Efficient biotransformation of D-fructose to D-mannose by a thermostable D-lyxose isomerase from Thermosediminibacter oceani

Article abstract of DOI:10.1016/j.procbio.2016.08.023

D-Mannose has prebiotic effect and potential medical application. Besides, it can be used as substrate to produce mannitol, a functional polyol widely used in food industry. As this result, it has attracted many researchers’ attention. In this work, a thermostable D-mannose-producing D-lyxose isomerase (D-LI) was characterized from a hyperthermophile, Thermosediminibacter oceani. The recombinant D-LI could be remarkably activated by Mn²⁺. It displayed maximal activity in presence of 1 mM Mn²⁺ at pH 6.5 and 65 °C, and was determined to be highly thermostable at 80 °C. The half-life was calculated to be 5.64, 2.82, 0.77, and 0.2 h at 70, 75, 80, and 85 °C, respectively. The enzyme showed the optimum activity using D-lyxose as substrate and could also effectively catalyze the isomerization between D-fructose and D-mannose. Under optimum conditions, 101.6 g/L D-mannose was produced from 400 g/L D-fructose after reaction for 9 h, giving a conversion yield of 25.4%.

Full text of DOI:10.1016/j.procbio.2016.08.023

G Model
PRBI-10782; No. of Pages8
ARTICLE IN PRESS
Process Biochemistry xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Process Biochemistry
journal homepage: www.elsevier.com/locate/procbio
Efficient biotransformation of d-fructose to d-mannose by a
thermostable d-lyxose isomerase from Thermosediminibacter oceani
Lina Yua, Wenli Zhanga, Tao Zhanga, Bo Jianga,b, Wanmeng Mua,b,∗
a State Key Laboratory of Food Science and Technology, Jiangnan University, Wuxi, Jiangsu, 214122, China
b Synergetic Innovation Center of Food Safety and Nutrition, Jiangnan University, Wuxi 214122, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 6 June 2016
Received in revised form 22 July 2016
Accepted 18 August 2016
Available online xxx
d-Mannose has prebiotic effect and potential medical application. Besides, it can be used as substrate to
produce mannitol, a functional polyol widely used in food industry. As this result, it has attracted many
researchers’ attention. In this work, a thermostable d-mannose-producing d-lyxose isomerase (D-LI)
was characterized from a hyperthermophile, Thermosediminibacter oceani. The recombinant D-LI could
be remarkably activated by Mn2+. It displayed maximal activity in presence of 1 mM Mn2+ at pH 6.5 and
65 ◦C, and was determined to be highly thermostable at 80 ◦C. The half-life was calculated to be 5.64,
2.82, 0.77, and 0.2 h at 70, 75, 80, and 85 ◦C, respectively. The enzyme showed the optimum activity using
d-lyxose as substrate and could also effectively catalyze the isomerization between d-fructose and d-
mannose. Under optimum conditions, 101.6 g/L d-mannose was produced from 400 g/L d-fructose after
reaction for 9 h, giving a conversion yield of 25.4%.
Keywords:
Thermosediminibacter oceani
d-Lyxose isomerase
d-Mannose
Thermostability
© 2016 Elsevier Ltd. All rights reserved.
1. Introduction
tion, this reaction theoretically results in 50% (W/W) and 25%
(W/W) mannitol from pure d-fructose [8] and invert sugar (50:
d-Mannose is the C-2 epimer of d-glucose and the aldose isomer
of d-fructose. It is a naturally occurring monosaccharide, existing
as the monomer unit in the polysaccharide mannan and an impor-
tant component of many glycoproteins, therefore, it theoretically
exists in a wide range of foods. d-Mannose has prebiotic effect and
induces expression of pro- and anti-inflammatory cytokines [1].
As a supplement, d-mannose could reduce Salmonella typhimurium
contamination to prevent urinary tract infections [2] and could
treat the phosphomannose isomerase deficiency disease [3]. Fur-
thermore, d-mannose is widely used as a cheap starting/basic
material for synthesizing vitamins [4], immunostimulatory [5] and
antitumor agents [6]. In addition, d-mannose is an important mate-
rial to produce mannitol, which is a high-value sugar alcohol widely
used in the food, pharmaceutical, medical, and chemical industries
[7].
50 d-fructose/d-glucose mixture) [9], respectively. However, if
d-fructose is previously converted to D-manose by isomerase,
then d-mannose can be further hydrogenated into 100% mannitol
[10,11].
Biological production of d-mannose has been studied using sev-
eral kinds of enzymes. d-Mannose isomerase (D-MI, EC 5.3.1.7)
catalyzes the reversible isomerization reaction between d-fructose
and d-mannose [12]. The conversion of d-fructose to d-mannose
has previously been performed using D-MIs from Escherichia coli
K12 [13], Pseudomonas cepacia [14], and Agrobacterium radiobac-
ter M-1 [15]. A cellobiose 2-epimerase (2-CE, EC 5.1.3.11) from
Caldicellulosiruptor saccharolyticus DSM 8903 is able to catalyze the
epimerization of d-glucose to d-mannose, but the reaction also
catalyzes the isomerization of d-glucose to d-fructose as a byprod-
uct [16]. The production of d-mannose has also been performed
by isomerization of d-fructose using d-lyxose isomerases (D-LI, EC
5.3.1.15) from Serratia proteamaculans KCTC 2936 [17] and Provi-
dencia stuartii KCTC 2568 [18].
Nowadays, mannitol is mainly produced by chemical hydro-
genation catalysis from d-fructose or invert sugar (d-fructose
and d-glucose mixture). But due to the selectivity of hydrogena-
D-LI is an aldo-keto isomerase catalyzing the isomerization
reaction between d-lyxose and d-xylulose. So far, D-LIs have
been characterized from Aerobacter aerogenes PRL-R3 [19], Bacillus
licheniformis DSM13 [20], Cohnella laevoribosii RI-39 [21], Dictyo-
glomus turgidum DSM 6724 [22], Escherichia coli O157:H7 [23], P.
∗
Corresponding author at: State Key Laboratory of Food Science and Technology,
Jiangnan University, Wuxi, Jiangsu, 214122, China.
