Katarzyna P Sokol1, William E Robinson1, Ana R Oliveira2, Sonia Zacarias2, Chong-Yong Lee1, Christopher Madden1, Arnau Bassegoda3, Judy Hirst3, Inês A C Pereira2, Erwin Reisner1. 1. Department of Chemistry , University of Cambridge , Lensfield Road , Cambridge CB2 1EW , U.K. 2. Instituto de Tecnologia Química e Biológica António Xavier (ITQB) , Universidade NOVA de Lisboa , Av. da República , 2780-157 Oeiras , Portugal. 3. Medical Research Council Mitochondrial Biology Unit , University of Cambridge , The Keith Peters Building, Cambridge Biomedical Campus, Hills Road , Cambridge CB2 0XY , U.K.
Abstract
The biological formate hydrogenlyase (FHL) complex links a formate dehydrogenase (FDH) to a hydrogenase (H2ase) and produces H2 and CO2 from formate via mixed-acid fermentation in Escherichia coli. Here, we describe an electrochemical and a colloidal semiartificial FHL system that consists of an FDH and a H2ase immobilized on conductive indium tin oxide (ITO) as an electron relay. These in vitro systems benefit from the efficient wiring of a highly active enzyme pair and allow for the reversible conversion of formate to H2 and CO2 under ambient temperature and pressure. The hybrid systems provide a template for the design of synthetic catalysts and surpass the FHL complex in vivo by storing and releasing H2 on demand by interconverting CO2/H2 and formate with minimal bias in either direction.
The biological formate hydrogenlyase (FHL) complex links a formate dehydrogenase (FDH) to a hydrogenase (H2ase) and produces H2 and CO2 from formate via mixed-acid fermentation in Escherichia coli. Here, we describe an electrochemical and a colloidal semiartificial FHL system that consists of an FDH and a H2ase immobilized on conductive indium tin oxide (ITO) as an electron relay. These in vitro systems benefit from the efficient wiring of a highly active enzyme pair and allow for the reversible conversion of formate to H2 and CO2 under ambient temperature and pressure. The hybrid systems provide a template for the design of synthetic catalysts and surpass the FHL complex in vivo by storing and releasing H2 on demand by interconverting CO2/H2 and formate with minimal bias in either direction.
Semiartificial
catalytic systems
combine synthetic and biological units to drive challenging reactions
and provide new concepts for catalyst design.[1] Such solar-driven systems have already demonstrated coupling of
water oxidation to the reduction of CO2,[2−4] and protons[4,5] for the production of chemical fuels. However, storage and transport
of energy vectors are also important components in energy production–utilization
cycles and their development will benefit from more advanced approaches
and model systems.H2 is a promising fuel in a carbon-neutral
economy and
its conversion to formate allows for easier storage and transport.
H2 and formate are therefore an attractive energy vector
pair. Furthermore, H2 gas cleanly separates from dissolved
formate, and their interconversion comes at little thermodynamic cost
(eqs –3).[6,7] Achieving kinetic efficiency in
HCO2–/H2 interconversion remains
a synthetic challenge. Artificial systems commonly compete between
decomposition of formic acid to CO and H2O (dehydration),
and CO2 and H2 (dehydrogenation), and rely on
precious metals, high temperature/pressure, organic solvents, and
light.[8−10]FHL complexes are biological machines for HCO2–/H2 interconversion.[11] They
are either membrane-associated complexes composed of a multisubunit
[NiFe]-H2ase coupled to an FDH,[11−13] or smaller
soluble complexes of an [FeFe]-H2ase and an FDH.[14,15] The E. coli FHL-1 complex, composed of the membrane-bound
[NiFe]-H2ase 3 (HYD-3/HycE) and FDH-H (FdhF; Figure a), represents a well-studied
FHL, evolving H2 under fermentative conditions.[11,12] The constituent enzymatic units of FHL-1 have been demonstrated
to be reversible electrocatalysts,[16−20] but the complex is catalytically biased toward H2 production from formate.[14,15,19] Interconversion of HCO2–/H2 has also been reported in whole-cell studies,[14,20] notably in sulfate-reducing bacteria in the absence of sulfate.[21,22]Desulfovibrio vulgaris Hildenborough can grow by
converting formate to H2,[23] with
formate oxidation catalyzed by a periplasmic FDH, and H2 produced either via direct (periplasmic H2ase) or transmembrane
electron transfer (cytoplasmic H2ase).[24]
Figure 1
(a) Biological E. coli FHL-1 complex. FdhF, [Mo]-FDH;
B/F/G, Fe–S cluster-containing proteins; HycE, [NiFe]-H2ase; HycD/C, membrane proteins.[17] (b) IO-ITO|FDH||IO-ITO|H2ase cell: IO-ITO|FDH wired to
IO-ITO|H2ase electrode. (c) FDH–ITO–H2ase nanoparticle (NP) system with enzymes immobilized onto
ITO NP in solution. Species size not drawn to scale.
