Enzymatic cleavage of lignin β-O-4 aryl ether bonds via net internal hydrogen transfer

Article abstract of DOI:10.1039/c3gc40295a

The current greening of chemical production processes going along with a rising interest for the utilization of biogenic feedstocks recently revived the research to find new ways for the degradation of the complex lignin-backbone by means of biocatalysis and combined chemo-enzymatic catalysis. Lignin, which accumulates in 50 million t/a, is regarded as a potential substitute for phenolic and other aromatic, oil-based chemicals in the upcoming post oil age. The cleavage of the β-O-4-aryl ether linkage is the most favoured, since it accounts for approximately 50% of all ether linkages in lignin. This enzymatic cleavage was proposed to be a part of the lignin catabolism in the proteobacterium Sphingobium sp. SYK6. Three enzymes, LigD, a Cα-dehydrogenase, LigF, a β-etherase and LigG, a glutathione lyase, are supposed to be involved in lignin degradation. We cloned and recombinantly expressed these genes in E. coli and determined their pH and temperature optima on the lignin model substrate 1-(4-hydroxy-3-methoxyphenyl)-2-(2-methoxyphenoxy) -1,3-propanediol 1. Using an NAD⁺ dependent glutathione reductase from Allochromatium vinosum (AVR) we established an efficient way to regenerate the co-substrates NAD⁺ and glutathione allowing for a self-sufficient balanced enzymatic cascade with net internal hydrogen transfer (hydrogen borrowing). We showed the capability of this enzyme system to release lignin monomers from complex lignin structures coming from differently prepared real lignin substrates. This novel enzyme system could become a useful tool to release lignin monomers from complex lignin structures.

Full text of DOI:10.1039/c3gc40295a

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Enzymatic cleavage of lignin β-O-4 aryl ether bonds via
net internal hydrogen transfer†
Cite this: Green Chem., 2013, 15, 1373
Jochen Reiter,a,b Harald Strittmatter,b Lars O. Wiemann,b Doris Schiedera and
Volker Sieber*a,b
The current greening of chemical production processes going along with a rising interest for the utiliz-
ation of biogenic feedstocks recently revived the research to find new ways for the degradation of the
complex lignin-backbone by means of biocatalysis and combined chemo-enzymatic catalysis. Lignin,
which accumulates in 50 million t/a, is regarded as a potential substitute for phenolic and other aromatic,
oil-based chemicals in the upcoming post oil age. The cleavage of the β-O-4-aryl ether linkage is the
most favoured, since it accounts for approximately 50% of all ether linkages in lignin. This enzymatic clea-
vage was proposed to be a part of the lignin catabolism in the proteobacterium Sphingobium sp. SYK6.
Three enzymes, LigD, a Cα-dehydrogenase, LigF, a β-etherase and LigG, a glutathione lyase, are supposed
to be involved in lignin degradation. We cloned and recombinantly expressed these genes in E. coli and
determined their pH and temperature optima on the lignin model substrate 1-(4-hydroxy-3-methoxy-
phenyl)-2-(2-methoxyphenoxy)-1,3-propanediol 1. Using an NAD+ dependent glutathione reductase from
Allochromatium vinosum (AVR) we established an efficient way to regenerate the co-substrates NAD+ and
glutathione allowing for a self-sufficient balanced enzymatic cascade with net internal hydrogen transfer
(hydrogen borrowing). We showed the capability of this enzyme system to release lignin monomers from
complex lignin structures coming from differently prepared real lignin substrates. This novel enzyme
system could become a useful tool to release lignin monomers from complex lignin structures.
Received 10th February 2013,
Accepted 19th March 2013
DOI: 10.1039/c3gc40295a
www.rsc.org/greenchem
(EC 1.10.3.2) are copper-containing enzymes mostly produced
by fungi.3 Using molecular oxygen as a final oxidant these
Introduction
Lignin is a branched polymer of high complexity that consists enzymes oxidize phenolic molecules via single electron trans-
of substituted phenylpropane units joined by carbon–carbon fers from the aromatic compounds to enzyme bound copper(II)
and ether linkages. The mechanism of lignin biosynthesis pro- ions forming oxidized aromatic radicals and enzyme bound
ceeds via polymerization of free radical forms of the three copper(I) ions. The latter are reoxidized by molecular oxygen.
monolignol precursors. The monomers include coniferyl, The oxidation capability of laccases results in breakage of aro-
p-coumaryl and sinapyl alcohols that are formed by radical matic and aliphatic C–C bonds and thus depolymerization of
mediated oxidative coupling of several different interunit lin- lignin.4 These enzymatic reactions find application in a variety
kages.1 The β-O-4 aryl ether linkages are the most abundant of technical fields including pulp and paper, textile and food
followed by β–β, β–5, 5–5 and 5-O-4 linkages.2 These binding industries. For instance laccases are supplemented to wood
types are specifically cleaved in vivo by soil microorganisms chips as biopulping agents to degrade lignin and loosen lignin
that play important roles in the planet’s carbon cycle. The structures.5 The separation of lignin from cellulose fibers is an
non-phenolic enzyme cleavage distinguishes this enzymatic important step in the processing of wood for manufacturing of
approach from the well investigated oxidative but unspecific paper pulp.5,6 The textile industry applies cell-free enzyme
laccase or peroxidase based lignin treatments. Laccases (Lac) extracts originating from numerous white-rot fungi which are
effective in the biodecolourization of many synthetic and
industrial dyes.7 Lignin peroxidase (LiP, EC 1.11.1.14) involves
aChair of Chemistry of Biogenic Resources, Technische Universität München,
Schulgasse 16, 94315 Straubing, Germany. E-mail: sieber@tum.de;
Fax: +49 9421 187 242; Tel: +49 9421 187 310
H2O2 as the co-substrate in addition to a mediator like veratryl
alcohol to degrade lignin and other phenolic compounds.