E-mail address: wmmu@jiangnan.edu.cn (W. Mu).
http://dx.doi.org/10.1016/j.procbio.2016.08.023
1359-5113/© 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: L. Yu, et al., Efficient biotransformation of d-fructose to d-mannose by a thermostable d-lyxose
isomerase from Thermosediminibacter oceani, Process Biochem (2016), http://dx.doi.org/10.1016/j.procbio.2016.08.023

G Model
PRBI-10782; No. of Pages8
ARTICLE IN PRESS
2
L. Yu et al. / Process Biochemistry xxx (2016) xxx–xxx
stuartii KCTC 2568 [24], and S. proteamaculans KCTC 2936 [17]. D-LI
exhibits broad substrate specificity with highest activity toward d-
lyxose and may also catalyze the isomerization between d-fructose
and d-mannose. However, only the D-LIs from S. proteamaculans
[17] and P. stuartii [18] have been used for biological production
of d-mannose from d-fructose [24]. Except the thermophilic D.
turgidum DSM 6724 and C. laevoribosii RI-39, all previously reported
D-LIs are from mesophilic microorganisms. For industrial applica-
tion of aldose isomerases generally thermostable enzymes, such
as D-xylose isomerase [25] and l-arabinose isomerase [26], are
required.
Thermosediminibacter oceani DSM 16646 is an anaerobic hyper-
thermophilic bacterium isolated from deep sea sediments of Peru
Margin and the optimal growth temperature is 65 ◦C [27]. The com-
plete genome sequence of the strain has recently been determined
and deposited in GenBank with accession No. NC 014377 [28]. The
genome sequence reveals the presence of a putative D-LI gene
(locus tag: CP002131.1; protein ID: ADL08607.1). In this study, the
putative D-LI gene was cloned and expressed in Escherichia coli
BL21(DE3). This recombinant D-LI was then purified by metal affin-
ity chromatography, characterized and the d-mannose production
from d-fructose was studied.
with a 2 s on: 3 s off cycle. The insoluble cell debris was removed by
centrifugation at 10,000 × g for 30 min. The recombinant enzyme,
expressed as 6 × histidine-tagged fusion protein, was purified by
nickel-affinity chromatography using an ÄKTA purifier system (GE
Healthcare, Sweden). The centrifuged supernatant (15 mL) as crude
enzyme was filtrated by a 0.22-␮m Millipore filter and then loaded
onto a 5-mL HisTrap HP Ni-NTA column (GE Healthcare, Sweden),
which was chelated with Ni2+ and pre-equilibrated with Binding
Buffer (50 mM sodium phosphate buffer, 500 mM NaCl, pH 6.5).
Then, the Binding buffer and Washing buffer (50 mM sodium phos-
phate buffer, 500 mM NaCl, 20 mM imidazole, pH 6.5) were used
in turn to remove the impurity protein. At last, the 6 × histidine-
tagged target protein was eluted from the column using an Elution
Buffer (50 mM sodium phosphate buffer, 500 mM NaCl, 500 mM
imidazole, pH 6.5). The collected enzyme solution was dialyzed
against dialysate A (50 mM sodium phosphate buffer, 10 mM ethyl-
enediamine tetraacetic acid (EDTA), pH 6.5) for 12 h, and then
against dialysate B (50 mM sodium phosphate buffer, pH 6.5) at
4 ◦C for another 12 h to remove EDTA.
2.5. Molecular mass determination
The subunit molecular mass was examined by sodium dode-
cyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) under
denaturing conditions. The separating and stacking gels were 12%
and 5% (w/v) of acrylamide, respectively. Coomassie brilliant blue
R250 was used to stain proteins for visualization.
The molecular mass of native enzyme was assessed by gel
filtration using high-performance liquid chromatography (HPLC,
Agilent 1200 LC Systems, Agilent technologies, Santa Clara, CA, USA)
equipped with a Diode Array Detector and a TSKgel G2000SWxl col-
umn (125 Å, 5 ␮m, 7.8 mm id × 30 cm, Tosoh Bioscience LLC, Tokyo,
Japan). The HPLC conditions were as follows: mobile phase, 0.1 M
phosphate buffer (pH 6.5) containing 0.05% (W/V) NaN3 and 0.1 M
Na2SO4; column temperature, 25 ◦C; flow rate, 1 mL/min; detection
wavelength, 260 nm.
2. Materials and methods
2.1. Chemicals and reagents
The carbohydrate standards, including d-fructose, d-mannose,
d-lyxose, and d-xylulose were obtained from Sigma (St. Louis, MO,
USA). Tryptone and yeast extract for Luria-Bertani (LB) broth were
purchased from Difco (Detroit, MI, USA). All other chemicals were
at least of analytical grade obtained from Sinopharm Chemical
Reagent (Shanghai, China) and Sigma (St. Louis, MO, USA).
2.2. Bacterial strain, plasmid, and culture conditions
The E. coli BL21(DE3) strain was used for heterologous expres-
sion and the plasmid pET-22b(+) was used as an expression vector.
The E. coli cells were cultivated in Luria-Bertani (LB) broth (pH 7.5),
composed of 10 g/L trypone, 5 g/L yeast extract, and 5 g/L of NaCl.
2.6. Enzyme assay
The enzyme activity was determined by measuring the accu-
mulation of d-fructose using d-mannose as a substrate. Unless
otherwise stated, the reaction was carried out in 50 mM sodium
phosphate buffer (pH 6.5) containing 10 mM d-mannose, 1 mM
MnCl2 and 0.2 U enzyme/mL at 65 ◦C for 10 min. After incubation,
the reaction was stopped by the addition of trichloroacetic acid to
final concentration of 150 mM. One unit of enzyme activity was
defined as the amount of enzyme that produced 1 uM d-fructose
from d-mannose per min at pH 6.5 and 65 ◦C.
2.3. Gene cloning and expression
The putative D-LI gene (locus tag: CP002131.1; protein ID:
ADL08607.1) from T. oceani DSM 16646 has been published in
the open GenBank database. In this work, the full length of the
target gene was commercially synthesized by Shanghai Generay
Biotech Co., Ltd (Shanghai, China), with an in-frame C-terminal
6 × histidine-tag sequence, and was cloned into the expression vec-
tor pET-22b(+) with NdeI and XhoI restriction sites at the 5ꢀ- and
3ꢀ-terminus.
