Reduction of CO2 to methanol by a polyenzymatic system encapsulated in phospholipids-silica nanocapsules

Article abstract of DOI:10.1039/c3nj00688c

By reversing the biological metabolic reaction pathway of three dehydrogenases, a formate dehydrogenase, a formaldehyde dehydrogenase and an alcohol dehydrogenase, it was possible to transform CO₂ into methanol by a cascade reaction. The activity of each enzyme was examined separately and then the relative amount of each enzyme for the cascade reaction was optimized. The enzymes consume one molecule of the NADH cofactor each to run which should be regenerated for cost reasons. Three different NAD⁺ regenerating systems were compared: 2 enzymes (phosphite dehydrogenase (PTDH) and glycerol dehydrogenase) and a natural photosystem extracted from spinach leaves (chloroplasts). PTDH was proven to be more efficient at neutral pH. The new polyenzymatic system (4 enzymes) was then encapsulated in silica nanocapsules (internal diameter 30 nm) nanostructured by phospholipids (NPS). This hybrid nanobioreactor showed an activity 55 times higher than the free enzymes in solution. A methanol production of 42 μmol g_NPS^-1 corresponding to 4.3 mmol g_{commercial enzymatic powder}^-1 in 3 h at room temperature and 5 bar was obtained.

Full text of DOI:10.1039/c3nj00688c

NJC
View Article Online
View Journal | View Issue
PAPER
Reduction of CO2 to methanol by a polyenzymatic
system encapsulated in phospholipids–silica
nanocapsules†
Cite this: NewJ.Chem., 2013,
37, 3721
a
a
a
b
b
´
Remi Cazelles, Jullien Drone, François Fajula, Ovidiu Ersen, Simona Moldovan
and Anne Galarneau*a
By reversing the biological metabolic reaction pathway of three dehydrogenases,
a formate
dehydrogenase, a formaldehyde dehydrogenase and an alcohol dehydrogenase, it was possible to
transform CO2 into methanol by a cascade reaction. The activity of each enzyme was examined
separately and then the relative amount of each enzyme for the cascade reaction was optimized. The
enzymes consume one molecule of the NADH cofactor each to run which should be regenerated for
cost reasons. Three different NAD+ regenerating systems were compared: 2 enzymes (phosphite
dehydrogenase (PTDH) and glycerol dehydrogenase) and a natural photosystem extracted from spinach
leaves (chloroplasts). PTDH was proven to be more efficient at neutral pH. The new polyenzymatic
system (4 enzymes) was then encapsulated in silica nanocapsules (internal diameter 30 nm)
Received (in Montpellier, France)
25th June 2013,
Accepted 19th August 2013
nanostructured by phospholipids (NPS). This hybrid nanobioreactor showed an activity 55 times higher
DOI: 10.1039/c3nj00688c
À1
than the free enzymes in solution. A methanol production of 42 mmol gNPS
corresponding to
À1
www.rsc.org/njc
4.3 mmol gcommercial enzymatic powder in 3 h at room temperature and 5 bar was obtained.
an expensive molecule that should be regenerated in situ in a
cost efficient process. This can be achieved through electro-
1. Introduction
Considering the need for environmentally friendly carbon
sources for chemical, petrochemical and pharmaceutical industries,
researchers have developed a process that produces C1 synthetic
building blocks from a greenhouse gas such as carbon dioxide.
CO2 is used as feedstock for the synthesis of a broad range of
compounds such as urea, salicylic acid, carbonates or an alcohol
like methanol.1–3 Different pathways to produce methanol from
CO2 have been investigated: chemical, electrochemical, photo-
chemical and enzymatic synthesis.4–8 Among them, a polyenzy-
matic process using 3 dehydrogenases, firstly described by Obert
and Dave,7 operated at ambient temperature and atmospheric
pressure. Its thermodynamic feasibility was proved by Wang and
coworkers.9 This low energy consuming process can be seen as a
green and sustainable strategy as compared to the other techniques.
The biocatalytic synthesis of methanol from CO2 involving the 3
dehydrogenases needs a cofactor (reduced Nicotinamide Adenine
Dinucleotide, NADH) to achieve the biotransformation. NADH is
chemical and photochemical reactions or by using another
enzyme, which converts the oxidized form of the cofactor
(NAD+) into its reduced state (NADH).10–12 Among these possibilities,
the enzymatic pathway is often more selective and provides the
highest reaction rates. Formate dehydrogenase and phosphite
dehydrogenase are the more efficient recycling systems reported
so far.12,13
Multi-enzyme cascade catalysis is an attractive alternative to
chemical catalysis for the production of a number of fine
chemicals, as it allows energy efficient conversion thanks to
mild conditions: ambient or near ambient temperature and low
pressure. Xue and Woodley already discussed the technology
options and strategies that are available for the development of
multi-enzymatic processes.14 Hold and Panke have reproduced
in vitro the glycolysis of Escherichia coli using several enzymes
coming from the same bacterium.15
In the present study, we used an artificial cascade reaction,
in which we combined enzymes from different organisms
which may not share either similar optimal pH or have similar
reaction rates at similar concentrations of reagents. Under the
natural conditions, biological pathways are able to shift the
equilibrium of some enzymes by controlling product concentra-
tions, which is complicated to reproduce in vitro. Moreover,
in vitro reactions take place at strong concentration of substrates,
a Institut Charles Gerhardt Montpellier, UMR 5253 CNRS-UM2-ENSCM-UM1,
ENSCM, 8 rue de l’Ecole Normale, 34296 Montpellier Cedex 5, France.
E-mail: anne.galarneau@enscm.fr
b
´
Institut de Physique et Chimie des Materiaux de Strasbourg, UMR 7504
CNRS-University of Strasbourg, 23 rue du Loess, 67034, Strasbourg Cedex 2, France
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3nj00688c
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem., 2013, 37, 3721--3730 3721

View Article Online
Paper
NJC
(YADH) enzymes were purchased from Sigma-Aldrich. These
commercial powders are not pure enzymes and the effective
enzyme content needs to be determined prior to use. The
genetic construct encoding the phosphite dehydrogenase
(PTDH) used in NADH recycling experiments was obtained by
courtesy of Prof. H. M. Zhao.13 In brief, Escherichia coli
BL21*(DE3) (Life Technologies) was transformed by the plasmid
pET15b-PTDH12x13 and used to overexpress the enzyme as pre-
viously described with slight modifications. Homogeneous enzyme
preparation was obtained after purification using IMAC Ni2+
(HisTrap 1 mL, GE Healthcare) followed by desalting (HiTrap
5 mL, GE Healthcare) with conservation buffer (50 mM MOPS,
10% glycerol, pH 7.25) under standard conditions. Aliquots of the
enzyme were stored at À80 1C until use. Oxidized and reduced
nicotinamide adenine dinucleotide (NAD+, NADH), mono- and
di-hydrogenated potassium phosphate used for buffers and
phosphite (Na2HPO3) were obtained from Sigma-Aldrich. For
the synthesis of silica nanocapsules for enzymes encapsulation,
tetraethoxysilane (TEOS), egg lecithin (60%), dodecylamine,
and b-D-lactose were purchased from Sigma. Potassium hydro-
gen carbonate (KHCO3) and Na2CO3 were used as gaseous CO2
sources (CO2(g)) and were obtained from Sigma. At pH 6.5 and
37 1C, dehydrogenated carbonate (CO32À) is likely nonexistent
(o1%) and soluble hydrogenated carbonates coexist with gaseous
Scheme 1 Representation of the enzymatic pathway of carbon dioxide con-
version into methanol by 3 dehydrogenase enzymes. Three moles of the reduced
nicotinamide dinucleotide (NADH) cofactor are needed to convert 1 mole of CO2
into methanol.