H2O2 is reduced to H2O by gaining an electron from the then
oxidized LiP. By acquisition of an electron from veratryl
alcohol and oxidizing it to veratryl aldehyde, oxidized LiP
returns to its native reduced state. Subsequently veratryl
bFraunhofer Institute for Interfacial Engineering and Biotechnology IGB Project
Group BioCat, Schulgasse 11a, 94315 Straubing, Germany
†Electronic supplementary information (ESI)
available. See DOI:
10.1039/c3gc40295a
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 1373–1381 | 1373

View Article Online
Paper
Green Chemistry
aldehyde gets reduced to its alcohol by gaining electrons from
lignin or analogous structures such as xenobiotic pollutants.8,9
Mangan peroxidase (MnP, EC 1.11.1.13) requires H2O2 as a co-
substrate as well as Mn2+. Oxidation of Mn2+ to Mn3+ leads to
Mn3+ chelate-oxalate, which consecutively oxidizes the pheno-
lic substrates forming radical cations that lead to bond break-
age and depolymerization. Mn3+ chelate-oxalate has the ability
to diffuse into areas inaccessible to the enzyme. This principle
allows degradation of lignin or xenobiotic pollutants, which
are not essentially available for the enzymes.9,10 Decolouriza-
tion of molasses spent wash (MSW), a by-product of sugar
mills and alcohol distilleries, and black liquor was shown to
be MnP-dependent in Phanerochaete chrysosporium.9,11
A new family of ligninolytic peroxidases is versatile per-
oxidase (VP, EC 1.11.1.16), which combine the substrate speci-
ficity of LiP and MnP thus being capable of using both Mn2+
as well as veratryl alcohol and other mediators like dimethoxy-
benzenes, different types of dyes, substituted phenols, and
hydroquinones.12,13 These established lignin degrading
enzymes are characterized by their technically unresolved lack
of cleavage specificity and their high predisposition to provoke
the repolymerisation of previously released monolignols. A
different approach using hydrolases (specifically lipases EC
3.1.1.3) to form peracids in situ in the presence of H2O2, first
described by Björkling and co-workers in the early 90s, was
applied to oxidize lignin and thus separate it from the cellulo-
sic fraction.14 Most recently Wiermans et al. (2013) advanced
this lipase-based approach and reported on the successful oxi-
dation of beech wood lignin into a non-aromatic lignin oil
using dimethyl carbonate as a solvent and immobilized
Candida antarctica lipase B as an enzyme.15
Recently Chen et al. showed the capability of Novosphingo-
bium sp. B-7, a close relative of Sphingobium sp. SYK-6, to
release lignin monomers with MnP and Lac from kraft lignin.
The effectiveness of the β-O-4, biphenyl-, coumaryl-, pinoresi-
nol-lignin degrading enzyme cascade of this strain could
not be elucidated.16 Specific cleavage of β-aryl ether bonds
accounting for up to 50% of the ethers in lignin possibly will
result in the release of monolignols and therefore be a versatile
source for widely used phenolic compounds in the chemical
industry. In this work, we characterize and apply a β-O-4 aryl-
ether cleaving enzyme system from the proteobacterium Sphin-
gobium sp. SYK-6, whose genome sequence was recently pub-
lished, on various lignin containing substrates.17 The three
proteins involved are LigD, a Cα-dehydrogenase, LigF, a beta-
etherase and LigG, a glutathione lyase. The Cα-dehydrogenase
LigD oxidizes the Cα of 1-(4-hydroxy-3-methoxyphenyl)-2-(2-
methoxyphenoxy)-1,3-propanediol 1 under consumption of
NAD+ to 3-hydroxy-1-(4-hydroxy-3-methoxyphenyl)-2-(2-methoxy-
Fig. 1 Reaction mechanism of the LigDFG enzyme system. LigD oxidizes the Cα
of 1 under consumption of NAD+. LigF cleaves the intermediate 2 to 3 and 4
with attachment of glutathione at the Cβ position. Glutathione is oxidized by
LigG and releases 5.
the omega class of glutathione transferases. Its hydrophilic
substrate binding site and its known ability to stereoselectively
deglutathionylate 4 broadened the definition of the omega
class.19 The lignin catabolism in Sphingomonas sp. is stereo-
specific, with another three enzymes responsible for the degra-
dation of the arylglycerol-β-aryl ether threo-isomers.20 LigD
converts just the (αR,βS)- and the (αR,βR)-isomers of 1 into the
(βS)- and the (βR)-isomers of 2, whereas the two other enzymes
LigL and LigN catalyze the conversion of the (αS,βR)- and the
(αS,βS)-isomers of 1 to the (βR)- and the (βS)-isomers of 2,
respectively. In the next step, the (βS)-isomer of 2 is specifically
processed and cleaved by LigF and the (βR)-isomer of 2 is the
stereospecific substrate for LigE and LigP.20,21 In this work we
analyzed the activity of the isolated enzyme system on the
model substrate 1 and 7-(2-(4-benzyloxy)-3-methoxy-phenyl)-2-
oxoethoxy)-4-methyl-2H-1-benzopyran-2-one 6.22 For the first
time, we overproduced and purified recombinant LigD, LigF
and LigG in E. coli and determined their enzymatic activities.