2.7. Effects of metal ions, pH, and temperature
The constructed plasmid harboring T. oceani D-LI genes was
transformed into E. coli BL21(DE3) strain for overexpression. The
recombinant strain was grown at 37 ◦C in LB medium contain-
ing 100 ␮g/mL of ampicillin. When the optical density at 600 nm
reached 0.6, isopropyl-␤-d-1-thiogalactopyranoside (IPTG) was
added to the culture medium at 0.1 mM to induce the enzyme
expression and the culture was further grown at 28 ◦C for 6 h.
To investigate the effect of divalent metal ions on enzyme activ-
ity, the enzyme assay was carried out after adding 1 mM of each
divalent metal ion such as CoCl2, MnCl2, NiSO4, MgCl2, CaCl2, ZnCl2,
CuSO4, FeSO4 or BaCl2 to the divalent metal-free enzyme. The reac-
tions were performed in 50 mM sodium phosphate buffer (pH 6.5)
at 65 ◦C. The activity of enzyme without adding any divalent metal
ion was taken as 100%.
2.4. Purification of recombinant D-LI
To study the effect of pH on the recombinant D-LI, pH was var-
ied from 4.5 to 9.0 using 50 mM acetate buffer (pH 4.0–5.5), 50 mM
phosphate buffer (pH 6.0 − 7.0), and 50 mM Tris-HCl buffer (pH
7.5–9.0). The effect of temperature on the enzyme activity was eval-
uated in 50 mM phosphate buffer (pH 6.5) at various temperatures,
from 50 to 85 ◦C. The activities at each pH or temperature were
relative to the highest activity value (100%).
All purification steps were conducted at 4 ◦C. The grown cells of
E. coli were harvested from 200 mL culture broth by centrifugation
at 10,000 × g for 15 min, and then were resuspended in 15 mL cell
lysis buffer (50 mM sodium phosphate buffer, 100 mM NaCl, pH
6.5). The harvested cells were disrupted by sonication for 12 min
Please cite this article in press as: L. Yu, et al., Efficient biotransformation of d-fructose to d-mannose by a thermostable d-lyxose
isomerase from Thermosediminibacter oceani, Process Biochem (2016), http://dx.doi.org/10.1016/j.procbio.2016.08.023

G Model
PRBI-10782; No. of Pages8
ARTICLE IN PRESS
L. Yu et al. / Process Biochemistry xxx (2016) xxx–xxx
3
Fig. 1. Multiple sequence alignment of D-LIs and their homologs. Amino acid sequence for D-LI from T. oceani DSM 16646 (TO-LI; GeneBank accession No: ADL08607.1) was
aligned with B. subtilis D-LI (BS-LI; AIY91703), B. licheniformis D-LI (BL-LI; AAU22106.1), D. turgidum D-LI (DT-LI; YP 002352606.1), and C. laeviribosi D-LI (CL-LI; ABI93960.1).
The strictly conserved residues are shown on a red background, and the highly conserved residues are shown in red type and boxed in blue. The secondary structure elements
and the residues involved in the metal coordinating sites (the symbol star above the residues) are symbolized based on the previously reported structure of B. subtilis D-LI
(PDB No. 2Y0O A) [29]. The alignment was performed using ESPript [42]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web
version of this article.)
The thermal stability of the recombinant D-LI was studied by
pre-incubating the enzyme in 50 mM sodium phosphate buffer (pH
6.5) at temperatures ranging from 70 to 85 ◦C. The residual activities
withdrawn at different time intervals were determined after the
isomerization reaction from d-mannose to d-fructose. The initial
activity of enzyme without incubation was taken as 100%.
used to measure the kinetic parameters. The reactions were
performed in 50 mM phosphate buffer (pH 6.5) at 65 ◦C contain-
ing 1 mM Mn2+
. The enzyme kinetic parameters, including
Michaelis–Menten constant (Km), turnover number (kcat),
and catalytic efficiency (kcat/Km), were determined using
Michaelis–Menten equation and Lineweaver-Burk plots.
2.8. Determination of kinetic parameters
2.9. Enzymatic production of d-mannose from d-fructose
Various concentrations (from 10 mM to 200 mM) of monosac-
charide substrates (d-lyxose, d-mannose, and d-fructose) were
Biological production of d-mannose was performed in 50 mM
phosphate buffer (pH 6.5) containing 1 mM Mn2+ and 4 U/mL
Please cite this article in press as: L. Yu, et al., Efficient biotransformation of d-fructose to d-mannose by a thermostable d-lyxose
isomerase from Thermosediminibacter oceani, Process Biochem (2016), http://dx.doi.org/10.1016/j.procbio.2016.08.023

G Model
PRBI-10782; No. of Pages8
ARTICLE IN PRESS
4
L. Yu et al. / Process Biochemistry xxx (2016) xxx–xxx
enzyme at 60 ◦C. 100, 200, 300 and 400 g/L of d-fructose were used
as initial concentrations of substrate.
2.10. Analysis of monosaccharides
The concentrations of monosaccharides were determined using
a Bio-LC system (Dionex ICS-5000, Sunnyvale, CA, USA) with a pulse
ampere detector and a CarboPac PA20 column (6.5 ␮m, 3 × 150 mm,
Thermo Fisher Scientific, Massachusetts, USA). NaOH (20 mM) was
used as the mobile phase with the flow rate of 0.5 mL/min. The
column temperature was 30 ◦C.
3. Results and discussion
3.1. Heterologous expression and protein purification
So far, D-LIs from C. laevoribosii (GenBank accession No.
ABI93960.1, 182 aa) [21], B. licheniformis DSM13 (AAU22106.1, 167
aa) [20], D. turgidum DSM 6724 (YP 002352606.1, 181 aa) [22],
Bacillus sutbilis strain 168 (AIY91703, 167 aa) [29], S. proteamac-
ulans (BAJ07463.1, 228 aa) [17], E. coli (Q8 × 5Q7, 227 aa) [23]
and P. stuartii KCTC2568 [24] have been identified in recombinant
forms. To find a D-LI with good thermostability, the protein blast
search using the known D-LIs as controls was performed and the
thermophilic bacterium sources were selected for further analysis.