the accumulation of the products can then lead to the inactivation
of the other enzymes. Furthermore, another limitation for the use
of CO2 is the low solubility of gaseous CO2 in water and therefore
its availability, a limitation which can be partly overcome by
increasing the pressure.16 Indeed CO2 in water decomposes into
carbonate species depending on the pH of the solution and an
equilibrium between carbonates and gaseous CO2 takes place. In
the artificial polyenzymatic cascade reaction we studied, the for-
mate dehydrogenase (FateDH) converts gaseous CO2 into formate,
then the formaldehyde dehydrogenase (FaldDH) converts formate
into formaldehyde and finally alcohol dehydrogenase (YADH)
converts formaldehyde into methanol (Scheme 1). Previous
studies have highlighted the crucial role of adding a regenerating
system for NADH recycling and of using an appropriate enzyme
immobilization technique to improve the reaction activity.7,17,18
Immobilization of the enzymes is a critical concern for
biochemical processes as it provides stabilization and easier
use, but also in the present case, improved activity. Dave and
coworkers found that these enzymes encapsulated in a silica
matrix by sol–gel show higher productivity of methanol compared
to free enzymes in solution.7,17 Jiang and coworkers investigated
the influence of enzyme immobilization into silica sol–gel and
alginate–silica nanoparticles as well as an encapsulation in multi-
compartmented titanate systems.8,19
Herein, we performed a systematic study of the polyenzy-
matic system and searched for the most effective conditions to
reach the highest activity: pH, the ratio between the three
enzymes, the nature of the NADH regeneration system and
initial NADH concentration. Finally, we encapsulated the poly-
enzymatic system by an optimized silica sol–gel technique using
natural phospholipids (egg lecithin) and lactose to protect the
enzymes leading to phospholipids–silica nanocapsules (NPS).
NPS have previously proven their efficiency for the encapsulation
of bienzymatic systems such as glucose oxidase (GOx) and
horseradish peroxidase (HRP) or GOx and bovine hemoglobin
(Hb) to generate in situ H2O2 and oxidize polycyclic aromatic
hydrocarbon pollutants in water using O2 as an oxidant.20,21
À
CO2 in equilibrium with 45% of CO2(g) and 55% of K+HCO3
(Fig. S1, ESI†), which are enough to carry out enzymatic reactions
with dissolved gaseous CO2.22,23
2.2. Enzyme activity measurement
2.2.1. Standard assays for commercial powder character-
ization. To determine the specific activity of the 3 commercial
dehydrogenases, it is first necessary to quantify the real amount
of enzymes contained in the commercial powders and then to
quantify the activity of each enzyme by specific assays. The
enzyme content was determined using the Thermo Scientific
Pierce BCA Protein Assay Kit. The catalytic activities of the
enzymes were determined under fixed conditions (V = 100 mL,
T = 37 1C) in potassium phosphate buffer (0.1 M, pH 6.5).
To characterize the activity of the formate dehydrogenase
(FateDH), a standard assay was conducted as follows: NAD+
(1 mM) and formate (HCOOÀ 500 mM) were mixed with FateDH
(10 mg LÀ1). For the formaldehyde dehydrogenase (FaldDH),
the standard assay was carried out using NAD+ (0.5 mM),
formaldehyde (HCHO 6 mM) and FaldDH (10 mg LÀ1).
For the alcohol dehydrogenase (YADH), the standard assay
was performed with NADH (0.025 mM), formaldehyde (HCHO
100 mM) and YADH (1 mg LÀ1). The activity of the enzymes was
determined by following the increase (for FateDH and FaldDH)
or the decrease (for YADH) of the absorbance band of NADH at
340 nm by UV-Vis spectroscopy. The pH meter used to adjust
pH stock solutions was from Eutech instrument, model pH510.
A mini hybridization oven from Appligen was used for all the
2. Experimental section
2.1. Materials
Formate dehydrogenase (Homo-dimer, 80.7 kDa) from Candida thermostated reactions and synthesis.
boidinii (FateDH), formaldehyde dehydrogenase (Homo-tetramer,
2.2.2. Optimization of the polyenzymatic system used to
168 kDa) from Pseudomonas putida (FaldDH) and alcohol dehydro- reduce CO2 into methanol. All reactions were carried out under
genase (Homo-tetramer, 141 kDa) from Saccharomyces cerevisiae a N2 or CO2 atmosphere in order to avoid oxidation of NADH.
c
3722 New J. Chem., 2013, 37, 3721--3730
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
Solutions of enzymes and chemicals were prepared in soni- incubation at 37 1C for 65 hours. The optimization of the amount
cated, demineralized H2O. The optimum pH for each enzyme of regenerating system (PTDH) was investigated by adding NADH
under their reductive reaction pathway was investigated using (0.01 M), Na2HPO3 (0.05 M) and a variable amount of PTDH from
different buffer solutions. Solutions of pH 4.5, 5.5, 6, 6.5, 7, 7.5, 0 to 6.1 g LÀ1 of pure enzyme. The influence of CO2 pressure to
8, 8.5, 9, 9.3 were prepared by mixing 1 M K2HPO4 solution with increase CO2 gaseous concentration in the aqueous reaction
1 M KH2PO4 solution, pH was adjusted using a pH meter at (Fig. S1, ESI†) was investigated for 3 h at 37 1C under 0.5 MPa
25 1C. The activity of FateDH (0.1 g LÀ1) was measured at 25 1C (5 bar) of CO2 by using the optimized ratio of the polyenzymatic
in 100 mL of 96-well plates with a 0.1 M phosphate buffer solution, system and NADH: FateDH, FaldDH, and YADH (0.01, 0.15, and
NADH (1.5 mM) and Na2CO3 (1 mM) used as the CO2 source. The 0.75 g LÀ1, of commercial enzymatic powder, respectively) and
activity of FaldDH (0.1 g LÀ1) was measured at 25 1C in 100 mL of PTDH (3.5 g LÀ1 of pure enzyme) in a phosphate buffer solution
96-well plates with a 0.1 M phosphate buffer solution, NADH (0.05 M, pH 6.5) containing NADH (0.1 M).
