Ionic liquid tolerant hyperthermophilic cellulases for biomass pretreatment and hydrolysis

Article abstract of DOI:10.1039/b916564a

One of the main barriers to the enzymatic hydrolysis of cellulose results from its highly crystalline structure. Pretreating biomass with ionic liquids (IL) increases enzyme accessibility and cellulose recovery through precipitation with an anti-solvent. For an industrially feasible pretreatment and hydrolysis process, it is necessary to develop cellulases that are stable and active in the presence of small amounts of ILs co-precipitated with recovered cellulose. However, a significant decrease in cellulase activity in the presence of trace amounts of ILs has been reported in the literature, necessitating extensive processing to remove residual ILs from the regenerated cellulose. Towards that end, we have investigated the stability of hyperthermophilic enzymes in the presence of the IL 1-ethyl-3-methylimidazolium acetate ([C2mim][OAc]) and compared it to the industrial benchmark Trichoderma viride (T. viride) cellulase. The endoglucanase from a hyperthermophilic bacterium, Thermatoga maritima, and a hyperthermophilic archaeon, Pyrococcus horikoshii, were over expressed in E. coli and purified to homogeneity. Under their optimum conditions, both hyperthermophilic enzymes showed significantly higher [C2mim][OAc] tolerance than T. viride cellulase. Using differential scanning calorimetry we determined the effect of [C2mim][OAc] on protein stability and our data indicates that higher concentrations of IL correlated with lowered protein stability. Both hyperthermophilic enzymes were active on [C2mim][OAc] pretreated Avicel and corn stover. Furthermore, these enzymes can be recovered with little loss in activity after exposure to 15% [C2mim][OAc] for 15 h. These results demonstrate the potential of using IL-tolerant extremophilic cellulases for hydrolysis of IL-pretreated lignocellulosic biomass, for biofuel production.

Full text of DOI:10.1039/b916564a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
Ionic liquid tolerant hyperthermophilic cellulases for biomass pretreatment
and hydrolysis†
Supratim Datta, Bradley Holmes,a,b Joshua I. Park,a,b Zhiwei Chen, Dean C. Dibble,
a,b
a,b
a,b
a,b
a,c
a,b
a,b
Masood Hadi, Harvey W. Blanch, Blake A. Simmons and Rajat Sapra*
Received 11th August 2009, Accepted 11th November 2009
First published as an Advance Article on the web 21st January 2010
DOI: 10.1039/b916564a
One of the main barriers to the enzymatic hydrolysis of cellulose results from its highly crystalline
structure. Pretreating biomass with ionic liquids (IL) increases enzyme accessibility and cellulose
recovery through precipitation with an anti-solvent. For an industrially feasible pretreatment and
hydrolysis process, it is necessary to develop cellulases that are stable and active in the presence of
small amounts of ILs co-precipitated with recovered cellulose. However, a significant decrease in
cellulase activity in the presence of trace amounts of ILs has been reported in the literature,
necessitating extensive processing to remove residual ILs from the regenerated cellulose. Towards
that end, we have investigated the stability of hyperthermophilic enzymes in the presence of the IL
1
-ethyl-3-methylimidazolium acetate ([C2mim][OAc]) and compared it to the industrial
benchmark Trichoderma viride (T. viride) cellulase. The endoglucanase from a hyperthermophilic
bacterium, Thermatoga maritima, and a hyperthermophilic archaeon, Pyrococcus horikoshii, were
over expressed in E. coli and purified to homogeneity. Under their optimum conditions, both
hyperthermophilic enzymes showed significantly higher [C2mim][OAc] tolerance than T. viride
cellulase. Using differential scanning calorimetry we determined the effect of [C2mim][OAc] on
protein stability and our data indicates that higher concentrations of IL correlated with lowered
protein stability. Both hyperthermophilic enzymes were active on [C2mim][OAc] pretreated Avicel
and corn stover. Furthermore, these enzymes can be recovered with little loss in activity after
exposure to 15% [C2mim][OAc] for 15 h. These results demonstrate the potential of using
IL-tolerant extremophilic cellulases for hydrolysis of IL-pretreated lignocellulosic biomass, for
biofuel production.
Introduction
markets given the role of corn as a food commodity. LC biomass,
on the other hand, is an abundant and potentially low cost
resource that does not compete with human needs. It includes
The use and depletion of fossil fuels are major environmental and
energy security issues for the twenty-first century. Renewable
biofuels produced from biomass are an important alternative to
petroleum use for transportation. Lignocellulosic (LC) biomass
is an abundant and, potentially, carbon-neutral energy resource
3
1
agricultural residues (corn stover, wheat straw and rice straw),
deciduous and coniferous woods, agricultural processing by-
products (corn fiber, rice hulls and sugar cane bagasse) and
4
energy crops like switch grass and miscanthus.
2
for the production of biofuels and chemicals. However, there are
Lignocellulosic biomass is made of three major components—
cellulose (35–50%), hemicellulose (20–35%) and lignin (10–
scientific and technological challenges that need to be overcome
before LC biomass-derived biofuels can be a significant and
economically competitive alternative to petroleum.
The production of ethanol from corn grain represents the
most convenient and technically advanced option for biofuels
in the United States today. However, the increased displacement
of available corn supplies for fuel has adverse affects on food
2
5%)—the relative concentrations of which vary according
5
to the source. For efficient utilization, the biomass has to
undergo pretreatment before it can be enzymatically hydrolyzed
to glucose and other constituent sugars. Cellulosic component of
LC biomass is highly crystalline with polymeric cellulose chains
6
held together by hydrogen bonds and van der Waals forces.
