. L. Stoddard, V.
Derbyshire), Springer, Berlin, 2005, pp. 233 – 255; b) M.-Q.
Chong, Xu, Homing Endonucleases and Inteins, Vol. 16 (Eds.:
M. Belfort, D. W. Wood, B. L. Stoddard, V. Derbyshire),
Springer, Berlin, 2005, pp. 273 – 292; c) K. Shingledecker, S.
Jiang, H. Paulus, Arch. Biochem. Biophys. 2000, 375, 138 – 144;
d) W. Y. Wu, C. Mee, F. Califano, R. Banki, D. W. Wood, Nat.
Protoc. 2006, 1, 2257 – 2262.
Na SO , carbon black, carbon fiber, and 1-butyl-3-methylimidazolium
2
4
chloride (BMIMCl) ionic liquid to form a conductive bioanode. A
Prussian blue cathode made with 100 mg of iron(III) chloride, 100 mg
of potassium ferricyanide, and 200 mL of BMIMCl was used. The
anode and the cathode were separated by Nafion 212 and allowed to
equilibrate for 3 h prior to the introduction of 100 mm fuel. Polar-
ization curves were obtained by using linear sweep voltammetry
ꢀ1
scanned (1 mVs ) from open circuit voltage (OCV) to 0 V with
a Digi-Ivy DY2100 potentiostat.
For the type 2 biobattery, a 25 wt% hydrogel was kneaded onto
2
a 0.125 cm stainless-steel mesh surface until the gel was equally
distributed across the surface. The full biobatteries were constructed
using hydrogel-coated anodes along with commercially available
cathodes which consisted of a Nafion 112 assembled gas-permeable
[10] P. K. Addo, R. L. Arechederra, S. D. Minteer, J. Power Sources
2011, 196, 3448 – 3451.
2
ELAT electrode with 0.5 mg platinum per 1 cm carbon (Fuel Cell
Store, San Diego, CA). Both the anode and cathode were exposed to
air. 100 mm of fuel (methanol, formaldehyde, or formate) was added
to the anode in a total final volume of 5 mL. Electroanalytical
[11] A. A. Karyakin, E. E. Karyakina, W. Schuhmann, H.-L.
Schmidt, S. D. Varfolomeyev, Electroanalysis 1994, 6, 821 – 829.
1
440 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1437 –1440