(1 mM) and HCOOÀ (2 mM) as substrates. The activity of YADH
(0.01 g LÀ1) was measured at 25 1C in 100 mL of 96-well plates with a
2.3. Enzymes encapsulation in silica nanocapsules (NPS) and
0.1 M phosphate buffer solution, NADH (1 mM) and HCHO (3 mM)
as substrates. The reactions were carried out in a 0.6 mL Eppendorf
tube placed in an incubator at 37 1C with stirring at 250 rpm, for
72 h under a CO2 atmosphere.
bioactivity
Silica nanocapsules (NPS) were synthesized as reported by
Galarneau et al.21 in a 1.5 mL Eppendorf tube. An organic
solution of lecithin (0.1 g), dodecylamine (7.5 mg) and ethanol
(0.290 mL) was first prepared. An aqueous solution was prepared
containing 0.15 mL of potassium phosphate buffer (0.1 M, pH 6.5),
b-D-lactose (1.25 mg), NADH (0.03 mM) and the optimized poly-
enzymatic system FateDH (0.05 g LÀ1), FaldDH (0.75 g LÀ1), and
YADH (3.75 g LÀ1). Another aqueous solution was also prepared by
adding the NADH regenerating enzyme PTDH in the following ratio:
FateDH (0.0066 g LÀ1), FaldDH (0.10 g LÀ1), YADH (0.50 g LÀ1) of
enzymatic commercial powder and PTDH (2.31 g LÀ1) of pure
enzyme. A volume of 43 mL of the organic solution was stirred
magnetically at 1300 rpm at 37 1C and then 150 mL of the aqueous
solution was added dropwise. After complete addition and 3 min
stirring, a homogeneous emulsion was obtained. Then TEOS (25 mL)
was added dropwise always under magnetic stirring at 1300 rpm
and homogenized for 60 minutes until the silica gelification occurs
and prevents the magnetic stirrer to rotate. The mixture was then
aged in a static environment for 22 h at 37 1C. A volume of 0.5 mL of
potassium buffer (0.1 M, pH 7) was added to the resulting gel and
was then centrifuged for 30 min at 4 1C and 10000 rpm. Then the
supernatant was filtered using 0.2 mm filters and passed through a
10 mL Sephadex G25 column to separate enzymes from chemicals
used for the synthesis. The Bradford assay was used to determine
the protein concentration in the supernatant. The washing step was
repeated 4 more times. After washing the NPS material was dried
under vacuum with P2O5 for 72 h at 20 1C. The resulting
as-synthesized NPS powders were stored at 4 1C.
To determine the optimal FaldDH/FateDH ratio, different
solutions were prepared using a fixed FateDH concentration
(0.1 g LÀ1) and a variable FaldDH concentration (from 0 to 1.5 g LÀ1
)
in a pH 6.5 buffer solution (KH2PO4–K2HPO4 at 0.1 M) with
the following amounts of cofactor and substrate: cofactor
[NADH] = 10 mM and [KHCO3] = 100 mM. To determine the
best YADH/FaldDH ratio, different solutions were prepared
using a fixed FaldDH concentration (0.1 g LÀ1) and a variable
YADH concentration (from 0 to 1 g LÀ1) in a pH 6.5 buffer
solution (KH2PO4–K2HPO4 at 0.1 M) with the following
amounts of cofactor and substrate: cofactor [NADH] = 10 mM
and [HCOOÀ] = 100 mM. The reactions were carried out in a
0.6 mL Eppendorf tube placed in an incubator at 37 1C with
stirring at 250 rpm for 22 h under a N2 atmosphere.
To investigate the NADH recycling system, two enzymes were
tested: the phosphite dehydrogenase (PTDH, 10 mg LÀ1) using
phosphite (Na2HPO3, 0.5 M) as a substrate and the glycerol
dehydrogenase (GlyDH, 10 mg LÀ1) using glycerol (0.5 M) as a
substrate. The two mixtures were incubated in potassium
phosphate buffer (0.1 M) with pH varying from 4.5 to 9.3. The
reaction took place in a microplate spectrophotometer at 37 1C
after adding the oxidized form of the cofactor NAD+ (1 mM).
Another NADH regenerating system was studied, a natural
photosystem of chloroplast suspension prepared from spinach
leaves as described by Berthold et al.24 A solution containing a
chloroplast suspension (0.3 gchlorophylls L
À1) in a 0.1 M phos-
phate buffer solution pH 6 was activated using a 40 W Neon
desk lamp in the presence of the oxidized form of the cofactor
NAD+ (1 mM). The NADH regenerating capacity of the 3 systems
2.4. Characterization
2.4.1. Reaction product characterization. Methanol amount
was followed by monitoring the appearance of an NADH band was determined using a Varian 3900 gas chromatograph with a
at the absorbance of 340 nm by UV-Vis spectroscopy.
polar column (DB 52-CB column 25 Â 0.32 Â 0.25) and a FID
2.2.3. Activity of the polyenzymatic system used to reduce detector by using 1-pentanol as internal standard. The injector
CO2 to methanol. The 3 enzymes, FateDH, FaldH and YADH in was set at 220 1C with a column temperature of 55 1C for 3 min,
optimized ratios of commercial powder (0.01, 0.15 and 0.75 g LÀ1
,
then from 55 1C to 220 1C at 20 1C minÀ1 under a hydrogen
respectively) were incubated in phosphate buffer solution flow of 1.2 mL minÀ1. The N2 flux was fixed based on the
(0.05 M, pH 6.5) containing KHCO3 (0.05 M) and NADH with Van Deemter curve of the methanol chromatographic peak.
a variable concentration from 0 to 0.2 M to identify the most Chromatograms show the separation of methanol from
suitable ratio of NADH needed for the reaction. The Eppendorfs formaldehyde (Fig. S2, ESI†). The quantification of form-
used for the enzymatic reaction were flushed with CO2 prior to aldehyde by gas chromatography was inefficient because of its
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem., 2013, 37, 3721--3730 3723

View Article Online
Paper
NJC
low coefficient response with a FID detector, that is why the are the enzymes originally used for the overall enzymatic
Nash method25 was preferred. conversion of CO2 to methanol.7 However, to the best of
2.4.2. Silica nanocapsule (NPS) characterization. XRD pat- our knowledge, no systematic study has been performed to
terns were recorded using a Bruker D8 Advance diffractometer determine the amount of active enzyme contained in the
with Bragg–Brentano geometry and equipped with a Bruker commercial enzymatic powders and the optimal conditions of
Lynx Eye detector. XRD patterns were recorded in the range reaction (pH and the relative amount of each enzyme in the
0.5–61 (2y) with an angular step size of 0.01971 and a counting polyenzymatic system) to run efficiently the cascade reaction.