The polymeric chains are made of anhydroglucose units joined
together by glycosidic bonds with degree of polymerization
a
Deconstruction Division, Joint BioEnergy Institute, 5885 Hollis St, 4th
7
(DP) ranging from 1000–20 000 units. Pretreatment is essential
Floor, Emeryville, CA, USA. E-mail: rsapra@lbl.gov;
Fax: +1 510-486-4252; Tel: +1 510-508-9956
to reduce the DP of the cellulose and make it amorphous
to facilitate access to the substrate for enzymatic cellulose
b
Biomass Science and Conversion Technology Department, Sandia
National Laboratory, Livermore, CA, USA
8
hydrolysis. Current strategies in pretreatment and recovery of
c
Department of Chemical Engineering, University of California,
cellulose from biomass have been primarily derived from pulp
and paper industry, where the downstream process of converting
cellulose to glucose is not a consideration and are incompatible
Berkeley, CA, USA
†
Electronic supplementary information (ESI) available: DSC, halotol-
erance and compositional analysis. See DOI: 10.1039/b916564a
3
38 | Green Chem., 2010, 12, 338–345
This journal is © The Royal Society of Chemistry 2010

View Article Online
with enzymatic hydrolysis; thereby requiring extensive cellulose
processing for enzymatic hydrolysis and resulting in an increase
in time and cost. It is therefore desirable and important to
develop a biomass pretreatment process that is compatible
with not only the enzymatic hydrolysis but also subsequent
fermentation of glucose to biofuels.
genome encode enzymes that are involved in the metabolism
of monosaccharides and polysaccharides. Tma Cel5A is a free-
acting endoglucanase containing only a catalytic domain.
25
Pyrococcus horikoshii (Pho), a hyperthermophilic archaeon
found in deep-sea hydrothermal vents, is an anaerobic het-
erotroph, which utilizes peptides as its main carbon source
◦
Current pretreatment methods include mechanical reduction
in biomass particulate size followed by dilute acid, flow-through
ammonia fiber expansion, ammonia recycle percolation, lime,
at temperatures approaching 100 C but has been shown to
contain an endoglucanase (EG) that is functional at an optimum
◦
temperature of 97 C.
9
steam explosion, or organosolv (OS) pretreatment. The draw-
Since cellulases from Trichoderma constitute an established
cellulase hydrolysis process for benchmarking, here we present
data comparing the efficiency of hydrolysis of Tma Cel5A and
Pho EG with the Trichoderma viride cellulase. Tma Cel5A and
Pho EG were purified to homogeneity and enzymatic hydrolysis
activity was measured in the presence of varying concentrations
of [C2mim][OAc]. Both enzymes showed significant specific
activity at 15% (v/v) [C2mim][OAc] with specific activities
decreasing at higher concentrations. Decreased specific activities
at higher concentrations showed a direct correlation to enzyme
stability as indicated by decreases in unfolding temperatures
by DSC measurements. Both enzymes were very stable at high
temperatures for extended periods of time and were tolerant to
2 M NaCl and KCl. These enzymes are also active on pretreated
biomass. Long half-lives and a near 100% recovery of activity
from ILs make them promising targets for further investigation
for use in large-scale saccharification reactions employing this
pretreatment technology.
backs of these methods include severe reaction conditions (high
temperature and/or high pressure and extremes of pH), and
high processing costs. A new pretreatment approach using ionic
liquids (ILs) has been shown to be effective in solubilizing
10
crystalline cellulose.
Room temperatures ILs are pure salts with melting points
◦
10,11
typically below 100 C.
While the properties of ILs are very
diverse, high thermal stability and non-volatility make this class
of compounds an attractive alternative to organic solvents in
12
several existing industrial processes. ILs depending on their
chemistry, can dissolve a wide range of polar and non-polar
organic and polymeric compounds, including cellulose, and
make the polysaccharide chains more accessible to enzymatic
13
hydrolysis. Of particular interest to the biofuels industry is
the ability of several imidazolium based ionic liquids with weak
conjugated-base anions to dissolve crystalline cellulose at mod-
13,14
erate temperatures.
The cellulose is subsequently recovered in
the amorphous form upon the addition of anti-precipitants like
1
5–17
water
rendering it more accessible to enzymatic hydrolysis
Results
18
and subsequent rapid hydrolysis to glucose using cellulases.
The dissolution of lignocellulosic materials such as corn stalks,
rice straw, bagasse, pine wood, and spruce wood in ILs followed
by cellulose hydrolysis with acid or enzymes has also been
Hyperthermophilic enzymes and specific activity assays
26
Previous studies by Chhabra et al. showed that the Tma Cel5A
from Tma is intracellularly expressed in T. maritima and has
18–20
recently reported in literature.
◦
an optimum temperature (Topt) of 80 C. The EG from Pho
Cost-effective enzymatic hydrolysis of cellulose to sugars
would benefit from the development of enzymes that are effective
under pretreatment conditions (which may be extreme pH or
temperature). However, significant decreases in cellulase activity
in the presence of trace amounts of ILs have been reported in
◦
has a even higher Topt of 97 C and specific activity of 8.5 U
-
1
-1
mg (1 U = 1 mmole of reducing sugar produced min using
◦
27
CMC as substrate) at pH 5.6 and 85 C. Thus, both of these
enzymes were attractive targets for our work. The enzymes were
expressed in E. coli in autoinduction media and purified by nickel
affinity chromatography to homogeneity as detected by SDS-
PAGE. Maximal amounts of soluble Tma Cel5A were obtained
21,22
literature.