(a) Biological E. coli FHL-1 complex. FdhF, [Mo]-FDH;
B/F/G, Fe–S cluster-containing proteins; HycE, [NiFe]-H2ase; HycD/C, membrane proteins.[17] (b) IO-ITO|FDH||IO-ITO|H2ase cell: IO-ITO|FDH wired to
IO-ITO|H2ase electrode. (c) FDH–ITO–H2ase nanoparticle (NP) system with enzymes immobilized onto
ITO NP in solution. Species size not drawn to scale.Redox biocatalysts, including H2ases and FDHs,
have
been coupled to other enzymatic processes via electron relays. H2ases have been connected to nitrate and fumarate reductases,[25] diaphorase moieties,[26] nicotinamide reductase, and alcohol dehydrogenase[27] via graphitic particles. Notably, coupling a H2ase to carbon monoxide dehydrogenase efficiently catalyzed the water–gas
shift reaction.[28] Enzymatic cascades have
linked FDH with formaldehyde and alcohol dehydrogenases for methanol
production.[29,30] However, the reversible interconversion
of substrate and product has not been previously accomplished with
such coupled enzymes in vitro.Here, a semiartificial
FHL complex mimic is presented by rewiring
FDH[31,32] and H2ase[33] from D. vulgaris Hildenborough into electrochemical
and colloidal systems (Figure b,c). These systems rely on efficient electrical contact of
the [W/Se]-FDH active site via four [Fe4S4]
clusters and the [NiFeSe]-H2ase active-site via three [Fe4S4] clusters with nanostructured ITO.Macro-mesoporous
inverse opal (IO) ITO electrodes (20 μm
film thickness; 0.25 cm2 geometrical surface area) were
assembled as previously reported.[34] IO-ITO|FDH
and IO-ITO|H2ase electrodes were prepared by drop-casting
an FDH solution (2 μL, 19 μM with 50 mM DL-dithiothreitol,
incubated for 15 min) and a H2ase solution (2 μL,
5 μM), onto IO-ITO.[31,34] Protein film voltammetry
(PFV) was recorded using a three-electrode configuration (Figures a and S1) in CO2/NaHCO3 solution.
Current densities (J) of −185 μA cm–2 (CO2 reduction to formate by FDH) and
−450 μA cm–2 (H+ reduction
to H2 by H2ase) were observed at an applied
potential (Eapp) of −0.6 V vs standard
hydrogen electrode (SHE). Addition of sodium formate (20 mM) to the
IO-ITO|FDH system resulted in formate oxidation to CO2,
and 300 μA cm–2 was reached at −0.2
V vs SHE. After purging the IO-ITO|H2ase system with H2 (0.4 bar), H2 oxidation to H+ was observed
and 440 μA cm–2 was reached at −0.2
V vs SHE. The voltammograms cut through zero current around the formal
potentials of the CO2/HCO2− (eq ) and H+/H2 redox couples (eq ), demonstrating reversible
electrocatalysis for both enzymes.[6,35]
Figure 2
(a) Three-electrode
PFV (ν = 5 mV s–1,
1st and 5th scan, increasing transparency) using IO-ITO|FDH or IO-ITO|H2ase working, Ag/AgCl (KClsat) reference and Pt
mesh counter electrodes. (b) Two-electrode PFV (ν = 5 mV s–1, 1st and 5th scan) of IO-ITO|FDH wired to IO-ITO|H2ase. (c) Two-electrode CPE of IO-ITO|FDH wired to IO-ITO|H2ase. Conditions: CO2/NaHCO3 (100 mM),
KCl (50 mM), 1 bar CO2 or 0.4/0.6 bar H2/CO2, pHinitial = 6.5–6.7, T = 25 °C, stirring. Substrates: formate (20 mM) and/or 0.4/0.6
bar H2/CO2.
(a) Three-electrode
PFV (ν = 5 mV s–1,
1st and 5th scan, increasing transparency) using IO-ITO|FDH or IO-ITO|H2ase working, Ag/AgCl (KClsat) reference and Pt
mesh counter electrodes. (b) Two-electrode PFV (ν = 5 mV s–1, 1st and 5th scan) of IO-ITO|FDH wired to IO-ITO|H2ase. (c) Two-electrode CPE of IO-ITO|FDH wired to IO-ITO|H2ase. Conditions: CO2/NaHCO3 (100 mM),
KCl (50 mM), 1 bar CO2 or 0.4/0.6 bar H2/CO2, pHinitial = 6.5–6.7, T = 25 °C, stirring. Substrates: formate (20 mM) and/or 0.4/0.6
bar H2/CO2.Multiple PFV scans of IO-ITO|FDH and IO-ITO|H2ase (Figure S2) showed minimal desorption/activity
losses. Controlled-potential electrolysis (CPE) of IO-ITO|FDH and
IO-ITO|H2ase was performed to measure H+/CO2 reduction (Eapp = −0.6
V) as well as H2/formate oxidation (Eapp = −0.2 V) (Figure S3).