time of 0.2 s per step. SEM images were obtained using a Here we report a systematic study of the combination of the
Hitachi S4800 microscope. Thermogravimetric analyses (TGA) three different enzymes to get an effective catalytic system.
were performed using a Perkin Elmer STA6000 thermogravimeter. Firstly we investigated the enzyme activity under standard
The materials were heated in a N2-flow at 900 1C with a temperature conditions to determine the amount of active enzymes in the
ramp of 5 1C minÀ1. Textural properties of the as-synthesized commercial powder. In terms of protein or enzyme amount, we
nanocapsules were determined by N2 adsorption–desorption at found that 160 mg, 140 mg and 630 mg of protein per g of
77 K on a Micromeritics Tristar 3000 apparatus. Samples were commercial powders were effectively present in FateDH, FaldDH
previously outgassed in a vacuum at 50 1C for 12 h. To characterize and YADH commercial enzymatic powders, respectively. This
their particle size, the silica nanocapsules were first sonicated for confirms that enzyme suppliers add a lot of additives (e.g. sucrose,
30 min in water before performing size measurement on a Zeta polyethylene glycol, NaCl) before lyophilization to stabilize the
sizer Nano ZS90 instrument.
enzymes into their commercial forms and therefore the exact
The electron tomography experiments were carried out on a amount of active enzyme has to be determined for each batch.
JEOL 2100F transmission electron microscope (TEM) with a The content of active enzymes may vary from batch to batch of
field emission gun operating at 200 kV, equipped with a probe commercial enzymatic powders and rather than mass, one should
corrector and a GATAN Tridiem energy filter. Prior to observa- use the activity expressed in Unit (U: micromoles of product formed
tion, the silica nanocapsules were dispersed in deionized water per min) to be accurate. For the enzymatic commercial powders
À1
and then sonicated for several minutes. Up to 5 droplets were we used, we fÀo1und specifics activities of 55 U mgFateDH
,
À1
further deposited onto a copper grid covered by a holey carbon 14 U mgFaldDH and 2300 U mgYADH expressed by mg of
membrane rendered hydrophobic by H2/Ar plasma cleaning. commercial enzymatic powder (corresponding to a Unit ratio of
Afterwards, the specimen was plunged into liquid ethane and 1/3.8/3136 U/U/U for FateDH, FaldDH and YADH, respectively).
mounted on the cryo holder. The latest manipulations were However all of the following weights of commercial enzymatic
carried out in liquid nitrogen (77 K) in a similar manner as for powder described in this paper should be recalculated from the
the biological specimens. Even so, the specimen morphology standard units when a new commercial enzymatic powder is
considerably changed under the electron beam. One should used. The activity of the 3 enzymes in the reductive reactions
keep in mind as well that during the acquisition of a complete (Scheme 1) was studied at different pH to determine the
tomography series the specimen is to be exposed to the electron optimum pH for the cascade reaction. FateDH and FaldDH
flux for durations of one hour or more. Therefore, we have cannot reduce CO2 and HCOOÀ, respectively, at pH > 8.5
tested several representative regions in terms of electron beam (Fig. 1). The overall reduction of CO2 to methanol should be
induced damage, by employing different electron doses and carried out at pH values between 6 and 7 to allow at least 80%
irradiation durations. Once these conditions were established, efficiency of each enzyme, this is concordant with the litera-
two tomography series were acquired by tilting the specimen ture.26 It appears that highest activities for all the enzymes were
over a range of Æ601, with an image recorded for every 21. The obtained at pH 6.5 (Fig. 1). As CO2 bubbled into water solution
images were aligned by using the cross-correlation algorithm gives immediately an equilibrium between soluble gaseous CO2
implemented in the IMODod software. The reconstructions
have been computed using 10 iterations within the algorithms
based on algebraic reconstruction techniques (ART) implemen-
ted in the TOMOJ software. The visualization and quantitative
analysis of the final volumes have been done by using the
ImageJ software. The statistic distribution of the particle sizes
has been built up by direct measurement of 50 nanocapsules
observed in the 2D slices extracted from reconstructed volume.
3. Results and discussion
3.1. Optimization of the polyenzymatic system FateDH–
FaldDH–YADH
Formate dehydrogenase (FateDH) from Candida boidinii, form-
Fig. 1 Relative activity as a function of pH for the 3 enzymes FateDH, (J),
aldehyde dehydrogenase (FaldDH) from Pseudomonas putida and
alcohol dehydrogenase (YADH) from Saccharomyces cerevisiae
FaldDH (&) and YADH (D) in 0.1 M potassium phosphate buffer at 25 1C for their
specific reductive reactions.
c
3724 New J. Chem., 2013, 37, 3721--3730
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
Fig. 2 (a) Influence of the relative amount of FateDH and FaldDH for the conversion of CO2(g) into formaldehyde at a fixed concentration of FateDH (0.1 g LÀ1) (72 h
with KHCO3 (100 mM) as the substrate). (b) Influence of the relative amount of FaldDH and YADH for the conversion of formate into methanol at a fixed concentration
of FaldDH (0.1 g LÀ1) (22 h with HCOOÀ (100 mM) as the substrate). All the reactions were carried out in phosphate buffer (0.1 M, pH 6.5) at 37 1C under a saturated N2
atmosphere with NADH (10 mM).
(CO2(g)) and hydrogen-carbonate, with a proportion of 45% of from formate (HCOOÀ) (Fig. 2b). The optimum ratio of the
CO2(g) and 55% of HCO3À at pH 6.5, we added KHCO3 instead three polyenzymatic systems FateDH, FaldH and YADH is
of bubbling CO2 for a better reproducibility of the reaction therefore 0.01, 0.15 and 0.75 g LÀ1 of commercial enzymatic
(Fig. S1, ESI†). The affinity constant of FateDH and FaldDH for powder, respectively (the conversion factor between commer-
CO2 is nearly two orders of magnitude lower than for HCOOÀ. cial enzymatic powder and pure enzymes is 0.16, 0.14 and 0.63
The Michaelis constants found for the oxidative and reductive
g
pure enzyme gcommercial enzymatic powderÀ1 for FateDH, FaldDH and
reactions of FateDH are KmCO = 30–50 mM c KmHCOOÀ = YADH, respectively). However, this optimum ratio differs from
2
0.5 mM,23,27 and for FaldDH KmHCOOÀ has never been the weight ratio previously used in the literature for this
described in the literature and KmHCHO = 0.09 mM.28 However polyenzymatic system, where for example the same mass of
the enzymes can be forced to catalyze their unfavorable each commercial powder was used.7,17,31 The reduction of CO2
reduction reaction by adding an excess of the NADH cofactor into methanol was performed in the previous experiments in
(NADH/NAD+ molar ratio Z 2000).29 Moreover care should be the presence of a concentration of NADH of 10 mM. We found
taken to measure the activity of the enzymes, which has been that the optimum concentration of NADH for this cascade
usually done by following the disappearance of the NADH reaction was 100 mM (Fig. 3) and corresponds to a productivity
band in the UV-Vis spectra at 340 nm, as we found that NADH in methanol of 0.05 mM in 65 h.