In order to avoid extensive processing and clean
up of the regenerated cellulose, the cellulases that are used
must be able to perform optimally in up to 15% concentrations
of IL that is co-precipitated with cellulose. The amount of
leftover IL would however depend on the cost of recovery
of ILs. Towards the goal of discovering cellulases that are
stable and active in the presence of trace amounts of ILs, we
investigated the stability of hyperthermophilic enzymes, broadly
categorized with extremophilic enzymes, as a function of IL
concentration. Since extremophilic organisms grow in extremes
of pH, temperature and/or salinity etc., these are a promising
source of enzymes for industrial processes. The enzymes derived
from hyperthermophilic organisms may be capable of tolerating
ionic liquids since the native growth condition of these organisms
◦
-1
following induction at 37 C, yielding 30 mg L . The yield of
-
1
Pho EG was around 2 mg L , under similar conditions except
that an additional anion exchange chromatography step was
added to obtain purer protein. Both cellulases were active on
carboxymethyl cellulose (CMC). In our experiments, the specific
-
1
activities of Tma Cel5A and Pho EG were 30 U mg and
-
1
1
.9 U mg , respectively (Table 1). The commercially available
-
1
◦
T. viride cellulase, has a specific activity of 10–12 U mg at 37 C
with CMC as the substrate, depending on the batch of enzyme
used.
Specific activity in the presence of [C2mim][OAc]
◦
23
is greater than 80 C in a strongly reducing environment.
A model extremophilic organism is the hyperthermophilic
bacterium Thermatoga maritima (Tma), a member of the order
thermotogales, which is an anaerobic heterotrophic hyper-
thermophile capable of fermenting both simple and complex
The purified recombinant enzymes were assayed in the presence
of [C2mim][OAc] and the specific activity results were compared
with the cellulase from T. viride. [C2mim][OAc] was selected as
the IL of choice. The assays were preformed under optimum
pH and temperatures for T. maritima and T. viride enzymes, as
24
sugars. Roughly 7% of the predicted coding sequences in its
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 338–345 | 339

View Article Online
Table 1 Biochemical properties of hyperthermophilic enzymes: optimum temperature and pH, specific activity, and stability in presence of
[
C2mim][OAc]. The % specific activities (last column) are reported as percentages of residual specific activity, taking the specific activity of IL
free enzyme as 100% activity
Properties
%
Activity recovered after
◦
a
Enzymes
T
opt/ C
pHopt
T
1/2/h in 0% IL
T
1/2/h in 15% IL
Specific activity
incubation in 15% IL for 15 h
b
T. viride cellulase
37
80
>95
4.5
4.8
6.4
9
0
11 ± 2
30 ± 2
1.9 ± 0.3
0
44
79
b
Tma Cel5A
20 ± 2
24 ± 4
19 ± 3
23 ± 4
c
Pho EG
a
Units is mmoles of reducing sugars min mg-1 with carboxymethylcellulose as substrate. Assays were done at Topt
-1
b
c
◦
.
Assays were done at 80 C.
◦
shown in Table 1. The Pho EG was assayed at 80 C where
of its activity while it was undetectable in the presence of 10%
(v/v) [C2mim][OAc]. Both the recombinant hyperthermophilic
enzymes behaved very differently. Tma Cel5A in the presence of
5% [C2mim][OAc] loses 24% of its specific activity. At 15% IL
(v/v), Tma Cel5A retains 50% of its specific activity. The highest
residual activities were observed with Pho EG which retains
◦
its specific activity is around 10% less than at 95 C (data not
shown). As seen in Fig. 1, the cellulase from T. viride rapidly
lost activity with increasing concentration of [C2mim][OAc]. In
the presence of 5% (v/v) [C2mim][OAc] the enzyme lost 60%
1
00% activity in 20% IL (v/v) (Fig. 1). Pre-incubation overnight
(15 h) in 15% [C2mim][OAc] resulted in an almost complete loss
of activity for the T. viride cellulases, as expected (Table 1) while
the Tma Cel5A and Pho EG retained 44% and 70% respectively
of their activity. The hyperthermophilic enzymes also exhibited
longer half-lives, around 20–24 h, compared to the 9 h for the
fungal T. viride cellulase, at their optimal pH and temperatures
(
Table 1).
Effect on enzyme stability
Fig. 1 Enzymatic hydrolysis of CMC by cellulase from T. viride (cross
line), Tma Cel5A (solid) and Pho EG (horizontal line), in the presence of
various [C2mim][OAc]: 2% CMC in and 0–50% of [C2mim][OAc] (v/v)
were incubated at 37 C (T.viride), 80 C (Tma EG, Pho EG). The specific
activities were calculated as mmoles of reducing sugars formed per min
per mg of enzyme. The specific activities are reported as percentages of
residual specific activity, taking the specific activity of IL free enzyme as
To understand the effect of ILs on protein stability, differential
scanning calorimetry studies were done on the T. viride cellulase
and compared to Tma Cel5A and Pho EG. The unfolding
temperatures of the enzymes are listed in Table 2 and the
thermograms for Tma Cel5A and Pho EG are shown in the ESI,
Figure 1A and 1B.† The decrease in unfolding temperatures for
◦
◦
◦
1
00% activity. Error bars indicate the standard deviation.