With this knowledge we studied the capability of this enzyme
system to release lignin monomers from complex lignin struc-
tures coming from differently prepared real lignin substrates,
such as kraft-lignin and organosolv lignin and looked at the
macromolecular size distribution after enzymatic treatment.
We also demonstrate the effectiveness of an Allochromatium
vinosum reductase (AVR) to recycle and supply this enzyme
system with the expensive co-substrates NAD+ and glutathione.
Materials and methods
Strains, plasmids and chemicals
phenoxy)-1-propanone 2. 2 is then cleaved at its Cβ ether to 2- The four genes were synthesized by Geneart GmbH (Regens-
glutathionyl-3-hydroxy-1-(4-hydroxy-3-methoxyphenyl)-1-propa- burg/Germany) with codon optimization for E. coli., and
none 4 and 2-methoxy-phenol 3, while glutathione is released cloned into a pET28b vector with a C-terminal His-tag to trans-
from 4 by the glutathione lyase LigG into 3-hydroxy-1-(4- form E. coli BL21 (DE3). Strains for all experiments were grown
hydroxy-3-methoxyphenyl)-1-propanone 5 and oxidized gluta- in LB medium with kanamycin (50 mg L−1) from OD 0.1 to OD
thione (GSSG) (Fig. 1).18 The X-ray structure of LigG was 0.6 at 37 °C and induced with 1 mM isopropyl-β-D-thiogalacto-
recently published and indicates that this enzyme belongs to pyranoside (IPTG). All chemicals were bought from Sigma
1374 | Green Chem., 2013, 15, 1373–1381
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
Aldrich. 1 was purchased from TCI Japan, with 98% purity, no Kinetics of LigDFG with compound 1
further data about isomer composition are available. Bagasse
The enzymes LigDFG were added with a final concentration of
0.5 mg mL−1 to substrate 1, NAD+, glutathione, each 0.1 mM
in 3 mL ammonium acetate buffer pH 9. Aliquots of 250 μL
were taken at 0 min, 10 min, 20 min, 30 min, 40 min, 50 min,
70 min, 100 min, 130 min and 12 h and stopped by filtration
with a 10 kDa filter (VWR). Triplicates were analysed via HPLC
(Shimadzu, CX18-column, Macherey&Nagel, Germany) with a
10 mM, pH 4.3 ammonium acetate buffer-/acetonitrile gradi-
ent. Commercially available 3 (Sigma, U.S.) and substrate 1
(TCI, Japan) samples were used for standardization.
organosolv lignin was obtained by the Fraunhofer ICT
(Germany), softwood and hardwood alkali lignin was kindly
sponsored as Lignex® powder by the Chemische Werke Zell-
Wildshausen GmbH (Germany).
Enzyme purification and activity assay
For the purification of enzymes, over-expressed cells were lysed
in binding buffer (Tris/HCl pH 7.2, 10% (v/v) glycerol, 20 mM
imidazole) in a cell disruption system (ThermoFisher, U.S.)
and the cell lysate was applied to a Ni2+-NTA column of the
Aekta® protein purification system (GE Healthcare, U.S.) at a
flow rate of 0.3 mL min−1 and then fractionated in elution
buffer (20 mM Tris/HCl pH 7.2, 10% (v/v) glycerol and 500 mM
imidazole). The elution fractions were analysed by SDS-PAGE
and Coomassie staining to test the purity. The protein concen-
tration of every fraction was determined by Bradford.23 5 μg of
each protein was added to 0.15 mM of substrate 1, 0.5 mM of
NAD+, 0.1 mM of glutathione in 1 mL of 20 mM Tris/HCL, pH
7.0 reaction mixture, filtered with a 0.2 μm PVTE filter and
analysed with HPLC. The same reaction was used with 10 mM
ammonia acetate buffer for analysing cleavage products with
LC-MS (Ultimate3000, Dionex, U.S.) with a flow rate of 0.3 mL
min−1 with ammonia acetate buffer/acetonitrile as mobile
phase.
Activity of LigG
Substrate 1, NAD+, glutathione, each 10 mM in 6 mL piper-
azine/glycylglycine buffer pH 9 were initially incubated with
LigDF to produce thioether 4 and filtered. The enzyme LigG
was added to the compound 4 stock solution with a final con-
−1
centration of 60 μg mLLigG and to different pH probes (pH
4–12) with a final volume of 1.5 mL. Aliquots of 300 μL were
taken at 0 min and 2 min, and stopped by filtration with a
10 kDa filter (VWR, Germany). Triplicates of this experiment
were analysed by HPLC as described.
Co-substrate recycling by AVR
AVR was produced in E. coli BL21 (DE3) and purified with a
His-tag. Activity was tested photometrically at 340 nm via
decrease of NADH concentration. The enzyme system LigDFG
with substrate 1, NAD+ and glutathione was incubated with
AVR. Addition of AVR led to decrease of NAD+ as 1 was the lim-
iting reaction partner. HPLC samples were filtered with 10 kDa
(VWR, Germany) and applied to HPLC as shown above.
Enzyme assays with LigD
Assays were performed in microplates for the determination of
pH-optima in CBP-Universal buffer (pH 4–12). Oxidation of
the Cα of substrate 1 was measured spectrophotometrically by
following the reduction of NAD+ to NADH at 340 nm. The
enzyme activity was analysed with different pH values in multi-
well plates in a Varioscan photometer (Thermo Scientific, U.S.)
with 0.5 mM of substrate 1 and 0.5 mM of NAD+. One enzyme
unit corresponds to the amount of enzyme that catalyzes the
oxidation of 1 μmol of 1 per minute. The kinetics was moni-
tored at pH 9, 45 °C pre-incubation with 0.125 mM, 0.25 mM,
Lignin analytics
The concentrations of guaiacol 3, ferulic acid 9, eugenol 10,
acetovanillone 11 and vanillin 12 in aqueous 200 g L−1 lignin
samples with 1 mg LigDFG were determined using a BPX5
column (SGE, 30 m, I.D. 0 : 25 mm, film 0 : 25 μm) connected
to a QP2010 Plus gas chromatograph with a Single Quad MS-
detector (both Shimadzu, Japan) and helium as carrier gas.