interacts with CO2 by most probably forming a complex
3.2. NADH cofactor regeneration systems
between NH2 groups of NADH and CO2 (as CO2 is very well
known to interact preferentially with amines),30 which dis-
places the band at 340 nm (Fig. S3, ESI†). This complexation
artificially increases the rate of the reaction. However another
UV-Vis band located at 290 nm characteristic of the probable
NADH–CO2 complex could be used to follow the consumption
of NADH. To be more accurate another analytic method (gas
chromatography or the Nash reagent) should be preferred for
this kind of reaction when using at the same time NADH and
CO2. In order to displace the reaction in favor of the reduction,
the addition of an excess of the second enzyme in the bienzy-
matic system is necessary. For the FaldDH–FateDH bienzymatic
system, an optimum activity corresponding to a productivity of
60 mM formaldehyde in 72 h was reached for an excess of
15 g LÀ1 of FaldDH for 1 g LÀ1 of FateDH (Fig. 2a).
The transformation of formaldehyde into methanol is more
favorable as methanol is not a suitable substrate for YADH
(KmMeOH = 130 mM). The equilibrium between MeOH and
HCHO is therefore naturally displaced toward the production
of MeOH. The optimization of the bienzymatic system YADH–
FaldDH has been performed. An excess of 5 g LÀ1 of YADH for
1 g LÀ1 of FaldDH was necessary to get the highest activity
corresponding to a productivity of 2.3 mM methanol in 22 h
NADH is an expensive cofactor, which is consumed stoichio-
metrically during the enzymatic reaction. Thus, for one molecule of
MeOH produced, 3 molecules of NADH will have to be consumed,
rendering the whole process prohibitive without recycling.7 Inter-
estingly, a photosystem PSII was attempted giving no by-products
except O2 produced from H2O.17 The authors claimed that the sole
Fig. 3 Influence of initial NADH concentration for CO2 conversion into metha-
nol with the optimized amount of the trienzymatic system: FateDH, FaldDH, and
YADH of 0.01, 0.15, and 0.75 g LÀ1, respectively (65 h at 37 1C in a phosphate
buffer (0.05 M, pH 6.5) with KHCO3 (0.05 M) as the substrate).
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem., 2013, 37, 3721--3730 3725

View Article Online
Paper
NJC
Scheme 2 Schematic representation of CO2 valorization into methanol using different NADH regenerating systems.
photosystem PSII was able to regenerate NADH with sunlight. GlyDH is more active at pH 8 to 9. GlyDH presents only 20% of its
However the real photosystem able to regenerate NADH is more activity at pH 6.5. Therefore, under the catalytic conditions needed
complicated. The active photosystem is in fact a chloroplast suspen- for the CO2 reduction into methanol (pH 6.5, 37 1C) PTDH is 4 times
sion containing five different proteins, which are present on thyla- more active than GlyDH (Fig. 4b). For the chloroplast photosystem,
koid membranes within the cell and can be extracted from spinach only 0.25 mM NADH was obtained in 30 minutes, which is below the
leaves.32,33 Several light harvesting complexes, an oxygen evolving NADH regenerating activity of PTDH. The low activity of the
complex, two photosystems (PSI and PSII), a plastoquinone pool, chloroplast could be explained by the poor stability of chloroplasts
cytochrome bf6 complex and a ferrodoxin reductase are required. with time in water and the oxidation of the NADH cofactor induced
The plastoquinone transfers electrons from one photosystem to the by the production of oxygen from the parallel reaction. An increase
other and the two essential macromolecular proteins are carrying in the natural photosystem stability could be achieved by immobi-
organometallic complexes bounded on each photosystem. The first lization of the photosystem and/or the thylakoid membrane into a
protein bounded to PSII contains an oxygen evolving complex (OEC), silica host.17 Recently, some stable inorganic photosystems based on
which can extract electrons from water and the second protein called hierarchical nanostructured carbon nitride materials (g-C3N4) pre-
ferrodoxin, bounded to PSI allows the cofactor regeneration with the pared by the Antonietti group11 have shown some efficiency to
concomitant use of a ferrodoxin reductase enzyme (EC 1-18-1-2).33,34 recycle NADH, but this system was active at pH higher than 8 and
In our study, we choose to compare this complex photosystem with formed by-products (oxidized triethanolamine). To be efficient at
two enzymes (Scheme 2): the phosphite dehydrogenase (PTDH), one pH 7, a mediator ([CpRh(bpy)(H2O)]) should be added, which is
of the most efficient enzymatic recycling systems reported so far known to be an inhibitor of some enzymes.36 Even if photosystems
(together with formate dehydrogenase12,13 but already used here in are very attractive, it is obvious that at this point, the most suitable
the opposite direction) which converts phosphite into phosphate, a system for NADH regeneration is based on phosphite dehydrogenase
chemical used to buffer the solution, and the glycerol dehydrogenase PTDH plus phosphite.
(GlyDH), which transforms glycerol into dihydroxyacetone (DHA, a
tan cosmetics). Glycerol is known to stabilize enzymes and could
also help to solubilize gaseous CO2 into water as CO2 is 11 times
3.3. Reduction of CO2 to methanol by the optimized polyenzymatic
system (FateDH, FaldDH, YADH, and PTDH) free in solution
more soluble in glycerol than in water (0.36 mol LÀ1 at 25 1C).35 The The regenerating system composed of PTDH plus phosphite
pH dependency of the two regenerating enzymes has been investi- was added in different amounts to the optimized polyenzymatic
gated (Fig. 4a). PTDH has an optimum activity at 6 o pH o 8 and system of FateDH, FaldDH and YADH in the presence of an
Fig. 4 (a) Relative activity of NADH formation from NAD+ with PTDH (J) and GlyDH (&) as a function of pH in 0.1 M potassium phosphate buffer at 37 1C at
À1
saturated substrate concentration ([Na2HPO3] or [Glycerol] = 0.5 M). At pH 6.5, VPTDH = 243 U mgpure enzyme and VGlyDH = 42 U mgpure enzymeÀ1; U = mmol minÀ1
.