Since the amino acid sequence of the D-LI from P. stuartii KCTC2568
is not available in GenBank, and the S. proteamaculans D-LI and
E. coli D-LI have much higher molecular weight and show no sig-
nificant amino acid sequence similarity with other characterized
D-LIs, those three are not used as controls. The putative D-LI from
a thermophilic strain, T. oceani DSM 16646 (ADL08607.1, 181 aa),
exhibited 69%, 65%, 65%, and 64% amino acid sequence identity to
the D-LIs from B. sutbilis, C. laevoribosii, B. licheniformis DSM13, and
D. turgidum DSM 6724, respectively (Fig. 1). The crystal structures
of D-LI from B. sutbilis and E. coli have recently been determined and
released with PDB No. 2Y0O and 3MPB, respectively [29]. The struc-
ture of B. sutbilis D-LI forms a dimer with a molecule related by the
crystallographic twofold axis. It contains two divalent metal coor-
dinating sites, which are composed of three histidines (H69, H71,
and H137) and one glutamic acid (E82). What’s more, these four
residues are strictly conserved across all D-LIs including T. oceani
DSM 16646 (Fig. 1). Although the E. coli D-LI shares a low sequence
identity (24%) with B. sutbilis D-LI, it is also a dimer in the asymmet-
ric unit. The overall fold of is a cupin-type ␤-barrel which forms a
deep hydrophobic pocket. Interestingly, the metal-binding site in
the hydrophobic pocket is also conserved. Similar to other D-LIs,
the metal is coordinated by H103, H105, E110, H171 and two water
molecules [23].
In this work, the putative D-LI from the thermophilic strain T.
oceani DSM 16646 was selected for study in order to obtain a ther-
mostable d-mannose-producing D-LI. The putative D-LI-encoding
gene (GenBank accession No. CP002131.1) was heterologously
expressed in E. coli BL21(DE3) with C-terminal 6 × histidine-tag.
The theoretical molecular weight of the recombinant protein was
calculated to be 21,664 Da based on the 181 amino acid residues
and six histidine residues, as measured using the ExPASy Computer
pI/Mw tool. As shown in Fig. 2, the target protein didn’t appear
without adding inducer. As the induction time increased, the band
of target protein became more and more obvious. After expression
by IPTG induction for 6 h at a low temperature (28 ◦C), a strong pro-
tein band of approximately 22 kDa appeared on the electrophoresis
gel of SDS-PAGE, compared to the electrophoresis result of the con-
trol E. coli BL21(DE3) (data not shown), indicating that the target
recombinant protein with the predicted molecular weight had been
overexpressed. The recombinant enzyme was purified to homo-
Fig. 2. SDS-PAGE analysis of proteins. lane 1, whole recombinant E. coli cells without
IPTG induction; lane 2–4, the cells after IPTG induction for 2, 4, and 6 h, respectively;
lane 5, purified recombinant T. oceani D-LI by metal affinity chromatography; Lane
6, protein markers. The proteins on gel were stained by Coomassie brilliant blue
R250.
geneity, exhibiting a single band of 22 kDa in SDS-PAGE (Fig. 2)
and the native purified enzyme existed as a homodimer with a
molecular weight of 44 kDa as measured by gel filtration chro-
matography, using TSK G2000SWxl column (data not shown). D-LIs
from C. laevoribosii [21], D. turgidum DSM 6724 [22], and P. stuar-
tii KCTC2568 [24] were also identified as homodimers with subunit
molecular weight of 21 or 22 kDa S. proteamaculans D-LI was identi-
fied as a homodimer as well [17], but the subunit molecular weight
(27 kDa) was much higher than other D-LIs and it displayed very
low amino acid sequence identity with other D-LIs as mentioned
above.
3.2. Effect of metal ions on the enzyme activity
To investigate the effect of metal ions, the purified recombi-
nant T. oceani D-LI was firstly treated with EDTA to remove the
divalent metal ions and dialyzed against divalent metal-free phos-
phate buffer (50 mM, pH 6.5) to remove EDTA. The relative activity
was detected in presence of various divalent metal ions at the final
concentration of 1 mM (Table 1). The enzyme displayed a maxi-
mal activity in the presence of Mn2+. The Mn2+-bound D-LI showed
1250% of the relative activity compared to the metal-free enzyme.
The Ni2+, Co2+ and Fe2+ increased significantly the catalytic activity
to 1179%, 913%, and 214% of the initial relative activity, respec-
tively. However, other tested metal ions, especially Cu2+ and Zn2+
inhibited the enzyme activity up to 87% (Table 1).
,
Based on the B. subtilis D-LI structure data, the metal coor-
dinating sites are present in the active center of the enzyme;
therefore, the metal ion probably has important effect on the cat-
alytic activity [29]. All the reported D-LIs have been identified to
be metal-dependent enzymes, which especially are activated by
Mn2+ and Co2+. The D-LIs from C. laevoribosii [21], B. licheniformis
DSM13 [20], S. proteamaculans [17], E. coli O157:H7 and P. stuartii
KCTC2568 [24] also require Mn2+ as an optimal metal cofactor. Fur-
Please cite this article in press as: L. Yu, et al., Efficient biotransformation of d-fructose to d-mannose by a thermostable d-lyxose
isomerase from Thermosediminibacter oceani, Process Biochem (2016), http://dx.doi.org/10.1016/j.procbio.2016.08.023

G Model
PRBI-10782; No. of Pages8
ARTICLE IN PRESS
L. Yu et al. / Process Biochemistry xxx (2016) xxx–xxx
5
Table 1
Effect of Metal Ions on the Recombinant T. oceani D-LI.
Metal ion (1 mM)
Relative activity (%)a
None
Mn2+
Ni2+
100
4
1250 59
1,179 68
913 42
214 18
Co2+
Fe2+
Ca2+
Mg2+
Ba2+
Cu2+
Zn2+
86
68
64
16
13
4
5
2
1
1
a
The activities were measured at pH 6.5 and 65 ◦C. Results were the mean values
of three experiments standard deviation.
Fig. 4. Effect of temperature on the activity of the recombinant T. oceani D-LI. The
relative activity was investigated in presence of 1 mM Mn2+ at pH 6.5 and vari-
ous temperatures. Values are means of three replications and error bar represents
standard deviation.
products [38]. Therefore, many studies focused on the identification
of aldose isomerases with slightly acidic pH optima [39] and the
molecular modification to shift the pH optima toward acidic side
[40].
3.4. Effects of temperature on the enzyme activity and stability
The activity was examined over a temperature range of 50 −
85 ◦C at pH 6.5 in presence of 1 mM Mn2+. The recombinant T. oceani
D-LI showed the highest activity at 65 ◦C, and retained 77% of max-
imal activity at 75 ◦C, but the relative activity dropped significantly
above 75 ◦C (Fig. 4). By comparison, the D-LIs from D. turgidum DSM
6724 [22] and C. laevoribosii [21] showed the highest relative activ-
ity at 75 and 70 ◦C, respectively, and the temperature optima of
other reported D-LIs were measured to be less than 50 ◦C (Table 2).