The samples were extracted with an SPME fiber (85 μm poly-
acrylate, Supelco/Sigma Aldrich, U.S.) on an autosampler
(CombiPal, Swiss) for 30 min at 60 °C and desorbed for 5 min
in an SSL-injector (290 °C) starting at 80 °C for 1 min, then
heated to 170 °C at 6 °C per min, then to 280 °C at 15 °C per
min and finally held for 6 min. Area integration of the tripli-
cates was calibrated with the area averages of pure substance
concentrations.28
0.5 mM, 1.0 mM, 1.5 mM, 1.75 mM and 2.5 mM
concentrations.
1
Synthesis of substrate 6
Synthesis of the lignin model compound 6 was performed
according to Weinstein, starting from 1-(4-hydroxy-3-methoxy-
phenyl)ethanone. After protection of the hydroxy-function and
bromination of the acyl group the compound was coupled
with compound 8.24
Enzyme assay with LigF
Cleavage of the reference substance 6 was analyzed fluorimetri-
GPC samples were filtered with 45 μm cellulose acetate
cally with an excitation at 360 nm and an emission at filters (VWR, Germany) and applied to HPLC upgraded with
450 nm.22 0.25 mM of 6 was cleaved in a multiwell plate at GPC (Shimadzu, Japan) with a Mcx10 column (PSS, Germany).
25 °C with dilutions of LigF and the release of compound 8 A 0.01 M phosphate buffer with 10% methanol was used with
was detected and calibrated with commercial compound 8 99% methanol as a solvent. The isocratic method (20% metha-
(Sigma Aldrich, U.S.) as the standard.
nol) had a flow rate of 0.9 mL min−1 at 60 °C.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 1373–1381 | 1375

View Article Online
Paper
Green Chemistry
to convert the (−)(αSβS)-isomer of 1 which is present in the
racemic commercial sample of 1. This is in accordance with the
reported high substrate specificity of LigD towards the (+)(αRβS)-
Results and discussion
Cloning in E. coli
The four target genes from the soil proteobacterium Sphingo- and the (−)(αRβR)-isomers of 1.21 Being such a highly specific
monas sp. SYK6 were commercially synthesized with codon secondary alcohol dehydrogenase LigD might be of interest
optimization for E. coli. The genes were cloned into a pET28b for fine chemical conversions. We therefore tested LigD on six
vector with a C-terminal His-tag under a lac promoter and commercially important chiral alcohols (1-phenyl-2-propanol,
used to transform E. coli BL21(DE3).
4-phenyl-1-butanone, 6-methyl-5-hepten-2ol, for the oxidation)
and 4-propylbenzaldehyde or ketones (3,4-dimethylbenzaldehyde,
acetophenone, 4-phenyl-2-butanone and 6-methyl-5-hepten-2-
one) for the reduction. No activity could be detected (data not
shown). The substrate and stereo-specificity of the active site of
this enzyme seem to be strongly restricted.
Purification and proof of principle for the overall activity of
the enzyme system
Subsequently, all three genes were overexpressed in E. coli
BL21(DE3) and enzymes purified with their C-terminal His-tag
on a Ni2+-NTA column with an automated protein purification
system. SDS-PAGE showed distinct and clean bands of all three
proteins with elution fractions of 5–20 mgprotein mL−1 with
33 kDa, 28.5 kDa and 33 kDa respectively (Fig. S1†). The
overall activity of the enzyme system after the purification step
was analyzed by detection of the cleavage products of the
lignin model substrate 1 by means of HPLC and LC-MS. After
one hour reaction time, an enzymatic cleavage of 37% of 1 was
observed by HPLC as shown in Fig. S2.† LC-MS confirmed the
mass peak of 1 and the formation of compound 5, when cofac-
tors and enzymes were present in the enzyme assay (Fig. S2,
S3†). The recombinant enzyme system LigDFG is therefore able
to cleave the beta-aryl ether of 1.
Determination of pH and temperature optima and kinetics
The purified enzymes were tested individually to get new
insights into the dependencies between enzymatic activity and
pH as well as temperature.
LigD. The activity of LigD, a Cα-dehydrogenase, was deter-
mined as mentioned above. The highest activity of 11.7 U
−1
mgLigD was shown to be at pH 9, above which there is a
strong decrease in activity possibly due to enzyme denatura-
tion (Fig. 4A). The optimum temperature was determined to be
−1
at 60 °C, pH 9 with 12.0 U mgLigD and almost complete in-
activation of LigD activity at 70 °C (Fig. 4B). These prerequisites
allowed the determination of the kinetics of LigD. Under vari-
ation of 1 from 0.125 to 2.5 mM the vmax/2 was determined to
be 3.93 U, resulting in a Km-value of 1.10 mM. The correspond-
ing kcat was 1330 s−1 resulting in an efficiency (kcat/Km) of
1204.5 mM−1 s−1 (Fig. 4C).