(b) NADH production from 5 mmol NAD+ for the two regenerating enzymes (pH 7, 18 1C) corresponding to 50 mmol NAD+ per gpure enzyme and for the photosystem
to 3.3 mmol NAD+ per g of equivalent chlorophylls at pH 6.
c
3726 New J. Chem., 2013, 37, 3721--3730
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
inorganic supports as immobilization increases the conversion
of CO2 to methanol in comparison to free polyenzymatic
systems. This activity enhancement could arise from a better
stability of the enzymes or a better conformation of the
enzymes inside the inorganic support or from a better adsorp-
tion of CO2(g) around the inorganic particles and inside the
pores increasing the availability of CO2(g) for the enzymes.
Dave17 used a classical silica sol–gel encapsulation procedure
and Jiang and co-workers8 found that titania oxide nano-
Fig. 5 Influence of phosphite dehydrogenase (PTDH) amount for CO2 conver- particles templated by protamine were most suited for CO2
sion into methanol with the optimized trienzymatic system: FateDH, FaldDH, and
conversion. On our side, we have developed for several years
a new way of enzyme encapsulation using a modified silica
sol–gel procedure combined with a double protection of the
YADH of 0.01, 0.15, and 0.75 g LÀ1, respectively. Reactions were carried out for
65 h at 37 1C in phosphate buffer (0.05 M, pH 6.5) with KHCO3 (0.05 M), NADH
(10 mM) and Na2HPO3 (0.05 M).
enzymes by the phospholipid bilayer and lactose to avoid direct
contact with the silica. This encapsulation process has led to
initial amount of 10 mM NADH. The conversion of CO2 into high activity for several enzymes (lipases, GOx, HRP, and Hb)
methanol reaches a plateau of highest methanol productivity and to the discovery of phospholipid–silica nanocapsules,
for a PTDH content of 3.5 g LÀ1 (Fig. 5). The optimized which were very efficient for bienzymatic systems encapsula-
tion (GOx–HRP, GOx–Hb).20,21,37–40 The nanocapsule formation
composition of the polyenzymatic system is: FateDH, FaldDH,
YADH, and PTDH of 0.01, 0.15, and 0.75 g LÀ1 of commercial
is explained by, first the stacking of adjacent phospholipid
enzymatic powder, respectively, and 3.5 g LÀ1 of pure PTDH.
bilayers and then the hemifusion of the bilayers to form
interconnected phospholipid nanocapsules. After TEOS addi-
The most suitable NADH concentration was then investigated.
In the polyenzymatic system without a NADH regenerating tion, a silica shell is formed around the nanocapsules (Fig. 7).
system, the maximum conversion was obtained for 100 mM Electron tomography image slices with thicknesses of 0.7 nm
NADH (Fig. 3). In the presence of PTDH (Fig. 6), an excess of extracted from reconstructed volume (Fig. 7) reveal diameters
NADH increases drastically the rate of reaction (Fig. 6a) and the of nanocapsules between 20 and 40 nm, which are intercon-
nected by phospholipid bilayer bridges. The silica shell thick-
ness is between 3 and 5 nm around the nanocapsules and the
silica shell thickness decreases at the vicinity of the intercon-
nection and then vanishes. The phospholipid thickness inside
same optimum 100 mM NADH was found (Fig. 6b). However
under these conditions, the reaction ends after 48 h. Since
formate and formaldehyde were not detected, the accumulation
of methanol is likely the reason for the process to be inacti-
vated. The process could be therefore improved if continuous the silica nanocapsules is estimated between 3 and 6 nm. In the
removal of methanol could be performed.
SEM image, NPS appear as an aggregation of nanoparticles
(Fig. S5, ESI†) and the nitrogen adsorption isotherm shows an
interparticular porosity of ca. 20 nm (Fig. S6, ESI†). After
sonication in water, DLS measurements show that NPS have
an average size of 30 nm (Fig. 8) in accordance with the statistic
size of the nanocapsules determined from the tomography
analysis, meaning that the phospholipid interconnections
3.4. Encapsulation of the polyenzymatic system in
phospholipid–silica nanocapsules (NPS) and their bioactivity
Previous studies by Dave and co-workers7,17 and Jiang and
co-workers8,19 have demonstrated the importance to immobi-
lize the polyenzymatic system FateDH, FaldDH and YADH in
Fig. 6 (A) Methanol production as a function of time for two different initial NADH amounts: 10 mM and 100 mM, for the polyenzymatic system: FateDH/FaldDH/
YADH: 0.01/0.15/0.75 g LÀ1 of commercial enzymatic powder and 3.5 g LÀ1 of pure PTDH with Na2HPO3 (0.05 M). (B) Methanol production as a function of initial
NADH amounts: (a) for the system: FateDH/FaldDH/YADH: 0.01/0.15/0.75 g LÀ1 of commercial enzymatic powder and 3.5 g LÀ1 of pure PTDH, after 48 h of reactions
using Na2HPO3 (0.05 M); (b) after 65 h of reaction. All the reactions were carried out at 37 1C in phosphate buffer (0.05 M, pH 6.5) with KHCO3 (0.05 M).
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem., 2013, 37, 3721--3730 3727

View Article Online
Paper
NJC
Fig. 7 (up) Proposed mechanisms for the formation of phospholipids-templated silica nanocapsule (NPS) enzymes. (down) Electron tomography: slices redrawn from
the reconstructed volume of FateDH–FaldDH–YADH phospholipid–silica nanocapsules (NPS) and the observed statistic size distribution of the nanocapsules as
identified from the sections within the volume.
encapsulated in the NPS. The loading of enzymes in NPS
containing the three dehydrogenases without the regenerating
system was 27.5 mgpure enzymes gNPSÀ1. The encapsulation of the
3 dehydrogenases plus the regenerating enzyme PTDH has
been performed and the resulting NPS nanocapsules contain
43.9 mgpure enzymes gNPSÀ1. The catalytic activity of the polyenzymatic
system, used in the presence or absence of the regenerating
enzyme PTDH, is reported in Fig. 9. Addition of the regenerating
system increases in all cases the productivity in MeOH by a factor of
2 for the free enzymes or a factor of 5 in the case of encapsulated
enzymes. The encapsulation in NPS, in the presence or absence of
the regenerating enzyme PTDH, increases the bioactivity of the
polyenzymatic system by a factor of 10 or 27, respectively. The
Fig. 8 (a) XRD of FateDH–FaldDH–YADH phospholipid–silica nanocapsules (NPS)
highest productivity in methanol was obtained for the polyenzymatic
and (b) the particle size distribution of NPS by DLS after sonication.
between nanocapsules are fragile and can be disconnected. By
XRD a characteristic distance at ca. 8 nm, most probably
corresponding to the repeated distance of phospholipid bilayer
and silica shell, would correspond to a phospholipid bilayer
thickness of 4 nm if an average of 4 nm for the silica shell
thickness is taken, in accordance with tomography observation.