T. viride cellulase between 0 and 5% [C2mim][OAc] is 7.2 C
-
1
-1
Table 2 Unfolding temperatures from differential scanning calorimetry thermograms of Tma Cel5A (5 mg mL ), T. viride cellulase (2.2 mg mL )
-1
and Pho EG (1.8 mg mL ). The decrease in unfolding temperature of Tma Cel5A, T. viride cellulase and Pho EG is compared to the decrease in
percent specific activity
◦
◦
Enzyme
% [C2mim][OAc] (v/v) Unfolding T/ C Decrease in % specific activity Decrease in unfolding T/ C
T. viride cellulase
0
5
64.2
57.0
52.4
49.4
47.5
—
60
100
100
100
—
7.2
11.8
14.8
16.7
1
1
2
0
5
0
T. maritima endoglucanase (Tma Cel5A)
P. horikoshii Endoglucanase (Pho EG)
0
5
92.0
89.8
89.3
88.5
87.2
67.0
—
23
35
52
58
98
—
2.2
2.7
3.5
4.8
25.0
1
1
2
5
0
5
0
0
0
5
102.3
99.3
97.5
94.9
91.8
66.1
—
0
0
5
10
100
—
3
4.8
7.4
10.5
36.2
1
1
2
5
0
5
0
0
3
40 | Green Chem., 2010, 12, 338–345
This journal is © The Royal Society of Chemistry 2010

View Article Online
◦
◦
compared to a 2.2 C drop in Tma Cel5A and a 3 C drop in
Pho EG (Table 2). This difference gets more pronounced at 10%
Discussion
◦
◦
[
C2mim][OAc] (11.8 C vs. 2.7 C) when the T. viride cellulase is
Hyperthermophilic proteins exhibit a remarkable balance be-
tween heat stability and functionality. This makes such proteins
suitable for industrial applications like enzymatic hydrolysis
of biomass for biofuels production. These enzymes generally
owe their stability to a higher number of ion pairs and the
◦
completely inactive. In the case of Tma Cel5A, there is a 25 C
drop in unfolding temperature at 50% [C2mim][OAc] (v/v) when
the specific activity of Tma Cel5A is undetectable using the DNS
◦
assay. Similarly there is a 36.2 C drop in unfolding temperature
30
of Pho EG when the specific activity becomes undetectable. To
understand the basis of tolerance to ILs we also looked at the
effect of high concentrations of salt on the specific activity of
the hyperthermophilic enzymes. To investigate whether the IL
tolerance of the Tma Cel5A can be correlated with ionic strength
of the buffer, Tma Cel5A was assayed in the presence of high
salt concentration. As seen in the ESI, Figure 2A and 2B,†
Tma Cel5A and Pho EG retained 65% and 85% of its activity
respectively, in the presence of 2 M NaCl and 2 M KCl.
polarity of exposed surfaces. We started our present studies
with two endoglucanase from hyperthermophilic organisms (one
characterized representative each from bacterial and archaeal),
to look into the effect of ILs with an eye on developing a large
scale bioprocess.
Current processing conditions do not allow an in situ pretreat-
ment and saccharification using an ionic liquid solvent, since ILs
31
has been shown to affect the activity of cellulases. Turner et al.
reported the inhibition of T. reesei cellulases by the IL 1-butyl-
-methylimidazolium chloride. A recent report by Pottk a¨ mper
3
32
et al. reported the IL tolerance of a salt-tolerant cellulase
Effect on pretreated biomass
33
from an uncultured microorganism and discovered that a GH5
family cellulase that was active at IL concentrations of up to 30%
(v/v). However the specific activity of these cellulases was very
Corn stover refers to stalks, leaves and cobs that remain in the
fields after the corn kernel harvest and is the largest quantity of
28
-1
-1
biomass residue in the United States and as such is also a LC-
biomass source for producing cellulosic ethanol in the United
low, in the range of 0.014 mmoles min mg with the assay
32
performed with cell extracts.
29
States. [C2mim][OAc] pretreated corn stover was used as a
substrate to verify enzymatic efficiency on IL-pretreated biomass
and a comparison made with [C2mim][OAc] pretreated Avicel.
Avicel is a commercially available microcrystalline cellulose
and a good model substrate for analysis. Similar assays were
done in the presence of [C2mim][OAc] to simulate real biomass
processing and saccharification scenarios. The assay products
for the reactions in the presence of IL were detected by high
performance anion exchange chromatography (HPAEC). The
enzymatic hydrolysis reaction was carried out for 8 h for the
T. viride cellulase and 15 h for the hyperthermophilic enzymes,
taking into account the half-lives of these enzymes. To compare
the amounts of total sugars formed between different reactions
we chose to calculate the sum of glucose, xylose and cellobiose
and report as mmoles of sugars formed per min per mg of enzyme
used (Table 3). A comparison of assay results between pretreated
and untreated substrate indicates a 2–6 fold increase in hydrolysis
products after pretreatment irrespective of the enzyme used,
thereby confirming that IL pretreatment can increase the
efficiency of hydrolysis of cellulose recovered from biomass.
The fungal enzyme was more active on untreated substrates
compared to the hyperthermophilic enzymes probably due to the
presence of cellobiohydrolases and xylanases in the enzyme mix
We wanted to conduct an extensive study to look into the
effect of ILs on hyperthermophilic cellulases with an eye on
developing a large scale bioprocess with the hypothesis that
hyperthermophilic enzymes would better tolerate ILs. Both
Tma Cel5A and Pho EG were found to retain between 44%
and 79% respectively of their activity in the presence of 15%
[C2mim][OAc] unlike the T. viride cellulase with greater than
85% recovery of activity after exposure to 15% [C2mim][OAc] for
15 h. To verify enzymatic activity on biomass, we compared the
efficiency of the recombinant cellulases to those from T. viride
cellulase using both corn stover and Avicel as substrates and
compared the results with assays on IL pretreated Avicel and
corn stover. The pretreated substrates were more efficiently
hydrolyzed compared to the untreated substrates. None of our
recombinant proteins contain carbohydrate-binding domains,
yet we detect the formation of sugars in the presence of untreated
Avicel or corn stover. Further studies are needed to differentiate
between the possibility of hydrolysis of amorphous cellulose
in these samples or the possible effect of hyperthermophilic
enzymes without carbohydrate-binding domains on insoluble
substrates. However, as would be expected from reactions with
insoluble substrates, enzymatic efficiency was less compared
to results obtained using CMC. These results also show the
importance of evaluating enzymatic efficiency with “real-world”
insoluble substrates as opposed to synthetic soluble substrates
that are easier to manipulate but cannot be used for translating
analysis to actual substrates. The hyperthermophilic enzymes
displayed higher activity compared to T. viride cellulase in the
presence of [C2mim][OAc] (Table 3) for IL-pretreated cellulose,
verifying our earlier results using a soluble substrate. The
CMCase activity results of the three cellulases are not indicative
of the activity on non-pretreated biomass where the T. viride
outperforms the extremophilic cellulases. However since IL
pretreatment increases the amount of sugars released, it would
thus be preferred to pretreat the biomass prior to enzymatic
hydrolysis, thus underscoring the need for IL-tolerant cellulase.