Activity of LigDFG
Consumption of substrate 1 and formation of compound 3 of
the enzyme system were determined as mentioned above. Cali-
bration with a standard of 1 and 3 and integration of the product
peak showed a decrease of the concentration of 1 in the enzyme
assay probe from 78.7 μmol to 50.9 μmol within one hour. This
corresponds to an enzymatic activity of 10.3 U mg−1. 3, the
product of the beta-aryl ether cleavage, increased at the same
time from 0.8 μmol to 27.8 μmol (Fig. 2). The maximum conver-
sion of 1 after 10 h was ca. 35%. We assume LigD being unable
LigF. For the activity determination of LigF, a glutathione
transferase, the model substance 6 (Fig. 3) was used. The
activity was determined at different pH-values and calibrated
with commercial compound 8. The highest activity of LigF was
determined at room temperature with pH 10 having 5.6 mU
−1
−1
mgLigF and increasing to 42 mU mgLigF at 60 °C (Fig. 5A
and 5B). Kinetics were determined under variation of 6 from
0.0625 to 0.75 mM resulting in a vmax/2 of 0.018, in a Km of
0.24 mM. The kcat was calculated to be 0.075 s−1 with a kcat/Km
of 3.19 mM s−1 (Fig. 5C).
LigG. The activity of LigG was detected indirectly via HPLC
by following the formation of 5 (Fig. 6). 4 as a substrate for
LigG is not commercially available and lacks photo- or fluori-
metrical detection capability to set up an easy assay. Therefore
LigG activity was tested for pH- and temperature optima via
Fig. 2 Kinetics of LigDFG with the lignin model substance 1. The triplicates
were tested in parallel and analysed via HPLC. Concentrations of 1 and 3 were
calibrated with reference substances, transformation of 1 by ligD is limited to
the (+)(αRβS)- and the (−)(αRβR)-isomers of 1. The erythro (−)(αSβS)-isomer of 1
isomer is not used by LigD.
Fig. 3 Cleavage of the model substance 6. LigF activity is determined fluori-
metrically with the synthesized reference substance 6 by detection of released
compound 8.
1376 | Green Chem., 2013, 15, 1373–1381
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
Fig. 4 Activities of LigD via detection of NAD+ formation at 340 nm at different pH (A) and temperature (B) in U mg−1 protein at (n = 3). The enzyme activities fol-
lowed Michaelis–Menten under increasing substrate conditions (n = 3) (C).
Fig. 5 Activities of LigF through cleavage of substrate 6 at Ex 360 nm, Em 450 nm at different pH (A) and temperatures (n = 3) (B). LigF kinetics with varying sub-
strate 6 concentrations (n = 3) (C).
Fig. 6 Activities of LigG via detection of the formation of 5 with HPLC in ΔArea at different pH (n = 3) (A). LigG shows highest activities at 60 °C (n = 3) (B). Indirect
LigG activity measurement via HPLC. The non-commercial intermediate 4 is catalysed by formation of 5 (n = 3).
HPLC without calibration. LigDF with cofactors was used to as shown in Fig. 6A. Highest activity was monitored at 60 °C
convert 1 to 4. Enzymes were removed and this highly concen- with 427.424 areas min−1 (Fig. 6B). A more detailed quantitat-
trated master mix was used for smaller pH assays (pH 4–12). ive analysis requires higher amounts of the thio-ether inter-
Due to the broad pH-range for LigG activity the wide range mediate as a substrate which was not available.
buffer piperazine/glycylglycine was used. The reactions were
Co-substrate recycling – a proof of principle
stopped by filtering aliquots to remove LigG to analyse the
initial conversion rates via HPLC. The kinetics shows a broad
The enzyme cascade LigDFG catalyzes the reductive cleavage of
activity plateau from pH 4 to pH 11 with its optimum at pH 7
an ether bond and oxidation of a secondary alcohol via the
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 1373–1381 | 1377

View Article Online
Paper
Green Chemistry
conversion of the cofactors NAD+ to NADH and two molecules reducing GSSG to glutathione via production of NAD+
of GSH to GSSG. For technical applications it is essential to (Fig. 8A). To evaluate the ability of AVR to regain NAD+ and
efficiently recycle the cofactors to reduce the cost. We used an GSH, 30 mM conversion of 1 was conducted with 3 mM initial
NADH dependent glutathione reductase from Allochromatium NAD+ and 7 mM GSSG. HPLC analysis showed significantly
vinosum (AVR) to regenerate the cofactor(s) of the LigDFG decreased amounts of 1 in the samples containing AVR com-
enzyme system. AVR was first described by Chung and Hurl- pared to the control probes without AVR (Fig. 8B). Based on
bert.25 We recombinantly expressed the enzyme, characterized the initial concentrations of 1 and coenzyme these results
it and found it to accept NADH instead of the common and clearly demonstrate the successful recycling of NAD+ and GSH.
more expensive NADPH as a substrate (manuscript in prepa- After seven days of reaction, 1 was eventually cleaved to the
ration). In a first experiment we determined the cleavage of 1 previously determined maximal conversion of 1 of around
by following the formation of NADH with and without AVR in 35% due to the stereospecific restriction of LigDFG (Fig. 8B).21
the system (Fig. 7). In the initial phase of the photometrical With the transfer of hydrogen from NADH to GSSG the ether
enzyme assay (detection at 340 nm) LigDFG converts 1 under cleavage reaction can be regarded as having a net internal
consumption of NAD+ and glutathione to 5. With further con- hydrogen transfer from Cα to Cβ. A similar result can also be
sumption of 1, the reaction stopped with a significant increase obtained using ruthenium(II) complexes as homogenous cata-
of NADH concentration. In the second step of the experiment, lysts.26 However, much higher temperatures are required
addition of AVR decreased the previously formed NADH, for full conversion here, in comparison to the enzymatic
approach.