The effect of temperature on the enzyme stability was studied
at 70 − 85 ◦C (Fig. 5). The enzyme retained 65%, 51%, and 50% of
its initial activity after incubation at 70 ◦C for 4 h, 75 ◦C for 2 h, and
80 ◦C for 1 h, respectively. For the determination of half-life, the
first order deactivation kinetic model was selected. The stability
decreased remarkably when incubated at 85 ◦C, and the residual
activity dropped to less than 50% after 10 min. The decay constant
kd was calculated to be 0.123, 0.246, 0.904, and 3.473 at 70, 75,
80, and 85 ◦C, respectively. Therefore, based on the calculation for-
mula of half-life (t1/2) = ln2/kd, the t1/2 was determined to be 5.64,
2.82, 0.77, and 0.20 h at these temperatures, respectively. The ther-
mostability was also studied at the optimal temperature (65 ◦C),
and results showed that the enzyme almost did not lose any activ-
ity after incubation for 5 h and retained 80% of initial activity after
incubation for 50 h (data not shown). This data indicated that the
recombinant T. oceani D-LI showed a very good thermostability. By
comparison, the half-lives of D-LIs from S. proteamaculans [17], P.
stuartii KCTC2568, and B. licheniformis DSM13 were 0.09, 1.4, and
7 h at 50 ◦C, respectively. The thermostability of D-LI from the ther-
mophilic C. laevoribosii RI-39 was not reported. The D-LI from the
hyperthermophile, D. turgidum DSM 6724, has the optimum tem-
perature at 75 ◦C and a half-life of 9.1 h at 60 ◦C [22]. Although the
optimum temperature of T. oceani D-LI (65 ◦C) was relatively lower
than that of D. turgidum D-LI, the T. oceani D-LI exhibited much
higher thermostability.
Fig. 3. Effect of pH on the activity of the recombinant T. oceani D-LI. The relative
activity was investigated in presence of 1 mM Mn2+ at 65 ◦C and different pH values.
Values are means of three replications and error bar represents standard deviation.
thermore, the Co2+ is the optimal cofactor of D-LI from D. turgidum
DSM 6724 [22], but strongly inhibited the activity of P. stuartii D-LI
[24]. The activity enhancement by Mn2+ or Co2+ was also found in a
wide range of aldose isomerases and ketose epimerases, such as d-
glucose isomerase [25], l-arabinose isomerase [26], and D-tagatose
3-epimerase family enzymes [30]. Moreover, some enzymes were
identified to be strictly metal-dependent, which did not display
any activity without divalent metal ion, such as L-rhamnose iso-
merase from Thermoanaerobacterium saccharolyticum NTOU1 [31],
l-arabinose isomerase from Geobacillus stearothermophilus DSM22
[32], and D-psicose 3-epimerases from Clostridium species strains
[33–35].
3.3. Effects of pH on the enzyme activity
The enzyme assay was performed in presence of 1 mM Mn2+
.
The purified recombinant enzyme showed an optimal activity at pH
6.5, and retained more than 50% of maximal activity at pHs ranged
from 5.0 to 8.0 (Fig. 3). These results were in accordance with other
studies (done by whom) which stated that D-LI from C. laevoribosii
exhibited maximal activity at pH 6.5, and the optimal pH values
of the D-LIs from B. licheniformis DSM13, D. turgidum DSM 6724,
S. proteamaculans, E. coli O157:H7 and P. stuartii KCTC2568 were
determined to be in the alkaline range of pH 7.5 − 8.0. Generally, a
slightly acidic pH optimum is considered as an advantage for enzy-
matic reaction of reducing sugars in industry [36,37]. The acidic
pH condition can remarkably reduce the non-enzymatic reactions
which easily happen at alkaline pH and lead to unwanted by-
In general, commercial applications of aldose isomerases, such
as d-glucose isomerase [12] and l-arabinose isomerase [41], require
the enzymes to have a high optimum temperature, a good ther-
Please cite this article in press as: L. Yu, et al., Efficient biotransformation of d-fructose to d-mannose by a thermostable d-lyxose
isomerase from Thermosediminibacter oceani, Process Biochem (2016), http://dx.doi.org/10.1016/j.procbio.2016.08.023

G Model
PRBI-10782; No. of Pages8
ARTICLE IN PRESS
6
L. Yu et al. / Process Biochemistry xxx (2016) xxx–xxx
Table 2
Comparison of Biochemical Properties of Various Microbial D-LIs.
Microorganism
Subunit
molecular mass
(kDa)
Total molecular
mass (kDa)
Optimum pH
6.5
Optimum
Optimum metal
Mn2+
Half-life (h)
Reference
Temperature (◦C)
T.
oceani
DSM
16646
C. laevoribosii RI-39
B.
licheni-
formis
DSM13
22
44
65
12.7 (70 ◦C)
6.8 (75 ◦C)
1.7 (80 ◦C)
0.5 (85 ◦C)
NRa
This
study
21
19.5
42
NR
6.5
7.5–8.0
70
40–45
Mn2+
Mn2+
[21]
[20]
140 (30 ◦C)
62 (35 ◦C)
30 (40 ◦C)
18 (45 ◦C)
7 (50 ◦C)
D. turgidum DSM 6724
S.
pro-
tea-
mac-
u-
Ela.ncsoli O157:H7
P.
22
27
44
54
7.5
7.5
75
40
Co2+
Mn2+
9.1 (60 ◦C)
84 (30 ◦C)
17 (35 ◦C)
2.6 (40 ◦C)
0.3 (45 ◦C)
0.09 (50 ◦C)
NR
[22]
[17]
NR
22
NR
44
7.5
7.5
50
45
Mn2+
Mn2+
[23]
[24]
36 (30 ◦C)
14 (35 ◦C)
8.9 (40 ◦C)
3.4 (45 ◦C)
1.4 (50 ◦C)
stu-
ar-
tii
KCTC2568
a
NR, not reported.