(
Table 3). The catalytic efficiency of the fungal enzyme however,
decreases with increasing IL concentration in the reaction mix.
After normalizing for the reaction time and enzyme amounts, the
amount of sugars formed is undetectable at 15% [C2mim][OAc].
The catalytic efficiencies of the hyperthermophilic enzymes with
insoluble substrates and in the presence of [C2mim][OAc] follow
a similar trend as seen with the soluble substrate CMC. The yield
of sugars from Tma Cel5A hydrolysis of pretreated Avicel in the
presence of 15% IL decreases by 43% compared to 0% IL, while
the decrease is only 10% in the case of Pho EG on the same
substrate under similar conditions (Table 3). Cellobiose was the
major product in the case of Pho EG, while a mixture of sugars
were obtained for Cel5A.
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 338–345 | 341

View Article Online
Table 3 Activity assay on insoluble substrates in the absence of [C2mim][OAc] and presence of [C2mim][OAc]. The hydrolysis reactions were
performed under pH and temperatures for each enzyme as listed in the Methods section. The reactions were run over a 14 h period with 5–10 mg of
enzyme with 6% (w/v) enzyme loading in optimum buffer for each enzyme and run in duplicate. The reaction products were measured on a HPLC
and HPAEC. The amounts of sugars detected are shown in mmoles and normalized with regards to amount of enzyme and time of reaction. Different
reactions under each enzyme were run under exactly similar conditions
a
-1
%
0
[C2mim][OAc]
Enzyme
Substrate
Sugars produced /mmoles min mg-1
T. viride cellulase
Avicel
pAvicel
Corn stover
pCorn stover
3 ± 0.5
5 ± 0.9
2.8 ± 0.8
6 ± 0.9
Tma Cel5A
Pho EG
Avicel
pAvicel
Corn stover
pCorn stover
1.1 ± 0.1
6.6 ± 0.8
3 ± 0.9
9 ± 1
Avicel
pAvicel
Corn stover
pCorn stover
0.54 ± 0.1
3.2 ± 0.2
0.58 ± 0.1
2.14 ± 0.1
5
T. viride cellulase
Tma Cel5A
Pho EG
Avicel
pAvicel
Corn stover
pCorn stover
2.1 ± 0.6
4 ± 0.9
2.3 ± 0.5
3.6 ± 0.6
Avicel
pAvicel
Corn stover
pCorn stover
1.0 ± 0.2
6 ± 0.5
2 ± 0.3
8 ± 0.9
Avicel
pAvicel
Corn stover
pCorn stover
0.5 ± 0.1
3.3 ± 0.2
0.59 ± 0.1
2.2 ± 0.14
1
0
T. viride cellulase
Tma Cel5A
Pho EG
Avicel
pAvicel
Corn stover
pCorn stover
0.2 ± 0.1
1 ± 0.1
0.6 ± 0.3
1.1 ± 0.3
Avicel
pAvicel
Corn stover
pCorn stover
0.9 ± 0.1
5.7 ± 1
2.4 ± 0.8
7 ± 1
Avicel
pAvicel
Corn stover
pCorn stover
0.52 ± 0.1
3.1 ± 0.2
0.57 ± 0.1
2.0 ± 0.2
1
5
T. viride celulase
Tma Cel5A
Pho EG
Avicel
pAvicel
Corn stover
pCorn stover
<0.01
<0.01
<0.01
<0.01
Avicel
pAvicel
Corn stover
pCorn stover
0.6 ± 0.1
4 ± 0.5
2.3 ± 0.5
5.9 ± 1
Avicel
pAvicel
Corn stover
pCorn stover
0.6 ± 0.08
2.9 ± 0.06
0.59 ± 0.1
1.9 ± 0.1
a
b
The sum of glucose, cellobiose and xylose obtained, normalized by enzyme amount and time. The reactions were spiked with 0 to 15% (v/v)
c
[
C2mim][OAc], everything else being the same for all substrates for an enzyme. The sum of glucose, cellobiose and xylose obtained, normalized by
enzyme amount and time.
3
42 | Green Chem., 2010, 12, 338–345
This journal is © The Royal Society of Chemistry 2010

View Article Online
Future studies would focus on using optimum sized particles,
solids loading during the saccharification reaction as well as
enzyme loadings.
PCR and cloning of endoglucanase from T. maritima and
P. horikoshii
The open reading frame of the endoglucanase from Tma
Cel5A (NP_229549), and Pho EG (UniProt ID: O58925) was
synthesized and codon optimized for protein expression in
E. coli (GenScript Corporation, Piscataway, NJ). Tma Cel5A
was cloned by ligation independent cloning method into the
expression vector, pCDF2 LIC/Ek (Novagen/EMD Chemicals,
Gibbstown, NJ), using the kit supplied by the manufacturer.