Some characteristic distances at about 5.1 and 3.2 nm are
detected and remain to be understood, which could result from
the impurities present in the lecithin.
The procedure was applied to encapsulate the optimized
polyenzymatic system FateDH, FaldDH and YADH in the phos-
Fig. 9 Production of methanol from CO2 after 3 h per mass of total commercial
enzymatic powders for FateDH, FaldDH, YADH trienzymatic system (0.01, 0.15,
0.75 g LÀ1, respectively) at 5 bar. (A) Free enzymes; (B) free enzymes with PTDH
addition (3.5 g LÀ1 of pure enzyme) to regenerate NADH; (C) enzymes encapsu-
lated in NPS, without PTDH; (D) enzymes and PTDH (3.5 g LÀ1 of pure enzyme)
encapsulated in NPS. Conditions: phosphate buffer (0.05 M, pH 6.5), NADH
(0.1 M) and Na2HPO3 (0.05 M) when PTDH is added.
pholipid–silica nanocapsules (NPS). TGA (Fig. S4, ESI†) shows
that NPS contains 470 mg of organics (mainly phospholipids)
per gram of NPS. The Bradford assay used to quantify the
enzymes in the supernatant showed no significant absorbance
at 595 nm after enzyme separation on a Sephadex-G25 exclusion
column. All enzymes used in the synthesis were therefore
c
3728 New J. Chem., 2013, 37, 3721--3730
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013

View Article Online
NJC
Paper
À1
system plus the regenerating enzyme PTDH encapsulated in slightly lower than the 4.3 mmol
NPS and leads to a productivity in MeOH of 42 mmÀo1l gNPS
gcommercial
enzymatic powder
MeOH
À1
obtained using NPS with PTDH as the NADH regenerating
corresponding to 4.3 mmol gcommercial enzymatic powder in 3 h system. In terms of CO2 consumed, however, the productivity
under 5 bar pressure.
achieved in the static system under 5 bar pressure is nearly
These results were compared to those reported earlier by other three orders of magnitude higher than in the open system with
groups. In terms of process, Dave17 and Jiang and co-workers19 CO2 bubbling.
used a continuous bubbling of CO2 into the reaction vessel. Such a
design allows to increase CO2 availability in water, but necessitates
high recycle rates to get significant CO2 reduction yields and costs
4. Conclusions
in the case of use of purified carbon dioxide. CO2 bubbling was not
possible to achieve in our case for the free polyenzymatic system
because a dense foam formed in the presence of YADH, flushing
away the solution from the reactor. We therefore used a static
process with a fixed initial CO2 amount. Jiang and coworkers8,19,41
have also tried bubbling under a pressure of 3 bar for a silica sol–gel
encapsulation and of 5 bar for alginate encapsulation. However the
best results they report in terms of activity were obtained for a
polyenzymatic system encapsulated in titanate/protamine particles
with a constant bubbling of CO2 at 1 bar. These authors demon-
strated that the type of encapsulation is more important than
pressure. In Fig. 10, we compared our results with those reported
by the above two groups. Jiang and coworkers8 in the absence of a
regenerating system with the polyenzymatic system encapsulated in
protamine templated titanate particles obtÀai1ned a methanol yield of
0.83 mmolMeOH gcommercial enzymatic powder close to that reached
with the NPS encapsulated system in the absence of NADH
regeneration (0.88 mmolMeOH gcommercial enzymatic powderÀ1). Dave,17
using a regenerating system composed of a ‘PSII’ suspensioÀn1,
obtained a MeOH yield of 3.3 mmolMeOH gcommercial enzymatic powder
The biocatalytic conversion of CO2 into methanol was investi-
gated using a cascade reaction of 3 dehydrogenases (formate
dehydrogenase
(FateDH),
formaldehyde
dehydrogenase
(FaldDH), and alcohol dehydrogenase (YADH)). To avoid any
accumulation of the different successive products, which can
denature the other enzymes, the rate of conversion of the third
enzyme has to be higher than the rate of conversion of the
second one, which has to be higher than the rate of conversion
of the first one. The optimized ratio for this polyenzymatic
system FateDH, FaldDH, and YADH was 0.01, 0.15, and 0.75 g
(commercial enzymatic powder) per L or to be more precise if
different batches or suppliers are used
a composition
corresponding to a ratio of 1, 4.4, and 790 Units per L. Units
(mmol minÀ1) have been calculated for each commercial powder
by defined standard assays. Different regeneration systems of
the NADH cofactor, among which enzymes and photosystems,
have been studied and the most efficient system was an enzyme,
the phosphite dehydrogenase (PTDH) producing phosphate
from phosphite as side reaction, which is very convenient in a
phosphate buffer medium. Furthermore PTDH is very active at
the optimum pH of the reaction (pH 6.5). Adding an efficient
regenerating system allows us to increase the activity of the
polyenzymatic system by a factor of 5. However the most
important improvement comes from the immobilization of the
polyenzymatic system into an inorganic matrix. By using an
efficient enzymatic encapsulation host such as phospholipids–
silica nanocapsules (NPS), the activity of the polyenzymatic
system was increased by a factor close to 30, leading to a
productivity in methanolÀ(i1n static at 5 bar for 3 h) of 4.3 mmol
gcommercial
equivalent to what was previously
enzymatic powder
reported using a continuous CO2 bubbling. The productivity in
methanol per mass of catalyst corresponds to 42 mmol methanol
produced per gNPS in 3 h. The reaction rate decreases after some
time, due most probably to the accumulation of methanol. To
improve this biocatalytic conversion process, continuous flow
catalysis should be performed to extract methanol in order to
avoid the denaturing effect of methanol on the enzymes. The
limitation associated with the low solubility of CO2 in water
should be overcome by using other media than water, such as
ionic liquids, glycerol, polyethylene glycol, supercritical CO2 or to
run the bioconversion in a continuous gas phase.
Fig. 10 Comparison with literature results of the production of methanol from
CO2 after 3 h per mass of total commercial enzymatic powder for FateDH,
FaldDH, YADH trienzymatic systems encapsulated with and without a NADH
regenerating system. (A) Jiang and co-workers:18 FateDH/FaldDH/YADH (1.35,
1.35, 0.3 g LÀ1) encapsulated in protamine templated-TiO2 particles and con-
tinuous bubbling of CO2; (B) enzymes encapsulated in NPS (this work) at 5 bar of
CO2; (C) Dave:17 FateDH/FaldDH/YADH (5, 5, 5 g LÀ1) encapsulated in silica
sol–gel and continuous bubbling of CO2; (D) Dave:17 FateDH/FaldDH/YADH
(5, 5, 5 g LÀ1) and a regenerating system (PSII) encapsulated in silica sol–gel
and continuous bubbling of CO2; (E) enzymes and PTDH regeneration system
encapsulated in NPS (this work) at 5 bar. Our conditions: phosphate buffer
(0.05 M, pH 6.5), NADH (0.1 M) and with Na2HPO3 (0.05 M) when PTDH is added,
FateDH/FaldDH/YADH: (0.01, 0.15, 0.75 g LÀ1) (g of commercial enzymatic
powder) and PTDH of 3.5 g LÀ1 (g of pure enzyme).