Application of LigDFG on real lignin substrates
While the effect of LigDFG cell extracts on a non-natural, per-
oxidase polymerized substrate 6 was described previously,27 we
have advanced this concept by applying characterized and pur-
ified LigDFG on complex lignin. Therefore we used the two
technical lignins, softwood/hardwood alkali-lignin and
bagasse organosolv-lignin as substrates, and measured the
effect of LigDFG on the monomer release and polymer distri-
bution compared to untreated control samples. GPC analysis
showed a minimal increase of the 120–200 MW area% distri-
bution in softwood and hardwood alkali-lignin after 7 days
compared to the controls by 1.8% and 2.1% respectively. A
slight increase of the 300 MW fractions by 1.5 area% could
also be detected in softwood alkali-lignin, whereas a decrease
could be observed in hardwood alkali-lignin (Fig. S4 to S5†).
GPC of bagasse organosolv showed no detectable differences
between the control and the sample, which might be due to
possible problems with the water solubility of the highly
Fig. 7 Reaction of the LigDFG enzyme system with AVR. All cofactors can be
recycled via reduction of GSSG with NADH.
Fig. 8 Activities of AVR by detection of NADH formation at 340 nm at pH 7.5. The first reaction converted the limiting concentration of 1 under consumption of
NAD+ and formation of GSSG with LigDFG. Addition of AVR after 2 min shows re-conversion of NADH to NAD+ (n = 3). The back reaction is diluted due to AVR
addition (A). Recycling via HPLC analysis. Transformation of 1 (30 mM) was detected over seven days with (S) and without (C) AVR addition. At the beginning of the
reaction 3 mM of NAD+ were added to both probes (n = 3) (B).
1378 | Green Chem., 2013, 15, 1373–1381
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
dispersed and partly insoluble substrate. It is noteworthy that alkali-lignin only showed a very small release of vanillin 12
the high molecular weight fraction of all three lignin samples from 0.47 mg L−1 after 2 h. Release of guaiacol 3 was even
increased, instead of being degraded or shifted in its mass dis- smaller (0.025 mg L−1 after 2 h) decreasing again to 0.0 mg L−1
tribution. This may be addressed to on-going radical coupling within 50 h (Table 1). Based on the structures guaiacol 3 and
under the aerobic assay conditions.
ferulic acid 9 are probably originating from the aryl part of the
A distinct monomer releases from the complex lignin mole- ether. Vanillin 12 and acetovanillone 11 in contrary could be
cules after addition of LigDFG and 10 mM NAD+ and GSH originating from either side of the ether bonded molecule.
were observed in softwood- and hardwood alkali-lignin, but Bagasse organosolv-lignin showed no detectable enzyme
not in bagasse-organosolv lignin (Table 1; Fig. 9). The area induced release of phenolic monomers.
ratios of the lignin monomers were calculated in mg L−1 by
LigDFG had no major detectable effect on the polymeric
standardization with commercial references following every fractions of all tested lignin samples. The reason for this is
GC-MS SPME run according to Kolb et al.28 The reaction most likely due to the low β-aryl ether content in kraft- and
batches contained definite amounts of free phenolic mono- organosolv lignin on the one hand and a possible inhibition
mers including our target monomers before starting the reac- of the enzymes by sulfides and/or solvent residues from the
tion. These monomers derive from the lignin pre-treatment pretreatment procedures on the other.29 In the alkali and
processes which typically consist of thermal, chemical or phys- organosolv process the β-aryl ethers are among the first
ical depolymerisation methods. Several additional peaks were bonding types being cleaved to more than 90%, while partial
detected but have not yet been analysed due to the lack of repolymerization occurs via C–C bond formation. Accordingly
specific reference substances for calibration. Softwood alkali- a complete depolymerisation of the technical substrates using
lignin showed a significant increase of monomer release after LigDFG was not to be expected. It is unclear whether LigDFG
enzyme treatment. For instance guaiacol 3 content increased serves as a lignin dimer/monomer processing tool in bacteria,
by 2.14 mg L−1 after 50 h. Similar effects could be observed for or enables the bacteria together with their cocktail of phenyl-
ferulic acid 9, being increased by 17.7 mg L−1 after 2 h and coumarane-, diarylpropane biphenyl-, pinoresinol- and aryl-
staying constant at that level for 50 h. Vanillin 12 increased by ether cleaving enzymes to degrade the complex lignin
58.2 mg L−1 within 2 h and by 89.8 mg L−1 after 50 h, aceto- polymer.30 Applying these enzymes in addition to LigDFG
vanillone 11 from 0.9 mg L−1 to 6.9 mg L−1 after 50 h. Hardwood could enhance polymer degradation and give rise to higher
monomer yields. It was thus shown that enzymatic cleavage of
ether linkages in technically pretreated lignin is only feasible
Table 1 Increase of lignin monomers in aqueous complex kraft-lignin samples. with a very small monomer release by the LigDFG enzyme
The values for released concentration of the probes were first subtracted from
system. Despite the problems of enantioselectivity, substrate
availability and radical coupling in the assay, these first experi-
ments showed promising activity and can be seen as an overall
proof of principle for the feasibility of the LigDFG etherase
system.