The Pho EG amplicon was gel-purified, and then cloned into
pDONR221 (Invitrogen, Carlsbad, CA) via BP recombination
reaction to create an entry vector. The Pho EG gene was then
inserted into pET DEST42 vector (Invitrogen, Carlsbad, CA)
as a fusion to a C-terminal V5 epitope and His(¥6) tags by LR
recombination reaction to create pET DEST42-PhoEG plasmid.
The DNA sequence of entry and expression vectors for Pho EG
and expression vector containing Tma Cel5A was confirmed by
DNA sequencing.
The DSC experiments indicate that at higher concentrations
of [C2mim][OAc] the enzymes unfold and are less stable. Thus,
the loss of activity could be accounted by the denaturation
of the enzyme or a disruption of the tertiary structure of the
protein. It has been speculated that halotolerant enzymes are
more stable in ILs in comparison to non-halotolerant enzymes,
because of the inherent ability of such organisms to live in
11
high salt concentration environments. However, we found that
the recombinant proteins Tma Cel5A and Pho EG from non-
halotolerant organisms maintained around 80% of their activity
in 1.5 M sodium or potassium chloride (ESI, Fig. 2).† This is
in contrast to general findings that non-halotolerant cellulases
21
show reduced activity at high salt concentrations. This implies
a different mechanism of denaturation depending on the identity
of the cation and anion of the IL. Further studies are needed
to better understand the differences in behaviour of Tma Cel5A
and Pho EG at higher concentrations of ionic liquids to get a
better understanding of the mechanisms of deactivation, and
to investigate if other IL’s used in biomass pretreatment would
better stabilize the cellulases.
Protein expression and purification
Escherichia coli BL21(DE3) star transformed with the plasmid
construct was grown in Luria–Bertani (LB) medium containing
The recombinant enzymes studied here are members of
the GH5 family, supporting the speculation that the higher
-1
◦
streptomycin (50 mg mL ) at 37 C, induced with 0.5 mM IPTG
(
◦
at A600 nm = 0.5–0.7), and then grown overnight at 30 C. Tma
32
resistance to ILs is due to a common structural motif. Based
Cel5A with a C-terminal His tag was purified from cells (10 g wet
weight) suspended in 100 mL of 10 mM potassium phosphate
32
on saturation mutagenesis studies, J. Pottk a¨ mper et al. have
suggested that motifs within the N-terminal part of the protein
and within carbohydrate-binding domain are responsible for IL
tolerance. We speculate that the hyperthermophilic properties
of these enzymes may lead to stability of the enzymes in ionic
liquids. Related studies with IL tolerance of other cellulases
from hyperthermophilic organisms support a similar trend
-1
buffer (pH 7.2) and containing 0.15 mg mL lysozyme and one
tablet of complete EDTA-free protease inhibitor (Roche). After
sonication, NaCl (150 mM) and imidazole (10 mM) were added
to achieve the specified final concentrations and loaded onto a
Ni-nitrilotriacetic acid column (GE Healthcare, Piscataway, NJ)
equilibrated with phosphate buffer containing 150 mM NaCl,
and 10 mM imidazole. The column was washed extensively, and
proteins were eluted with a 10–400 mM imidazole gradient. The
protein eluted between 100 and 350 mM imidazole and were
pooled and dialyzed against 100 mM Acetate buffer (pH 4.8).
The Pho EG was purified under similar conditions with an
additional purification step by Q Sepharose High Performance
column (GE Healthcare, Piscataway, NJ) using manufacturer
recommended protocol and dialyzed against 100 mM MES
buffer (pH 6.4). The protein concentrations were measured by
Bradford Assay for both recombinant and commercial enzymes
used in this study. The recombinant protein purity was visually
assessed using SDS-PAGE.
(
S. Datta, J. I. Park and R. Sapra—unpublished results).
Conclusions
While there are a number of reports in literature on using ILs
to deconstruct cellulosic biomass, the present work reports on
the effects of ILs on purified and active recombinant cellulases.
Our work suggests that hyperthermophilic cellulases can tolerate
around 15% (v/v) of ILs, which reflects the IL content that
may be present from a large-scale commercial IL pretreatment
process. Fungal enzymes studied did not possess this tolerance
to ILs. Further work is needed to discover, evolve and engineer
cellulases with longer half-lives that are more stable and active
in higher concentrations of IL, towards the ultimate goal of
cheaper and efficient saccharification of biomass.
Avicel and corn stover pretreatment by [C2mim][OAc]
[
C2mim][OAc] (1 kg, as received, < 0.2% moisture specified)
◦
was heated to 130 C in a 1.5 L glass reaction vessel with
mechanical stirring and 8% w/w (corn stover, 4.8% moisture)
or 10% w/w (Avicel PH-101, 3% moisture). After around 3 h
when the majority (corn stover) or all (Avicel) had dissolved,
the dissolved corn stover or Avicel was allowed to cool to below
Experimental
Chemicals
All chemicals used were reagent grade. Trichoderma viride cel-
lulase (Sigma Cat # C9422), sodium acetate, 3,5-dinitrosalicylic
acid (DNS), sodium hydroxide, sodium potassium tar-
trate and carboxymethylcellulose (CMC) was obtained from
Sigma (St. Louis, MO). 1-ethyl-3-methylimidazolium acetate
◦
80 C and was added to 2 L of 95% ethanol with rapid agitation
to induce the precipitation of dissolved materials. The resulting
slurry was filtered under pressure through polypropylene filter
cloth and the solids redispersed in 2 L of additional ethanol. The
filtration and redispersion steps were repeated twice to remove
(
[C2mim][OAc]) was obtained from BASF, USA.
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 338–345 | 343

View Article Online
most residual ionic liquid and the filter cake dried under vacuum
at 40 C to yield a free-flowing powder product. Residual free
ionic liquid remaining in the dried biomass is estimated to be
less than 4% (w/w).