Following equilibration, both electrodes retained good activity after
24 h in both directions. Faradaic efficiencies (ηF) for formate and H2 production were determined to be
76% and 77%, respectively. Efficiency losses may be attributed to
the capacitive background current of porous IO-ITO,[34] undetected trapped product, and a contribution from ITO/FTO
degradation.[36,37]The comparable formal redox
potentials of H+/H2 and CO2/HCO2– conversion
(eq -3), reversible catalysis of the individual enzymes, high and
matching current densities, and good stability make this enzyme pair
a promising candidate for assembling a reversible HCO2–/H2 interconversion system.[6] Thus, the IO-ITO|FDH (working electrode) was wired to the
IO-ITO|H2ase (counter electrode) in a two-electrode configuration
(Figure b). When no
additional substrate was present (only buffering CO2 and
H+), only a noncatalytic current attributed to IO-ITO capacitance
was observed. Upon addition of formate, an oxidative current was observed
(formate oxidation to CO2 and H+ reduction to
H2) at a positive applied voltage (U >
0 V); 250 μA cm–2 was reached at U = 0.2 V. Addition of H2 resulted in a reductive current
(H2 oxidation to H+ and CO2 reduction
to formate) at a negative voltage with −250 μA cm–2 obtained at U = −0.2 V.To achieve reversible formate/H2 interconversion (eq ) both formate and H2 were added in addition to CO2 and H+. A reversible voltammogram was observed, with zero current at approximately U°′ at 0.02 V. A marginally more positive or
negative voltage drove the reaction in either direction, demonstrating
reversible unbiased electrocatalysis. 200 μA cm–2 and −200 μA cm–2 were reached at U = 0.2 V and −0.2 V, respectively. Multiple PFV
scans of the IO-ITO|FDH||IO-ITO|H2ase cell (Figure S4) showed stability of the system with
marginal losses. Control experiments with IO-ITO|FDH (or ITO|H2ase) wired to IO-ITO (Figure S5) gave only a small capacitive current in the presence and absence
of substrates (H2/formate).CPE over 2 h at Uapp = 0.2 V with the
IO-ITO|FDH||IO-ITO|H2ase cell with formate present (Figure c) produced H2 (5.84 ± 0.88 μmol cm–2) with
ηF of (79 ± 11)%. Similarly, CPE at Uapp = −0.2 V for 2 h with H2 present generated formate (5.00 ± 0.80 μmol cm–2) with ηF of (81 ± 15)%. This semiartificial
electrochemical FHL system exhibited good stability, retaining >95%
of its initial activity after 2 h in both directions. After equilibration,
the cell exhibited high bidirectional stability for >1 day (Figure S6). For formate oxidation (Uapp = 0.2 V), H2 (36.28 μmol cm–2) was detected with ηF = 72%. For H2 oxidation
(Uapp = −0.2 V), formate (42.80
μmol cm–2) was detected with ηF = 77%. Similarly to the three-electrode systems, capacitive currents
and FTO/ITO dissolution[36,37] might have decreased
the product yield.To further investigate the system’s
reversibility without
electrochemical wiring, FDH and H2ase were coassembled
on ITO nanoparticles (NPs) (0.3 mg mL–1) (Figures and S7) dispersed in electrolyte solution (see Supporting Information). Solutions of FDH (19
nM, incubated as above) and H2ase (3.4 nM) were added to
the vessel, which was sealed and purged with CO2. Either
formate or H2 was introduced to the vessel. FDH:H2ase molar ratios (Figure S8) and total
concentrations (Figure S9a,b) were screened
for the optimum H2 evolution rate. The optimal system contained
an enzyme loading of approximately 40 FDH and 7 H2ase particles
per ITO NP, based on the adsorption surface area of 27 m2 g–1, ∼31 400 nm2 per
NP (assuming a 50 nm diameter sphere), and an enzyme footprint of
∼100 nm2.
Figure 3
Product quantification of the colloidal FDH–ITO–H2ase NP system: using ITO NPs (0.3 mg mL–1), FDH (19.0 nM) and H2ase (3.4 nM). (a) H2 production in the presence of 10 mM formate and 1 bar CO2. Vheadspace = 1.72 mL. (b) Formate production
in the presence of 0.4/0.6 bar H2/CO2. Vsolution = 2 mL. Conditions: CO2/NaHCO3 (100 mM), KCl (50 mM), 1 bar CO2 or 0.4/0.6 bar
H2/CO2, pHinitial = 6.5–6.7, T = 23 °C, stirring.