the control probes without enzymes, and then values at 0 h were subtracted
from the 2 h and 50 h values to eliminate effects of unspecific binding of
released substances to the proteins. Measurements were done in triplicate with
standard deviations up to 24%. All values shown in mg mL−1
2 h
50 h
Softwood
Guaiacol 3
Ferulic acid 9
Eugenol 10
0.0
17.7
0.002
2.1
17.1
0.1
Conclusion
Acetovanillone 11
Vanillin 12
0.9
58.2
6.9
89.8
The enzymes LigD, LigF and LigG were successfully expressed
and purified in E. coli and their proper functioning according
to Sato et al.20 could be confirmed. This enzyme system was
able to specifically cleave the β-aryl ethers of model lignin sub-
strates. The specific activities of every enzyme were deter-
mined. We successfully established activity assays for all three
Lig enzymes. While LigD and LigF could be analysed colouri-
metrically (LigD spectrophotometrically via NADH-formation
and LigF fluorometrically by cleavage of a model substance),
LigG activity had to be determined by HPLC analysis. Further-
more, we were able to show that an efficient enzymatic recy-
cling of all involved co-substrates is possible. Following the
green chemistry concept of hydrogen borrowing and being
able to be applied at ambient temperatures the combined
(LigDFG and AVR) system could become a useful tool in lignin
degradation for example for the synthesis of natural aroma
compounds. Certainly the reaction cascade has to be highly
improved for the conversion of complex lignin substrates to be
Hardwood
Guaiacol 3
Vanillin 12
0.03
0.5
0.0
0.7
Fig. 9 Chemical structure of lignin monomers, released after LigDFG incu-
bation; guaiacol 3, ferulic acid 9, eugenol 10, acetovanillone 11, vanillin 12.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 1373–1381 | 1379

View Article Online
Paper
Green Chemistry
technically feasible. Methods of enzyme engineering should
be most helpful here and will be applied. We have started to
successively improve the reaction conditions, e.g. temperature,
shaking and pH, for complex lignin enzyme assays to test the
upcoming mutant enzyme libraries with their stereospecific or
general process related capabilities. The crystallization of
LigD, LigF and LigG as basis for further optimizations is on-
going and will most likely give new insights to improve the
lignin degradation process with respect to its process related
properties (overall activity, temperature stability, enantioselec-
tivity and efficiency of recycling their co-substrates NAD+ and
glutathione). In addition the price for the co-substrates,
especially for NAD+, has to be considered. NAD+ is very expens-
ive and the cost should be prohibitive even though NAD+ is
recycled in the reactions. The utilization of alternative small
molecules as electron carriers such as N-benzyl-nicotinamide
4 A. P. Virk, P. Sharma and N. Capalash, Biotechnol. Prog.,
2012, 28(1), 21.
5 R. T. Mendonça, J. F. Jara, V. González, J. P. Elissetche and
J. Freer, J. Ind. Microbiol. Biotechnol., 2008, 35(11), 1323.
6 C. Eggert, U. Temp and K. E. Eriksson, Appl. Environ. Micro-
biol., 1996, 62(4), 1151.
7 P. Ollikka, K. Alhonmaki, V. M. Leppanen, T. Glumoff,
T. Raijola and I. Suominen, Appl. Environ. Microbiol., 1993,
59(12), 4010; K. L. Min, Y. H. Kim, Y. W. Kim, H. S. Jung
and Y. C. Hah, Arch. Biochem. Biophys., 2001, 392(2), 279;
J. A. Libra, M. Borchert and S. Banit, Biotechnol. Bioeng.,
2003, 82(6), 736.
8 R. ten Have and P. J. Teunissen, Chem. Rev., 2001, 101(11),
3397.
9 C. Raghukumar, D. D’Souza-Ticlo and A. Verma, Crit. Rev.
Microbiol., 2008, 34(3–4), 189.
derivatives can be of advantage here. These molecules are 10 G. R. Tortella, M. C. Diez and N. Duran, Crit. Rev. Micro-
more stable than the natural cofactor, much cheaper to biol., 2005, 31(4), 197.
produce and have been shown to be accepted as alternative 11 B. Dehorter and R. Blondeau, FEMS Microbiol. Lett., 1993,
cofactors by enzymes from different classes.31 In addition it
109(1), 117.
has been repeatedly shown that this acceptance can be highly 12 G. A. Martinez, P. M. Civello, A. R. Chaves and M. C. Anon,
improved by methods of enzyme engineering.32 That small
artificial molecules can work as unnatural cofactors in lignin
Phytochemistry, 2001, 58(3), 379; P. Du, J. R. Collins and
G. H. Loew, Protein Eng., 1992, 5(7), 679.
processing even today is demonstrated by the laccase mediator 13 A. Heinfling, M. J. Martinez, A. T. Martinez, M. Bergbauer
system in the pulp and paper industry.33
and U. Szewzyk, FEMS Microbiol. Lett., 1998, 165(1), 43;
A. Heinfling, M. J. Martinez, A. T. Martinez, M. Bergbauer
and U. Szewzyk, Appl. Environ. Microbiol., 1998, 64(8), 2788;
A. Heinfling, F. J. Ruiz-Duenas, M. J. Martinez,
M. Bergbauer, U. Szewzyk and A. T. Martinez, FEBS Lett.,
1998, 428(3), 141.
Another question is whether the application of four
different enzymes could be too expensive. The production cost
for enzymes is steadily decreasing34 and such application of
enzyme cascades using four and more enzymes in in vitro
systems, free of cells and “designer bugs” is finding more and
more acceptance even for the production of low value chemi- 14 F. Björkling, S. E. Godtfredsen and O. Kirk, J. Chem. Soc.,
cals and of energy from biomass.35 So even though it will
require much more work and optimization, the lignin degrad-
ing enzyme cascade shown here might be of technical rel-
evance sometime in the future.