CMC in 100 mM sodium acetate buffer (pH 4.8) at various
temperatures. Similar experiments in 50 mM MES buffer, pH 6.4
was done for Pho EG. Thermostability was determined by
◦
◦
incubating the enzyme at 80 C in 100 mM acetate buffer, pH 4.8
for Cel5A and 50 mM MES buffer, pH 6.4 for Pho EG and for
24 h prior to determining specific activity.
Enzyme activity assays on soluble and insoluble polysaccharides
Enzymatic activity was measured on soluble polysaccharide
substrates, CMC, using the 3,5-dinitrosalicylic acid (DNS)
Differential scanning calorimetry
34
reducing sugar assay. Briefly, 2% CMC (w/v) in 120 mL acetate
buffer, (100 mM, pH 4.5 and pH 4.8 for T. viride and Tma,
respectively, and 100 mM MES buffer, pH 6.4 for P. hori EG)
-1
Unfolding temperatures of Tma Cel5A (5 mg mL ) in the
absence and presence of 5–50% (v/v) of [C2mim][OAc] were
determined with a multi-cell differential scanning calorimeter
◦
was incubated at 80 C for 30 min for Tma Cel5A and Pho
(
Calorimetry Sciences Corporation, Lindon, UT, USA). The
◦
◦
enzymes and 37 C and 10 min for T. viride. The solution was
[
C2mim][OAc] containing samples were dialyzed overnight
cooled down to 4 C and 80 mL of DNS solution was added. The
against 100 mM sodium acetate buffer (pH 4.8) in the presence
of similar amounts [C2mim][OAc] as in sample. The dialyzed
◦
reactants were incubated at 95 C for 5 min and cooled down
to room temperature before the absorbance was read at 540 nm.
The reducing sugar concentration in the sample was calculated
from its absorbance using the standard curve of D-glucose and
cellobiose. All experiments were run in triplicate. For hydrolysis
reactions in the presence of [C2mim][OAc], control reactions
with CMC in the presence of IL but without no enzyme was
subtracted from each measurement.
◦
enzyme was scanned between 25 and 100 C using a scan
◦
-1
rate of 0.5 C min . The enzyme scans were corrected with
-1
a buffer–buffer baseline. The T. viride cellulase (2 mg mL ) was
run under similar conditions in the presence of 0–20% (v/v)
[
C2mim][OAc]. Unfolding temperatures of Pho EG (2.6 mg
-1
mL ) were determined similarly to the Tma but on a NanoDSC,
Differential Scanning Calorimeter (TA Instruments—Waters
LLC, New Castle, DE). The protein sample was dialyzed
overnight against the presence of similar amounts of IL as in
The solid substrate (Avicel and corn stover) hydrolysis reac-
◦
tions were conducted at 80 C for 15 h for the recombinant
◦
enzymes and at 37 C for 8 h in the case of T. viride cellulase
◦
-1
sample in the pHopt buffer. A scan rate of 2 C min was used
and the data analyzed using the NanoAnalyze software supplied
by the manufacturer.
at the optimum pH for each enzyme. The reaction products
were monitored using an Agilent 1200 HPLC equipped with
Varian 380-LC Evaporative Light Scattering Detector. The
total reaction volumes were 500 mL and shaken at 900 rpm at
temperature controlled shakers. Enzyme loadings were 0.4 mg
per g of glucan and each data point was measured in triplicate.
Separation was achieved using a Varian/Polymer labs Hi-Plex
Pb carbohydrate analytical column (300 ¥ 7.7 mm) with a
Acknowledgements
The DSC experiments were performed in the laboratory of Prof.
Francis C. Szoka, Jr., Department of Bioengineering and Ther-
apeutic Sciences, University of California, San Francisco, with
the help of Dr Zhaohua Huang and with the help of Dr Mara
Bryan at Energy Biosciences Institute, Berkeley. This work was
part of the DOE Joint BioEnergy Institute (http://www.jbei.org)
supported by the U. S. Department of Energy, Office of Science,
Office of Biological and Environmental Research, through
contract DE-AC02-05CH11231 between Lawrence Berkeley
National Laboratory and the U. S. Department of Energy.
◦
guard column (50 ¥ 7.7 mm) at 85 C (Polymer Laboratories,
Varian Inc., Shropshire, UK). The mobile phase was deionized
-
1
water with a flow rate of 0.6 mL min . Hydrolysis reactions
in the presence of [C2mim][OAc] were monitored by either
by High Performance Anion Exchange Chromatography with
Pulsed Amperometric Detection (HPAEC-PAD) on a Dionex
DX600 equipped with a Dionex Carbopac PA-20 analytical
column (3 ¥ 150 mm) and a Carbopac PA-20 guard column
(
0
3 ¥ 30 mm) (Dionex, Sunnyvale, CA). Eluent flow rate was
.4 mL min- and the temperature was 30 C. A gradient
1
◦
References
consisting of a 12 min elution with 14 mM NaOH followed
by a 5 min ramp to 450 mM NaOH for 20 min, then a return
to the original NaOH concentration of 14 mM for 10 min prior
to the next injection. Product concentrations were determined
by the integrations of the appropriate peaks from the HPLC or
HPAEC chromatograms.
1
M. Klare, Blood and Oil: The Dangers and Consequences of America’s
Growing Dependency on Imported Petroleum, Macmillan, London,
2005.
2
3
J. Barnett, Global Environ. Change, 2003, 13, 7–17.
B. Simmons, D. Loque and H. Blanch, GenomeBiology, 2008, 9,
2
42–247.
4 J. D. Wright, Energy Progress, 1988, 8, 71–80.
5
6
L. R. Lynd, P. J. Weimer, W. H. van Zyl and I. S. Pretorius, Microbiol.
Mol. Biol. Rev., 2002, 66, 506–577.