Fig. 5. Effect of temperature on the stability of the recombinant T. oceani D-LI. Thermal stability was studied by incubating the enzyme at various temperatures for different
time intervals at pH 6.5. The parameter kd represents decay constant.
mostability, and a slightly acidic pH optimum. Elevated reaction
temperatures could increase the carbohydrate solubility, reduce
viscosity of reaction mixture, and decrease the possibility of
microbial contamination. High thermostability is helpful for the
biocatalysts to maintain activity and extend the reaction period.
Slightly acidic pH optimum reduces the non-enzymatic reactions
between the proteins and carbohydrates forming brown-colored
by-products. In this study, the recombinant T. oceani D-LI showed
the optimum pH and temperature at pH 6.5 and 65 ◦C, and was the
most thermostable of the D-LIs reported to date, indicating its good
potential for industrial application.
3.5. Substrate specificity and kinetic parameters
The specific activity of the recombinant T. oceani D-LI was
determined for d-lyxose, d-mannose, and d-fructose as substrates
(Table 3). The isomerization reactions showed only one product
from each substrate, and the optimum substrate was d-lyxose.
The specific activities were measured to be 7.1 0.2, 5.3 0.1, and
1.2 0.1 U/mg, for d-lyxose, d-mannose, and d-fructose, respec-
tively.
The kinetic parameters for the three monosaccharides were cal-
culated (Table 3). The Km for d-lyxose was much lower than those
for d-mannose and d-fructose, indicating that the recombinant
Please cite this article in press as: L. Yu, et al., Efficient biotransformation of d-fructose to d-mannose by a thermostable d-lyxose
isomerase from Thermosediminibacter oceani, Process Biochem (2016), http://dx.doi.org/10.1016/j.procbio.2016.08.023

G Model
PRBI-10782; No. of Pages8
ARTICLE IN PRESS
L. Yu et al. / Process Biochemistry xxx (2016) xxx–xxx
7
Table 3
Specific Activities and Kinetic Parameters of the Recombinant T. oceani D-LI.a
Substrate
Product
Specific activity (U/mg)
Km (mM)
kcat (min−1
)
kcat/Km (mM−1 min−1
)
d-lyxose
d-mannose
d-fructose
d-xylulose
d-fructose
d-mannose
7.1 0.2
5.3 0.1
1.2 0.1
15.5 1.2
32.8 1.8
27.3 1.3
3,108 32
5,686 39
1,030 26
201
173
38
6
5
1
a
Results were the mean values of three experiments standard deviation.
d-mannose concentration could still retain 52% of its initial value
after 35 cycles [17].
Compared with the conversion rate of other d-mannose produc-
ing enzymes (almost 20% − 25%), the conversion rate of T. oceani
D-LI was acceptable. Besides, some other rare sugar isomerase also
displayed the similar conversion rate, that D-psicose 3-epimerase
could convert d-fructose to D-psicose with the rate of 20% − 30%.
In the industrial production of rare sugars, high thermostability
and slightly acidic pH optimum were required, because high ther-
mostability could improve the utilization efficiency of the enzymes
and slightly acidic pH optimum was beneficial for reducing non-
enzymatic reactions and unwanted by-products. Interestingly, the
thermostable T. oceani D-LI exhibited maximum activity at acidic
pH (pH 6.5), which could meet the requirements of industrial pro-
duction.
4. Conclusion
Fig. 6. Enzymatic production of d-mannose from d-fructose by the recombinant T.
oceani D-LI. d-mannose was produced by 4 U/mL enzyme in 50 mM phosphate buffer
(pH 6.5) containing 1 mM Mn2+ at 60 ◦C, from four concentrations of d-fructose, 100,
200, 300 and 400 g/L. Values are means of three replications and error bar represents
standard deviation.
A thermostable d-mannose-producing D-LI was characterized
from a thermophilic strain, T. oceani DSM 16646. The purified
recombinant T. oceani D-LI showed a slightly acidic optimum pH
at 6.5 and a relatively high optimum temperature at 65 ◦C. It exhib-
ited a very high thermostability compared to other reported D-LIs
and was stable up to 80 ◦C for 1 h of incubation. All these data indi-
cated that the T. oceani D-LI had a great potential for industrial
d-mannose production.
T. oceani D-LI showed the highest affinity toward d-lyxose. The
kcat/Km values were 201 6, 173 5, and 38 1 for d-lyxose, d-
mannose, and d-fructose, respectively. All these results indicates
that d-lyxose was the optimum substrate for T. oceani D-LI. The kcat
and kcat/Km values for d-mannose were much higher than those for
d-fructose, the reaction from d-mannose to d-fructose was favor-
able. But d-mannose was much more expensive than d-fructose,
and it was meaningful to carried out the reverse reaction from d-
fructose to d-mannose. In addition, interestingly, the Km values of T.
oceani D-LI for d-lyxose, d-mannose, and d-fructose were slightly
lower than most other known D-LIs., while the kcat and kcat/Km
values were not prominent (data not shown).
Acknowledgements
This work was supported by the NSFC Project (No. 21276001),
the 863 Project (No. 2013AA102102), the Fundamental Research
Funds for the Central Universities (No. JUSRP51304A), the Sup-
port Project of Jiangsu Province (No. BK20130001), and the project
of outstanding scientific and technological innovation group of
Jiangsu Province (Jing Wu).
3.6. Production of d-mannose from d-fructose
References
d-Mannose production by the recombinant T. oceani D-LI was
performed at pH 6.5 and 60 ◦C in 100 mL of reaction volume (Fig. 6).
Upon adding 100, 200, 300 and 400 g/L d-fructose, the enzymatic
reaction reached the equilibrium at 2, 4.5, 6 and 9 h, exhibiting
conversion rates of 26.8%, 26.1%, 25.8% and 25.4% respectively.
As the substrate concentration added, the amount of produc-
tion increased, which led to the end-product inhibition and slight
decrease of conversion rate. During high-level d-mannose produc-
tion by T. oceani D-LI, 26.8, 52.2, 77.4 and 101.6 g/L of d-mannose
was produced from 100, 200, 300 and 400 g/L d-fructose, respec-
tively. By comparison, the S. proteamaculans D-LI produced 100 g/L
d-mannose from 500 g/L d-fructose in 5 h, with conversion ratio
of 20% [17]. d-Mannose production from d-fructose was also stud-
ied using free and immobilized P. stuartii D-LI. The free P. stuartii
D-LI produced 150 g/L d-mannose from 600 g/L d-fructose in 2 h,
with conversion ratio of 25%. The immobilization of D-LI displayed
a good potential for the industrial application. The immobilized P.
stuartii D-LI with Duolite A568 as carrier exhibited a 25% (W/W)
conversion yield of d-fructose to d-mannose after 23 cycles and the
[1] O.S. Korneeva, I.V. Cheremushkina, A.S. Glushchenko, N.A. Mikhailova, A.P.