Chem. Commun., 1990, 1301; F. Björkling, H. Frykman,
S. E. Godtfredsen and O. Kirk, Tetrahedron, 1992, 48, 4587;
S. Duncan, Q. Jing, A. Katona, R. J. Kazlauskas, J. Schilling,
U. Tschirner and W. Wafa AlDajani, Appl. Biochem. Biotech-
nol., 2010, 160(6), 1637; D. Yin, Q. Jing, W. W. AlDajani,
S. Duncan, U. Tschirner, J. Schilling and R. J. Kazlauskas,
Bioresour. Technol., 2011, 102(8), 5183.
Acknowledgements
15 L. Wiermans, M. Pérez-Sánchez and P. Domínguez de
María, ChemSusChem, 2013, 6(2), 251.
Special thanks to Manuela Kaiser, Fraunhofer IGB BioCat, for
synthesis of substrate 6, to Petra Lommes, for assistance and
help at the GC-MS analytics, Anja Philip for her analytical
advices and Benjamin Kick, all TUM Chair of Chemistry of bio-
genic Resources. We also thank the Bavarian government for
funding this project within the BayernFit program and both
the Chemische Werke Zell-Wildshausen GmbH (Germany) and
the Fraunhofer ICT (Germany), for providing lignin samples.
16 Y. Chen, L. Chai, C. Tang, Z. Yang, Y. Zheng, Y. Shi and
H. Zhang, Bioresour. Technol., 2012, 123, 682.
17 S. D. Copley, J. Rokicki, P. Turner, H. Daligault, M. Nolan
and M. Land, Genome Biol. Evol., 2012, 4(2), 184.
18 E. Masai, A. Ichimura, Y. Sato, K. Miyauchi, Y. Katayama
and M. Fukuda, J. Bacteriol., 2003, 185(6), 1768.
19 E. Meux, P. Prosper, E. Masai, G. Mulliert, S. Dumarçay,
M. Morel, C. Didierjean, E. Gelhaye and F. Favier, FEBS
Lett., 2012, 586(22), 3944.
20 Y. Sato, H. Moriuchi, S. Hishiyama, Y. Otsuka, K. Oshima,
D. Kasai, M. Nakamura, S. Ohara, Y. Katayama, M. Fukuda
and E. Masai, Appl. Environ. Microbiol., 2009, 75(16), 5195.
Notes and references
1 W. Boerjan, J. Ralph and M. Baucher, Annu. Rev. Plant Biol., 21 K. Tanamura, T. Abe, N. Kamimura, D. Kasai,
2003, 54, 519.
2 E. Adler, Ind. Eng. Chem., 1957, 49(9), 1377.
3 G. Palmieri, J. Biol. Chem., 1997, 272(50), 31301.
S. Hishiyama, Y. Otsuka, M. Nakamura, S. Kajita,
Y. Katayama, M. Fukuda and E. Masai, Biosci., Biotechnol.,
Biochem., 2011, 75(12), 2404.
1380 | Green Chem., 2013, 15, 1373–1381
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
22 Y. Otsuka, T. Sonoki, S. Ikeda, S. Kajita, M. Nakamura and 30 E. Masai, Y. Katayama and M. Fukuda, Biosci., Biotechnol.,
Y. Katayama, Eur. J. Biochem., 2003, 270(11), 2353.
23 M. M. Bradford, Anal. Biochem., 1976, 72, 248.
24 G. Weinstein, Holzforschung, 1979, 33(4), 134.
25 Y. C. Chung and R. E. Hurlbert, J. Bacteriol., 1975, 123(1),
203.
Biochem., 2007, 71(1), 1.
31 C. E. Paul, S. Gargiulo, D. J. Opperman, I. Lavandera,
V. Gotor-Fernández, V. Gotor, A. Taglieber, I. W. C.
E. Arends and F. Hollmann, Org. Lett., 2013, 15(1), 180;
H. C. Lo and R. H. Fish, Angew. Chem., Int. Ed., 2002, 41(3),
478.
26 J. M. Nichols, L. M. Bishop, R. G. Bergman and
J. A. Ellman, J. Am. Chem. Soc., 2010, 132(36), 12554; T. vom 32 E. Campbell, M. Meredith, S. D. Minteer and S. Banta,
Stein, T. Weigand, C. Merkens, J. Klankermayer and
W. Leitner, ChemCatChem, 2013, 5(2), 619.
Chem. Commun., 2012, 48(13), 1898; J. D. Ryan, R. H. Fish
and D. S. Clark, ChemBioChem, 2008, 9(16), 2579.
27 T. Sonoki, Y. Iimura, E. Masai, S. Kajita and Y. Katayama, 33 H. P. Call and I. Mücke, J. Biotechnol., 1997, 53(2–3), 163.
J. Wood Sci., 2002, 48(5), 429.
34 D. Klein-Marcuschamer, P. Oleskowicz-Popiel, B. A. Simmons
and H. W. Blanch, Biotechnol. Bioeng., 2012, 109(4), 1083.
35 Y.-H. P. Zhang, ACS Catal., 2011, 1(9), 998; J.-K. Guterl,
D. Garbe, J. Carsten, F. Steffler and B. Sommer, et al.,
ChemSusChem, 2012, 5(11), 2165; J.-K. Guterl and V. Sieber,
Eng. Life Sci., 2013, 13(1), 4.
28 M. Kolb, V. Sieber, M. Amann, M. Faulstich and
D. Schieder, Bioresour. Technol., 2012, 104, 298.
29 Taschenbuch der Papiertechnik: Mit 85 Tabellen, ed.
J. Blechschmidt, Fachbuchverl. Leipzig im Carl-Hanser-
Verl., München, 2010.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 1373–1381 | 1381