Product quantification of the colloidal FDH–ITO–H2ase NP system: using ITO NPs (0.3 mg mL–1), FDH (19.0 nM) and H2ase (3.4 nM). (a) H2 production in the presence of 10 mM formate and 1 bar CO2. Vheadspace = 1.72 mL. (b) Formate production
in the presence of 0.4/0.6 bar H2/CO2. Vsolution = 2 mL. Conditions: CO2/NaHCO3 (100 mM), KCl (50 mM), 1 bar CO2 or 0.4/0.6 bar
H2/CO2, pHinitial = 6.5–6.7, T = 23 °C, stirring.Upon formate addition to the FDH–ITO–H2ase
system (Figure a),
H2 was produced with a reaction rate (Figure S9c) of 0.24 ± 0.01 μmol H2 h–1 during the first 8 h [turnover number,
TON = (23.0 ± 1.5) × 103 and turnover frequency,
TOF = 6.4 ± 0.4 s–1 for the H2ase],
after which the rate started to decrease (Table S1). Equilibrium was reached after ∼72 h (5.82 ±
0.24 μmol H2, pH 6.88, T = 23 °C),
in agreement with calculations (5.95 μmol, 2.97 mM of H2; see Supporting Information).[7]In the presence of H2, the FDH–ITO–H2ase system (Figure b) produced formate with an initial reaction rate of 1.33
± 0.01 μmol formate h–1 [TON = (15.8
± 5.4) × 103 and TOF = 4.4 ± 1.5 s–1 for the FDH] for the first 8 h (Figure S9d). Equilibrium was reached after ∼96 h (36.16 ± 1.47
μmol formate, pH 6.99, T = 23 °C), consistent
with calculations (37.11 μmol, 18.56 mM of formate).[7] Control experiments with no ITO NPs, omitting
an enzyme or with denatured enzymes (Figure S10), showed only negligible H2 and formate production (<0.2
μmol) (Tables S2 and S3). Therefore,
the ITO NPs act as a semiheterogeneous electron relay facilitating
electron transfer between electroactive FDH and H2ase.In D. vulgaris cells, the two periplasmic enzymes
exchange electrons through the type-I cytochrome c3 (TpIc3) electron acceptor.[24] We therefore studied the activity of these enzymes
in solution with TpIc3. A high concentration
of the cytochrome (1.9 μM, 100-fold excess vs FDH) was required
to achieve comparable kinetics of H2 and formate production
(Figure S11a,b), revealing the superiority
of coimmobilizing the two enzymes on synthetic ITO NPs to achieve
efficient electron transfer.In summary, we have presented how
semiartificial systems consisting
of FDH and H2ase from D. vulgaris wired
to ITO can mimic the biological FHL complex. The semiartificial FHL
systems are based on a bottom-up design that employs a pair of reversible
redox enzymes immobilized on conductive scaffolds to enable an overall
catalytic reaction to proceed to thermodynamic equilibrium. The semiartificial
FHL concept can be deployed in either an electrochemical cell or a
self-assembled colloidal suspension, providing versatility for applications
in different contexts. The design concept of linking two half-reactions
via a conductive scaffold also provides a blueprint to develop improved
synthetic H2/formate cycling catalysts in future development.
Authors: Marta C Marques; Cristina Tapia; Oscar Gutiérrez-Sanz; Ana Raquel Ramos; Kimberly L Keller; Judy D Wall; Antonio L De Lacey; Pedro M Matias; Inês A C Pereira Journal: Nat Chem Biol Date: 2017-03-20 Impact factor: 15.040
Authors: Sofia M da Silva; Johanna Voordouw; Cristina Leitão; Mónica Martins; Gerrit Voordouw; Inês A C Pereira Journal: Microbiology Date: 2013-05-31 Impact factor: 2.777
Authors: Jennifer S McDowall; Bonnie J Murphy; Michael Haumann; Tracy Palmer; Fraser A Armstrong; Frank Sargent Journal: Proc Natl Acad Sci U S A Date: 2014-08-25 Impact factor: 11.205
Authors: Katarzyna P Sokol; William E Robinson; Ana R Oliveira; Julien Warnan; Marc M Nowaczyk; Adrian Ruff; Inês A C Pereira; Erwin Reisner Journal: J Am Chem Soc Date: 2018-11-27 Impact factor: 15.419
Authors: Esther Edwardes Moore; Samuel J Cobb; Ana Margarida Coito; Ana Rita Oliveira; Inês A C Pereira; Erwin Reisner Journal: Proc Natl Acad Sci U S A Date: 2022-01-25 Impact factor: 12.779