Baturo, E.E. Romanenko, S.A. Zlygostev, Prebiotic properties of mannose and
its effect on specific resistance, Zhurnal Mikrobiol. Epidemiol. Immunobiol. 5
(2012) 67–70.
[2] B.A. Oyofo, J.R. Deloach, D.E. Corrier, J.O. Norman, R.L. Ziprin, H.H.
Mollenhauer, Prevention of Salmonella typhimurium colonization of broilers
with d-mannose, Poult. Sci. 68 (1989) 1357–1360.
[3] P. de Lonlay, N. Seta, The clinical spectrum of phosphomannose isomerase
deficiency, with an evaluation of mannose treatment for CDG-Ib, Biochim.
Biophys. Acta 1792 (2009) 841–843.
[4] F. Chen, J. Zhao, F. Xiong, B. Xie, P. Zhang, An improved synthesis of a key
intermediate for (+)-biotin from d-mannose, Carbohydr. Res. 342 (2007)
2461–2464.
[5] K. Ranta, K. Nieminen, F.S. Ekholm, M. Polakova, M.U. Roslund, T. Saloranta, R.
Leino, J. Savolainen, Evaluation of immunostimulatory activities of synthetic
mannose-containing structures mimicking the ␤-(1, 2)-linked cell wall
mannans of Candida albicans, Clin. Vac. Immunol. 19 (2012) 1889–1893.
[6] S.S. El-Nakkady, M.M. Hanna, H.M. Roaiah, I.A.Y. Ghannam, Synthesis,
molecular docking study and antitumor activity of novel 2-phenylindole
derivatives, Eur. J. Med. Chem. 47 (2012) 387–398.
[7] B.C. Saha, F.M. Racine, Biotechnological production of mannitol and its
applications, Appl. Microbiol. Biotechnol. 89 (2011) 879–891.
[8] A.W. Heinen, J.A. Peters, H. van Bekkum, Hydrogenation of fructose on Ru/C
catalysts, Carbohydr. Res. 328 (2000) 449–457.
Please cite this article in press as: L. Yu, et al., Efficient biotransformation of d-fructose to d-mannose by a thermostable d-lyxose
isomerase from Thermosediminibacter oceani, Process Biochem (2016), http://dx.doi.org/10.1016/j.procbio.2016.08.023

G Model
PRBI-10782; No. of Pages8
ARTICLE IN PRESS
8
L. Yu et al. / Process Biochemistry xxx (2016) xxx–xxx
[9] M. Makkee, A.P.G. Kieboom, H. van Bekkum, Production methods of
d-mannitol, Starch–Stärke 37 (1985) 136–141.
[10] D.K. Mishra, J.S. Hwang, Selective hydrogenation of d-mannose to d-mannitol
using NiO-modified TiO2 (NiO-TiO2) supported ruthenium catalyst, Appl.
Catal. A: Gen. 453 (2013) 13–19.
[11] F. Devos, Process for the production of mannitol, US Patent (1995) 5466–795.
[12] W. Mu, L. Yu, W. Zhang, T. Zhang, B. Jiang, Isomerases for biotransformation of
d-hexoses, Appl. Microbiol. Biotechnol. 99 (2015) 6571–6584.
[13] F.J. Stevens, P.W. Stevens, J.G. Hovis, T.T. Wu, Some properties of d-mannose
isomerase from Escherichia coli K12, J. Gen. Microbiol. 124 (1981) 219–223.
[14] P. Allenza, M.J. Morrell, R.W. Detroy, Conversion of mannose to fructose by
immobilized mannose isomerase from Pseudomonas cepacia, Appl. Biochem.
Biotechnol. 24–25 (1990) 171–182.
[15] J. Hirose, K. Maeda, H. Yokoi, Y. Takasaki, Purification and characterization of
mannose isomerase from Agrobacterium radiobacter M-1, Biosci. Biotechnol.
Biochem. 65 (2001) 658–661.
[16] C.S. Park, J.E. Kim, J.G. Choi, D.K. Oh, Characterization of a recombinant
cellobiose 2-epimerase from Caldicellulosiruptor saccharolyticus and its
application in the production of mannose from glucose, Appl. Microbiol.
Biotechnol. 92 (2011) 1187–1196.
[28] S. Pitluck, M. Yasawong, C. Munk, M. Nolan, A. Lapidus, S. Lucas, T.G. Del Rio,
H. Tice, J.F. Cheng, D. Bruce, C. Detter, R. Tapia, C. Han, L. Goodwin, K. Liolios,
N. Ivanova, K. Mavromatis, N. Mikhailova, A. Pati, A. Chen, K. Palaniappan, M.
Land, L. Hauser, Y.J. Chang, C.D. Jeffries, M. Rohde, S. Spring, J. Sikorski, M.
Goeker, T. Woyke, J. Bristow, J.A. Eisen, V. Markowitz, P. Hugenholtz, N.C.
Kyrpides, H.P. Klenk, Complete genome sequence of Thermosediminibacter
oceani type strain (JW/IW-1228P(T)), Stand. Genom. Sci. 3 (2010) 108–116.
[29] J. Marles Wright, R.J. Lewis, The structure of a d-lyxose isomerase from the
sigmaB regulon of Bacillus subtilis, Proteins 79 (2011) 2015–2019.
[30] W. Mu, W. Zhang, Y. Feng, B. Jiang, L. Zhou, Recent advances on applications
and biotechnological production of D-psicose, Appl. Microbiol. Biotechnol. 94
(2012) 1461–1467.
[31] C.J. Lin, W.C. Tseng, T.H. Lin, S.M. Liu, W.S. Tzou, T.Y. Fang, Characterization of
a thermophilic l-rhamnose isomerase from Thermoanaerobacterium
saccharolyticum NTOU1, J. Agric. Food. Chem. 58 (2010) 10431–10436.