Acknowledgements
The authors thank Dr Joel Chopineau and Dr Eric Dubreucq
for fruitful discussions and Dr Huimin Zhao from the Univer-
sity of Illinois, Department of chemical and biomolecular
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013
New J. Chem., 2013, 37, 3721--3730 3729

View Article Online
Paper
NJC
21 L. T. Phuoc, P. Laveille, F. Chamouleau, G. Renard, J. Drone,
B. Coq, F. Fajula and A. Galarneau, Dalton Trans., 2010, 39,
8511–8520.
engineering, for providing us with the pET15b-PTDH12x
construct. This work was financially supported by the Region
Languedoc-Roussillon through the ‘‘Chercheur d’Avenir
22 S. K. Lower, Carbonate equilibria in natural waters, Simon
Fraser University, 1999, pp. 1–26.
´
Confirme’’ grant and by the CNRS. The authors acknowledge
financial support from the French METSA network (CNRS-
FR3507 and CEA).
¨
23 U. Rusching, U. Mu¨ller, P. Willnow and T. Hopner, Eur. J.
Biochem., 1976, 70, 325–330.
24 D. A. Berthold, G. T. Babcock and C. F. Yocum, FEBS Lett.,
1981, 134, 231–234.
References
1 M. Aresta and A. Dibenedetto, Catal. Today, 2004, 98, 25 T. Nash, Biochem. J., 1953, 55, 416–421.
455–462.
26 J. Shi, X. Wang, Z. Jiang, Y. Liang, Y. Zhu and C. Zhang,
2 S. Klaus, M. W. Lehenmeier, C. E. Anderson and B. Rieger,
Coord. Chem. Rev., 2011, 255, 1460–1479.
3 C. Song, Catal. Today, 2006, 115, 2–32.
4 G. A. Olah, A. Goeppert and G. K. S. Prakash, J. Org. Chem.,
2009, 74, 487–498.
5 S. N. Riduan, Y. Zhang and J. Y. Ying, Angew. Chem., Int. Ed.,
2009, 48, 3322–3325.
6 C.-C. Lo, C.-H. Hung, C.-S. Yuan and Y.-L. Hung, Chin. J.
Catal., 2007, 28, 528–534.
Bioresour. Technol., 2012, 118, 359–366.
27 T. Schmidt, C. Michalik, M. Zavrel, A. Spieß, W. Marquardt
and M. B. Ansorge-Schumacher, Biotechnol. Prog., 2010, 26,
73–78.
28 S. Ogushi, M. Ando and D. Tsuru, Agric. Biol. Chem., 1984,
48, 597–601.
29 U. Ruschig, U. Muller, P. Willnow and T. Hopner, Eur. J.
Biochem., 1976, 70, 325–330.
30 N. McCann, D. Phan, X. G. Wang, W. Conway, R. Burns,
M. Attalla, G. Puxty and M. Maeder, J. Phys. Chem. A, 2009,
113, 5022–5029.
7 R. Obert and B. C. Dave, J. Am. Chem. Soc., 1999, 121,
12192–12193.
8 Q. Sun, Y. Jiang, Z. Jiang, L. Zhang, X. Sun and J. Li, Ind. Eng. 31 B. C. Dave, M. S. Rao and M. C. Burt, WO Pat., 2007/022504,
Chem. Res., 2009, 48, 4210–4215. 2007.
9 F. S. Baskaya, X. Y. Zhao, M. C. Flickinger and P. Wang, 32 S. O. W. Vishniac, J. Biol. Chem., 1952, 195, 75–93.
Appl. Biochem. Biotechnol., 2010, 162, 391–398.
10 Y. H. Kim and Y. J. Yoo, Enzyme Microb. Technol., 2009, 44,
129–134.
33 J. P. Dekker and E. J. Boekema, Biochim. Biophys. Acta,
–Bioenerg., 2005, 1706, 12–39.
34 W. Vredenberg, Biosystems, 2011, 103, 138–151.
11 J. Liu and M. Antonietti, Energy Environ. Sci., 2013, 6, 1486–1493. 35 O. Aschenbrenner and P. Styring, Energy Environ. Sci., 2010,
12 A. Weckbecker, H. Groger and W. Hummel, Biosystems Engi-
neering I: Creating Superior Biocatalysts, 2010, pp. 195–242.
13 T. W. Johannes, R. D. Woodyer and H. M. Zhao, Appl.
Environ. Microbiol., 2005, 71, 5728–5734.
14 R. Xue and J. M. Woodley, Bioresour. Technol., 2012, 115,
183–195.
15 C. Hold and S. Panke, J. R. Soc., Interface, 2009, 6, S507–S521.
16 Z. Duan and R. Sun, Chem. Geol., 2003, 193, 257–271.
17 B. C. Dave, US Pat., 6,440,711, 2002.
18 Y. J. Jiang, Q. Y. Sun, L. Zhang and Z. Y. Jiang, J. Mater.
Chem., 2009, 19, 9068–9074.
3, 1106–1113.
36 M. Poizat, I. Arends and F. Hollmann, J. Mol. Catal. B:
Enzym., 2010, 63, 149–156.
37 M. Mureseanu, A. Galarneau, G. Renard and F. Fajula,
Langmuir, 2005, 21, 4648–4655.
38 A. Galarneau, M. Mureseanu, S. Atger, G. Renard and
F. Fajula, New J. Chem., 2006, 30, 562–571.
39 A. Galarneau, G. Renard, M. Mureseanu, A. Tourrette,
C. Biolley, M. Choi, R. Ryoo, F. Di Renzo and F. Fajula,
Microporous and Mesoporous Mater., 2007, 104, 103–114.
40 A. Galarneau, F. Sartori, M. Cangiotti, T. Mineva, F. Di
Renzo and M. F. Ottaviani, J. Phys. Chem. B, 2010, 114,
2140–2152.
19 S.-w. Xu, Y. Lu, J. Li, Z.-y. Jiang and H. Wu, Ind. Eng. Chem.
Res., 2006, 45, 4567–4573.
20 P. Laveille, L. T. Phuoc, J. Drone, F. Fajula, G. Renard and 41 Z. Y. Jiang, H. Wu, S. W. Xu, S. F. Huang and Q. Xie, Chin. J.
A. Galarneau, Catal. Today, 2010, 157, 94–100. Catal., 2002, 23, 162–164.
c
3730 New J. Chem., 2013, 37, 3721--3730
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013