Estimation of temperature dependence and thermostability
R. Atalla and D. VanderHart, Science, 1984, 223, 283–286.
◦
7 D. Klemm, B. Philipp, T. Heinze, U. Heinze, and W. Wagenknecht,
Comprehensive Cellulose Chemistry, Wiley–VCH Verlag GmbH,
Berlin, 1998.
T. viride enzyme was assayed at 37 C and pH 4.5 as suggested
by the manufacturer. The pH dependence of the recombinant
enzymes was determined by measuring specific activities of the
enzyme on CMC in a series of pH values between 4 and 5 with
8 T. A. Hsu, M. R. Ladisch and G. T. Tsao, Chem. Technol., 1980, 10,
315–319.
9
N. Mosier, C. E. Wyman, B. Dale, R. Elander, Y. Lee, M. Holtzapple
and M. Ladisch, Bioresour. Technol., 2005, 96, 673–686.
1
00 mM sodium acetate buffer, and between 5 and 8 with sodium
phosphate buffer. The optimal temperature of Tma Cel5A was
determined by measuring specific activity of the enzyme on
1
0 R. P. Swatloski, S. K. Spear, J. D. Holbrey and R. D. Rogers, J. Am.
Chem. Soc., 2002, 124, 4974–4975.
3
44 | Green Chem., 2010, 12, 338–345
This journal is © The Royal Society of Chemistry 2010

View Article Online
1
1
1
1
1
1
1
1 F. Van Rantwijk and R. A. Sheldon, Chem. Rev., 2007, 107, 2757–
785.
2 R. A. Sheldon, R. M. Lau, M. J. Sorgedrager and F. van Rantwijk,
Green Chem., 2002, 4, 147–151.
24 K. Nelson, R. Clayton, S. Gill, M. Gwinn, R. Dodson, D. Haft,
2
E. Hickey, J. Peterson, W. Nelson, K. Ketchum, L. McDonald, T.
Utterback, J. Malek, K. Linher, M. Garrett, A. Stewart, M. Cotton,
M. Pratt, C. Phillips, D. Richardson, J. Heidelberg, G. Sutton, R.
Fleischmann, J. Eisen, O. White, S. Salzberg, H. Smith, J. Venter and
C. Fraser, Nature, 1999, 399, 323–329.
3 S. Zhu, Y. Wu, Q. Chen, Z. Yu, C. Wang, S. Jin, Y. Ding and G. Wu,
Green Chem., 2006, 8, 325–327.
4 Y. Fukaya, K. Hayashi, M. Wada and H. Ohno, Green Chem., 2008,
25 S. Arumugam Mahadevan, S. Gon Wi, D.-S. Lee and H.-J. Bae,
FEMS Microbiol. Lett., 2008, 287, 205–211.
26 S. R. Chhabra, K. R. Shockley, D. E. Ward and R. M. Kelly, Appl.
Environ. Microbiol., 2002, 68, 545–554.
27 S. Ando, H. Ishida, Y. Kosugi and K. Ishikawa, Appl. Environ.
Microbiol., 2002, 68, 430–433.
28 Robert D. Perlack, Lynn L. Wright, Anthony F. Turhollow, Robin L.
Graham, Bryce J. Stokes, and D. C. Erbach, (DOE/GO-102005-
2135), Oak Ridge National Laboratory, Oak Ridge, TN, 2005.
29 R. L. Graham, R. Nelson, J. Sheehan, R. D. Perlack and L. L. Wright,
Adv. Agron., 2007, 92, 1–11.
30 L. D. Unsworth, John van der Oost and Sotirios Koutsopoulos,
FEBS J., 2007, 274, 4044–4056.
31 M. Turner, S. Spear, J. Huddleston, J. Holbrey and R. Rogers, Green
Chem., 2003, 5, 443–447.
32 J. Pottk a¨ mper, P. Barthen, N. Llmberger, U. Schwaneberg, A. Schenk,
M. Schulte, N. Ignatiev and W. Streit, Green Chem., 2009, 11, 957–
965.
1
0, 44–46.
5 A. P. Dadi, S. Varanasi and C. A. Schall, Biotechnol. Bioeng., 2006,
5, 904–910.
6 A. P. Dadi, C. A. Schall and S. Varanasi, Appl. Biochem. Biotechnol.,
9
2
007, 137–140, 407–421.
7 M. J. Earle, J. M. Esperan c¸ a, M. A. Gilea, J. N. Lopes, L. P. Rebelo,
J. W. Magee, K. R. Seddon and J. A. Widegren, Nature, 2006, 439,
8
31–834.
1
1
8 S. Singh and B. A. Simmons, Biotechnol. Bioeng., 2009, 104, 68–
7
5.
9 I. Kilpel a¨ inen, H. Xie, A. King, M. Granstrom, S. Heikkinen and
D. S. Argyropoulos, J. Agric. Food Chem., 2007, 55, 9142–9148.
0 C. Li, Q. Wang and Z. Zhao, Green Chem., 2008, 10, 177–182.
1 M. B. Turner, S. K. Spear, J. G. Huddleston, J. D. Holbrey and R. D.
Rogers, Green Chem., 2003, 5, 443–447.
2
2
2
2
2 H. Zhao, C. L. Jones, G. A. Baker, S. Xia, O. Olubajo and V. N.
Person, J. Biotechnol., 2009, 139, 47–54.
3 P. Turner, M. Gashaw and Eva Nordberg Karlsson, Microb. Cell
Fact., 2007, 6, 9.
33 S. Voget, H. Steele and W. Streit, J. Biotechnol., 2006, 126, 26–36.
34 G. L. Miller, Anal. Chem., 1959, 31, 426–428.
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 338–345 | 345