[32] D.W. Lee, E.A. Choe, S.B. Kim, S.H. Eom, Y.H. Hong, S.J. Lee, H.S. Lee, D.Y. Lee,
Y.R. Pyun, Distinct metal dependence for catalytic and structural functions in
the l-arabinose isomerases from the mesophilic Bacillus halodurans and the
thermophilic Geobacillus stearothermophilus, Arch. Biochem. Biophys. 434
(2005) 333–343.
[17] C.S. Park, S.J. Yeom, Y.R. Lim, Y.S. Kim, D.K. Oh, Substrate specificity of a
recombinant d-lyxose isomerase from Serratia proteamaculans that produces
d-lyxose and d-mannose, Lett. Appl. Microbiol. 51 (2010) 343–350.
[18] C.S. Park, H.J. Kwon, S.J. Yeom, D.K. Oh, Mannose production from fructose by
free and immobilized d-lyxose isomerases from Providencia stuartii,
Biotechnol. Lett. 32 (2010) 1305–1309.
[33] W. Zhang, D. Fang, Q. Xing, L. Zhou, B. Jiang, W. Mu, Characterization of a
novel metal-dependent D-psicose 3-epimerase from Clostridium scindens
35704, PLoS One 8 (2013) e62987.
[34] W. Mu, W. Zhang, D. Fang, L. Zhou, B. Jiang, T. Zhang, Characterization of a
D-psicose-producing enzyme D-psicose 3-epimerase, from Clostridium sp,
Biotechnol. Lett. 35 (2013) 1481–1486.
[19] R.L. Anderson, D.P. Allison, Purification and characterization of d-lyxose
isomerase, J. Biol. Chem. 240 (1965) 2367–2372.
[20] D.H. Patel, S.G. Wi, S.G. Lee, D.S. Lee, Y.H. Song, H.J. Bae, Substrate specificity of
the Bacillus licheniformis lyxose isomerase YdaE and its application in in vitro
catalysis for bioproduction of lyxose and glucose by two-step isomerization,
Appl. Environ. Microbiol. 77 (2011) 3343–3350.
[21] E.A. Cho, D.W. Lee, Y.H. Cha, S.J. Lee, H.C. Jung, J.G. Pan, Y.R. Pyun,
Characterization of a novel d-lyxose isomerase from Cohnella laevoribosii
RI-39 sp. nov, J. Bacteriol. 189 (2007) 1655–1663.
[22] J.G. Choi, S.H. Hong, Y.S. Kim, K.R. Kim, D.K. Oh, Characterization of a
recombinant thermostable d-lyxose isomerase from Dictyoglomus turgidum
that produces d-lyxose from d-xylulose, Biotechnol. Lett. 34 (2012)
1079–1085.
[23] L.M. van Staalduinen, C.S. Park, S.J. Yeom, M.A. Adams-Cioaba, D.K. Oh, Z. Jia,
Structure-based annotation of a novel sugar isomerase from the pathogenic
E. coli O157: H7, J. Mol. Biol. 401 (2010) 866–881.
[24] H.J. Kwon, S.J. Yeom, C.S. Park, D.K. Oh, Substrate specificity of a recombinant
d-lyxose isomerase from Providencia stuartii for monosaccharides, J. Biosci.
Bioeng. 110 (2010) 26–31.
[35] W. Mu, F. Chu, Q. Xing, S. Yu, L. Zhou, B. Jiang, Cloning, expression, and
characterization of a D-psicose 3-epimerase from Clostridium cellulolyticum
H10, J. Agric. Food Chem. 59 (2011) 7785–7792.
[36] D.K. Oh, Tagatose: properties, applications, and biotechnological processes,
Appl. Microbiol. Biotechnol. 76 (2007) 1–8.
[37] P. Kim, Current studies on biological tagatose production using l-arabinose
isomerase: a review and future perspective, Appl. Microbiol. Biotechnol. 65
(2004) 243–249.
[38] J. Keramat, H.E. Nursten, The relationship between the coloured compounds
present in the pressed liquor of cane sugar manufacture and those formed in
Maillard reactions, in alkaline degradation of sugars, and in caramelisation,
Food Chem. 51 (1994) 417–420.
[39] W. Zhang, H. Li, T. Zhang, B. Jiang, L. Zhou, W. Mu, Characterization of a
D-psicose 3-epimerase from Dorea sp. CAG317 with an acidic pH optimum
and a high specific activity, J. Mol. Catal. B: Enzyme 120 (2015) 68–74.
[40] C. Fan, W. Xu, T. Zhang, L. Zhou, B. Jiang, W. Mu, Engineering of
Alicyclobacillus hesperidum l-arabinose isomerase for improved catalytic
activity and reduced pH optimum using random and site-directed
mutagenesis, Appl. Biochem. Biotechnol. (2015) 1–13.
[25] W. Mu, X. Wang, Q. Xue, B. Jiang, T. Zhang, M. Miao, Characterization of a
thermostable glucose isomerase with an acidic pH optimum from
Acidothermus cellulolyticus, Food. Res. Int. 47 (2012) 364–367.
[26] C. Fan, K. Liu, T. Zhang, L. Zhou, D. Xue, B. Jiang, W. Mu, Biochemical
characterization of a thermostable l-arabinose isomerase from a
thermoacidophilic bacterium, Alicyclobacillus hesperidum URH17-3-68, J. Mol.
Catal. B: Enzyme 102 (2014) 120–126.
[41] Z. Xu, S. Li, X. Feng, J. Liang, H. Xu, l-arabinose isomerase and its use for
biotechnological production of rare sugars, Appl. Microbiol. Biotechnol. 98
(2014) 8869–8878.
[42] P. Gouet, E. Courcelle, D.I. Stuart, F. Metoz, ESPript: analysis of multiple
sequence alignments in PostScript, Bioinformatics 15 (1999) 305–308.
[27] Y.J. Lee, I.D. Wagner, M.E. Brice, V.V. Kevbrin, G.L. Mills, C.S. Romanek, J.
Wiegel, Thermosediminibacter oceani gen. nov. sp. nov. and
Thermosediminibacter litoriperuensis sp. nov., new anaerobic thermophilic
bacteria isolated from Peru Margin, Extremophiles 9 (2005) 375–383.
Please cite this article in press as: L. Yu, et al., Efficient biotransformation of d-fructose to d-mannose by a thermostable d-lyxose
isomerase from Thermosediminibacter oceani, Process Biochem (2016), http://dx.doi.org/10.1016/j.procbio.2016.08.023